Effect of an industrial chemical waste on the uptake J. Serb. Chem. Soc. 87 (0) 1–17 (2022) Original scientific paper JSCS–12062 Published 14 December 2022 1 Screening the binding affinity of bile acid derivatives for the glucocorticoid receptor ligand-binding domain SRĐAN BJEDOV1*, SOFIJA BEKIĆ2, MAJA MARINOVIĆ2, DUŠAN ŠKORIĆ1, KSENIJA PAVLOVIĆ1, ANĐELKA ĆELIĆ2, EDWARD PETRI2 and MARIJA SAKAČ1 1Department of Chemistry, Biochemistry and Environmental Protection, Faculty of Sciences, University of Novi Sad, Trg Dositeja Obradovića 3, 21000, Novi Sad, Serbia and 2Department of Biology and Ecology, Faculty of Sciences, University of Novi Sad, Trg Dositeja Obradovića 2, 21000, Novi Sad, Serbia (Received 12 September, revised 22 September, accepted 28 September 2022) Abstract: The necessity of anti-inflammatory drugs such as glucocorticoids has been evident during the COVID-19 pandemic. Glucocorticoids, are the stan- dard therapy for the treatment of moderate and severe COVID-19 patients. However, serious side effects limit the use of these drugs, and anti-inflam- matory drugs with better pharmacological properties are urgently required. Bile acids are of interest, because of their anti-inflammatory and immunomodul- atory properties, facilitated through an unclear mechanism involving trans- membrane and nuclear receptors. In this work, we screened the binding activity of a number of bile acid derivatives, for the ligand-binding domain of glucocor- ticoid receptor (GR-LBD), the most important receptor for anti-inflammatory processes. Tested compounds include oximes, lactones, lactams, tetrazoles, dienones, C-24 alcohols and cholic acid amides. Cholic acid oxime, deoxycho- lic acid dienone, 3-keto-24-cholic alcohol and cholic acid amide showed best binding affinities for GR-LBD among tested compounds. The in silico mole- cular docking explanation is provided. SAR analysis showed that expansion of B and C steroid rings or attachment of heterocycle to C ring is not beneficial for binding; side chain should contain hydrogen donor group; the GR-LBD tolerate well different functionalities on C-3 position. These results provide valu- able information toward synthesis of the new glucocorticoids based on bile acids. Keywords: organic synthesis; docking studies; molecular modelling. INTRODUCTION Glucocorticoids (GCs) are a class of steroid hormones that are among the most commonly prescribed drugs used for the treatment of allergic, inflam- * Corresponding author. E-mail: srdjan.bjedov@dh.uns.ac.rs  Serbian Chemical Society member. https://doi.org/10.2298/JSC220912078B https://doi.org/10.2298/JSC220912078B 2 BJEDOV et al. matory, and immune disorders such as rheumatoid arthritis, asthma, brain edema, shock, and various blood cancers.1,2 More recently, they were also the first drugs shown to reduce deaths from COVID19.3 GCs are the most effective, cost-effi- cient, and necessary anti-inflammatory and immunomodulatory drugs available. However, the use of GCs is limited by serious adverse effects, such as diabetes, osteoporosis, muscle wasting, hypertension, and glaucoma. Also, certain groups of patients do not respond well to GC therapy.4 Thus, there is an urgent medical need for new molecules with both enhanced GC therapeutic activities, and fewer or less severe side effects. GCs mediate their effects via the glucocorticoid rec- eptor (GR), a ligand-activated transcription factor. GC activated GR performs its anti-inflammatory functions through a number of mechanisms, among which the most important are transrepression and transactivation.5,6 As steroid molecules, bile acids (BAs) play important roles as hormones that regulate a large number of metabolic processes, including inflammation. BAs are enzymatically synthesized in hepatocytes from cholesterol, conjugated with gly- cine or taurine, and stored in the gallbladder. When food is ingested, liver-syn- thesized primary BAs are secreted in the small intestine where they emulsify dietary lipids and lipid-soluble vitamins, enabling their absorption. A fraction of primary BAs is converted to secondary BAs by intestinal microbiota. Secreted BAs are reabsorbed and returned to the liver by a very efficient process known as enterohepatic circulation. The hormonal role of BAs is mediated through the BA- -activated nuclear receptor farnesoid X receptor (FXR), which controls gene transcription in BA homeostasis and via G-protein coupled receptor TGR5. TGR5 is broadly expressed in humans and is involved in various physiological and pathological processes, including energy expenditure, glucose homeostasis, obesity, and bile acid homeostasis.7 The anti-inflammatory properties of BAs were