key: cord-004879-pgyzluwp
authors: nan
title: Programmed cell death
date: 1994
journal: Experientia
DOI: 10.1007/bf02033112
sha: 
doc_id: 4879
cord_uid: pgyzluwp

nan

It is widely held that all developmental cell death is of a single type (apoptosis) and that neuronal death is primarily for adjusting the number of neurons in a population to the size of their target field through competition between equals for target-derived factors. We shall draw on our research and on that of others to criticize these views and replace them by the following.

At least three types of neuronal death occur, only one of which resembles apoptosis; a neuron can choose between several self-destruct mechanisms depending on the cause of its death.

The purpose of the death is to regulate connectivity, not neuron number. Competitors for trophic factors are unequal, and many losers have made axonal targeting errors.

A neuron's survival and differentiation depend on multiple anterograde and retrograde signals. Activity affects retrograde signals and some but not all anterograde ones. The pattern of activity is more important than the overall amount.

In rodents, the period of naturally occuring cell death of motoneurons is followed by a period of supersensitivity to axonal injury. Thus, in newborn rodents lesion of the facial nerve leads to a rapid degeneration of the injured motoneurons. We have tested whether overexpression, in rive, of the bcl-2 proto-oncogene was capable of preventing death of axotomized motoneurons. To address this question we used transgenic mice whose motoneurons overexpress the Bcl-2 protein. One of the two facial nerves of newborn mice was transected on the 2nd-3rd post-natal day. Seven days after the lesion, the morphology of the facial nuclei was analyzed. In control mice, and when compared to the intact nucleus, 70 to 80 % of axotomized motoneurons had disappeared. In contrast, in the transgenic animals, the number of motoneurons on the lesioned side remained unchanged when compared to the eontralateral nucleus. Furthermore, their axons remained visible up to the distal lesion site. These experiments show that, in rive, motoneurons overexpressing the Bcl-2 protein survive after axotomy, and suggest that, in rive, Bcl-2 protect neurons from experimentally induced cell death and could be a target for treatment of motoneurons degenerative diseases.

Messmer S., Mattenberger L., Sager Y., Blatter-Garin M-C., Pometta D., Kate A., James R.W. Drpt de Mrdeeine, Drpt. de Pharmacologie, Div. de Neurophysiologie Clinique, Facult6 de Mrdecine, Gen~ve.

Clusterin is a widely expressed glycoprotein, highly conserved across species. Numerous functions have been postulated for this protein. The most important are roles in lipid transport, as elusterin is associated with apolipoprotein AI in HDL, complement regulation and tissue remodelling, in particular during cell death and differentiation. Using cultures of rat spinal cord neurones (90% neurons and 5-10% non-neuronal cells), we have studied the expression of clusterin and ape E in glutamate-induced neuronal cell death to examine potential roles in lipid management. Up-regulation of the two proteins was observed. Clusterin and ape E appear in the conditioned medium respectively 15h and 7.5h after incubation with glutamate. Control studies, in the presence of a noncompetitive NMDA receptor agonist showed the secretion of clusterin and ape E to be diminished by >60%. No up-regulation of either protein was observed in complementary studies with exclusively non-neuronal cell cultures. The cellular origin of the 2 secreted proteins is presently under investigation. Programmed cell death and tissue remodelling are consequences of hormonally induced restructuring of the rat ventral prostate after castration and the rat mammary gland after weaning. We used the "Differential Display"-method (Liang and Pardee, 1992, Science 257:967) to detect and isolate eDNA fragments whose corresponding RNAs are regulated either coincidentally, or in an organ specific fashion during mammary gland involution and postcastrational prostate regression. Partial sequencing of 12 clones revealed high, but not absolute homology of 5 fragments with sequences, previously characterized in different biological contexts. These five encode functions which could be anticipated to be important for cell growth and/or programmed cell death, We are presently investigating the functions of several of these transcripts in cell culture and in rive. Antisense oligos are being employed in vivo to determine whether these genes contribute to the phenotype of programmed cell death. B epitopes derived from the envelope gp52 glycoprotein (EP3) or from the viral superantigen of MMTV have been incorporated into inert or live vaccines. The inert vaccine consists of purified chimeric proteins which contain the B epitopes alone or fused to multimeric promiscuous T helper epitopes from tetanus toxin. Mice were immunized subcutaneously with these chimeric proteins. The live vaccine consists of an avirulent strain of Salmonella typhimurium which expresses the MMTV epitopes in the form of chimeric proteins fused to the nucleocapsid protein of Hepatitis B Virus. This vaccine is given to mice in one oral dose. The level, duration and isotype of the immune response generated by each vaccine have been measured and compared. The level of protection has been investigated by systemically challenging immunized mice with the relzovims. A reduced binding of oxytocin (OT) occurs with aging in some, but not all, areas of the rat brain (Arsenijevic et al., Experientia 1993, 49, A75) . The candate putamen showed the most impressive loss of OT receptors. Two other regions, the hypothalamic ventromedial nucleus (VMH) and the islands of CaUeja (ICj) had also an important deficit of OT binding sites. On the other hand, these two regions were known to be sensitive to sex steroids. In the present work, we treated from 20 month old rats during one month with testosterone propionate (2 #g/kg s.c., once every 3 days) dissolved in oil. Three rats of the same age injected with oil only served as controls. We labelled OT receptors throughout the brain of old rats using a 125I-labelled ligand specific for OT receptors. Analysis of autoradiograms by an image analyzer revealed that the testosterone treatment increased OT binding sites in the VMH, in the ICj, and, to a lesser extent, in the bed nucleus of the stria terminalis, a region also sensitive to sex steroids, By contrast, in the caudate putamen, the disappearance of OT receptors was not compensated. In conclusion, the decrease of OT receptors occurring in VMH and ICj with aging can be reversed by administration of gonadal steroids. In contrast, the loss of OT receptors in the striatum appears to depend on another mecanism. Vasopressin (AVP) receptors are expressed transiently in the facial nucleus during development (Tribollet et el., 1991, Dev. Brain Res., 58, 13-24) . AVP may therefore play a role in the maturation of neuromuscular connexions in the neonate rat, and possibly in the restanration of these connexions after nerve lesion in the adult. In order to investigate the latter proposition, we have sectionned the facial nerve in adult rats and used quantitative autoradiography to look at AVP binding sites in the facial nucleus at various postoperative times. We observed a massive and transient increase of AVP binding sites on the operated side. The number of facial AVP binding sites reaches a maximum about one week after nerve section, remains stable during 2-3 weeks, then begin to decrease towards control level. The induction of AVP receptors is markedly delayed if the proximal stump of the nerve is ligated. To assess whether other motor nuclei would also react to axotomy by up-regulating the expression of AVP receptors, we have sectionned the hypoglossal nerve and the sciatic nerve. In both cases, the binding of AVP receptor ligand increases massively in the respective motor nuclei, with a time-course similar to that found in the facial nucleus. Altogether, our data suggest that central AVP could be involved in the process of nerve regeneration.

CYTOTOXIC T-CELL MEDIATED APOPTOSIS Schaerer,E, Karapetian,O.,Adrian,M. and Tschopp,J. Inst.de Biochimie, Univ.de Lausanne, 1066 Epalinges. An apoptotic cell death mechanism is used by cytolytic T cells (CTL) to lyse appropriate target cells. CTL harbor cytoplasmic storage compartments, containing the lytic protein perforin and serineproteases (granzymes), whose content is released upon target cell interaction. We show that these granules are multivesieular bodies and that degranulation releases these intragranular vesicles (IGV) having granzymes, T-cell receptor and yet undefined proteins associated. Isolated IGVs and perforin induce DNA breakdown in target cells within 20 minutes. Microscopic analysis demonstrates that IGV specifically interact with target cell via the T-cell receptor and that their contents is taken up by the target cell. Already 15 min. after interaction, 3 distinct IGV proteins are found in the nucleus of the target cell.One of the molecules has been identified to be granzyme A, previously reported to be involved in apoptosis. We propose that lymphocytes transfer apoptosisinducing proteins to the nucleus of the target cells using vesicles as vehicles for delivery.

Cytotoxic T cells kill their targets by a mechanism involving membranolysis and DNA degradation (apoptosis). Recently, two sets of proteins have been proposed as DNA breakdown-inducing molecules in T cells: granzyme A, B and TIA-I. In this study, we cloned and further characterized the TIA-I mouse homologue. Aa sequence comparison with the human TIA-1 showed an overall identity of 93%. Devoid of a signal peptide, TIA is yet localized to cytotoxic granules, probably targeted via a Gly-Tyr-motif. As TIA-I, its mouse homolcgue contains three RNAbinding domains.

Expression of TIA during development shows a very strong signal in the brain and weaker signals in thymus, heart and other organs.

During embryonic development several structures that contribute to organogenesis form transiently and are later eliminated by apoptosis. This pattern of TIA expression could indicate its involvement in apoptosis. Prostate involution occurs after castration in rats and is associated with the death by apoptosis of a large fraction of the epithelial cells. We have isolated several genes from a prostate involution bacteriophage lambda library using differential screening methods. Among these clones, one d~monstrated an especially strong signal when used as a probe against Northern blots of prostate mlhNA obtained before, and at different times after castration. This gene is down-regulated after castration by 40-fold within 5 days. Intramuscular injection of a testosterone depot resulted in complete restoration of expression within 24 hours. Upon sequencing it became apparent that this clone has a high degree of homology to a known NDAH dehydrogenase encoded in mitochondrial DNA. The clone failed to hybridize to any transcripts from rat organs other than prostate. We are now in the process of isolating the htm~n hc~olog to this gene for use as a biomarker in study of benign hyperplasia and developing carcinoma. This gene is a possible indicator for testosterone-independent cell populations or of cells lacking ftl~ctional testosterone receptor. During the first three postnatal weeks the rat lung undergoes the last two developmental stages, the phase of alveolarization and the phase of microvascular maturation. The latter involves a decrease of the connective tissue mass in the alveolar septa and a merging of the two capillary layers to a single one.

Speculating that programmed cell death may play a role during this remodeling, we searched for the presence of apoptotie cells in rat lungs between days 10 and 24. Lung paraffin sections were treated with Y-terminal transferase, digoxigenin-dUTP, and anti-digoxigeninfluorescein-F(ab)-fragments, and the number of fluorescent nuclei was compared between sections at different days. While the number of apoptotie ceils was low until the end of the second week and at day 24, we observed an about eight fold increase of fluorescent nuclei towards the end of the third week. We conclude that programmed cell death is involved in the structural maturation of the lung.

Brunner, A., Wallrapp, Ch., Pollack, I, Twardzik, T. and Schneuwly, S. Lehrstuhl Genetik, Biozentrum Universit~t W~rzburg, Mutants in the giant lens (g/l) gene show a strong disturbance in ommatidial development. In the absence of any gene product, additional phetoreceptors, cone cells and pigment cells develop. Opposite effects can be seen in flies in which the gene product of the giant lens gene can be ectopically expressed by heat shock. A second very typical phenotype is the disturbance of photoreceptor axon guidance. Molecular analysis of gil shows that it encodes a secreted protein of 444aa containing three evolutionary conserved cystein-motives very similar to EGF-like repeats. We propose that gil functions as a secreted signal, most likely a lateral inhibitor for the development of specific cell fates and that gil, either directly or indirectly, is involved in targeting photoreceptor axons into the brain.

THE DECREASE IN CELLULARITY DURING SCAR ESTABLISHMENT IS MEDIATED THROUGH APOPTOSIS DESMOULIERE, A., REDARD, M., DARBY, I., AND G. GABBIANI Department of Pathology, CMU, 1 rue Michel Server, 1211 Gen~ve 4 Dudng the healing of an open wound, granulation tissue formation is characterized by replication and accumulation of fibroblastic cells, many of which acquire morphological and biochemical features of smooth muscle cells and have been named myofibroblasts (Sch0rch et el., Histology for Pathologists, t992). As the wound evolves into a scar, there is an important decrease in ceUuladty, including disappearance of myofibroblasts. The question adses as to which process is responsible for myofibroblast disappearance. During a previous investigation on the expression of (z-smooth muscle actin in myofibroblasts, we have obsewed that in late phases of wound healing, many of myofibroblasts show signs of apoptosis end suggested that this type of cell death is responsible for the disappearance of myofibroblasts (Darby et al., Lab. Invest. 63:21, 1990) . We have tested this hypothesis by means of electron microscopy and morphometry and by in situ end-labeling of fragmented DNA (Wijsman et al., J. Histochem. Cytochem. 41:7, t 993) . Our results show that the number of apoptotic cells increases as the wound closes and suggest that this may be the mechanism for the disappearance of myofibroblasts as well as for the evolution of granulation tissue into a scar. (Supported by the Swiss National Science Foundation, Grant n~ S01-16

R. Jaggl, A. Marti and B. Jehn. Universit~t Bern, AKEF, Tiefenaustr. 120, 3004 Bern At weaning the mammary gland undergoes a reductive remodelling process (involution) which is associated with the cessation of milk protein gene expression and apoptosis of milk-produclng epithelial cells. This process can be reversed by returning the pups to the mother within 1 day. Elevated nuclear protein kinase A (PKA) activity was observed from one day post-lactation, paralleled by increased c-los, junB, ]unD and to a lesser extent c-]un mRNA levels. AP-1 DNA binding activity was transiently induced and the AP-1 complex was shown to consist principally of cFos/JunD. Oct-1 DNA binding activity and Oct-1 protein were gradually lost from the gland over the first four days of involution, whereas Oct-1 m_RNA levels remained unchanged. Comparing nuclear extracts from normal mammary glands with nuclear extracts from glands which had been cleared of all epithelial cells three weeks after birth revealed that PKA activation, AP-1 induction and Oct-1 inactivation are all dependent on the presence of the epithelial compartment. The increased Fos/Jtm expression and the inactivation of Oct-1 may be consequences of the increased PKA activity. When involution is reversed, both, PICA activity and AP-1 DNA binding activity (and fos andjun mRNA levels) are reduced to basal levels. Our data suggests a role for PKA and AP-1 on progranlmed cell death of manlnmry epithelial ceils.

Bcl-2~ does not require membrane attachment for its survival activity C. Borner*, I. Martinout, C. Mattmann*, M. Irmler*, E. Sch&rrer*, J.-C. Martinou-j-, and J. Tschopp*. * Institute of Biochemistry, University of Lausanne, 1066 Epalinges, 1 Institute of Molecular Biology, Glaxo Inc.,1228 Plan los Ouates.

8cl-2(z is a mitochondrial or perinuclear-associated oncoprotein that prolongs the life span of a variety of cell types by interfering with programmed cell death. How it exerts this activity is unknown but it is believed that membrane attachment is required. To identify critical regions in bcl-2o~ for subcellular localization and survival activity, we created by site-directed mutagenesis, various mutations in regions which are most conserved between the different bcl-2 species. We show here that membrane attachment is not required for the survival activity of bcl-2o< A truncation mutant of bcl-2(z lacking the last 33 amino acids (T3) including the hydrophobic domain is soluble, yet fully active in blocking apoptosis of sympathetic neurons induced by NGF deprivation or L929 fibroblasts induced by TNFc~ treatment. We further provide evidence for a putative functional region in bcl-2 which lies in the conserved domains 4 and 5 upstream of the hydrophobic COOH terminal tail. The breakdown of nuclear DNA is considered to be a hallmark of apoptosis. We previously identified the perinuclear membrane localized DNase I as the endonuclease involved in the formation of oligonucleosomal-sized fragments (DNA ladder). It is not clear how the nuclease is activated and has access to the DNA. We show that in thymocytes induced to undergo apoptosis, lamin breakdown preceded DNA laddering.

By transfeeting HeLa cells with a constitutively active cdc2 mutant, nuclear envelope breakdown and typical apoptotic features (ehromatin condensation) were observed. Moreover, co-transfection with cdc2 mutant and DNase I led to DNA degradation. We propose that apoptosis can be induced by wrongly timed and hence abortive mitosis leading to uncontrolled nuclear membrane disintegration. S02-01 S02-04

Platelet-derived growth factor (PDGF) is thought to play an active role in fibrosing diseases. Bronchiolitis obliterans-organizing pneumonia (BOOP) is a condition characterized by intraluminal proliferation of connective tissue inside distal air spaces. To evaluate PDGF expression in BOOP we performed immunohistoehemistry on lung biopsies from 20 patients and 10 controls free of fibrosis. Sedal sections were stained with an antibody against either PDGF or the monoeyte/macrophage marker CD68, In both groups the PDGF ~9 cells were essentially tissue macrophages. Using point counting to measure volume fraction (Vv) , PDGF-pesitive cells represented 4.65+1.63% (mean+SD) of the volume occupied by lung tissue in the BOOP cases, and 2,12+0.65% in the controls (!0<0,001). Similarily, 10.73+4.69% of the lung tissue was occupied by CD68 E~ macrophages in the BOOP cases, compared to 5.37:~3.73% in the controls (p<O.005). A positive linear correlation was found between the V V of PDGF ~ cells and the V V of CD68@ cells on the 30 cases (r=g.50 p<0.0O5). We conclude that BOOP is characterized by a recruitment of tissue macrophages and that a significant proportion of them express PDGF. As POGF can also induce the macrophage chemo-attractant MCP-1, we speculate that maerophages are involved in a positive feed back regulation, and may play a pivotal role in the pathogenesis of BOOP.

BARRIER LESIONS IN HYDROSTATIC PULMONARY EDEMA D.Wu, H.Bachofen, U.Gyger and E.R. Weibel Department of Anatomy, University of Bern, CH 3012, Bern Pulmonary edema induced in isolated perfused rabbit lungs by moderate elevation of perfusion pressure (14 tort ) results in the formation of characteristic bleb-like extensions of the thin cytoplasmic leaflet of alveolar epithelium ( Bachofen et al. Am Rev Resp Dis. Vo1.147. pp989-996,1993) . In this study, we show by means of electron microscopy and morphometry that the frequency of the blebs correlate with the distribution of alveolar edema fluid. We find that 5% of the epithelial surface is involved in the formation of blebs. Three-dimensional reconstruction showed that blebs had near-spherical shape and that their opening toward the interstitium is irregular. These finding demonstrate a great plasticity of epithelial cell extensions when subjected to moderate pressure stress. At high pressure the cell linings demonstrate clear breaks. Two major roles have been defined for NO: cellcell communication mediated by the stimulation of cGMP synthesis, and cytotoxicity by direct interaction of the free radical NO with cellular target. To investigate these effects in human epithelia, which constitute a major source of NO in man, we introduced the murine iNOS sequence into SV40-T immortalized human bronchial epithelial cells (Beas-2B). iNOS activity was higher than 100 pmol citrulline/min/ mg prot in the first passages of iNOS transfected cells, but it decreased by subculturing. No iNOS activity was detected in cells transfected with control vector. In iNOS transfected cells, NO stimulated a cGMP synthesis and c-los expression which could be inhibited by addition of LNMA. Thus, in human epithelial cells NO is able to significantly alter signal transduction pathway. Partial pneumoneetomy, i.e. the resection of lung tissue, is known to initiate compensatory growth in the remaining lobes. Under appropriate conditions this results in a complete restoration of lung parenchymal structure and function. Unfortunately the molecular mechanisms controlling the onset of this compensatory growth are poorly understood. The aim of our study was to find and eharacterise genes which are up or down regulated during this time.

For isolating such genes, we chose the differential display approach (P.Liang and B.Pardee 1992, Science 257:p 967). Therein a 3'end fragment of a subset of reverse transcribed mRNAs is amplified by the polymerase chain reaction. The fragments can be separated on a DNA sequencing gel. Genes differentially expressed in pneumonectomised and sham-treated rats show different patterns on these gels. The fragments of such genes can be isolated from the gel and cloned into vectors in order to provide tools to characterise their gene. A set of such genes, some up-regulated and others downregulated following pneumonectomy, have been isolated, partially sequenced and characterised.

time periods. Our new plasma marker consisting of highly concentrated gold nanospheres was infused via a femoral vein catheter into the right atrium of anaesthetised and ventilated rabbits.

Five or 10 seconds after start of infusion, the blood flow was stopped and the lung was fixed by instillation. Silver enhanced paraffin sections revealed labeled and unlabeled blood vessels. Electron-microscopic sections showed various plasma gold particle concentration. Morphometric analysis of electron micrographs showed an increase of the portion of marked capillaries from shorter to longer circulation times. Based on our results, we conclude the existance of inhomogeneous plasma flow velocities within the pulmonary microvascular bed.

(Supported by SNSF grant 31-30946.91) In Schizosaccharomyces pombe pairing of meiotic chromosomes is not accompanied by the formation of a tripartite synaptonemal complex (SC), a structure which connects the homologs in most other eukaryotes. Meiotic nuclei revealed novel structures ("linear elements") that might be related to the axial elements ($C precursors) of other organisms and function in meiotic chromosome pairing, recombination, and segregation (BAler et al. (1993) JCB 121:241). To get further insight into the behavlour of meiotic chromosomes, we performed fluorescence in situ hybridizations on spread nuclei with probes from different chromosomal regions. Interestingly, all centromeres and telomeres become clustered during meiotic prophase. The centl:omeres are paired already before meiosis, whereas the telomeres initiate pairing specifically during meiotic prophase, followed by pairing of interstitial regions. Painting of whole chromosomes revealed that the homelogs occupy together a specific territory in the nucleus. An expression vector containing the VAI gene read by RNA polymerase Ill, injected into Xenopus zygotes, yields high levels of RNA in virtually all embryonic cells. Anti sense FINA is produced from a segment of the XhoxlA gene inserted in opposite orientation into the VAI gene. Immunological staining of whole mount embryos shows that target mRNA level and XhoxlA protein expression of XhoxlA protein is diminished in about half of the antisense treated individuals. Accordingly, nearly half of the embryos develop typical malformations in regions where the XiToxlA gene is normally expressed. Somites el the anterior trunk are heavily disorganized and mesodermal denvalives surrounding the intestinal tract are reduced or missing. Antisense inhibition, through production of RNA in situ from expression vectors, thus represents a useful tool to study the functional role of zygotic genes in Xenopus development.

Cloning and characterization of the mouse homolog to S.cerevisiae SEP1 encoding a DNA strand exchange and homologous pairing protein Vladimir Bashkirov and Wolf-Dietrich Heyer. Institute of General Microbiology, University of Bern, Baltzer-Strasse 4, 3012 Bern.

Homologous pairing and DNA strand exchange are the central steps postulated in the current models of genetic recombination. To date no mammalian gene encoding a homologous pairing protein has been cloned. To get insight into the mechanism of recombination in higher eukaryotes we were interested to clone such genes. Protein sequence comparison of the S.cerevisiae homologous pairing protein, SEPI, and its homolog from S.pombe, p140 ex~ revealed several conserved stretches of amino acids in their NH2-terminal regions. Using fully degenerate primers and mouse testes eDNA as a template for the polymerase-chain-reaction a product with the expected length has been amplified. This 360 bp DNA was shown to encode a putative polypeptide with high degree of homology to both proteins from the two yeasts. The mouse PCR-product was used as a hybridization probe for screening a 1 gtlO mouse testis cDNA library. After several rounds of successive screening a total of about 4 kb of mouse eDNA (mSEP1) was cloned and sequenced. The putative translation product revealed high homology throughout the entire sequence to its S.pombe and S.cerevisiae counterparts, 38 % and 41% of amino acid identity, respectively. The mSEP1 sequence was shown to be unique in the genome and the chromosomal gone is likely to contain many introns. We want to establish a functional full length eDNA clone of mSEPI and characterize the gone and protein product. We have earlier developed a cell free system to study recombinational repair of DNA double strand breaks and deletions. The method allowed us to purify and characterize a high molecular weight multiprotein complex (RC-1), which catalyzes DNA recombination. A DNA polymerase and DNA ligase as well as other enzymatic activities copurify with RC-1.We have commenced an investigation of homologous DNA recombination in nuclear extracts which were prepared fzom normal and mutated mouse thymocytes. The scid mutation in mice causes an aberrant V(D)J joining reaction, an elevated sensitivity to ionizing radiation and a defect in recombinational repair. Moreover, the normal mouse thymocytes were sorted into the CD4/CD4 double negative (DN), double positive (DP), and single positive (SP) subclasses and their nuclear extracts, and extracts from RAG-2 -/-mice, were tested. Only the scid mice and the CD4/CD8 SP thymocyte extracts were inactive in the recombinational repair reaction. This indicates a development stage specific activity and a scid specific defect of recombinational repair in thymocytes. Restoration of the recombination activity was observed in thymocyte extracts derived from scid, which express a Wansgene for the T-cell receptor I~ chain. A protein could be purified from normal thymus cells to near homogeneity which specifically rescues the recombination activity in scid extracts. This protein, SRSP ~cid Recombination Stimulatory ~otein), migrates as a single band of app. 7 2 kDa in SDS polyacrylamide gel electrophoresis. The ade6-M26 mutation in the ade6 gene of the fission yeast Schizosaecharomyces pombe gives rise to a hot spot for meiotic homologous recombination. To address whether transcription plays a role in M26 activity, the effect of the deletion of the ade6 promoter in combination with either the M26 and M375 mutations on recombination frequency was studied at the tetrad level. Deletion in eis of the promoter abolishes M26 hot spot activity relative to the M375 control deletion strain. It is possible that deletion of the promoter has eliminated a sequence necessary for M26 hot spot activity, rather than lack of transcription being responsible for the loss of M26 activity.

The role of transcription in M26 hot spot activity is currently being examined. We are analysing meiosis in a temperature sensitive top2 mutant. Complementary temperature-shift experiments showed that topoisomerase II is required for both meiotic divisions; early meiotic events are not impaired. In a meiotic time course, DAPI staining revealed that the cells are blocked before the first meiotic division. Interestingly, in meiotic spreads we found that the spindle pole body (SPB) cycle is not affected by the block: the final arrest point showed 1 nucleus with 4 completely separated SPBS. We want to analyse the number of SPBs and spindle formation with immunofluorescence. We are planning to analyse the course of meiosis in a top2 rec7 double mutant (rec7 is recombination deficient) to see if topoisomerase II activity is necessary for separation of recombined chromosomes. Analysis of meiosis in a top2 rec8 double mutant (rec8 is recombination deficient and does not form linear elements) could give us a clue about the function of the linear elements.

Genetic recombination is able to induce cancer and to mediate its further progression. In cells from cancer predisposed individuals, mitotic recombination can lead to loss of heterozygous tumor suppressor genes as is e.g. observed in the hereditary form of retinoblastoma. One mechanism for gene loss is inter-or intrachromosomal recombination involving short duplicated sequences. Our aim is to identify xenobiotics that are able to induce mitotic recombination and to elucidate the involved molecular mechanisms. For that purpose we have constructed transgenic D.

melanogaster (1). rn.). cell lines. Cells were stably transfeeted with an extrachromosomal DNA element providing a putative recombination substrate+ "l%e constructed element contained two ta'uncated fragments of the neomycin resistance gene (nee R) interrupted by a hygremycin resistance marker (hygR). The hyg R insert was flanked on both sides by identical 352 bp fragments of nee R, Different D. m. promoters shown to be constitutively active in various D. m. tissues were inserted 5' to both E. coil genes. Restoration of a functional nee R gone may occur by deletion of the hyg R after recombination between the homologous regions. Stably transfected Schneider line 2 cells were obtained and preliminary results concerning chemical induction of genetic rearrangements will be shown.

Fleck, 0., Sch~r, P. and Kohli, J., Institute of General Microbiology, Baltzerstr. 4, 3012 Bern

To detect mismatch-binding proteins of S. pombe we performed band shift assays with radiolabeled oligonucleotides containing a defined mismatch. We found two distinct mismatch-binding activities.

One shows high affinity to most of the mismatches, but is absent for C/C. This protein might belong to the general mismatch-repair system, which is thought to be related to the bacterial MutS dependent pathway. The second activity preferentially binds to those mismatches containing a cytosine and therefore represents a novel type of mismatch-binding protein.

Baur M., SCh~r P. and Kohli J., Institute of General Microbiology, Baltzerstrasse 4, CH-3012 Bern

Homologs of the Salmonella typhimurium (mutL, S and 59 and Streptococcus pneumoniae (hexA, B) genes involved in mismatch repair have been identified in several distantly related organisms. So far no mismatch repair gone is known in Schizosaccharomyces pombe. In order to get more insight into the ~chanisaLs of mismatch repair in S. pombe we started the cloning of a mutL homolog. Degenerated oligonucleotide primers based on conserved regions of mutL homologs were used in the polymerase chain reaction (FCR) to amplify a corresponding DNA fragment from S. pombe. A DNA fragment of about 220 bp was amplified whose deduced amino acid sequence shared a high degree of homology with PAiql of Saccharomyces cerevisiae and mutL, hexB. With the help of this DNA fragment the gone was isolated from a genomic DNA library by colony hybridization. Sequence analysis of this mlhl gene (mlh = mut~ homolog) shows an ORF of the expected size with three putative introns. While the deduced amino acid sequence of the 5' region shares strong homology with the procaryotic genes mutL, hexB and the eucaryotic gene PMSI, the amino acid sequenoe of the 3' region shows homology only to PMSI and hexB. The central region of the protein seems to be conserved only weakly.

Parisi S., and Kohli J., Institute of General Microbiology, University of Bern, CH-3012 Bern

Rec7 and rec8 are supposed to be genes with major functions in meiotic recombination since mutations in these genes reduce the frequency of meiotic intragenic recembination ~1000 fold. These genes have been isolated and sequenced but no enzymological activity could be assigned to the protein. A first step toward enzymological and biochemical characterization of rec7 and rec8 is to make beth protein products available in sufficient quantities and in reasonable pure form. For this purpose both genes will be cloned in a S. /x~nbe over-expression system using the strong and inducible nmtl promoter.In parallel both genes will also be expressed in E. coli as GST fusion proteins in order to produce polyclonal antibodies in rats and rabbits. These antibodies can then be used for protein purification in S. pombe as well as for immunofluorescence experiments. Induction of expression in E. coli produces a band of the expected size of 56kDa for Rec7 and of 73kDa for Rec8. To remove the GST carrier, the Rec7 and Rec8 fusion proteins can successfully be cleaved by thrombin or factor Xa respectively. Production of antibodies is in progress.

D6bbeling, U., nobi, R,, Berchtold, M.W. and Kuenzle, C.C. Institut fdr Veterinarbiochemie, Universit6t ZUrich, Wintertharerstrasse 190, Normally, only pre B-and pre T-cells can rearrange their immunoglobulin genes by V(D)J recombination, but when other cell types are transfected with the rag-i and rag-2 genes they also become able to perform V(D)J recombination. By transiently transfeeting NIH 3T3 fibroblasts with both intact rag genes we found da recombination rate of 0.06%. No recombination was detected, when we transfected the intact rag-1 gene with a mutant rag-2 gene, which lacks 89 C-terminal amino acids, or the intact rag-2 gene with a rag-I gene containing a deletion in the topoisomerase IZ homology region. These experiments show that these two regions of the rag genes are essential for V(D)J recombination. When the L4 fibroblast cell line, which has been rendered recombination competent by stable transfection with the rag genes was overtransfected with both wild type rag genes, we found in contrary to an expected increase of V(D)J recombination a reduction by a factor of 2-2.5. This decrease was not observed with the mutant rag genes.This result indicates, that there exists an optimal intracellular RAG concentration for efficient V(D)J recombination.

M. Fischer, A. Sailer, H. Blieler, T. Riilicke*, A. Aguzzi +, P. Autenried and C. Weissmann; 1nstitut fiir Molekularbiologie L Universitfit Ziirich, H6nggerberg, 8093 Ziirich, *Biologisches Zentrallabor and + lnstitut fiir Neuropathologie , Universitdtsspital Ziirich, 8091 Zfirich, Switzerland The infectious agent of transmissible spongiform encephalopathies such as Creutzfeldt-Jakob disease in man or scruple in sheep is thought to be prpSc, a posttranslationally modified form of the normal host protein prpC We have shown that mice devoid of PrPC (Prn-pOlO) are completely resistant to scrapie. Here we report that also heterozygous Prn-p o/+ mice with reduced levels of PrP C show enhanced resistance to scrapie, as manifested by a long delay in the onset and slow progression of clinical disease. Conversely, Prn-pO/O animals overexpressing PrP transgeues exhibit shortened incubation times and accelerated progression of the disease. Propagation of the scruple agent is completely abolished in Prnpolo animals. High titers of infectivity, about 108 logLD50 units/g of brain, are detected in Prn-pO/+ mice 33 weeks after inoculation (the earliest time point measured), 24 weeks or more before the animals die of scrapie. In contrast, wildtype animals survive no more than 16 weeks after reaching similar titers of infectivity. A practical implication of our findings is that moderate inhibition of prpC synthesis could mitigate disease progression in spongiform encephalopathies. The RuvB protein has been shown, biochemically and genetically, to be involved in the process of homologous recombination in E. coli. One of its functions is to promote branch migration within Holliday junctions formed during the process of recombination. To investigate the mechanism by which RuvB promotes branch migration we have used conventional electronmicroscopy, cryo-electron microscopy, scanning transmission electronmicroscopy, and image analysis to study the interaction of RuvB protein with DNA. We observed that RuvB forms multimeric rings on dsDNA which encircle the DNA. We also observed that such rings tend to localize at the Holliday junctions. We propose that RuvB rings can actively move along DNA using the energy of ATP hydrolysis.This movement continues upon encountering a Holliday junction, thus forcing branch migration along the DNA. Gschwind M. and Huber G., Pharma Division, Preclinical Research, In some families with early onset AIzheimer's disease mutations on the 8-APP gene have been identified which cosegregate with the disease. So far all these pathogenic mutations have been identified on exons 16 and 17 in humans framing the sequence of the 8-amyloid itself.A genomic clone was isolated from mouse strain C57 Black/6 containing the coding regions of the last three exons (16) (17) (18) according to the human sequence. Sequencing of all three exons including the large 3' non coding region showed a 100% homology to the previously described Balb/c mouse cDNA. Exon-intron boundaries were determined at the same positions as in human and rabbit genome and the flanking regions in the introns are also very similar, Not only is the homology between the human and mouse f$-APP very high (97.6%) but also the genomic organisation of their genes is nearly identical. Therefore homologous recombination can be applied to introduce molecular 8-APP changes and to study their functional consequences on the protein level. We systematically investigated the ability of RecA protein to push the DNA strand exchange reaction through heterologous regions. We observed that RecA can push the strand exchange through substantial heteroiogy (up to 150 bp) when the heterologous insert was placed on the so called proximal end of the linear duplex. In the case of the medial beterology the strand exchange reaction could overcome heterologies of up to about 50 bp. Heterology at the distal end was most difficult for RecA to resolve, and already 22 bp long insert was completely blocking the progress of the strand exchange. These results, together with earlier electron microscopy studies, allow to propose a molecular mechanism by which RecA can exchange strands between partially heterohigous DNA molecules.

Institut for Veterin~r-Virologie, Universit~t Bern, CH-3012 Bern Bovine viral diarrhae virus is a positive-stranded RNA virus of the pestivirus group. Two biotypes, cytopathic and non-cytopathic in cell culture, can be distinguished. All cytopathic strains characterized so far carry insertions in their genomes. They probably result from a strand-switching activity of the viral RNA polymerase during viral replication. To study the switching capacity of the viral enzyme, RNA isolated from cells infected with two different strains of bovine viral diarrhea virus were analysed for homologous recombination in an RT-PCR assay. Performing the assay we found that fragmented RNA present during reverse transcription, the reverse transcriptase enzyme itself and viral RNA present in the PCR reaction result in the production of artificial recombination during the assay. It was not possible to detect a homologous recombination activity of the viral polymerase above the background recombination frequency of the RT-PCR assay. Our results strongly question the use of the RT-PCR assay for the study of recombination of RNA viruses. Transpositional rearrangements mediated by insertion sequence (IS) elements represent an important source of genetic variation within populations 1. We analyze DNA rearrangements at the level of the complete genome by RFLP, using 7 IS elements as probes for hybridisation to Southern blots Samples taken at different time points from 12 independent cuttures of E coli B grown by serial transfer for 10'000 generations 2 and stored at -70~ are used to study structure and evolution of genetic diversity within populations. The seven IS elements contribute differently to genetic variation and lead to a succession of mutants affecting the genetic structure of the population. We now search for correlations between RFLP patterns and the increase of relative fitness over time observed previously 2 Naas T., 81o~ M, Fitch W.M and Arber W, (1993) The complete DNA sequence of the Locusta migratoria mitochondrial genome has been derived from four cloned mtDNA fragments. The sequence is 15.2 kb in length, contains 13 peptide coding sequences, and genes for 22 tRNAs and two rRNAs.

The organisation of the genome shows a strong resemblance to Drosophila, the gene order and orientation being the same except for the positions two tRNAs. This contrasts with the only other non-dipteran insect mtDNA to have been sequenced, Apls mellifera, which shows differences in the positions of ii tRNAs. This finding is discussed in relation to the AT content of the different genomes. The sequences of the three mtDNA genomes are also compared directly and the results related to existing knowledge concerning the evolution of insects. Udem, New Jersey Medical School, Newark, NJ 07103-2714, USA Nonsegmented negative strand RNA viruses are so far not amenable to reverse genetics. The genomes must associate with nucleocapsid (N), phosphoprotein and polymerase proteins to form a functional RNP. As a step towards reverse genetics of measles virus (MV), plasmid vectors were constructed allowing synthesis of RNAs corresponding precisely to a MV genome in which all coding and intercistronic regions are replaced by the chloram-phenicol acetyl transferase (CAT) coding region, Transfection of these negative polarity transcripts into MV infected HeLa or 293 cell lines gave rise to CAT activity which could be serially transferred and massively amplified together with progeny helper virus in cell cultures. Both expected MV/CAT chimeric mRNA and replication product were identified in the progeny. Preliminary experiments indicate that only RNAs in which the total number of nucleotides is a multiple of six replicate and transcribe efficiently, consistent with the data on natural and modified copyback DI RNA of Sendal virus (Calain and Roux, J. Virol. 67 4822 (1993) ). Precise end to end encapsidation of RNA, each N protein contacting six nucleotides, might be required. The embryonic tissue specificity of the human cytomegalovirus immediate-early (hCMV-IE) promoter/enhancer (-522 to +13) was examined by l~-galactosidase staining of transgenic mouse embryos carrying hCMV-IE regulated lacZ genes. Embryos of three transgemc lines were analyzed between 8.5 -13.5 dpc. For all lines, hCMV-IE transcription was primarily confined to subsets of neural and vascular cells. However, extents of transgene expression along the embryonic primary axis were slightly and greatly restricted in two of the lines relative to the third line at all stages analyzed. These differences most likely represent integration site-dependent chromatin constraints that affect levels of hCMV-IE-directed lxanscripfion. Together, the three lines can be taken to define a graded potential for hCMV-IE transcription along the anteriorposterior axis of the embryo with greatest permissivity in mid-axial structures which include the dorsal neural tube at the level of the hindbraln and upper cervical spinal cord, the inner ear and vestibular acoustic ganglia, branchial arteries, aortic sac, and lung buds. The cell typespecificity and axially graded permissivity for hCMV-IE transcription in the mouse embryo provide a basis for understanding the pathology and relative frequencies of different types of birth defects caused by congenital hCMV infection in humans.

Stauffer, Y., 0fford, E. and Beard, P., Swiss Institute for Experimental Cancer Research (ISREC), CH-I066 Epalinges, Switzerland.

HPVI8 is associated with genital carcinoma. Like other human papillomaviruses, HPVI8 cannot be easily propagated in cell culture, because the viral growth cycle is dependent on the differentiation of its host cell, the keratinocyte. Moreover, very little viral capsid protein is found in vivo in infected tissues. The HPVI8 L1 and L2 genes code for the major and minor capsid proteins, respectively. To obtain the viral capsid proteins we made constructs with the L1 and L2 genes for expression in E. coli (the pQE system) and in recombinant vaccinie virus (a modification of the pHGS-I system) infected cells. The L1 and L2 proteins expressed in E. cell contained a histidine tag encoded by the vector. This allowed us to purify the proteins on a sickelaffinity column in order to raise antibodies. These will permit us to detect the presence of capsid proteins expressed from recombinant vaccinia viruses. We expect the L1 and L2 proteins expressed together from the vaccinia-based vector to selfassemble to form virus-like particles. to papain-like proteinases.

The role of these putative catalytic residues in the activity of the proteinase was studied by site-directed mutagenesis.

A minimal proteinase has also been constructed by N-and C-terminal deletion to determine the boundaries of the proteolytic domain, Sequence analysis of the C-terminal cleavage product should indicate the site of cleavage.

CELL ENTRY, UNCOAT~NG AND NUCLEAR TRANSPORT OF ADENOVIRUS Z .

U. F. Greber and A. Helenius. Yale University School of Medicine, New Haven, CT, U.S.A.

To enter cells, viruses must achieve three major goals: transfer their genome across the plasma membrane into the Cytosol, target the genome to the replication site and decondense it for replication. Adenovirus, an icosahedral nonenveloped particle with a double stranded genomic DNA and ii to 15 structural proteins~ enters into human KB cells by a sequential disassembly program during receptor-mediated endocytosi~, acid-dependent penetration into the cytosol, and targeting to the nucleus. During internalization, the fibers, primarily responsible for virus binding to the cell surface, are detached. In early endosomes, the capsid-stabilizing proteins Ilia and VIII are released. Fenton base, a vertexassociated protein implicated in penetration across the endosomal membrane, is removed in endosomes, if penetration is blocked with lysosomotroplc reagents. In the absence of inhibitors, a large fraction of penton base stays associated with the cytoplasmic capsid. The viral DNA is disconnected from the shell after proteolysis of the internal protein VI~ a capsid and DNA binding component. Removal of protein IX, a capsid binding factor, further destabilizes the coat, and the remaining 'stripped-down' virus particles associate with nuclear pore complexes to release their DNA-genomes into the nucleoplasm. This stepwise dissociation program is thought to provide the high efficiencies of the entry process needed to successfully deliver the viral DNA across multiple barriers to the cell nucleus, the site of replication.

Ch. Schmidt and W. Arber; Biozentrum der Universit~t Basel, Abt. Mikrobiologie, Klingelbergstr. 70, CH-4056 Basel Bacteriophage P1, carried as a single copy plasrnid in E. col~, is widely known for its horizontal gene transfer ability (transduction) in Enterobscteria. As a temperate phage, it can lysogenize its host and thereby provides it with accessory functions (restriction of foreign DNA by EcoP1, functions expressed by transposons carried on the P1 genome or integrative suppression). These properties reflect both symbiotic and evolutionary functions of PI. P1 regulates expression from its genes by two different mechanisms: trans(:ription of early regulatory genes is repressed by the phage encoded repressors C1 and Bof, whereas transcription of its morphogenetic genes from P1 specific late promoters is induced by the early expressed gpl0. By site-directed mutagenesis the late promoter PS, composed of the -22 inverted repeat AAGTTACTT and the -10 box, was functionally characterized. P5 and its derivatives compare well with several other late promoters of Pt with regard to their sequence dependent strengths: the strong promoters PS, Pdar and LPr2b all share the perfect inverted repeat around position -22, whereas the weaker promoters LPr91, LPr82a, LPr82b and LPr83 contain mismatches in their inverted repeats or the inverted repeat is positioned mere upstream relative to the transcription initiation site. homologies. However, the phages formed two immunity groups. The results may suggest a common genetic trait providing a selective advantage to lysogenic Aa.

By using a DNA substrate with defined gap size we found that human immunodeficiency virus type 1 reverse transcriptase (HIV~RT) was able to perform strand displacement DNA synthesis. This activity was not affected first by calf thymus proliferating cell nuclear antigen and replication factor C and second by Escherichia coli single-stranded DNA binding protein~ which together allow DNA polymerase ~ to perform strand displacement DNA syrlthesis (Podust , V. and H~bscher, U. (1993) Nucl. Acids Res. 21, . 3'-azido-2',3'-dideoxy-thymidine tripbosphate inhibited displacement completely indicating that DNA synthesis is required for this reaction. The HIV-RT p66 polypeptide alone could perform limited strand displacement DNA synthesis, whereas the HIV-RT p51 polyDeptide was completely inactive, likely due to its inability to replicate extensively on a MI3 DNA template, on the other hand the HIV-RT pSl polypeptide e~hanced the strand displacement activity of the HIV-RT D68 subunit at a molar ratio of 4:1 mainly by chasing short products into longer ones. Furthermore kinetic experiments after complementation of HIV=RT p66 with KIV-RT pSl indicated that HIV-RT pSl can restore rate and extent of strand displacement activity by HIV-RT p66 compared to the HIV-RT heterodimer D66/D51, suggesting a function of the 51 kDa polypeptide, The mouse mammary tumor virus proviral DNA contains an open reading  frame in the 3' long terminal repeat which can code for a 36 kDa polypeptide  with a putative transmembrane sequence and five N-linked glycosylation sites. This gene is known to code for a superantigen which deletes a specific subset el CO4 + T-lymphocytes in vivo, The superantigen encoded by the exogenous mouse mammary tumor virus of the GR strain acts specifically on V[314 bearing T-cells. We produced recombinant vaccinia viruses to express either the complete or a truncated ORF protein after infection of primate cells in culture. The complete err gene in mammalian cells leads to the production of a 47 kDa protein which is specifically detected with an anti-ORF-peptide antiserum. The 47 kDa protein can be labeled with D-[2-3H]mannose and its synthesis is inhibited by tunicamycin, an N-linked glycosylation inhibitor, indicating that it is a glycoprotein. The truncated ORF protein beginning at the second ATG of the open reading frame is also modified, but the C-terminal half of ORF, starting at the fifth ATG (ORF5), has the expected size of the non modified polypeptide. Pulse-chase experiments indicate that the 47 kDa ORF g}ycoprotein has a short half-life of about 1.5-2 hours in this system whereas the ORF5 truncated protein is even less stable; its half-life is about 20 minutes. The cellular localization of the ORF protein and the role it could play in vivo are currently under investigation. conclusion was based on a restriction fragment length polymorphism (RFLP) with probes of IS-elements. The mutants displayed different growth behavior. The evolutionary stability of 32 of these mutants was further tested by competing a mixture of them for ten days. Polymorphism was maintained during this competition but two fitter mutants took over 60% of the population. Implications of these results will be discussed in terms of generation and selection of genetic variants for the adaptation of microbial population. Mouse mammary tumor virus (MMTV) is produced in the mammary gland of infected females and transmitted by the milk to suckling new-boms. The passage of MMTV from the gut to the mammary gland is still poorly understood, the nature of the tissue tropism of MMTV has not been elucidated, A single receptor present on mammary epithelial and lymphoid cells might serve for the entry of the virus. A cell-type specific entry mechanism could in part be responsible for this phenomenon. Alternatively different surlace molecules might be involved in its uptake. We have studied the cell susceptibility to MMTV using a sensitive method of detection of newly acquired exogenous proviral DNA sequences, prior to viral integration and transcription. Thereby we confirm that MMTV infects mouse, rat and monkey cell lines of different tissue origin in vitro demonstrating the lack of a strict host range specificity in tissue culture as previously described.

We have also observed the infection by MMTV of various human cell lines. Furthermore various infection susceptibilities are observed for cell lines of a same species.

R. Cattaneo, P. Spielhofer, K. Kaelin, M.A. Billeter Mo/eku/arbio/ogie /, Universit~t ZOrich, HSnggerberg, 8093 ZOrich Subacute sclerosing panencephalitis (SSPE), a rare, lethal disease is caused by clonally selected defective measles viruses (dMVs) characteristically mutated. This emerged from sequencing of five MV genes from SSPE cases and by functional tests of the envelope glycoproteins hemagglutinin (H) and fusion protein iF). These expressed H and F variants migrate poorly to the cell surface but maintain some cell fusion activity whereas the envelope associated matrix protein (M) appears dispensable. Thus, the spread of dMV genomes through the brain seems to occur by membrane fusion at local cellular contacts, essentially hidden from the massive immune responses.

In more than half of SSPE cases, the propagated dMVs contain M genes in which up to 50% of U residues are replaced by Cs. Such "biased hypermutations" are likely due to the simultaneous conversion of many A residues to inosine in double stranded RNA regions accidentally formed, mediated by an ubiquitously occurring adenosine deaminase. Similar events at less spectacular degrees have now been observed in functional genes of several negative strand RNA viruses. Thus, this enzyme might be an important evolutionary driving force for Mononegavirates.

Pull, I., Wieland, S., Nieder~st, B., Keist, R., Walter, E., and Blum, H.E. Department of Medicine, University Hospital Z'tirich Infection by HBV-positive organ donors remains a serious problem in transplant surgery. Here we present a case of a 58-year-old woman who received a heart transplant from a hepatitis B surface antigen (HBsAg) positive donor. The patient was negative for all HBV markers prior to transplantation. Two months after transplantation she became HBsAg positive and died from subacute liver failure about seven months later.

Cloning and sequencing of HBV DNA revealed several HBV mutants in the serum of the recipient, each with a 108 bp in-frame deletion in the pre-S 1 region. Transfection studies in HUH-7 cells with the predominant pre-S 1 deletion mutant showed a higher replication competence compared to wild-type HBV. Sequence analysis of PCR products from serum of the donor showed mainly wild-type HBV DNA in the pre-S region without the pre-S 1 deletion.

These data demonstrate that mutant viruses accumulate in immunesuppressed patients. Specific mutants may play a critical role in the severe or fatal clinical course of HBV-infection in transplanted patients. In non-neuronal cells Herpes Simplex Virus (HSV) establishes a productive lyric infection starting with transcription of its immediate early (IE) genes by corecruitment of a viral protein (VPI6) and a cellular factor (Oct-l) onto characteristic oetamer-like sequence motifs fTAATGARAT) in the promotors of the IE genes. In neurons, however, HSV establishes latent infections whereby transcription of the IE genes is prevented and the lyric program aborted. Our efforts are focused towards determining which neuronal factors are involved in the regulation of HSV IE gene transcription and whether they may be involved in HSV latency or reactivation in neurons. In electrophoretic mobility shift assays (EMSA) using mouse brain nuclear extracts we found that neuronal Oct factors (NOct-2, NOet-3 and NOct-4) bind to TAATGARAT sequences from the viral ICP0 and ICP4 promoters. An additional brain protein also binds to these sequences but not to a consensus octamer sequence (ATGCAAAT) as do neuronal Oct factors. EMSA assays with mutated TAATGARAT probes show that the TAAT sequence is crucial for this binding and we have tentatively named this brain protein TAAT-1. We are currently testing the NOct factors in in vivo and in vitro transcription assays to study their involvement in HSV IE promoter regulation.

BIOCHEMICAL AND FUNCTIONAL ANALYSIS OF A VACCINIA VIRUS RECOMBINANT CONTAINING TWO COPIES OF THE P37K GENE.

Schmutz, C., and Wittek, R. Institut de Biologie Animate, Univei-sit~ de Lausanne, CH-1015 Lausanne.

The most abundant protein of the vaccinia virus envelope is the palmitated 37K protein. This protein is required for envelopment and subsequent release of virus from infected cells. The amount of extracellular enveloped virus (EEV) produced varies considerably depending on the virus strain and host cell line but remains in any case low (<30% of the virus production). If P37K is a limiting factor for envelopment, its overexpression might be a taeans to increase the production of EEV. For this purpose a vaccinia virus recombinant containing a second copy of the P37K gene was enginee,'ed. Western blot analysis clearly showed increased levels of this protein. Titration assays revealed slightly lower yields of both intracellular naked virus (INV) and extracellular enveloped virus (EEu in ceUs infected with recombinant virus. Unexpectedly, with recombinant virus, the ratio of EEV versus INV was significantly lower when compared to wild-type virus infection, despite the higher level of P37K expression. We are currently testing whether this is a consequence of a lower production, or a reduced infectivity of EEV particles in cells infected with recombinant virus.

Hanspeter Stalder, Christian Hertig and Ernst Peterhans Institut f~r Veterin~r-Virologie, Universit-st Bern, CH-3012 Bern Bovine viral diarrhoea virus (BVDV) causes acute transient infection in animals exposed to virus postnataly and persistent infection in animals infected in utero in the early phase of gestation. The latter is associated with specific immunotolerance to the infecting viral strain. Animals infected in utero may be born normal and remain healthy despite the persisting infection with a noncytopathic biotype of BVDV. Mutation to the cytopathic biotype, or superinfection with an antigenically similar cytopathic biotype, results in lethal mucosal disease. We have PCR amplified and sequenced parts of the region coding for the surface glycoprotein E2 and p80, a nonstructural protein associated with enzymatic activities. Over a time of one year, virus obtained from an immunotolerant persistently infected animal remained identical in its nucleotide sequence in the analyzed regions. Contrasting with this remarkable evolutionary stasis over some 600-800 virus generations is the heterogeneity of viral isolates obtained from different persistently infected animals. We propose that the muitilineage distribution of BVDV is the result of antigenic drift in the population of acutely infected animals, combined with evolutionary stasis in immunotolerant animals. in 20 to 30% of naturally infected goats, The sevedty of disease in infected animals correlates with the antibody titers to the viral envelope protein (enw) and especially to the gp 38 transmembrane podion (TM). In order to analyze linear B epitopes of env, we produced a random library of CAEV env polypeptides fused to ~-galactosidase and expressed in E. coli. The complete env gene obtained from an infectious clone of CAEV-CO was subcloned in a pUC18 plasmid and DNAse digested, Random fragments of 200 bp were inseded in ~,gtl 1 DNA and expressed in E. coil Y1090. As the first step of B epitope mapping and in order to identify immunodominant group epitopes we used high titer sera from naturally infected Swiss goats to screen the library. Six immunoreactive clones were obtained and subsequently sequenced. All six mapped to the TM region of env. A fine mapping of these immunoreactive regions was performed using pepscan technology and ELISA with synthetic peptides. Four epitopes could be identified including the eysteine loop corresponding to an immunodominant epitope in other lentiviruses. These peptides were used in vitro to test the reactivity of sera from naturally and experimentally CAEV-infected goats and in rive to modulate the immune response of goats to TM. The minute virus of mice (MVM) has a 5Kb linear single stranded DNA genome. At both termini, short selfcomplementary sequences are folded into stable hairpin structures. The 5' (right) hairpin is a single stem structure with only three unpaired bases within the stem. These unpaired form a bubble structure. We examined the requirement in MVM lytic infection for the bubble structure in the 5' terminal hairpin Of MVMp. MVMp RF DNA containing the entire 3' extremity and extending to the Hha I site downstream of the axis of symmetry of the 5' extremity was cloned into pGEM-SZf(+). This construct contains more than half of the 5' palindrome, however it lacks the sequence information required to form a bubble. Murine fibroblast A9 cells were transfected with this DNA and the 5' terminal sequence of resulting virus was analysed. A bubble was seen to be generated, however the sequence of nucleotides contributing to the bubble was of n0n-viral, plasmid origin. The nucleotides in question were removed from the original plasmid construct and the experiment repeated. Virus was obtained as before. Sequence analysis of the 5' end indicated that both flip and flop forms were present. The corresponding 5' hairpin deduced from these two forms contains no unpaired bases at the position where the bubble is normally located in wild type MVMp. The immediate early lIE) gene promoters of herpes simplex virus type-1 share a similar sequence which in some cases overlaps with a binding site for octamer binding proteins. This sequence confers responsiveness to a viral transaetivator, VP16. On these promoters, VP16 combines with multiple cellular proteins to form an activating complex. We set up a reconstituted in vitro system consisting of HeLa nuclear extract and recombinant Oct-1/POU domain protein (without the activation domain of Oct-1 ) and VP16 where we showed that the activation of IE genes is not mediated by Oct-I itself. Rather, Oct-1 serves to recruit VP16 that has no intrinsic DNA binding capacity. VP16, however, possesses a stong acidic C terminal activation domain of -80 amino acids which is indispensable for the in vivo function of the protein. As this and other acidic activation domains possess only a relatively weak activating capability when we test them in in vitro transcription assays we are interested in investigating the role of this acidic activation domain in vitro i.e. under condition of cell free extracts and on naked template DNA. We are using recombinant Oct-1/POU domain protein and recombinant VP16 that has been truncated at the C terminus. Additionally, we are also testing different truncations of VP16 in vitro that have been produced in cos cells. Furthermore, we are also trying to see whether these different truncations already affect the assembly of the actvating multiprotein complex. We have studied the kinetics of synthesis of transcripts initiated from the vaecinia virus 7.5 kD gene early promoter. Unexpectedly, after a first burst of RNA synthesis early in infection, the promoter showed a second peak of activity in the late phase.Reactivation of transcription was neither dependent on the presence of the sequences upstream of the early promoter, nor on the location of the promoter in the genome. Reactivation seems to be dependent on virus assembly since it is prevented by rifampicin, that specifically inhibits an early step of viral morpho-g~nesis. Analysis of several other early genes showed that reactivation is not unique to the 7.5 kD early promoter. analyzing, by in situ hybridization using digoxigenin-labelled riboprobes, the expression of the viral genome relative to that of an array of cytokines including TNF-(z; IFN-% IL-I~, IL-2, IL-8 and GM-CSF. Initial experiments suggest that viral genome expression is restricted in the inflammatory infiltrates whereas certain cytokines, particularly TNF-oq are strongly expressed in the synovial membranes. Overall, the results point to a prominent role of macrophage activation in the maintenance of chronic inflammation and possibly also in the development of arthritic lesions. The core protein of the hepatitis B virus (HBV) interacts with the viral RNA pregenome and the reverse-transcriptase.JDNA-polymerase to form a core particle which allows for replication of the viral genome. Deletion analysis of the core protein revealed, that the Arg-rich carboxyterminal domain is required for RNA encapsidation and productive viral DNA synthesis.

We demonstrate, that the duck hepatitis B virus (DHBV) core protein can be derided into two domains comprising amino acids 1-67 (N-ter) and 67-262 (C-ter), respectively. The co-expression of these two core subdomains leads to the formation of replication competent core particles. Furthermore, we demonstrate that the carboxy-terminal domain of DHBV core (C-ter) can be functionally replaced by the corresponding domain of the human HBV. The complete HBV core protein, however, does not form replication competent core particles with the DHBV pregenome and RT.

These results imply that the carboxy-terminal part of the HBV core proteins define an exchangeable, functionally conserved domain. Influenza HA mediates the adhesion and penetration of host cell membranes. Acylation of three conserved cysteine residues in the membrane spanning region of several HA subtypes has been shown. The biological significance of this hydrophobic modification remains to be elucidated. A possible role for the membrane fusing activity has been controversially discussed in the past. Removal of covalently bound fatty acid from protein by incubation with hydroxylamine is a commonly applied method. As we shall show for a number of different envelope viruses with acylated glycoproteins, this treatment leads to quite different consequences in a functional seiase depending on the virus used. As an alternative approach to study the biological significance of palmitoylation we expressed wild-type influenza HA and an HAmutant lacking the fatty acid binding sites using the baculovirns system. Both wild type and FA--mutant HA, after expression in insect cells bind to red blood cells, induce hemolysis, and fuse efficiently with fluoreseently labeled erythrocyte ghosts. Hydroxylumine treatment of both recombinant wt-and mutant-HA leads to a loss of hemolysis and of fnsion activity. At least in the ease of FA--I-IA its inhibitory action must be related to some other effect than fatty acid cleavage. These results indicate that hydroxylamine treatment may not always be suitable for the generation of fatty acid free glycoproteius or virus particles for functional studies. Gesemann and J.G. Nicholls, Biocenter Univ. Basel, 4056 Basel, Switzerland Depolarization of leech neurons leads to growth cone collapse and neurite retraction. This response to depolarization is influenced by the substrata on which the cells are growing and is mediated by the influx of calcium (S. Grumbacher-Reinert and Nicholls, 1992, J. Exp. Biol. 167:1-14; Neely, 1993 , I. Neurosci. 13:1292 -1301 . The morphologieal changes induced by depolarization of leech neurons growing on leech extracellular matrix extract is accompanied by a loss of microfilaments in the growth cones. Leech neurons growing on a different substrate, the plant lectin coneanavalin A (ComA), did not retract their neudtes or lose rnierofilaments, but continued to grow after depolarization. We attribute this to the reduced calcium influx during depolarization of neurons on ConA (Ross et al., 1988, PNAS 85:4075-4078) . Increasing the intraeellular calcium concentration by incubating neurons with the calcium-ionophore A23187, led to cessation of motility and disruption of microfilaments but not microtubutes in growth cones on ComA. To investigate the mechanism by which calcium affects the organization of microfilaments we stained neurons with antibodies against gelsolin, a protein that severs microfilaments when it is activated with calcium. We have observed that these antibodies colocalize with microfilaments in leech growth cones, We are now analyzing the nature of this antigen that cross-reacts with the anti-gelsolin antibodies with the goal of establishing its potential role in the calcium-induced disruption of microfilaments in leech neuronal growth cones. This work was supported by a grant from the Swiss Nationalfond No. 3127814.89 to J.G. Nicholls. 

Activation of metabotropic glutamate receptors (mGluRs) was studied in voltage-clamped CA3 cells in rat hippocampal slice cultures using the whole-cell pateh-clamp configuration. In saline containing 16 mM K +, 1S,3R-ACPD (50 ~tM), a mGluR agonist induced an inward current associated with an increase in membrane conductance. The reversal potential, assessed with a voltage ramp protocol from -80 to -60 mV and calculated by linear regression, was -15.7 • 18.7 mV, and was shifted to -53.4 • 6.4 mV when the concentration of external monovalent cations was halved. These results indicate that the conductance underlying this current is selective for cations (mainly K + and Na+). The steady-state I/V curve for the 1S,3R-ACPD-induced inward current showed a slight inward rectification. In most of the cells, this inward current was followed (or overlapped) by an outward current concomitant with a decrease in a K + conductance (Ere v = -51.2 ~ 5.6 mV in [K+]o = 16 mM and shifted to -74.0 :t: 2.8 mV in [I4+]o = 8 raM). This K + current was voltage-independent in the membrane potential range of-120 to -20 mV. Similar results were obtained with methacholine, a cholinergic muscarinic agonist. The non-specific cationic current, seen only in the presence of elevated extracellular K + concentration, could play an important role during intense neuronal activity. The c-Myc protein has been implicated in the regulation of cell growth and differentiation, c-Myc specifically interacts with Max which in turn forms heterodimers with Mad and Mxil and homodimera with itself. We present evidence that this network o! proteins is regulated both at the level of relative protein concentration as well as by post-translational mechanisms. In the hematopoietic cell lines U-937, ML-1 and HL-6O e-myc mRNA and protein are rapidly downregulated after induction o| differentiation whereas max expression is not significantly altered. In contrast the expression of mad is induced with properties similar to immediate early genes, taxi1 is induced to a lesser degree and with slower kinetics, In an effort to study the regulation of o-Myc we have mapped all the major in vivo phosphorylation sites in both c-Mye and Max. Two sites in the N-terminal transactivation domain are important for the regulation of the transforming potential el c-Myc both in established cell lines as well as in primary rat embryo tibroblasts. However, no difference in the transactivating potential of corresponding mutants was observed. Phosphorylatien of Max at sites close to the basic region increases both on-and off-rates of either Myc-Max hetero-as well as Max homodimers to DNA.

Brain Research Institute, University of Z~rich, CH-8029 Z~rich, Switzerland GABA B and adenosine receptors act via pertussis toxinsensitive G-proteins of the GI/G o class to mediate longlasting inhibition in the CNS. This class of G-protein is negatively coupled to the enzyme adenylyl eyclase. Our aim was to determine whether the inhibition of adenylyl nyclase mediated by these receptors has functional consequences for electrical signaling in hippocampal neurons. The calciumdependent K + current, I~, underlying adaptation of cell firing, was used as an electrophysiological bioassay for adenylyl cyclase activity, as it is very sensitive to intracellular levels of cAMP. Accordingly, the ~-adrenergic agoniat iaoproterenol (5-500 nM), that activates G,-linked receptors, reduced the amplitude of I~ by 51.5 • 3.6% in voltage-clamped CA3 pyramidal cells in TTX (n=20). Reduction of Imm by isoproterenol was curtailed to 27.3 • 2.9%, (p < 0.001) in the presence of baclofen (I0 ~M), acting at GABA B receptors, or adenosine (50 NM). Thus, the inhibition of adenylyl cyclase activity due to activation of GABA a and adenosine receptors can modulate responses to other neuretransmitters by an interaction Occurring at the level of intracellular signal transduction. Furthermore, stimulation of these receptors will tend to enhance I~, suppressing repetitive cell firing. This represents a further mechanism which will result in prolonged inhibition of hippocampal neurons. A study of the possible modulation by nitric oxide (NO) of endogenous dopamine (DA) release was performed in bovine retina in vitro. Isolated half retinas were incubated in KKG at 37 ~ for 3 successive 30-rain incubations, and DA was measured in the incubation medium by HPLC with EC detection. When retinas were incubated in the presence of the NO-generating compound hydroxylamine (300 ttM), a decrease in either basal or K +-(50 raM) stimulated DA release was observed during the 2 nd 30-rain period. Tested under similar conditions, dibutyryl cGMP (1 mM) was ineffective. The NO synthase inhibitor L-NG-nitro-arginine methyl ester (L-NAME, 100 gM) increased basal DA release during the 1 st and 2 nd incubation, whereas it did not affects the K+-stimulated DA release. In addition, its effect were potentiated in calcium-free medium. Finally, we have also shown in cell-free experiments that retinal NO synthase activity was totally calcium-dependent and blocked by L-NAME. Taken together, these results suggest that endogenous NO regulates DA release via a cGMP independent mechanism. Division of Endocrinology, University Hospital, CH-1211 Geneva 14 While the stimulation of aldosterone biosynthesis by angiotensin II (AngII) in bovine glomerulosa cells clearly depends upon a sustained influx of Ca 2+, the pathway for Ca 2 § entry is not yet accurately defined. At least two mechanisms seem to be involved: 1) the opening of voltage-operated Ca z § channels, and 2) the stimulation of a "capacitative" Ca 2+ entry regulated by intraocllular Ca 2 § stores; however the relative contribution of these pathways to the stimulation of steroidogenesis needs to be determined. Two pharmacological tools were used: nieardipine, a blocker of To and L-type Ca 2+ channels, and thapsigargin, which releases Ca ~+ and therefore induces a capacitative Ca ~+ influx. Nicardipine (1 laM) completely blocked the cytosolic Ca 2+ response to K +, which exclusively activates voltage-operated Ca 2. channels, but only partially (22 %) the response to AnglI. In the presence of nicardipine, Angll was unable to stimulate a Ca 2+ influx aiter depletion of Ca 2 § stores with thapsigargin, demonstrating that AnglI principally stimulates a capacitative Ca 2+ entry pathway. In addition, nicardipine completely blocked the steroidogenic response to K +, but only 30% of the response to Angll. Thapsigargin-induced aldosterone formation, as the response to Angll, was markedly potentiated by low concentrations of K +. In conclusion, this study shows that in bovine glomerulosa cells, AnglI activates a sustained Ca 2+ influx, principally through a capacitativc pathway, leading to aldostcrone formation. This finding could partially explain the poor efficiency of dihydropyridine Ca 2 § antagonists for preventing aldosterone secretion.

Harald Holder and John M. Lucocq*. Institute of Anatomy, University of Berne, CH 3012 Berne *Depanment of Anatomy and Physiology, University of Dundee, UK Dundee DD1 4HN. Subjection of Hela cells to 45~ for 30 minutes leads to a complex activation pattern of MAP-kinases as revealed with renaturadon gels. Exponentially growing HeLa cells show activallon of a myelin basic protein (MBP) phospborylating kinase migrating with an apparent molecular mass of around 60kDa upon heat shock induction. In contrast, in cells arrested in Go by serum starvation for 24h, essentially three different proteins phosphorylafing MBP are activated subsequent to heat shock treatment. These kinases migrate with apparent molecular masses of around 32kDa, 44kDa and 55kDa. Western blot analysis with antibodies recognizing the human boraologues of erkl and erk2 revealed that considerable pools of erkl and erk2 show retarded elec/rophnretic migration behavioar indicating that they are phosphorylated and probably activated in heat subjected and growth-arrested HeLa cells. Down-regulation of protein kinase C (PKC) by a prolonged treatment with 12-O-tetradecanoyl-phothol-13-acetate (TPA) did neither change the major activation pattern observed upon separation of cell lysates on renaturation gels nor that one observed on western blots decorated with anti-erkl or anti-erk2 antibodies. These results suggest that heat shock induced activation of at least erkl, erk2 and the 55kDa MBP-kinase is not mediated via activation of PKC. This work was supported by Swiss NSF grant no. 31-27636.89.

Effects of dominant negative glucocorticoid receptor mutants in cell cultures and transgenic animals Stefano Brenz Verca, Stefan Schneider, Sandro Rusconi, Institut fiir Molekularbiologie H der Universitat Zf~rich, Winterthurerstr. 190, Switzerland The glueocorticoid receptor (GR) is involved in a large number of developmental and biochemical processes. Recently, we obtained a trans-dominant negative GR mutant (TDN) -called GR[AIa] -by a frame shift of the CAG-repeat of the rat GR (see poster by Hug et al.) . The GR[Ala]-mutant shows normal llgand-binding and DNA-binding but cannot stimulate transcription (Lanz et al., Steroids; in press) . Modulatable versions of the TDN-mutant have also been created by substituting the HBD of the GRmutant with the HBD of sex-hormone receptors. These constructs have only been tested so far in transient cotransfection assays. Therefore, we are currently trying to express them permanently in cell lines (rat hepatoma cell line FTO-2B) in order to verify whether they can affect the transcription level of a known endogenous GR-dependent target-gene, like the tyrosine aminotransferase (TAT) gene or a permanently transformed MTV-IacZ reporter. So far no rapid test-system has been described, that allows the quantitative analysis of such effects in transient expression assays. We are about to set up such a rapid and generally applicable transient test-system. In the future we plan to express dominant negative GR mutants in a tissuespecific manner in transgenic mice. It will be particularly interesting to investigate their effects on the immune system or on liver functions. These studies investigated growth cone responses m brief local encounters with surface molecules implicated in neuronal pathfinding using chick DRG neurons. We tested two interrelated hypotheses: 1) Discrete points of information (guideposts) override sustained surface information and dominate growth cone behavior and 2) guidance molecules provide such information by activating second messengers instead of simply increase adhesivity. The potential guidance molecule lamialn was covalentiy coupled to polystyrene beads in order to mimic guidepost cells. Encounters with laminin model guideposts significantly altered the behavior and morphology of growth cones advancing on fibronestin or polylysine. Growth cones steered towards theses beads in a saltatory medea once a long lasting contact with a single filopodium had been established. Subsequent to a pause at the bead, growth cones advanced with a significantly increased growth rate for the next llmin. The average increase in rate stimulated by bead contact was 5 times the basal rate on polylysine and 2.5 times that on fibronectin. Positioning of multiple model guideposts along a path and within filopodial reach maintained this bead stimulated rate of outgrowth and directed growth cone advance. The laminin induced pattern of events was totally blocked in the presence of either H7 or sphingosine, two PKC inhibitors, or Neomycin, a PLC inhibitor. Neither of these inhibitors affected neurite outgrowth on the substrata alone. Our results demonstrate the necessity of filopodial contacts, the importance of spatially restricted stimuli and the activation of transducing pathways as key mechanisms in neuronal pathfinding.

Mechanisms of glucocorticoid receptor desactlvation by homopolymeric alanines Martin Hug, Manuela H&fferer and Sandro Rusconi, Institut fiir Molekularbiologie 11 der Universitat UZ Zart'ch, Winterthurerstrasse 190, 8057 Ziirich, Switzerland The rat glucocorticoid receptor (GR) eDNA contains in its Y-portion a (CAG-repeat that encodes a siring of 22 Gin residues interrupted by 1 Arg. A mutant with the frame-shifted repeat which codes now for poly-Ala residues (called GR[AIa]) is completely trans-inactive, is more cytoplasmitally located in absence of ligund and exhibits dominant-negative properties in presence of ligand (Lanz et al., Steroids, in press) . We generated GR recombinants with increasing numbers of CAG-Triplets in three possible reading frames (oligo-Gln, -Ser or -Ala). The GR-oligo[Ala]~-~.23 mutants showed an activity comparable to wild type GR, whereas the GRoligo[Ala]>_25 were inactive. The severity of the effect of oligo A.la su'ings seems to be dependent on the position of the repeat along the primary sequence (e.g. progressive decrease of the effects by moving toward more earboxy-position). We have also examined the effects of the oligo[Ala]21, 23, 25, 27, 31 on GAL4 chimeric factors and found that the inhibition by Ala repeats is probably factor-specific. Our results made us speculate that possible differences in post-translational modifications may be at the root of the negative effects by the Ala repeats. We believe that these effects are due to transient association of the nascent GR[AIa] mutant proteins with intraceUular membranes. These hypotheses are currently tested.

The Activation of Steroidogenesls in Calcium-Clamped Bovine Glomerulosa Cells and its Modulation by Angiotensin II and cAMP, Python CP, Laban OP, Rossier MF, Vallotton MB and Capponi AM Division of Endocrinology, University Hospital, Geneva, Switzedand. The Ca2+-messenger system plays a crucial role in the regulation of steroid secretion in adrenal zona glomerulosa cells, as it is known to mediate the action of angiotensin II and potassium ion. In the present study, we used intact isolated glomerulosa cells in which the cytosolic free Ca 2+ concentration ([Ca2+]c) was clamped at various levels with the Ca2+-ionophore, ionomycin, in order to locate the sites of action of Ca 2+, to determine their sensitivity towards Ca 2+ and the modulation by other physiological factors. By measuring simultaneously steroid synthesis and [Ca2+]c, we show that Ca2+-clamping (50-860 nM) induced a concentration-dependent increase in the production of both pregnenolone (506___70 % of the basal level) and aldosterone (255+27 % of the basal level, EC50 = 273 nM). By contrast, Ca 2+ did not influence the conversion of 1 t-deoxycorticosterone into aldosterone at concentrations up to 600 riM, although at higher concentrations we observed a modest but significant inhibition. In addition, Ca2+-clamping did not affect the formation of pregnenolone from a freely diffusible analogue of cholesterol, 25hydroxycholesterol, indicating that Ca 2+ acts at a step upstream of cholesterol side-chain cleavage. Both angiotensin II and a membrane-permeant surrogate of cAMP, dibutyryl cyclic AMP, potentiated the steroidogenic response to increases in [Ca2+]c . In summary, these studies reveal an exquisite sensitivity of the glomerulosa cell steroidogenic pathway toward very small physiological changes in [Ca2+]c. Radtke, Rainer Heuchel, Oleg Georgiev, Gerlinde Stark#, Michel Aguet# & Walter Schaffner Institut fiir Molekularbiologie 11, UniversiLat Ziirlch, 8057 Z'6rich # lnstitut Rir Molehdarbiologie I, Universitiit Ziirich, 8093 Ztirich We have previously described and cloned a factor (MTF-1) that binds specifically to metal-responsive DNA sequence elements in the enhancer/promoter region of metallothionein genes. MTF-1 is a protein of 72.5 kDa that contains six zinc fingers and multiple domains for transcriptional activation. Here we report the disruption of both alleles of the MTF-1 gene in mouse embryonic stem ceils by homologous recombination.

The resulting null mutant cell line fails to produce detectable mounts of MTF-1. Moreover, due to the loss of MTF-1 the endogenous metalinthionein-I gene is silent, showing that MTF-I is required for both basal and metal-induced transcription. Accordingly, reporter genes containing either synthetic or natural metal-reslxmsive promoters were not transcribed after transfection into the null mutant ceils. Cotramsfection of the MTF-I expression vector rescued both basal and heavy metal-induced transcription. These results demonstrate that MTF-1 is essential for normal metallothionein gene regulation. Rat pheochromocytoma cells (PC12) were differentiated with NGF for 3-4 days, a time sufficient for the formation of growth cones and elongation of neurites. Addition of the phosphatase inhibitor calyculin A (CL-A) (0.2.50.0 nM) led to a concentrationdependent collapse of growth cones and retraction of the neurites within 15 minutes. After the retraction, the cell bodies showed the appearance of a grape-like domain opposite to the cell nucleus. They recovered to normal shape within about 30-60 minutes if the inhibitor was washed out. The neurite retraction was further analysed using both low-light level fluorescence video-microscopy and fura-2 single cell microfluorimetry. The onset of retraction started in the central part of the growth cone prior to a complete collapse of filopodia. The basal intracellular free calcium concentration ([Ca2+] i) remained constant during neurite retraction. As a measure for the viability of CL-A-treated cells, agonist-isduced responses of [Ca2+] i were analysed. Ca2+ entry across voltage-activated Ca 2-~ channels was inhibited by 50% after 30 minute treatment with CL-A, while the ATP-induced Ca2+ entry via Ca2*-permeable cation channels was not significantly reduced. However, the onset of the ATP-induced rise in [Ca2+]i started at the grape-like domain. The speed of the Ca 2+ wave across the cell body was slower than in control cells. The addition of 0.5 gM okadaie acid, a different type of phosphatase inhibitor, caused no cell shape change or neurite retraction within 24 hours. Taken the specificity of the two phesphatase inhibiters into account, the observed effects seem to be induced by inhibition of a PPl-class phosphatase. 

Myosin in nematocytes of Hydra Michel Nakano & Robert P. Stidwill Dept. Zoology, University of Ztirich, Switzerland

The functions of nonmuscle myosins during cell locomotion are not clear. Several models have been presented reaching from very active participation of the molecules in different parts of cells to only minor or no roles at all. There is increasing evidence that the functions of myosin II (capable of forming bipolar filaments) are quite distinct from the ones of myosin I and generally that the different members of the growing family of myosin-like proteins may not all have identieal functions.

We have attempted to demonstrate the occurrence and determine pattern of myosins in tissue of Hydra and especially in migrating nematoeytes mainly for two reasons. First, nematocytes are fairly rapidly moving cells which can be studied in vitro and, second, the question is of interest from a evolutionary point of view.

We have utilized several antibodies against vertebrate and invertebrate myosins for immunofluoreseenee (confocal laser scanning microscopy) and Western blotting studies.

In addition to these results findings on the screening of a ~.ZAP c-DNA library are presented. The JAKI protein tyrosine kinase is a member of a family of intracellular kinases characterized by having two kinase catalytic domains: a bona fide tyrosine kinase domain and an amino terminally positioned kinase-related domain. Recently it has been shown that JAKI and the other members of this family (JAK2, TYK2) are intimately involved in cytokine and growth factor signalling. The presence of two putative nuclear localizing sequences in the amino terminus of the protein prompted us to examine the subcellular localization of this protein. Immunohistochemical and biochemical analysis revealed that a part of JAKI was localized to the nucleolus. Metabolic labelling showed that (a) the JAKI protein partitioned rapidly into the nucleus and (b) the nuclear JAKI became smaller with time. The size difference was not due to either co-translational glycosylation or phospherylation. These results suggest that part of the JAKI protein is rapidly translocated to the nucleoli where it becomes processed. revealed that this clone is expressed predominantly in spleen, mammary gland and thymus as two transcripts, a major species of approximately 5kb and a minor species of approximately 4kb. Nucleotide sequence analysis revealed a 84% homology to the porcine syk gene. The syk gene is expressed as a major 3.0kb transcript and has been reported to be lymphoid specific. Due to the discrepancy of transcript size and number, we suggest that our cDNA clone represents a novel member of this family of PTKs. We are currently isolating a full length sequence from a mouse spleen cDNA library. In addition, we want to characterize which cell(s) in the mammary gland are responsible for the observed expression. Expression and function of G-protein isoforms were studied in differenflaring (6C8) and nondifferentiaflng ((33) subclones of RED-1 Rauscher murine erythroleukemia cells. Erythroid differentiaflon induced by the combined action of erythropoietin (Epo) and DMSO was associated with a selective loss (>70%) of immunoreactive Gai3. Simultaneously, a Iruncated form of G~2 accumulated in the cytosol, while the membrane concentrations of Gai2, Gets and Gaq/11 remained unchanged. Nondifferenfiating (33 cells contained sigrtificanfly higher levels of most Get subunit isoforms that remained unchanged upon Epo/DMSO treamaent. Changes in adenylyl cyclase activity and in intracellular Ca ion levels ([Ca]i) resulting from activation of G-protein-coupled receptors were monitored for differentiation-dependent effects on transmembrane signaling. ADP-mediated changes of [Ca]i in native and differenflaflng 6C8 cells were consistent with P2T receptor coupling to Gai3. Thrombin receptors seemed to couple preferentially to Gai in G3 cells and to Gaq in 6C8 cells. Our results document substantial alterations in G-protein-mediated signaling during erythroid differentiation. To investigate the involvement of protein tyrosine kinases (PTKs) in the growth control of the mammary gland, we have used PCR based cloning to identify PTKs expressed during normal mammary gland development. This approach led to the isolation of three members of the eph-related family of receptor PTKs: myk-;, a novel member expressed predominantly in lung, heart and mammary gland; m-eck, the murine homologue of the human eck gene and mek5, the murine homologue of the chicken cek5. All three PTKs were expressed in a distinct and narrow window of mammary gland development: puberty and estrus cycle. No expression was found either in late pregnancy or in the end differentiated state of lactation.

0vet-expression was found in the undifferentiated and invasive tumors of transgenic mice expressing the H-ras oncogene. In contrast, no elevated expression of all three genes was detected in the well differentiated, non-invasive mammary tumors characteristic of c-myc expressing transgenic mice. These results suggest that members of this family of PTKs are involved in the control of proliferation of the mammary gland but are incompatible with its differentiation. NP-TCII is a lymphocyte specific transcription factor (Lattion A.-L. et el., Mol. Cell. Biol., 1992, 12:5217-5227 a nucleotide receptor that mediates phosphoinositide and phosphatidylcholine hydrolysis by phospholipases C and D, respectively. Here we report that ATP and UTP potently stimulate mesangial cell proliferation. Both nucleotides stimulate phosphorylation and activation of MAP kinase and the up-stream MAP kinase kinase. Activation of MAP kinase by both nucleotides is dose-dependently attenuated by the P2 receptor antagonist suramin. Stimulation of protein kinase C (PKC) by phorbol ester increased MAP kinase phosphorylation and activation, and downregulation of PKC partially inhibited ATP-and UTP-induced MAP-kinase activation. Inhibitors of PKC which display a selectivity for the Ca2+-dependent isoenzymes (c~, [3, 7) , as compared to the Ca2+-independent isoenzymes (5, e, ~, rl) such as staurosporine and the specific PKC inhibitor CGP 41251 did not inhibit nucleotide-stimulated MAP kinase phosphorylation, thus implicating the involvement of a Ca2+-independent PKC isoform. In conclusion, these data suggest, that ATP and UTP trigger the activation of the MAP kinase signalling cascade in mesangial cells and this may be responsible for the potent mitogenic acitivity of both nucleotides.

SUBCELLULAR ION-GRADIENTS IN CA SIGNALING P. Lipp & E. Niggli, Department of Physiology, Univexsity of Bern Recent experimental evidence has indicated an important role for submicroscopic concentration gradients during Ca-signaling in various cell types. In heart muscle cells influx of Ca via L-type Ca channels triggers Ca release from the sarcoplasmic reticulum (SR) and links electrical excitation to mechanical contraction (ec-coupling). We used a ratiometric confocal method (Fluo-3 and Fura-Red) to follow the intraceUular Ca-concentration while Ha-and Ca-currents (INn, Ica) were elicited in the whole cell mode of the patch-clamp technique. Both, Icand IN, were able to trigger Ca release from the SR. Kinetic differences between INI and Ic -induced Ca-transients, Li-substitution experiments and the existence of a residual INn-induced Ca transient in the absence of SR function suggested that these Ca signals were mediated by the Ha-Ca exchange running in the Ca influx mode. Control experiments showed that INs-induced Ca transients did not result from spurious activation L-type Ca channels. The significant increase of the Na concentration required to activate the Ha-Ca exchange can only occur provided cytosolic diffusion of Na is restricted in the vicinity of the exchangers. This limited diffusion results in significant Ha-gradients in the subsarcolemmal space. In conclusion, I~iinduced Ca influx may represent a safety factor for ec-coupling while Ic, ensures sustained Ca-release and is the source for refilling the SR with Ca.

Supported by SNF. The regenerative Ca-induced Ca release (CICR) mechanism is an important amplifier in cardiac signal transducilon. The CICR contributes to the Ca transient responsible for mechanical activity, but also generates Ca waves propagating within the cytosol. We investigated subcellular Ca signals in neonatal rat and adult guinea-pig ventricular myocytes using ratiometric confocal microscopy with a mixture of two fluorescent Ca indicators. Individual resting myocytes exhibited spontaneous Ca release events from the sarcoplasmic reticulurn (SR) that were attributable to three distinct patterns: i) local Ca release events without propagation; ll) planar Ca waves propagating across the cell; iii) spiral Ca waves spinning around a nucleus or shifdng along a subcellular region without entering it. Transitions between these patterns were frequently observed, suggesting that a particular release pattern is not the property of an individual cell but reflects the functional state of the SR. The existence of focal release and refractory zones within a single cell indicates that the SR is not uniform on the subcellular level. The SR seems to he a network of elements with distinct functional properties. A suhcellular variability in the positive feedback of the CICR mechanism can account for the observed features of Ca signaling. The degree of positive feedback exhibited by a single SR element may depend on its Ca content. To characterise the intracellular signalling and regulatory properties of the cloned parathyroid hormone (PTH) receptor we have stably transfected two renal epithelial cell lines with cDNA's encoding the PTH receptor (provided by Dr G. Segre). The proximal tubular, porcine (LLC PK1 [clone 4]), and the distal tubular, Madine Darby canine (MDCK) cell lines were transfected since the culture of either cell line on permeant filter supports leads to the development of polarised epithelial cell layers that mimic the epithelial organisation in vivo. Transfection of the cloned PTH receptor into each of these cell lines allowed us to examine, independently, the functional properties of the receptors expressed at the apical and/or the basolateral cell surface. While the basolateral application of PTH produces a large increase in cAMP production in both cell lines, the relative efficacy of an apical application of PTH differs between the cell types. The transfected LLC PK1 cells exhibit a greater response to apical PTH compared to the transfected MDCK cells. PTH-mediated regulation of intrinsic phosphate transport was also examined in polarised, transfected MDCK cells. Neither apical or basolateral application of PTH affects intrinsic apical or basolateral Ha-dependent or phosphate transport activities. Similar experiments are being performed with the transfected LLC PK1 cells. A. and Preiss, A. Biozentrum der Universit&t Basel, CH 4056 Basel Hairless, a recessive larval lethal mutation, causes dominant defects in sensory organs of adult Drosophila flies.Manifold genetic interactions indicate that Hairless antagonizes neurogenic gene function suggesting an involvment also in neurogenesis. In order to gain insight into its function we have cloned the Hairless gene which encodes an extremely basic, very serine rich protein of ii0 kD calculated molecular weight. No significant homologies to proteins of known function could be identified. Therefore, we are concentrating by various means on a structure -function analysis of the Hairless protein. Firstly, we are testing a series of Hairless deletion constructs for rescue capacity and generation of anti-Hairless phenotypes after overexpression compared with the wild type gene. Secondly, we are analysing HAIRLESS protein distribution throughout development, also at the subcellular level. Heat induced HAIRLESS protein runs at ~ 150 kD, possibly due to phosphorylation which might be intrinsic to Hairless function, a hypothesis we are currently scrutinizing. Thy-1 and CD26 surface glycoproteins are both capable of delivering proliferative signals to T cells upon cross-linking with appropriate antibodies. The tyrosine phosphatase CD45 is a key regulator of T cell proliferation as it controls the activity of src family kinases p56/~k and p59fY n. Associations of Thy-1 with CD45, and of CD26 with CD45 have been reported after chemical cross-linking of intact cells and it is likely that such associations constitute part of the structural basis for the T-cell stimulating capacity of Thy-1 and CD26 accessory molecules. We report that Thy-1, CD26 and CD45 can be specifically crosslinked at the cell-surface and isolated as multimolecular complexes containing a 78 kDa protein that is recognized by an anti-BiP (Grp78) antibody. In vitro kinase assays show the selective association of p56 Ick and p59fY n with Thy-1 and CD45 contained in multimolecular complexes. The multimolecular structure we describe could representa transducing unit incorporating surface receptors and regulators of tyrosine phosphorylation that are stabilized through interaction with the BiP protein in the plasma membrane.

STRUCTURE-FUNCTION-ANALYSIS OF HAIRLESS, A GENE INVOLVED IN DROSOPHILA NERVOUS SYSTEM DEVELOPMENT Maier, D.,

FUNCTIONAL COUPLING OF PPARs AND TRs IJpenberg, A., Wahli, W., Desvergne, B., Institut de Biologie Animale, 1015 Lausanne We are interested in a possible functional coupling of thyroid hormone (TR) and peroxisome proliferator-activated receptors (PFAR). To this end, we performed cotransfection experiments in NIH 3T3 cells using CAT reporter constructs containing the promoter of either the T3 responsive malic enzyme (ME) gene, or the PPAR regulated acyl CoA oxidase (Aco) gene.

The ME reporter construct showed very moderate response to the activated mouse PPAR (mPPAR), whereas combination of TRy1 and unactivated mFPAR potentialized the T3 induction. A putative PPAR response element was identified within this promoter and will be further investigated.

The Aco reporter constructs were not responsive to TRs. In constrast, a TR and T3 dependent inhibition of the FPAR mediated activation was observed.

The mechanism of functional coupling of these receptors, which may possibly involve the 9-cisretinoic acid receptor RXR, will be further investigated by molecular analysis. Unliganded steroid receptors form a complex with HSP90 via their hormone binding domain (HBD). The receptor activation involves a hormone-induced release of HSP90. We genetically addressed the role of HSP90 in budding yeast by analyzing three kinds of HSP82 (the essential yeast homologue of HSP90) mutants. These mutations affect either the quantity of the protein (low versus normal levels), its integrity (deletion analysis) or its nature (HSP82 homologues from other species). HSP82 mutant are examined for their ability to complement a HSP82 deletion mutant. Moreover, they are tested for functional interaction with a HBD. Interestingly, a viable deletion of a charged region, suspected to be involved in HSP82-HBD interaction, only slightly interferes with hormonal activation of both estrogen or glucocorticoid receptors. Furthermore, a short deletion in the last third of HSP82 leads to the production of a dominant negative mutant which prevents ceil growth at elevated temperature. The ability of various HSP82 homologues to palliate or not the absence of the yeast protein allows us to define, using chimeras, the regions of the protein which are necessary and sufficient to support cell life.

Zhang,H.,Reynaud, S. and Spohr,G. D~partement de Biologie Cellulaire, Sciences III, 30, quai Ernest-Ansermet, CH-1211 Gen&ve 4

Id cDNAs expressed during early Xenopus development have been isolated and sequenced. The encoded protein is homologous to the proteins described in higher vertebrates. Northern blot analysis reveals that transcription starts soon after mid blastula and decreases to some extent after tailbud-stage. Lower levels of transcription are detected also in adult frog tissues such as liver and heart.

Microinjection into oocytes of in vitro-synthesized MyoDor/and Id-mRNA, along with a CAT reporter gene whose expression is under the controll of an a3-actin promoter supplemented with four E-boxes shows that Myo D function is repressed by Id.

To investigate the importance of negative regulatory sequences we have isolated and characterised the Id promoter. As a strategy for identifying cis-regulatory elements, chimeric genes consisting of different 5' elements of the promoter were attached to the CAT coding sequences and microinjected into embryos. We are studying the tissue distribution of the rat Peroxisome Proliferators Activated Receptor (PPAR~), a member of the nuclear hormone receptors superfamily, during development and in adults.

High levels of PPAR~ mRNAs are detected in adult liver, kidney and heart. Muscle, stomach, and intestine show moderate amounts of the PPAR~, whereas brain, testis and lung present a very low expression. During postnatal development, we observe a continuous increase of the PPAR~ expression in the kidney, in contrast to a steady level in the liver. Anjard, C., Groux, B ., Gamboni, S. and Reymond, C.D., Institut d'histologie, UNIL, Rue du Bugnon 9, 1005 LAUSANNE.

The cAMP dependent protein kinase (cAPK) from Dictyostelium is composed of a single catalytic (C) and a single regulatory (R) subunit. The C subunit we characterised, PkaC is about twice the size of its mammalian counterpart. We showed that it possesses all properties of a catalytic subunit.

It associates with the R subunit in absence of cAMP, it copurifies with cAPK activity and an increased activity is observed in cells over-expressing PkaC. Furthermore, we dissected its role during development and cell differentiation.

Overexpressing pkaC under cell type specific promoters allowed to show the requirement for its activity during spore differentiation. PkaC seems to play a more complex role in prestalk cells.

We are now dissecting the functional regions of the PkaC protein, making use of the known tertiary structure of the C subunits of mammalian enzymes.

DNA BINDING PROPERTIES AND STIMULATION OF GENE TRANSCRIPTION OF BACULOVIRUS EXPRESSED xPPAR~. Hihi, A.K, Mermod, N., Medin, J., Ozato, K. and Wahli, W., Inst. de Biol. Animale, Uni Lausanne, CH-1015 Lausanne. Lab. of Mol. Growth Reg., NIH, Bethesda, MD, 20892 USA. The peroxisome proliferators activated receptors (PPARs) are members of the steroid/thyroid nuclear receptor superfamily. So far, they have been found in amphibians, rodents and humans. In Xenopus laevis, 3 subtypes (xPPARa,~.y} have been isolated. In order to study the mode of activity of the ~ form, the xPPAR~ gene product was overexpressed using a baculovirus expression system. The nuclear protein obtained has the correct molecular weight of 46 KD. Gel shift assays showed that nuclear extracts containing the recombinant protein are able to specifically bind the 'PPAR response element' which is a direct repeat of the core AGGTCA motif. This DNA binding activity is potentiated by another nuclear receptor, the mouse RXR~ and is inhibited by phosphatase, suggesting a role for phosphorylation in PPAR binding to DNA. A functional approach, using an in vitro transcription system, was chosen to define PPAR and RXR action on transcription stimulation, as well as the role of their respective activators. Arachidonic acid (0.5 -20 raM) induces various patterns of Ca2+i signal in bronchial smooth muscle cells, as revealed by single cell dynamic video imaging of Fura-2 loaded cells. Most frequently, Ca2+i oscillations were observed, although other patterns (a single Ca2+i transient; a single peak followed by a sustained elevated level, or a sustained Ca2*i elevation solely) were occasionally seen. The amplitude of Ca2*i oscillations was related to the concentration of arachidonic acid. The frequency histogramm was noticeably shifted to higher frequencies by nordihydroguaiaretic acid (NDGA, 0.2 rtM, a lipoxygenase inhibitor), whereas it was markedly shifted to lower frequencies by indomethacin (50 pM, a cyclooxygenase inhibitor) and by 5, 8, 11, 14-eicosatetraynoic acid (ETYA, 3 ~tM, a lipoxygenase and cyclooxygenase inhibitor). These observations suggest that: (1) arachidonic acid per se elicits Ca2+ i oscillations in these ceils; (2) cyelooxygenase activity products modulate the frequency of these oscillations.

PROMOTER ANALYSES OF A GROWTH FACTOR REGULATED GENE T. Triib, M. Kalousek, R. Kessler, and R. Klemenz Dept. of Pathology, University Hospital, 8091 Ziirich Stimulation of quiescent cells to enter the cell cycle results in altered gene expression. Some of the first genes to be induced (immediate early (i.e.) genes) encode transcription factors which are thought to pass on the mitotic signal to the delayed early (d.e.) genes. We have analysed the promoter elements required for growth factor mediated induction of the d.e. mouse gene T1 which encodes a secreted glycoprotein of the immunoglobulin superfamily. A 448 bp region located between 3.6 and 4.0 kb upstream of the transcription start site is sufficient for growth factor mediated inducibility. It contains an AP-1 binding site which is absolutely required for T1 gene expression. It is surrounded by 3 E-boxes. At least 2 of them are essential for efficient gene induction. Thus the i.e. proteins encoded by the fos and jun gene family which form the AP-1 complex can activate the d.e. gene T1 in collaboration with helixloop-helix transcription factors binding to the E-boxes. The increase of radiolabelled inositol phosphates ([3H]IPs) production in agonist-stimulated human airway smooth muscle cells (ASMC) was determined by HPLC. In order to observe the role of calcium, PKC and PLA2 on PLC activity during agonist stimulation, the effects of pharmacological agents (able to alter calcium, PKC and PLA2) were tested on IP production elicited by a 5 sec stimulation with histamine. The inhibitor of PKC staurosporine increased, while PMA (an activator) decreased the histamine-stimulated production of [3HI 1,4,5-IP3 and of its degradation products. An inhibition was also observed with the two PLA2 inhibitors, quinacrine and p-bromophenacyl bromide. In calciumfree buffer, no difference in the IP increase was shown, but an inhibition was observed when thapsigargin, an inhibitor of the Ca 2+, Mg 2+-ATPasc of the sarcoplasmic reticulum, was added in the same conditions. The calcium ionophore ionomycin increased the lP production in presence of external calcium, whereas it completely blocked the stimulation in calcium-free buffer. The results suggest that PLC activity, in human ASMC, is modulated by a positive feedback of calcium and by a negative feedback of PKC and PLA2.

REGULATION OF HORMONE-STIMULATED CALCIUM SIGNALS M. Boolman, AFRC, Laboratory of Molecular Signalling, Dept. Zoology, Cambridge, UK Stimulation of cells with hormones that activate inositol 1,4,5-trisphosphate (InsP~) formation, leads to an increase in the intracellular Ca 2. concentration ([Ca2+]~). These hormone-stimulated [Ca2 § signals have a complex temporal and spatial regulation, with the [Ca2*]~ increase often taking the form of a series of repetitive spikes or oscillations, arising from a steady basal level. The spatial counterpart of a [Ca2+]~ spike is a wave, where [Ca2 § is first elevated in a localised region of a cell, and then spreads throughout the cell in a regenerative manner. The [Ca~*]~ spike frequency is modulated by several environmental factors including hormone concentration, extracellular calcium and thiol reagents. Current evidence suggests that a cell can interpret these complex [Ca2+]~ changes to regulate a diverse range of cellular activities. Although many of the biochemical steps between hormonereceptor activation and the [Ca2 § response are known, the precise mechanism generating these complex [Ca2+]~ signals is unclear. In order to understand this mechanism more clearly, we have investigated the regulation of InsP~-mediated Ca ~ § release from intracellular stores in intact cells. Our current data suggests that under certain conditions, the InsP3-sensitive intracellular stores behave as functionally discrete units, and that during a Ca 2 § spike these stores become functionally coupled by the diffusion of Ca z+ from one store to the next, triggering a regenerative Ca 2+ release.

Phenylarsenoxide as a tool for identifying proteins involved in the secretory process Wiedemann. C., Schaefer, T.,*Gitler, C., Burger, M. M. Friedrich-Miescher institut, P.O.Box 2543, CH-4002 Basel "Department of Membrane Research and Biophysics, Weizmann Institute of Science, Rehovot 76110, Israel Phenylarsenoxide (PAO) at 20btM blocks exocytosis in chromaffm cells of the bovine adrenal medulla. This inhibition can be reversed by small dithioIs, such as dithiothrcitol or 2,3~limercaptopropanol. Because trivalent arsenicals interact specifically with dithiol4containing proteins, we conclude that dithiol-proteins do play an important role in regulated secretion in chromalYm cells. To identify dithiol-proteins putatively involved in the secretory process, we compare PAO-binding proteins from chromaffm and PCI2 cells and from fibroblasts as well as from sulx:ellular fractions by 2-dimensional gelelectrophoresis. The proteins are identified by radioactive PAO bound to the dithiol in a complex that witltslands SDS-PAGE. Affinity chromatography on PAO-Scpharose with various protein fractions is used to purify the dithiol-proteins of interest. Enzymatic activity, e.g. kinase or phosphatase activity, can be measured in the eluates of such a column to give further hints at the possible function of dithiol-proteins. RAC-PKs (~elated to PKA_ and PKCC protein kinnses) represent a serine/threonine kinase family whose catalytic domain is closely related to those of protein kinases A and C. They contain a PH (pleckstrin homology) domain N-terminal to the catalytic domain. In addition, RAC-PK~ was shown to be a proto-oncogene. In order to investigate the role of RAC-PK in cellular signallimg we undertook genetic and biochemical analysis of the Drosophila homolog (DRAC-PK). The DRAC-PK gene encodes multiple classes of transcripts. Antisera raised against the DRAC-PK recombinant protein specifically recognize 3 distinct molecular weight forms (68, 85 and 120 kDa). All forms are expressed throughout development and are both maternally and zygotically regulated. The DRAC-PK gene is localized at 89B4-I0 region of the third chromosome, betwee~ the pannie~ and the Stubble genes. Universit6 de Lausanne, rue du Bugnon 9, 1005 Lausanne When spinal meninges homogenates were incubated with reduced glutathione (GSH), a cofactor of prostaglandin (PG) biosynthesis, [14C] arachidonate (AA) was bioconverted into a major and a minor product which comigrated on TLC with PGE2 and PGD 2 respectively. In absence of GSH: 1) PGD2 could not be detected; 2) the level of PGE2 was dramatically reduced but surprisingly counterbalanced by accumulation of an unusual [I4C] AA metabolite which exhibits particular traits of migration on TLC and of retention time on HPLC. This AA metabolite: 1) does not correspond to a degradation product of PGE2; 2) crossreacts with PGE 2 antibody; 3) fits the conditions required for a 15 hydroperoxy PGE2 (chemical degradation into PGE2 by hydroperoxyde reducing reagents such as SnC12 or GSH-hemin). It is therefore inferred that depletion of GSH favours accumulation of hydroperoxyprostaglandins which would participate to oxidative injury.

The ecdysteroid receptor (EcR) and ultraspiracle (USP) from both, Chironomus tentans and Drosoplu'la melanogaster, were expressed in E. coil as fusion proteins with glutathione-S-trsusferase. The identity of the expressed recombinant proteins was confirmed by Western blot analysis. The Drosophila EcR binds to its cognate hormone response element as a heterodimer with USP as a partner. We now want to compare the capabilities of the two receptor partners from both insect species in regard to DNA binding and heterodimerization by means of gel mobility shift assays. The use of naturally occurring andartificial hormone response elements will allow to define a hormone response element consensus sequence for EcR-USP heterodimers and, if existing, for EcR and USP homodimers, respectively. For other members of this receptor family it was shown that the DNA binding domain and adjacent sequences alone sufficiently defines response element specificity for both homo-and heterodimers. The DNA binding domains of EcR and USP from the two species show a high similarity (92% and 85% identity, respectively). The corresponding sequences were selected for overexpression in E. coli. In combination with immtmoprecipimtion of receptor-DNA complexes and PCR amplification steps, the expressed proteins will be used for the detection of naturally occurring hormone response elements within genomic DNA.

University of Fribourg, CH-1700 Fribourg. Various agonist-induced cell responses in neutrophils and fibroblasts, such as chemotaxis and cytoskeletal rearrangements, have been shown to parallel the synthesis of PtdIns(3,4,5)P 3. But the importance of this rise is not clear. We show here that wortmannin blocks PDGF-induced production of PtdIns(3,4,5)P 3 in human foreskin fibroblasts with an IC50 of about 5 nM. A similar inhibition was observed in in vitro assays (ICs0=lnM) with PI 3-kinase iI~nunoprecipitated by antibodies directed against the 85 kD subunit (p85). On the other hand, wortmannin did not affect PDGF-mediated phosphorylation of p85, indicating the correct interaction of p85 with the PDGF-~ receptor. The pl10/p85 complex of the PI 3-kinase remained intact in the presence of ~M concentrations of wortmannln. These results are consistent with a direct, specific inhibition of PI 3kinase by wortmannin at concentrations relevant for its previously reported effects on cellular responses. When stimulated with PDGF, human foreskin fibroblasts form circular membrane ruffles rich in filamentous actin. The fact that wortmannin inhibits these PDGF-mediated aotin rearrangements suggests the need of PI 3-kinase activity as a signal for this cell response. Caicineurin is a calcium/calmodulin-regulated protein phosphatase which is comprised of a catalytic subunit A and regulatory subunit B. Although calcineurin was discovered in brain, the calmodulin stimulated protein phosphatase which has caicineurin like structure has been described in other tissues. The protein structure show that the calcineurin B is very conserved in different tissues and species, Using anti-calcineurin B monocicnal antiserum, the tissue extracts from rat have been explored. Immunoreactivity was obverved in all tissues, although its intensity was different. The highest intensity was found in brain. Spleen, Heart and thymus had an intermediate level intensity. The results were supported by the quantitative measurement of calcineurin. The caicineurin concentration was 10 to 50 fold higher in brain than that in other tissues. Furthermore, the distribution of the calcineurin activity in rat tissue was studied using dephosphorylaticn assay of calctneurin. The highest caicineurin activity was found in brain too, but was only 5 to 15 fold more than that in other tissues. The specific activity of calcineurin was different in tissues. These results indicate the caicineudn activity might be regulated in tissue specific fashion.

BURST AND CYTOSKELETON Arcaro A. and Wymann M.P., Institute of Biochemistry, University of Fribourg, CH-1700 FRIBOURG Chemoattractants like N-formyl-Met-Leu-Phe (fMLP) stimulate in neutrophils a rapid and transient increase in the levels of PtdIns(3,4,5)P3 (PIP3) which is due to the activation of Ptdlns(3)-kinase. PIP 3 has been proposed to be a second messenger molecule, but its cellular targets are presently unknown.Here we report that pretreatment of human neutrophils with wortmannin inhibits the fMLP-etimulated production of PIP 3 in a dose-dependent manner (IC50 = 5 nM). Furthermore, PtdIns(3)-kinase activity immunoprecipitated from resting cells is totally abolished by I0 nM wortmannin (IC50 = 1 nM). These results show a potent and direct effect of wortmannin on PtdIns(3)-kinase. Wortmannin also inhibits the respiratory burst of neutrophils at nanomolar concentrations, which suggests that PIP3 is involved in the signalling pathway controlling activation of the NADPHoxidase.When pretreated with wortmannin and stimulated with fMLP, neutrophils display oscillatory changes in Factin content that correlate with oscillations in cell ~hape. This, and the inhibition of PtdIns(3)-kinase, suggest a modulatory role for PIP 3 in cytoskeletal rearrangements.We are currently investigating the effects of PIP 3 on the functions of gelsolinl a protein that is proposed to mediate actin polymerization and can be regulated by polyphosphoinositides and Ca 2+.

Matthias Gstaiger, Walter Schaffner and Christopher Hovens Institut f'tir Molekularbiologie ILl der Universitat Z'tirich, CH 8057 Ziirich

The octamer motif ATGCAAAT plays a central role in mediating the activity of a number of both lymphoid specific and ubiquitous promoters and in the immunoglubulin heavy chain intronic enhancer. The activity of thils motif in lymphoid cells is presumably mediated by the lymphoid cellspecific factor Oct-2A. When ectopically expressed in non-lymphoid ceils, Oct-2A is not sufficient to mediate enhancer activity of a multimerized 50 bp enhancer core element derived from the immunoglobulin heavy chain intronie enhancer, a segment, which is however highly active in B-cells. This and other findings support the supposition that additional B-cell specific factor(s) can account for the specificity and extent of the Octdependent transcription observed in lymphoid cells. To further our understanding of the molecular mechanisms involved in mediating lymphoid-specific expression, we have employed the yeast two hybrid system to identify human proteins capable of physically associating with human Oet-2A. To this end, a Oct-2A expressing yeast strain has been constructed which bears two integrated reporter genes, HIS3 and lacZ under the control of the octamer motif. Transformation of this strain with a cDNA library has allowed us to isolate clones interacting with Oct-2A. We are currently analysing these candidates to determine the veracity of these interactions in higher enkaryotic cell background. In general, untransformed nuclear hormone receptors are predominantly confined to either the nucleus or the cytoplasm of a cell. Receptors of the latter class are translocated to the nucleus upon interaction with ligand. Our studies of the use of the ecdysteroid receptor complex from insects, consisting of a heterodimer formed between EcR and USP, as a tool for target gene activation in vertebrate cells have revealed a novel mechanism of nuclear translocation of EcR that does not depend on hormone action. After transient transfection of respective expression plasmids into HeLa or CV-1 cells, the subeellular distribution of the individual receptor types was analyzed by indirect immanofluorescence staining. While EcR alone was detected in both cellular compartments, nucleus and cytoplasm, USP was predominantly found in the nucleus of either host cell. Cooxpression of both EcR and USP resulted in an efficient transloeation of EcR into the nucleus. USP appears to trap EcR in the nucleus, presumably by the formation of a heterodimeric complex. Thus, heterodimer formation of EcR and USP not only appears to be required for high affinity binding to cognate hormone response dements (and subsequently for transactivation of an adjacent target gene) and for binding to ecdysteroid but in addition for an enhanced translocafion of EcR into the nucleus. Activation of rodent macrophages with cytokines or bacteria is accompanied by the induction of a Ca2+-independent nitric oxide (NO) synthase which plays a key role in antimicrobial and antitumoral activity. Firm evidence for expression of a similar enzyme in other mammals or in man has been lacking. We now show that bovine bone marrow-derived macrophages produce nitrite in an L-arginine-dependent manner upon stimulation with heat-killed grampositive or gram-negative bacteria.

Homologous interferon-7 and tumor necrosis factor-~ induced little NO 2-production, but primed macrophages for enhanced N0 2-production induced by

Salmonella dublin or LPS. This is one of the first demonstrations of NO 2-production by nonrodent macrophages and encourages a search for an involvement of reactive nitrogen species in the killing of parasites or tumor cells by macrophages from mammals other than rodents. Lectins are used in many analytical methods to study the carbohydrate structure of glycoproteins. We report here a technique which combines the high resolution of two-dimensional gel electrophoresis (2-DPAGE) to separate plasma proteins, the specificity of lectins to bind carbohydrates and the sensitivity of chemiluminescence. This method is compared with that of a commonly used colorimetrio reaction.

Since the reference plasma protein map obtained by 2-D PAGE is available (i), the technique described here allows a quick and specific identification of plasma glycoproteins. Therefore any ma j or glycosylation modification which may happen in some diseases can be detected. (i) Electrophoresis 1992; 13:707-714. This work was supported in part by FCAR (Quebec, Canada). In our laboratory we use genetic and biochemical approaches to study translation initiation, in particular the initiation factor eIF-4A in yeast. This factor from higher eukaryofic cells is known as an RNA dependent ATPase and functions as a RNA helicase together with another initiation factor, elF-4B. It belongs to a family of putative RNA helicases, the D-E-A-D proteins. To find new genes involved in translation initiation, suppressors of a temperature sensitive eIF-4A mutant were isolated. One of the suppressors (STM1) is a single copy gene which encodes a protein resembling the human translation initiation factor eIF-4B. Disruption of the gene is not lethal to the cell, but affects growth. Polysome analysis of strains with a deleted STM1 gene have shown that this new gene is also involved in translation initiation. It carries six repeated sequences of 25 amino acids of unknown function and has -like the human eIF-4B -a RNA binding domain. The second suppressor (STM2) is also a novel gene. It encodes a large protein which shows no homology to other proteins present in the data libraries. It contains high portion of charged amino acids and is rich in glutamines and serines. Its function in translation initiation is currently under investigation.

TOBACCO TRANSLATIONINI~ATIONFACTORE~4AISENCODED BY A LARGE ANDDIVERSE GENEFAMILY Karl Brander, Therese Mandel, Isabelle Lutziger, George Owttrim and Cris Kuhlemeier, Institute of Plant Physiology, University of Berne, Altenbergrain 21, CH-3013 Berne Using heterologous probes we isolated cDNAs coding for translation initiation factor eIF-4A from tobacco, eIF-4A is encoded by a large multigene family of which at least nine members are expressed in leaves and most if not all other organs of the tobacco plant. While screening genomic libraries we found several additional genes. Two of them code for genes highly homologous with eIF-4A throughout the coding region, but with changes in the GKT, PTRELA and DEAD-box motifs. A third gene is expressed exclusively in the developing male gametophyte. This eIF-4A-related gene may have a role in the well-documented regulation of translation in pollen. In order to study the function of these genes we have begun altering their expression in transgenic tobacco plants.

Seauence reauirements for a non-AUG protein initiaf~jon

Non-AUG initiation codons are still infrequent and were initially observed in animal viruses, but a number of cellular examples have now also been reported. Several years ago we reported the use of an ACG in the P/C mRNA of Sendal Virus (SEN) to initiate the non-structurar C' protein. We recently described a C' protein in the Human Parainfluenza Virus Type 1 (hPIV1), which is initiated from a GUG. In both cases, the C' protein is an N-terminally ebngated form of the C protein which is initiated at the second AUG, in the +1 reading frame relative to the first AUG. This AUG is used to initiate the P protein, an essential component of the viral RNA polymerase. Unlike the ACG in SEN, the GUG in hPlVt is clearly not a weak start codon since C' is the most abundant protein expressed from this mRNA both in rive and in vitro. It appears that not any upstream GUG in an optimal context will function as a start codon. For instance, the P gene of hPIV3 cannot express a C' protein (the ORF is closed by stop codons) but it does contain a GUG in an optimal context upstream of the P protein AUG, which could encode a P' protein. Nevertheless, we have been unable to demonstrate that this GUG functions as a start cedon. With the use of chimeric mRNAs between hPIVl and 3, we have shown that positions +5 and +6 (the first G of the GUG triplet being +1) are the major determinants of the efficiency of GUG as an initiaton codon for protein synthesis. We are currently investigating the possibility that these nucleotides may increase the initiation rate of weak non-AUGs such as the ACG of SEN.

Biological activity of the heterodimeric subunit SRP9/14 of the Signal Recognition Particle is retained in 2 fusion proteins Fabrice Bovia & Katharina Strub, D6partement de Biologie Cellulaire, Universit6 de Gen~ve, CH-1211 Gen~ve 4

The targeting of secretory proteins to the membrane of the endoplasmie reticulum is mediated by a cytoplasmic ribonucleoparticle, the signal recognition particle. It is composed of 6 proteins and f RNA molecule (SRP RNA). The 2 smallest proteins SRP9 and SRP14 are required for the translational control function of SRP. It effects a specific arrest in the biosynthesis of ER-targeted proteins. These 2 polypeptides bind to the SRP RNA exclusively as a heterodimer (SRP9/14). We have constructed single-chain variants of this dimer using a short peptide linker of 17 aa to connect the C-terminus of the 14 kD subunit to the N-terminus of the 9 kD subunit (14-9) and vice versa (9-14). We found that both proteins bind SRP RNA with a similar efficiency as the wild type protein.