reported by Hench in the 1930s. He observed alleviation of rheumatic symptoms in patients with the onset of severe jaundice, a side effect associated with increased BA serum concentra- tions.8 BAs were first used as starting compounds for the synthesis of cortisol, which resulted in the 1950 Nobel Prize in Physiology or Medicine for Hench, Kendall and Reichstein.9 FXR is involved in the pathophysiology of several inf- lammatory diseases, including non-alcoholic fatty liver diseases, inflammatory bowel disease and atherosclerosis.10,11 Activation of TGR5 in macrophages and monocytes suppresses lipopolysaccharide-induced production of cytokines and prevents liver damage. The beneficial effects of TRG5 activation were noticed in multiple inflammatory diseases, including diet-induced obesity, atherosclerosis, colitis and steatohepatitis. More information about TGR5 and immunometa- bolism can be found in an excellent review by Perino and Schoonjans.12 Although the hormonal and anti-inflammatory activity of BAs via FXR and TGR5 pathways is well established, there is reasonable evidence that some of the BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 3 anti-inflammatory activities of BAs is also mediated through GRs. It was found that 5β-cholanic acid can bind to GR and modulate GR signaling in cell models of Parkinson’s disease.13 Taurochenodeoxycholic acid also exhibited anti-inflammat- ory and immunomodulatory properties by inhibiting transcription and expression of AP-1 via stimulation of the GR.14 Ursodeoxycholic acid exerts immune-suppres- sive effects by reducing IFN-γ production by liver lymphocytes, such as NK and NKT cells, in a GR-dependent manner, which may be an important immunological mechanism.15 The promiscuity of bile acids for FXR, TGR5 and GR could be exploited in an anti-inflammatory and immunomodulatory manner. In the present work, we explored the GR binding affinity of BA derivatives as a screen for unexplored BA structural features that may be important for GR binding. This work is aimed at the development of anti-inflammatory compounds that could lead to drugs with fewer side effects than current GCs. EXPERIMENTAL General synthetic methods 1H- and 13C-NMR spectra were recorded on a Bruker Avance III HD 400 (400 MHz 1H, 101 MHz 13C) apparatus using tetramethylsilane as an internal standard. HRMS spectra (TOF) were recorded on a 6210 time-of-flight LC/MS Agilent Technologies (ESI+) instrument. IR spectra were recorded on a Perkin Elmer Spectrum Two FT-IR spectrometer, and Thermo Nexus 670 FT-IR spectrometer, and melting points were determined on Stuart MP-10 apparatus. Flash chromatography was performed on silica gel 60 (0.04–0.063 mm, Merck). Synthesis of compounds 28 and 29 was done according to a procedure described in our earlier publication,16 and compound 27 was obtained following the Iqbal and Elliott pro- tocol.17 For synthesis of 25, we used a different route then Leppik,18 and 1 was obtained by a method described by Hüttenrauch.19 Spectral data of the compounds are given in the Supplementary material to this paper. (3EZ, 7Z, 12Z)-3,7,12-Trioximino-5β-cholan-24-oic acid (1) NaOAc (1.09 g, 13.3 mmol) and NH2OH×HCl (0.45 g, 7.09 mmol) were added to a suspension of dehydrocholic acid (DCA, 0.63 g, 1.6 mmol) in MeOH (20 mL). The mixture was refluxed for 40 min, and the resulting suspension filtered and washed with cold water. Compound 1 was obtained as a pure mixture of geometric isomers, as a white powder with a yield of 0.6541 g (91 %; thermal decomposition before melting started at 240 °C). Methyl (7Z,12Z)-3,3-dimethoxy-7,12-dioximino-5β-cholan-24-oate (2) NaOAc (0.80 g, 9.8 mmol) and NH2OH×HCl (0.25 g, 3.9 mmol) were added to a solution of 3 (0.44 g, 1 mmol) in MeOH (20 mL). The mixture was refluxed for 1 h, and the resulting sus- pension filtered and washed with cold water and dried. Pure compound 2 was obtained as a white powder at a yield of 0.35 g (72 %; thermal decomposition before melting started at 240 °C). Methyl 3,3-dimethoxy-7,12-dioxo-5-cholan-24-oate (3) TsOH (0.16 g, 0.9 mmol) and 2,2-dimethoxypropane (7 mL, 57 mmol) were added to a solution of DCA (1 g, 2.5 mmol) in MeOH (50 mL). The mixture was refluxed for 3 h, evaporated under vacuum, dissolved in EtOAc, washed with cold water (2×15 mL) and dried. 4 BJEDOV et al. Pure compound 3 was obtained as a white powder at a yield of 1.1 g (93 %; thermal decom- position before melting started at 225 °C). 