Glutaraldehyde cross-linking experiments and sedimentation analysis indicate that the 2 fusion proteins fold into a similar structure as SRP9/14 and bind SRP RNA as a monomer. Furthermore, they can functionally replace SRP9/14 in the elongation arrest and translocation assay. Since the 2 fusion proteins are circular permutations of each other, we can conclude from this results that N-and C-termini of both proteins have no essential role in folding, RNA-binding and in mediating their biological activity. Furthermore, the possibility to express an oligomeric protein as a single polypeptide facilitates the analysis of its functions as well as its structure.

A24 S06-08

Studer, R. and Hunziker, P. E., 8iochemisches Institut der Universit~t Z0rich, CH-8057 Z0rich

Metallothioneins (MTs) are small Cys-rich proteins which are inducible by a variety of agents such as bivalent transition-metal ions, tumor promoters, cytokines and hormones. The biological role of MTs is still unclear. Initially postulated to serve in the sequestration of the toxic heavy metals such as cadmium and mercury, they are now believed to play a major role in the cellular metabolism of essential metals such as zinc and copper. To explore this function further we have now established a method to quantify basal MT production in several human cell lines. Thus, cellular isoMT contents were monitored by h.pJ.c, and the rate of synthesis was followed by the incorporation of ~SS-cysteine. The results show that maximum MT synthesis and accumulation occurs when the cells enter the exponential growth phase. With increasing celldensity the MT content decreases progressively until confluence is reached. In correlation with preliminary measurements our results suggest a close relationship between the MT content, the cell growth rate and the intracellular abundance of zinc.

Vincent Bernard, Adrian Etter, Heinz Tobler and Fritz Mtiller. Institute of Zoology, University of Fribourg, 1700 Fribourg (Switzerland). The nematode Ascaris lumbricoides undergoes chromatin diminution which causes a loss of DNA in all somatic ceils. We have identified a ribosomal protein (rp) gene (coding for ALEP1 = Ascaris lumbrieoides eliminated protein 1) which is located within the germqine specific material. ALEP1 shows strong homologies with the ribosomal protein S19 (rpS19). The transcription of its gene takes place only in the germ-line. 2D-gel electrophoresis on germ-line and somatic purified ribosome fractions shows that the ALEP1 protein (rpS19) is present only in the germ-line ribosome.

Does the ALEP1 germ-line specific protein have a homologue in the somatic ribosomes? We have cloned a homologue of ALEP1 from Ascaris somatic tissue named rpS19', We are currently studying the expression pattern of this rpS 19'. These results indicate the existence of a developmentally regulated ribosome heterogeneity between the germ-line and somatic ceils.

What is the function of rpS 19? Since Ascaris is not suitable for genetic analyses, we isolated rpS19 from the nematode Caenorhabditis elegans. The C. elegans rpS19 cDNA shows only 60% homology at the nucleic acid level. RpS19 is located on the C. elegans chromosome IV. The identification of a rpS19 mutant in C. elegans will indicate its function.

Favre, D, l, Pavlovic, j2 and Michel, M.R. t 1 Institute of Medical Microbiology, University of Beme, CH-3010 Berne. 2 Institute for Immunology and Virology, CH-8006 Ziirich

We have successfully purified the recombinant C (recC) protein from Spodoptera frngiperda (Sf9) insect cells which were infected with AcNPV-C (A). The recC protein retained the biological activities of native C protein obtained from wild type SFV (B). Our results suggest that the C protein is responsible for the cellular shut-off of host protein synthesis by inhibiting the initiation of translation. We are currently investigating the molecular mechanisms involved in this shut-off. Tif3 is a new eukaryotic initiation factor which shows 26% identity with the sequence of mammalian translation initiation factor eIF-4B. The TIF3 gene is not essential for growth; however, its disruption results in a slow growth and coldsensitive phenotype. In vitro translation of total yeast RNA in an extract from a TIF3 gene-disrupted strain is reduced as compared to a wild-type extract.

The translational defect is more pronounced at lower temperatures and can be corrected by the addition of purified yeast Tif3,or mammalian eIF-4B.

In vivo translation of ~galactosidase reporter mRNAs with varying degree of RNA secondary structure in the 5' leader region in a TIF3 gene-disrupted strain show preferential inhibition of translation of mRNA with more stable secondary structure. Chloroplast protein synthesis requires the import of elongation factors EF-Tu (tuf-gene) and EF-G (fusgene). It seems that the expression of both genes is light regulated. We recently cloned and sequenced a chloroplast specific nuclear fus gene in soybean (Biochim.Biophys Acta 1174 : 191-194, 1993 . Further analysis revealed that a second very similar (sequence, anatomy) nuclear fus-2 gene exists which, however, shows strong sequence diversion in the 3'trailer region. Several cDNA clones were partially sequenced but sofar only transcripts from the fusi gene were retrieved.

We currently study the promoter regions of either gene to test possible differential expression of fus genes. Comparative mapping and sequencing studies of the two genes in the 3"-region indicate that a major DNA insertion occurred distal to the coding part of both fus genes what may also influence gene expression.

THE RNA-BINDING DOMAINS IN THE 9KD AND 14KD SUBUNITS OF THE SIGNAL RECOGNITION PART~CLEHAVEA pUTATIVE~-H~LICALSTRUCTURE. Nazarena Buiand Katharina Strub Dpt de Biologie cellulalre, Science Ill, Universit4Gen~ve

The signal recognition particle (SR2), a cytoplasmic ribonucleoprotein is important for targeting nascent secretory proteins to the endoplasmic reticulum. SRP9and SRPl4are constituents of SRPandare, together with the Alu sequences of SRP RNA, required for the elongation arrest activity of the particle. SRP9 and SRPI4 form a heterodimer in the absence of the RNA and bind the RNAexclusively as a heterodimer. The primary sequence of SRP9 and SRPI4 revealed that both proteins are highly basic and lack structural similarities to other characterized RNA-binding proteins. We have therefore been interested in characterizing the molecular nature of RNA-protein interactions.

The analysis of N-and C-terminal deletion mutants of the proteins have shown that both proteins contribute to the formation of an RNAbinding pocket, The RNA-binding and dimerization domains in both proteins, are found in regions that have a predicted ~~helical structure. In SP~PI4, this u-helical region is located betwen aa 72and I00 at the C-terminus; it contains the RNA binding and dimerization domains adjacent to each other. Our results support the model that the hydrophobic face of the putative u-helix is implicatedinprotein-protein interactions and the positively charged face in RNA-protein interaction. In SP~Pg, the two RNA-binding domains found N-and C~ terminally, also have a predicted amphipatic s-helical structure. It has been proposed that the decline in protein synthesis observed in aging organisms may result from a decrease in the protein synthesis elongation factor EF-la. John Shepherd's finding that transgenic Drosophila melanogaster carrying an additional copy of the EF-la gene have an extended lifespan further indicated that the EF-la gene may play an important role in determining longevity (Shepherd et al., 1989) . In order to test this hypothesis, we have quantitated EF-la mRNA, EF-la protein, as well as catalytic EF-la activity in John's flies. Young and aging EF-hx transgenic (EF) and control lines ((2) were examined. Because the the additional EF-lc~ copy is under the control of the hsp70 promoter, flies were aged either at 25~ or at 29~ The results show that transgenic EF flies do not express more EF-lct than the control flies, neither at the RNA nor at the protein level. The question arises, whether the lransgene can be induced at all. This was tested by doing RNase protection assays. Transgenlc message can be detected only in EF flies heat shocked at 37~ but not in EF flies grown at 29~ nor in control flies at 37~ From this experiment we conclude that the transgene is indeed inducible, but is not expressed in detectable amounts at 29~

We conclude that the difference in the lifespan does not result from the overexpression of the EF-la transgene. Phenobarbital (PB) induces the expression of liver enzymes involved in the metabolism of drugs and other foreign compounds (xenobiotics). These enzymes include several cytochromes P450, NADPH:P450 reductase, aldeyde dehydrogenase, epoxide hydrolase, UDP-glucoronyltransferases and glutathione-S-transferases. Many other proteins not yet recognized may however be induced by PB. Our first approach therefore is aimed at identifying gene products which respond to PB treatment. As model systems we are using either chicken embryos (induction in ovo) or primary cultures of chicken embryo hepatocytes (induction in culture) (Althaus et al., JBC 254 pp 2148 , 1979 .

To differentially display mRNA of induced vs non-induced cells we are using a RT-PCR technique with sets of random primers. Electrophoretic separation of amplification products allows to identify mRNA species which are induced (or decreased) following treatment. Data obtained from these experiments and sequence information revealed from extracted PCR products indicate that PB treatment results in the transcriptional avtivation (or repression) of a variety of so far unknown genes.

ANTITHROMBIN -BINDING HEPARAN SULFATES FROM OVARIAN GRANULOSA CELLS Hosseini G., de Agostini, A., Clinique de St6rilit6, HCUG, Gen~ve. At the time of ovulation, several serine proteases of the plasminogen activator and coagulation cascades are activated in the ovary. These proteolytic activities are tightly controlled in time and space, and the heparin-activated serpins plaminogen-aetivator-inhibitor-1, protease nexin-1 and antithrombin (AT) are present in the follicular environment. We have observed that granulosa cells produce a subset of heparan sulfate proteoglycans that bind and activate AT. These AT-binding heparan sulfates (aHSPGs) endow the vascular endothelium with antithrombotic properties. We are now examining the role of aHSPGs in the extravascular compartment formed by the ovarian follicle. Using 125I-AT binding assays, we have detected aHSPGs on granulosa cell surfaces and culture media, in amounts comparable to those found for endothelial ceils. After 48h of stimulation by the gonadotropin FSH (1-100 ng/ml) granulosa cells increased the amounts of aHSPGs they released in culture media by 2-6-fold, while keeping their cell surfaceassociated aHSPGs constant. Analysis of 35S-labelled HSPGs revealed that heparan sulfates constitute about 40 % of the surface-associated and 10% of the soluble granulosa cell glycosamineglyeans. Finally, using 125I-AT autoradiography on rat ovary cryosections, we have localized aHSPGs on the granulosa cell layers of large antral follicles, while aHSPGs could not be detected in smaller follicles. These data suggest that granulosa cell aHSPGs could be critically located to modulate proteolytic activities in the changing environment of the growing follicle.

Division, CIBABasel, Switzerland. The development of chelators which can mobilise excess storage iron (Fe) and permit its excretion is necessary in the treatment of Fe overloaded diseases. Although oral administration of such a pharmaceutical is highly desirable, no orally active Fe chelator is so far in regular clinical use. Searches for orally active and safe Fe chelators require appropriate animal models of Fe overload in order to evaluate the potential of these drugs.

An animal model is particularly useful for testing the capacity of new Fe chelating ce~oounds in enhancing the mobilisation of clinically relevant storage iron. Iron loading in animals was achieved by dietary supplementation with carbonyl Fe, Fe fumarate and 3,5,5-trimethylhexanoyl ferrocene (TMH-ferrocene). Liver Fe concentrations were increased 6-21 times above normal levels. TMH-ferrocene was the most efficient compound to induce stable liver Fe overload. The orally given con'pounds induced a predominantly hepatocellular iron distribution and was found distributed among reticuloendothelial cells only at a later stage. This pattern of liver Fe storage is similar to that observed in the early stages of primazy ha6~ochromatosis and late stage of transfusional hae~ochromatosis. Eeeently, we and ethers have cloned and characterized novel transcription factors called peroxisome proliferator-actirated receptors (PPAK). They belong to the superfamily of nuclear hormone receptors as the steroid, thyroid and retinoid hormone receptors. PPARs are activated, beside xenobiotic peroxisome proliferators, by natural fatty acids and induce, together with the receptor for 9-cis retinoic acld (RXR), the transcription of genes involved in the ~-and ~-oxidation of fatty acids. PPAR and RXR bind as heterodimers to the PPAR response element (PPRE), consisting of a direct repeat of AGGTCA-Iike sequences with one intervening nucleotide. Now, we show by gel retardation analysis and transient transfection assays that PPAR/RXR heteredimers also bind to and transactivate gene transcription through estrogen response elements (ERE). The ERE is a palindrome with AGGTCA half-sites separated by three nucleotides. The selectivity and specificity of the transactivation through an ERE was demonstrated by the testing of various related palindromic and inverted palindromic repeats of AGGTCA sequences, which were all inactive. The possible activation of estrogen responsive genes by fatty acids via PPARs in vivo is discussed especially with regard to a debated role of fatty acids in breast cancer. The purpose of this study was to evaluate whether intermittent doxorubicin (DOX) treatment in vitro selects for multidrug resistance in RPMI 8226 myeloma cells and, if so, to determine the underlying mechanism(s). Cells were exposed to constant doses of DOX for 4 days alternating with growth in DOX-free medium for 17 days. After > 9 months of treatment, the cells developed an atypical MDR phenotype with 4-to 6-fold resistance to topo II poisons. Cross-resistance to vincristine and taxotere was < 2-fold, sensitivity to cisplatin and melphalan remained normal. Efflux of MDR1 drugs was incresed with no effect of verapamit; subcellular DOX distribution was normal. Expression of topo II a was found to be reduced by around 50 and 60-70% at the RNA and protein level, respectively. Minimum MDRt RNA overexpression was detected when using RT-PCR; P-glycoprotein was not detectable. MRP RNA expression was increased by 60-90% relative to the wild type cells without detectable p190. This data suggests that atypical MDR might play a role in acquired DOX resistance in human myeloma. Mucosal surfaces cover a vast area in the human body. They satisfy its need to communicate with its environment, e.g. in terms of nutrient uptake and sexual reproduction. However, at the same time they offer a vulnerable gate for invasion by pathogenic microorganisms. For immunoprotection of these regions, large amounts of IgA antibodies are produced in the bloodstream and transported across the epithelial barrier to the lumen by means of polymeric Ig receptor. There the receptor is cleaved, and its extracellular portion remains associated as secretory component (SC) to the secretory IgA complex.

In the context of a project aiming to mass production of secretory IgA as a passive vaccine, we evaluated a series of eukaryotic expression systems for production of SC. The cDNA encoding polymeric Ig receptor was engineered in a way that only its extracellular domains corresponding to SC were expressed in yeast, in insect cells infected with recombinant baculovirus, and in mammalian cells infected with recombinant vaccinia virus. Furthermore, a C-terminal poly-histidine tag was introduced for simple purification of the products. We will show optimization of expression levels for each of the systems as well as a quantitative and qualitative comparison of the products.

Recombinant vaccinia viruses (rVV) that express gone products from other organisms has become a popular and widely used laboratory expression system. Our current interest is directed toward the production of secretory component (SC), a subunit of secretory IgA antibody complex involved in mucosal immunity. Toward this goal, we constructed rVV governing SC transcription under the control of two viral-based and the T7 bacteriophage promoters. Expression of the protein was assessed in several mammalian cell lines, such as human TK-, human HeLa (grown in suspension or as a monolayer) and monkey CV-I. Optimization of infection parameters and culture conditions were also performed.

Institut de Biologic animale de rUniversit~ de Lausanne

Protection of mucosal surfaces is mediated by secretory IgA ant/bodies (algA) constituted of dimeric IgA and secretory component (SC) . Toward the aim of reconstituting slgA complexes in vitro and using them for passive oral immunization, SC has been produced in yeast, insect and mammalian expression systems. The gone encoding SC has been engineered so that a) the C-terminus of the protein carries a 6xHis tag and b) the newly synthesized polypeptide is released in the culture medium. Culture conditions have been established to permit recovery of SC in serum-free medium. Purification procedures based on nickel chelate and classical chromatographies have been optimized to yield SC under native conditions. Finally, in vitro reassociation with purified dimeric IgAs obtained from hybddomas has been used to assess biological activity of recombinant SC. Using a Magnet-Assisted Subtraction-hybridisation Technique (MAST), 17 cDNA clones were isolated that are expressed in nQrm~l l~nq %issue but n__D_i expressed in the corresoondina NSCLC tissue, ii of the 17 eDNA clones are highly homologous to eDNA sequences for the following proteins: Pulmonary surfactant proteins SP-A and SP-B; Receptor for Advanced Glycosylated Endproducts of proteins (RAGE); Calmodulin-like protein; Natural killer gone 5 (NKG5); Matrix Gla protein (MGP); Glutamine synthetase; ubiquitin; Vascular smooth muscle alpha-actin; Cytoskeletal beta-actin; Vimentin.

The remaining 6 eDNA clones may represent parts of still unknown genes. The MAST eDNA clones were derived from a patient who developed squamous adenocarcinoma. Northern blot analysis of normal/tumour RNA from three other NSCLC patients showed that the lack of gone expression was independent of NSCLC type and stage.

Enzymatic methylation of cytosines in mammalian DNA is believed to play a fundamental role in basic gene regulation. Methylation in the vertebrate genome is restricted to CpG dinucleotides at the C-5 position of cytosine. In vitro studies have shown that methylation can drastically influence the DNA structure. CpG islands remain unmethylated in normal cells, whereas several immortalised, transformed cell lines contain partially methylated CpG islands. Transcription of genes can be inhibited by methylation of CpG's. This modification contributes e.g. to the stable repression of genes on the inactivated X chromosome of females. The collagen VI genes show typical CpG islands in their promoter region and they were shown to be down regulated at the transcdptional level in src or myc transformed cells. Given these facts we dee~ded to investigate the effects of methylation in the collagen VI gwene. e could demonstrate that the promoter activity of (x2(VI) chicken collagen gone was strongly diminished by methylation in vitro. Furthermore DNA methylation completely prevented binding of a novel transcription factor to the promoter, whereas binding of SP1 was not affected. To show evidence of methylation in vivo, we are currently Iocalising the exact positions of methylated CpG's in normal and transformed cells by genomic sequencing.

ALTERATIONS OF PUTATIVE TUMOUR SUPPRESSOR GENES IN HUMAN NON-SMALL CELL LUNG CARCINOMA (NSCLC). R. Shipman and C.U. Ludwig, Molecular Oncology, Lab 405 (ZLF), Basel, CH-4031 Chromosomal subregion llp13 displayed non-random allelic loss in 61 NSCLCs. llp13 contains the genetic loci for catalase (CAT), the Wilms' tumour gene (WTI) and the follicle-stimulating hormone ~-chain (FSH~). 76% of the NSCLCs were deleted at CAT suggesting that loss of a gene(s) near CAT may be involved in the progression of NSCLC. YAC clones containing the CAT locus are being prepared for use in a direct selection procedure for cDNAs encoded at this region. Although 38% of the NSCLCs were deleted at WTI, no mutations were detected in the remaining WTI allele. WTI expression was subsequently demonstrated in fetal lung, fetal hematopoietic tissue, normal adult lung and NSCLC. Allelic deletion at 17p13, immuno-histochemical and sequence analysis of the p53 gene were also examined in our NSCLCs. These analyses support the contention that aberrant expression of the p53 gene is a pivotal event in the progression of NSCLC.

$08-11 Recently, a number of advances have been made which enable the experimentalist to study the effects of low levels of ionizing radiation. The results suggest a threshold to radiation damage exists and that above a critical level, recovery mechanisms in the cell are induced and the biological response falls. Thus low levels of radiation cause greater than expected levels of biological response. This has been demonstrated in studies of mutations, cell survival and cancer induction. Because the low dose-rates and the low doses per fraction result in the total doses being small, the magnitude of the biological effects induced is insufficient to warrant alarm.

Identification of genes associated with the revertion of the malignant phenotype of a rhabdomyosarcoma cell line. Michele Genini, Sabina Solinas Toldo*, Ruedi Fries* and Beat W. Schafer, Dept. of pediatrics, University of Ziirich, Steinwiesstr. 75, 8032 Ziirich, and *Institut for Nutztierwissenschaften, ETH Ztifich, 8092 Ztirich.

The human rhabdomyosarcoma cell line RD is tumorigenic and differentiates very poorly despite the expression of the myogenic factors myf3 and myf4. Therefore it can be regarded as natural mutant. To identify genes which are involved in suppression of tumorigenicity or enhance differentiation we applied two different approaches. Since the development of human embryonal rhabdomyosarcoma has been associated with abnormalities of chromosome 11 band p15, in the first approach we transferred a normal human chromosome 11 into RD ceils via microcell fusion. The microcell hybrids did not show a higher degree of differentiation. Suppression of tumorigenicity was observed in one clone but is probably independent of chromosome 11, as it was shown by chromosome 11 specific painting. ' In the second approach we applied a subtractive hybridization procedure between primary myoblasts and RD cells to find genes specific for the normal phenotype. These genes will be tested in vivo for induction of differentiation and/or tumorsuppression.

Deficient synthesis of the GPI anchor is the molecular basis of idiopathic paroxysmal nocturnal hemoglobinuria ("PNH") and a similar phenotype can accompany aplastie anemia CPNH-Iike"). The population expressing CD48, a GPI-anehored membrane glycoprotein, was quantified by flow cytometry in samples from patient P.G. (PNH) and patient M.M. (PNH-like) and amounted to 90% (P.G.) and 78% (M.M.), respectively. T lymphocytes from both patients were isolated, expanded and cloned. From both patients, clones expressing the CD48 marker (CD48+) and CD48 deficient clones (CD48-) were obtained. Takeda et al. (Cell 73, 1993, 703-11) showed that in some (CD59-) Bcell clones, a deletion of the P1G-A gene product is responsible for the expression of the PNH-phenotype. PCR analysis of T cell mRNA from extracts of both CD48+ and CD48-clones of patient M.M. (PNH-like) using PIG-A specific primers revealed a band pattern from which can he concluded that no deletion is present in the PIG-A gene product. This result suggests that CD48-T lymphocytes of the PNH-like patient express a normal PIG-A gene product; a disorder in regulation of PIG-A expression or an unrelated biosynthetic defect may explain the CD48phenotype in these T cells. Further work is aimed at defining the differences in biosynthetic defects of PNH and PNH-like CD48-T cell clones. Supported by grant 31-30757-91 of the SNSF to EGB. W@ have recently employed an experimental approach to turn the function of "nuclear messengers" such as c-Fos (and c-Jun) on and off at will in polarized me/nmary epithelial cells. To this end we used constructs encoding the c-fos (and c-jun) genes fused to the hormone-binding domain of the human estrogen receptor, designated c-FosER (and c-JunER), We could show that short-term activation (30 mins.) of c-FosER by estradiole (E2) led to the disruption of epithelial cell polarity within 24 hours, as characterized by the expression of apical and basolateral marker proteins. During the next 3-5 days, however, the cells fully regained their polarized phenotype.

Conversely, long-term activation (24-48 hours) caused the irreversible loss of epithelial cell polarity, leading ultimately to an epithelial-fibroblastoid cell transition. These cells underwent dramatic changes in gene expression including the complete loss or reduction of epithelial markers such as E-cadherin/uvomorulin, ZO-I and cytokeratins, and the de novo expression of mesenchymal proteins like vimentin, type I collagen and fibronectin. Expression of the v-Ha-ras oncogene (acting upstream of C-FOS) did not markedly influence epithelial cell organization when the cells were grown on plastic. However, when cultured within type I collagen gels in the presence of serum, or when injected into the mammary glands of nude mice, rapid cell growth and a clear epithelial-fibreblastoid cell transition was observed.

Growth properties determine the organ colonization ability of a metastatic murine melanoma cell line.

The ability of metastatic tumor cells to specifically colonize distant organ sites strongly depends on the interaction of the metastatic cells with the local organ environment. We have selected a B16 routine melanoma cell fine (B16-LS9) which has an increased ability to colonize the liver, and compared its behavior with either its parental (B16-F1) or inng-specific (B16-F10) cell line. We have found that adhesion in vitro and organ lodgement in vivo were not different for LS9 and F10. In contrast, B16-LS9 grew at slower rates than the other lines both in vitro and in vivo after subcutaneous or intra-foot-pad injection. Analysis of tyrosine-phosphorylated proteins indicated a higher phosphorylation activity in B16-LS9 cells, which might suggest an altered regulation of some kinase(s) in this cell line. Intercellular contact with hepatocytes in vitro, or the liver in vivo could restore an efficient growth of B16-LS9, enabling thus these cells to colonize this organ much better than others. Hepatocytes or liver plasma membrane (p.m.) extracts conserved the growth stimulatory activity towards B16-LS9 ceils. Chromatographic fractionations of these extracts showed that a major growth sfimulatory protein in this fiver p.m. fraction turns out to be ttansferrin, which in the liver appears to exist in at least two different mature forms. Protein import into mitochondria requires ATP in two locations: outside the organelle, and in the matrix space. Depletion of matrix ATP arrests translocation of precursors across the inner membrane, but does not prevent precursors from crossing the outer membrane. As a result, proteins that reach the inner membrane or intermembrane space without passing through the matrix are often sorted normally in ATP-depleted mitochondria. External ATP is used to maintain precursors in a translocationcompetent state. However, some precursors can fold and still remain translocation-competent, and import of these precursors is independent of external ATP. With the folded cytochrome b2 precursor, it is matrix ATP rather than external ATP that drives translocation across the outer membrane. Thus matrix ATP generates a pulling force that can drive both the translocation and the unfolding of precursor proteins. Mitochondrial hsp70 probably functions as the ATP-dependent import motor.

Arsenijcvic D, Dulloo AG & Girardicr L, Dpt of Physiology, CMU. Geneva, CH. The relationship between body weight, daily energy expenditure (assessed by indirect calorimetry), and immune status (determined by lymphocyte proliferation tests) was investigated in Swiss Webster mice maintaining a stable body weight several weeks after infection with Toxoplasma gondii(Me49). Following an initial period of weight loss (infection-induced cachexia), the surviving mice could be differentiated into two main groups: (i) those showing a partial regain in body weight (the infected gainers) and eventually stabilizing al a mean body weight 25% below a non-infected control group, and (if) those showing no weight regain (the infected non-gainers) and eventually stabilizing at 45% below control levels. Immune function was found to be markedly suppressed in the infected non-gainers, whereas it was normal (and similar to control levels) in the group of infecled gainers. Daily energy expenditure (and food intake), after normalisation for differences in body weight, was found to be the same in both infected gainers and non-gainers, but lower by 15% when compared to controls (P<0.01); thereby indicating that the infected groups were able to increase their overall efficiency of food utilization in response to the low food inlake. In conclusion, this study conducted during a weight stable phase of chronic infection shows a clear-cut association between the inability to regain body weight and impaired immune function.

In contrast, no relationship was found between metabolic rate and body weight nor between metabolic tale and immune function, but the chronically infected mice exhibited the same phenomenon of enhanced metabolic efficiency characteristic of chronic underfeeding per se.

Identification and functional analysis of chaperonin 10, the groES homolog of yeast mitochondria.

Biozentrum, University of Basel, Klingelbergstrasse 70, CH-4056 Basel, Switzerland. Phone ++41-61-2672171, FAX ++41-61-2672148.

The mitochondrial chaperonin system consists of chaperonin 60 (cpn60; also termed hsp60), which is homologous to E. cull gruEL, and chaperonin 10 (cpnlO), which is homologous to E. coil groES. We have used a functional assay to identify cpnl0 of yeast mitochondria. When dimeric Rubisco is denatured and allowed to bind to yeast cpn60, subsequent refolding of the enzyme is strictly dependent upon yeast cpnl0. In the presence of MgATP, yeast cpn60 and yeast epnl0 form a stable complex that can be isolated by gel filtration. The ATPase activity of hsp60 is not inhibited by complex formation with cpnl0. A different result is obtained with the E, coil system, where groES is able to completely inhibit the ATPase activity of gruEL (1). To test the in vivo role of cpnl0, we have cloned, sequenced and disrupted the corresponding nuclear gene CPNIO. This gene encodes a protein of 11,372 Da that is imported into the mitochondrial matrix without detectable cleavage. Haploid cells lacking a functional copy of CPNIO fail to grow at temperatures between 23 ~ C and 37 ~ C (2).

(1) Rospert, S. et al. (1993) Proc. Natl. Acad. Sci. 90, in press.

(2) Rospert, S. et al. (1993) FEBS Lett., in press.

Characterization of yeast mitochondrial heat shock protein 70 , L. Bolliger, B. S. Glick and G. Schatz, Biocenter, University of Basel, 4056 Basel. Mitochondrial hsp70 (mhsp70) is thought to use the energy of ATP hydrolysis to pull precursor proteins across the mitochondrial membranes. To facilitate the purification of yeast mhsp70, we used the cloned gene (a gift of Dr. N. Morishima) to modify mhsp70 with a six-histidine tag at its C-terminus. This construct supports growth of yeast cells lacking wild-type mhsp70. We developed a purification procedure that allows us to recover this tagged protein with a purity of about 95%. Experiments are in progress to characterize the nucleotidedependent interaction of purified mhsp70 with unfolded proteins. In addition, we are looking for partner proteins that may modulate the action of mhspT0. We have copurified a protein of about 20 kD together with mhsp70. The 20 kD protein could be released from mhsp70 by ATP or ADP. When tryptic fragments of this protein were subjected to microsequencing, one fragment showed strong homology to the GrpE protein of E.coli. The transmembrane electrical potential of mitochondria (AWmit) can be assessed in single hepatocytes by using epifluorescence microscopy. For this end the rhodamine 123 fluorescence of a mitochondria-rich (Fr) and a mitochondria-poor region (Fp) are measured. The ratio of these signals depends on the AWmit according to a modified Nernst equation (Reber B .F.X., Somogyi R. & Stucki J.W. (1990) , BBA, 1018, 190-193) . To calibrate this method the cell membrane was permeabifized for monovalent cations with nystatin and gramicidin. Then the cytosolie K + was replaced by Na + from a K+-free bath solution. Subsequent addition of valinomycin made the mitochondrial membrane permeable for K + ions. By the addition of different K + concentrations to the bath, the AWmit could be clamped to defined values. The relation between the ratio Fr/Fp and the AWmit could be well fitted to our modified Nernst equation. However, the calibration curve flattens out at potentials higher than about 110 inV. Therefore the detection limit of changes of A~mit within the physiological range is about 10-20mV. Maximal stimulation of the Na+/K + ATPase by addition of nystatin to the Na + containing bath medium caused a drop of AWmit of about 20inV. Subsequent addition of ouabain resulted in an almost complete reversal of this drop. This shows that changes in ATP consumption can be assessed via changes of AUdmit. Schneiter Ph., Di Vetta V., J6quier E. and Tappy L. Institut de physiologie de l'UniversitG Bugnon 7,1005 Lausanne. The metabolic effects of an exercise performed in the fasting or fed state were studied in 8 subjects over a 8 hour period in a respiratory chamber. A mixed meal containing 13C labeled carbohydrates was ingested at 9:30 am and an exercise session (walking at 5 km/h 45 rain, slope 10 %) was performed at 8:15 am (fasting) or 11 am (postprandial). Net carbohydrate (net CHO ox) and lipid oxidation (indirect calorimetry) and oxidation of exogenous carbohydrates (exo CHO ox, breath 13CO2) were measured. Glycogen breakdown was calculated as net CHO ox -CHO exo ox, and glycogen synthesis as CHO in -CHO exo ox. Energy expenditure over 8 hours was similar but net CHO ox was 16 % lower, lipid ox 61% higher and exo CHO ox 39 % lower when exercise was performed in fasting vs fed state; moreover, glycogen synthesis was increased by 40 % (p<0.0001), while glycogen breakdown was increased by 12 % (p=0.06).

In conclusion, a 45 min exercise in the fasting vs fed state a) increases utilization of endogenous lipids, b) enhances muscle glycogen turnover over a 8 hour period.

Hormonal regulation of e-Abl tyroslne kinase by fusion to a steroid binding domain T. Mattioni, O. 8chini Hooft van Huijsduijnen, P.K. Jackson and D. Picard D4partement de Bielogie Cellulaire, Universit4 de Geneve, CH-1211 Geneve 4

The activities of a wide variety of proteins can be subjected to hormonal control by fusion to the hormone binding domain of steroid receptors. We show that this strategy can also be applied to a tyrosine kinase. In particular, transforming derivatives of c-Abl become hormone-inducible oncogenes by fusion to a steroid binding domain. It is noteworthy that the hormone-inducible transformation of NIH3T3 cells is completely reversible upon removal of the specific Ugand. Reversion experiments reveal another facet of Abl: its ability to inhibit cell proliferation. 48 hours after hormone depletion, cells are not only morphologically reverted, but the cell cycle appears to be blocked in G1. A cytostatic effect of Abl overexpression has been reported by several groups but, until today, it could only be examined by comparing different Abl derivatives or the same derivatives in different cellular contexts. In contrast, our hormone-regulable system has the advantage that the two distinct functions of Abl (transformation versus growth inhibition) can be studied using the same derivative expresed in the same cell line. We are using the hormone inducible system to investigate the mechanisms regulating the transtorming and the cytostatic effects of Abl in a variety of cell lines.

Jiewu YANG and Herbert Zuber Institute for Molecularbiology and Biophysic, ETH, 8093 Z'tirich Enzymes from thermophilic organisms show higher thermostability and temperature optima than those of mesophilic organisms. It has been shown that these properties are based on a specific structure of the protein. The primary structure of thermophilic (B. stearothermophilus) and mesophilic(B, megaterium) triosephosphate isomerase(TIM) have been determined by cloning and sequencing of the TIM genes. Comparison of the primary structures of the thermophilic and rnesophilic enzyme showed specific amino acid substitutions, particularly in the c~-helices of the TIM barrel, which could play a critical role with respect to thermostability. We have consmacted mutants by exchanging (x-helices between TIM of B. stearothermophilus and B. megaterium. Helix H1, H2 and H7 of TIM from B. megaterium, corresponding to amino acid residues 13-36, 44-55. and 220-227 respectively, have been exchanged. The properties (thermostability and temperature optimum) of the hybrid-mutants were compared with the wild type enzymes. In a comparative investigation of structural and functional parameters related to the oxidative substrate metabolism in skeletal muscle cells of dogs and goats, a close relationship was found between intracellular lipid stores and mitochondria. As revealed in serial sections with subsequent EM analysis, each lipid droplet was in direct contact to one or several mitochondria. Quantitatively, 23% of the lipid droplet surface in goats and 40% in dogs were in direct contact to outer mitochondrial membranes or -in other terms -1% of outer mitochondrial membranes in goats and 2% in dogs were in direct contact to lipid deposits. These findings were in good agreement with the physiological data showing a 2 times higher oxidation rate of fatty acids drawn from intracellular stores for dogs than for goats.

Human muscle adapts to altered load with changes in the expression of enzymes involved in energy metabolism. Endurance exercise in humans is known to increase the oxidative capacity of skeletal muscles (Hoppeler, Int. J. Sports Med. (1986), 7, 187) . At the ultrastructurai level, a higher mitochondriai content contributes to the higher oxidative capacity of skeletal muscles. Little is known about the molecular mechanisms that lead to such adaptations in humans. The expression of mitochondriai components involves complex regulation of both mitochondriai and nuclear encoded genes. We have developed a PCR approach to quantify DNA and RNA from human skeletal muscle cryostat sections and addressed the question, whether the structural adaptations in endurance trained athletes are accompanied by concomitant changes in respective RNA steady state levels. Preliminary data indicate a higher concentration of COXIV and COXl (1.8 and 1.6 times, respectively) in trained subjects. No differences were observed for mtDNA, despite a two fold higher mitochonddal content. Human skeletal muscle contains three main fiber types which are based on the myosin heavy chain composition of the individual fiber, The isoforms of other contractile proteins within a given fiber type can vary leading to a continuum of different phenotypes. We are studying the fiber type specific expression of the mRNAs of myosin alkali light clmins (MLC) using in situ hybridization, especially the rarer slow form MLC lsa. Differences were found between shoulder muscles and leg muscles: m. deltoideus yielded stronger signals for MLC lsa and MLC lsb in the least glycolytic type I fibers (IA), weak but detectable signals for MLC lsb and MLC lf/f mRNA in more glycolytic type I fibers (IB) and strong signals for MLC lf/3f in type II fibers. In m. vast. lat. MLC lsb was strongly expressed in type IA as well as IB fibers, MLC lsa mRNA was only found in some IA fibers. MLc Isa has been shown to be expressed at a particular time point during muscle development and regeneration. In a biopsy of a Becker Dystrophy patient, we found very strong signals in some small fibers displaying the morphology of regeneralmg fibers. Thus this probe could be useful to dete~t regenerative processes in pathological muscle samples. The aim of this study was to validate the complementarity of two new non-invasive methods for the assessment of autonomic regulations. Reactions to a moderale exercise (75 W on a cycloergometer) were compared in cardiac transplant recipients (CTR) who modulate their heart rate (HR) by means of circulating calecholarnines, their heart being donervated, and in normal subjects where such exercise level doesn't stimulate the sympathetic system. The methods used were: 1) the spectral analysis of HR variability where the high frequency component (HF, above 0.15 t4z, related to the respiratory frequency) is a marker of the vagal activity, the low frequency component (LF; at about 0.1 Hz) depends on both sympathetic and parasympathetic activities and the LF/RF ratio indicates changes in the sympatbo-vagal balance; and 2) a pure sympathetic index obtained from the amplitude of the T-wave of the ECG (TWA) which is decreased by adrenergic stimulation of the myocardium. Our results indicate that in both groups HR was increased by the exercise. In CTR, a net decrease in TWA (-39_+ 10%, n=6) was measured. Contrasting with this, in normal subjects, no change in TWA (+2_+12%, n=6) but a decrease in beth HF (-67_+13%, n=7) and LF (-85_+5%, n=7) without change in LF/HF ratio (-19+_.26%, n=7) were observed. These results indicate that CTR subjects increase their HR by an adrenergic stimulation, whereas the normal subjects regulate their HR for such moderate exercise only by redudng their vagal tone. There was a significant relationship (r=0.79, p<0.0005) between PS and the REE. The slope of the regression line indicated a net cost of PS of 3.6 kcal/g. We conclude that obesity in children is associated with an absolute increase in whole-body protein turnover consistent with the increased FFM and REE.

THE "GLUCOSE PARADOX" IN THE ANOXIC AND REOXYGENATED EMBRYONIC MYOCARDIUM Raddatz, E., Tran, L., Rochat, A.-C., Kucera, P. and de Ribaupierre, Y. Institut de Physiologie de l'Universitd, CH-1005 Lausanne. The hearts of 4-day-old chick embryos explanted in vitro react rapidly, reversibly and reproducibly to anoxia and reoxygenation. This work deals with the effects of exogenous glucose on the mechanical activity of the hearts submitted to strictly controlled normoxia-anoxia-normoxia transitions. The anoxic periods lasted from 10 to 60 seconds. Instantaneous heart rate, amplitude of contraction, velocities of contraction and relaxation of the ventricle wall were determined optically. In presence of 8 and 20 mmol/1 of glucose 1) the arrest of cardiac activity under anoxia was delayed by 60 and 110 %, respectively, 2) the duration of the postanoxic cardioplegia was reduced and 3) recovery was faster. Paradoxally, these concentrations of glucose induced arrhythmias both during anoxia and reoxygenation. The incidences of such arrhythmias increased with the duration of anoxia. The absence of glucose or blockade of glycolysis suppressed arrhythmias. Image analysis showed that the observed myocardial dysfunctions originated in the atrium and were sometimes associated with disturbances of propagation. The microinjeetion of okadaie acid into incompetent oocytas leads to the release of the prophase block with entry into M-phase. These G2/M-like alterations include chromatin condensation, germinal vesicle breakdown (GVBD), microtubules reorganization and formation of an abnormal (often multipolar) spindle of meiosis I, with chromosomes not aligned on the metaphase plate. The polar body is never extruded. Moreover okadaic acid microiujection induces the emergence of phosphorylated cytolasmic foci as detected by the monoclonal antibody MPM-2, known to react with mitotic phosphoproteins associated with centrosomes. The kinetics of GVBD following mieroinjection is extremdy retarded (24-48hrs) when compared to spontaneous maturation (30ran-lhr) and depends on the oocyte diameter. Interestingly, when inc~ompetent oocytes are cultured during 48hrs and then microinjeeted with okadaic acid meiosis resumes within a short delay (3-rhrs). MPM-2 epitopes are present on ecntrosomes only after okadaic acid injection. Following GVBD one of the step depending on protein synthesis, the expression of tissue type plasminogen activator, is triggered. Indeed OA-injected incompetent oocytes produce small amounts of this activator. Altogether these results argue for an effect of okadaic acid favoring entry into Mphase including microtubule organization, centrosomes phosphorylatiou and the recruitment of one of the masked mRNAs (tPA). St~rillt~, NCb~, C~-121l G~. ~t4RS-INSE~, Monf~llier, France. Meiotic reinitiation of the mouse oocyte is caracterized by a slow entry into metaphas~ I, b6~jinnin9 with germinal ve~ir break@own (GVBD) and ending with spindle Formation. It is accompanied by a cascade of protein kinases and phosphatases increasir~g protein phosphorylation. The activation of histone HI kinase (maturation promoting factor, MPF) and of the 1342 HAPK have been compared during spontaneous or okedalc acid (OA)-induced rhetoric reinitiation. In spontaneously maturing oocytes, hist6ne HI kinase activity increases before GVBD (2X) and is a~sociated with the disappearance of the upper (tyrosine phosphorylated) migrating form of p34 cdc2, Following GVBD, histone HI klnase activity culminates (8X) in metaphase I. Activation by phesphorylation of p42 MAPK occurs after GVflD. In contrast, no increase of histone HI kinase is detec~cable in oocytes induced to reinitiate meiosis by a transient exposure %o OA, neither before GVBO nor during the following 7 hours of culture. The upper migrating form of p34 cdc2 is present for 8 houra. The activation of p42 MAPK b~lins before GVBD. Furthermore, when OA is applied on Coclrces microinjected with p13 sucl, neither iocrease of histone HI kinase activity nor p34 cdc2 dephosphorylation are observed although GVBD is induced; p42 MAPK is activated.

Altogether these results suggest that GVBD may or may not be associated with a detectable activation of histene HI kinase, depending on the experimental conditions. Activation of pS4 cdc2 and p42 MAPK are separable events. We propose that, in the mouse c~yte, OA might be able to aetivaf8 an alternative pathw~ leading to GVBD that is cdc2-independent and that involves p42 MAPK activation ensuing MPF-independent phosphorylations. S 10-04 Urich, K., 4002 Basel, Switzerland Transformation of cells by polyomavirus is mediated by middle-T antigen. This viral protein is able to form complexes with src family tyrosine kinases (e.g. c-Src), phosphatidylinositol-3-kinase (PI 3-K) and phosphatase 2A (PP2A). C-Src associated with middle-T is constitutively dephosphorylated at tyrosione 527, a site negatively regulating the activity of this enzyme. This results in high tyrosine kinase activity in interphase and mitotic polyoma-transformed cells. Middle-T is transiently hyperphosphorylated during mitosis as reflected by an increase in the apparent lVl, on SDS acrylamide gels. Two putative phosphorylation sites for a cyclin-dependent kinase present in middle-T, threonine 160 and threonine 291, are also phosphorylated by purified p34 ~a~ in vitro. We constructed mutant middle-T genes where individual phosphorylation sites were converted to alanine residues. Mutants lacking threonine 160 were still able to associate with all cellular enzymes described above yet defective for Gell transformation. Interestingly, the defect of this mutation could be compensated by additional changes in the middle-T protein sequence introduced further downstream in a domain suspected to play a role in the targeting of this protein to intracellular membranes.

$10-05

Schmidt, S., Fa~khauser, C., and Viesturs, S. ISREC, 1066 Epalin~es, Switzerland Little is understood of the mechanisms which regulate cytokinesis, or how it is integrated with mitosis. We have cloned a number of genes from the fission yeast S. po~be which are required for the initiation of septation and are characterising them. Defects in the cdc7 and cdcll genes result in continued nuclear division without cytokinesis, while cdcl6 mutants undergo multiple rounds of septum formation without cell cleavage. Our analysis suggests that phosphorylation will play an important role in the regulation of cytokinesis and its integration with mitosis. A functional cdcl6 product is required for maintenance of p34 r activity in mitosis and the primary sequence of the cdc7 protein indicates that it encodes a protein kinase. Data concerning the role of the cdc7, cdcll and cdcl6 genes will be presented. RFC was originally identified in human cell extracts as one of the cellular factors essential for the replication of SV40 origin-containing DNA in vitro. It is a five subunit protein with one large subunit of 100-I40 kDA and four small subunits of 36-40 kDa. RFC binds specifically to primertemplate structures at the 3'-end of the primer. Together with proliferating cell nuclear antigen (PCNA) it serves as a primer recognition factor for DNA polymerase ~ and ~. To show the invobement of RFC in cellular DNA replication we are currently cloning S. cerevisiae RFC. The amino acid sequences of a number of tryptic peptides have been obtained. This information was used to amplify parts of the S. cerevisiae genes by PeR and to screen a genomic library with these probes. Three genes were cloned so far. The large subunit, RFC1, codes for a 95 kDa polypeptide that is 36% identical to the large subunit of human RFC (liRFC). The sequences for two of the small subunits were also obtained. RFC2, coding for a 40 kDA polypeptide, is 50% identical to the hRFC 37 kDA subunit. RFC3, coding for a 38 kDA polypeptide, is 51% identical to the hRFC 36 kDa subunit. There is also considerable similarity between the different subtmits. RFCJ, RFC2 and RFC3, as well as the hRFC subunits, have consensus sequences for a ATP/GTP binding site. All three genes are essential in S. cerevisiae. p53 is a major tumor suppressor gene which spefically interferes with the onset of E phase. We have therefore cxa~ined the possibility that p53 modulates the normal functions of enzymes and proteins directly involved in DNA replication. We have shown that human wtp53 protein binds physically to calf thymus single stranded DNA binding protein A, RP-A, We have also found that p53 blocks DNA helicases in a typical strand displacement assay. This may reflect the fact that p53 is able to reanneal complementary DNA single strands. Since both wt and some mutant forms of p53 share this function, they are unlikely to be related to the p53 tumor suppressor activity. To our surprise, among the helicases which were tested we determined that the p53 inhibition could be released by RP-A in the case of cellular but not in the case of viral DNA helicases. Additionally, we found that p53 can partially inhibit DNA polymerase ~ and s holoenzymes on singly primed MI3 DNA. This inhibition, however, was only evident when E. coli SSB was substituted for RP-A. Our data thus show that p53 exerts an inhibitory effect on several enzymes involved in DNA replication and DNA repair. The p53-RP-A interaction may function to protect replication enzymes from inhibition by p53 under certain physiological conditions. catalytic subunit showed that pol ~ is the most conserved replicative pol (Cullmarm G., Hindges R., Berehtold M.W. and Htibscher U., GENE, in press). We synthesized PCR primers and cloned three fragments spanning the whole catalytic subunit. The resulting PCR products, called N-term, Middle and C-term have a size of 1600, 1570 and 1280 bp, respectively. The primer for the N-term and Middle fragments contained a histidin tag in order to purify the recombinant proteins by using a Ni-NTA column. These fragments were cloned into the expression vector pGEMEX174. Because of the natural termination codon in the C-term fragment, the histidin tag was placed directly in the vector in front of the fusion protein, generating a new vector, pRHH1S. All three fragments were expressed in E.coli. The fusion proteins could be purified in a single step and were used to raise polyclonal antibodies. Finally, the entire pol 8 sequence was joined together and could be expressed. Data on the characterisation of the antibodies and the proteins will be shown.

Hsp 70 from calf thymus can influence DNA polymerase under heat shock conditions. We have generated antibodies against DnaK protein, the Hap70 homolog from Escherichia coil, and used them for detecting of Hsp70 protein during its purification from calf thymus tissue. The apparently purified protein has a molecular weight of 70kDa, and has been recognized by antibodies against DnaK and eucaryotic Hsp72/73, It possesses a very

week ATPase activity, which can be stimulated by different poly-and oligopeptide substrates. Results will be presented showing the influence of Hsp70 on DNA polymerase r from calf thymus under heat shock conditions.

Identification of a vertebrate cdc2 mutant which is unable to complete the G1/S transition. Nicole Sr and Viestars Simanis. Chemin Boveresses 155, 1066 Epalinges,.ISREC. S. pombe strains have been constructed which should permit the generic and molecular analysis of the events which occur in late GI when cells become committed to the mitotic cell division cycle and the initiation of DNA synthesis. A number of mutants of the chicken cdc2 gene were eonsUueted during the course of analysis of phosphorylation sites some of which lead to the interesting phenotype of cold sensitivity when expressed in a cdc2 null background, when the only source of p34 cde2 is the chicken homologue of the gene, expressed from a muldcopy plasmid. In order to create a genetically tractable strain, the wild-type chicken cdc2 gone and one mutant into the S. pombe genome to create a strain in which cell cycle progression is dependent upon the chicken cdc2 gene. Analysis of this strain indicates gnat the cells block predominantly before replication of DNA. In order to determine whether the cells were arrested before or after the execution of the start control their ability to conjugate at the arrest point was assessed. Consistent with a block before the traverse of start and contmim~ent to S-phase, cells were able to conjugate with high efficiency. Further support for the view that this mutant is defective only for the G1-S transition is provided by the observation that it is able to complement a cdc2A2l mutation in trans. This mutant is defective only for (32 function and not traverse of start. Tl~e double mutant is viable at the restrictive temperature of either of tile parents alld is no longer cold sensitive. Further work will concentrate on cloning genes involved in the traverse of GI into S phase in fission yeast. (Masson and Kreis, 1993, J. Cell Biol. 123:357) where it is localized on a subset of microtubulea. When Caco-2 cells are grown to confluency and become polarized, E-MAP-115 expression levels increase concomitant with a higher microtubule stability. Transfection of fibroblastic cells (Veto), which do not contain any detectable E-MAP-115, with cDNA encoding this protein, results in stabilization of microtubules to nocodazole treatment. These data suggest that E-MAP-115 may be involved in the stabilization and reorganization of microtubules in polarizing epithelial cells. Since E-MAP-115 is a stabilizing protein, its interaction with microtubulen should presumably be modified at the onset of mitosis to allow disassembly of the interphase microtubule network and formation of the mitotic spindle. Indeed, immunolabeling for E-MAP-t 15 varies during mitosis. In early prophase, no E-MAP-115 can be detected on the microtubules of the forming mitotic spindle. The protein is observed on the spindle microtubules of metaphase cells and staining becomes bright on the new interphase microtubules of telophase cells. Analysis of the protein in cells blocked in mitosis by low concentrations of nocodazole indicates that E-MAP-115 is hyperphosphorylated. This hyperphosphorylated form of E-MAP-115 is no longer able to co-sediment with microtubules in in vitro microtubule-binding assays. Phosphopeptide map and phosphoamino acid analysis show the existence of novel phosphorylation sites in E-MAP-115 during mitosis compared to the protein during interphase, We are currently characterizing the kinases involved in the modulation of E-MAP-115 activity. Little is known about the specific functions of calcium-binding proteins in epithelial cells ~md in particular in tumoral cells of epithelial origin. Calretinin (CR) and parvalbumin (PV) are supposed to be involved in cell proliferation. We observed the endogenous expression of CR in several cell lines from human colon adenoearcinomas, and we obtained stably transfected cells for PV in a human ovarian adenocarcinoma cell line. We studied the cell cycle in correlation with the endogenous or ectopic expression of CR and PV, respeetiveIy. G1 and mitotic celts are strongly labelled with the CR-antiserum 7696, while PV immunoreactivity is dominant in interphasic, but not in mitotic cells. In CR expressing cells, the cell cycle seems to be normal and the rate of proliferation to be proportional to the percentage of positive cells. The cell cycle is inhibited at mitosis after the addition of CR antisense oligo nucleotides to the culture medium. The cell cycle is blocked in GI/S and G2 in the cells eetopically expressing PV. The results support the hypothesis that these two calcium.binding proteins, CR and PV, could intervene at different moments and probably with different mechanisms on the mechanism of the cell cycle in the tumoral cells studied.

Production of polyclonal antibodies against calretinin CR-22k Gander, J.-Ch., Gotzos, V., Cello, M.R. and Schwaller, 8. Institute of Histology and Gen. Embryol., University; 1700-Fribourg A mRNA for an alternatively spliced form of calretinin, ealretinin-22k (CR-22k) was detected in WiDr cells (a cell line from a human adenocarcinoma). We therefore decided to produce antibodies specific for this protein. It is directed against the 14 amino-acids localized at the C-terminus of the truncated protein which are unique for this protein and not present in full-length calretinin. Two different approaches have been chosen to obtain a specific antiserum. We have produced a chimeric protein containing the Cterminal region of CR-22k and the (H)6(NANP)19 polypeptide as carrier. The fusion protein was expressed in E. coil and purified on a nickel chelate column. As a second method, a synthetic peptide (corresponding to the last 15 amino acids of CR-22k) was crosslinked with the carrier protein KLH (keyhole limpet hemocyanin). The antigens were used to immunize two rabbits. After the first boost, the 2 sera were tested. We have observed that both antisera recognize besides the protein used for immunization also CR-22k (expressed in E. coil) in a Western blot specifically. No other protein bands of either E.coli extracts or from eytosolic extracts of WiDr cells crossreact with the 2 antibodies. Both antisera detected the protein CR-22k in immunohistochemieal stainings of WiDr cells confirming the presence of CR-22k in WiDr cells. The TCTP (P23) is a growth related tumour protein which is expressed under translational control. Homologous acidic proteins have been found in higher plants and in Saccharomyces cerevisiae (YKL312) . The striking conservation of this polypeptide between these very divergent organisms suggests some important but still, unknown cellular function. To address this question, a null mutation was constructed in yeast, by deleting almost all the YKL312 structural gone. The deleted strain shows no detectable phenotype. Therefore, we conclude that, in yeast, this gene is dispensable. The protein YKL 312has been overproduced in E. coil and injected to rabbits to raise antibodies. We are now characterizing the YKL312 expression at RNA and protein level.

Ellenrieder,C., Forrer,P., Kosovsky,J., Rischle, M., Nueller, R. and Jaussi, R., Institut f~r Medizinische Radiobiologie, Paul Scherrer Institut & Universitgt ZH, CH-5232 Villigen Radiation therapy of tumors could be improved by influencing cellular radiation sensitivity. Yeast strains with defective cell cycle checkpoint genes (e.g.

radl, rad3, radg) show increased radiation sensitivity. Cross-hybridization of a probe from the tad9 gene on human mRNA Northern blots yielded a single prominent signal. Experiments to clone the corresponding 4 kb cDNA are underway.

We have cloned a hamster cDNA encoding the homologue of human CDK2. Probably, this cDNA encodes the p40 protein from hamster whose phosphorylation responds to radiation checkpoint signals and who does not bind to p13 sucl beads. Treatment of cells with CDK2 antisense oligonucleotides is hoped to modulate the radiation sensitivity. Proliferation of smooth muscle cells (SMCs) is a major event in vascular development and atheromatous plaque formation. In order to characterize SMC replicative potential, newborn rats have been injected with 3H-thymidine (3H-TdR) during their first week of life when most of SMCs were proliferating; then, 3H-TdR labelling was evaluated in adult rats. Depending on the number of mitosis that SMCs had accomplished after the first week of life, four different SMC subpopulations could be defined indicating that rat aortic SMCs are heterogeneous in their replicative activity. 5-Bromo-2'-deoxyuridine (BrdU) incorporation after balloon induced endothelial denudation of rat aorta in adult rats showed that SMCs entering into the cell cycle were mainly devoid of 3H-TdR labelling, This category could derive from two distinct SMC subpopulations: SMCs which were arrested in their proliferation before or just after birth or SMCs which had actively replicated during development. Thus, one (or two) subpopulation(s) of rat aortic SMCs, characterized by a particular replicative activity during development, is (are) selectively activated after balloon induced endothelial denudation. The situation in vivo of dermal fibroblasts embedded within the connective tissue can be emulated in vitro by growing the cells in collagen gels. We have investigated how fibroblasts cultivated within a matrix react upon wounding of this dermal equivalent. Human skin-derived fibroblasts (KD-cells) growing within attached collagen matrices were monitored over a period of 8 days. Fluorescently labeled cytoskeletal elements (tubulin, vimentin, acfin) and fibronectin served as phenotypica/ markers. The 3D-arrangement of fibroblast microtubules, intermediate filaments and microfilaments, as well as of fibronectin was visualized by CLSM and optical sectioning. Cells at the wound margin, adjacent to the wound, and in non-wounded controls, up to a depth of about 100gm were analyzed. Upon wounding, i.e. upon excision of a 1 mm 0 "punch biopsy" from the center of the collagen matrix, fibroblasts in the wound region oriented towards the matrix defect. Layer-like structures of increased cell density evolved over time at the wound margin. The tissue defect was covered by fibroblasts, reminescent of the situation in vivo.