12α-Hydroxy-3-oxo-5β-chola-4,6-dien-24-oic acid (25) Compound 27 (0.4985 g; 1.05 mmol) was dissolved in methanolic KOH (1 g; 17.8 mmol in 50 mL MeOH) and refluxed for 30 min. After reaction completion, the reaction mixture was poured into water (200 mL) and acidified with HCl (1:2) to pH 1. The resulting precipit- ate was filtered, washed with brine to neutral pH and dried. The raw mixture was purified by flash chromatography (CH2Cl2/EtOAc 94:6). Compound 25 was obtained as white needle-like crystals (after recrystallisation from acetone) at a yield of 0.3642 g (90 %); mp 251 °C, mp lit.20 249–252 °C. 12α-Hydroxy-3-oxo-5β-chola-4,6-dien-24-oic acid (25) and ethyl 12α-hydroxy-3-oxo-5β- -chola-4,6-dien-24-oate (26) Compound 27 (0.4872 g; 1.02 mmol) was dissolved in ethanolic KOH (0.19 g; 3.4 mmol in 70 mL EtOH) and refluxed for 1 h. After reaction completion, reaction mixture was poured in water (200 mL) and acidified with HCl (1:2) to pH 1. The resulting precipitate was filtered, washed with brine to neutral pH and dried. The raw mixture was purified by flash chroma- tography (CH2Cl2/EtOAc 94:6). Compound 25 was obtained at a yield of 0.1116 g (28 %) and compound 26 (mp 136 °C) at a yield of 0.2872 g (67 %). 2-(5β-chol-3-ene-7α,12α,24-triol)-N-(1-hydroxy-2-methylpropan-2-yl)acetamide (30) Triethylamine (0.3 mL), 2-amino-2-methyl-1-propanol (0.2 mL, 1.1 mmol) and EEDQ (0.23 g, 0.9 mmol) were added to a suspension of compound 29 (0.299 g, 0.69 mmol) in EtOAc (10 mL). The reaction mixture was refluxed for 5 h. After cooling to room tempera- ture, the reaction mixture was washed successively with 3 M HCl (2×4 mL), H2O (4 mL), 10 % NaHCO3 (2×10 mL) and then with water to neutrality (3×5 mL). The organic layer was dried and evaporated in vacuum to give an oily residue, which was further purified by flash column chromatography (CH2Cl2/MeOH 96:4). Pure 30 was obtained at a yield of 0.2612 g (75 %). Fluorescent assay in yeast A yeast-based fluorescent screen was applied for testing the relative binding affinities of BA derivatives 1, 2, 4–23 and 25–31 for the GR ligand binding domain (LBD). Assays were conducted following our previously published procedure; optimized for identification of stero- idal ligands of the GR.21,13 Saccharomyces cerevisiae FY250 strain (MATα, ura3-52, his32Δ00, leu2Δ1, trp1Δ6) and expression vector pRF4-6-rGR LBD-EYFP were kindly pro- vided by Dr. Blake Peterson from The University of Kansas.22 Yeast cells were transformed with plasmid DNA by treatment with lithium acetate and polyethylene glycol, to improve the efficiency of exogenous DNA uptake, following the procedure of Gietz et al.23 Transformed yeast cells were then incubated at 30 C until the appearance of transformant colonies on agar plates. Selection medium supplemented with 2 % raffinose was then inoculated with recom- binant yeast cells grown to saturation in a Biosan orbital shaker-incubator ES-20/60. Saturated yeast cells were diluted in fresh medium and grown to mid-logarithmic phase (OD600 nm  0.5), monitored using a Nicolet Evolution 100 UV–Vis spectrophotometer. Protein expression was induced by addition of galactose to a final concentration of 2 %. Bile acid derivatives, prednisolone or estradiol (high- and low-affinity GR ligands) were added to a final concen- tration of 100 μM. Stock solutions of all tested compounds were freshly prepared in DMSO. Incubation was continued for 15 h at 25 C. Resulting fluorescence intensity was detected by BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 5 fluorimeter (Fluoroskan Ascent FL) and fluorescence microscopy (Olympus BX51) using a FITC filter. For fluorimetry, 150 μL of cell suspension was added to a microplate in triplicate and fluorescence was recorded using an excitation and emission filter set of 485 and 538 nm. Growth medium without cells served as a blank. To normalize fluorescence intensity to cell number, the optical density of yeast cells was measured at 600 nm using a Thermo Lab sys- tems Multiscan EX spectrophotometer. Ligand binding affinity was calculated as the fold fluorescence change between cells exposed to test compounds and those treated with negative control ligand, estradiol. Histograms showing the relative binding affinity of bile acid derivat- ives and control ligands for GR-LBD were created in Origin Pro 8 with included error bars, obtained by propagation of standard error of the mean. Additionally, the fluorescence intensity distribution of recombinant yeast cells expressing GR LBD-YFP treated with bile acid deri- vatives or control compounds, was visualized by fluorescence microscopy. Molecular docking Coordinates for the GR-LBD receptor were converted to PDBQT format in the program VegaZZ using the available “receptor.c” script.24 Structural coordinates for compounds 1, 25, 28 and 31 were created in the program Avogadro 1.0.3;25 Avogadro: an open-source mole- cular builder