Global yeast genome expression analyzed by 2-D PAGE after TCTP homolog gene (YKL312) disruption. Sanchez A translationally controlled tumor protein(TCTP) was identified and microsequenced from a human liver sample. Previous publications described the presence of a gene related to TCTP in plants, mouse erythroleukemia and ascetic tumor, human mammary carcinoma and more recently in the yeast. TCTP has no known function but its high degree of homology from plants to human underlines its probable crucial role in cell function. In order to study multifactorial chan~es in protein expression as a function of the YKL312 gene disruption m the yeast, we used high resolution two-dimensional gel electrophoresis combined with an efficient computer system (MELANIE). Two groups of gels (YKL312 +(n=10) and YKL312 -(n=l 0)) were analyzed, compared and classified using correspondence analysis. There was no significant difference between the two populations of gels according to either qualitative and quantitative spots expression except for two spots which completely disappeared. The first one corresponded to YKL312 gsne product. The second one corresponded to a 28 KDa peptide with an isoelectric point of 4.6. This YKL312 associated peptide still needs to be characterized and linked to the yeast genome. Demyelinating and neurodegenerative diseases are associated with altered cellular responses including processes mediated by receptor protein tyrosine kinases. Using a PCR cDNA cloning approach we have identified a novel member of the eck/elk/eph genefamily in myelinating serum-free brain cell cultures. Alternatively spliced cDNAs encoding complete open reading frames for putative transmembrane proteins have been isolated from both rat and human brain. A recombinant protein derived from the rat cDNA was demonstrated to have protein tyrosine kinase activity. An affinity purified antiserum to this protein was used to identify a glycoprotein of 120 kD molecular weight in brain cultures, and developping and adult brain tissue. The protein expression is most prominent during embryonal development and during the first three weeks of postnatal life. By in-situ mRNA hybridisation the transcripts were found predominantly in restricted neuronal populations. In order to investigate systematically the effects of geometrically defhied abrupt tissue expansion on the characteristics of impulse propagation, we constructed expanding cellular structures in vitro using patterned monolayer cultures of neonatal rat heart cells (photolithographically patterned substrates.) The preparations, which consisted of narrow cell strands (40-80 tim wide) inserting into large reetangular cell areas (side lengths >1500 ~m), were stained with the voltage-sensitive dye di-g-ANEPPS and were imaged onto a 12 X 12 photodiode matrix array (maximal spatial resolution 151.tin X 151am in the object plane). While impulse propagation was uniform in linear cell strands (average conduction velocity 0.35 m/s), a transient decrease in conduction velocity hi the transitional region was invariably observed in the suddenly expanded structures. This decrease was accompanied by marked distortions of the local action potential shapes due to bidirectional electrotonic interactions across the transition: upstream from the expansion, the repolarization phase was interrupted, a few ms after its inception, by a small secondary depolarization ("spike and dome" configuration), while, downstream, the action potential upstrokes were preceded by prominent prepotentials. In those cases where conduction failed at the tlansition, action potential durations upstream were dramatically abbreviated. These findings showed that phenomena similar to those recorded in intact Purkinje-fiber-ventricular preparations can be observed in cell ensembles in vitro having comparable geometlinS but consisting of cells with uniform active membrane properties. Supported by Swiss NSF grant 823A-028424 (SR) and USPHS grant NS 16824 03MS). It is a noncollagenous glycoprotein with a molecular mass of 148 kDa consisting of three identical subunits. CMP is selectively extracted with EDTA-containing buffer what indicates a divalent cation dependent anchorage of the protein in cartilage matrix. Electron microscopy revealed the presence of three compact globular subdomains and sequence analysis indicated the presence of a coiled-coil c~-helical assembly domain formed at the C-terminal end of each subunit. The trimeric structure was retained after complete reduction under native conditions which reveals that the subunit structure of CMP is not only stabilized by disulfide bridges but also by the coiled-coil assembly domain. Tissue distribution studies in mouse revealed that CMP is selectively expressed in some but not all cartilages. Adult rat cardiomyoeytes (ARC) in culture degenerate the myofibrillar apparatus and after attachment to the substrate they grow and reassemble new rnyofibrils. The appearance of new, well organized rnyofibrils can be observerd in several distinct regions. They appear close to the membrane proximal to the culture substrate and in the perinuclear region close to the substrate.

Previous studies have shown that embryonic cultured cells assemble new myofibrils which insert in adherens junctions at sites of cell-cell contacts and in sub-sarcolemmal adhesions plaques (SAPs) at sites of cell-substrate contacts. This SAPs are thought to function as the nucleation sites for myofibril formation.

We have investigated the formation of this SAPs in cultured adult rat cardiomyocytes. Using confocal light microscopy we can show that the interaction between rnyofibrils and membrans occurs close to the membrane proxirnal to the substrate and that this SAPs are flanking both nascent myofibrils and SFLS, the latter structures being believed to serve as scaffold for myofibrillogenesis, Additionally we have used electron microscopy to investigate the formation of this structures at higher resolution.

$11-05 Muscle satellite cells (SC) are quiescent myoblasts, ready to be activated and to regenerate new muscle fibers in case of muscle damage. In vitro, single human muscle SC, cultivated as clones, give rise in proliferating conditions to two subpopttiations of cells, cells expressing both ~-striated muscle actin and desmin (c~-SR+DSM+), and cells expressing desmin alone (DSM+). In culture conditions promoting differentiation, a clone .typically gives rise to a fusing progeny, yielding c~-SR+DSM + myotubes, and to non fusing myoblasts (NFMB). The latter are DSM +, a phenotype similar to quiescent SC found in situ.

In order to determine whether NFMB are the in vitro equivalent of in situ quiescent SC, this first generation of NFMB were selectively collected and subcultured. In proliferating conditions, NFMB were able to resume proliferation and to give rise to a progeny with both c~-SR+DSM + and DSM + cells. When cultured in differentiating conditions, NFMB formed ct-SR+DSM + myntubes, and a new generation of DSM + NFMB. When NFMB of the second generation were again selectively collected and subculture& they resumed proliferation and ~re again able to yield both myetubes and a third generation of NFMB. NFMB of the first generation were also cultivated as clones, and 5 to 25% of them were able to resume extensive proliferation, yielding in their progeny both c~-SR+DSM + and DSM + cells, which differentiated into myotubes as well as NFMB.

These results strongly suggest that SC have the stem cell property, of selfrenewal, yielding in their progeny both cells ready to differentiate into muscle and ceils similar to themselves. Extra cellular matrix (EC~ regulates the expression of B-casein in cultured mouse mammary epithelial cells. We have developed a functional ~ epithelial cell strain, which expresses high level of milk proteins, forms alveolar-like structures when plated onto a reconstituted basement membrane and secretes casein unidirectior~lly into a it~aen. We have further shown that FflV~dependent regulation of B-casein occurs mainly at the transcriptional level and that 5' sequences play an inloortant role in this regulation. We have located a 160bp transcriptional enhancer (BCEI) within the 5' flanking region of the B-casein gene. Using functional assays, we show that BCEI contains responsive elements for ~ and prolactin-dependent regulation. BCEI placed upstream of a truncated inactive B-casein promoter reconstitutes a promoter even more potent than the intact prcmoter, which contains BCEI in its natural context more than 1.5kb upstream. This small fusion promoter reconstitutes the nonaal regulation pattern. We show that BCEI mediates F/24 dependent regulation even when linked to a het~rologous viral promoter. Purified satellite cells (SC) were obtained from human skeletal muscle biopsies following cell sorting by flow eytometry. A chemiluminescent assay for acetylnholine (ACh) revealed the presence of 1,3 nmoles/weU of 100.000 SC. In elonal cultures of proliferating SC, this intracellular level of ACh was 0.l nmoles/well. When cells were cultured in the presence of an esterase inhibitor (10 IxM phospholine), the ACh amount was enhanced 2 fold. Conversely, cultivating the cells in presence of a potent inhibitor of choline acetyl transferase (2 gM BromoACh), gave a 50% decrease of ACh content Finally, the release of ACh was detected in the supernatant of cultured SC, following a 15 min incubation, and the measured level of ACh was 3 pmoles/well/min.

The presence of an ACh-like compound in myogenic cells in both freshly isolated and in proliferating SC suggests that SC may contain ACh in situ. In additiou, the fact that ACh was present in the extracellular medium suggests that ACh can be released spontaneously by the cultured SC. Type XII collagen is an extracellular matrix protein associated with collagen fibrils in vivo. As observed for several other extracellular matrix proteins, the synthesis of type XII collagen by chick fibroblasts cultured in 0.1% FCS is stimulated by TGF-I~I. Two splice variants of type XII collagen are known, with subunits of either 220 kDa or 350 kDa. By using antibodies specific for the large (350 kDa) form of type XII collagen, we could show a more restricted distribution of the large variant compared to the smaller form in the embryo. While only the large variant carries chondroitin sulfate, both variants are identical in their collagenous domain. It is thought that via this domain, type XII collagen binds to collagen type I fibrils. We demonstrate here that type XII collagen can bind to collagen I fibrils also in vitro. By neutralizing acid soluble collagen type I, we could coprecipitate type XII collagen together with newly formed collagen type I fibrils. This interaction occurs in physiological saline but is highly salt dependent. In further studies we investigated wether 35S-methionine labeled type XII collagen treated with chondroitinase ABC, ct-chymotrypsin, or collagenase can still be precipitated together with type I fibrils. In addition, we found that type XII collagen also affects the rate of type I collagen fibril formation.

H.U. Keller Department of Pathology, University of Bern, CH-3010 Bern Locomoting blebbing cells Colchicine (10-SM) can induce locomotion associated with blebbing in Walker carcinosarcoma cells. Blebs expand at a rate (about 2~m/sec) which is much faster than known rates of actin elongation. This suggest that the membrane is directly pushed forward by forces other than actin elongation, possibly by hydrostatic pressure. This interpretation is suggested by the observation that there is significant F-actin staining all along the cell membrane but not with the cytoplasmatic content the blebs. Blebbing is suppressed by 0.5 M sorbitol. The finding that cytochalasin D suppresses blebbing shows that actin polymerization is nevertheless indirectly instrumental in bleb formation, possibly by generating a high intracellular pressure. A tentative model explaining locomotion of blebbing cells is presented. Osteopetrosis encompasses a family of diseases with different causes and the common phenotype of impaired bone resorption. The osteopetrotic mouse mutant oo/oo is deficient in CSF-1, the growth factor for the cells of the mononuclear phagocytic system. The phenotype of oo/ou mice is characterized by a low number of peritoneal macrophages and peripheral monocytes, and a virtual absence of osleoclasts, leading to the osteopetrotic phenotype. Injections of CSF-1 into op/oo mice reversed the osteopetrotic phenotype, proving the requirement for CSF-1 during the process of osteoclastogenesis.

By in situ hybridization, expression of CSF-1 was demonstrated during bone development. Osteoclast precursors and mature osteoclasts were identified as putative target cells for the growth factor, since these cells express CSF-1 receptors, which is encoded by the proto-oncogene c-fm~. Subsequently, specific binding of CSF-1 to osteoclasts was shown. Expression of c-fms parallels the expression of CSF-1 in bone both in time and place, suggesting a local action of this growth factor in osteoclast formation, but also in the modulation of osteoclastic activity. Different transcripts, raised by alternative splicing of a commmon nuclear RNA precursor, have been found to encode a secreted and a membrane associated forms of CSF-I. The secreted form can be modified by attachment of a glycosaminoglycan side chain, which serves as an anchor to integrate the peptide into the extracellular matrix. Investigation of the role the different CSF-1 forms play during recruitment of osteoclasts and regulation of their activity will provide further insights into the mechanisms governing these processes. Comparative sequence alignments of known voltage-gated K + channels revealed two conserved cysteine residues in the putative transmembrane segments $2 and $6. If the cysteines were connected by a disttifide bridge, it would put a structural consllaint on possible channel models by placing $2 next to $6. We used site-directed mutagenesis to study systematically the potential roles of these invariant cysteines. Fourteen of 17 substitutions in $2 (C232), and 7 of 19 substitutions in $6 (C393) maintained channel function in Xenopus oocytes microinjected with mutant cRNA. Therefore, the conserved cysteines are not essential for K + channel expression in Xenopus oocytes. However, electrophysiological characterization of the cysteine replacement mutants revealed distinct roles for the two cysteines. Inactivation, deactivation, and ion permeation did not considerably change in $2 mutants. In $6, in contrast, cysteine replacement by leucine, asparagine, glycine and valine accelerated inactivation and deactivation kinetics substantially, whereas serine and threonine showed opposite effects. Furthermore, the voltage dependence of deactivation was differently affected. In summary, it appears that the side-chain at position 393 in $6 of Kv2.1 is involved in channel gating by participating in the transitions from the open to the closed and inactivated state of the channel. (Supported by grants from the NIH and MDA to RHJ, and from the SNF and the Swiss Foundation for Medical-Biological Grants to RDZ.) The segment between $5 and $6 of voltage-gatedK + channels, called H5 or P region, has been implicated to form part of the ion conduction pathway. Little is known about the conforroation of this region although various models have been proposed including a J3 barrel-like structure formed by four adjacent antiparanel J3 sheets. To gain insight into the secondary structure of this region, we used cysteine substitution mutagenesis and sulfhydryl-specific, membrane-impermeant reagents to probe the accessibility of amino acid side chains in and around the P region of Kv2.1 (DRKI). Twemy-eighi positions from K356 to T383 were each mutated to a cysteine. After expression of mutant K + channels in Xenopus oocytes, cysteine side chain accessibilities were probed by superfusion with CH3SO2SCH2CH2NMe3 + (MTSET), MTSET can form a mixed disulfide with an accessible cysteine, and this may lead to current reduction if the covalently modified side chain is in or close to the ion conduction pathway. Thus far, we have identified four mutations that showed K + current reduction after superfusion with 3 mM MTSET. The mutants P361C, I379C, Y380C, and K382C showed 87%, 99%, 39%, and 21% reduction in current amplitude, respectively. In coutrast, mutants $363C, A367C, T368C behaved like wild type Kv2.1 with less than 5% current reduction. Our results suggest that the side chains of P361,1379, Y380, and K382 are directly accessible from the extracellular environment. Taken together, these residues may, therefore, face the lumen of the ion channel pore. Furthermore, the degrees of inhibition are in agreement with a model in which the ion conduction pathway narrows from residue K382 to Y380 to I379. (Supported by grants from NIH and MDA to RHJ, and from the SNF and the Swiss Foundation for Medical-Biological Grants to RDZ.)

Motejlek K., H,,iuselmann R., and Liig:her B.; Pharmakologisches Institut der UniversitiIt Ziirich, Winterthmtr. 190, CH-8057 Zlirich Comparison of 5' flanking sequences of three GABAA-receptor subunit genes (atl, ~, 8) revealed a novel conserved purine sequence dement with the consensus sequence GAGAGGGGAGAOGA GAGAG(GG/AA)G. This element is present once each in the (zl and 72 subunit gene promoters and in seven tandem copies in the 8 gene promoter. A novel DNA binding activity (BSF1) was identified that binds to various versions of this purine element. BSF1 was found ordy in brain but not in any other tissues tested. Furthermore, the factor is distinct of BETA, another brain-specific DNA binding .protein with a purine-rich DNA recognition sequence. The expression of BSF1 during differentiation of eerebeUar granule cells in vitro correlates with the expression of GABAA-receptor subunit genes. This correlation suggests a role for BSF1 as a transcriptional activator of neuronal genes. Therefore, BSF1 may be important for cell ty~specific regulation of GABAA-receptor gone expression. We also found that BSF1 binds to wheat germ agglutinin, suggesting that this protein is glycosilated. Taking advantage of this property, we are in the course of purifying BSF1, The question, whether BSF1 is indeed a transcription factor is being investigated in vivo and in vitro. To produce a simple method of estimating the free magnesium concentration ([MgZq) in solution, we have manufactured dip cast Mg 2+ macroelectrodes using the neutral Mg carrier ETH 7025.We have used the elecmides to measure the dissociation constant (K~) for MgATP in a background solutions mimicking the intracellniar milieu. For the mcasmement of the big ATP dissociation constant, the background solution contained 2 mmol/l N%ATP. This solution was titrated with MgCI 2 and the changes in [Mg ~] above 0.05 retool/1 monitored with the maeroelectrode. During the titration the pH was maintained constant. Fitting such titration curves with the standard hyperbolic binding equation, after correction for zero drift, gave the following mean+SD values for the Ka (pmol/1 at 37~ pH 6.4, 103.3-+-25,3 (n=9); pH 7, 59.7:k-27.4 (n=6) and pH 7.4, 30.g~15.4 (n=7). Elevation of the [Naq to 50 m~l/1 and reducing the [Kq to 1(30 mmol/l was without effect at pH 6.4 (n=6). Reducing the temperature to 25~ increased the K a at 7.4 to 49.0:k21.6 (n= 5). Mg =+ macroelcctrodes provide an easy way of measuring [Mg z § in solution.

[Mg ~] can also be measured in a background solution containing I mn~I/l Ca, although the electrode response is reduced. Interference by protein to date prevents their use in plasma. Similarily to the block by Zn 2+ and Ca 2+, protons only partially block the channel. The ability for all mutated channels to. conduct Na + current, and the partial block induced hy Zn 2+, Ca 2+ or H + indicate that Y401 is not located deep in the pore but rather in the extracellular mouth of the channel.

Progesterone modulates the activity of glycine receptor expressed in colliculli neurons of neonatal rats.

Maury, K. & Bertrand, D. Dpt of Physiology, CMU, 1211 Geneva 11. Steroids are potent narcotics and have been shown to act on ligand-gated channels activated by GABA or nicotine. However, little is known about their possible actions on other receptor types. For instance, while it was demonstrated that progesterone inhibits glycine receptors expressed by spinal cord neurons isolated from chick embryos, nothing has yet been reported for brain receptors. In this study, we have determined the properties of glycine channels expressed by brain neurons of neonatal rats using the whole-cell voltage-clamp technique. In inferior colliculi neurons, glycine elicited little-desensitizing inward currents which reversed around -40 mV. These currents were blocked by strychnine in the nanomolar range. Surprisingly, however, we found that progesterone, in the micromolar range, potentiated the glycine evoked currents. These results, contrasting those obtained in chick spinal cord neurons, suggest that progesterone enhances or antagonizes glycine currents depending on the subtype of the receptor.

The mammalian auditory organ (organ of Corti) relies on two groups of sensory cell for its normal hearing: Inner Hair Cell & Outer Hair Cell. IHCs are mainly innervated by afferent fibers while the OHCs essentially receive efferent fibers. The OHCs possess unique electromofile properties which are thought to enhance the motion the basilar membrane and to refine the mapping of the sound frequency. We found, using whole ceil recordings, that 5-10% of the fleshly dissociated OHCs displays fast inward currents. The magnitude of peak currents which can be as large as 2nA vary from cell to cell. Further experiments have revealed that this current is sensitive to TTX and strongly reduced when extracellular sodium is replaced by choline. Its kinetic is relatively slow compared to that of sodium current of typelspiral ganglion cell, and has a more negative inactivation. Other voltage activated currents and ligand-gated currents are under inverstigation. These results will provide further insights in the hearing transduction mechanism.

The formation and the properties of homotypic and heterotypic gap junctions were studied using two types of insect cells, C6/36 and SF9. Two single cells were pushed against each other and the subsequent de novo formation of gap junction channels was assessed by means of the dual voltage-clamp method (Bukauskas and Weingart: Pfl~g. Arch. 423:152, 1993 It was symmetrical for homotypic junctions. The Vj-sensitivity was less pronounced in SF9 cells. Hence, the relationship for heterotypic junctions was asymmetrical.

All pairs examined revealed a S-shaped relationship between gj(ss) and V, which was virtually superimposable (gj(ss) declined upon depolarization).

Each channel contains a Vm-gate and two ~-gates. The Vj-gates are operated independently. They close when their intracellular aspect is made positive. shaped curve compatible with a two-state Boltzmann process. Half maximal inactivation was reached at -77 mV and +66 mV, implying an asymmetrical gating behavior. In case of an asymmetrical protocol (~ and Vewas stepped simultaneous, but in opposite direction), the q-dependent asymmetry disappeared (half maximal inactivation at -59 mV and +60 mv). The asymmetry in gj-gating, attributable to the pulse protocol, was also reflected in the time constants of Ij inactivation (r177 Cerebellar granule cell cultures were used to study the regulation of the NMDA receptor expression. We have shown previously that chronic membrane depolarisation (25 him K +) or treatment with NMDA {i00 ~M) promotes the functional expression of NMDA receptors, as assessed by NMDA-evoked 45Ca2+ influx. We have now developed a RNase protection procedure for the quantitative determination of the mRNA levels of the NMDA receptor subunits known so far (NR-I,2A,2B,2C,2D). The growth conditions mentioned above failed to alter the mRNA levels of the different subunits. At the protein level, NMDA receptors were evaluated quantitatively by the labeling pattern obtained by the new photoaffinity ligand 125I-CGP55802A. The intensity of the phctolabeled bands, thought to represent NR-I and NR-2A subunits, was increased by treatment with high K + or NMDA compared with control cultures. This result suggest an increase in receptor protein by high K + or NMDA treatment. Thus, posttranscriptional mechanisms seem to play an essential role in the modulation of the NMDA receptor activity. The gene encoding the pore-forming subunit of the rat epithelial Na + channel (~rENac) was cloned recently and shown to belong to a novel gene family coding for cation channels (Nature 361, p467-470 (1993) . Transport of Na + ions through these channels is the rate-limiting step of Na + absorption and thereby controls osmotic balance of body fluids and secretions. In order to study the implication of ~rENac in regulating sodium balance, blood volume and blood pressure and therefore its involvement in hypertension, we expressed ~rENac under the control of the human CMV promoter in transgenic mice. Several lines of mice showing expression in different tissues were obtained. Progress in the analysis of these mice will be discussed. Activation of metabotropic glutamate receptors increases the excitability of neurones in the lateral septum of the rat. To elucidate the mechanism(s) of this action, we used whole-cell patch-clamp recordings from coronal brain slices of young animals. In voltageclamped cells, the selective agonlst (IS,3R)-ACPD, at 1-50 /~M, had the following effects, a) It evoked a sustained inward current, which was resistant to TTX and to cadmi~am and which persisted in neurones loaded with BAPTA, a calcium chelator. This current displayed inward rectification and was reduced, or suppressed, when extracellular sodium was partially replaced with N-methyl-D-glucamine. b) It elicited a TTX-insensitive, voltage-dependent inward current, which activated at -50/-40 mV and which reversed at aound 0 mV. This current was suppressed in a low-calcium/high-magnesium perfusion solution and was undetectable in the presence of intracelhilar BAPTA. We suggest that ACPD exerts a dual action on lateral septal neurones. It causes a steady depolarization by generating a sodium-dependent current and it triggers transient plateau potentials by inducing a calcium-dependent cationic current. Interaction between these currents results in a powerful, self-reinforcing neuronal excitation. We have investigated the effects of protein kinase C (PKC) activators (4[~-PMA, OAG) and of phosphatase inhibitors (okadalc acid, calyculin A) on voltage-gated Ca 2+ and K + channels in NGF-differentiated PC12 ceils. Whole-cell Ba 2+ and K + currents were recorded with the patch-clamp technique. By using specific Ca 2+ channel blockers (co-conotoxin (CgTX), isradipine) we found 3 types of Ba 2+ currents (IBa): a) a ~CgTX-sensitive IBa; b) an isradipine-sensitive IBa; and c) a t0-CgTX plus isradipineresistant IBa. 4[~-PMA and OAG specifically down-modulated the isradipine-sensitive IBa. The inhibition of IBa was prevented by staurosporine and PKC (19-31) (2 PK inhibitors). The delayed rectifier K + current was unaffected by PKC activators. Applied externally, okadaic acid and calyculin A inhibited the total IBa by affecting several components of the Ba 2+ currents. However, the 0~-CgTX plus isradipine-resistant IBa was unaffected by okadaic acid. In conclusion, our results suggest a differential modulation of voltage-gated Ca 2+ and K + channels by the PKC signalling pathway in NGF-differentiated PC 12 cells. Agrin, a protein isolated from basal lamina extracts of the electric organ of the marine ray, is thought to mediate the motor neuron-induced aggregation of acetylcholine receptors (AChRs) in muscle fibers at their neuromuscular junction. Recent cloning in the marine ray, rat and chick has revealed that agrin can be alternatively spliced at two sites in its C-terminal half. Site A encodes either 0 or 4 amino adds and site B encodes either 0, 8,11 or 19 (8+11) amino acids. Studies of agrin mRNA expression indicate that early in synaptogenesis, chick motor neurons contain high levels of A4B19. In contrast, non-neuronal cells and muscle calls contain A4B0 and AOB0 mRNA. In the adult ray, electric lobe motor neurons that innervate the electric organ contain A4B8, whereas agrin mRNA in the electric organ again lacks both sites (McMahan et al. (1992) , Curr. Up, Cell Biol. 4:869-874). To investigate the functional properties of agrin isoforms in more detail, we have now compared their specific activity to aggregate AChRs on cultured chick myotubes. Heterologous expression of the C-terminal half in COS-7 and 293 cells revealed that chick agrin isoforms A4B8 and A4B19 were highly active, while A4Bll was up to 40 fold lower in activity. No activity could be detected for the A4B0 and AOB0 isoforms. In agreement with these findings the A4B8 isoform of the marine ray also showed high AChR-clustering activity. Therefore, we conclude that motor neurons throughout development synthesize highly active agrin isoforms while the postsynaptic target ceils do not play a primary inductive role in the formation of synapses. We have used whole-cell patch-clamp recordings and hypothalamic slices in order to characterize the effect of N-methyl-D-aspartate (NMDA) on suprachiasmatic neurones. In ceils clamped at or near their resting membrane potential, NMDA (50-100 #M) generated an inward current of 10-60 pA, which was insensitive to TTX and which reversed at about 0 inV. The NMDA current-voltage (I-V) relation contained a region of negative slope conductance. The NMDA current was reduced or suppressed by D(-)-2-amino-5-phosphonopentanoic acid (D-APS) or by MK-801. It was potentiated by reducing the extracellular magnesium concentration from 1 to 0.01 raM, or by adding glycine (10 /~M) to the perfusion solution. In a majority of neurones, lowering the extracellular calcium concentration from 2 to 0.01 mM caused a 1.5to 4-fold enhancement of the NMDA current. I-V relations indicate that in the low-calcium solution, the region of negative slope conductance was attenuated. This effect was due to extraceUular calcium, since it persisted in neurones loaded with the calcium chelator BAPTA. We conclude that NMDA channels present in suprachiasmatic neurones may be modulated by extraeellular calcium. Institut, and Abt. Pharmakologie*, Biozentrum, Basel. During innervation of skeletal muscle, the myonuclei underlying the synapse begin to express acetylcholine receptor (AChR} ~-subunit gene and they remain activated after the nerve is removed. Our experiments show that this is due to a factor in the synaptic portion of the muscle fibre basal lamina (BL). One candidate for this factor is agrin, which regulates AChE clustering in the synaptic membrane: i) Unlike in untreated muscle, the level of s at synapses was strongly reduced in cultured rat muscle fibres after proteolysis of synaptic BL. 2) Conversely, when rat myotubes were cultured on BL isolated from adult muscle, they preferentially expressed AChR accumulations and ~-mRNA at sites where they contacted the synaptic portions of the isolated BL. 3) When various substrates were impregnated locally with agrin A4BIg, an isoform expressed by motor neurones in fetal spinal cord, it locally induced in the myotubes the expression of S-mRNA.

Art increase of functional nicotinic acetylcholine receptors precedes the fusion of cultured human muscle satellite cells.

R. M. Kranse, C.-R. Bader and L. Berrtheim. Department of Physiology and Division of Clinical Neurophysiohigy, University Medical Center, 1211 Geneva 4, Switzerland Nicotinic acetylcholine receptors (nACtLR) are expressed on embryonic myoblast and ACb may play a role in mechanism of fusion. Our study focuses on the appearance of functional nAChR in freshly isolated and cultured satellite cells (SC) from normal human muscle biopsies. SC cart be conditioned to either proliferate or fuse to form myotubes depending upon culture media. Presence of ACh-activated current was investigated using the whole-cell and single.channel patch-clamp technique. In freshly isolated SC (whose properties should be close to the quiescent in vivo SC/no nAChR were observed (n=16). In the proliferating state, 54% (n=35) of the satellite ceils (which are also called myoblasts) displayed a small ACh-activated current (10 pA/pF) and, as expected, after fusion of salellite cells into myotabes, 100% of the ceils displayed an ACh-activated current (n=6; 26 pA/pF). To examine, whether the increased expression of nAChR precedes SC fusion, SC were cultured in a medium which promotes fusion, but were prevented from fusing by keeping them at a low density. In this culture condition, 93% of the SC (n=15) displayed an ACh-activated current and, in addition, these cells expressed a 4 times higher ACh-current density.

(40 pA/pF) than the proliferating SC. We also observed that, in high density cultures, a small population of SC do not fuse even atter 3 months in the medium promoting fusion. These non-fusing myoblasts (NFMB) had no nAChR. Our results suggest that the appearance of nAChR may be related to the process of SC fusion as i) quiescent SC in vivo do not express nAChlL ii) nAChR expression increases in conditions promoting fusion and iii) no nAChR were observed in NFMB. The latter cells may be equivalent of quiescent SC in vivo.

$12-20 We have investigated the sensitivity of neuronal nicotinic acetylcholine receptors (nAChRs) of known subunit composition to various cholinergic antagonists. Neuronal nAChRs were expressed in Xenopus oocytes after injection of pairwise combinations of (x2 or c~3 with either 1~2 or 1~4 cRNAs. The two-electrodes voltage-clamp technique was used to measure currents induced by rapid application of low concentrations of acetylcholine (ACh) together with increasing concentrations of antagonists. The response of ct31~4 neuronal nAChRs to ACh was halfmaximally inhibited by following antagonist concentrations (IC50): hexamethonium (0.3 I.tM), mecamylamine (0.2 p.M), pentolinium (0.2 I.tM) and trimetaphan (1.2 ~tM). With (x3132 nAChRs the IC50s of the same antagonists were about 10 times higher. Finally, (+)-tubocurarine was a conventional competitive inhibitor of ACh in ~3132 and t~2~2 nAChRs. In contrast, low concentrations of (+)-tubocurarine increased the response to low ACh concentrations in a3(34 and t~2~4 nAChRs. These results further underline the importance of [l subunits in determining the functional properties of neuronal nAChRs.

Supported by the Hochschnlstiftung and NF grant 31.31018.91 to ABC. Previously, the role of the transglial tubular system was shown for the squid giant axon: In Na + free (TRIS) solutions the series resistance increased in correlation with a decrease of the tubular opening density (TOD). In the crayfish giant axon, which show similar tubules, we correlate the change in TOD, measured in freeze-fracture preparations, witl] the conduction velocity. The control TO'D was estimated to be 16 per #ms. After 30 and 75 rain TRIS-Inkubation the TOD decreased to 10,4 and 1,5, respectively. This reaction was reversible when these axons were reincubated in Na~-c41ntaining solution. After 120 rain reincubatiou the TOD was 11.8 per tam'Z. To determine the influence of decreased TOD on the impulse conduction velocity, the nerve was incubated for 30 min in TRIS solution followed by reincubation. Impulse propagation then recovered in two phases. During an initial fast phase with a short time constant of 1-2 rain Na + diffused back into the periaxonai space and allowed impulses to occur with a low conduction velocity. In the second phase with a long time constant of 7 min, conduction velocity recovered slowly to normal values. This time course was comparable to the recovery of TOD, as estimated from recent freeze-fracture preparations after short reincubation. These results indicate that the normal TOD is a necessary condition for the normal excitability of the axon and determines its conduction velocity. Most spinal cord neurons respond to the inhibitory action of GABA and glycine, suggesting co-expression of GABA A-and glycine receptors 4n individual ceils. While glycine-receptors are exclusively found in post-synaptic densities, the cellular localization of GABA Areceptors has not yet been characterized.

In this study, the distribution of GABAA-receptors in the spinal cord was analyzed with an antiserum recognizing the (xl-subunit.

Double-and tripleimmunofluorescence staining were employed to identify synaptic receptors apposed to GABAergic terminals (immunoreactive for glutamic acid decarboxylase) and to assess their co-localization with glycine-receptors (visualized with an antibody to the 93 kDa protein gephyrin).

Staining for the GABAA-receptor Ctl-subunit decorated the soma and dendrites of numerous neurons in laminae III-VIII and X, revealing their morphology in clear detail. By contrast, laminae 1I and IX contained little immunoreactivity for these GABAA-receptors. Most GABAA-receptor-positive cells also exhibited a prominent glycine-receptor immunoreactivity. Both types of receptors had very similar distribution patterns and were frequently co-localized in sites apposed to GABAergic boutons. These results indicate that GABA Aand glycine-receptors may co-exist within single post-synaptic densities, suggesting a possible synergism between GABA and glycine neurotransmission in spinal cord neurons. Prenatal benzodiazepine (BDZ) exposure changes both behavioral and neuiochemical parameters of developing Iats. These changes are not a consequence of remaining BDZ in brain tissue, but rather indicate alterations in BDZ receptol binding and function. Since the BDZ binding site is located on the GABA~ receptor complex, it is possible that prenatal BDZ treatment influences the expression of GABA~ receptor subunits. To evaluate effects of pienatal BDZ exposure on mRNAs expression specific for several GABA~. receptor subunits (~i, ~5, ~ & Y2), we treated plegnant rats with diazepam (l.25mg/kg/d; s.c.) from gestational day 14 to 20. We then analyzed specific mRNA levels in offspring at four diffeIent developmental stages (GD20, PNS, PNI5 & adult) by in situ hybridization. Pleliminary data flom optical density measurements indicate a decIease in expression of mRNA in particular for ~-subunit of GAB~ receptors in different brain reglons of plenatally diazepam-treated zats. Processing of sensory information within the auditory system has been well characterized using histological and eleetrophysiologlcal techniques. However, relatively little is known about the neurotransmitters involved. The present study was performed to elucidate the possible role of excitatory (EAA] and inhibitory (i.e., GABA) amino acids in neurotransmission within the primary auditory cortex (AlL The main afferent to the AI, the ipsilatera} medial geniculate body (MGB), was electrically stimulated and evoked responses were recorded intracellularly in the AI region in halothane anesthetized cats. A seven-barrelled iontophoresis pipette glued alongside the recording electrode allowed localized application of EAAergic and GABAergic compounds onto the recorded neuron. In most AI neurons, MGB stimulation evoked a short latency, short duration EPSP followed by a long-lasting IPSP. Pattern, duration, and amplitude of synaptic potentia{s were highly variable and strongly dependent on stimulation intensity, Reduction of GABAA receptor-mediated inhibition by iontophoretic application of either bicuculline or SR95531 markedly enhanced MGB stimulation-evoked EPSPs. Only the late component of this enhanced EPSP was reversibly blocked by the competitive NMDA receptor antagonistl AP7 at currents which selectively blocked neuronal excitation induced by iontophoretic application of NMDA, Our results demonstrate that in the eat 1) NMDA receptors in AI are activated following MGB stimulation and that 2) a dominant GABA~ receptor mediated inhibition usually masks this NMDA receptor activation under our experimental conditions, A highly purified and specific cell wall degrading endo-l,5-u-L-arabinanase was isolated from an A. niger pectinase preparation by low and medium (FPLC) pressure column chromatographic separation methods. The isolated enzyme was most active on linear 1,5-~-L-arabinans (-90 I.U./mg), whereas branched arabinans from sugar beets were degraded to a lesser extent (-14 I.U./mg). The enzyme was shown to be electrophoretically pure after silver staining (SDS-PAGE) and its identity was confirmed through specific binding to an antiserum directed against endo-l,5-~-L-arabinanase. The major physico-chemical characteristics of the enzyme were the following: Mr 42'000, IEP ~ 3.0, pH optimum 4.8, temperature optimum 55~ pH stability 3.5-8.0, temperature stability S 45oc, Km = 0.205 mg/ml, Vmax = 17.7.10 -3 ~unol/min. Zn and Hg showed potent inhibitory effects. 3.2) is a specific enzyme of the glyoxylate cycle, which plays a key role in the initiation of gluconeogenesis in germination oilseeds, This pathway is localized in glyoxysomes, single membrane organelles which are converted into peroxisomes when lipid reserves are depleted. The glyoxylate cycle enzymes have been the subjects of numerous conflicting reports typically addressing the problems of their synthesis (on free or bound ribosomes), targetting and suborganelar localization (membrane bound or matrix proteins). Recent biochemical studies have shown that MS from germinating soybean cotyledons is an interconvertible enzyme (membrane bound, aggregated or soluble fomls) which is significantly affected by its ionic environment (Henry et al., 1992, Plant Science 82:21-27) . Microscopy studies have now been carried out using immunofluorescence staining or immunogold-silver staining for light microscopy, and immunogold labelling for electron microscopy. All results consistently characterize MS as a glyoxysomal maWix enzyme. Chorismate mutase (EC 5.4.99.5) catalyzes the first step in that branch of the shikimate pathway which leads to the aromatic amino acids phenylalanine and tyrosine. We have isolated a cDNA for this enzyme from the higher plant Arabidopsis thaliana by complementing a yeast strain (aro7) with a cDNA library from A.

This is the first chorismate mutase cDNA isolated from a plant. The A. thaliana chorismate mutase expressed in yeast revealed allosteric control by the three aromatic amino acids as previously described for plastidic chorismate mutase isozymes. An attachment of Rubisco to chloroplast membranes under Cu++-stress has been described for wheat (Mehta et al., J. Biol. Chem. 267, 2810 -2816 . The displacement to the membranes has been discussed as a possible step in the degradation of this predominant stromal protein, e.g. during leaf senescence. In our recent experiments it became evident that such interactions are not restricted to Rubisco. Several other stroreal, and even a peroxisomal enzyme (glycolate oxidase), were also detected in the membrane fraction after 6 to 30 hours of oxidative stress induced in wheat seedlings by a treatment with 10 mM CuSO 4. The velocity and the degree of membrane attachment varied for different enzymes. Based on these results it was interesting to check the solubility of Rubiseo and its previously described 45 kD fragment which accumulates during the incubation of bean leaf discs under oxygen deficiency (Hildbrand and Feller, Experientia 47, A67) . Neither Rubisco nor the breakdown product were found to accumulate in the membrane fraction. Thus, at present, the steps involved in Rubiseo catabolism cannot be generalized. Different mechanisms depending on the metabolic situation and on environmental conditions should be considered.

Overexpression of the four parsley phenylalanine ammonia-lyase (PAL) isoenzymes in E. coli. Characterization of the enzymes and kinetic analysis of the inhibition by 2-aminoindan-2-phosphonic acid. Christoph Appert, J/irg Schmid, Jerzy Zorn* and Nikolaus Amrhein Institute of Plant Sciences, ETH-Z/Jrich, 8092 ZiJrich. *Technical University Wroclaw, 50-370 Wroclaw, Poland.

All four phenylalanine ammonia-lyase (EC 4.3.1.5) isoenzymes of parsley (Petroselinum crispum Nym.) were expressed as glutathione S-transferase fusion proteins in E. coil After affinity purification, the glutathione Stransferase moieties were cleaved off with factor Xa and the phenylalanine ammonia-lyases were characterized. The proteins form enzymatically active tetramers, even as fusion proteins. The four isoenzymes have comparable Km values for L-phenylalanine (15gM to 24.5/.tM), similar temperature (58~ and pH optima (8.5). The aminooxy-and phosphonic-analogues of L-phenylalanine are competitive inhibitors of the enzymes. 2-Aminoindan-2-phosphonic acid (J. Zo{a & N. Amrhein, Liebigs Ann. Chem. 1992,625) was found to be a potent slow-binding inhibitor of these and other phenylalanine ammonnia-lyases, both in vivo and in vitro. Chlorophyll a fluorescence has been used to compare the drought resistance of two varieties of Solanum tuherosum (AVR/I)C-12ST-19 AND SIBTEMA).

Fast fluorescence rise kinetics were measured during the first second of illumination with a time resolution of l0 Ms and 1Z bits in fluorescence intensity. The rise kinetics shows the typical steps called Fo -J -I -P. With this values different indexes have been defind with the goal to compare the behaviour of these two varieties of potato. Salicylic acid (SA) was proposed as a putative signalling molecule for the induction of systemic acquired resistance (SAR) in infected plants.

We studied its biosynthesis as follows.

Cotyledons of cucumber plants were injected with a solution containing Pseudomonas lachr~ans and fed with radioactive SA precursors by injection of 14C-benzoic acid (14C-BA) or 14Cphenylalanine (14C-Phe). Alternatively, leaf 1 of cucumber plants were inoculated with tobacco necrosis virus (TNV) and leaf disks were then vacuum-infiltrated using 14C-BA or 14C-Phe.

Free and bound 14C-phenolies were quantified by HPLC.

Incorporation of 14C into SA was found in the two plant-pathogen systems. I~C-BA gave mostly 14C-SA and an unknown polar compound.

14C-Phe gave also some 14C-SA, in addition to 14C-labelled lignin precursors like ferulic and p-coumaric acids.

The biosynthetic pathway of SA from Phe through cinnamic acid and BA will be discussed. After infection with a necrotizing pathogen, systemic acquired resistance (SAR) is often observed in plants. In cucumber, salicylic acid (SA) increases in the lower infected leaf as well as in the upper uninfected leaves. SA was also found in the phloem sap of infected cucumber plants and was proposed as a signalling molecule for the induction of SAR. We intend to clarify whether SA is translocated from the site of infection to the upper leaf, or wether it is made in the phloem tissue upon the action of a primary systemic signal.

One cotyledon of cucumber plants was first infected with Pseudomonas lachrymans and then fed with 14C-SA, 14C-benzoic acid (BA) or 14C-phenylalanine. After different times, cotyledons and first leaves were collected and the radioactive ~henolics were analyzed by HPLC. In some cases, 4C-SA was found in the first leaf. In addition, two unknown radioactive compounds were detected in leaf 1 after treatment with 14C-SA and 14C-BA respectively. We will discuss signalling processes for the induction of SAR in cucumber. One modern approach of creating virus resistant grapevine plants is the introduction of resistance genes into existing grapevine varieties by genetic engineering. The goal of this work is to use the coat protein (CP) mediated strategy to induce resistance to nepovirus in Vitis spp. As a first step, several chimeric nepovirus CP genes, were constructed by addition of various promoter regions upstream of the GFLV and ArMV CP regions.

Their ability of conferring resistance to nepovirus infection in transgenic Nicotiana benthamiana and N. tabacum plants will be presented. The best constructions will be used to transform several grapevine varieties in order to create nepovirus resistant grapevine plants.

$13-12

The bctalains axe a class of natural pigments found only in plants of the order Caryophyllales and in some fungi. The first step in betalain biosynthesis is the conversion of tyrosine to DOPA. Subsequently , DOPA is transformed to betalamic acid, the bctalain chromophorr through the action of DOPA-4,5-dioxygenase.

We have purified and characterised a copper-containing enzyme of -38kDa from Amanita muscaria pileus that catalyses the hydroxylation of tyrosine to DOPA. Considering that the enzymatic activity was restricted to the coloured parts of the mushroom, we postulate that it is involved in betalain biosynthesis. The enzyme featured a broad substrate specificity, and also oxidised the diphenols to their corresponding ortho-quinones, a reaction typical for tyrosinases. The implications of our findings on betalain biosynthesis axe discussed. Chlorophyll a fluorescence was measured under steady state conditions of pea and tomato leaves adapted to low light intensity (30 Wm-2s -I) at different temperatures (Havaux and Strasser, Z. Naturforsch. 45C, 1133 -1141 , 1990 ). When leaves were exposed for a short period (i0 min) to heat (25 to 45~ in darkness, the level of variable fluorescence decreased. When the heat stress was imposed in presence of low light, the variable fluorescence was much less affected and virtually no effect of heat treatment was observed until 37~ This protecting mechanism by low light was absent in algae and mostly absent in submerged water plants but fully present in free floating plants on the water surface. We conclude that a mechanism has been developed for higher land plants during evolution which protects the plants against heat damage on warmer days. Caryophyllales, e.g. Portulaca grandiflora, and in a few fungal species, e.g. Amanita muscaria. We analysed the similarity of a key enzyme of the pathway, DOPA-4,5-dioxygenase, in plants and fungi both at the protein and the DNA level. Antibodies against purified DOPA-4,5-dioxygenase from the fungus A. muscaria crossreacted with protein from P. grandiflora petals and two cDNA clones were obtained from this plant. Their similarity with fungal sequences and with other dioxygenases is discussed.

Kinetics of prollne hydroxylaUonj Intrucellnlur transport and C-terminal processing of the tobacco vacuolar cbltlnase.

Ernst Frcydl, Thomas Boiler and Jean-Marc Neuhaus Botenisehes Instimt, Abt. Pflanzenphysiologie, Hehoistxasse 1, CH-4056 Basel

The vacuolar chitinase A of tobacco (Nicotlana tabacum) is syndietlzed as a preproprotein with ma N-teaminal signal peptide which causes its synthesis in the ER mad a C-termlnal extension, which has been idcmified as the vacuolar targeting peptide (VTP) (1) and whleh is cleaved off from the maybe chitinase. It has recently been shown that mature chltin:~e A contains several hydi'oxyprolines within the short peptide spacer that links the N-terminal cysteine-rieh chitln-bindlng domain to the catalytic domain (2). We received specific antibodies against mature chitinasr (a kind gift from Dr. F.Meins. F/vii) and raised antihodiea against a synthetic VTP. We performed pulse-chase experiments and cell [ractlonation with 1) stably transformed tobacco plants expressing chitinase A or mutants lacking either the chitin-binding domain and spacer (ChitAH) or the vacuolar targeting poptid 9 (ChitAVTP) and 2) transient expression of the same conswacts in Nicotiana plumbaginifolia protoplasts. In both systems, proline hydroxylatino in chltinase A was detectable after 30 vain. and complete after 90-120 rain. The ehltinase intermediate forms were detected in the mlcrosomal and soluble fractions for up to 90-120 rain. The mature chitinase continuously accumulated only in the soluble fraction. In both systems ChitAH showed the same kinetics of intraceilular transport and processing. whereas the transport of CbitAVTP seemed more rapid.These results indicate that in vivo cbitinase A is synthetized as a proprotein that is than modified by proline hydroxylation. This second intermedinte form is then transported to the Oolgi apparatus and sorted to the vacuole. C-terminal processing occurs late in this pathway or in the vacuole. Infection of bean roots with the soil-borne phytopathogenic fungus Fusarium solani f.sp. phaseoli leads to a rapid increase of chitinase activity (~five-fold). Part of this increase in enzyme activity is due to transcriptional activation of the basic class I chitinase isoenzymes.

Semi-native polyacrylamide gels stained for chitinase activity using the substrate glycol-chitin revealed the appearence of three additional chitinase isoenzymes that are absent from uninfected control roots. Conventional protein purification techniques showed that Fusarium solani infected bean roots contain two basic and two acidic chitinase isoenzyrnes. Their physiochemical properties and possible biological function will be discussed. Proteasomes are multicatalytic protease complexes that function as a major nonlysosomal proteolytic system. They exhibit multiple endopeptidase activities that promote the intracellular turnover of abnormal polypeptides and short-lived regulatory proteins. Moss proteasome has been purified from protonemata by successive chromatography on MacroPrep Q, Bio-Gel A-1.5, Bio-Gel A-5 and HA. The molecular mass was estimated to be 1,300 kD by gel filtration. Gel electrophoresis of proteasome under nondenaturing condition gave a single band and two-dimensional gel electrophoresis identified the moss proteasome to be composed of 14 different subunits with molecular weight in the range 22-35 kD. Electron microscopy in aqueous solution showed that it is a cylindrical particle composed of four stacked rings with a diameter of 9 nm and a height of 14 rim. End-on projection established a 7-fold symmetry of the outer disks. Substrate specificity of proteasome indicates that it contains endopeptidase activity against substrates bearing hydrophobic, basic, acidic and glycine residues immediately preceding the cleavage site. Antibodies raised against moss proteasome reacted with proteasomes of other eukaryotes, including human.

$14-01

Hunger R.E., Hess M.W., Laissue J,A,, Mueller C.

The NOD (non obese diabetic) mouse, a widely used animal model for insulin dependent diabetes mellitus, shows in addition to mononuclear cell infiltration of the islets of Langerhans, massive cellular infiltrates of the submandibular and lacrimal glands with considerable tissue damage. Several lines of evidence suggest that both insulitis and sialadenitis represent autoimmune disorders, To investigate a possible role of tumor necrosis faetor-a (TNF-c0 and activated cytotoxic cells (NK cells, T cells) in the inflammatory process, tissue sections of NOD mice were hybridized with radiolabeled RNA probes specific for the detection of mRNA for TNF-r and the serine proteases granzyme A and B (GrA and Gr8) and perforin which are expressed in activated cytotoxic cells. TNF-e expressing cells were mainly located in infiltrates and are absent in non affected glands, whereas GrA, GrB and perforin expressing cells were distributed over the whole section with highest densities in the zones adjacent to parenchyma. The finding that activated cytotoxic cells are present in early stage of disease development indicates that cell mediated cytotoxicity may play a crucial role in initial tissue destruction. The role of TNF-a in autoimmune diseases is not known yet. The appearance, however, of cells expressing this gene in the infiltrate indicates a possible role of this cytokine in the progression of the inflammatory reaction, e.g. by increasing the lymphocyte traffic to this organ. We further demonstrated that the induction of iNOS was at the transcriptional level. In order to understand the underlying mechanisms we characterized the promoter region of the rat NOS gene. A 9enomic rat library was screened using a cDNA-fragment coding for the if-end of the iNOS cDNA (nucleotides 1-817). A 1,8 kb Hinc-II fragment containing the 5"-flanking region of the iNOS gene was characterized by DNA-sequencing. The transcriptional start site was determined by primer extension. Sequencing analysis revealed a multitude of possible cis-acting elements homologues to consensus sequences for the binding of different transcription factors including IFN~' response element (IRE), nuclear factor-KB (NF-~B), tumor necrosis factor response element (TNF-RE), 7f-activated sites (GAS) and one X-box. NF-KB, a transcription factor involved in signaling and immediate early gene activation during inflammatory processes was induced by treatment of mesangial cells with 2 nM of IL-113. This was shown by the appearence in nuclear extracts of the DNA binding activity of NF-~B using electrophoretic mobility shift assays (EMSAs) with a 32p-labeled ~r DNA probe. Pyrrolidinedithiocarbamate (PDTQ), an inhibitor of NF-KB, suppressed the expression of iNOS mRNA in mesangial cells without affecting the Dibutyryl-cAMP-triggered increases in NOS mRNA levels. This data suggest that the IL-I ~ signalling pathway responsible for NOS expression requires NF-K8 activation and is definetly different from the signalling cascade directed by cAMP. Recent findings indicate that the multifunctional cytokine IL-6, which plays a key role in irm-nune and inflammatory responses, also exerts specific effects in the central nervous system (CNS). For example, IL-6 promotes neuronal survival, induces differentiation and modulates neurotrophin production. Using the very sensitive technique of reverse transcription combined with polymerase chain reaction the developmental profile of IL-6 and its receptor mRNAs was analyzed in various rat brain regions. Our results indicate that both genes are expressed in a region-specific manner and are developmentally regulated. These findings support the hypothesis that IL-6 is involved in differentiation and maintenance of neuronal subpopulation in the CNS. Interleuldn 2 (IL-2) expression is strictly dependent on a signal transmitted by the T cell receptor upon antigen stimulation. This signal can be mimicked in vitro with phorbol esters and calcium ionophores. We have found that stimulation with Ca-ionophore alone induces expression of IL-2 mRNA, but does not induce secretion of IL-2. In polysome gradient fractionation of cells stimulated with phorbol esters and ionophore, the fractions containing IL-2 rrtRNA are found at the bottom of the gradient, bound to polysomes. However, in ionophorestimulated cells the fractions containing IL-2 mRNA are clearly shifted to the top of the gradient, and therefore are not lzanslated. This effect is specific for IL-2, as other mRNA's like 1~2-microglobulin are not regulated. This data suggests that IL-2 gene is translationally regulated. In addition, in vitro experiments have shown that the lack of translation of IL-2 mRNA in ionophore-stimulated cells is due, most likely, to the presence of a translational regulator that specifically inhibits translation of IL-2 mRNA. Assuming the presence of a translational regulator that specifically represses IL-2 mRNA translation, we can predict that, if the represser is in excess, even upon a second stimulus that is by itself able to induce secretion of IL-2, there will be no translation of IL-2 mRNA. The resuhs of such an experiment confirmed our prediction. Polysome gradients showed that after ionophore stimulation, IL-2 mRNA gets "stacked" with one ribosome, and no further ribosome loading takes place. If malaria is highly endemic, every febrile child is a presumptive malaria case, and there is no differential diagnosis based on clinical, immunological or parasitological grounds.

We evaluated the diagnostic usefulness of interleukin-2 receptor (IL-2R), tumor necrosis factor-receptors 55 (TNF-R55) and 75 (TNF-R75) .

Sera came from Tanzanian pediatric fever patients and controls, all enrolled in the Kilombero Malaria Project. We found that all 3 receptors marked pediatric fever episodes, whereas stability of observed levels was highest for TNF-R75 and lowest for TNF-R55. TNF-R55 levels reflected severity of the fever attack. Regardless of the presence of a febrile illness, TNF-R75 levels were strongly associated with parasite density. Immunological markers can be applied in rapid pre-and post-intervention morbidity assessments, validation of health interviews and monitoring of convalescence. We have recently shown that human eosinophils produce mRNA for interlenkin (IL)-8 when stimulated with calcium ionophom (Eur. J. Immunol. 23 (1993), 956). We now present evidence that human eosinophils contain preformed IL-8 proteIn although they do not express mRNA for IL-8 as measured by RT-PCR. IL-8 protein expression was determined using specific anti-human IL-8 monoclonal antibodies by Western blotting, FACS analysis and immunohistochemistry. Moreover, preformed IL-8 was released into supernatant by 99% pure eosinophils after priming with GM-CSF or IL-5 and subsequent 25-min stimulation with PAF, RANTES, ionomycin or PMA, however not with IL-1 or IL-2, as measured by ELISA. This observation implies that activation of protein kinase C and/or inWacellular calcium mobilization am necessary events for IL-8 release in human eosinophils. The determined amounts of preformed IL-8 protein was higher in patients with asthma compared to normal individuals suggesting a role for eosinophil derived IL-8 in asthma. Since the eosinophil is the predominant cell in the asthmatic airways, we determined 1L-8 concentrations in bronchoalveolar lavage (BAL) fluids from normal individuals and asthmatic patients. Indeed, IL-8 concentrations in BALs from both groups were similar to those seen in the supematants released by activated eosinophils. The role for 1I-,-8 in asthma remains to be determined. To study the effect of canthaxanthin in vitro, the hydrophobic compound was introduced in vivo into chick high density lipoproteins (HDL) by canthaxanthin feeding. The effects of HDL and HDL associated canthaxanthin on two types of cultures have been tested: flat sedimented cell cultures of embryonic chick neuronal retina, retinal Figment epithelium (RPE), brain and meninges, and in reaggregate cell cultures of the neuronal retina. At high canthaxanthin concentrations the formation of colored, birefringent entities were induced in the neuronal retina in vitro. In line with human data, parameters of cellular function and cell differentiation remained unaffected by canthaxanthin at these concentrations.

By contrast, FIDL was found to induce various cell type dependent effects. Proliferation of meninges cells was decreased at low HDL concentrations as concluded from.light microscopic examination and indicated by protein content, and lysosomal and mitochondrial activity of the cultures (IC50:15-30 mg HDL apoprotein/L medium). These parameters, however, were increased in brain cell cultures, while they remained unaffected in cell cultures of neuronal retina and RPE. Concentrations above 0.3 g. HDL apoprotein/L caused a reduction of glial cell differentiation. $14-08

TNF-c~ had close-dependent antiproliferative activity on hormone-dependent human breast cancer cells that represent an early stage of the disease, whereas hormone-independent cells representing a late stage were not affected. However, in the presence of 1000 U/ml interferon ?(INF), high concentrations of TNF-c((lnM) were able to inhibit the growth of hormone-independent cells. Using specific monoclonal antibodies in ftow cytometry, no significant changes of either the p75 or the p55 TNF receptors were observed upon INF treatment of hormone-independent cells. This indicates that the increased responsiveness of these cells for TNF-c~ is not due to a change in available TNF receptors. The affinity of the receptors for TNF-R are one order of magnitude different for the two cell types. INF treatment shifted the ECs0 of hormone-independent cells towards the ECs0 of hormone-dependent cells. Thus, the basis for the increased TNF-sensitivity of hormone-independent cells in the presence of INF might be the interconversion of TNF receptors from low-to high-affinity. The adhesion of tumor cells to endothelium is the first step in the processus of metastasis, and is frequently dependent of cytokinemediated activation of endothelial cells. Cytokines are also involved in nitric oxide (NO) production by various cells. These experiments were designed to evaluate NO production during tumor cell adhesion to endothelium. Rat brain-derived endothelial and rat colon carcinoma cell lines were used during these experiments. NO was evaluated with the Griess reagent. Activation of endothelial cells with cytokines (TNF-d, and IFN-~ ), only slightly increased (10-20 %) the adhesion of tumors cells to endothelium. Activation of endothelial cells with TNF-oL and IFN-~, induced a 4-8 fold increase of NO production. However, addition of tumor cells to the endothelial monolayer, either directly or in a semi-permeable membrane, decreased the cytokine-induced NO production. These results indicate that NO production is modulated during the proeessus of adhesion of tumor cells to endothelium. is induced in rat mesangial cells by inflammatory cytokines such as interleukin 1 ]3 (IL-1 ~) and requires BH 4 as a cotactor. 2,4-Diamino-6-hydroxypyrimidine (DAHP), a selective inhibitor of GTP cyclohydrolase I, the rate-limiting enzyme for BH 4 synthesis, potently suppressed IL-I~induced NOS activity, measured as nitrite production. Inhibition of NO synthesis by DAHP was reversed by sepiapterin, which provides BH 4 via a salvage pathway. Sepiapterin dose-dependently augmented IL-1 I],-stimulated NO synthesis, indicating that availability of BH 4 limits the production of NO in cytokine-stimulated mesangial cells. N-acetylserotonin, an inhibitor of the BH 4 synthetic enzyme sepiapterin reductase, completely abolished IL-lJ~-induced NO formation whereas methotrexate, which inhibits the pterin salvage pathway, displayed only a moderate inhibitory effect, thus suggesting that mesangial cells predominantly synthesize BH 4 tom GTP. In conclusion, these data demonstrate that BN 4 synthesis is an absolute requirement for cytokine induction of NOS in mesangial cells. Inhibition of BH 4 synthesis may provide new therapeutic approaches to the treatment of pathological conditions mediated by NO. -1 [3) can induce a macrophage-type of nitric oxide synthase (iNOS). Northern-blot analyses of iNOS mRNA-levels and nuclear run-on experiments of control and IL-1 ~ stimulated mesangial ceils revealed that the induction of iNOS was at the transcripUonal level.