and visualization tool based on the structure of dexamethasone, taken from PDB 1M2Z.26 Ligand geometries were optimized in Avogadro 1.0.3 using an MMFF94 force field and 500 steps of conjugate gradient minimization with a convergence setting of 10e-7. Non- -polar hydrogens were merged and gasteiger partial charges added in VEGAZZ 3.1.0,27 using the “ligand.c” script, and resulting ligand coordinates were converted to PDBQT format for Autodock. Grid maps for atoms present in the tested compounds were created using the pro- gram AutoGrid, with a grid center of 31.04, 7.76, 12.52, grid spacing of 0.0375 nm and diel- ectric of –0.1465. Docking simulations in Autodock 4.2,29 were conducted using the Lamarck- ian genetic algorithm with the following parameters: GA population size 150. GA num evalu- ations 250000. GA num generations 27000. AutoGrid calculations and Autodock simulations were conducted using the PyRx virtual screening tool (version 0.8).27 Results were visualized in the program PyMol (v0.99).28 RESULTS AND DISCUSSION Chemistry Compounds used in the present study were chosen to examine the influence of easy-to-achieve modifications of the BA molecule on GR binding affinity. Structures of compounds investigated in the present work can be categorized into roughly seven categories: oximes, lactones, lactams, tetrazoles, dienones, C-24 alcohols and cholic acid amides. Structures of oxime derivatives are shown in Fig. 1. Synthesis of oximes 1 and 2 is shown in Scheme 1. The reaction of dehydro- cholic acid (DCA) with hydroxylamine hydrochloride afforded a mixture of geo- metric isomers 1, at the C-3 oxime group. The mixture of isomers 1 was obtained with a yield of 91 % and individual isomers were not separated. The isomer ratio of 1 could be speculated to be 3:2 based on the intensity of C-3 signals (156.87 and 156.67 ppm) present in 13C-NMR spectra. Compound 2 was obtained in a two-step sequence. First, the carbonyl group at C-3 was regioselectively trans- 6 BJEDOV et al. formed into acetal, while the carboxyl group was esterified which afforded 3 in excellent yield (93 %). This reaction represents a very good way for regioselect- ive protection of the C-3 bile acid carbonyl group. Next, C-7 and C-12 carbonyl groups were converted into oxime to obtain dioxime 2. Fig. 1. Structures of oximes. Scheme 1. Reagents and conditions: a) NH2OH×HCl, NaOAc, MeOH, 40 min, reflux; b) MeOH, 2,2-dimethoxypropane, TsOH, 3 h, reflux. Unlike 1, compound 2 was obtained as a stereochemically pure compound. The stereochemistry of oxime groups at C-7 and C-12 was determined as Z-based on the absence of cross-peaks in NOESY NMR spectra that originate from, the through space, interaction of oxime hydrogens with any steroid skeleton hydro- gens. Only the Z configuration for the oxime groups could provide enough dis- BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 7 tance to explain the lack of NOE interactions. Stereochemistry could be exp- lained by the larger volume available for hydroxyl groups if the oximes have a Z configuration (Fig. 2). Oximes 4,29 530 and 631 were prepared by procedures found in the corresponding literature. Fig. 2. Different perspectives of 2 show the available space for Z-oriented oxime groups at C-7 and C-12. Bile acid derivatives with a lactone or lactam moiety in B or C steroid ring (Fig. 3) 7–14 were synthesized according to published procedures,31–33 while the stereochemistry and physicochemical properties of these compounds were des- cribed by Poša et al.34 Fig. 3. Structures of bile acid lactone and lactam derivatives 7–14. The structures of tetrazoles 15–23 used in the present work are shown in Fig. 4. Tetrazole rings are fused on B and C steroid rings. The general synthesis of tetrazole compounds is shown in Scheme 2, using the synthesis of 16 as an example. Bile acid derivative 24, with C-3 acetoxy, C-24 ethoxycarbonyl groups, and a carbonyl group at C-12, was subjected to Schmidt reaction condition (Scheme 2) with trimethylsilyl azide as an azide source and tri- 8 BJEDOV et al. methylsilyl trifluoromethanesulfonate as an amide intermediate activator. Details of the synthesis of the tetrazole compounds used in the present study have been published by our group.35 Fig. 4. Structures of the bile acid tetrazole derivatives 15–23. Scheme 2. The general synthesis of tetrazole compounds 15–23 is illustrated using the synthesis of 16 as an example. Reagents and conditions: TMSN3, TMSOTf, ACN, rt, 3 h (a). Structures of dienones 25 and 26, alcohols 28 and 29 and amides 30 and 31 are shown in Fig. 5. Dienones 25 and 26 were obtained by the following procedures: when com- pound 27 (Scheme 3) was refluxed in methanolic potassium hydroxide solution, elimination and hydrolysis occurred to afford 25 at a yield of 90 % (78 % rep- orted by Leppik et. al. 18). Reaction samples were collected every 5 min and since BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 9 only one product was detectable by TLC, it appears that both elimination and hydrolysis reactions occur simultaneously. The same result was obtained upon treatment of 27 with methanolic KOH solution for five days at room temperature, or 2 M NaOH solution for five hours at room temperature, or by reflux with HCl solution (1:3) in water/acetone 5/7 for five hours. When ethanol was used instead of methanol, in addition to 25 (28 %), transesterification occurs to afford ethyl ester 26 (67 %). Signals corresponding to vinylic protons in 25 are easily visible by 1H-NMR as overlapping singlet peaks of H-4 and H-6 at 6.16 ppm, and a singlet from H-7 at 5.61 ppm. An additional quartet at 4.03 and a triplet at 1.16 in the 1H-NMR of 26 suggests the presence of an esterified carboxyl group. Fig. 5. Structures of dienone, alcohol and amide derivatives 25, 26, 28–31. Scheme 3. Reagents and conditions: a) KOH, MeOH, 0.5 h, reflux; b) KOH, EtOH, 1 h, reflux; c) HCl, H2O. Easily obtainable and well-known 3-dehydrocholic acid (3DCA) was trans- formed into the bile alcohol 28 using acetal protection as seen in 3, and sub- sequent reduction (Scheme 4). The same bile alcohol 28 was transformed into 29 10 BJEDOV et al. using a Wittig–Horner–Emmons reaction. Details concerning the synthesis of 28 and 29 are available in our previous publication.16 Scheme 4. Synthesis of bile alcohol 28. Compound 30 was obtained in the good yield (75 %) by reaction of 29 with 2-amino-2-methyl-1-propanol using EEDQ as a coupling agent (Scheme 5). Scheme 5. Reagents and conditions: 2-Amino-2-methyl-1-propanol, EEDQ, TEA, EtOAc, 5 h, reflux (a). The cholic acid amide 31 (Fig. 5) was synthesized according to a procedure reported by Miljković et al.36 Fluorescent assay in yeast In order to test the relative binding affinities of the synthesized BA derivatives, we used a yeast-based fluorescent screen that is optimized for identification of stero- idal ligands of the glucocorticoid receptor, as previously described.13,21 Briefly, the fluorescence intensity of yeast cells expressing the ligand-binding domain (LBD) of the glucocorticoid receptor (GR) fused to a yellow fluorescent protein (YFP) has been shown to respond in a dose-dependent manner to treatment with GR agonists such as prednisolone.13,21 Because the assay is measured in yeast, the concentration of ligand required to elicit a response may not necessarily be the same concentration that would affect GR activity in vivo. During optimization of the assay, we measured dose response for a positive control ligand (prednisolone, Fig. S-1 of the Supple- mentary material). As assay sensitivity was highest at a final prednisolone concen- tration of 100 μM, bile acid derivatives were tested at 100 μM. In the present study, bile acid derivatives 1, 2, 4–23 and 25–31 were tested against GR-LBD, by mea- BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 11 suring fold fluorescence changes between cells exposed to test compounds compared with those treated with a negative control ligand, E2 (estradiol), or a positive control GR agonist (prednisolone). It can be seen in Fig. 6, that bile acid derivatives 1, 25, 28 and 31 displayed fold fluorescence changes indicative of moderate binding affinity for GR-LBD compared with prednisolone or estradiol. Fig. 6. Bile acid derivatives 1, 25, 28, and 31 showed fold fluorescence changes consistent with moderate binding affinity for GR-LBD-YFP based on a fluorescent screen in yeast following 15 h exposure at a final concentration of 100 μM. Ligand binding affinity was calculated as the fold fluorescence change between cells exposed to test compounds compared with those treated with negative control ligand, E2 (estradiol), measured by fluorimetry. Prednisolone (P) was tested as a positive control GR agonist. To validate these results, yeast cells treated with the same compounds were also visualized by fluorescence microscopy. As can be seen in Fig. 7, treatment of recombinant yeast expressing GR-LBD-YFP with the GR agonist predniso- lone resulted in a strong increase and relocalization of fluorescence intensity, while treatment with compounds 1, 25, 28 and 31 resulted in a more moderate increase in overall fluorescence intensity compared to cells treated with negative control estradiol. Fig. 7. Recombinant yeast cells expressing GR-LBD-YFP treated with 100 μM estradiol (E2, negative control GR ligand), prednisolone (P, positive control GR ligand) or test BA derivatives 1, 25, 28 and 31 for 15 h, visualized using a fluorescence microscope. The intensity of cell fluorescence is proportional to GR ligand binding affinity. 