Here we demonstrate that platelet-derived growth factor BB (PDGF-BB) can suppress the IL-113 dependent inducUon of the iNOS gene. Nortllern-blot analyses of cellular RNA isolated from mesangial ceils that were coincubated with IL-1 [3 (2nM) and PDGF-BB (10 ng/ml, 100 ng/ml and 300 ng/ml) revealed a dosedependent suppression of iNOS mRNA-levels. Using run-on experiments with nuclei from mesangial ceils after stimulation with IL-I~ (2nM) and PDGF-BB (100 ng/ml) we demonstrate that the inhibition occurs at the transcriptional level. Basic fibroblast growth factor (bFGF), on the other hand, potentiates the IL-1 dependent stimulation of iNOS. Coincubation of mesangial cells with IL-113 (2nM) and bFGF (3ng/m[, 10ng/m[, 30ng/m[, 100ng/m0 dose-dependently superinduces iNOS mRNA levels as shown by Northern-blot analyses of total cellular RNA form control and stimulated cells. Run-on experiments with nuclei from mesangial cells stimulated with IL-113 (2nM) and bFGE (100ng/ml) revealed an enhanced transcriptional activity of the iNOS gene. Thus, PDGF-BB and bFGF differentaity modulate the IL-I~ dependent expression of the iNOS gene and are therefore an excellent model system to study the crosstalk between signal transductJon pathways. 

We report that IFNyR-/-mice have an increased resistance to lipopolysaccharide-induced toxicity (LPS). LPS-induced lymphopenia, thrombocytopenia and weight loss seen in wild type mice were attenuated in IFNTR-/-mice. IFNTR-/mice survived in the D-galactosamine-LPS model when conditions were 100% lethal for wild type mice, correlating with serum TNF levels of up to 10 fold higher in wild type mice. Bone marrow and splenic macrophages from IFNyR-/-mice had a 3 to 5 fold decreased LPSbinding capacity, with serum from these mice lowering macrophage LPSbinding by a further 50%. Thus, depressed TNF synthesis, diminished expression of LPS receptors and low plasma LPS-binding capacity in the mutant mice likely combine to manifest in the resistant phenotype of IFNy R-/mice to endotoxin. In a complementary approach we have screened the 5' portion (-7.8 kb/+4.1 kb) of the gene for tissue specific and/or inducible DNaseI hypersensitive sites. In chromatin from fibroblasts or kidney epithelial cells this segment of the IL2Ra gene does not contain any DNaseI hypersensitive sites.

In contrast, in resting T-and B-lymphocytes, as well as in activated T cells, two sites (DHI, -0.05 kb; DH3, -5.8 kb) were found. A third site (DH2, -1.35 kb) is detectable in T cells that have been activated with concanavalin A and IL2. This site maps to the same position as cisacting regulatory sequences that are required for the response of the IL2Ra gene to signals from the IL2 receptor. Interleukin-6 (IL-6) production in cultured human dermal fibroblasts can be stimulated by interleukin-i (IL-I} added to the culture medium. Absence of fetal calf serum and of growth factors (insulin, EGF, T3) reduced the stimulation by more than 50 %. EGF and T3 and even more choleratoxin increased the IL-6 production in absence of fetal calf serum. The effect was dose dependent. 2% human AB serum substituted for i0 % fetal calf serum. Pretreatment of the cultures with serum or growth factors prior to stimulation with IL 1 had a stimulatory effect (serum, T3, EGF) or an inhibitory effect (hydrocortisone, choleratoxin). Our results suggest that the IL-I signal transduction to IL-6 is modulated by serum components and growth factors as well as through c-AMP produced by choleratoxin. Interleukin-6 (IL-6) is produced by a variety of cells and plays a central role in host defense mechanisms. Its function include induction of IL-2 and IL-2 receptor expression, proliferation and differentiation in T-cells. In cocultures of human dermal fibroblasts and allogenic human T-cells a cell number dependent 10 to 100 fold increase of the IL-6 production could be demonstrated after 24h and 48h. Conditioned medium from Tcells and from fibroblasts failed to induce IL-6 secretion in fibroblast and T-cell cultures, respectively. No up-regulation of IL-6 production was found in cocultures of fibroblasts with Tcells killed by ethanol and in T-cell culture incubated with recombinant hlL-lc~. After separation of the cells IL-6 production in fibroblast and T-cells returned to precoculture levels. Our results indicate that cell-cell contact more than soluble factors must be involved in the up-regulation of IL-6 synthesis in cocultures of T-cells and fibroblast. Using cocultures of autologes human T-cells and fibroblasts, we are currently investigating whether the cell contact essential for the IL-6 production depend on immunolcical interactions of T-cells and fibroblasts. Rapid growth of tumor often results in oxygen and nutritional deprivation leading to extensive necrosis, which is one of the characteristics of glinblastomas (Gbl). Neoplastic cells surrounding necrotic areas often present a peculiar arrangement of cells called "pseudo-palisading" -as a hallmark of this entity, together with abundant neovascularization. A special biological milieu is probably formed in these palisading areas by certain unknown cytokines and/or growth factors induced by the necrotic process. Plate et al. (1992) reported the expression of vascular endothelial growth factor (VEGF) in the palisading cells. We previously demonstrated that Gbl cells produce IL-8, which is known to be an angiogenic factor. The aim of this study is first to assess if IL-8 is expressed in the palisading ceils, and second to determine if hypoxia can trigger IL-8 gene expression. In rive observation using in siru hybridization and immunohistuchemistry showed that IL-8 is highly expressed specifically in the palisading cells. We also demonstrated the mRNA expression of IL-8 receptor in the endothelial cells adjacent to necrotic area. To simulate the in vivo condition, two in-vitro models are currently developed to determine if IL-8 is induced by hypoxic insult on cells. First, Gbl cell lines are cultured in the absence of oxygen and IL-8 expression is assessed by Northern blot analysis. Preliminary results have demonstrated the induction of IL-8 by hypoxia. Secondly Gbl cells are grown in spheroids until development of central necrosis. The IL-8 expression will be assessed by in situ hybridization combined with immunohistochemistry to determine if the ceils surrounding necrosis express IL-8.

It has been proposed by Taniguchi and his colleagues that activation of the interferon (IFN)-I3 gene by virus or double-stranded RNA involves induction and modification of IRF-1. Overexpression of IRF-1 can induce expression of the IFN-B gene in certain cells. A role of IRF-1 in the activation of IFN-a genes has also been proposed. Furthermore, IRF-1 has been claimed to play the role of a tumor suppressor gene. We have generated mice in which both IRF-1 alleles were disrupted and found them to develop normally. No spontaneous tumors have been observed so far. Injection of poly(I)-poly(C) resulted in the same levels oflFN in blood of wildtype and null mice, and equal IFN-ct mRNA levels in spleen. Induction of wild type and IRF ~ embryo fibroblasts (MEFs) with virus resulted in the same levels of IFN-a and IFN-I] mRNA respectively. In the case of polyO)-poly(C) induction, the levels of both 1FN mRNAs were higher in wild type than in mutant cells; this difference was abolished if the cells were first primed with type I IFN. We conclude that IRF-1 does not play an essential role in the induction oflFN genes.

Mitsuhiro TADA, Annie-Claire Diserens, Isabelle Desbaillets, Marie-France Hamon, Nicolas de Tribolet, Erwin Van Meir; CHUV, Lausanne There is increasing evidence that a variety of cytokines produced by glioblastoma (Gbl) cells modulate the tumor growth by affecting the host's immune response and by stimulating neovascularization. Cytokines such as IL-I, IL-6, IL-8, MCP-I, IL-10 and TGF-132, affect each other's production and receptor system and seem to form a cytokine network. Among them, IL-1 may play a pivotal role, since IL-I can induce or increase the production of other cytokines (IL-6,IL-8,MCP-I,TGF-132) and modulate their receptor expression. To test this hypothesis, we investigated co-expression of cytokines and their receptors in 15 Gbl cell lines and 15 Gbl tissues using RT-PCR, Northern blot analysis, immunnhistnchemistry and ELISA quantification. A majority of the Gbl cell lines and Gbl tissues expressed IL-I~, IL-113, type I and type II receptors, indicating the presence of an IL-1 autocrine loop. IL-1 stimulated growth of some of the cell lines. A positive correlation of IL-6 and IL-8 expressions with the presence of an IL-I autocrine loop in the cell lines was found. Immunohistochemical studies showed the in rive co-expression of the IL-I family and the secondary cytokines (IL-6, IL-8) in Gbls. Antisense oligonucleotide to IL-lc~ and/or IL-113 partly suppressed this secondary cytokine expression. These results demonstrate that the IL-1 autocrine loop plays a central role in the formation of the cytokine network in Gbls. Double-stranded RNA-dependent protein kinase (PKR) is induced by type I interferon (IFN) and is implicated in the establishment of the antiviral state. Upon activation, PKR phosphorylates the eukaryotic initiation factor EIF-2, thereby throttling protein synthesis. It was suggested that PKR may be essential for the induction of IFN-13 gene expression by both virus and double-stranded RNA and for the activation of at least some IFNinducible genes. Recently it was proposed that PKR is a tumor suppressor gene because 3T3 ceils overexpressing inactive human PKR gave rise to tumors in nude mice (Koromilas et al., 1992; Meurs et al., 1993) . We have produced homozygous PKR knockout (PKR ~176 mice and found that they develop normally and have no striking phenotype. Induction of IFN-c~ and IFN-13 gene transcription by virus was unimpaired in fibroblasts derived from PKR ~ mice, as was the induction of several IFN-stimulated genes. So far, no spontaneous tumor formation was observed. PKR o/o embryonal stem ceils showed no increased malignancy in nude mice as compared to their wild type counterparts. We conclude that PKR does not play an essential role in viral induction of the IFN genes and that its tumor suppressing effect, if any, is redundant. TGF-13 plays an important role as a negative regulator of hepatocyte proliferation in liver regeneration. Exposure of rodents to nongenotoxic carcinogens like Cyproterone Acetate (CPA), Thioacetamide (TA) and Phenobarbital (PB) causes abnormal hepatocyte proliferation and liver tumors after long term treatment. This suggests that these chemicals have either a direct mitotic activity or impair the negative regulatory system for growth.

Therefore the inhibitory activity of TGF-6 on DNA-synthesis induced whith CPA was investigated in cultured rat hepatocytes. After a 48h exposure to CPA, a dose dependent increase in DNA synthesis (3H-TdR incorporation) was observed. The maximum effect was attained with 12 pM CPA (up to 4.3 fold) which was stronger than with 10 ng/ml EGF (up to 2.8 fold). TGF-81 and TGF-132 inhibited this response dose dependently (IDso = 0.1 ng/ml) and reduced it to control levels at 1 ng/ml. These findings show that the TGF-6 mediated pathway for negative growth control in hepatocytes is still responding after CPA treatment. Mouse Mxl protein is an interferon (IFN) induced GTPase with intrinsic antiviral activity against influenza A viruses. It accumulates in the cell nucleus and inhibits the transcription of influenza virus genes, Recently, we have shown that Mxl exerts its inhibitory activity by interfering with the function of the viral polymerase subunit P82. PB2 is responsible for recruiting 5" ends of cellular mRNAs as primers and for the elongation of nascent viral transcripts. To elucidate which PB2 dependent step of viral mRNA synthesis is the target of Mxl action, we examined the activity of recombinant Mxl protein in an in vitro transcription system of influenza virus. First, we expressed wildtype and mutant Mxl proteins, carrying a hlstidinetag at their N-terminus, in E. coil and purified them under nondenaturing conditions by Ni-chelate affinity chromatography. Mxl efficiently inhibited viral transcription in vitro, while mutant Mxl proteins, lacking antiviral activity in vivo ,were inactive. Moreover, the use of capped synthetic RNA primers revealed, that Mxl almost completely abrogated the elongation of viral transcripts, whereas the the cap-binding and endonuclease activity of PB2 was not affected. A50 $14-28 Neutralization rates of TNF-a from eight animal species by a monoclonal antibody against human TNF: implication for receptor action?

We have recently developed a sensitive bioassay for measuring porcine TNF-a based on homologous PK(15) cells. This bioassay is 100-to 1000-fold more sensitive than the widely used L929 bioassay, Also, human TNF-a and human TNF-8 were detected with a slightly higher sensitivity in the new test compared to the L929 bioassay, We now show that also canine, ovine, bovine, equine and caprine TNF-a can be detected with the PK(15) bioassay. We tested a murine monoclonal anti-human TNF-a antibody for its capacity to neutralize TNF-a from eight different animal species. The action of this antibody revealed that neutralization of TNF bioactivity is a complex phenomenon, indicating species-specific differences in the interaction with TNF receptors. To study this differential neutralization, we are cloning the porcine TNF receptors for expression in a heterologous cell system.

Administration.of high dose r-TNF(~ in combination with IFN 7 in isolation and perfusion of the limbs in melanoma patients has shown to be very promising with a response rate greater that 80%. This observation prompted us to investigate the possible expression of the TNF receptors in melanoma cells using the monoclonal antibodies UTR-1 specific for the TNF Type A (55 kDa) receptor and HTR-9 specific for the TNF Type B (75 kDa) receptor. Flow cytometric analysis of cultured melanoma cells showed the presence at low level of the TNF Type A and to a slightly higher level of the Type B receptor. Similar results were obtained in vivo by immunohistochemistry on fresh tumor raateriai, Treatmem of melanoma cells in culture with the-AMP induced up to a 2 fold increase in the number of Type A TNF receptors with only a minimal change in the number of Type B receptors. This increased expression of TNF receptors is confLrmed by direct binding experiments using 125I-labeled TNFtx. The number of CPM's bound on dbc AMP treated cells was about 0.5-3 fold higher than on untreated control calls, likewise incubation of melanoma cell lines with IFNy increased the specific binding of subsequently added 125I-labeled TNFct. From flow cytometric analysis using the two anti-TNF receptor antibodies it became evident that fliNT was able to increase the expression of both Type A and B receptors depending on the cell line used.

Characterization of a murine tumor necrosis factor (x-LacZ reporter construct Tumor necrosis factor ot (TNFcQ is a prominent mediator of a variety of different pathologies such as endotoxic shock syndrom and rheumatoid arthritis. It also plays a crucial role in host defence against bacterial and viral infection. Up to now, only few data about signal pathways leading to TNFo~ induction are available. An easy to handle TNFc~ -LacZ reporter assay will represent a useful tool to study signal transduction on the one hand and provide data about compounds with potential TNFo~ inducible activity on the other hand.

In order to develop a routine macrophage cell line capable to easily report TNFc~ induction, different TNFe-LacZ reporter constructs were assembled. Data about functionality of the different constructs in routine macrophages will be presented in the context of TNFot expression.

CLONING OF A NOVEL PROTEIN BINDING TO LYMPHOKINE, FOS AND MYC mRNA WITH UNEXPECTED ENZYME ACTIVITY J. Nakagawa, H-P. Waldner, S. Meyer-Monard, and Ch. Moroni Inst. Medizinische Mikrobiolegie, Univ. Basel An AU-rich sequence in the 3' untranslated region of lymphokines, c-los, and c-myc proto-oncogene mRNA is responsible for their rapid decay.

We have purified a 32 kd protein by an AUUUA affinity column from human brain.

Partial amino acid sequence was obtained and the corresponding eDNA has been cloned. Unexpectedly the eDNA sequence exhibited significant homology to enoyl CoA hydratase and bacterially expressed recombinant protein indeed showed the activity. While rat hydratase did not show RNA binding activity, the recombinant protein bound to cfos, c-myc, AUUUA cluster of IL-3 mRNA, and Adeno virus IV, but not to mutated Ad-IV, nor irrelvant transcript. The binding was competed out by poly U.

Interestingly the protein seems to be processed from a larger precursor.

We suspect that the protein plays a role in mRNA turnover and perhaps in oncogenesis.

Institute for Medical Microbiology, University of Basel In T cells, the immunsuppressive agent CeA is known to inhibit Ca 2+ dependent induction of lymphokine mRNA transcription. In contrast, the immunsuppressive drug rapamycin inhibits the lymphokine induced proliferation. We have analysed the effect of these drugs on a v-H-ras induced autocrine mast cell tumor line which expresses IL-3 constitutively. CsA and rapa inhibited autocrine growth by different mechanisms,, because added IL-3 could antagonize inhibition by CsA, but not by rapa. Whereas CsA acted by downregulation of IL-3 mRNA expression, rapa blocked IL-3 induced proliferation. CeA also inhibited IL-3 superinduction by Ca-ionophore, whereas rapa did not. Nuclear run on assay indicated that the mechanism is posttranscriptional. Therefore, we have introduced IL-3 transgenes with and without the AU-rich region in the 3' UTR of the mRNA into a tumor line. In contrast to the wild type, the expression of the mutant construct was insensitive to CsA. Since the mutant lacks sequences known to regulate rapid mRNA decay, we suggest that CsA affects the regulation of IL-3 mRNA stability. Some non-hematopoietic cell lines produce both SCF and its receptor, the c-kit protein (c-kit). Testing the hypothesis that SCF may be involved in secondary growth of NB bone marrow metastasis, we found and previously communicated (Beck D. et al, Proc AACR, 34, 52, 1993) a low or absent expression of c-kit in NB cells. The mRNA and surface expressions of SCF gene in NB cell lines grown in chemically-defined medium were evaluated by northern blot analysis and flow cytometry. The SCF antigenic concentrations in culture supernatants were determined by ELISA and growth-inhibition experiments performed by pre-incubating NB cells with anti-SCF antibodies prior to 3H-thymldine uptake assay. The SCF mRNA was expressed in 4 of 5 tested lines but no membrane-bound antigen could be detected on those cells. Detectable levels of SCF in the supernatants were found in 5/7 lines (range : 39-96 pg/mL). Blocking experiments resulted in a decrease of DNA synthesis in i out of 7 lines only. From these and previous results, we conclude that SCF is synthesized and released by many NB cells in vitro but the functional ~ole of it remains undetermined since SCF does not appear to be usually involved in autocrine or paracrine growth stimulation of NB cells (supported by Swiss FNRS grant 3200-031005). Expression of the v-H-ras oncogene in the IL-3 dependent PB-3c mast cells leads to the generation of two classes of IL-3 autocrine tumors in vivo. Autocrine IL-3 expression in class-1 tumors results from increased IL-3 mRNA stability due to the loss of negative trans-acting posttranscriptional control. This defect can be corrected by somatic cell fusion to the nontumorigenic parental PB-3c resulting in downregulation of oncogenic IL-3 expression and concomitant tumor suppression. Expression of the v-H-ras oncogene remains unaffected in class-1 hybrids.

In contrast, class-2 tumors are characterized by transcriptional activation of the IL-3 gene due to the insertion of an endogenous retroviral element (intracisternal A-particle) which cannot be overcome by ceil fusion (see Hirsch et aL, 1993, J.Exp.Medicine 178, 403-411) . Although v-H-ras is required for generation of either class of tumors, expression of anti-sense ras constructs inhibited only the proliferation of class-I tumor cells. Experiments are underway to characterize the target and the nature of the class-1 alteration. Nitric oxide (NO) promotes vasorelaxatlon of renal vessels in vitro. The purpose of the present study was to assess the role of NO in regulating renal hemodynamics in vivo. Renal blood flow (RBF) and glomerular filtration rate (GFR) were studied in anesthetized mechanicallyventilated newborn rabbits both before and after the administration of a NO synthesis inhibitor, NG-nitro-Larginine methyl ester (L-NAME), at a dose of 50 pg/kg followed by 10 Jzg/kg x rain. Such a dose of L-NAME did not modify mean arterial pressure or pulse rate. By contrast, the renal vascular resistance increased by 31 + 9% (P < 0.005) while RBF decreased by 20-+6% (P < 0.02). GFR and urine flow rate remained constant. The overall results suggest that in normal conditions NO may play a role in decreasing the renal vascular resistance of the immature kidney, without altering systemic blood pressure.

Interleukin-4 (IL-4) is a T-cell derived lymphokine which, in addition to its effects on the immune system, is also able to suppress the growth of certain tumor cells. IL-4 reduced the growth rate of four human colon tumor cell lines up to 70% in a dose-dependent manner. Three other cell lines showed no significant alteration of 3H-thymidine incorporation or colony formation in the presence of 100 u/ml IL-4. Responder and nonresponder tumors were immunopositive for IL-4 receptor (IL-4R) expression in flowcytometric analysis, using monoclonal antibodies raised against recombinant IL-4R and bound biotinyated IL-4. Responsive tumors expressed more receptors.

11-4 binds with high affinity to a 130 kDa transmembrane receptor (IL-4R), through which the extracellular signal has been shown to be transduced. Coimmunoprecipitation and chemical crosslinking studies have enabled us to demonstrate IL-4R target proteins. Tyrosine phosphorylation of various proteins between 70 and 170 kDa appears to be initiated during IL-4mediated signaling events in colon tumor cells. The cloning of human IL-4R target proteins will contribute to the understanding of the IL-4 signal transduction pathway at the initial stage. The renal effects of endothelin-1 were investigated in 16 anesthetized and meehanleally-ventilated newborn rabbits. Each animal acted as its own control. In 8 newborn rabbits, a bolus injection of 5 nmol.kg "1 of endothelin-1 caused an initial fall in mean blood pressure (MBP) followed by a gradual but significant increase in MBP that lasted for 45 rain. The dramatic increase in renal vascular resistance (+ 28 -+ 4%) induced by endothelin led to a fall in glomerular filtration rate (-12 -+ 4%) and renal blood flow (-16 _+ 3%). In spite of the reduction of GFR and RBF, urine flow and sodium excretion rates increased significantly (+ 20 _+ 5% and + 49 _+ 9%, respectively). In 8 additional newborn rabbits, a bolus injection of lnmol.kgaof endothelin-1 -a dose that usually induces marked renal and systemic vasoconstriction in adult models -did not affect systemic or renal hemodynamics. In conclusion, endothelin induces renal and systemic vasoconstriction, and affects water and sodium homeostasis during the neonatal period. However, these effects occur under higher doses than those used in adult animals, possibly reflecting receptor immaturity and/or interference of high levels of counteracting hormones.

Ontogeny of the endocrine pancreas in Xenopus laevis C, Maake 1 , W. Hanke 2 and M. Reine~ke 1 , 1 Institute of Anatomy, University of Z0rich, Switzerland and Department of Zoology II, University of Kaflsruhe, F.R.G.

The pancreatic islets of anurans show insulin (INS), glucagon (GLUC), somatostatin (SOM) and pancreatic potypeptide (PP) containing cells. Since only limited information exists on the ontogeny and on the presence of insulin-like growth factor 1 (IGF-1), we studied the development of the gastro-entero-pancreatic (GEP) system in Xenopus laevis using immunohistochemical techniques. Singular INS-immunoreactive (-IR) cells were observed in the pancreas as early as on stage 41. At stage 43, the first PP-IR cells were found in the GEP system followed by SOM-IR and GLUC-IR cells. The first pancreatic islets consisting of INS-IR and GLUC-IR cells occurred around stage 50. Between stage 54 and 60, i.e. after the onset of metamorphosis, the endocrine cells decreased in number and the islets partly disintegrated. Starting with stage 61, the islets reformed and the amount of endocrine cells increased reaching the adult status at stage 63. Around stage 52, IGF-l-immunoreactions appeared. IGF-1immunoreactivity was found in PP-IR and GLUC-IR cells but not in INS-IR and SOM-IR cells. It is assumed that islet-derived IGF-1 may p~ay an important role during metamorphosis in Xenopus. 

T. CATHOMEN AND R. CATTANEO, Institut for Molekularbiologie I, Universit~t ZOrich, Hfnggerberg, CH-8093 ZOrich The signals used for synthesis of the measles virus fusion (F) protein are poorly understood: a long (>570 bases) untranslated region is followed by three in frame AUGs at codons 1,4 and 15. F is a type I transmembrane protein with a postulated signal sequence of 31 amino acids. We confirmed that the 46 aminoterminal residues contain a signal peptide by transfering them to another protein. With the other envelope protein hemagglutinin, F protein induces virus-cell and cell-cell fusion. Mutagenesis of the three AUGs indicates that proteins initiated at codon 1, 4 or 15 are all functional, but show slightly different fusion efficiencies. Forms translated from AUG 1 or 4 appear larger in electrophoresis than forms initiated at AUG 15, suggesting that two different signal peptide cleavage sites are used. We are currently investigating whether both F protein forms are incorporated in viral particles. The production of protein isoforms with staggered amino-termini is common in cytoplasmic and type II transmembrane proteins, but signal sequences usually preclude a similar organization of type I transmembrane proteins.

Beguin P., Beggah A.T., Rossier B.C., Jaisser F. and Geering K.

Institut de Pharmacologie et de Toxicologie de l'Universit6, CH-1005 Lausanne Recent experimental evidence suggests that Na,K-ATPase might be a candidate for regulatory phosphorylation by protein kinases A and C (PKA and PKC).

To verify this hypothesis, we have mutated several serine and threonine residues in consensus phosphorylation sequences of the ~subunit of Bufo marinus Na, K-ATPase. The mutants were expressed in X~nopus ooeytes and phosphorylation was studied in oocyte homogenates upon stimulation of oocyte, PKA and PKC.

Our results indicate that a unique phosphorylation site for PKA is located at serine 943 in the C-terminus of the ~-subunit but its phosphorylation can only be revealed in the presence of detergent.

On the other hand, mutations of several serine and/or threonine residues located in consensus sequences for PKC phosphorylation did not or only partially abollsh phosphorylation by PKC.

In particular, mutations close to the catalytic phosphorylation site of the ~-subunit influenced phosphorylation by PKC, suggesting that either this region contains a real phosphorylation site or is part of a conformational domain necessary for phosphorylation by PKC.

The heat-stable antigen (HSA or mouse CD24) gene is differentially regulated but has a housekeeping promoter.

Roland H. Wenger*, Georges KShler and Peter J. Nielsen. Max Planck Institut f(Jr Immunbiologie, St0beweg 51, D-79108 Freiburg i. Bsg. "Present address: Physiologisches Institut der Universit&t Zi.)rich, Winterthurerstrasse 190, CH-8057 Z0rich, Expression of the GPI-anchored murine glycoprotein heat-stable antigen (HSA) shows tissue-specific as well as developmental regulation. During the maturation of several hematopoietic lineages, HSA expression is generally high in immature precursor ceils and low or absent in terminally differentiated cells. We present evidence suggesting that this regulation of the HSA gene (Cd24a) occurs at the transcriptional level. In addition, sequence and methylation analysis of the Cd24a promoter revealed characteristics of both "housekeeping" and tissue-specific promoters, including a methylation-free, Hpall Tiny Fragment (HTF) island, multiple putative SP1 and AP-2 consensus binding sites and a TATA box. Functional analysis of a 0.6 kb DNA fragment containing these elements fused to the CAT reporter gone in transient transfection experiments showed activity in both HSA expressing and non-expressing cell lines with a strength similar to that of the HSV TK promoter. Large fragments from the flanking region of the Cd24a promoter did not influence the ubiquitous nature of this promoter, Finally, the Cd24a, Cd24b and Cd24c genes were mapped to mouse chromosomes 10, 8 and 14 respectively. We have cloned, sequenced and characterized the gone for the m__itochondriat NADH-cytochrome b5 reductase (MCRI) of Saccharomyces cerevisiae. Surprisingly, this gone encodes two mitochondrial isoforms of the flavoenzyme: a 34 kDa integral protein exposed on the outer surface of outer mitochondrial membrane (Mcrlp34), and a 32 kDa soluble protein of the intermembrane space (Mcrlp32). The smaller intermembrane space isoform is generated from the larger form by the action of the inner membrane protease I (IMPI) on the outer surface of the inner membrane. This is the first demonstration that a single gone encodes proteins which are located in different compartments of the same organelle. with special interest in any with a mitochondrial localization. Degenerate primers were designed on the basis of high homology between members of the ABC family, and PCR was performed on yeast genomic DNA. Ten DNA fragments bearing significant homology to the ABC family were amplified, nine of which were previously unidentified. Disruptions of five of the corresponding genes were performed. Disruption of one gene led to markedly impaired growth on rich medium and cessation of growth on minimal medium. The gene was cloned and found to encode a protein of 694 amino acids, a "half-transporter" which likely forms a dimer. IBI~tlnofluorescence on yeast expressing the gene cmyc-tagged at the C-terminus reveals co-localization of the protein with porin and with mitochondrial DNA, visualized by DAPI staining. Nycodenz-purified mitochondria are enriched for the protein by immunoblotting. Additionally,the c-myc epitope is protease-protected in mitoplasts, indicating an inner membrane sub-localization and a probable orientation with the C-terminus in the matrix. The gene, termed ATMI for ABC Transporter of Mitochondria, thus encodes the first member of the large ABC transporter superfamily to be localized to this organelle. Present work is aimed at revealing the substrate of this transporter. cycle in cardiac myocyte. The molecular study of this protein has been difficult even after its over-expression was made possible by the cloning of the corresponding cDNA. The chief problem is the lack of convenient domains which could be used in the purification of the molecule.

We have previously reported the efficient expression of the cardiac Na+/Ca 2+ exchanger in mammalian cell lines using the T7 RNA polymerase-hybrid vaccinia virus system. We have now constructed the recombinant virus for the exchanger with a 6xHIS-tag at the C-terminal. This tag has enabled the purification of the protein through a Ni-NTA agarose column.

The expressed tagged protein induced Na-dependent Ca uptake which was not different from that in intact cells, i.e., cells in which the exchanger did not have the tag, The most important aspects of the purification produced and those of the reconstitution of the expressed protein will be described.

$15-08 

The neural cell adhesion molecule L1 is involved in the formation and specification of cell contacts in the central and peripheral nervous system during development, and thus may also play a role in synaptic plasticity of the adult central nervous system, Using extracellular recordings of evoked field potentials in the rat hippocampal slice, we found that locally applied polyclonal anti-L1 and fusion proteins containing the Ig-domains I-VI of L1 specifically block the stabilization of LTP in the CA1 region. Baseline synaptic transmission was not affected by the presence of the anti-L1 antibodies. The anti-Ll-induced effect was dependent on the antibody concentration and anti-L1 antibodies had no effect on previously established LTP. LTP was also reduced by an antiserum against the recombinantly expressed Ig-domains I-VI of L1, whereas controt antibodies against liver cell membranes, which bind to neuronal membranes in the hippocampus, exhibited no effect on LTP.

The contribution of L1 to stabilization of LTP within the CA1 synapses suggests that members of the Ig-supertamily of cell adhesion molecules are involved in the structural modifications which are associated with synaptic plasticity in the mammalian central nervous system. An increasing number of surface proteins are described as being attached to the membrane by a glycosyl-phosphatidylinositol (GP1) anchor instead of the conventional transembranous hydropbobic domain. They are initially synthesized with a COOHterminal polypeptide extension which is cleaved in the E.R. and replaced by the preformed glycolipid. This processing event is directed by a signal, located in the C-terminal region of the protein, which requires a group of three small amino acids positioned 10-12 residues upstream of an hydrophobic C terminal sequence. We previously transformed the transmembfanous glycoprotein D of Herpes simplex type I into a GPI-anchored molecule by deleting its cytoplasmic domain and thus creating a molecule, gDww63, which meets the requirements for anchoring via a GPI structure. We now demonstrate, using GPI-deficient cell lines and biochemical studies, that a hybrid protein consisting of the Thy-I ectoplasmic domain and of the C-terminal region of gDww63 can be alternative[y expressed in a GPI-anchored or in a transmembranous form. Service de P6diatrie, CHUV, CH-1Oll Lausanne Proteolipid protein (PLP) is a major myelin protein in the central nervous system (CNS). A number of X-linked dysmyelinating disorders have been identified as mutations in the PIp gene. We have identified a novel Plp mutation in paralytic tremor (It) rabbit, characterized by body tremor, spastic limb paresis and hypomyelination in the CNS. Reduced levels of PLP protein and its mRNAs were observed in the CNS as well as in the PNS. Plp sequence analysis revealed a single nucleotide change in exon 2 which results in the substitution of a histidine by a glutamine at position 36. Histidine 36 is localized on the boundary between the first transmembrane region and the intracellular part of the protein. Therefore, its position can be crucial for the efficient PLP interaction with other proteins and lipids, and correct incorporation of the PLP molecule into the membrane. Histidine 36 is also part of a short fragment of ten amino acids which is conserved in all species studied. The same amino acid is altered in another hypomyelinated mutant, shaking pup, stressing its functional importance. Furthermore, the Pt mutation affects the PLP "premyelin" functions, including oligodendrocyte maturation. Equilibrium constants for octamer dissociation as well as dissociation rates in vitro increased in correlation to the number of charged residues eliminated, e.g. mutant 4-7, in which all four charged residues in the N-terminal heptapeptide are substituted with uncharged amino acids, showed a 100-fold higher equilibrium constant for octamer dissociation and a 145-fold increased dissociation rate compared to wild type. This strongly suggests that the charged amino acids in the N-termlnal heptapeptide of Mib-CK, and therefore an ionic interaction, play an important role in forming the oetameric molecule. Gangliosides and neutral glycosphingolipids (GSLs) cell surface molecules are involved in a variety of biological processes and are assumed to play an important role in the mechanisms of HIV entry into lymphoid and epithelial cells. The total lipid-bound sialic acid (LBSA) content and the composition of gangliosides and GSLs were investigated on PBMN cells from HIV+ve and normal donors. LBSA level was quantified by a fluorimetric assay, while gangliosides and GSLS were assayed with high performance thin layer chromatography (HF'TLC) after multistep lipid extraction of 100 0013 x g membrane fractions. In normal PBMN cells the LBSA content accounted for 2.6/~g/108 cells or 20 nmol/mg protein whilst in PBMN cells from HIV+ donors the LBSA level decreased to 1.6 ,ug/108 cells or 13 nmol/mg protein. The qaantitation of neutral GSLs and gangliosides was carried out by scanning spots revealed on HPTLC plates and the integrated areas computed with a dedicated software. Our findings showed that the most relevant ganglioside present in normal PBMNC was GM3 (65-80 % of total gangliosides) followed by small amounts of OD3 (5 -15 %) and other acidic glycosphingolipids. Neutral GSLs included GL1, GL2, GL3 and GL4-hexosylceramide. HIV+ PBMN cells were characterized by a decrease of neutral glycosphingolipids as well as GM3 content which represented only 55 % of the ganglioside fraction, indicating an alteration in the glycolipid composition in the plasma membrane of infected cells. The neuropeptide Y (NPY) is one of the most abundant peptides of the mammalian brain. Upon stimulation of cell surface receptors, NPY generates diverse physiologic responses. NPY acts on the cardiovascular system. It is a vasoconstrictor by itself and in addition, it potentiates the action of vasoactive substances such as angiotensin II or norepinephrine. NPY also inhibits glucose induced insulin secretion and is a potent inducer of appetite. To facilitate the development of NPY antagonists we are investigating the interaction of NPY with its cell surface receptor.We first constructed a series of mutants in which negatively charged residues present in putative extracellular domains of the Y1 receptor were systematically replaced by alanines. The mutant cDNAs were transiently expressed in HeLa ceils using a vaccinia virus derived expression system and the abi}ity of the encoded proteins to bind NPY was evaluated. The capacity of the mutant proteins to be recruited to the cell surface was assessed by confocal microscopy. We identified initially 4 spacially clustered extracellular acidic residues of the human Y1 NPY receptor essential for NPY binding. Subsequently, we mutated 36 additional residues. Based on these data a detailed computer model of the interactions between NPY and its receptor was built.

FUNCTIONAL ACTIVITY TO ENERGY METABOLISM. L. Pellerin and P.J. Magistretti. Institut de Physiologie, Universit6 de Lausanne, Switzerland. Astrocytes play a central role in brain energy metabolism. For example glucose, the exclusive brain energy substrate, is avidly taken up by astrocytes (PNAS 90:4042, 1993) . Application of glutamate (Glu), the main excitatory neurotransmitter, stimulates in a concentration-dependent manner 3H-2-deoxyglucose (2-19(3) uptake by astrocytes in primary culture with an EC50 of ~ 100 ~M. The effect is not receptor-mediated since it can not be reproduced by glutamatergic agonists such as (NMDA, kainate, quisqualate, t-ACPD) nor prevented by antagonists (APV, CNQX, AP3). Instead, it involves glutamate uptake since the effect is blocked by DL-threo-J3hydroxyaspartate, a potent inhibitor of the Glu transporter. Replacement of Na + in the medium by choline also prevents the effect, suggesting a Na+ driven process. Finally ouabain, a selective inhibitor of the Na+/K + ATPase, completely prevented the effect of Glu. These observations suggest that when glutamatergic synapses are active, glutamate, taken up by astrocytes, stimulates 2-DG uptake by astrocytes whose end-feet are known to surround capillaries, i.e. the source of glucose. They also provide a simple explanation for the mechanism linking functional activity to energy metabolism.

Cloning and molecular characterization of the mouse astrocyte glycogen synthase. G. Pellegri, C. Rossier, S. van Berchem, P.J. Magistre~i and J.-L. Martin. Institut de Physiologie, Universit6 de Lausanne, Switzerland. In the brain glycogen is predominantly localized in astrocytes, although its presence has been detected in certain large neurons, in ependymal and choroid plexus cells. Glycogen is synthesized by the enzyme glycogen synthase (GlyS) from UDP-glucose. Out of the different glycogen synthase isozymes, only the muscle and the liver forms of GlyS have been cloned. We have isolated and characterized from a mouse astrocyte ~.ZapII cDNA library, a cDNA clone encoding the mouse cerebral cortical astrocyte GlyS. The 3.5 kb clone isolated contains an open reading frame of 2214 bp encoding a protein of 738 amino acids. The coding region of the mouse astrocyte GlyS cDNA shares 87% and 86% identity with the nudeotide sequences of the human muscle and the rabbit muscle GIyS cDNA respectively. The cellular distribution of the GlyS mRNA studied by Northern blot shows the presence of a single transcript of 4 kb in cultures of astrocytes and, to a lesser degree, of neurons. The distribution of glycogen phosphorylase, which was also examined by Northern blot shows the presence of a 4.2 kb transcript both in cultured astrocytes and neurons with however higher levels expressed in astrocytes. The functional molecular size of the vacuolar H+-pyrophosphatase (H +-PPase EC 3.6.1.1) from maize (Zea mays L.) roots was estimated by radiation inactivation, both for substrate hydrolysis and for proton transport. Light microsomal vesicles enriched in tonoplast membranes were prepared from young (2 days) maize roots using sucrose step gradients. These vesicles were still competent for PPi-dependent proton transport after freeze-drying and rehydration of the membranes. Frozen native or freeze-dried samples were irradiated with 6~ for various periods of time. After thawing the samples, the activities of glucose-6phosphate dehydrogenase (added as an internal standard) and H+-PPase (hydrolysis of pyrophosphate (PPi) and PPi-dependent proton transport) were tested. By applying target theory, the functional size for hydrolysis was found to be around Mr 155 000 when the native or freeze-dried samples were irradiated. No PPi-dependent proton transport activity was measurable after irradiation of the native samples. On the contrary, the freeze-dried membranes were still pumping protons after irradiation and a target size of around Mr 250 000 was measured for the transport activity. These experiments suggest that the native H+-PPase from maize roots may exist as a dimer (at least) of the Mr 80 800 catalytic subunit.

$15-20 The Na,K-pump moves 3 Na + out and 2 K + ions into the cells.

The domains of the protein involved in ion translocation have not been determined. We have studied the role of the N-terminal region in the Na + translocation by measuring the kinetics of charge movements under Na/Na exchange conditions in wild type (WT) and two N-terminally truncated forms of the Bufo ~ subunit with deletions of the 31 (T31) and 40 (T40) first amino acids expressed in Xenopus oocytes. We have developped a new technique to perform fast (< I ms) voltage clamp on whole oocyte. The membrane on one side of the oocyte is permeabilized by digitonin allowing to measure current flowing across the other half. We observed presteady state ouabain-sensitive currents similar to that reported by other techniques (J.Gen. Physiol. 101:117,1993) . The estimated forward charge translocation rate was lower in both N-truncated mutants (WT 430 • 14, T32 215 Z 8, T41 124 Z 3 s-l). The midpoint potential of the charge translocation (best fit to Boltzman equation) was -79 , -45, and -28 mV in the WT, T31 and T40 groups, respectively. The highly charged Nterminus of the Q subunit has a significant role in the forward translocation of Na + ions. 

Deletion mutants of human small intestinal lactase-phlorizin hydrolase (LPH) were constructed to determine the role of protein subdomains in the intracellular transport of LPH. The mutant LPH1646MACT (236 aa deletions), which contains the membrane anchor (MA) and the cytoplasmic tail (CT), was transported to the cell surface in a fashion similar to wild type LPH. By contrast, LPH1646 lacking the transmembrane domain persisted as a mannose-rich polypeptidc. This suggests that the membrane anchor and/or cytoplasmic tail are implicated in the ER-egress of LPH. Another mutant, LPH1559MACT (323 aa deletions), was not efficiently transported to the Golgi apparatus. Epitope mapping with monoclonai antibodies as well as protease-sens{tivity assays provided an evidence that the tertiary structure of LPH1646MACT is very similar to that LPH1559MACT. In view of the variations in the proportions of the complex glycosylated molecules of the two mutants we conclude that the domain between LPH1646MACT and LPHI559MACT may play an essential role along the secretory pathway of LPH.

CONTAINING TYROSINE. Hussein Y. Naim, and Michael G. Roth; Dept, of Biocherni~try, UT-Southwestem Medical center, at Dallas, Dallas,Texas-USA We have investigated an artificial internalization signal created by site-directed mutagenesis of the short cytoplasmic sequence of the influenza virus hemagglutinin (HA). Mutation of cysteine 543 to tyrosine converted HA from a protein essentially excluded from coated pits to one that internalized at rate similar to some cellular endocytic receptors. To determine whether or not the HA-Y543 signal is a degenerate form of the internalization signal found in proteins such as the transferrin receptor and the mannose-6-phosphate/IGFII receptor, we have mutated amino acid positions in the HA-Y543 shown to be important for internalization of the two receptors. We have changed amino acids on either side of Y543 to those that either fit, or break the pattern of aromatic and hydrophobic residues proposed as consensus sequence. Our results indicate that the HA-Y543 mutant contains a sub-optimum sequence for a tyrosine-based internalization signal similar to those found in the receptors for transferrin, LDL, and mannose-6-phosphate/IGFIl. However, amino acids with side chains having different chemical properties functioned well in positions that are important for the internalization signal, indicating that the consensus sequence for this signal is more variable than previously proposed. N-linked glycesylation is an essential protein modification occuring in all eukaryotic cells. Core-oligosaccharides are transferred by the enzyme N-oligosaccharyl-transferase frc~ the donor Glc Man GIcNAc -Dolichol to selected asparagine 9 2 .

resldues of t~e nascent polypeptlde chazn. The donor is build up in a stepwise fashion at the ER-membraneby the products of the ALG-genes (asparagine-linked-~lycosylafion) Mutants defective in this pathway are available, but often lack a phenotype suitable for isolation of the corresponding genes. We found that double mutants of alg5 or alg8 with a wbpl mutation (defective oligosaecharyl-transferase) are severely growth retarded at 30oC. The double mutant phenotype allowed the isolation of both ALG5 and ALG8 genes which have been refractory to cloning so far. Both genes encode potential trarsmenbrane proteins and are able to complement the biochemical defects of the corresponding alg5 or alg8 mutants. The Na,K-ATPase produces the driving force for the regulated transport of Na across epithelial ceils. It is composed of an c~-subunit which carries the catalytic and ouabain binding sites and a B-subunit. To test the role of the usubunit in transport regulation, we transfected A6 cells with the B. marinus al-subunit (txl-TBM). This subunit confers a 300-fold higher resistance to ouabain (K i = 50/~M) compared to the endogenous pump. Taking advantage of this difference, stable transfectants were selected with ouabain. Expression of the cd-TBM subunit was visualized by Western blotting. Approximately half of the positive cell lines showed a high electrical resistance similar to untransfected cells, when cultured on filters. Na-transport activity, measured as Isc, was lower even in the absence of ouabain, but the relative increase in response to aldosterone was maintained. Functional pumps containing the exogenous or endogenous al-subunit were quantified by equilibrium ouabain binding. Interestingly, cells cultured on plastic dishes had a large fraction of low affinity binding sites (txI-TBM), while most sites were of the high affinity type when the cells were grown on filters. We conclude that the ouabain-resistant Na,K-ATPase cd-subunit of B. marinus can be expressed in A6 cells and used as a selective marker. However, the results suggest that the expression of functional ouabain-resistant pumps containing the al-TBM subunit could be influenced by the degree of cellular differentiation. In infected cells fully functional Na/Pi-cotransport was found in a time dependent manner; up to a 50-fold stimulation of Na/P;-cotransport compared to uninfected cells was observed aft6r 3-4 days. Transport of phosphate by infected cells was highly dependent on the presence of sodium, exhibited a KmtD~ of around 0.16 mM and was dependent on extraceUular I~H'. In agreement with expressed transport of phosphate, infected cells produced an immunoreactive polypeptide of 66 kDa that represents a non-glycosylated form of the 80 to 90 kDa mature Na/P;-cotransporter. We conclud$ that the protein NaPi-2/rat when expressed in Sf9 cells exhibits Na/P-eotransport with similar kinetic characteristics as observdd in prox ma tubular apical Na/Pcotransport. Therefore the NaPi-2 protein as expressed h Sf9 cells can be used for further structural and functional studies. Talin interacts in vitro with actin, vinculin, integrins and bilayers of acidic phospholipids, and nucleates actin filament growth at the bilayer surface. It is therefore believed to be a key protein mediating aetin-membrane linkage. We have analyzed functional properties of proteolytic fragments of purified platelet talin. Using limited proteolysis two fragments of 47 and 190 kD were generated. Preliminary results, obtained using viscometry and F-aetin-nucleating assays, suggest that the 190 kD fragment contains the actin binding site. We also assessed the lipid-binding capacity of the fragments. When a mixture of the two fragments was centrifuged in the presence of large multilamellar liposomes containing phosphatidylserine, only the 47 kD fragment, but not the 190 kD domain, co-sedimented with the liposomes, suggesting the presence of a specific lipid-binding site in the 47 kD domain. Interestingly, this fragment contains the N-terminus of talin which shows homologies to the membrane-binding regions of protein 4.1.

We are now analyzing membrane-binding properties of talin domains in more detail using hydrophobic photolabelling. Recent studies suggest a role of the neural cell adhesion molecules L1 and NCAM in mechanisms of memory formation (Doyle E, et al, J. Neurochem. 59:1570 -1573 , 1992 Scholey A.B. et al, Neuroscience 55:499-509, 1993; LiJthi A. et al. unpublished data and see USGEB 94) . in the present study we analyzed the effect of chronic intraventricular infusion of polyclonal antibodies against L1 (anti-U) or NCAM (anti-NCAM) on the performance of male Wistar rats during the acquisition and retention of a spatial learning task. Briefly, rats had to remember the position of a submerged escape platform in a water tank (Morris water-maze). When the escape platform was removed after acquisition training (= retention test), the performance of animals chronically injected with anti-L1 showed an impaired search pattern when compared with that of salineqnjected animals (p>0.05, Mann-Whitney). Whereas control animals spent up to 45% of their time searching for the platform in the correct (= training) quadrant, the time anti-L1 animals swam in the correct quadrant was closer to chance level (31%). Although monitoring penetration of antkNCAM into the hippocampus was almost as efficient as that of anti-L1, the performance of the anti-NCAM-group was highly variable and did not differ from the performance of controls. These results agree with recent findings on the involvement of neural cell adhesion molecules in other forms of learning. We have isolated cDNAs from mouse and from Xenopus encoding the ~-subunit of the gastric H,K-ATPase, which is expressed in the parietal cells of the stomach mucosa.

The gastric H,K-ATPase transports H + into the stomach lumen with the counter transport of K +, all at the expense of ATP hydrolysis.

When Xenopus oocytes are injected with cRNA encoding ~H,K from either species, as well as a gastric ~H,K subunit, the two subunit proteins are synthesized, assemble, and form functional pumps located in the plasma membrane as demonstrated by 86Rb+ uptake.

The sensitivity to the gastric H,K-ATPase inhibitor SCH28080 was determined, with the Ki for beth species found to be in the low ~M range. Spodoptera Frugiperda ceils are often used to overexpress eucaryotle proteins using Baculovirus vectors. The atomic force microscope (AFM) is a device which allows the observer to obtain high resolution images of biological samples immersed in a fluid medium; this instrument has therefore the potential to visualize overexpressed proteins on the surface of living cells. However, the interpretation of the images is sometimes difficult and requires a good knowledge of the specimen topography, i.e. the "normal" plasma membrane of the cell used for the overexpressisn. On our poster we present AFM images of the plasma membrane of both fixed and living Spodoptera Frugiperda ceils.

In additon, we discuss the possibility of achieving high resolution in the imaging of dynamic changes which affect /iving systems. The deduced protein sequence consists of 295 amino acids with about 55% amino acid sequence identiCy when compared to mammalian ~H,K-subunits.

Data from other species and from the structurally similar Na,K-ATPase has shown that the ~-subunit is necessary for the formation of a functional pump, which consists of an ~/~ heterodimer.

Following co-injection of Xenopus oocytes with cRNA for both the ~-and ~subunits, we have observed by Rb + uptake H,K-ATPase activity in the plasma membrane.

We are presently studying the role of the ~-subunit in the maturation and function of the H,K-ATPase. The atomic force microscope (AFM) is a new type of microscope which allows visualisation of inorganic and organic samples with a very high resolution. A sharp tip fixed at the end of a cantilever is used as a topographic sensor. By scanning the sample with this device, a computer records the minute deformations of the cantilever and computes the topography of the sample. The instrument allows also the measurement of the interaction between the tip and the sample as a function of the distance. The knowledge of these interactions allows us to compute several mechanical properties like indentation, elasticity and stiffness. On this poster we report the results of these measurements on cells (sfg, C131), bacteria (E. coli, B. subtilis) and proteins (actin). In addition we discuss the use of the AFM as a sensor for probing the mechanical properties of biological systems at a nanometric scale. Neonatal thymectomy of BALB/c mice induces AIG.

We have cloned the ~-and ~subunits of the BALB/c H,K-ATPase and expressed the proton pump in cRNA injected oocytes.

Both and ~ proteins are made, form functional pumps, and are immunoprecipated using auto-immune sera.

The ~-subunit can also be in~unoprecipated independent of the ~-subunit when oocytes are injected with ~ cRNA alone.

BALB/c mice will be immunized with oocyte membranes containing either the ~/~ or ~ antigens in order to study the role of each H,K-ATPase subunit in triggering AIG.

The prion protein is expressed in both astrocytes and oligodendrocytes of the neonatal brain and is down-regulated in adulthood.

M. Moser, Colello, R, , Pott, U, and B. Oesch Institut for Hirnforschung der Universit~t ZOrich, August Forel Str. 1, 8029 ZOrich The infectious agent causing transmissible spongiform encephalopathies consists largely or entirely of PrP sc, a modified isoform o| normal PrP (prPc).

PrpC is most abundant in the brain. The development of the disease strictly depends on the presence of PrPC. and incubation times are shortened by overexpression of PrpC. The normal function of prpe is not known and its cellular Iocalisation in the CNS is poorly characterised, except for its evident presence in neurons. Here we describe the expression of PrP c mRNA in both oligodendrocytes and astrocytes during neonatal development of hamster and rat. In adult animals, expression is down-regulated in gila but not in neurons.

Our findings provide a link to the previously described accumulation of prpSC in astrocytes and the apparently faster course of prion disease in neonatal hamster. Our results argue for a major involvement of non-neuronal cell types in prion diseases. such structures we raised Mab against membrane protein fractions. A ~lab against a vitronectin receptor containing fraction was found, which if applied in vitro to BI6FI mouse melanoma cells, influenced growth, adhesion and morphology. ~Tnen BI6FI cells prior to tail vein injection were treated, 2 different outcomes were observed. Short treatment (lh) resulted in 85% inhibition of lung colonization whereas long-ter~ treatment ( 2weeks) resulted in 10-15 times more lung colonies. A candidate molecule likely involved in these observed effects could be identified as a 150 kDa cell surface protein. Sequence information from peptides revealed in several cases 100 % homology to mouse centrosomin A, i.e.a 32 kDa cytosolic protein involved in the organization of the spindle apparatus. These data suggest that we may have found a protein which may participate in a link between the extra-cellular space and the microtubular system. We are currently trying to clone this 150 kDa protein from a BI6FI c-DNA library.

Wahlberg, J.M., and Spiess, M. Dept.Biochemistry,Biozentrum, University of Basel. Signals for correct basolateral sorting of subunit HI of the human asialoglycoprotein receptor are located in the cytoplasmic tail, involving a tyrosine at position 5. Removal of this residue results in nonpolarized transport of the protein to the surface in MDCKII-celIs. A deletion mutant of HI, lacking 30 of the normal 40 residues of the cytoplasmic tail, including the tyrosine, has been found not to be transported to the cell surface, but to be accumulated intracellularly.

The mutant protein is complex glycosylated and sialylated, indicative of passage into or through the trans-Golgi.

Immunofluorescence analyses show defined juxtanuclear, tubular structures, which do not colocalize with markers for ER, early endosomes or lysosomes, but which show similar staining pattern as markers for late endosomes and TGN. To define the determinants responsible for the intracellular accumulation a series of deletions and fusions have been constructed. The results suggest that the length of the cytoplasmic domain is one important determinant for missorting of HI. 

Salerr~ N., Medilansld, J., Pellegrinelli, N. and Eder-Colli, L. Departement of Pharmacology, CMU, 1211 Geneva 4 In chollnergic neurons the enzyme choline acetyltransferase (CHAT) exists as cytosolic(80-90%) and membrane-bound activity (10-20%}. Are these two activities encoded by a single mRNA? Is membranebound enzyme preferentially associated with a given cellular membrane? In order to investigate these questions we isolated a fulllength cDNA (4.2 kb) encoding Drosophila ChAT and used it to transfect Xenopus oocytes, rat fibroblasts and human neuroblastoma cell lines. Triton X-114 fractionation of transfected cells showed that hydrophilic and amphiphilic ChAT activity were produced. The sp act. of hydrophflic enzyme reached 2, 0.8 and 0.25 ~mol ACh/h/rng protein respectively in oocytes, fibroblasts and neuroblastoma whereas the sp. act. of amphiphilic enzyme was 1.8, 0.4 and 0.125 HInol ACh/h/mg protein, respectively. Amphiphilic ChAT was less sensitive to inhibition by the product acetylcholine and to heat denaturation than was hydrophilic enzyme. Similar results were observed for the native Drosophila CHAT. Amphiphillc enzyme sedimentation was affected by the type of detergent present in linear density gradients of sucrose. Transfected oocytes were subjected to subfractionation. Besides a large amount of soluble ChAT activity, a significant proportion of enzyme was found associated with a fraction enriched in endoplasmic reticulum (ER). Preliminary results using transfected mammalian cells indicate that apart cytosolic CHAT, plasma mernbranes+Golgi enriched fractions as well as ER-enriched fraction contain ChAT activity. We have characterized a Drosophila melanogaster gene encoding a very hydrophobic protein. The amino acid sequence shows considerable homology (33-42% identity) to neurotransmitter transporters. The gene is, however, not expressed in neural tissue, but in the male germline. Transcription and translation occurs in early spermatocytes. Antibody staining data suggest the following pathway of the protein during spermatogenesis: The protein is synthesized in early spermatocytes and is transported into the nucleus during prophase of RT i. At meiosis it associates with the mitochondria and it stays associated while they fuse to the Nebenkern. During individualization it is stripped off in the cystic bulge. Inspection of the putative protein sequence shows that it possesses alle the necessary targeting signals to allow its transfer from the cytoplasm through the nucleus into the mitochondria: three nuclear targeting signals, a mitochondrial import signal and, moreover, two PEST-sequences for rapid protein degradation are present. We shall present a, currently tested, working hypothesis. In Hg-treated toad skins, water permeability is high or low depending on whether the major anion of the inner bathing medium is SO~ or CI. We report here the results of experiments designed to determine if, in skins bathed with SO4-Ringer, net water flow (J~,) could be diminished by agents added to the outer medium. Toad skins were mounted inbetween glass chambers and exposed to a standard osmotic gradient of 200 mOsm. Jw was continuously monitored by automatic, volumetric techniques. Exposure to lmM HgClz or CHzCIHg (external side) led to the typical anion-induced J,, changes. Re-exposure to Hg caused a mild (10%), although significant (N=8, P<0.001), fall in Jw. Since CuSO4 is a SH-reagent like Hg, we looked at the effects of CuSO4 (lmM). There was a 68% reduction of J, (N=7, P<0.001), an effect totally reversible by simple washing of the skins. Likewise, addition of dithiothreitol (DTT, 5mM) in the presence of Cu, caused a 90% reversal of the Cu effect ; there was no change in Jw if DTT was given before Cu, or if both agents were added simultaneously. Finally, in Hg-treated, glutaraldehyde-fixed skins, lmM and 5ram CuSO4 caused 51% and 92% Jw decrements, respectively, that were totally reversible. In conclusion, Cu causes a DTT-sensitive inhibition of Jr, in Hgtreated skins. Further work is needed to establish if the Cu effect is exerted on SH-groups of apical aquapories of toad skin.