12 BJEDOV et al. Molecular docking Based on fluorescent screening in yeast, compounds 1, 25, 28 and 31 have moderate binding affinity for GR-LBD compared with prednisolone, a strong GR agonist. To visualize these results in a molecular framework, binding poses and binding energies were predicted for compounds 1, 25, 28 and 31 by molecular docking in the program Autodock.37 Because the structure of GR in complex with prednisolone has not been published, coordinates for GR-LBD in complex with another GR agonist, dexamethasone (PDB 1M2Z), were used as ‘receptor’ for docking.26 For all docking simulations, the receptor was kept rigid and the ligand was allowed to rotate around flexible bonds. To validate the Autodock protocol for our system, control redocking simulations were conducted using dexamethasone as a positive control “ligand”. Autodock correctly reproduced the X-ray structure with an RMSD of < 0.06 nm2 and a strong binding energy of – 57.20 kJ/mol. Based on the X-ray structure, dexamethasone (DEX) is in part held oriented in the GR ligand binding site via polar contacts with arginine 611, glutamine 570 and asparagine 564. As can be seen in Fig. 8, very similar binding poses and intermolecular contacts are formed by test compounds 28, 25 and 31. Fig. 8. Molecular docking poses and predicted binding affinities for test compounds 25, 28 and 31 compared with dexamethasone (DEX), a positive control GR ligand. The X-ray structure of GR ligand binding domain in complex with dexamethasone (PDB 1M2Z) was used as receptor for Autodock simulations. Top ranking poses for the following compounds are shown: A) dexamethasone, B) 28, C) 25 and D) 31. All three compounds bind in the same orientation, forming interactions with Asn564, Thr739 and Gln642 from the D-ring side, and Arg611 and Gln570 from the A-ring side of these BA derivatives. Consistent with experimental results, moderate binding energies of <−41.84 kJ/mol were predicted for each of these BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 13 test compounds: for compound 25, −46.36 kJ/mol; compound 28 −45.94 kJ/mol, and compound 31 −42.16 kJ/mol. None of the test compounds had a predicted binding affinity greater than prednisolone, as expected. Interestingly, although compounds 1 displayed moderate affinity for GR-LBD in fluorescence experi- ments, docking of C-3 Z-isomer of 1 failed to predict binding to GR under the simulation conditions used. In general, Autodock is capable of correctly predict- ing the binding affinity and geometry for a set of compounds if ligand binding is not associated with significant backbone and side chain conformational chan- ges.37 Thus, binding by compound 1 likely induces conformational changes to the GR active site that are different from the GR conformation present in complex with dexamethasone, preventing Autodock from correctly estimating the binding pose and affinity for compound 1. For compounds 25, 28 and 31, molecular docking results could be used in future studies as a starting point to design molecules with increased binding affinity, and/or specificity for GR-LBD. CONCLUSION Compounds used in the present work were designed to examine the influ- ence of specific functionality modifications of BAs on their GR binding pro- perties. Bile acids have a suitably functionalized steroid skeleton for modification in strategic regions for GR interactions. Our glucocorticoid receptor binding assay provided insight into structural features important for binding of the tested molecules to GR. The results of the SAR analysis are shown in Fig. 9. Fig. 9. Results of SAR analysis. Four differently functionalized compounds 1, 25, 28 and 31 showed moderate binding affinity for GR-LBD compared with prednisolone or estradiol. The oxime geometric isomers in 1 showed the best relative binding affinity, not just among 14 BJEDOV et al. tested oximes but among all tested compounds. Although we were not able to sep- arate the geometric isomers in mixture 1 and molecular docking did not enable prediction of possible interactions with the enzyme, binding affinity could be attri- buted to the C-24 carboxylic group and oxime groups. The hydroxyl group at C-11 with β orientation (11β-OH) has been previously shown to be very important for GR-glucocorticoid interactions,38 and the relative binding affinities measured for 1 suggest that the C-12 oxime group could be exploited as an alternative hydrogen bond donor to the 11β-OH