$15-44

The amino acid sequences of the PC-645 e-subunits and other cryptomonad biliprotein ~-subunits pose the most intriguing question about the phylogenetic origin of these unique chromopeptides. Cryptomonad PC-645 ~-subunits seem not to be closely related to the known phycobiliproteins, linker polypeptides (LPPs), BPE or any other light-harvesting chl polypeptides from the thylakoid. For most of them the number of identical amino acid residues was 15%-20%. Sequence alignment of C-PE associated LPPs with T-subunits from cyanobacteria and red algae and the cryptophytan ~-subunits however show that the identity between C-PE associated LPPs and cyanobacterial x-subunits is still significant, whereas it reaches the limit of significance between cyanobacteril and eucaryotic x-subunits. Nevertheless it has to be assumed that x-subunit derive from cyanobacterial C-PE associated LPPs as well as cryptomonad ~-subunits may derive from T-subunits. Whereas the the phycobiliprotein ~-and 8subunits remained structurally conservative during evolution (e.g.

-70% identity between cyanobacterial and red algal PE-subunits) the LPPs and T-subunits show drastical changes in their chromophore binding domains.

$15-45

Functional and evolutionary relationships of the components of the primary events in photosynthesis have been traced using amino acid sequence comparison. Basically, the question was posed regarding the interrelationship of the apoproteins of different photosynthetic organisms which gather light and/or separate charges across membranes. Do their protein-chemical structure and organisation decipher us certain clues and parts of the path by which they have been selected and derived from a possible primordial reaction centre polypeptide? Within that context the light-harvesting-and energy transfer systems of purple bacteria offered as ideal model components. A data set of more than 70 apoprotein sequences has pointed to some keystructural elements which have been partially verified as such by limited proteolysis and genetic engineering. A careful inspection of the determined bacterial antenna-structures revealed some remarkable similarities to eukaryotic antenna and reaction centre apoproteins indicating possible basic structural motifs for complexing pigment molecules, like chlorophylls or bacteriochlorophylls in very distant representatives of phototrophs. A number of glycoproteins of the nervous and/or immune system that are involved in cellular recognition and/or adhesion share the common LZ/HNK-1 carbohydrate structure. In a large proportion of patients with peripheral demyelinating neuropathy associated with paraproteinemic gammopathy (PPN), monoclonal IgM antibodies (M-IgM) react with the LZ/HNK-1 determinant of MAG and P0. We plan to test the hypothesis that anti-MAG M-IgM autoantibodies may alter the expression of HNK-1 bearing myelin proteins such as MAG, P0, MOG, NCAM or PMP22. By using monoclonal antibodies to myelin proteins on tissue sections, preliminary results indicate that MBP protein content seems to be strongly downregulated, whereas GalC and $100 protein do not appear to be affected by the loss of myelin in PPN nerves. GFAP protein expression appear to be upregulated in demyelinating biopsy specimens from PPN patients with anti-MAG M-IgM gammopathy. In conclusion, the expression of some L2/HNK-1 negative proteins might be disregulated by the presence of circulating anti-MAG M-IgM in the serum of PPN patients. We have isolated a human liver Na+-taurocholate cotransporting polypeptide (NTCP) by screening a human liver eDNA library with a rat eDNA probe derived from Ntcp (PNAS 88, 10629, 1991) . The cDNA codes for a protein of 349 amino acids which shows 77% amino acid homology with the rat Ntcp. Expression of NTCP in X. laevis oocytes resulted in Na+-dependent bile acid uptake which exhibited saturability with an apparent Km for taurocholate of 6 ~M. NTCP mediated taurocholate uptake into oocytes was inhibited by all major bile acid derivatives, bumetanide and bromosulphophthalein. Southern blot analysis of genomic DNA from a panel of human/hamster somatic cell hybrids mapped the human NTCP-gene to chromosome 14. In conclusion, these studies supplement the previously suggested selective mammalian distribution of the Ntcp gene family and provide the possibility for direct molecular characterization of human bile acid transporter genes. Diverse cell surface molecules of the nervous system play an important role in specifying ceil interactions during development. With the aid of a polyclonal antibody against L1, a 1.Skb eDNA clone closely related to L1 (CH-L1) was isolated from a ~. gtl 1 library derived from poly(A)* l~qA of day 8 mouse brain. Using this eDNA as probe the remaining cDi'IA 5' sequences were cloned out of a plasmid library constructed from postnatal day 6-14 mouse brain poly (A)+ P, NA. Two independent full length clones of 4.2 and 4.4 kb encompassing the entire coding region of CHL] were isolated. These represent putative alternatively spliced mP, NAs. The deduced primary structure of CHL1 reveals that, like the cell adhesion molecules (CAMs) mouse L1, chicken Ng-CAM, and Nr-CAM (BRAVO), CHL1 is composed of six Ig-like domains, a transmembrane domain, and a short cytoplasmatic region. In contrast to the other L1 family members, only four and a half instead of five fibronectin type Hi-like repeals are found in CHL1. The fibronectin type [H-like repeats were expressed in E. col[ and the protein fragment was used for immunization. As observed for LI, Ng-CAM, and Nr-CAM, CHL1 is susceptible to proteolytic cleavage. Bands of 185kD, 165kD, 125kD, and 50kD could be detected byWestern blot analysis. By in situ hybridization, a predominant expression by neurons was found in the adult mouse brain. Western blot analysis showed expression of CHL1 protein in brain and heart, but not lung, kidney and intestine. In liver, a 50 kD immunoreactive band was found. The relationship of CHL1 to molecules known to be involved in cell adhesion and neurite outgrowth, combined with its pattern of expression, suggests a role for this molecule in cell interactions dndng neural development. Work from this laboratory revealed that in Hg-treated, glutaraldehydefixed toad bladders, the water permeability coefficient (Pf) is high in SOd-Ringer and low in CI-Ringer ; in all these studies the osmolality of the serosal medium was 220mOsm. As an attempt to determine the mechanism(s) underlying such Pf changes, experiments were carried out with bladders whose serosal surface was exposed to [so-, hypo-or hyperosmotic media ; the outer surface was always exposed to a hyposmotic solution (22 costa). Net water flows were measured continuously with automatic, volumetric techniques. After exposure to 1 mM CHzCIHg (10', outer medium) followed by fixation with 0.25% glutaraldehyde (5', serosal medium), the bladders were thoroughly washed. In SOd-Ringer, Pf (/~m/sec) was as follows (N=6) We wanted to determine the molecular weight (MW) of Ntcp and its topology in the plasma membrane (PM) of rat liver. A rabbit antiserum generated against the C-terminus of Ntcp was used to determine its MW on Western blots and its subcellular distribution in liver and in hepatocytes by immunofluorescene (IF). Site directed mutagenesis and deletion experiments were used to determine the N-linked glycosylation sites. Ntcp encodes a 50 kD protein in rat liver PM vesicles. The basolateral localization of Ntcp in liver was confirmed by IF. Ntop could only be immunostained in permeabilized hepatocytes suggesting an intracellular localization of the Cterminus, cNDA constructs showed the Ntcp is glycosylated at positions 5 and ii. The data show a selective basolateral expression of Ntcp in rat liver and they suggest that the two ends of Ntcp are located on opposite sides of the PM. Zymogen granule membranes (ZGM) are known to contain proteins with nucleoside phosphatase activity, but their exact nature and function are unknown. The activities require Ca ++ and Mg ++ and are sensitive to ATPase inhibitors but not to inhibitors of Na/K-ATPase activity. The aim of this study was to isolate individual proteins of pig pancreatic ZGM, to attribute enzymatic activity to individual proteins and to characterize them with various substrates (GTP, ATP, AMP etc.), inhibitors (AMP-CP, AMP-CPP, GTPgS etc.) and lectins (MAA, DSA, GNA etc.). We have subjected highly purified ZGM solubilized in CHAPS to IEF on Immobiline Dry Strips | (Pharmacia) with s pH range 3--10.5 (Ist dim.). These strips were then electrophoresed in a polyacrylamide gradient gel containing 0.2% CHAPS and Tris/HCI at pH 9.5 (2nd dim.). Nucleoside phosphatase activity was detected histochemically in the 2D gel by incubation with substrate in the presence of lead nitrate. Focussed strips were also subjected to SDS-PAGE for comparison.

Several proteins showed strong activity to GTP, ATP and ITP. The role of these phosphatases are discussed in the context of the role of GTP-binding proteins and the de-/phosphorylating events on the ZGM in intracellular targeting and exocytosis. The disease-specific isoform of the priori protein (PrP Sc ) is an essential part of the infectious particle causing scrapie or BSE. PrP Sc differs from PrP of normal animals (PrP C ) by its relative protease resistance. The physical nature of this difference is still unknown. Here we analyzed the protease-resistance of PrP Sc quantitatively. PrP Sc was rendered completely protease-sensitive at alkaline pH or in guanidinium thiocyanate (>1.5 M). Denaturation in 2M guanidinium thiocxanate (GdnSCN) completely abolished protease resistance of PrP bc within 15 min while denaturation in 6 M urea showed a much slower time course. Denaturation cuwes were used to calculate the Gibbs free energy (AS D ) in dependence of GdnSCN or urea at different concentrations. The linear relationship between AGD and the denaturant concentration is suggestive of a two state model involving the conformational change of a single protein domain. This is also reflected in the small number of side chains (11,6) additionally exposed to the solvent upon conversion of PrP Sc to its proteasesensitive isoform. Our results suggest that minor rearrangements of the structure of PrP Sc are sufficient to abolish its protease resistance. of Physiology, University of Zurich, Winterthursrstr. 190, CH-8057 Zt~rich In contrast to mammalian kidneys where inorganic phosphate (Pi) is reabsorbed unidirectionally, flounder renal epithelial cells both re.absorb and secrete Pi. In order to investigate on a molecular level the cellular Pi handling we have cloned a re,absorptive NalP i transport system from flounder renal tubules. Based on homology with the cloned Na/P i cotransporter from rat we have isolated a eDNA clone (flounder NaPi-II, 2424 bp) encoding for a protein of 637 amino acids. In the eight transmembrane spanning domains the protein is 78% identical with the corresponding system from rat kidney, in the hydrophilic regions only 30%. Expression of in vitro transcribed RNA in Xenopus oocytes specifically stimulated Na-dependent Pi transport. Binding properties of Na and Pi as well as the pH-dependency were similar to the ones found in mammalian systems (rat, human, OK cells). By Northern analysis three transcripts, 1.9, 2.7, and 4.2 kb in length, could be detected. RNA from flounder renal tubules contained all of the transcripts, intestinal RNA only the two lower bands, and non-epithelial tissue from kidney expressed only the 1.9 kb transcript. Mammalian systems share only the 2.7 kb transcript but not the related species. The close functional relationship of flounder NaPi-II with the known NaJP i transport systems and the pronounced differences in primary structure provide the basis for structure function analysis of Na/P i cotransport. The amyloid 13/A4 protein precursor (APP) is a transmembrane protein expressed in different isoforms throughout the organism. Our work has consisted in a detailed study of the structure and biological function of APP. Specifically, we have shown that the secreted form (sAPP) of APP-695 is involved in the growth regulation offibroblast% and also stimulates neurite extension in B 103 neurnblastoma cell line. Functional mapping studies have shown that the domain responsible for both the growth-regulating and the neurotrophic activities of sAPP-695 is the RERMS site, Arg-328 to Ser-332. The presence ofsAPP binding sites on the surface of B 103 ceils (through the active RERMS site) indicate that the neurotrophic activity of sAPP is mediated by a receptor, and the subsequent activation of a signal transduction mechanism. In addition, we have shown that the RERMS site is also involved in the neuronal survival properties of sAPP-695. Finally, we have initiated an in vivo study of APP biological function, and shown that a peptide representing the neurotrophic domain ofsAPP-695 can strengthen memory retention in rat through stabilization of synaptic structures in the frontal cortex. This work was supported by grants from the Swiss National Science Foundation (823A-028366), and NIH (AG 05131). We previously showed that in smooth muscle cells, the (~I-AR is rapidly phosphoryMted following exposure to agonist. To understand this biochemical event, we investigated the phosphorylation of the c(1B adrenergic receptor stably expressed in Rat1 fibroblasts (3.3 pmol/mg prot.). The =(1B-AR was solubilized from 32p-labeled Rat cells and immunoprecipitated with antibodies raised against the N-terminus part of the receptor. Treatment of ceils with 100 ~.M epinephrine showed a rapid and significant increase in receptor phosphorylation above basal. Treatment of cells with a specific pKC inhibitor (RO-318220) did not affect agonist.promoted phosphorylation while it completely abolished phorbol esterinduced receptor phosphorylation. This strongly suggests that pKC is not involved in agonist-induced c{1B-AR phosphorylation. The elB-AR contains several putative phosphorylation sites in both its third intracellular loop and C-terminus tail. To investigate the role of these domains, a trunoated receptor (lacking its last 147 amino-acids) was constructed and stably expressed in Rat cells (1.3 pmol/mg prot). This receptor mutant displayed similar ligand-binding properties and abirity to activate phospholipase C as compared to the wild-type receptor. Phosphorylation experiments revealed that this mutant is not at all phosphorylated, neither by agonist nor by phorbol esters. Agonistinduced decrease in cell surface receptors, measured by binding of the antagonist 3H-prazosin at 4~ was however similar for both the wild-type and mutant receptor. This suggests that loss of phosphorylation sites does not affect receptor internalization. Human intestinal lactase-phlorizin hydrolase (LPH) is synthesized as a precursor molecule (pro-LPH), which undergoes proteolytic processing during maturation. LPH expressed in COS-1 cells was enzymatically active, The electrophoretic properties of LPH-species synthesized by transfected COS-1 cells correspond to those in organ cultured human intestinal explants, in Caco-2 ceils and in transfected MOCK ceils. They included the pro-LPH and a proteolytically processed form, which had similar electrophoretic properties to the mature enzyme. The appearance of the putative mature form was inhibited by Brefeldin A, ammonium chloride and chloroquine. In addition, only complex glycosylated pro-LPH (pro-LPHc) was inserted into the cell surface membrane. These data indicate that in COS-1 cells pro-LPH is inserted into the cell membrane, is internalyzed and enters lysosomes where proteolytic events lead to the appearance of a mature-like enzyme.

$15-59 ASSEMBLY AND TRANSPORT OF PABA PEPTIDE HYDROLASE ALPHA AND BETA SUBUNITS IN COS-1 CELLS Eldedng, J.A., Gr0nberg, J. and Sterchi, E., Institut for Biochemie und Molekularbiologie, Universit&t Bern. Paba peptide hydrolase (PPH) (EC 3.4.24.18), a metalloendopeptidase from epithelial cells of human small intestinal mucosa, consists of two subunits, ~ and 13, which form homodimers and oligomers, cDNA cloning has revealed homologies with developmentally important proteins from different species and has led to the definition of a new family of metalloendopeptidases, the astacin family (J. Biol. Chem. 266, 21381, 1991) . Here, we report on the expression of the cDNAs of PPHcz and PPHJ3 in COS-1 cells. When expressed on its own, PPH(z is synthesized in a core glycosylated form only, suggesting that it is transport-incompetent and not able to exit the ER. Immune-fluorescence studies have confirmed that PPH~ is not expressed on the surface of COS-1 cells. PPHL3, on the other hand, matures and is transported to the cell surface. When the two subunits are coexpressed, PPHo~ can also be detected on the cell surface, suggesting that a heterooligomeric assembly is necessary for the surface expression of PPHtz. Truncated forms of both PPHc~ and PPHI~ could be detected in the medium of transfected COS-1 cells. In our previous work, we demonstrated by immunocytochemical reaction that sensory neurons expressed nuclear triindothyronin receptors (NT3R) in embryonic as well as adult rats. In contrast, in sciatic nerve, Schwann cells exhibited only transient NT3R immunoreactivity from E17 to postnatal day 10. In the present work, we attempt to demonstrate by radioautography that the detected NT3R are effectively the binding sites of [125I] labeled triiodothyronin. Cryostat sections of dorsal root ganglia (DRG) or sciatic nerve were incubated with 0.1 nM of L, 3,5,3'-[t25I]triiodothyronin 2200 Ci/m mol (NEN), while the controls were incubated with a large excess (llxM) of unlabeled trfiodothyronin (T3). In aduIt rat DRG, radioautographs revealed that silver grains were exclusively restricted to neuronal cell bodies, while the Schwann cells ensheathing the axons were free of significant label. In newborn rat sciatic nerve, silver grains accumulated over Schwann cells; in contrast in adult rat sciatic nerve, Schwann ceils were free of label. In control DRG or sciatic nerve sections incubated in large exceas of unlabeled "1"3, all the radioautographs exhibited only scattered silver grains.

In conclusion, our results show that sensory neurons and Schwann cells are able to express functional NT3R which bind thyroid hormones (SNF 31-33671-92).

Characterization of the maturepreeursor glycophospholipid for glycosylphophatidylinositoi anchors of Saccharomyces cerevisiae.

Many attempts to isolate the mature yeast precursor GPI glycophospholipid ready to be attached to newly translated proteins for GPI-anehoring have failed although such precursors were previously identified in mammalian cells and protozoa. Here we show that metabolically labeled glycolipids of the expected structure can be observed if the incorporation, their accumulation and their extraction are optimized. Two very polar, (3H)-mannose-labeled glyeolipids named CPI and CP2 were identified and purified. They were structurally characterized using phnspholipase treatments, partial deacylation with methanolie ammonia, hydrofluoric acid dephosphorylation, nitrous acid deamination, acetolysis, exoglycosidase treatments and combinations thereof to produce labeled fragments which could be analyzed by paper chromatography or thin layer chromatography. CP1 and CP2 both contain the identical core oligosaccharide

Mangl,2(X->PO4-6)Manal,2Mamxl,6(Y->)Mana-GlcN-lansitol, X and Y being hydrofluoric acid-sensitive substituents (most likely ethanolamine and phospheethanolamine, respectively). The inositol is acylated although no acyllaositol is present on protein-bound yeast GPI-anchors. CP1 and CP2 can also be observed after lableling with (3H)myo-inositol, The lipid moieties of CP1 and CP2 can be completely removed by mild alcaline hydrolysis altough the protein-bound GPI-anchors made by the pmi210 cells under identical labeling conditions are completely mild base resistant. This finding reinforces the notion that the ceramide moiety typically found on the majority of yeast GPI-proteins are added only after addition of the GPI precursor glycolipid to proteins.

The distributions of the Mannose-6-Phosphat e/IGFll-Recept or and a2,6Slalyltransferase in HepG2 cells: No evidence for co-localization by confocal laser scanning microscopy.

The organization of the trans Golgi apparatus (GA) with respect to (recycling) man-6-P/IGFII receptor (MPR) and (resident) c~2,6sialyltransferase (ST) has been investigated by double immunofluorescence laser scanning confocal microscopy in HepG2 cells. They were chosen because biochemical evidence suggested sialylation of MPR upon recycling (Duncan & Kornfeld, J. Cell Biol. 1988, 106, 617-28) implicating colocalization of both proteins. In the steady state ST was confined to the GA whereas total (endogenous) MPR was localized to coarse vesicles without evidence for colocalization by confoeal microscopy. Endocytosis of surface-labeled MPR was monitored in presence and absence of Brefeldin A (BFA): Without BFA, early endosomes (2' of warming up) showed no, late endosomes (20') little if any colocalization with ST. In presence of BFA, the ST-compartment disappeared whereas late endosomes persisted with some tubular emanations. Under similar conditions, in "absence of BFA, endogenous MPR concentrated in the GA region suggesting colocalization with ST. In presence of BFA the MPR compartment formed a redculum whereas the ST compamnem disappeared. In conclusion, within the limits of the techniques used, exogenous MPR was not detectable in the ST compartment whereas endogenons MPR concentrated in the GA region but obvious co-localization was exceptional. Supported by grant 31-30757.91 of the SNSF to EGB. Cultures of dissociated striatal neurons prepared from fetal rats were grown in the presence of neurotrophin-4/5 (NT-4/5) as well as the other known neurotrophins, NGF, BDNF and NT-3. We found that acute administration of NT-4/5 to seven days old cultures stimulates the hydrolysis of phosphatidyllnositol, an event involved in neurotrophin signal transduction. Growth of stdatal cultures in presence of NT-4/5 resulted in increased cell survival as indicated by elevations in total cell number, protein content and number of viable cells. NT-4/5 exposure increased GABA uptake and staining intensity in GABA immunocytochemistry in these cultures indicating atrophic action on GABAergic neurons. To further identify responsive cell populations we analyzed for calretinin, a calcium binding protein known to colocalize with GABA in a number of neuronal cells. NT-4/5 strongly increased the number of calretinin positive cells in cultures prepared from rats of embryonic day 15, as we]l as calretinin levels determined by Western blot analysis. When cultures were prepared from embryonic day 18 rats, NT-4/5 very strongly increased the morphological differentiation of calretinin positive cells. While all effects produced by NT-4/5 were mimicked by BDNF with similar potency, NT-3 was found to be only marginaUy effective, Our findings identify NT-4/5 as a potent neurotrophie factor for striatal GABAergic neurons. Fusion of cells and organelles is a general patho-biological phenomenon: Muscle cells, transport vesicles, Langhans cells and virus induced fusion from within (FFWI, during replication) and without (FFWO, by the particles). In vivo and after virus isolation only little fusion can be seen, only after cultivation fusing HSV can be detected; a selective pressure must be released. 6 syn loci are known on the genome; the syn 3 locus is located inside of the cell in the carboxy terminal part of glycoprotein B, all other fusion domains are outside the ceil. Fusion inducing mutations are in aa 816, 853, 854, 856, 857 and 872 and were correlated to fusion from within and without as well as to selective Cyclosporin A effects. A model is presented for the 3 syn locus. -By recombination the fusion capacity could be transfected to fusion negative viruses. -HSV penetrates the cellmembrane by fusion by 2 ligand-recepter interactions. FFWO+-strains penetrate at 0~

Cell/cell fusion was analysed and found to need also 2 receptor-ligand interactions if induced by syn 3 mutations. Timing analysis of fusion events and by certain inhibitors allow further conclusions. quenching of LHCII isolated from dark-adapted (LHCII-D) or preilluminated (LHCII-L) rye leaves was similar in both preparations, but reversibility of this process within two hours of darkness was slower in LHCII-D.

No differences in fluorescence quenching and full reversibility was observed for both LHCII preparations in reverse micellar solution. Xanthophyll/Chl concentrations in LHCII were consi-derably decreased under this conditions. Excitation difference spectra (before and after illumination) of LHCII in asolectin liposomes showed typical changes of energy transfer between carotenoids and Chl. A model is proposed according the xanthophyll cycle controls reversibility of light-driven excitation quenching as well as the xa~a~q~_x~tentof ~,,~nchi~g i/~ LHC~Z. Myelinogenesis is a complex, developmentally regulated process involving coordinate expression of myellnafion-related proteins. The myelin forming cells are the oligodendrocytes in the central nervous system (CNS) and the Schwann cells in the peripheral nervous system (PNS). We differentially screened a rat postnatal day 16 (P16) spinal cord eDNA library with probes derived from normal (plus probe) and from myelin-free (minus probe) P16 spinal cord mRNAs. We succeded in the isolation of several novel eDNA clones whose corresponding mRNAs were expressed either selectively by oligodendroeytes or by oligodendroeytes and Sehwann cells.

Here we describe one of the eDNA clones (tentatively denominated NS 2) whose mRNA is specifically expressed by oligodandroeytes and Schwann cells. The 3.5 kD transcript is expressed at highest level during myellnation and at much lower levels in the adult as it is known for other myelin-specific mRNAs. Sequence analysis of the full length eDNA sequence showed a 50% identity to the rat liver UDPglucuronosyltransferase family, involved in the detoxificafion system. The 541 amino acid open reading frame encodes a 62 kD protein with a putative signal peptide and two transmembrane domains, Whether this NS 2 protein is involved in a detoxification reaction or in glycosilation of proteins and lipids with glueuronie acid is under investigation. Neuroblastoma (N8) are sympathetic tumors of childhood characterized by MYCN amplification in advanced stages. CD44 standard (CD44s) and splice variants (CD44v) represent a group of surface glycoproteins whose expression has been recently linked to metastasis. We analysed the expression of CD44s and CD44v on N8 tumors and cell lines by immunohistology and Northern blot. Results showed high levels of CD44s on 33/44 NB, 6/10 N8 cell lines and 4/4 ganglioneuromas (mature, benign sympathetic tumors). Lack of CD44s staining was observed on stage IV, MYCN amplified tumors only. In cell lines, expression was not detected on cells with a neuronal phenotype while high levels of CD44s were expressed by cells with a glial differentiation, independently of MYCN amplification. Up-regulation of CD44s was obtained with agents that induce a glial differentiation, while neuronal differentiation by retinoic acid was not accompanied by a change in CD44s expression. No GD44v were detected on tumors or cell lines. We conclude that CD44s expression is down-regulated in a subset of NB ceils and tumors and that MYCN product and additional mechanisms responsible for control of differentiation are involved in this regulation. The glycoprotein gC of herpesviruses is known to be non-essential for virus replication. Non-essential herpesvirus proteins are suggested to perform "luxury functions" which may influence the outcome of an infection. We are aiming to analyse the function(s) of IBC specified by bovine herpesvirus 1 (BHV1), bovine herpesvirus 5 HV5) and caprine herpesvirus 1 (CapHV1). These viruses are closely related but differ in their pathogenic potential. In a first step we have cloned and sequenced the individual genes. The nucleotide and the deduced amino acid sequence homologies of two 8HV1 reference strains was found to be >98%. The sequence homologies between BHV1 and BHV5 were >75% and those between BHV1 and CapHVl were >65%. Comparisons on the amino acid sequence level revealed that the differences were most prominent in the N-terminal parts, whereby the potential signal sequence region might be concerned. The putative glycosylation sites, however, were not affected, and the transmembrane sequences were similar in size and location. In order to characterize the significance of these differences, we are constructing gC-deletion mutants and recombinant viruses carrying a foreign gC gene. The in vitro and in vivo behaviour of these viruses will be compared to the wildtype virus.

$15-67

O. MOULLET and J.-L. DREYER, Instimt de Biochimie, Unlversit6 de Fribeurg, CH-1700 Fribourg cAMP has been recently shown to promote a cell survival response by retarding apoptosis (Berridge, M.V& el. (1993) Exp. Hematol. 21,269-276) . On the other hand quinones are widely distributed substances of often potential toxicological significance. We have tested the effects of quinones on adenylate cyclase. The enzyme is rapidly inhibited by pbenzoquinone (IC50=40-45 ktM) or dicnlorophenol-indopbennl (/6'50=200 ILM), but 2-substituted quinones are inactive. The lC50's are decreased 3-10 times after membrane soinbilization but are not affected by GTP, GDP or analogues nor by cholera and pertussis toxins. The inhibition is not mediated by a G-protein or by the activation of a defined receptor. Quinone inhibition stoichiometrically competes with forskolin activation of adenylate cyclase, equimolar concentrations of beth quinone and forskolin restoring the enzyme activity to its basal value. The cytotoxicity of benzoquinone, tested in vivo on Hep-G2, a human bepatocellular carcinoma cell line, could be prevented with forskolin. In plasma membranes quinones tightly bind to only one major and two minor proteins that were purified by PAGE electrophoresis under native conditions. Together these observations indicate that quinone action can be attributed to targetting to a limited number of proteins at tile plasma membrane in a highly selective way. By affecting adenylate eyclase, a modulation of the cAMP pool induces apoptosis as a result of the cytotoxicity via. T.congolense is an African,unicellular hemoflagellate, pathogenic for domestic animals such as cattle.Within the host bloodstream,parasites adhere to the wall of the microvasculature,eausing severe organ damage.We have set up and characterized an in vitro assay in order to study the metabolic requirements,reeeptor-ligand interactions and the ultrastructure of this adhesion phenomenon.Our findings suggest that adhesion is an active process requiring metabolic energy on part of the parasite, but is independent of the target cells.We also show that T.congolense possess a leetin-like domain localized at specific sites along the surface of the anterior end of the flagellum, which interacts with specific carbohydrate receptors, probably sialic acid and/or N-aeetylglueosamine residues,on the endothelial cell surface. This suggests that adhesion is likely to be mediated by a trypanosomal surface component distinct from the variable surface glycoproteins(VSGs). We also suggest that the cytoskeletal protein aetin is involved in this process. Local immune response based upon intestinal parasitespecific T and B cells to Giardia lamblia may be impomrant in parasite clearance. Experimental infections Qf mice (CR:NIH:S) with G. lamblia (clone GeM-H7 which e~presses a major 72kD antigen on its surface recognized b~ MAb G10/4) have demonstrated that the parasite undergoes surface antigenic variation in rive. Experimental i~fection of immu/%odeficient mice (nu/nu mice and sci~d mice) and oontrol nu/+ littermates provided insights into associated immunological mechanisms: dominant surfade epitope-specific slgA plays a major role in modulatin6] surface antigenic variation, and thymus-dependent T-lyn~phecytes in the induction of selfcure. Athymio nu/nu mioe (ZU. ICR-strain) were reconstituted with immune Peyer:~ patch lymphocytes obtained from self-healed littermat~ nu/+ mice. Intestinal B-cells from these nu/+ mice @s well as from reconstituted athymic nude mice synthesized in vitro parasite-specific IgA. This IgA showed a predominant immunoblet reaction with the major surface antigen (a Mr 72,000 polypep-tide) characterizing the Giardia clone in question. Nu/nu mice reconstituted wit~ immune cells acquired the potential to decrease significantly their intestinal parasite mass. Thus the hypothesis on the causative role of intestinal IgA in tl~ control of intestinal G. lamblia infection has found further experimental support.

Stephan Jentsch, Heinz Tobler and Fritz M011er, Institute of Zoology, University of Fribourg, P4rolles, "1700 Fribourg

The process of chromatin diminution in the parasitic nematode Ascaris lumbricoides leads to the formation of somatic cells that contain less DNA than the germ-line cells. During this process, which occurs between the third and the fifth embryonic cleavage divisions in five different blastomeres, the termini of the chromosomes are cut off and eventually degrade in the cytoplasm. To analyze this process at the molecular level we constructed a XZap somatic telomere library. The analysis of three cloned telomeres and their corresponding germ-line genomic regions revealed that chromosomal breakage takes always place within short specific chromosomal regions (CBRs) and is followed by the addition of the telomeric sequences (TTAGGC)n to the breakage sites. The different CBRs do not crosshybridize to each other. A comparative nucleotide sequence analysis is now being performed to screen for the existence of conserved sequence motifs.

In a parallel approach we analyze the chromatin structure of the CBRs and their flanking regions in developmental stages before and during the elimination process, Putative recognition or regulation sequences important for the elimination process might be recognized as DNasel hypersensitive sites. Once such sequences will be identified, it should be interesting to establish their role in the elimination process and to look for specific binding factors. Chromatin diminution takes place in all presomatie ceils of the early embryo of Ascaris lumbricoides and is followed by the addition of many repeats of the telomeric sequence TTAGGC. Chromosomal fragmentation is developmentally regulated and occurs within specific chromosomal regions (CBRs). One of these CBRs (CBR1) was analyzed in detail. Agene located close to the CBR1 encodes a putative GTP-binding protein, whose promoter region is located within a distance of only 2 kb from the telomeric TTAGGC repeats of the corresponding somatic chromosome. A homologous gene was isolated from C. elegans. Most interestingly, the C. elegans GTP-binding protein gene is also located at a telomeric position, namely at the end of chromosome V. In Ascaris, transcripts of this gene are present in all developmental stages, though they seem to be stronger expressed in oocytes and early embryos than in later developmental stages and somatic tissues. A high degree of sequence conservation of these genes between the two nematode species and man (a corresponding partial cDNA clone was identified in a human heart cDNA library) suggests an important biological function. Currently, we are using genetic methods to determine the possible function of this gene in C. elegans and to test whether its telomeric localization in both nematode species has any functional importance. it encodes a nuclear protein which is associated with she chromatin together with other probeins of the poiyoomb group. The mechanism of gene regulation caused by Pelycomb seems to be similar to that of the modifier genes of the position variegation effect which also encode structural proteins of the heterochromatin. One of these proteins is HPI which shares with Polycomb a region of similarity at the amino terminus called the ~omatin ~rganisation M~difier domain (chromo domain). The particular function of the chromo domain is not yet understood, but it seems to be important for the protein-protein interactions in chromatin associated complexes. Furthermore, chromobox cDntai~ing genes were found in several species, suggesting that they are widespread in the animal and plant kingdoms. Here we show chat ~. clemens contains at least one chromobox containing gene (cDNA: cm08h7). Length and szrusture of the putative ~. elegs~ chromo domain protein are similar to those of the chromo domain containing prozeins of other species. The expression pattern of cm08h7 is s:udied with antibodies raised against the putative chromo domain protein and by injecting a 5-ga! fusion construct inns the germ line of ~. eleman$. Chromatin diminution in the parasitic nematode Ascaris lumbricoides is a complex molecular mechanism, involving cleavage of the chromosomes at specific breakage regions, degradation of the eliminated chromatin and new telomere formation in all presomatic cells during the early embryonic development. The aim of the present project is to reproduce this DNA elimination process in vitro, step by step or all in one, by using cell-free extracts of A. lumbricoides eggs. Therefore, 4-8 cell extracts were established and their quality was assessed by the ability to transcribe 5S rRNA genes (pol III) and SL RNA genes (pol lI). Faithful extracts were then tested for cleavage activity on DNA substrates derived from different chromosomal breakage regions. Finally, preliminary results revealed a possible non processiv RNAse sensitive telomerase activity in the in vitro extracts, capable of adding specific nucleotide sequences to an oligonucleotide primer containing four repeats of the telomeric sequence TrAGGC. As de novo synthesis of several kilobases of somatic telomere is likely to require a strong telomerase activity, we assume that this A. lumbricoides in vitro system is a good tool to isolate the telomerase protein and its corresponding gene or other implicated factors, S16-07

University of Berne, CH 3010 Berne A common feature of all mycobacteria -whether pathogenic or not -is their richness and variety of lipids. Among numerous lipids widely distributed, pathogenic mycobacteria exhibit a group of sulfated surface lipids thought to be determinants of virulence. Therefore, we studied sulfolipid-bioslrnthesis in pathogenic and apathogenic (opportunistic) mycobacteria species by monitoring the kinetics of [35S]sulfate incorporation into cellular lipids, Pathogenic mycobaeterla (two virulent M. tuberculosis strains) showed a marked sulfolipid-synthesis with highest rates during exponential growth, while apathogenic ones (an avirulent M. tuberculosis strain and an opportunistic species) manifested negligible incorporation of the label. The partial characterization of the [35S]-sulfated lipids of pathogenic mycobacteria by thin layer chromatographic separation, combined with autoradiographic detection, revealed the presence of several radiolabeled lipid components of different polarities and with altering expressionpatterns during growth. These results strengthen the hypothesis, that sulfolipids are involved in the pathogenesis of mycobacterial disease. maiada (1,2) . Some features of human Pf maiada are observed in mudne malaria, although the pattern of sequestration changes due to the different plasmodial species involved, In mice, we showed by immunohistocbemistry that the expression of VCAM-1 and ICAM-1 is upregulated in brain, lung and heart of P/asmodium bergheiinfected Baib/c and also in LPS-pdmed 8CID mice. In $GID mice injected with labeled human erythrocytes (E), a higher rate of sequestration to brain was observed with PfE than with uninfected E. LPS-pdming increased the retention of PfE in the brain and lungs significantly (t test), but decreased it in the spleen. We conclude that PfE express surface structures which allow them to adhere to the vasculature of the brain more efficiently than uninfected E. LPS increases the number of adhesive molecules on the host endothelium. Interestingly, sequestration in the 8CID mouse resembles that of PfE in man. 

F. Roth, F. Riniker, K. Schopfer and T. Burkart Inst. Med. Microbiology, Univ. of Berne, CH 3010 Berne It has recently been shown that bacterial lipids cancarry antigenic determinants. The study of antibody specificity to such lipid epitopes in vitro turns out to be difficult since hydrophobic antigens have to react with watersoluble molecules. We have developed a solide phase lipid antigen immunoassay (LAIA) that allows to quantify the antibody reactivities with an array of lipid antigens derived from mycobacterial cells. Defined amounts

Of extracted lipids were immobilized as equidistant lines on a nitrocellulose sheet. Small strips of nitrocellulose, each carrying the same sets of lipids were incubated with different human sera. Bound antibodies were then reacted with [125-I]-Iodine labelled antihuman antibodies and visualized by autoradiography.

The reaction was quantified by densitometry.

Assay conditions were optimized for antigen and antibody concentrations and appropriate incubation times. In addition, the effects of different ingredients (salts, proteins, detergent) on antigen immobilization and epitope accessability were studied. The presented LAIA-system is rapid, sensitive and reproducible. It allows to compare the specificities of different bacterial lipid antigens and to quantify their reactivity with serumantibodies. The atomic force microscope is a very convenient tool for studying hard and flat samples with atomic resolution. Biological objects, usually soft and rough, are sometimes difficult to image using this technique. In contrast bacteria, which are too small to be observed with high resolution using the optical microscope, are hard enough to be observed with an AFM at a relative good level of magnification. This kind of microorganisms can be prepared for AFM imaging using very rapid and simple techniques. This method could be a convenient tool to observe the effect of bioactive substances such antibiotics. We show in this poster an exemple with the action of penicillin on B. subtilis. Within the family Trypanosomatidae, Leishmania is a protozoan pathogen producing a broad spectrum of human disease. Its life cycle is divided in two stages. Promastigote is the flagellated form which colonizes the gut of the sandfly vector, and amastigote is the non-flagellated form that is specialized for growth and survival in the vertebrate host macrophage. To understand gene regulation during the transition between the promasttgote and the amastigote stage, we were interested in the characterization of stageregulated genes. We isolated a gene, sw3, from Leishmania major whose mRNA is more abundant in amastigote compared to promastigote. Structural analysis of this gene showed that an additional polypeptide could be encoded by the opposite strand of sw3. We confn'med the presence of an open reading frame on the opposite strand by in vitro translation of the sw3 antisense RNA. In vivo, an amastigote SW3 antisense transcript and the corresponding polypeptide were also detected suggesting that the sw3 is transcribed in both directions and that stage specific transcripts and polypeptides could be produced. Analysis of other gene segments tend to indicate that an intensive usage of the Leishmania genome to produce various polypeptides could be a general phenomenon in trypanosomatidae. The four variants and / or posttranslational modifications of histone H1 of procyclic Trypanosoma brucei brucei (Protozoa, Kinetoplastida) were purified by HPLC reversed phase chromatograpy and characterized by their amino acid composition and partial amino acid sequence. Differences in sequence of up to 45% as compared to histone H1 of higher eukaryotes exist. Alkaline phospatase digestion and analysis of H1 by means of HPCE was used to demonstrate the presence of phosphorylated modifications. The unique biochemical properties of H1 of T. b. brucei contribute to the structural differences seen in the chromatin of the parasite as compared to that of the higher eukaryote host. Larvae of the parasitic wasp Chelonus inanitus (Braconidae, Hymenoptera) develop inside embryonic and larval stages of the host Spodoptera littoralis (Noctuidae, Lepidoptera). Parasitism induces in the host a precocious onset of metamorphosis and developmental arrest in the precocious prepupa. The wasp injects, along with the egg, PDV and venom into the host egg. Injection experiments with PDV and venom revealed that PDV play a crucial role in altering host development and in suppression of the host's immune system. The genome of the PDV of C. inanitus consists of at least i0 segments of double-stranded circular DNA ranging in size from 7-30 kb. We analyzed the expression of viral genes in the course of parasitization with cDNA made from mRNA at various developmental stages of S.littoralis. We also investigated the localization of the expressed viral genes on the various segments. In addition, these bacteria were isolated and cultured from brain tissue. To identify the infectious agent we used, among other techniques, an atomic force microscope (AFM). When the. isolation fluids derived from the cortex of three Alzheimer's cases were examined with AFM and scanning electron microscopes, we observed bacteria whose size and morphology fit to the order of Spirochaetales. These images are presented and discussed on our poster.

~ottstein Bruno, Institute of Parasitology, CH-Berne. %lveolar echinococcosis AE is due to infection with the metacestode (larval stage) of Echinococcus multilocularls. An immunodominant antigen Em2 of E. multilocularis Was characterized by MAb-GII, Respective studies revealed, that (a) antigen expression metacestode stage-sDeci-fic and (b} that expression was restricted to the laminated layer. The Em2-antigen i s a lectin-binding carbohydrate linked to a lipid residue. The "Em2positive" laminated layer proved to be a crucial factor in protecting the parasite from host immune-effector mechanisms: only E. multilocularis larval structures exp ressing Em2 were able to induce secondary alveolar chinococcosis in rodents. Subsequent experimental studies in resistant (C57BL/10) versus susceptible (C57BL/ 6J and AKR) mouse strains showed that resistance was as-Sociated with synthesis of anti-Em2 IgG 3 and IgG I. Sus-Ceptibility of C57BL/6J-mice was associated anti-Em2 s and no IgG 3 . The parasite-specific in vitro lym-~hoproliferative i~une response remained stationary ~ith time after infection in C57BL/10 and decreased in r and AKI< mice. Day 30 p.i. CD4 § -lymphoblast cells dominated in all meuse strains; day 90 p.i. an in-Creased nu/0ber of CD4-/CD8 + and CD4+/CD8 + cells were observed. Only susceptible mice exhibited a significantly ~eereased response of lymph node cells to ConA-stimula-~ien with time p.i. In conclusion, subtypes of parasitespecific cellular and humoral host immune responses seem Leishmania species are protozoan parasites causing a spectrum of clinical manifestations in man ranging from cutaneous lesions to morbidity and death. The insect stage (promastigote) and intracellular stage (amastigote) differ in terms of morphology, physiology, antigenicity and gone expression. It is clear that the amastigote stage is uniquely adapted for survival within the inimical environment of the macrophage phagolysosome and that mechanisms regulating these adaptations must be called into play during the early stages of infection when the parasite undergoes transformation. In an effort to elucidate these regulatory mechanisms, we have constructed a "subtracted" cDNA library which is specifically enriched for parasite transcripts expressed during the first 7 hours of infection ("switch" genes). Several clones have been characterized by sequence, Northern, and Southern analysis. Three clones, Sw3, Sw44, and Sw20, are expressed at a several-fold greater level in the amastigote stage and potentially encode proteins which interact with nucleic acids. Sw3 predicted protein is homologous with histone H1 proteins and Sw44 and Sw20 potentially encode ribosomal proteins. Analyses of transcripts from this library are yielding clues not only to mechanisms underlying the regulation of parasite transformation, but also insights into post-transcriptional and translational control of gene expression in Leishmania.

$16-17

Microbiology, University of Geneva, Medical School, C.M.U., 9 avenue de Champel, 1211 Geneva 4.

A viral RNA representing the sequence of a defective interfering RNA, naturally selected for efficient replication, i.e. only containing the replication promoter, was cloned under the control of the T7 RNA polymerase promoter. The expressed 17 transcript was then replicated in rive by supplying in trans the replication functions, equally expressed from plasmids. This system, completely accessible to genetic manipulations, allowed us to define a basic rule directing Sendal virus replication, the "rule of six" (Calain and Roux, J.Virol. 67 : 4822-4830, Ig93). A second defective RNA, representing a shorter version of the viral RNA, in that it contains not only the replication but also the transcription promoters, was then cloned and replicated in the same system, although with lower efficiency. The replication of this "transcribing" viral RNA was found to obey the "rule of six" as well.

replicationitranscdpfion promoters were then achieved to define the cissequence requirements needed for efficient replication or for replication /transcdption. Results obtained with these chimerical viral RNAs will be presented and their contdbufion to the comprehension of the paramyxovirus replication and transcription processes Bovine T-cells that become infected with the intracellular parasite Theileria parva undergo lymphoblastoid transformation and acquire the ability to proliferate continuously. They cease to require specific antigenic stimulation, become independent of exogenous growth factors and cause tumors in nude mice. The IL2 and IL2R genes are constitutively expressed and the transcriptional activator NFwd3 which regulates these two genes is continuously activated.

These processes are all strictly dependent on the presence of the parasite in the host cell cytoplasm and cease upon the specific killing of the parasite by the theilericidal drug BW720c.

We have shown that the presence of the parasite results in the activation of the protein tyrosine kinases fyn and ick, which can participate in early T cell receptor signalling, suggesting that the parasite may use this signal transducti0n pathway to induce continuous proliferation.

Cysteine proteinases of Leishmania as targets for chemotherapy. Jacques Bouvier*t and Judy Sakanari*.

*Department of Pathology, UCSF School of Medicine, San Francisco, CA, and tAnimal Health, Ciba-Geigy, CH1566 St. Aubin, Switzerland.

The overall goal of the study is to apply the structure-based approach to develop lead compounds and design novel enzyme inhibitors as potential therapeutic agents against leishmaniases. In this regard, we have identified and characterized the cysteine proteinases of leishmania major, and showed that they consist of excellent targets for drug design. We have purified a membrane-boand cysteine pmteinase not previously described in other trypanosomatids or protozoans and obtained amino acid sequence information from the purified enzyme. In addition, we have also characterized and purified the soluble cysteine proteinases of the parasite. Consequently, we have isolated two genes encoding the soluble and membrane proteinases. The soluble cysteine proteinase gene occurs in multiple copies of 2.9 kb with flanking regions of 1.2 and 2.2 kb and is localized on one chromosome of approximately 440 kb. The membranebound enzyme gene is 3.1 kb and occurs as a single copy on a chromosome of approximately 1100 kb.

A three dimensional computer model of both genes wiU be generated and enable us to scan the vast chemical database for new lead componds. We akeady have identified several cysteine proteinase inhibitors that are effective compounds in killing both promastigote and amstigote stages of the parasite in the micromolar range in vitro without affecting the host cells.

Dollenn~ier, G., Cassinotti, P. and Weitz, M., Institute for clinical Microbiology and irananology, 9000 St .Gallen/SWit zerland HAV is a ~r of the picornaviridae. Its RNA gencrne rF~y be divided into 3 coding regions (PI, P2 and P3), qhe P3 region contains a protease 3cP ro that generates mature proteins by cis-cleavage of the initial polyprotein precursor.

Catalytic activity of 3CP r~ has been d6~aonstrated in a vacciniavirus/T7 ~qA-polymerase hybrid expression system. (balOuter aided ali~ts of HAV 3CP r~ sequence with that of other picornaviruses imply that amino acid residues His-44 and/or H) the complete HAV open reading frame, HAV specific expression products were identified by radioJ_rmmnoprecipitation and lyy SDS-PAGE. The analysis with an antibody specific for putative nonstructural protein 2B revealed a 27,5 kDa peptide. This is in contrast to its predicted size (12 kDA). However, analysis with anti-(putative)2A antiserum confirmed an upstream location of the 2B N-terminus. The new 2Bpeptide was also identified in lyeates of HAV infected MRC-5 cells. Hence, recombinantly expressed HAV polyprotein underwent authentic processing and the predicted P2 organization has to be modified. We are interested in the development of animal models for one of the two major pathological changes found in the brains of patients with AIzheirnar~s Disease, the formation of fleuroitbrillary tangles, a main component of which is abnormally phosphorylaled Tau-protaln. (Tau itself is a protoin associated with mlcrotubuli.) Two fundamental questions may be answered with our models: First of nit we woald like to reproduce the formation of tangles, and secondly we hope to answer the question, whether the formation of these pathological structures is sufficient for the bevalopment of Senile Dementia ot the Alzheimer Type (SDAT). At present there are only a tow putative candidate genes known which may be in-voNed in the hyperphosphor~lation ot Tau in v/vo, one of which is the MAP-Kinase ERK2. We have cloned this candidate kinase from rat txaln RNA via an RT-PCR approach and expressed the cONA under,the co.rat Of tour different promoters in transgenis mice. These promoters were tested for maximal expression. A similar approach was chosen to express the human Taut0 isofonn in the brain ot transganic mice. We have stated to analyze different Tau an~ ERK2 expressing lines, Northern blot analysis has revealed high levels of expression for all transgenes (up to fNetoId {n comparison to endocjenous levels). The neuronal spectficRy of the transgenio ERK2expression was c~nfirmsd by in situ hybridization analysis. Ir~ addition several TAUexpi'esalog tines were analyzed in Wealernblots to determine the relative amount of hTau= protein in the brain. Sections of the brain were stained with Tau-sp~c~c ardib~iies to identify the neurons expressing Tau. From intemrosses of Tau-with ERK2-expressing lines, we have already identified double-positive mice which we currer~y analyze by immunohistocbemistry and in Western Nots to examine the phosphorylaiton status of Tau.

We are well aware of the possibility that only animals trans~enic for both Tau and ERK 2 might beveiop tang{es, whereas single transgenics mlsht not. Pairs of individual neurons were recorded with sharp electrodes or in whole-cell configuratio~ in hippocampal slice cultures of rat. Synaptic connections were studied between pre-and postsynaptie cells in cA3 and CA1 hippneampal fields by eliciting single aetinn potential in the presynaptic cell. Monosynapfic excitatory connections were highly probable between CA3 and CA1 pyramidal ceils (p=0.76, n=82) with almost no feed-forward inhibition (p=0.01). By contrast, disynaptic IPSPs, blocked by CNQX or bieuculline were observed with a very high probability (9--0.6) between CA3 pyramidal cells. Monosynaptic excitation was found ha 56% of cases between CA3 neurons (n=43) but only in 27% of cases (n=ll) between two CA1 cells. Monosynaptie IPSPs with very short lateneies (<lms) and blocked by bicuculline, but not by CNQX, were recorded within area CA3 (5/6) and within CAI (2/3), but not between CA3 and CAt fields (2/2). Probability of transmitlzr release, studied in recordings where failures could be unambiguously distinguished from spontaneous or small PSPs, was found to be close to 100% at both excitatory and inhibitory synapses. Specific inhibitors of presynaptie release at excitatory and hahibitory terminals (i.e. adenosine and opioid peptides, respectively) gradually reduced the anaplitude of the PSPs indicating that the release at excitatory and inhibitory unitary synapses is multiquantal. The CNS of a neonatal opossum, Monodelphis domestic& shows profuse growth of fibers across a lesion. As in other mammals, such repair is not seen in adults. Experiments have been made to define the age of the animal and stage of development at which repair ceases to occur. The entire CNS was removed from animals aged 3-6 days or 11-14 days after birth. The spinal cord was crushed and growth of fibers assessed 5 days later by electrical recording of impulses conducted through the crush. Repair after lesions to spinal cords of younger animals occurred in 38% of the trials (45 animals, aged 3-6 days). In nervous systems from older animals (11-14 days after birth) the success rate was low (only 4 out of 38 preparations). Similarly, the carbocyanine dye Dil labeled numerous fibers in spinal cords of younger animals (12 out of 21), whereas only one preparation out of 15 showed scant outgrowth into lesions in the older animals. Our results suggest that there is a critical period for neuronal repair; at about 12 days nerve fibers lose the ability to grow across lesions. It is an advantage that at this later stage the nervous system is still small enough to survive in vitro. Features correlated with these changes are oligodendrocyte differentiation and myelin formation. With immunofluorescent staining as well as electron microscopy myelination becomes apparent at 8 days after birth. We are now using time lapse-videomicroscopy to analyze molecular mechanisms that could play a part in promoting or preventing repair. Supported by Grants to J.G.N from the Swiss Nationalfonds (31-3626292) and from the Internat. Res. Inst. for Paraplegia. i. Impulse conduction in axons has integrative functions. The dorsal root ganglion (DRQ) of the embryonic rat put into culture forms a monolayer and Is thus accessible to eonventionalmieroelectrode technique, which allows testing the above hypothesis. We have chosen a dual approach by correlating ext~rimental el~ctrophysiological data from the chlture with results obtained from a compartrhental computer model. ii. The safety factor for conduction was found to be low. Thus failures of AP invasion of the DRG cell soma could occur at sites of impedance mismatch when a second stimulus felt into the relative refractory period of the first or when the axon was repetitively stimulated. The electrotonic resiuues (E.Rs) of the failed APs had discrete amplitude levels, suggesting that the failures were always caused at the same site along the axon. These sites of low safety factor were found to be the branch point in the unipglar DRG cell and the entrance of the stem piece into the soma in both cell types, the bipolar as well as the unipolar. A systematic comparison of 15oth cell types showed that the AP conduction through the latter is more reliable.'The length of this stem piece is inversely' correlated to the delay caused at the branch point, as tile electrical load of the soma is more efficiently masked by a long stem piece. The dendrites of motoneurons (and subsequently many other neurons) have been assumed to be electrically passive since the work of Coombs, Ecc]es and Fatt in the 1950's. However, direct evidence has been impossible to come by until recently with the advent of appropriate dyes and equipment for measuring very small and very fast optical signals. We examined the propagation of action potentials, both spontaneous and evoked, in the dendritic trees of spinal cord neurons using a 12 • 1 2 array of photodiodes and a CCD camera as weft as conventional microelectrode techniques. Organotypic slice cultures of embryonic rat spinal cord were stained with a voltage-sensitive dye (di-8-ANEPPS) or a calcium dye (Fluo-3) and recordings were made from motoneurons identified by their location and morphology. The results indicate that there is in fact an increase in calcium in the dendrites that accompanies action potentials whether synaptically evoked or evoked by current injection at the soma. Spikes in the dendrites up to 1 60 i~m from the soma are much larger than could be explained by models assuming passive dendritic trees. Work is currently underway to test the possibility that sodium conductances might contribute to the propagation of action potentials in the dendrites. Cat auditory cortex contains multiple cochlear representations in both hemispheres that are interconnected through the corpus callosum (CC). The distribution of auditory callosal axons was studied on the mediosagittal level after injections of biocytin at electrophysiologically identified locations. Single injections were done in AI (2 cats), AAF, AII and PAF; multiple injections in AI & AAF in one cat. An additional cat received 1 injection in SII. The mediosagittal image of CC was subdivided into 30 segments along a rostro-caudal axis. Labelled axons crossing the midline were counted in coronal or horizontal sections. Axons from auditory fields were found in the posterior 2/3 and those from the somatosensory cortex in the anterior third of CC. Axons from a discrete injection (ca Imm in diameter) spread over as much as i/5 of CC. The rostrocaudal distribution of axons at the midline reflected roughly the rostrocaudal position of the field of origin, but apparently not the tonotopic order of a given field. The major efferent projections of substantia nigra pars reticulata (SNR) convey information to the thalamus, to the tegmentum, to the superior colliculus, and to the spinal cord. Such a range of targets suggests a highly organized activity within SNR and this study investigates the temporal organization of spike trains. In the portion of SNR between +2.7 and +4.5 mm interaural levels, we recorded 40 single units along 6 electrode tracks in 4 young female Wistar rats. Most units (n=32) fired in a non-regular Poisson process, although only a small number (7/32) was characterized by an average burst size with more than 3 spikes. Crosscorrelograrns were computed for 25 pairs of units simultaneously recorded from the same electrode and for 26 pairs simultaneously recorded from distinct electrodes. Most crosscorrelograms (27/51) showed a symmetrical peak centered on time zero, which indicates a tendency to synchronous firing. When units were recorded from the same electrode tip we observed that 1/8 to 1/40 of their spikes were synchronous. About half of the significant interactions were characterized by a relatively long duration (>250 ms) and are likely to correspond to a different mechanism of synchronization. These results support the hypothesis that SNR cells receive common afferences of two different types and their fine temporally organized activity yields coordinated activity between the target structures. The neuronal microtubule-associated protein MAP2 binds to microtubules via a domain containing 3 or 4 repeats of an 18 amino acid sequence. We have previously shown that when cultured nonneuronal cells are transfected with MAP2 their microtubules become stiffer and are then capable of supporting process outgrowth when their cortical actin cytoskeleton is depolymerised. We hypothesise that this stiffening effect of MAP2 is caused by the 18 amino acid sequences binding to neighbouring tubulin subunits in the walls of the microtubules thus reducing their freedom of movement relative to one another. To investigate this further we have created MAP2 mutants in which 1 or 2 repeats of the tubulin-binding motif were deleted. These were tested by transfection and expression in cultured cells. The results confirm that the number of repeats affects the stability, stiffness and the cytoplasmic arrangement of cellular microtubules. We were previously able to demonstrate the presence ot kinin-like immunoreactive structures in the mouse brain. In order to characterize them, we have now performed a donble-immunofluorescence study using both polyclonaI auti-Srudykinin (BK) and mouse monocloual anti-neeron-specificenolase (NSE) antibodies. The latter is considered to be a specific neuronal marker. After fixation with Bonin-Hollande solution, the brains were removed, embedded in paraffin and serially sectioned at 10 am. The sections were incubated with the antibodies, which were visualized by an indirect immunofluorescence technique using FITC for NSE and by an avidin-biotin technique using Texas Red for BK. Double-immunofluorescent cells were found in the thalamus ventralis, in the nucleus cechleuris ventralis and in the nucleus mesencephaficns nervi trigemini (V). BK-positive cells in the nucleus principalis V were NSE-negative ih spite of their typical neuronal aspect. The other BK-positive structures (pia, ependyma, hnic~es) also were alwa~ NSE-negatlve. Most of the NSEpbsifive BK-negative cells were seen in the cortex. The neuronal nature of many kinin-like immunoreacfive structures in the mouse brain is thus ascertained. In contrast to the amphibian optic nerve (ON), ON of adult mammals is unable to regenerate after injury. Astrocytes disappear from and macrophages accumulate at the lesioned site (Blaugrund et al., J. Neurocytol, 118, 105, 1992) , We have followed the distribution of astrocytes and microglial cells in Xenopus tadpole ONs over 10 d after mechanical crush. Astrocytes were stained for cytokeratin and vimentin, and microglial cells for Griffonia Isolectin B4. Soon after crush, immunoreactivity for intermediate filament proteins was strongly increased. Astroeytic processes extended into the lesion from both proximal and distal stump. At ,5 d, astrocytes had crossed the injured region. In the distal stump, some of them contained vacuoles indicating phagocytic activity. While in the normal ON microglial cells were ramified and scarce, 2.5 d after crush they were roundish and vacuolized, and scarce within the lesion but numerous distal to it. At 10 d, their number was decreased and most of them had reacquired ramifications. Conclusively, in the tadpole ON, astrocytes are virtually never absent from the lesion site but rather extend rapidly into it. Astrocytes may thus provide a guiding support for outgrowing axons, Microglial cells that are frequent in the distal stump, do not accumulate at the lesion. The behaviour of the two cell types differs profoundly from that in mammals and is likely to be one of the factors that may contribute to successful neuronal regeneration of the amphibian ON.

CI4mence, J. F. 1, Ranieri, J. p.2, Aebischer, p.2, and Sigrist, H. 1, 1Institute of Biochemistry, University of Berne, CH-3012 Berne; 2Division Autonome de Recherche Chirurgicale, CHUV, CH-1011 Lausanne.

Small peptidic sequences of laminin (YIGSR, IKVAV) are known to promote attachment and/or neurite outgrowth of various neuronal cell lines (PC12, NB2, NG108-15). New and established heterobifunctional photo crosslinkers were used to immobilize these peptides in a topically defined fashion onto biocompatible surfaces such as hydroxylated fluorinated ethylene propylene (FEP-OH) and pely(vinyMlcohol) (PVA) . N-[m[(3-trifluoromethyl) diazirine-3-yl]phenyl]-4-maleimido-butyramide (MAD) was thermochemically linked to the synthetic peptide CDPGYIGSR via its N-terminal cysteine, leaving the bioactive part (YIGSR) free for cell receptor interaction. MAD-CDPGYIGSR was radioactively labeled by reductive methylation and purified by reversed phase HPLC. Patterned (20 and 300 #m) peptide immobilization was attained by photocoupling onto PVA and FEP-OH and visualised by autoradiography. Light-structured biomaterial surfaces with molecularly oriented nerve growth promoting peptides were investigated for spatially controlled neuronal cell attachment and differentiation. It is striking to observe that these inhibitory or stimulatory effects were corroborated by Binduced decrease or increase in 2-deoxyglucose uptake, respectively. This gives B a putative role in the limbic regulation. The sheet-like processes of oligodendrocytes wrap themselves spirally around central axons, thus formtng the myelin sheath. A single oligodendrocyte may donate internodal segments of myelin to each of 20-30 or more adjacent axons. It is not known, if one oligodendrocyte picks out axons randomly or if it prefers axons with a certain diameter or with a certain chemical specificity. We have studied this last possibility by combining intracellular injection of a fiuorochrome, and immunohistological detection of calciumbinding proteins (CaBP's), markers for the axons ef certain subpopulafions of nerve ceils.

Lucifer Yellow was injected into oligodendrocytes of the optic nerve, of living rat brain slices. The oligodendrocytes were identified by their electrophysiologieal properties. Slices containing the iniected cells were immunostained /or parvalbumin or calretinin using second antibodies labelled with Texas red. Using co.focal laser-scanning microscopy combined with a three*dimensional reconstruction programm, the relationship between oligodendrocyte processes and axons positive for one of the CaBP's was determined. For ultrastructural confirmation, single, Lucifer-Yellow injected, oligodendrocytes were photoconverted, and the CaBP's-positive axor~ were revealed by gold-labelled antibodies.

The few oligodendrocytes injected show the classic morphology, that is of a small cell body and few processes running parallel to the course of the axons. The proximity between glial processes and positive axons can be easily appreciated in confocal laser-scanning images. A preferential assodation of Oligodendrocyte processes with parvatbumin-positive axons has not as yet been found.

Thus, the confirmation of the hypothesis that oligodendrocytes choose their axonal targets according to chemical cues awaits further work including the injection of a lar~er number of these cells. To initiate a molecular genetic analysis of brain development in insects, we are characterizing the mechanisms of embryonic brain development in the fly Drosophila. Early in embryogenesis, a small number of molecularly diverse brain neuroblasts generates the neurons of the brain.

Subsequently, pioneering brain neurons initiate axonal outgrowth along glialbound brain compartments.

These pioneering neurons establish the primary brain tracts. The pioneering brain axons express the cell-cell adhesion molecule fasciclin I dynamically during outgrowth and initial fasciculation; a subset of the pioneering axons expresses fasciclin II. The axonal framework of the embryonic brain tracts is used for outgrowth and fasciculation by subsequently differentiating axons and, thus, prefigures the tracts of the mature brain.

A comparison of early brain development in Drosophila and in vertebrates reveals common mechanisms for brain development. Supported by the Swiss NSF. Institute ot Histology, University of Fribourg, P~rolles, CH-1700 Fribourg Calretinin (CR), an EF-hand Ca2+-binding protein, is expressed in CajaI-Retzlus cells during development of the rat cerebral cortex. These cells are located in the marginal zone and are the ear{iest cortical neurons to be generated. They have been consldered as unusual on account of their morphology and their fate during corticogenesis. The functional significance of CajaI-Retzius cells and their destiny in adulthood is still controversial. In order to approach these questions we have applied organotypic culturing methods.

Slices of neocortex from brains of 0-6 days-old rat pups were cultured for 10-21 days applying the interphase technique (Stoppini etaL, J. Neurosci. Methods 37, 1991) or the roller-tube technique (G&hwiler, J. Neurosci. Methods 4, I981) . The fixed cultures were immunolabe0ed with an antibody against CR. The obtained results showed that CR positive structures develop in vitro like in vivo. However, in organotyp[c cultures their distribution corresponded to a more mature stage than the situation on the day of the explantations. Many CR-positive cells were seen in layer L They were horizontally oriented or had a pyriform shape. Based on their morphology these ceils were identified as CajaI-Retzius cells. ORpositive cells were numerous in layer I1-l[I and showed mostly a bipolar morphology. Some muRipo[ar cells could also be observed. A fine net of CRpositive fibers and puncta was found in layers [ and iV. The demonstration that CajaI-Retzius celts are detectable in these cultures will allow us to follow the fate of these cells dynamically and [nte~ene in their function by applying exogenous factors.

Cervical primary afferent input to vestibule-spinal neurones projecting to the dorsal horn: A double labelling study in the rat.

Wheat germ agglutinin-horseradish peroxidase (WGA-HRP) was injected by pressure into cervical spinal ganglia C2 or C3. In the same experiments Fluorogold (FG) was iontophorefically injected into the ipsilateral dorsal horn of different cervical spinal cord segments. Anterogradely labelled fibers (WGA-HRP) were present mainly in the external cuneate nucleus, but also to a lesser extent in the caudal part of the medial and descending vestibular nuclei (MVN, DVN) . The FG injections, restricted to the cervical dorsal horn, revealed retrogradely labelled cells in the central part of the caudal MVN. A double exposure (fluorescent and polarisadon optics) under the microscope showed primary afferent fibers surrounding like baskets retrogradely labelled FG cells. The projection from cervical ganglia cells to the MVN represents a pathway for direct afferent information from neck receptors (in particular muscles) to vestibular nuclei and supplies the MVN with afferent information which could enable the vestibulospinal nenrones, projecting to the dorsal horn, to influence the local information processing. In the sciatic nerve, all myelinated and non-myelinated axons were SNAP-IR. In peripheral tissues, all nerve endings were positive. The preterminal Schwann cells of motor axons were also SNAP-25-IR. After sciatic nerve section, many myelinating Sehwann cells also expressed SNAP-25-IR. In conclusion, SNAP-25 is expressed by neurons in the PNS.

It seems to be also expressed by preterminal Schwann cells and by myelinating Schwann cells during regeneration.

$17-19 REPAIR OF COMPLETELY TRANSECTED SPINAL CORD OF NEONATAL OPOSSUM IN CULTURE H.A. Vischer, Dept. Pharmacology, 8iocenter, Uni. of Basel, Switzerland Woodward et al. (J. exp. Biol. 1993 , 1993 have shown that fibers grow rapidly and profusely across a lesion made by crushing the spinal cord of the neonatal opossum Monodelphis domestica. In such experiments careful controls must be made to establish that fibers were in fact broken by the lesion. Tests have now been made to determine whether similar repair can occur after a more drastic lesion involving complete transection of the cord. After dissection of the entire GNS from animals aged 3-8 days, the cord was cut into two with scissors at the C5 -C7 level. The two-ends were glued together with matrigel on a sylgard mold, which encompassed and held the CNS. The rostral end was labeled with the fluorescent carbocyanine dye Dil in order to visualize growing fibers. In 33 preparations fibers grew over the cut into the separated part of the spinal cord. Such growth could extend to 500 Ixm. Fibers grew straight, branched and multidirectionally, or in fascicles. In another set of experiments cords were combined from different animals of the same or different ages. Again, fibers could grow across the cut but they did so less frequently. These results set the stage for investigating fiber growth after rotating the two ends at a defined angle and for analyzing factors that influence the direction of growth. MAP la is a microtubule associated protein of 360 KD. In rat brain two monoclonal antibodies, A and BW6, recognized MAP la in neurons and their axonal and dendritic processes. In mouse cerebellum, monoclonal A recognized Purkinje ceils and granule cells uniformly, as already described for rat brain. In contrast, monoclonal BW6 stained selectively some dendrites of Purkinje cells, forming bands in the molecular layer. We compared this striation with that obtained with a monoclonal antibody, anti-zebrin II, which recognized aldolase C. In the mouse vermis, zebrin stained in a striated pattern, but complementary to BW6. These results demonstrate that MAP la is present in forms which are differentially distributed in the mouse cerebellum. These observations may be explained either by differences in metabolic states of neurons, or by differences in their regional functions, or by differences in the regional stability of microtubules. This work was supported by grant No 31-33447.92 from the Swiss National Science Foundation. Serum free aggregating brain cell cultures prepared from 16 day old fetal rat telencephalon are used as a model to study brain development. In these cultures it is possible to distinguish ontogenetic events such as cell proliferation, differentiation and myelination, In the present study we examined synaptogenesis by analyzing the expression of synaptic proteins such as synaptophysin, SNAP-25, synapsin I, and brain spectrin.

different developmental stages were analyzed by western blotting. Immunoreactivity for synaptic proteins was detectable already at day 12 in culture, suggesting that the first synapses are already present at this early stage. The expression of synaptic proteins strongly increased between day 20 and day 30 in culture, reflecting a period of intense synaptogenesis. Treatment of the culture with an elevated concentration of KCI (30mM) increased the expression of synaptic proteins, suggesting a stimulation of synaptogenesis. These findings demonstrate the utility of this approach to study brain synaptogenesis, and possibly synaptic plasticity. Drenhaus, U.I Gunten, A.v., Rager, G. ; Institute of Anatomy, University of CH-1700 Freiburg The retina of Tupaia comprises three types of ganglion cells (RGCs) analogous to X-, Y-, and W-cells found in other mammals. These cells differ in size and in their distribution pattern: large RGCs are frequent in temporal, small RGCs in nasal retina. As the fiber diameter correlates with the size of its respective RGC, it can serve as a parameter for the topographic representation of RGCs in the nerve. To study this we analyzed the optic nerves of three Tupaia. Using the electron microscope we recorded the density, diameter, and the position of fibers in the nerves. We found, that fibers with the largest diameter are located dorso-temporally and close to the center of the nerve. They are surrounded by zones of smaller fibers. The diameter decreases gradually towards the periphery and is smallest in the ventro-nasal region of the nerve. Since the fiber diameter distribution corresponds to that of the size of RGC-types in the retina, we assume that the topographic relationships in the nerve and the retina are similar. (supported by S 17-23

Stoppini Luc, S. Duport, L. Parisl, C. Oropes~, Departement of Pharmacology, CMU, 1211 Geneva 4 The micro environment of the central nervous system is important for neuronal function. In this context, the blood-brain (BBB) provides and maintains the extracellular medium that is compatible with normal neuronal and synaptic activity. Due to the difficulties inherent using whole animal as an experimental model to study permeability and metabolic processes at the cellular level, major efforts have been engaged In recent years, In order to bring up suitable /n vitro models. We are presently developping an In vitro BBB by interfacing a coculture of endothelial cells monolayem and nervous organotypic cultures. The main idea of this study is based on the premise that the complex Intercellular Interactions between the different cell types pertaining to the nervous tissue and the neighbeurlng endothelial cell is of fundamental Importance to promote a realistic BBB system. We expect that erganotypic culture of nervous tissue, combined to endothelial cell monolayers, will initiate and maintain a BBB phenomena/n vilro. We will test more specifically the hypothesis that neuronal activities [spontaneous or elicited) will influence gila] cell response which will In turn modify endothelial properties. PrelimInary results clearely show a good nervous tissue and endothelial cell survival. Tight junction llke structures could be identified usIng freezefracture or conventional transmission electron microscopic thechniques. {Work supported by Chemodyne Gent~ve ). The process of myelination is a prerequisite for the proper function of the brain since it enables rapid saltatory conduction of axons. In the central nervous system (CNS) myelination is performed by oligodendrocytes. A differential screening approach was used to isolate new rat cDNA clones that are expressed in this glial cell type. Here we report the further characterization of two of these novel cDNAs which appear to be expressed specifically in oligodendrocytes.

The two characterized cDNA clones (tentatively called CNS1 and COP-25.1) share an approximately 420 bp region of sequence identity which suggests that they are dedved from the same gene by alternative splicing, Within this area a putative open reading frame is located. We demonstrated by Northern blot analysis and in situ hybridization that the two mRNAs of the novel gene are expressed exclusively in oligodendrocytes. However, the mRNAs show a different specific spatial localization within the cell. We compared mRNA expression of myelin basic protein (MBP) and proteolipid protein (PLP), with that of CNS1 and COP-25.1 in brain and optic nerve during development. It appeared that the mRNAs of the novel gene were delayed significantly as compared to MBP and PLP mRNAs. Streit. Phyaiologisches Institut, Btihlplatz 5, 3012 Bern In organotypic slice cultures of the embryonic spinal cord, rhythmic spontaneous activity arises when the inhibitory synaptie transmission is blocked. The rhythmic activity consists of bursts of regular oscillations of activity at 4 -5 Hz. This study investigates the origin of the regular oscillations. When stimulated electrically at 5Hz, the synaptic transmission in the cultures showed strong depression by about 50% on average. The synaptie depression was highly variable among individual connections and tended do be less pronounced in frequently used connections. When random depression ratios with an average of 50% at 5Hz were implemented in a computer model consisting of 100 randomly connected excitatory cells, regular oscillations of activity did not arise, even in the presence of synchronous pacemaker cells. However, when individual depression ratios were adapted according to the rate of activity of individual connections, regular oscillations at 4 -5 Hz arose in the presence of few unsynchronized pacemakers. This finding suggests that the regular oscillations in the cultures originate in a network formed by the plasticity of synaptic depression. Maier A., schlumpf M., Beer H.F., Sehubiger P.A. and Lichtensteiger W., Inst. of Pharmacology, Univ. of Zurich, The ontogeny of expression of Dopamine D 1 receptor mRNA in the rat brain and binding to the receptor was studied at 12 time points from gestational day (GD) 13 to postnatal day (PN) 60 by quantitative receptor autoradiography and in situ hybridization.Long Evans rat pups and fetuses of time pregnant rats were frozen and sectioned coronally and sagitally at i0 um. D 1 recepto~o~inding , determined with the selective antagonist ~JI-SCH 23982 was first noted on GD 18 in the developing striatum, basal ganglia and the olfactory tubercle, reaching adult levels at around PN 14.

The expression of D I receptor ~NA was studied by using a mixture of 3-~JS-labelled oligonu-specific cleotides. On GD 13 messages were noted in the developing stiatum, olfactory tubercle and retina. This study shows a good regional correlation between the development of receptor binding and mRNA, however, mRNA precedes detectable binding to the receptor by several days. Earlier work had shown that Q~ replicase forms complexes with Q~, plus strand RNA via interactions at two internal binding sites, the S-and the M-site, while binding of the 3'-end is mediated by host factor. In contrast, the 3'-end of the minus strand appeared to be directly accessible by replicase. We have prepared a series of plus and minus strand variant RNAs containing either internal deletions or terminal deletions extending from the 5'-end. The template activities of the variants were determined by single-round replication assays. For the plus strand we find that while deletion of the S-site remains without effect (as expected from previous results), deletion of the M-site reduces template activity to 25 % or less compared to wild-type RNA; the residual activity shows a decreased dependence on the presence of host factor. The results agree with an important role of the M-site interaction both with replicase and with host factor for the formation of productive initiation complexes. The template activity of the minus strand was unexpectedly found to be strongly dependent on the presence of a segment between nt 76 and 210 from the 5'-end. This shows that recognition of the minus strand by rep(icase does not only involve interactions near the 3'-end but requires a previously unknown structural feature near the 5'-end of this template.

$18-02

Karsten Graning and Angela Kr&mer D~partement de Biologie Cellulaire, Universite de Gen~ve, 30, quai Emest-Ansermet, 1211 Gen~ve 4 Splicing of nuclear pre-mRNAs takes place within multicomponent complexes (spliceosomes) that are assembled by interactions between the pre-mRNA, snRNPs and protein splicing factors. We have purified two splicing factors that function in the formation of the pre-splicing complex. SF3a consist of three subunits of 60, 66 and 120 kDa and corresponds to three proteins present in the functional 17S but not in the 12S U2 snRNP. It binds to the 12S U2 snRNP only in the presence of SF3b, suggesting a two step assembly pathway of 17S U2snRNP: binding of SF3b to the 12S U2 snRNP generates a 15S intermediate particle that is converted to the active 17S U2 snRNP by the subsequent association of SF3a. Comparison of proteins enriched in the SF3b fraction with polypeptides found in the purified 15S U2 snRNP suggests that SF3b comprises at least five of the other 17S U2 snRNP specific proteins and together with SF3a promotes the U2 snRNP/pre-mRNA interaction. By eDNA cloning, two of the subunits of SF3a have been identified as homologs of well characterized yeast proteins that function at the same stage of spliceosome assembly in yeast. Splicing factor SF1, a heat-stable 75-kD protein, functions early during spliceosome assembly. Tryptic peptides of SF1 were sequenced and cDNA libraries were screened with degenerate oligonucleotides. The information obtained by comparing the sequences of three overlapping cDNAs suggests the existence of at least three mRNAs that could be derived from a common pre-mRNA by alternative splicing. In agreement with this assumption mRNAs of the expected sizes that are differentially expressed depending on cell line or tissue are detected by Northern blotting. In theory, three polypeptides could be generated that differ by the presence or absence of 7 internal amino acids and in their Ctermini. The deduced amino acid sequence is rich in prolines and contains several possible phosphorylation sites and a putative leucine zipper. Proline-rich domains, which are also present in splicing factors PSF and SAP62, as well as leucine zippers have been found in transcription factors and could mediate protein/protein contacts, suggesting that SF1 could function during splicing by interaction with other spliceesomal proteins. Pre-mRNA splicing is catalyzed by a multicomponent complex (the spliceosome) that consists of small nuclear ribonucleoprotein particles (snRNPs) and non-snRNP protein factors. The spliceosome is assembled in a stepwise fashion on the pre-mRNA. Chromatographic fractionation of HeLa cell nuclear extracts and subsequent reconstitution of splicing in vitro has allowed the separation and isolation of snRNPs and several protein factors. SF4 triggers the transition between splicing complex B, which contains all spliceosomal snRNPs but is not competent for catalysis, and complex C, the active spliceosome. This transition involves a conformational change that leads to altered base pairing interactions between snRNAs and pre-mRNA. The purification of SF4 has been impeded by its low abundance; however, a correlation between SF4 activity and a polypeptide of 50 kD could be established in the most purified fractions. We are currently investigating whether this protein represents SF4. Interestingly an ATP-dependent RNA-helicase cofractionates with SF4 through several purification steps. Whether or not this activity is relevant for the splicing process is under investigation.

SELECTION OF IRON-RESPONSIVE ELEMENT RNAs WITH HIGH AFFINITY FOR IRON REGULATORY FACTOR Henderson, B.R., Menotti, E., Bonnard, C. , and Kith_n, L.C., ISREC, CH-I066 Epalinges Iron regulatory factor (IRF) post-transcriptionally regulates iron homeostasis via binding to mRNA iron-responsive elements (IREs). The IRE loop sequence (5'-CAGUGN-3') and "bulge" cytosine are phylogenetically conserved. We prepared a pool of 16,384 IRE molecules randomized at these 7 nucleotide positions, and employed in vitro selection to identify optimal sequences which bind human IRF. We define two major classes of high affinity RNA ligand, the optimal loop sequences of each are 5'-CAGUGN-3' (wild-type) and 5'-UAGUAN-3'. This novel finding predicts base-pairing within the loop between positions I and 5. All nucleotide substitutions in the "bulge" or at loop positions 2,3 and 4 decreased binding by 36% to 99%. In addition, binding specificity of rat IRF differed with that of the related rat IRE-binding protein, IRF B. In vitro analysis of IRE-like stem-loops identified by computer search

has not yet revealed any new IRE-containing genes.

A73 $18-08 3' Processing of histone pre-mRNAs: A phylogenetic comparison Reto Kohler, Petra Duda and Daniel Schiimperli. Zoologisehes Institut der Universtit/it Bern, Abteihing fiir Entwicldungsbiologie, Baltzerstr. 4, CH-3012 Bern, Switzerland. We are using a phylogenetic approach to study the biochemical reaction by which animal histone pm-mRNAs are processed at their 3' ends. In mammals, the efficiency and specificity of this reaction is known to depend absolutely on the U7 snRNA which interacts with conserved spacer sequences downstream of the proc~sing site. A highly conserved hairpin element which precedes the cleavage site serves as a binding site for an additional processing factor, but its importanea for efficient processing varies greatly between different historic genes and between extracts of different mammalian cell lines. To determine whether the relative importance of the hairpin and spacer dements have been conserved during evolution, we analysed the processing of histone genes from three different animal phyla in vitro, i.e. vertebrates (mouse), arthropods (Drosophila melanogaster) and nematodes (C. elegans). In the mouse system, processing was strongly dependent on the presence of a mouse spacer element. In nuclear extracts of Drosophila Ke cells, processing occurred exclusively when a Drosophila spacer element was present in the RNA. Processing efficiency was not reduced by foreign (mouse, C. elegans) hairpins. In whole cell extracts of Ascaris lumbricoides embryos, an inverted situation was observed. 3' processing products were only produced by RNAs that included an upstream segment derived from a C. elegans histone gene, irrespective of the spacer sequences present. These surprising results correlate with certain unique features in the C. elegans upstream element. We are currently in the process of cloning Ascaris lumbricoides histone genes, which will allow us to verify our results in a homologous system. In Xenopus laevls ooeytes, wild type mouse U7 RNA gets assembled into functional U7 snRNPs, both when transcr~ed from an injected gene or when injected as/n vitro synthesized RNA. If the Sm binding site of U7 RNA is converted into the canonical Sm sequence (U7 Sm OPT), the RNA assembles into a particle which accumulates more efficiently in the nucleus but wbAch is nonfunctional. This U7 Sm OFF particle inhibits the function of preformed endogenous U7 snRNPs, most likely by nonproductive binding to histone pre-mRNAs, Histone pre-mRNA processing can be demonstrated in c/s if U7 RNA is placed 3' of hlstone pre-mRNA and injected into oocytes. Only U7 Sm WT sequences are active in this c/s processing. Three proteins can be photoaffinity cross]inked to U7 Sm WT RNA, the common snRNP proteins G and B/B' and a U7-specific protein of 40 kD (50 kD in mouse). These crosstinks map to closely spaced positions within the Sm binding site with the U7-specific crosslink being located most 3'. U7 Sm OPT RNA does not become crosslinked to the U7specific protein and is also more tightly associated with Sm proteins. We now investigate the in vitro assembly of U7 snRNPs using these characterized U7 RNAs and a preparation of highly enriched snRNP proteins. U7 Sm WT and U7 Sm OPT RNAs associate with common Sm proteins in this system, but the interaction with the U7-speciflc protein could not be demonstrated. The double-stranded (ds)RNA modifying enzyme, which can convert adenine to inosine, was first identified in Xenopus laevis, but has since then been detected in different species and in mammalian cells. The enzyme is a specific adenine deaminase which uses dsRNA as substrate and recently, it has been postulated to be responsible for a specific mRNA editing reaction. We have purified this enzyme to homogeneity, approximately 400,000 fold from calf thymus whole cell extracts. The protein was purified primarily by ion exchange chromatography over six columns, with the final step being chromatography on a dsRNA affinity column. The purified protein has a molecular weight of 115 kD and is the only protein present when enzymatically active fractions are visualized on an SDS-polyacrylamide gel stained with silver. The dsRNA modifying enzyme is a very low abundant protein. We are in the process of further characterizing the purified enzyme and of cloning cDNAs coding for it.

Detailed mutational analysis of histone RNA 3' end formation Carmen Spycher, Andr6 Furger, Adrian Streit, Daniel Albrecht, D. Schiimperli, Zoologlsehes Iustitut der Universtit/it Bern, Abteilung thr Entwicklungsbiologie. Baltzerstr. 4, CH-3012 Bern, Switzerland Histone RNA 3' ends are formed by cndonucleolytic cleavage resulting in poly(A)" mRNAs with a highly conserved 3'-terminal hairpin loop structure. For the mouse H4-12 gene major and minor processing sites are located 3 and 5 nucleotides downstream of the hairpin, respectively. For the cleavage reaction, the U7 snRNP interacts with the spacer element of histone pre-mRNA by base-pairing through the 5' end of its RNA moiety, U7 RNA. We have observed that this base-pairing potential extends further than previously recognised in either direction. We have made site-directed rnutagenesis in fltis potential hybrid region and around the processing site. RNA processing and complex formation with the U7 SURNP were analysed by incubation in nuclear extract from mouse K 21 cells. Our results indicate that base-pairing with U7 RNA both in the classical spacer element and downstream of it are very important for histone RNA processing. In contrast, sequences upstream of the spacer element do not seem to be involved in base-pairing with U7 RNA, but mutations in this region may affect processing by other mechanisms. Systematic mutations of the highly conserved four nucleofides immediately following the hairpin show no qualitative or quantitative effects in the processing reaction. Alteration of nucleotides 5 to 7 around the minor processing site, however, result in a dramatic decrease in processing efficiency and alter the specificity of the cleavage site. In further experiments we will introduce specific nuclcetidc modifications at the two processing sites, which should allow us to characterize potential cleavage factors by chemical crosslinking. EVOLUTION Callaerts P., Glardon S., J.,oosli F., Quiring R. and Gehring W. Cell Biology, Biozentmm, Basel. The Pax genes encode a family of DNA binding factors which share the paired-box and play an important role in the control of development. They can be subdivided into four different classes which differ in the presence or absence of other conserved sequence motifs coding for a paired type homeobox or an octapeptide. The Pax-6 gene, which contains a paired-box and a homeobox, has been cloned from humans, mice, rats and zebra fish. Recently, we have cloned the Pax-6 homologue of Drosophila melanogaster. The five genes share a very high degree of sequence identity both in the paired-domain and the homeodomain.

Similar to its mammalian counterparts, the Drosophila gene is expressed in the eye primordia and the nervous system. In addition, mutations in Pax-6 affect eye development: Aniridia in humans, Small eye in mouse and rat, and probably eyeless in Drosophila. This would imply that the Pax-6 homologues share a conserved function in eye morphogeuesis in both vertebrates and invertebrates. Furthermore, this suggests that Pax-6 may have been present in a common ancestor. Therefore, we are now trying to isolate Pax-6 homologues from other, more primitive organisms by means of the polymerase chain reaction, and to determine whether they are implicated in eye development. Putative candidates have been identified from a number of species and are being characterized. Their sequence similarity and some evolutionary implications will be discussed.

Keegan Liam and Gehrin~ Walter J., Biozentrum, University of Basel, Klingelbergstr. 70, CH-4056 Basel, Switzerland

Using an enhancer map screen we have identified two genes that may be direct targets of Antennapedia involved in forming the adult leg. spalt major is repressed in the leg imaginal disc and tea, shirt is activated by Antennapedia. We wish to determine whether the control of these genes by Antennapedia is direct, spalt major is expressed in a ring in the antennal disc and is repressed by Antennapedia in the leg disc. The antennal ring is repressed by ectopie Antennapedia expression that transforms the antenna to a leg. We are defining an enhancer at spalt major that is required for antennal ring expression and repression by Antennapedia. We are using various approaches to show that Antennapedia binds to the antennal ring enhancer at spalt major. These include polytene chromosome banding studies using antibodies to Antennapedia as well as in vitro DNA-binding and genetic experiments. In search for novel regulators potentially involved in myogenesis, we identified a homeobox of the paired type in human muscle. Subsequent cDNA cloning revealed that we had cloned the full length coding region of human PAX7. Our human sequence extends the known mouse eDNA both on the 5' and the 3' end. As a first step in order to test for a functional role of PAX7 in myogenesis, we analyzed its expression pattern during chicken development. Transcripts are present in the neural tube and in the dermamyotome of the developing somites. Therefore, the pattern of expression of Pax7 is conserved between mouse and chicken. Next, we assessed the expression of Pax7 in different mouse and human cell lines. We found Pax7 transcripts specifically in myogenic cells and not in any other cell types. Interestingly, Pax7 is already present at the myoblast stage. Moreover, 10TI/2 fibroblasts converted to myoblasts by either transfection of myoD or treatment with 5-azacytidine expressed Pax7, whereas the parental cells did not. While myoD was able to activate Pax7 in 10T1/2 cells, expression of Pax7 was not sufficient to induce the myogenic phenotype. Nevertheless, our expression data are consistent with a role of Pax7 during the mesodermal commitment to the myogenic lineage. The segmental organization of the Drosophila head is achieved by a flow of positional information from maternal gene products to the zygotic gap genes. Recent genetic analysis has identified three new gap genes involved in head development. One of these genes is the empty spiracles (ems) gene. Mutations in eras cause severe defects in the head and the Filzk~rper at the posterior end are missing. The eros gene encodes a putative transcription factor with a homeodomain. The eros RNA is expressed in two phases of embryonic development. First expression is seen at the blastoderm stage in a single anterior band, correlating with its function as an anterior gap gene. Later during embryogenesis eros is expressed in the posterior spiracles as well as lateral regions of each segment where the tracheal pits form and lateral neuroblasts originate. Using g-gal fusions we could identify at least five different regulatory elements in the eros promotor region responsible for tissue specific expression of the gene. A 300 bp element responsible for the early expression which is dependent on the maternal gene bicoid, the key gene of the anterior system, and the terminal gene tailless was identified and studied in detail. This element contains two bicoid and two tailless binding sites. The effects of muations in these binding sites on eros expression will be discussed.This analysis should allow us to elucidate the molecular mechanisms by which the morphogen bicoid regulates subordinate target genes like ems in the Drosophila embryo. The Pax-6 gene of the mouse encodes a transcription factor with both a paired-and a homeodomain. Pax-6 is affected by several allelic mutations in the Small eye (Sey) gene. Sey mutations cause a reduction of the eyes in heterozygotes, and homozygotes lack both eyes and nasal openings and die as newborns. The corresponding mutation Aniridia in humans affects eye and cranlofacial development in a similar manner. We have isolated a Pax-6 homolog from Drosophila (DPax-6), which shares more than 90% sequence identity in the paired-and the homeodomain with the mammalian genes. Also outside of these domains we find considerable sequence conservation arguing for a true Pax-6 orthologue. The Drosophila gene spans ~16 kb and is expressed in the brain and the ventral nervous system of the embryo, and in the eye imagJnal discs and the brain of the larva. The gene was mapped to the eyeless (ey) locus.

Mutations at the ey locus cause either the partial or complete absence of the compound eyes or embryonic lethality. In one ey allele we have identified a transposable element in the first intron of DPax-6, which affects DPax-6 transcription in the eye discs, suggesting that DPax-6 is encoded by the ey gene. This implies that Pax-6 (Sey) in the mouse is homologous to ey in Drosophila. Thus, despite of the different modes of eye morphogenesis, the same gene seems to control eye development in insects and vertebrates. The Drosophila homolog of the vertebrate serum response factor (SRF) was isolated by low stringency-hybridization. Nucleotide sequence analysis revealed that the Drosophila SRF homolog (DSRF) codes for a protein which displays 93% of sequence identity with human SRF in the MADS domain, the region required for DNA binding, dimerization and interaction with accessory factors. The DSRF gene is expressed during several phases of embryonic development. Both the RNA and the protein are maternally provided to the egg and slowly decrease in their levels during gastrulation. After germ band retraction, high levels of zygotic expression were observed in a distinct subset of peripheral tracheal cells distributed throughout the embryo. Many of these cells are at the tip of tracheal branches and are in direct contact with the target tissues these branches tracheate. The DSRF gene was mapped to position 60C on the second chromosome, and overlapping deficiencies which remove the gene were identified. Analysis of tracheal development in embryos carrying these deletions revealed a degeneration of most of the major branches of the tracheal system. Although the initial migration of tracheal cells was not affected in those deficiency embryos, many tracheal cells appeared not to maintain their formerly correct position and continued to migrate. Thus, the DSRF gene might play a role in the proper formation and maintenance of the major branches of the trachea.

S 19-08 Our experiments would be expected to provide insight into such problems as the evolution and function of transcription factors with multiple DNA binding domains. Ideally, we would be able to mimic "gene splitting" which occurred during evolution. The poan gene plays an important role during Drosophila neurogenesis.

It is expressed in specific subsets of sensory mother ceils (SMCs) and

neuroblasts. The SMCs that express poxn differentiate into polyinnervated external sense (p-es) organs. In the absence ofpoxn, SMCs differentiate into mono-innervated rather than poly-innervated external sense organs. As the poxn gene contains a paired-box, it probably encodes a transcription factor. Since the function ofpoxn in the central nervous system is not known and no po~n mutants are available, an EMS mutagenesis screen was initiated to isolate such mutants.

Based on the assumption thatpoxn null alleles are lethal, we screened for lethals uncovered by the deficiency Df(2R)WMG, which includes poxn, but survive over the deficiencies Df(2R)XTE-18 and Df(2R)KL-32 flankingpoxn. In a further test, such lethals are screened for the absence of embryonic p-es organs.

Ifpoxn mutants are not lethal, they will escape this screen. Therefore, we are inducing local hopping of P-elements that have inserted in the vicinity of poxn. The offspring of flies that display an altered eye phenotype are screened for P-element insertion into poxn.

We hope that these approaches will generate mutants that will be crucial for future studies ofpoxn function in neurogenesis.

Chromatin structures and transcription of rDNA in yeast SaccharomyCes eerevisiae Reinhard Dammaun, Renzo Lucchini, Thee Keller and Jest M. Sogo; Institute of Cell Biology, ETH-HOnggerberg, CH-8093 Ziirich

The chromatin structure of yeast ribosomal DNA was analyzed in rive by cross-linking intact cells with psoralen. We found that in exponentially growing cultures the regions coding for the 35S rRNA precursor fall into two distinct classes. One class was highly accessible to psoralen and associated with nascent RNAs, characteristic for transcriptionally active rRNA genes devoid of nucleosomes, whereas the other class showed a cross-linking pattern indistinguishable from that of bulk chromatin and was interpreted to represent the inactive rRNA gene copies. By cross-linking the same strain growing in complex or minimal medium, we have shown that yeast ceils can modulate the proportion of active (non-nueleosomal) and inactive (nucleosomal) rRNA gene copies in response to variations in environmental conditions which suggests that yeast can regulate rRNA synthesis by varying the number of active gene copies, in contrast to the vertebrate ceils studied so far. Whereas intergenie spacers flanking inactive rRNA gene copies are packaged in a regular uucleosomal array, spacers flanking active genes show an unusual cross-linking pattern suggesting a complex interaction of regulatory factors and histories with DNA.

$20-04

R. Felleisen & B. Gottstein; Institute of Parasitology, University Bern, Bern, Switzerland Most patients suffering from alveolar echinococcosis (AE) respond to infection with a marked IgG synthesis directed against E.multilocularis metacestode antigen II/3, which represents a novel member of the family of cytovillin related proteins. Although the respective gene basically is also present and expressed in E.granulosus, most of the cystic echinococcosis (CE) patients do not recognize the antigen. This phenomenon was tackled by generating cDNA derived from full length II/3 genes from both Echinococcus species, performed by RT-PCR. Sequence analysis revealed a very high degree of conservation of the primary sequence (98.6% homology), cDNA fragments were expressed as recombinant proteins and were comparatively assessed in ELISA respective to antibody-binding characteristics. Sera from patients suffering from CE were showing no significant differences in reactivity with the antigens derived from both species. Therefore, parameters others than minor differences in the primary sequence seem to be responsible for the lack of antigen recognition with respect to CE.

$20-02

Patrick Rigoni, Sandro Rusconi, Institut f Molekularbiologie H der Universitat Zarich, Winterthurerstr. 190, 8057 Z~rich, Switzerland. Trinuclcotide repeats are often found in the coding sequence of transcriptional regulators in both mammals and yeast, However, their function is yet unclear and data collected in our laboratory on the gheocorticoid receptor even suggest that they may not have any, and that their presence is probably the result of a selfish replicative behavior. Nonetheless, we believe that these motifs can be used as tags to identify flexible protein domains typical of modular proteins such as transcription factors. Another kind or repeat (coding for a poly-glutamine/alanine) has been discovered in the yeast regulators GALl1 and SSN6. From Northern blots, we know that such CAGGCN repeats are present also in higher eukaryotes and we want to isolate them by constructing an enriched eDNA library using a 5' RACE protocol. So far, no mammalian protein bearing the motif GlrdAla has been characterized. Among the positive clones we hope to find the homologous or paralogous of GALl1 and SSN6 that have escaped so far conventional cloning techniques. Meanwhile, we have been testing the effect of coexpressed yeast GALl 1 on different reporter-activator combinations in mammalian cells. Preliminary results show that Galll but not the mutant form GalllP can inhibit the activation properties of some (not all) Gal4 chimeras. We are also producing antibodies directed against oligo-Ala and oligo-Gln and oligo-Gln/Ala motifs in order to better characterize natural proteins containing these repeats. We are interested in the early development of C. elegans, particularly in the contribution of genes whose homologues in vertebrates and D. melanogaster mediate positional infolzzations during development. Therefore, we are working on ceh-13 which belongs to a cluster of at least 4 homeobox containing genes. This cluster is considered to be equivalent to the homeotie cluster of Drosophila and to the Hox clusters in vertebrates. By generating ~-galactosidase transgenie lines, we have observed that ceh-13 is expressed very early during embryogenesis, Embryonic expression appears to be restricted to the posterior half of the embryo, which is rather surprising, since the ceh-13 hemologues in Drosophila and vertebrates are anterior-specific. During larval stages, ceh-13 is expressed in neuronal cells. These results are now in the process of being confirmed by immunolocalization of the ceh-13 protein product with a polyclonal antiserum. In order to clarify the function of ceh-13 in the C. elegans development, we have isolated a ceh-13 mutant from a Tcl insertion library, in collaboration with R. H. A. Plasterk from Amsterdam. From this worm, which looks WT, we are now trying to obtain a deletion derivative. Previous studies have shown that Bovine herpesvirus 1 (SHV-1) infected cell pratein (BICP) 0 acts -depending on the promoter -as a strong transactivator or as a repressor in transient expression assays. The 81CP0 polypeptide contains a cysteine-rich zinc finger domain (C3HC4) which is conserved in a number of viral and cellular regulatory proteins including the 8ICP0 homologs ICR3 of herpes simplex virus type 1 and protein 61 of varicella zoster virus. This type of zinc finger (the so-called RING tinge0 was shown to bind zinc ions but functional requirement for zinc has not yet been demonstrated, a gap which we aimed to fill with the present study. Transient expression assays were performed in oocytes which had been microinjected with thionein to chelate and deplete the intracellular pool of zinc. 81CP0-induced CAT activity of a promoter-CAT construct was 72-fold higher than the basal CAT activity of the reporter plasmid, In the presence of thionein, BICP0-induced transaotivatlon was reduced 54old. With a set of control experiments we excluded that thionein might affect transcription and protein synthesis in general, To our knowledge this is the first demonstration of zinc-dependence for a member of the RING finger family.

We have identified by PCR and eDNA cloning five POU genes expressed during early embryogenesis of zebrafislx Four of these genes show extended homology to the brn-1 class of POU genes. Preliminary evidence suggests that the brn-1like POU genes overlap with most of their expression domains.The gene studied in most detail so far (ZP-50) is first expressed on the ventral side of the future fore-and midbrain. Slightly later, expression is also found in rhombomeres 3 and 5 in the hindbraln. These rbombomeres were identified by the specific expression of the Krox-20 marker using a novel in situ hybridization protocol which allows the simultaneous detection of two different transcripts using different color substrates. In the 24 hours old embryo a complex expression pattern is found involving a variety of brain structures and the spinal cord. The distribution of the transcript suggests that ZP-50 is mostly expressed in glia cells. Another POU gene we have identified (GP-9) defines a novel class of POU proteins. As a maternal mP, NA it is initially ubiquitously present. After gastrnlation its transcript is found most notably in the developing hindbrain in rhombomeres 2 and 4 for which so far no good molecular markers were known. We are investigating the roles the identified POU genes may have in zebrafish hindbrain segmentation. We are currently attempting to manipulate gene expression in developing embryos by injection of RNA or by the production of transgenic fish. We are also screening the zebrafish genome for new developmental marker genes. Progress about these experiments will be presented.

$20-07 Willimann, T. and Trueb, B. To gain insight into the regulatory mechanisms of collagen VI synthesis we have characterized the cis-acting elements of the chicken ctl(Vl) collagen promoter. Footprinting experiments with nuclear extracts from chicken embryos revealed three distinct elements, designated A, B and C, that were protected from DNase I digestion. The nuclear proteins that interact with the three sites were identified by gel retardation assays in combination with the use of various oligonucleotide competitiors as well as specific antibodies raised against well-characterized transcription factors. Site A was found to be a target for transcriptional activator APf, whereas sites B and C were shown to be recognized each by two distinct nuclear proteins which belong to the Spl multigene family. To address the question whether the three sites alone are able to direct transcription, a minipromoter construct was created in which the sequences of sites A, B and C were placed in front of a reporter gene. After transfection into chicken fibroblasts, this construct exhibited a high relative promoter activity when compared to a large genomic fragment containing the basic ctl(VI) collagen promoter. Thus, the three sites are sufficient to induce transcription of this gene. Octamer transcription factors (Oct or OTF) are a subset of the POU family of transcription factors which regulate expression of cellular and viral genes by binding to the octamer sequence ATGCAAAT. Neurons and astroglial cells harbour, in addition to the ubiquitous Oct-I factor, brain specific factors termed N-Oct 2, 3, 4 and 5. In the present study we determined the chromosomal localization of the gene encoding the N-Oct 3 transcription factor and characterized the structure of isolated N-Oct 3 genomie clones. The chromosomal mapping was performed by analysis of somatic cell hybrid panels and radioactive and fluorescence in situ hybridization of human metaphase chromosomes. It is interesting that the 5' end of the N-Oct 3 coding sequence contains repetitive CAG and GGC residues. Several disorders have been discovered to be related to expansion of trinucleotide repeats. We are presently investigating if the N-Oct 3 CAG or GGC clusters are hot spots for such mutation mechanisms and if so, which diseases could be associated with it. Two DNA regions upstream of the distal glucocorticoid receptor binding site interact with nuclear proteins that are tissue-specific (region a) or ubiquitous (region c). The characteristics of the DNA-protein interactions have been studied in vitro. These sequences, in different combinations with the natural MMTV promoter, were tested in transfection assays of fibrcblast or mammary cell lines. We show that they are able to modulate the level of glucocorticoid-stimulated transcription. The proximal region of the MMTV promoter has binding sites for the transcription factors CTF/NF-I and oct-1. P~asmids with a deletion or mutations in the oct-1 site were tested in stable transfections, that are most representative of the state of proviral DNA with respect to both number of integrated DNA templates and chromatin organization. We show that the oct-1 site is important for the basal level of promoter activity. The data further indicate that the functional outcome may depend beth on the relative ratio of factors to DNA templates and on the relative affinity of binding sites, as determined by oligonucleotide competition footprinting. Besides the fact that telomeres represent specialised structures important for chromosome stability, the particular significance of cloned C. elegans telomeres is that they will contribute to the completion of the physical mapping of chromosomes and that they provide one set of elements for the construction of artificial C. elegans chromosomes. The cloning of C. elegans telomeres, however, is impeded by the fact that tandem repeats of the sequence TTAGGC are not only located at the ends of the C. elegans chromosomes, but also at many internal chromosomal sites. Therefore, we have developed a special protocol for the construction of a C. elegans endlibrary in 2~ Zap. The resulting telomeric library was screened for recombinants containing TTAGGC repeats and yielded about 70 positive clones. So far, 30 were analysed by partial sequencing and twelve of them satisfy all criteria required for telomeric clones. Their TTAGGC tandem repeats are located at the expected position, namely just adjacent to the blunt end cloning site in the polylinker. Furthermore, the G-rich strand of the repeats is oriented 5' to 3' towards the end of the fragment, thus corresponding exactly to the defined orientation of eukaryotic telomeric sequences. Finally, hybridisation data with one of these putative telomeric clones show that a subtelomeric fragment hybridises to Bal31-sensitive fragments on a Southern blot with C. elegans DNA, indicating that this clone represents a C. elegans telomere.

$20-09 To investigate whether chromatin structure or transcription can interfere with replication, derivatives of the yeast TRP1ARS1 minichromosome were constructed that contain either the DED1 or PET56 promoter firing against the origin of replication ARS1. While the PET56 constructs transformed yeast at high frequencies and were maintained as high copy number minichromosomes, the DED1 constructs transformed at low frequency and the constructs were integrated in the genome suggesting that ARS function was impaired. Insertion of the H4-ARS, a second origin of replication, rescued a DED1 construct as a minichromosome (YRpDH3).The chromatin structure mapped at low and high resolution of the ARS1 region in YRpDH3 was indistinguishable to that of the PET56construct YRpFT61. Ananlysis of RNA showed that transcription was going through ARS1 in YRpDH3 and in YRpFT61, but the levels of transcripts in YRpDH3 were much higher. We conclude that transcription through ARe1 interferes with replication and prevents extrachromosomal maintenance.

DNA SEQUENCE OF A REPEATED DNA SEGMENT ON CIRCULAR AND LINEAR PLASMIDS OF BORRELIA BURGDORFERI W.R. Z%ckert and J. Meyer, Abt. PZMOM, Zahn&rztliches Institut der Universit~t Basel A plasmid-associated repeated DNA segment of B. burgdorferi strain B31 was cloned. At least two copies of this segment appear to be present on the 29 kb circular plasmid, whereas one copy is carried on the 50 kb linear plasntid of this strain. DNA sequence analysis revealed a region containing a novel 906 bp putative open reading frame (orfl) on the 50 kb linear plasmid, orfl displayed extensive sequence homology (95%) to putative open reading frames present on the clones obtained from the 29 kb circular plasmid. Heterogeneity is mainly caused by 3rd-base-wobbllng. Flanking sequences share 85-95% homology, including the 5' end of an additional putative open reading frame irmaediately downstream of orfl. Whether orfl and/or the related sequences are being transcribed and yield, in the case of orfl, an approximately 36.5 kD, lysine-rich polypeptide, is yet unknown. The repeated sequence seems to be specific for B. burgdorferi sensu lato and therefore may be useful in nucleic acidbased diagnostics of Lyme disease.

$20-16 Similar to type IB oculocutaneous albinism in humans, overall production of pigment is greatly reduced in dark eyed albino and only obvious in eyes. We have studied the molecular basis of the c 44H mutation and show that expression of the tyrosinese gene is not affected. After sequencing tyrosinase cDNA isolated from c44H/c44H homozygotes we uncovered a single base alteration from wild type leading to a serine to isoleucine exchange. The importance of this mutation was demonstrated by generating transgenic mice containing a mutated tyrosinase minigene. This showed that the single base change is sufficient to severely depress pigment production in transgenic mice. We therefore conclude that the point mutation is responsible and sufficient to generate the dark eyed albino phenotype. Members of the POU-family of transcription factors are involved in developmental and differentiation processes. Using a PCR-based cloning strategy with degenerated oligonucleotides we identified several POUgenes from the lactating and involuting mouse mammary gland. Three of these cDNA clones which contained as yet unknown sequences were chosen for further investigations: clone 5.80 which has a high homology to Pit-l, clone 5.54 which is related to the Brn-3 gene and clone 5.48 which belongs to the class III of POU-proteins. The expression levels were analyzed in different mouse organs and in several developmental stages of the mammary gland, including the postlactational process of involution. Because of the low abundance of the specific transcripts we used a semiquantitative, reverse transcriptase-mediated PCR assay to investigate the mRNA levels of the three novel POU-family members. Distinct and specific expression patterns for all three members were obtained in the different investigated organs. Interestingly, the expression of the Pit-1 related cDNA clone 5,80 was elevated in all developmental stages of the lactogenic-hormone dependent mouse mammary gland and in sceletal muscle, whereas its expression was low in other organs. Repetitive sequences in DNA can allow ectopic (outof-register) homologous recombination, which is often undesirable and can even lead to disease in humans. To prevent this, long homologies should often be interrupted during evolution. Accelerated mutation of repetitive sequences by so-called ripping may be one of the mechanisms used to accomplish this in fungi. We are investigating if introns in vertebrate genes have an undiscovered role as interrupters of homology within and/or between genes, in addition to their established role in exon shuffling. By inserting homology-poor introns in an otherwise homology-rich region the genome could prevent ectopic homologous recombination. Such a mechanism could be especially useful where there is an advantage in encoding highly repetitive protein sequences. It appears that within the collagen gene family there is indeed a correlation between repetitiveness and intron density. The influence of prenatal diazepam exposume (i,25 mg/kg/d, s.c.) from gestational day 14 to 20 on the development of the ~-opioid binding sites in striaturn, nucleus accumbens and midbrain of PN 14, PN 28 and 8 week old male and female rats was studied. 8 week old prenatally diazepam exposed male rats showed a significant decrease of K-opioid binding sites in the nucleus accumbens. A sex-dependent difference in K-opioid binding site densities could also be detected between PN 28 prenatally diazepam exposed male and female rats. We now investigate by in situ hybridisation whether changes in the level of mRNA encoding fo~ prodynorphin, the precttrsor for various endogenous ~-opioid agonists, could be responsible for the decrease of ~-opioid binding sites in the nucleus accumbens of 8 week old prenatally diazepam exposed male ~ats.

$21-04 DISTRIBUTION AND MORPHOLOGY OF NITRIC OXIDE-POSITIVE NEURONS IN THE MARMOSET CEREBRAL CORTEX DURING PRE-AND POSTNATAL DEVELOPMENT J.P. Hornung* and J. Schultz** *Institute of Anatomy, University of Lausanne, 1005 Lausanne, and **University of Fdbourg, 1700 Fribourg Nitric oxide, synthesized by the enzyme nitdc oxide synthase (NOS), is a neuroactive substance and a limited number of neurons were shown to express this enzyme either by histochemistry or by immunocytochemistry. Both techniques were performed on brains of marmosets with ages ranging from 6 weeks before birth to adult. The morphology and distribution of NOS-positive neurons were analyzed in all lobes of the cerebral cortex. The same pattern was revealed by histochemistry or immunocytochemistry. Three populations of NOS-positive neurons were revealed. First, a transient population of neurons in the molecular layer of the embryonic cortex, many having the morphology of the RetzJus-Cajal cells. A second population was made of large multipolar neurons in the deep layers of the cortex and the underlying white matter, which persisted from embryonic to adult life. A third, permanent, population was made of small, weakly reactive neurons in the supragranular layers appearing postnatally, This study demonstrates that NO can exert its actions in the cerebral cortex through several pathways at different developmental stages. Trimethyltin (TMT), a well-known neurotoxicant, was used to study the sensitivity of the different brain cell types (i.e. neurons, astrocytes, oligodendrocytes and microglial cells) to a neurotoxic insult. Aggregating brain cell cultures of fetal rat telencephalon were treated during 10 days with TMT at different concentrations, ranging from 10 -10 M to 10 -6 M. Microglia were found to be the most sensitive cell type, since already at 10-10M of TMT an increased number and clustering of Griffonia simplicifolia-positive ceils could be observed. At 10 -8 M of TMT astrocytes showed increased staining for glial fibrillary acidic protein, characteristic of gliosis. Neurons and oligodendrocytes appeared to be the least sensitive cells, since a decrease in the activity of cell type-specific enzymes was observed only at 10 -6 M of TMT. These results show that microglial cells and astrocytes respond to a toxic insult before any neuronal changes could be detected, and that the microglial reaction may be the first target of TMT. This reaction could then trigger the astrocytic response. VIF is widely distributed in brain with suggested functions related to development. VIP expression was studied during rat brain development using hybridization histochemistry with a 48met probe. VIP mRNA was found in thalamic nuclei on El7, later recognized as the ventrolateral and reticular nuclei after further maturation during prenatal period, vIP mRNA was found in hypothalamus, especially suprachiasmatic nucleus on El9 and its expression matured over the next days. Few cortical neurons contained VIP mRNA on E21, they continuously increased in number and signal intensity over the perinatal period. VIP gradually matured in the first three postnatal weeks and adult-like patterns were found on P 22, when cerebral cortex, ventrolateral and reticular thalamic nuclei, hypothalamus, esp. suprachiasmatic nucleus, were the regions with most prominent VIP expression. These results demonstrate the relatively late appearance of VIP gene expression in the rat forebrain, as compared with peptides like SRIF and CCK, not suggesting a major role in earlv brain maturation. In primary cultures of mouse cerebral cortical astrocytes, a rapid glycogenolysis followed by a massive glycogen resynthesis ( six-to ten-fold over basal levels after 9 hr) are induced by vasoactive intestinal peptide (VIP) or noradrenaline (NA). Both actions of these neurotransmitters are mediated by cAMP. Since the induction of glycogen resynthesis triggered by VIP or NA is abolished by inhibition of transcription and translation, we applied the 2-D gel electrophoresis technique to search for astrocytic proteins induced by VIP or NA. The comparison of 35S-labeled proteins from primary astrocyte cultures treated or not with VIP 1 ~tM revealed two newly synthesized proteins: a 36 kDa protein (pI=5) and a 23/24 kDa protein (pI=6). Only the latter is visible on a silver stained 2-D gel, which is a prerequisite to consider its purification and microsequencing. Induction of the 23/24 kDa protein by VIP was time-dependent with a maximum at -12 hr. Preliminary results indicate a similar induction by NA and isoproterenol. In humans, the central analgesic effect of tramadol (T) is only weakly reversed by the ~t opioid antagonist naloxone (Nx) (Br J Clin Pharmacol 1993; 35:73P). T analgesia may thus result from an action on monoaminergic pathways as suggested by in vitro and animal data. We therefore investigated the effect of ct2-adrenoeeptor antagonism by yohimbine (Y) on T analgesia. According to a randomised double-blind crossover and placebo (P) controlled design, healthy volunteers (n=10) received T (100mg orally), followed (+ 3 h) by Y (0.1mg/kg intravenously), and Y + Nx (0.8mg intravenously). Analgesia was assessed over 8 h by subjective pain threshold (numerical scale -NS-) and objective pain threshold (Rill nociceptive reflex -RIII-) monitoring. Peak analgesic effect was observed at ca. 3.25 h (RIII +8.8 +SEMI.6 mA and NS +8.4 +1.8) vs P (RIH -0.3 +1.2 and NS +2.3 +1.6, P<0.05) and lasted ca. 6 h. Y reversed T analgesic effects during ca. 1 h (R/II -2.9 +1.8, P<0.05 and NS +2.8 +1.5, P<0.05), whereas Y + Nx abolished T effects thoughout the study period (Rill -1.9 4-1.1 and NS +0.6 +1.3, P<0.05), Y alone tended to lower pain thresholds (Rill -4.1 +1.5 and NS -1.9 +1.4, P>0.05 ANOVA). Yohimbine, an r antagonist, reverses tramadol effects, thus pointing to the major role of monoaminergic modulation in tramadol anfinociception. A monoclonal antibody (IN-l)was raised against neurite growth inhibitory proteins present in rat CNS myelin (Caroniand Schwab, Neuron 1: 85, 1988) . IN-1 neutralizes the inhibitory ettect of CNS tissue in vitro and in vivo, thus permitting regeneration of certain lesioned CNS fiber systems. We have used this antibody to immunostain cryostat sections of the nervous system of the rat. We found that IN-1 stains white matter and myehnated fiber tracts in the CNS. The staining pattern corresponds to that of the known myelin specific antigens MBP (myelin basic protein), MAG (myelin associated glycoprotein) and MOG (myelin oligodendrocyte glycoprotein). The staining of IN-1 in the CNS is found in regions of the adult nervous system which have been shown to be inhibitory for axonal growth. Interestingly, the regions which show a high abundance of IN-1 antigens are low in staining for GAP-43, a marker for growth and plasticity in the developing and adult brain. In contrast, GAP-43 is strongly expressed in unmyelinated fiber tracts or nuclei. Prevention of oligodendrocyte development in rat spinal cord resulted in suppression of IN-1 expression and elevated GAP-43 levels. This suggests that inhibitory myelin proteins may negatively influence plasticity in the rat CNS. To date, the existence of ANP and NPY in the adrenal medulla has been shown only for a few mammalian and anuran species. Thus the present immunohistochemical investigation attempts to localize ANP-like and NPY-like peptides n the adrena gland of representative mammals, birds, reptiles, amphibia and bony fish by the use of antisera specific for mammalian ANP and NPY. Furthermore, the catecholamine-containing cells were characterized by antisera against enzymes of catecholamine synthesis. ANP-and NPY-immunoreactivies were detected in adrenal chromaffin cells of all vertebrates studied while no immunoreactivities were observed in the adrenal cortex or in its homolog, the interrenal. The representatives of the different vertebrate classes exhibited marked differences in the proportions of ANP-immunoreactive and NPY-immunoreactive cells, in the presence of ANP-and NPY-immunoreactivities in noradrenalineand/or adrenaline-containing cells and in the amount of of cells containing both ANP-and NPY-immunoreactivities. Our results give evidence for a long phylogenetie history of adrenal ANP-and NPY-like peptides. Thus, they stress the physiological impact of these peptides as adrenal paracrine and/or endocrine hormones. Human neuronal growth inhibitory factor (GIE) is involved in the regulation of neuronal growth and is deficient in the brains of Alzheimer's disease victims. This brain-specific metalloprotein consists of 68 amino acids out of which 20 are Cys and has been found to contain copper and zinc. Proteins with similar primary structure were isolated from bovine and equine brain. The amino acid sequence of these proteins shows about 70 % homology to those of mammalian metallothioneins (MT), with two conserved inserts of 1 Thr and a glutamic acid rich hexapeptide in the N-and C-terminal regions, respectively. In contrast to MTs, which usually contain only Zn(I1), the presence of Cu(I) and Zn(II) (6-7 moles total) in all GIF isolations suggests the functional importance of both metals ions. To spectroscopically probe the native Zn-binding sites, the Zn(II) ions were replaced by Cd(II). Both bovine and equine Cd/Cu-GIF derivatives exhibit similar absorption and CD spectra with features characteristic of Cd-thiolate clusters. The release of 3H-arachidonic acid evoked by glutamate was characterized in cerebral cortical neurons in primary culture grown under conditions that prevent glial cell proliferation. In these cultures, 99% of the total ceils are immunocytochemicaUy characterized as being positive for specific neuronal markers such as neuron-specific enolase and neurofilament. Specific markers for astrocytes, oligodendrocytes or microglia were not detected. Glutamate induces a concentration-dependent release of 3H-arachidonic acid with a EC50 of 3 pM. The pharmacological profile of this glutamate response shows the involvement of two receptor types: the NMDA and the AMPA ionotropic receptors. The glutamate-evoked release of 3H-arachidonic acid is pHsensitive. When the incubation buffer is alkalinized from 7.25 to 7.65, both basal and glutamate evoked release of 3H-arachidonic acid are enhanced. The glutamate response is also enhanced as intracellular pH is alkalinized by pharmacological treatments: such as inhibition of C1-/HCO3-antiport by DIDS. These results further stress the role of intracellular pH in glutamate-mediated neurotransmission.