group. Lactone, lactam, and tetrazole derivatives did not show any significant binding affinity. It seems that the B or C steroid ring expan- sion or attachment of heterocycle at the C ring is not beneficial for binding affi- nities. Compounds 28, 25 and 26 have a keto or enone moiety, which is known to be involved in anchoring steroids in the receptor pocket.39 However, it seems that the hydrogen bond donating group at C-24 is very important for binding to GR. This was observed for all of the tested compounds. Ethyl ester 26 has a much smaller GR binding affinity compared to free acid 25. Compound 28 with the C-3 keto and C-24 hydroxyl group has a GR binding affinity similar to 1, among the best of the tested compounds. There is a degree of plasticity in the GR-LBD for the steroid A-ring and C-3 functional groups, since compounds 1 with C-3 oxime E and Z groups; 29, with carboxymethyl group at C-3, and even compound 30 with the corresponding amide at C-3 have moderate or less than moderate GR-LBD binding affinity. This work gives important insight into structural modification of bile acid steroid skeleton that could be used for development of new small mole- cule drug for treatment of inflammatory diseases. SUPPLEMENTARY MATERIAL Additional data and information are available electronically at the pages of journal website: https://www.shd-pub.org.rs/index.php/JSCS/article/view/12062, or from the corres- ponding author on request. Acknowledgements. The authors acknowledge financial support from the Ministry of Education, Science and Technological Development of the Republic of Serbia (Grant No. 451-03-9/2021-14/ 200125). И З В О Д ИСПИТИВАЊЕ АФИНИТЕТА ДЕРИВАТА ЖУЧНИХ КИСЕЛИНА ЗА ВЕЗИВАЊЕ ЗА ЛИГАНД-ВЕЗУЈУЋИ ДОМЕН ГЛУКОКОРТИКОИДНОГ РЕЦЕПТОРА СРЂАН БЈЕДОВ1, СОФИЈА БЕКИЋ2, МАЈА МАРИНОВИЋ2, ДУШАН ШКОРИЋ1, КСЕНИЈА ПАВЛОВИЋ1, АНЂЕЛКА ЋЕЛИЋ2, ЕДВАРД ПЕТРИ2 и МАРИЈА САКАЧ1 1Департман за хемију, биохемију и заштиту животне средине, Природно–математички факултет, Универзитет у Новом Саду, Трг Доситеја Обрадовића 3, 21000, Нови Сад и 2Департман за биологију и екологију, Природно–математички факултет, Универзитет у Новом Саду, Трг Доситеја Обрадовића 3, 21000, Нови Сад Антиинфламаторни лекови као што су глукокортикоиди били су неопходни током пандемије COVID-19 за лечење пацијената са умереним и тешким облицима COVID-19. Међутим, озбиљни нежељени ефекти ограничавају употребу ових лекова и хитно су пот- https://www.shd-pub.org.rs/index.php/JSCS/article/view/12062 BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 15 ребни антиинфламаторни лекови са бољим фармаколошким својствима. Жучне кисе- лине привлаче све више интереса, због својих антиинфламаторних и имуномодула- торних својстава, испољених засад нејасним механизмом који укључује трансмембран- ске и нуклеарне рецепторе. У овом раду смо испитали афинитет везивања низа деривата жучних киселина за лиганд-везујући домен глукокортикоидног рецептора (GR-LVD), најважнијег рецептора за антиинфламаторне процесе. Испитивана једињења укључују оксиме, лактоне, лактаме, тетразоле, диеноне, C-24 алкохоле и амид холне киселине. Међу тестираним једињењима: оксим холне киселине, диенон деоксихолне киселине, 3- кето-24-алкохол и амид холне киселине показали су најбоље афинитете везивања за GR- LVD. Објашњење везивања поменутих деривата помогли су in silico доковање. Анализа односа структуре и активности је показала да експанзија B и C стероидних прстенова или везивање хетероцикла за C прстен није од користи за везивање; бочни ланац треба да садржи водоник-донорску групу; и GR-LVD добро толеришу различите функционал- ности на позицији C-3. Ови резултати пружају вредне информације за прављење нових антиинфламаторних молекула базираних на жучним киселинама. (Примљено 12. септембра, ревидирано 22. септембра, прихваћено 28. септембра 2022) REFERENCES 1. N. Sundahl, J. Bridelance, C. Libert, K. De Bosscher, I. M. Beck, Pharmacol. Ther. 152 (2015) 28 (https://doi.org/10.1016/j.pharmthera.2015.05.001) 2. F. Buttgereit, R. H. Straub, M. Wehling, G. R. Burmester, Arthritis Rheumatol. 50 (2004) 50 3408 (https://doi.org/10.1002/art.20583) 3. RECOVERY collaborative group*, N. Engl. J. Med. 384 (2021) 693 (https://doi.org/10.1056/NEJMoa2021436) 4. J. Souffriau, M. Eggermont, S. Van Ryckeghem, K. Van Looveren, L. Van Wyngene, E. Van Hamme, M. Vuylsteke, R. Beyaert, K. De Bosscher, C. Libert, Sci. Rep. 8 (2018) 12894. (https://doi.org/10.1038/s41598-018-31150-w) 5. J. Vandewalle, A. Luypaert, K. De Bosscher, C. Libert. Trends Endocrinol. Metab. 29 (2018) 42 (https://doi.org/10.1016/j.tem.2017.10.010) 6. A. Louw. Front. Immunol. 