$21-09

Multiceli culture system for the study of drug kinetics Wechsler D., *Schittny J.C. and Honegger U. E., Depts. of Pharmacology and *Anatomy, University of Bern, CH-3010 Bern A cell culture system was developed for the investigation of cellular drug kinetics in cultures with up to 4 cell types. It was used for the study of uptake, competitive uptake, release and redistribution of drugs. Individual cell types (human skin fibroblasts, rat astrocytoma cells (C6) and rat hybfidoma ceils (oligodendrocyte x astrocytoma, ROC)) were separately grown to confluency on glass circle sectors. In the same petri dish 4 sectors in various combinations were exposed to media containing radiolabelled chlorpromazine (CPZ). Single and repetitive uptake and release of CPZ were measured in each cell type after individual exposure or exposure in any combination of cell types: In 2 hour competitive uptake studies fibreblasts reached 1.7 and 2.6 times the concentrations of C6-and ROC-cells, :respectively. In redistribution experiments the exchange of CPZ from preloaded to unloaded cells was tested. The exchange rate was dependent on cell type and loading period. It decreased with increasing time of exposure suggesting the formation of more stable CPZ stores. We have previously demonstrated that certain neurotransmitters such as VIP, noradrenaline (NA) and adenosine promote glycogenolysis in mouse cerebral cortical astrocyte cultures (Brain Res. 563: 227-233, 1991) . The question that we then addressed was the nature of the metabolic substrate released by astrocytes following glycogenolysis. We therefore determined the presence of various metabolic substrated released from astrocytes under static conditions and in the superfusate of a laminar flow system developed in our laboratory. Neither glucose nor any intermediate of the tficarboxylic acid cycle could account for the decrease in glycogen stores; astrocytes however were shown to release great quantities of pyruvate and lactate at a rate of 50 nmol/mg protein/minute. Noradrenaline but not VIP promotes a significant increase (42%) in lactate release. The effect of noradrenaline is mimicked by isoproterenol and inhibited by propranolol, suggesting a I~adrenergic mediation. When astrocytes are incubated in an isotonic solution containing no energy suhstrate, the glycogen stores are rapidly depleted and lactate, but not glucose, is released in the medium. In conclusion, lactate appears to play a major role in cerebral energy metabolism homeostasis, as substrate released by astrocytes to prey!de energy fuel for neurones and other neighboring cells. It is known so far that calcium-binding proteins (CaBPs: calbindin, pan, albumin and ca/retinin) are expressed in some cells of the pineal gland of laboratory mammals. The exact phenotype of these cells is unknown. Nevertheless, according to recent studies, it seems that some of them are ganglion cells. Therefore we studied the pineal gland of different species (gerbils, rats, goats, cows and humans) using calretiain (CR) antibody which is considered as marker for neurons (Andressen & al, Cell & Tissue Rss, 27t:181, 1993) ,

The CR-expressing cells were found in all oar species: in gerbils, rats, cows and humans they were scattered throughout the parenchyrna, whereas in goats they lined mostly the pericapillary spaces. According to our findings, it seems that most of the reactive ceils are rather neuron-like elements since morphological criteria for true neurons (axosomatic synapses, Nissl bodies, bundles of neurofilaments) are missing. However one cannot exclude that among these cells some are intrapineal ganglion cells, because of Ranvier's nodes found along the processes of some CRpositive cells.

Thus, it is possible that these cells are involved in the modulatory control mechanism of the pineal neurohumoral activity and/er by accumulation of intracellular calcium in the formation of pincal calcifications. Besides the excitatory amino acid transmitter glutamate, its sulfur containing analogue homocystcic acid (HCA) could play a role in excitatory processes in the central nervous system. HCA is present in and released from nervous tissue and is a potent neuronal excitant, predominantly activating N-Methyl-D-Aspartate receptors (Do et al., 1992) . These properties are suggestive of a classic synaptic neurotrausmitter role. On the other hand, HCA seems to be localized not in neurones but in glial cells (Grandes et al., 1991; Tschopp et al. 1992) . The efflux of HCA is temporally delayed following activation of specific pathways (Klancnik et at., 1992) ; these latter observations are not in accordance with the classic concepts of synaptically-mediated neurotransmission. A laminar flow superfnsian system of monolayer cultures was developed and the released materials from mouse cortical astrocytes, previously preincubated with [35S]-methioulne, were investigated. During application of either noradrenaline or the B-adrenergie agonlst isoproterenol, the extracellalar level of [35S]-methianine was decreased and that of labelled HCA was increased. This effect was not observed with the -adrenergic agonist methoxamine, and could be blocked by the ll-adrenergic antagonist propanoloL These results, combined with the delayed HCA release, the glial localisatian of HCA and its neuroactivity, strongly suggest a potential role of HCA as a gliotransmitter, modulated by noradrenaline inputs. Vasoactive intestinal peptide (VIP) induces long lasting metabolic effects in the mouse CNS. We have investigated a shorter neuromodulatory action of VIP on the synaptic responses in coronal slices of adult mouse cingulate cortex by means of intracelhilar current clamp recordings. Electrical stimulation of the underlying white matter evoked a multiphasic synaptic potential consisting of early EPSP~ early IPSP, late EPSP and late IPSP mediated by non-NMDA, NMDA, GABA A and GABA B receptors, respectively. VIP (0.1 ~d~l) superfused for 5 rain. had no effect, whereas concentrations over 0.2 IxM selectively and reversibly depressed the NMDA-medlated responses in 12/16 cells. In the presence of CNQX (10 [tM) and biencuUine (10 ~tM), the isolated NMDA-mediated late EPSP was still attenuated by 0.3 ~M VIP in 8/10 cells (mean:-42% al~er 30 rain.). In 3 other cells VIP induced a spreading depression (SD) and a strong, reversals inhibition of the late EPSP (mean:-64 % after 10 min.).These results indicate that in the absence of GABAergic inhibition VIP can induce two types of effects in the mouse cingulate cortex, a prolonged decrease of the NMDAmediated syaaptic response and a striking SD.

Interaction between VIP and glutamate on c-fos mRNA expression. |.-L. Martin, D. Gasser and P.] . Magistretti. Institut de Physiologie, Universit4 de Lausanne, Lausanne, Switzerland. The regulation of c-los mRNA expression by VIP and glutamate was examined in primary cultures of neurons originating from the mouse cerebral cortex. In primary cultures of cortical neurons, VIP increases cAMP levels and stimulates c-los mRNA expression. Interestingly VIP stimulation is completely blocked by MK-801, an NMDA receptor antagonist but not by CNQX or AP3, which antagonize AMPA/kainate and metabotropic receptors respectively, c-los expression stimulated by glutamate shares the same pharmacology. These results suggest an involvement of endogenously released glutamate in the stimulation of c-fos mRNA expression evoked by VIP. Furthermore, when added together, VIP and glutamate interact synergistically to increase cfos mRNA levels. Consistent with the pharmacology of VIP and glutamate, the synergistic interaction between VII' and glutamate on c-fos mRNA expression is also inhibited by MK-801. Interestingly, this synergism is completely inhibited by H-89, a protein kinase A inhibitor, whereas staurosporin and KN-62 which block protein kinases C and Ca++/calmodulin dependent respectively exhibit smaller inhibitory effects. Presynaptic proteins involved in neurotransmitter release (i.e. synapsin, B-50) and some postsynaptie GABA A receptor subunits possess phosphorylation sites for cAMP-dependent protein kinase (PKA) and/or protein kinase C (PKC). The functional consequences of phosphorylation by these kinases is not known. We have studied the action of specific activators of PKC and PKA, phorbol ester (PDBu) and forskolin, respectively, on GABAergic synaptic transmission in area CA3 of hippocampal slice cultures. PDBu significantly enhanced the amplitude of evoked monosynaptic IPSCs (in CNQX and AP5), and both PDBu and forskolin increased the frequency of spontaneous miniature IPSCs (m[PSCs) in the presence of CNQX, AP5 and TTX. This effect by PDBu was antagonized by staurosporin, a protein kinase inhibitor. PDBu and forskolin did not change the amplitude or decay of mIPSCs, suggesting that only presynaptic kinases are involved. Furthermore, PDBu enhanced mlPSC frequency by a comparable amount in the presence of the Ca 2+ channel blocker Cd 2+, indicating that PDBu did not enhance GABA release from presynaptic endings by facilitating Ca 2+ influx into axon terminals. Valiunas, V., Sc~ilinsky Fluri, G. and Weingart, R.

Arachidonic acid (AA) reversibly impairs the gap junction conductance (g~) in cardiac cells. Now, we examined whether AAacts directly or via its catabolites. Experiments were performed on pairs of neonatal rat myocytes using a dual voltage-clamp method. We used blockers which interfere with enzymes of various catabolic pathways. Pretreatment with i0 ~MPOCA (blocks carnitine acyltransferase I) did not prevent the decline in g9 by i0 ~M AA; i.e. intermediates of p-oxidation are not involved.

Similarly, preexposure to I0 ~M indomethacin (blocks the cyclooxygenase pathway) did not prevent AA from uncoupling; this rules out an involvement of prostaglandins and thromboxanes. Exposure to 10 ~MNDGA (blocks the 5-1ipoxygenase pathway) per se produced a decrease in gj. Likewise, exposure to i0 ~M ETYA (blocks the 15-1ipoxygenase pathway) also impaired gj. These effects cannot result from accumulation of endogenous AA because preexposure to 50 ~M 4BPB (blocks phospholipase A 2) did not prevent NDGA-or ETYA-induced uncoupling. Exposure to 75 ~M SKF 525-A (blocks the epoxygenase pathway) also produced a decrease in gj. Hence, NDGA, ETYA and SKF 525-A, compounds with different biochemical actions and of different chemical structure, act on g~ either directly or via an unknown mechanism.

$21-19 POTENTIAL CHANGES ASSOCIATED WITH ELECTRICAL ACTIVITY IN NONMYELINATED NERVE FIBRES A. Robert and P. Jirounek D~partement de Pharmacologie, CMU, 1211 Gen~ve 4 Simuheneous measurements of the extracellular concentration of K + ([K+]e), and of the concomitant changes in the membrane potential (Era) were measured in the rabbit vagus nerve during and after a period of activity. During a 15 Hz stimulation there was a large increase in the [K+]e , accompagned by a change in Era, which roughly followed the depolarization expected from the increase in [K+] e . At the end of the stimulation, although the [K+]e was still elevated, the nerve developed a posttetanic hyperpolarization (PTH). The PTH was generated by the electrogenicity of the Na+-K + pump, but its amplitude and time-course were strongly affected by two depolarizing currents: (i) an inward Ba2+-sensitive current of K + and (ii) an outward current of CI-. We hypothesize that during activity and during the early phase of recovery there was a Ba2+-sensitive uptake of potassium by the Schwann cells. The CI" current, by short-circuiting the Na+-K § pump, contributes to the control of the E m, and by this way to the driving force for the inward K § current. overdose of ifosfamide has ooo/red in a canc~c patient (25g i.v. IY0/24 hours with 20 g as ~utectarfE). Urir~ collecE{on was obtaJ/3ed in 24 hour ~ fcr t~o days. C~ was measured ~ a gas liquid d~romatogr~ method of the U5~ ~ (S~@pl. V, pag.2627-262g) ~sir~ a ~E~s ar~ ni~, sensitive ths~mi(~3ic d~.

~ U~s p~mt ~ ~ ~ ~ ~ ~ 1500 m~/24h and 207 r~/24h at day one ard day res~ec~vely. 9~ _h=~=mmtly ~ has also bsm foa~ in mti~s ~ mgh e~e ~ ~ (~/o~e). W~n ~ ~s am/nist~ to rats (i~ m~/~ ~c 200 m~f~ intraperit~neally), no ~ omild be detected in urine within 24 ht~r~ after drug administl-4(cicn. From these ~ data we conclude that C~ is extensi~_ly metabolized in rats and ~ mscbard_sm of ~ el ~mlnaticn in pati~rfc t=cJ_ne m~cjlr[c reflect a flmIcticrsl der~,~t of C~ m~t~0olisn in man to ifo~=~de ~tion. There are numerous reports of improved memory functions in rats following a dietary supplementation with choline. In some cases, this improvement can be related to enhanced cholinergic transmission (Mack et al., Behav Neurosci., 103, 1234 -1241 , 1989 ). But it is not clear whether there is a general improvement in memory or whether specific aspects of learning and memory are affected. This work was aimed at analysing the effects of perinatal cholinergic supplementation on the development of spatial abilities and upon adult performance.

Choline supplementation (3.5 g./l. in 0.02 M saccharine solution in tap water) was maintained during two weeks before birth. Additional supplementation was maintained from the 5th to the 1Oth week postnatal. Spatial learning capacities were studied at the ages of 26, 65 or 80 days in a cimular swimming pool (Morris place navigation task) and at the age of of 7 months on a homing board. Adult and aged rats were tested on a radial maze.

Selective aspects of the performance were markedly improved. This treatment had no specific effect upon spatial memory per se, but, instead, affected learning abilities by promoting efficient behavioural strategies hypothetically based on active representation and anticipation processes. The effects of the treatment remained evident up to 6 months following the supplementation. Barcelona, E-08193 Bellaterra Females of both lines (n=B/grp) were injected sc daily between days 15-20 of pregnancy with vehicle (V), I(DI) or 3(D3) mg/kg diazepam, or not(C).

In these studies(A,B) their 5-7 mo old offspring were subjected to a 50-run session in a shuttlebox(A:12 males, 6 females of each line/condition) or a 6min session in a hexagonal tunnel maze with strategically-placed barriers and a lit central arena(B_:6-7 males of each line/condition).

A: The freezing behavior of LD3 rats was reduced ~s LC, this behavior reverting more to escapes in males & avoidances in females, although V had an effect in the latter also. No within-line differences in inter-trial responses were seen, but pre-session activity was increased in female LVs and LDIs, returning to LC levels for LD3s. No differences existed among the H groups. B: Whereas H maze activity exceeded that of L, no wTthin-line differences appeared. HD3s entered the lit arena less than HCs or HVs, indicating an increased timidity after PND, and a trend in the same direction existed for LDI/LD3 vs LV. Two groups of abstinent female smokers, performing either a fixed rate or a subject paced version of a rapid information processing task were compared. Both groups participated in two test sessions (in which the task was performed twice) to compare two different types of cigarettes which were smoked before and during the second task trial: a) the subject's habitual cigarette with a nicotine yield of at least 0.7 mg/cigarette and b) a nearly nicotine free test cigarette with a tar yield comparable to that of the habitual cigarette. Reaction times to correct detections (series of three odd or even digits in a pseudorandom sequence of single digits) were longer with the subject paced version and not affected by the type of cigarette but shorter with the fixed rate version and even more decreased with the smoking of the habitual cigarette. On the other hand, the pre-to posttreatment increase in the subject paced processing rate was greater with the habitual than the test cigarette. It was concluded that the two task versions assess different cognitive functions. Whereas the fixed rate version primarily assesses vigilance performance and seems to be more suitable to measure changes in reaction time, the subject paced version is more sensitive to changes in speed capabilities in rapid information processing. Heart rate, physical activity, and cigarette consumption were continuously recorded under field conditions, whereas subjective craving was assessed six times per day and saliva cotinine was measured once daily. Abstinence, habitual cigarettes (with a nicotine yield of at least 0.7 mg/cigarette), and nearly nicotine free test cigarettes but with a tar yield comparable to that of the habitual cigarettes were compare~) in a sample of twelve female smokers for two days each. Whereas physical activity was similar for the three conditions, heart rate was highest with the habitual cigarettes, lowest on abstinence days and in between with the test cigarettes. Cigarette consumption was similar for both types of cigarettes and subjective craving was higher on abstinence than on smoking days but no differentiation between the two cigarette types was obtained. Saliva cotinine values were highest on the days with the habitual cigarettes but lower and similar with the test cigarettes and on abstinence days. It was concluded that heart rate and saliva cotinine depended only on the amount of nicotine absorbed, whereas subjective craving was reduced by smoking independently of the actual nicotine yield of the cigarette. (PTPS) is the rate limiting enzyme in the synthesis of BH, in man, a cofactor for several hydroxylases involved in catecholamine and serotonin biosynthesis. The human and rat liver cDNAS encoding the 16 kDa subunit of PTPS were expressed and the recombinant enzymes purified to homogeneity. The apparent K~ for the substrate, pI and heat stability of the re-combinant enzymes were similar to the native enzymes. The rat enzyme was crystallized and the X-ray structure showed a hexameric native form arranged as a sandwich of two trimers. Three histidine, a glutamic acid, and -also shown by site-directed mutagenesis -a cysteine residue characterize the active center of the enzyme thought to be Mg2+-dependent. According to the proposed ligands we replaced Mg 2* by Fe 2 § or Mn ~+ and observed an enhanced activity up to 100%.

M. Neerman-Arbez, V. Cirnlli and P. Halban. Laboratoires/eantet. Centre M4dical Universitaire, 1211 Gen~ve 4. PCI(3) and PC2, members of the mammalian family of pro-protein convertases homologous to the yeast kex 2, are both expressed in pancreatic islets of Langerhans and are thought to be responsible for the conversion of proinsulin to insulin and C-peptide in beta cells. However, the insulin secreting beta ceils are not the only ceils present in these complex microorgans, prompting us to evaluate the expression of PC1 and PC2 in islet beta and non-beta cells. Rat islet cells were sorted by autofluorescence activated flow cytometry to separate beta cells from non-beta ceils and conversion cndoprotease levels were analysed by Western blotting. In beta cells, PC1 levels were higher than PC2 (PC1/PC2=2.6+0.2) whereas the opposite was found for non-beta cells (PC1/FC2=0.05• confirming that PC1 is important for proinsulin conversion while suggesting a role for PC2 in the conversion of proglucagon, prosomatostatin and propancreatic polypeptide. Transformed murine cell lines (insulin-producing beta-TC and glucagonproducing alpha-TC) were found to faithfully reflect the primary rat cells in terms of their PC1/PC2 ratios. Finally, post-translational modification of the convertases themselves was found to differ between cell-types, in particular, a precursor 75 kD form of PC2 accumulated in beta cells whereas only the fully processed 67 kD form was detected in the non-beta cells. The kringle (2+3) domains (K2+3) were successfully expressed in E. coil A N-terminal hexahistidine tag was fused to insure the isolation of K2+3 by affinity chromatography on a Ni-2+-NTAagarose-column. To be able to remove the hexahistidine tag a factor Xa cleavage site was introduced between the 6 histidine residues and the N-terminus of the kringle structures. The correct arrangement of the disulfid bonds was determined by amino acid-and sequence analysis. The lysine binding site was proven by affinity chromatography on iysine-Bio-Gel, as previously shown for K2. The dissociation constant of K2+3 was measured by intrinsic fluorescence titration with 6-aminohexanoic acid.

A replication protein A copurifying DNA helicsse from calf thymus Anthl Georgakl, Narendra TuteJa, Blrglt Sturzenegger and Ulrich HObscher Department of Veterinary Biochemistry University of Z0rich-lrchel Winterthurerstrasse f 90 CH-8057 Z0rich, Switzerland A DNA helicase from calf thymus copurified with replication protein A through several steps of purification including DEAE-Sephacel, hydroxyapatite and single stranded DNA cellulose. It is finally separated from replication protein A on FPLC Mono Q where the DNA helicase elutes after replication protein A. We named this new calf thymus enzyme DNA helicase F. Characterization of the helicase F by affinity labeling with [a32p]ATP indicated that the enzyme has a catalytic subunit of 72 kDa. Gel filtration experiments suggested that DNA helicase F can exist both in a monomeric and an oligomedc form. The enzyme unwinds in the 5' ->3' direction in relation to the DNA strand it binds. All eight deoxyribonucleoside-and ribonucleosidetriphosphates could serve as an energy source. Testing a variety of DNNDNA substrates indicated that the DNA heitcase F preferentially unwinds very short substrates and is slightly stimulated by a single stranded 3'-tail Replication protein A allowed the DNA helicase F to unwind longer DNA substrates up to 400 bases suggesting that copurification of replication protein A with the DNA helicase might be of functional relevance. 2,4,5,2',4',5'-hexachlorobiphenyl (6-CB) shows limited elimination and extensive storage in adipose tissue in vivo (lipophilie unmetabolizable compound) while chlorpromazine (CPZ) is known for its lysosomal storage. Uptake and release of 6-CB and CPZ were studied in 3T3-preadipocytes (P) and 3T3-adipocytes (A). 3T3-Cells were cultivated in Dulbecco's minimal essential medium supplement with 10% fetal calf serum and grown to eonfluency. Differentiation was stimulated by dexamethason and bezafibrate exposure for 48h. Radiolabelled 6-CB (20~tM) or CPZ (5~tM) was added to the medium. Following a single dose 70-80% of CPZ were taken up within an hour into P and A. In contrast, 6-CB reached an intracellular plateau of 30-50% of the dose after lh in P whereas dependent on the triglyceride content A accumulated up to 95% of the dose. 6-CB uptake into A was temperature dependent and not saturable with repetitive daily doses up to 1 week. 6-CB uptake into P was fully reversible while the release from A was limited to 10% of the accumulated drug. In contrast, the release of CPZ was higher from P than from A as a consequence of the different lysosomal activities of the respective cells. The use of 3T3 cells allowed to show a remarkable difference in cellular handling of neutral and basic lipophilic drugs and may be helpful to further elucidate basic mechanisms involved.

Chronic granulomatous disease: An attempt to reconstitute the function of the X-linked form of the disease. Chronic Granulomatous Disease (CGD) is a group of congenital immunodeficiencies that predispose patients to recurrent severe bacterial and fungal infections. The cause of the disease is lack or impairment of 02-production in phagocytic cells due to mutations in any of the four components of the superoxide producing system. Approximately 70% of all CGD patients suffer from an X-linked form of the disease where the phagocyte oxidase gp91phox is affected. This project concentrates on the reconstitution of the deficient gene in phagocytic patient cells. Expression of recombinant gp91 was assayed in vitro and in vivo. Attempts to "tag" gp91 with a foreign epitope were unfortunately so far unsuccessful. The target cells for gene reconstitution are non-dividing monocytes. Thus, we have chosen an Adenovirus over other viral systems because of: a) its ability to infect nondividing cells, b) its genetic stability and, c) lack of human adenovirus related malignancies or side-effects. We constructed two recombinant Adenoviruses: one recombinant expresses a LacZ gene (control), another the gp9 lphox gene. We are trying the expression of recombinants Adeno/LacZ and Adeno/gp91 in normal and patient-derived cells (for example EBVtransformed normal and patient B-cell lines), as well as the reconstitution of NADPH oxidase function in peripheral blood monocytes of CGD patients.

KINETICS OF MOLECULAR CHAPERONE ACTION Schmid, D., Gehring, H., *Baici, A. and Christen, P., Biochemisches Institut der Universit&t Z~rich, CH-8057 ZOrich; *Rheumaklinik, Universit&tsspital, CH-8091 Z%rich Molecular chaperones of the Hsp70 type transiently sequester unfolded segments of proteins and promote their correct folding. Target peptides labelled with an environmentally sensitive fluorophore allowed to follow their binding to the molecular chaperone DnaK of Escherichia coli in real-time. The two-step process is characterized by relaxation times Z~ = 27 s and z2 = 200 s at 2 ~M DnaK and 0.i ~M ligand at 25~

In the presence of ATP, the formation of the complex is greatly accelerated and follows a single-exponential process with z : 0.4 s. The rate of dissociation of the complex was even more increased than that of association resulting in a decreased net affinity for ligands in the presence of ATP. The binding-release cycle of DnaK thus occurs in the time range of polypeptide chain elongation and folding and is too fast to be stoichiometrically coupled to the ATPase activity of the chaperone (ko~ ~ = 0.13 min i at 30"C). Supported by the Olga Mayenfisch Stiftung, Z%rich, and the Hartmann M~ller-Stiftung, Z~rich.

Comparison of the activities of native bovine seminal ribonuclease and that secreted from E. coil Schein, CH* (SIAT, Technopark, Pfingstweidstr. 30, CH8005 Ziirich) and Haugg, M (Lab. Org. Chemistry, E.TM., .

Seminal ribonuclease (BS-RNase) differs from the pancreatic RNase A in that it forms a cysteine-linked dimer and has a relatively higher specific activity in the cleavage of double-stranded (ds-) RNA substrates. We expressed BS-RN in a secretion system similar to that described previously (Biochem. J. 283:377-344, 1992) using the signal sequence of routine pancreatic ribonuelease. About I mg BS-RN was isolated per liter culture supernatant. Human and bovine recombinant interfcron-~ (IFN-r) inhibit the degradation of ss-and ds-RNA by bovine pancreatic RNase while they activate the activity of BS-RN on the same substrates (FEBS Letters, 270:229-332, 1990) . Recombinant bovine or routine pancreatic RNase (produced in our secretion system) are inhibited by IFN-y, while the recombinant BS-RN or a cysteine-linked dimer of RNase A is also activated by IFN-y. IFN-y activates BS-RN even in the presence of inhibitory concentrations (0.1-1 raM) of mononueleotides. Thus the mode of activation cannot be attributed to alleviation of product inhibition. Kinetic studies using 300 BP ds-RNA as substrate suggest that IFN-~ interacts directly with the BS-RN to increase the rate of highmolecular weight producffsubstrate release, as the apparent Ks increases proportionately to the increase in kcat. Significantly smaller movements were observed in the simulation with the liganded enzyme. The structural differences between the protein in the crystal and the ensuing calculated structures might arise from the absence of lattice forces in solution.

EVOLUTIONARY RELATIONSHIPS AMONG PYRIDOXAL-5'-P-DEPENDENT AMINO ACID DECARBOXYLASES Sandmeier, E., Hale, T.I. and Christen, P. Biochemisches Institut der Universit~t ZGrich, 8057 Z~rich. The pyridoxal-5'-P (PLP)-dependent amino acid decarboxylases (De) can be subdivided into four apparently unrelated groups. Group I is represented by glycine DC, group II comprises glutamate, histidine, tyrosine and aromatic-L-amino acid DC, group III procaryotic ornithine and lysine DC as well as the procaryotic biodegradative type of arginine DC, group IV eucaryotic ornithine and arginine DC as well as the procaryotic biosynthetic type of arginine DC and diaminopimelate DC. (N-l) Profile analysis, a more stringent application of profile analysis [Gribskov et al. (1990) Methods Enzymol. 183, 146-159] established the homology among the enzymes of each group. A search with the profile of group II indicated a distant relationship with aminotransferases and thus with the ~ family of PLP-dependent enzymes. No evidence was obtained that groups I, III and IV are related with other PLP-dependent enzymes or any other protein in the database. Apparently, the amino acid DC are of multiple evolutionary origin, in some cases even if they have the same substrate specificity.

A. Lo Russo, A.C. Passaquin, U.T. R~eg~, Ecole de Pharmacie, %hiv. lausanne, C~-1015 lallsaraqe Drug-induced local vasoccmstricticn appears to be respmnsible for the hypertensive side effect of the imn/nosti~pressant cyclosporin A (CsA). We have therefore ex~rnir~ the [Ca 2+] i increase caused by the vasoccr~trictor hormcme [Ar~]vascpressin (AVP) in vascular ameoth ~uscle cells (VSMC) treated with CsA.

[Ca2+] i levels were m~nitored either with a fluoresc~qce analyser using fura 2 or by 45Ca 2+ efflux.

Pretreatment of V~MC with CsA increased the AVP induced rise in [Ca2+]i. A leftward and upwsxd shift of the ccncentraticm-respmnse curve of the AVP-induced rise in 45Ca2+ efflux was also observed after CsA treatxr~%t. ilzg/cticn of CsA Dote~tiaticn w-as detectable after 3 rain, took 1 to 2 h to become fully established and was reversed after washout. CsA potentiaticm was cc~centraticn-deperdent with a threshold at 10 -7 M.

%]]is potentiating effect of CsA may be the underlyin~ cause for CsA-ind/ced hypert~3sicn. 92) Barja, F. 1 and Turian, G. I, iLab. Gen. Microbiology, 2Dept. Biochemistry, 3Station Exp. Zoology, University Of Geneva, 30 Quai Ernest-Ansermet, 1211 Geneva 4 The actin has been found from the simplest forms of cell to the highly evolved ones. In higher organisms it is transcribed by multigene families but in yeast and several filamentous fungi there appears to be only one actin gene. It was therefore of interest to study the expression of actin gene in N. crassa in which three actin isoforms have been characterized (Barja et al., 1991, FEMS Left. 77:19-24) . After isolating genemic DNA from wild type N. crassa a PCR was performed with primers corresponding to actin consensus sequences and an amplified fragment of the gene (verified by sequencing) was obtained. This fragment was 'used as a probe in a Southern to assess the number of aetin gene(s). Our results suggest that N. crassa has alsc only one actin gene. The blocking effect of the N-terminal decapeptide of msmooth muscle actin (SMA) (Ac.EEEDSTALVC) on the monoclonal antibody anti-aSM-1 was compared with that of synthetic peptides modified by changing the N-terminal acetyl group or by substituting a single amino acid in position 1 to 5. Using immunofluorescence or immunoblotting techniques, anti-c~SM-1 activity was abolished after incubation with the native peptide and peptides modified in position 1 (only when E~A) or 5. On immunoblots running BSA crosslinked peptides on denatured gels, only the natural peptide and peptides with substitution in position 1 or 5 were detected by the antibody. Our results indicate that Ac.(E)EED is the epifope for anti-c~SM-l. Binding of anti-~SM-1 to SMA increased actin polymedzation by decreasing the critical concentration. As control this action was not exerted on skeletal actin. This is the first example of the role of a precise N-terminal sequence in the polymerization of a single actin isoform. Double immunofluorescence for mSMA and total actin on cultured aortic SM cells microinjected with the native peptide showed a selective disappearance of mSMA staining suggesting that this peptide traps a protein involved in mSMA polymerization. Work is in progress to search for the putative actin-binding protein(s) physiologically interacting with the N-terminal sequence of o~-SMA. ( Within mitochondria, the mechanism of selective protein degradation is poorly understood. While the bulk of mitochondrial proteins are long-lived, some are degraded rapidly. The selective degradation of mitochondrial proteins is likely to be mediated by proteases similar to those found in bacteria; this is based on the hypothesis that mitochondria have evolved from bacterial endosymbionts, in conjunction with the finding that mammalian mitochondria contain an ATP-dependent proteolytic activity similar to that in bacteria. One ATP-dependent protease from bacteria, Lon, is of particular interest because it is involved in regulated protein turnover.

Degenerate oligonucleotide primers corresponding to two regions of the bacterial Lon gene were used to PCR amplify a 1 kb fragment from yeast DNA that encoded an open reading frame with striking similarity to the bacterial Lon protease. By screening a yeast cDNA library, we isolated a 3.6 kb clone potentially encoding a protein of 1133 amino acids. Haploids bearing a disruptedLON gene were unable to respire or to grow on ethanol/glycerol. Fractionation of rnJtochondrial matrix from wild-type cells revealed a peak of AIP-dependent proteolytic activity which was absent in the disruptant. Furthermore, we have identified matrix proteins that are degraded with a half-time of ~60 min in intact wild-type ceils but are stable in the Ion disruptants. Electron microscopy of lon disruptants revealed the presence of many electron dense bodies in the mitochondrial matrix which are not found in Lon + cells.

HYDROLYSIS OF HIGH MOLEKULAR WEIGHT KIN1NOGEN BY PLASMA KALLIKREIN Benner, S. and Duffer, H., Laboratory for Organic Chemistry, ETH H6nggerberg, CH-8093 Z/iNch High Molecular Weight Kininogen (HMWK) is cleaved by the serineprotease plasma kailikrein to generate the hypotensive peptide bradykinin in human blood. We established a measuring system for the time course of the hydrolysis of HMWK in its native and deglycosylated form and in the presence or absence of C~-inhibitor. Analysis of the bradykinin release and the yielded protein fragments lead to the three following results: 1. The hydrolysis did not follow plane Michaelis-Menten kinetics due to a HMWK-fragment operating as a competitive inhibitor. 2. Addition of the Ci-inhibitor stopped the HMWK-hydrolysis immediately, which is in contrast to the physiological role of plasma kallikrein in blood. 3. So far we have no evidence that the high content of O-glycosylated side-chains of the HMWK-light chain has an effect on the kinin-liberation as proposed in literature.

( Tetrahydrobiopterin (BH4) is the essential cofactor for the hepatic phenylalaoine hydroxylase, but also for tyrosine and tryptophan hydroxylases that are responsible for the production of nettrofransmitter precttrsors. Furthermore, BH4 is required for the deavage of gIyceryl ethers aztd is i~vo[ved in the biosynthesis of nitric oxide by the nitric oxide synthase. A vazia~t type of hyperphenylalani~emia is caused by a deficiency of BH4. The most frequent form of this cofactor deficiency is due to lack of 6-pyruvoyl-tetrahydropterin synthase (PTPS) activity. The human cDNA for PTPS was isolated, and the recombinant protein was found to be active when expressed in E. coil, thus allowing us to perform hmctional t e,3ting of mutant proteins. Primary skin fibroblasts derived from ~everal PTPS deficient patieIlts were investigated for the molecular nature of this autosomal recessive disorder. All fibroblasts had F1"PS mRNA expressed, and subsequent cDNA sequence analysis revealed different mutatkons in the coding region of PTPS (codon replacemenls, deletions, or nonseme codons) responsible for a lowered or no enzyme activity. In order to potentially correct PTPS activity (and restore BI-I 4 biosynthesis) we are establishing a recombhtant rel~oviral-mediated gene transfer system to efficiently introduce the wild-type human PTPS-cDNA in these fibroblasts. The thylakoid membrane (TM) contains I0 molecular species of phosphatidylglycerol (PG). The relative amount of these species depends on the plant species and growth conditions. Using phospholipase A2 (PLA2) at 0~ to digest preferentially PG in the outer monolayer of spinach TM, we have shown previously that this monolayer was enriched in PG (ca 70~). The purpose of this investigation was to determine the transmembrane distribution of each PG molecular species. To this aim, we have studied the hydrolysis kinetics of each species in the presence of PLA 2 in spinach and squash TM containing different relative proportion of molecular species. The hydrolysis rate and extent of the molecular species depended on the unsaturation degree of the fatty acid at the sn-I position as well as on the presence or absence of 16:1(3t) at the sn-2 position. For instance, the hydrolysis rate of PG ig:3/16:l(3t) was greater than that of PG 16:0/16:0. These results will be discussed in terms of the thylakoid transmembrane distribution of each PG molecular species. (Supported by the SNSF 3100-33693.92). From previous electrophysiological evidence, we concluded that metal ions (HgZ+,Cd2+,Ag +) at low concentrations (I0-6M) increase rapidly (10-60 s) cation (Na + , K + ) conductance at the apical membrane of epithelial cells. We investigated here the effects of these metals on [Ca2+]i in MDCK-I cultured cells, for a potential correlation between functional membrane changes and [Ca2+]i . Metals (10-4-10 -6 M) were added at the apical side of the epithelium, and [Ca2+]i was measured with fura-2. Hg 2+ induced a rapid (peak 5-10 s) and marked increase in [Ca2+]i, whereas Cd 2+ entailed a slower (peak 2-5 min) and marked response. The time-course of effects for both metals was similar to the electrophysiological changes. In contrast, the pattern of effects of Ag + on [Ca2+]i differed markedly from that on apical membrane conductance.

Thus, the effects of metals on [Ca2+]i are conspicuous but not tightly associated with changes in membrane conductance.

C. BULLIARD and L-L. DREYER, Institut de Biochimie, Universit6 de Fribourg, CH-1700 Fribourg Trans-plasma membrane oxydoreductases (PMO) play a significant role in energy transduction and post-receptor signal transmission, but the enzymes have not been characterized so far. We have achieved the purification of PMO from rat brain synaptic vesicles and from synaptic plasma membranes. The membrane enzymes were extracted by 1% Lubrol XL TM, precipitated with ammonium sulfate (between 40 and 70% saturation) and purified by means of hydrophobic chromatography on butyl-agarose and gel f'fltration on Superose-12, followed by PAGE under native conditions. Specific enzymatic staining of native PAGE gels yields two homogenous diaphoraselike activities, PMO-I and PMO-II . SDS-PAGE of the enzymes showed that PMO-I is made of two subunits of 52 kDa and 40 kDa whereas PMO-II consists of a single sub-unit of 52 kDa. The 52 kI)a subunits from PMO-I and PMO-[I are probably identical from their respective properties. Partial N-terminal sequencing (13 AA) of the 40 kDa subunit from PMO-I showed 84% homology with rat brain fructose-l,6-bisphophate aldolase. These data indicate that PMO is closely associated with the glycolytic enzyme that co-purifies in all steps under suitable conditions. The biochemical functions of PMO's remain to be established. Ferredoxin:thioredoxin reductase (FTR) is the key enzyme in the ferredoxin/thioredoxin system, the light-dependent regulatory system in oxygenic photosynthesis. It is a nucleus encoded ironsulfur protein, composed of two dissimilar subunits, one of which is the catalytic subunit containing a redox-active disulfide bridge functional in the reduction of thioredoxins and a 4Fe-4S cluster. Using PCR we have isolated and characterized a cDNA clone of 738 bp from spinach and a clone of 911 bp from corn coding for the catalytic subunits including the transit peptides. The first 31 (spinach) resp. 38 (corn) amino acids after the start exhibit typical features of chloroplast transit peptides. Following the transit peptides are 113 (spinach) or 114 (corn) ami[~o acids representing the catalytic subunits. The primary structures are highly conserved between these two species with 91% similarity and 81% homology. Seven of the eight Cys residues found in the spinach subunit and important for the catalytic activity are at conserved positions. (SNF 31-28811.90 and 31-37725.93) $23-20

R. ZURBRIGGEN and J.-L. DREYER, Institut de Biochimie. Univarsit6 de Fribourg, CH-1700 Fribourg Most mammalian cells display transplasma membrane oxidoreductase activity (PMO). The enzymes use intracellniar reduced pyridine nucleoticle to reduce an unspecific extracellular electron accepter as substrate. In order to set up a methodology for purifying, characterizing and quantifying PMO's, the plasma membrane from a neuroblastoma cell line, NB41A3, has been biotinylated. Subsequently the biotinylated proteins were purified by immanoprecipitation with avidin, antiavidin-antibodies and insoluble protein A. The protein recovery of an immunopurified membrane preparation was <0.15% of the protein content in the cell extract. Specific PMO activity was increased 15 to 20 fold compared to the activity in whole cells. This approach demonstrates the presence o1' PMO within the cell plasma membrane. This activity accounts for about one third of the total cellular diaphorase activity and cannot be attributed to an increased perraeabilization of the plasma membrane induced upon biotinylation nor to intracelluiar activity from lysed cells. PMO also promotes cell growth at low substrate concentrations (0.1-ll.tM). Native gel electrophoresis of iarinobiotinylated and affinity purified plasma membrane extracts displays two diaphorase-positive bands, indicating that a homogeneous cell population may express several PMO activities at the plasma membrane. Fluvastatin (FS) is a new HMG-CoA reductase inhibitor. Its affinity for 3 major human drug metabolizing P450 monooxygenases (CYP2C9, CYP2D6 and CYP3A4) was determined in liver microsomes. FS showed low affinity (Ki > 50 ~tM) for CYP2D6 and CYP3A4 whereas CYP2C9 was selectively and competitively inhibited (Ki = 0.1 ~tM). The potential for in vivo FS interactions with CYP2C9 substrates was therefore investigated in 14 volunteers using dicinfenac (DF) as a probe (25 mg orally) on days 0, 1 and 8. DF and 4'-HO-DF kinetics were determined by HPLC/UV. DF Cmax increased over time (0.28 [SD 0.12], 0.38 [0.20] and 0.45 [0.4] mg/L on day 0, 1 and 8, resp.). Oral clearance was reduced on days 1 and 8 (14 % and 15 %, resp.). A time dependent decrease in urinary metabolic ratio (4'-HO-DF/DF) was noted (1.07 [0.34], 0.90 [0,23] and 0.70 [0.18] on day 0, 1 and 8, resp. (p < 0.0001)) only over the first 4 hours. Fluvastatin therefore exhibits a two-phase interaction in vivo: an early transient and a late cumulative inhibition. The transient phase matches inhibition profiles predicted by quantitative models integrating in rive pharmacokinetics and in vitro inhibition data (Q-DIPS). Simple competitive inhibition by FS cannot explain the late phase. Other mechanisms (i.e. FS or metabolite cumulation, MI complexation) are hypothesized. In some situations FS is expected to cause interactions with CYP2C9 substrates, which are partially predicted by quantitative modeling of in vitro data (SNSF Project N o 32-36600.92). 1211 Gen~ve, Switzerland. Previously, a parvalbumin mutant, PVF~02w, was constructed with a unique reporter Trp in the middle of the hydrophobie center. In the present study three new parvalbumin mutants, derived from PVFt~w and containing alterations essential for Ca2+-binding in either one, PV.c• and PV_m, or both, PV-co~, of the two Ca2+-binding sites, were analyzed in order to study the metal binding properties of individual Ca~+-binding domains and their contributions to the structure and stability of the protein. Both, PV_ct, and PV.~, bind I Ca ~ § with affinity constants, Kc,, of 1.1,107 and 3.2-106 M "~ in the absence of Mg 2+. Mg 2+ shifts the Ca~+-binding isotherms of PV.co to higher free [Ca z+] according to the competition equation with KM~ = 2 9 103 M t. PV.m~ is much less effected by Mg z* (KMg = 80 M "t) and can be considered as possessing a Ca2+-specific site. Nevertheless, in the absence of Ca 2 § both, PV.co and PV~, bind I mole of Mg 2+ with much higher affinity than predicted from the competition studies. PV_cD;~ binds neither Ca 2. nor Mg 2+. Trp-fluorimetry and UV-difference spectroscopy revealed that the Ca~*-loaded conformations of PV_co, PV.EF and PVr~o~w are similar, with the Trp-102 deeply buried in the hydrophobie core. However, striking differences between the mutants are found in the metal-free and Mg~+-loaded forms. The conformation of PV~t~;~ is not affected by Ca 2~ or Mg z*. Our results indicate that destroying the Ca2+-binding of the EF loop, but not of the CD loop, strongly destabilizes the protein in the absence of Ca z § . Biochemlsches Institut der Unlversitfit Zfirich, Switzerland, CH-8057.

The sea urchin, Strongylocentrotus purpuratus, metallothionein gene, MTa, was constructed and expressed under the control of a heat shock promoter in a protease-deficient strain of E, coll. The nascent recombinant protein was stabilised by the addition of exogenous Cd. The Cd-containing protein was precipitated with ethanol and subsequently purified to homogeneity by gel permeation and ionexchange chromatography. The purified protein was identified as the desired gene product by tryptic peptide map analysis, sequence analysis and mass spectroscopy. Typical yields of protein were 2mg per litre of culture. The presence of 7 Cd ions per molecule was confirmed by the SH/Cd ratio and by the existence of 7 H3Cd NMR resonances. The apparent association constants of sea urchin MT with Cd, which has a charge of -8, were found to be approximately 15 times weaker than for rabbit MT, at pH 7. (MT) are small metalloproteins displaying as their main feature clusters of d I~ metal ions bound to the thiolate ligands of the abundantly occurring cysteine residues (Cys). From a set of over 85 sequences a general multialignment of 62 class I MT has been obtained on the basis of the Needleman & Wunsch algorithm. All Cys are conserved in the man%mals and over 83% of them in other vertebrates and invertebrates. On the basis of similarity/identity scores we can isolate groups of sequences.

Evolutionary relationships between species and groups have been calculated with the maximum parsimony method of Felsenstein and with the distance matrix method of Fitch and Margoliash.

The divergences observed among the mammalian MTs indicate a partition into four distinct subforms which all may have arisen before the separation of the species in evolution and which in part may differ in function.

$23-28

Bourque, A., Singh, A., Dykeman, A., MacMahon*, A., and Foster*, W. Atlantic Veterinary College, Charlottetown, Canada, and *Reproductive Toxicology Section, Environmental Health Centre, Tunney's Pasture, Ottawa, Canada. Hexachlorobenzene (HCB) is a known environmental pollutant that is produced as an industrial by-product of certain chemicals. When administered in high dosage, HCB induces alterations in the rhesus monkey ovary. A purpose of the study was to determine a no observable adverse effect level (NOAEL) for the compound. Twenty, cynomolgus monkeys, 6-13 years old were placed in five groups. HCB was given orally in concentrations of 0.01, 0.1, 1.0, and 10 mg/kg b.w. in gelatin capsules, daily, for 13 weeks. One group of animals fed glucose only, served as the control. Ovary specimens fixed in 2% buffered glutaraldehyde were prepared by conventional methods for transmission electron microscopy. Ultrastructural alterations were revealed in the developing ova and follicular cells in all the ovaries from HCB-treated animals in a dose-related manner. Clinical chemistry was unaffected by HCB. A NOAEL for this reproductive toxicant is yet to be determined. A.B. was a recipient of the NSERC Undergraduate Student award. An acidic haem-and zinc-containing protein has been isolated from bovine brain. Native and SDS-polyacrylamide-gel-electrophoresis reveal a monomeric protein with an apparent molecular weight of 47,2 kDa. The absence of co-staining with tetramethylbenzidine implies that the haem group is non-covalently bound. The electronic absorption spectrum, with a Soret-band absorption maximum at 412 nm and c~-and I~-bands at 540 and 575 nm, respectively, is similar to that of Fe(ll)-myoglobin, consistent with the presence of haemiron in the ferrous oxidation state. In air the protein was easily oxidized to the Fe(lll) form with a subsequent shift of the Soretband maximum to 405 nm. Atomic absorption measurements indicate approximately 1 mole of zinc and 1 mole of iron per mole of protein. The amino acid composition is similar to that of indoleamine-2,3-dioxygenase, although no such enzymic activity has been detected. In this study, we looked for this substance and its metabolites in young and old bovine lenses, because of their possible role in cataract formation. The substances were detected by HPLC analysis. The kynurenine aminotransferase (KAT) activity was determinated by the method of Tobes (Methods Enzymol 1987~ 142:217) . The fluorescent substance formed from 3-HK was characterized by thin layer chromatography followed by reaction with rtinhydrin, UV and fluorescence spectrometry, and atom bombardment for molecular mass determination. 3-HK at concentrations of 0.08-+0.01, 0.2_+0.04, and 1_+0.4 btg/g of tissue was detected in iris/ciliary body, retina, and calf lens respectively. This substance is deaminated by KAT. The activity of this enzyme was 2.6_+ 0.3, 3.7_+0.2, and 9.6+_2 ~tmol/g of tissue/hour in retina, iris/ciliary body, and lens respectively of old bovine eyes, but it was not found in calf eyes. The deamination of 3-HK resulted in the formation of a fluorescent substance which was identified as oxo-xanthurenic acid (OXA) with a molecular mass of 205 Daltons. The accumulation of OXA acid possibly interacting with lens proteins could induce cataract formation.

SNF 32-36058.92, EMDO, and Hartmann-Mtiller. The goal of covalent immobilization of oligonucleotide capture probes to solid supports has been accomplished by light-dependent, photolinker polymer mediated immobilization strategies. Synthetic oligonucleotides were immobilized by facile layer coating procedures. The procedure does not require functionalization of the oligonucleotide probe or the matedal surface. Oligonucleotides were linked to polystyrene with a biopolymer which was multiply substituted by photoactivatable diazirines. Capture probe photoimmobilization (350 nm irradiation) was strictly light and diazirine dependent. Using radiolabeled oligonucleotides as capture probes and polystyrene as material surface, the light dependent coupling efficiency was 40%. Nonspecific oligonucleotide binding was below 2 %. PhotoJmmobilized oligonucleotides retained the ability to form stable complexes with complementary DNA strands, and target oligonueleotides of varying length were captured as well as DNA fragments produced by PCR techniques. Indoleamine 2,3-dioxygenase (IDO), a superoxide radical scavenger, is present in transparent human lenses and produces 3-hydroxykynurenine, a UV-B filter. The synthesis of 3-hydroxykynurenine by IDO and its degradation by kynurenine aminotransferase (KAT) were measured in transparent lenses and cataracts. Post-mortem eyes of elderly persons between 72 and 84 years of age with transparent lenses were obtained from the eye-bank of the Department. Fresh cataracts were obtained from cataract surgery patients. IDO activity was measured by the method of Yamazaki et al (Biochem J 1985; 230:635) adapted for HPLC. KAT activity was measured by the method of Tobes (Methods Enzymol 1987; 142:217) . IDO activity found in post-mortem transparent lenses (cortex and nucleus) was 0.7_+0.3 nmol/mg protein/hour. In cataracts a low 1DO activity of 0.3_+0.2 nmol/mg protein/hour was found in the cortex but not in the nucleus. KAT activity, found in the eight cataracts examined, was 1.6_+0.9 nmol/mg protein/hour. Inhibition of IDO activity and induction of KAT activity seems to be involved in human cataract formation SNF 32-36058.92, EMDO, and Hartmarm-M011er Photolinker polymer mediated covalent immobilization of antibodies and F(ab') fragments has been achieved by newly established lightdependent coupling procedures. Anti alpha-l-fetoprotein monoclonal antibodies and F(ab') fragments derived from anti prostate specific antigen monoclonal antibodies were covalently linked to microplates. Diazirine derivatized bovine serum albumin served as multifunctional light activatable linking agent. Both immunoreagents, monoclonal antibodies and F(ab') fragments remained immunologically active after 350 nm irradiation (irradiance 0.7 mW cm -2 for 20 minutes). Covalency of antibody binding was inferred from i) the photoreagent dependence, ii) the light dependence of the immobilization process, and iii) the reversibility of immunocomplexation after acid treatment.

$23-33 TWO ~ OF EXTRACELLUIAR ELECIRICAL RESISTANCE IN V~kWRICULAR MYOCARDIL~ Fleischhauer J.C., Kl~ber A.G., Dept. of Physiol. Bern In myocardial tissue, the electrical resistive properties of the extracellular space are important for local current flow during e~citation and therefore influence impulse conduction and the magnitude of the extracellular electrical field. To assess the r~tlti~t nature of the resistance of the extracellular space, electrical cable analysis was applied on an arterially perf~-sed rabbit papillary mascle. ~he resistivity of the intravascular space was changed (70 to 220~I~n) by variation of hematocrit frcrn 0 to 60% in the perfusate. In order to change the voltm~ and hence the electrical resistance of the interstitial space, colloidosmotic pressure in the perfusate was varied by changing dextran {M r 70000) concentration (i0 to 80g/l). Altering the interstitial space volume had a marked influence on the extracellular resistance (ro) , conduction velocity and the magnitude of the extracellular field. Variations of the resistive properties of the intravascular space had no significant influence on ro, conduction velocity and the magnitude of the extraeellular electrical field. Our results suggest that the microvascular tree is insulated from the interstitial space and local electrical current flow during excitation in heart muscle is confined to the narrow interstitial clefts.

S04-25, S 15-69

S05-18, S05-19

S 11-05, S11-06

Bchini Hooft van Huijsduijnen

S08-12

S05-30

S15-45

S08-13

S08-05, S08-06

S01-08

S15-67, $23-19

S13-03 nook, EK

S01-15, S10-15

S05-02, S05-05

S 15-03, S15-29, S15-30

S19-03, S19-04, $19-05

S16-02, $16-15

S12-22

S15-48

S05-24

S03-01, S03-02, S03-03

S03-01 lrmler

S05-28, S09-06

S03-07, S03-10, S03-11

S18-05

Lipp, P. S05-26

SO 1-02

S15-15, S15-16, $21-03, $21-11, $21-12, $21-14

$23-29,$23-30

S01-17

S 15-29, S15-30

S12-12, S12-24 Molar

S15-19

S08-05, S08-06

E S09-06, S09-07

S15-03, S15-29, S15-30

E. S03-05

S05-09, S05-10, S05-51

S05-52

S10-16

S08-14

S04-23, S04-29, S05-13, S05-53, $20-15

S10-16, S10-19

S09-01, S09-02

S15-68, S 17-24

Stalder H. S03-19

S10-21

S09-09, S13-10, S13-13

S10-21 teHeesen

S01-10, S01-11, S01-18

S15-16 van Dillewyn

S05-32, S05-37, S05-38

S08-05, S08-06

S05-18, S05-19

Huber D., Ojha M. and Turian G. Laboratory of General Microbiology, University of Geneva, Sciences III, 30 Quai Ernest-Ansermet, 1211 Geneva 4, Switzerland. Ca2+-dependent protease in Allomyces is predominantly localized in the apical region of the exponentially growing hyphae (Huber and Ojha, submitted) . Many cytoskeletal proteins like actin and tubulin are also mainly localized in this growing region (Heath, l~91). It is known that the Ca2+-dependent protease proteolyzes a number of cytoskele~ ton proteins in order to regulate the plasticity of the cell (Mellgren, 1987) . We have studied the proteolysis of some cytoskeletal proteins by this protease purified from the aquatic fungus Allomyces erbuscula. Actin, tubulin and desmin, were found to be rapidly proteolyzed in vitro at a high substrate to enzyme ratio. Immunofluorescence studies of proteolyais made in situ in exponentially growing hyphae showed modification of apieally localized cytoskeletal proteins. This colocalization of the Ca2+-dependent protease and cytoskeletal proteins in the same apical region is inferred in relation to hyphal elongation. JBiological databases grow exponentially. In order to provide access, as much data las needed and as little volume as possible shall be returned upon a query. The IcurrentIy available DNA and Protein sequence databases are updated in a sophisticated network of procedures and expose a considerable amount of redundancy. Research is performed on optimizing the creation of non-redundant data sets. The resulting data are distTibuted in Switzerland, or made accessible to researchers who cannot utilize local resources. Transparent access to the Swiss resources, as well as paneuropean services, is provided with a job scheduling system which was developed in Basel, the Hierarchical Access System for Sequence Libraries in Europe ("HASSLE"), and various other computer programs which allow comprehensive browsing such as "GOPHER" and "WWW". Training courses to use the resource effectively are running throughout the year.