10 (2019) 1693 (https://doi.org/10.3389/fimmu.2019.01693) 7. E. Lontchi-Yimagou, E. Sobngwi, T. E. Matsha, A. P. Kengne. Curr. Diab. Rep. 13 (2013) 435 (https://doi.org/10.1007/s11892-013-0375-y) 8. P. S. Hench. Br. Med. J. 20 (1938) 394 (https://doi.org/10.1136/bmj.2.4050.394) 9. The Nobel Prize, https://www.nobelprize.org/prizes/medicine/1950/summary/ (20.06.2022.) 10. R. M. Gadaleta, M. Cariello, C. Sabbà, A. Moschetta, Biochim. Biophys. Acta 1851 (2015) 30 (https://doi.org/10.1016/j.bbalip.2014.08.005) 11. J. Hageman, H. Herrema, A. K. Groen, F. Kuipers, Arterioscler. Thromb. Vasc. Biol. 30 (2010) 1519 (https://doi.org/10.1161/ATVBAHA.109.197897) 12. A. Perino, K. Schoonjans, Trends. Pharmacol. Sci. 12 (2015) 847 (https://doi.org/10.1016/j.tips.2015.08.002) 13. B. Vasiljević, E. Petri, S. Bekić, A Ćelić, Lj. Grbović, K. Pavlović, RSC Med. Chem. 12 (2021) 278 (https://doi.org/10.1039/D0MD00311E) * Complete list of collaborators in the RECOVERY trial is available at: https://www.nejm.org/doi/suppl/10.1056/NEJMoa2021436/suppl_file/nejmoa2021436_appen dix.pdf https://doi.org/10.1016/j.pharmthera.2015.05.001 https://doi.org/10.1002/art.20583 https://doi.org/10.1056/NEJMoa2021436 https://doi.org/10.1038/s41598-018-31150-w https://doi.org/10.1016/j.tem.2017.10.010 https://doi.org/10.3389/fimmu.2019.01693 https://doi.org/10.1007/s11892-013-0375-y https://doi.org/10.1136/bmj.2.4050.394 https://www.nobelprize.org/prizes/medicine/1950/summary/ https://doi.org/10.1016/j.bbalip.2014.08.005 https://doi.org/10.1161/ATVBAHA.109.197897 https://doi.org/10.1016/j.tips.2015.08.002 https://doi.org/10.1039/D0MD00311E https://www.nejm.org/doi/suppl/10.1056/NEJMoa2021436/suppl_file/nejmoa2021436_appendix.pdf https://www.nejm.org/doi/suppl/10.1056/NEJMoa2021436/suppl_file/nejmoa2021436_appendix.pdf 16 BJEDOV et al. 14. L. Li, C. Liu, W. Mao, B. Tumen, P. Li, Molecules 24 (2019) 4513. (https://doi.org/10.3390/molecules24244513) 15. T. Takigawa, H. Miyazaki, M. Kinoshita, N. Kawarabayashi, K. Nishiyama, K. Hatsuse, S. Ono, D. Saitoh, S. Seki, J. Yamamoto, Am. J. Physiol. Gastrointest. Liver Physiol. 305 (2013) G427 (https://doi.org/10.1152/ajpgi.00205.2012) 16. M. Poša, S. Bjedov, V. Tepavčević, M. Mikulić, M. Sakač, J. Mol. Liq. 303 (2020) 112634 (https://doi.org/10.1016/j.molliq.2020.112634) 17. M. N. Iqbal, W. H. Elliott, Steroids 53 (1989) 413 (https://doi.org/10.1016/0039- 128X(89)90022-6) 18. R. Leppik, Steroids 41 (1983) 475 (https://doi.org/10.1016/0039-128X(83)90087-9) 19. R. Hüttenrauch, Arch. Pharm. Pharm. Med. Chem. 294 (1961) 366 (https://doi.org/10.1002/ardp.19612940608) 20. K. Tamaki, J. Biochem. 45 (1958) 299 (https://doi.org/10.1093/oxfordjournals.jbchem.a126869) 21. S. Bekić, M. Marinović, E. Petri, M. Sakač, A. Nikolić, V. Kojić, A. Ćelić, Steroids 130 (2018) 22 (https://doi.org/10.1016/j.steroids.2017.12.002) 22. S. Muddana, B. Peterson, Chembiochem. 4 (2003) 848 (https://doi.org/10.1002/cbic.200300606) 23. D. Gietz, A. St Jean, R. A. Woods, R. H. Schiestl, Nucleic. Acids. Res. 20 (1992) 1425. (https://doi.org/10.1093/nar/20.6.1425) 24. S. Dallakyan, A. J. Olson, Methods. Mol. Biol. 1263 (2015) 243 (https://doi.org/10.1007/978-1-4939-2269-7_19) 25. M. D. Hanwell, D. E. Curtis, D. C. Lonie, T. Vandermeerschd, E. Zurek, G. R. Hutchison, J. Cheminform. 17 (2012) (https://doi.org/10.1186/1758-2946-4-17) 26. R. K. Bledsoe, V. G. Montana, T. B. Stanley, C. J. Delves, C. J. Apolito, D. D. McKee, T. G. Consler, D. J. Parks, E. L. Stewart, T. M. Willson, M. H. Lambert, J. T. Moore, K. H. Pearce, H. E. Xu, Cell 110 (2002) 93 (https://doi.org/10.1016/s0092-8674(02)00817-6) 27. A. Pedretti, A. Mazzolari, S. Gervasoni, L. Fumagalli, G. Vistoli, Bioinformatics 37 (2021) 1174 (https://doi.org/10.1093/bioinformatics/btaa774) 28. PyMOL, http://www.pymol.org/pymol (15.09.2021.) 29. N. Meanwell, H. Roth, E. Smith, D. Wedding, and J. Kim Wright, J. Org. Chem. 56 (1991) 6897 (https://doi.org/10.1021/jo00024a036) 30. Y. Huang, J. Cui, S. Chen, C. Gan, Q. Yao, Q. Lin, Bioorg. Med. Chem. Lett. 23 (2013) 2265 (https://doi.org/10.1016/j.bmcl.2012.08.064) 31. H. H. Abdu-Allah, T. T. Chang, W. S. Li, Steroids 112 (2016) 54 (https://doi.org/10.1016/j.steroids.2016.04.013) 32. I. S. Zharinova, A. A. Bilyalova, S. I. Bezzubov, Acta Crystallogr., E 74 (2018) 816 (https://doi.org/10.1107/S2056989018007259) 33. M. I. Duran, C. González, A. Acosta, A. F. Olea, K. Díaz, L. Espinoza, Int. J. Mol. Sci. 8 (2017) 516 (https://doi.org/10.3390/ijms18030516) 34. M. Poša, V. Tepavčević, Lj. Grbović, M. Mikulić, K. Pavlović, J. Phys. Org. Chem. 34 (2021) e4133 (https://doi.org/10.1002/poc.4133) 35. D. Škorić, O. Klisurić, S. Jakimov, M. Sakač, J. Csanádi, Beilstein J. Org. Chem. 17 (2021) 2611 (https://doi.org/10.3762/bjoc.17.174) 36. D. Milijkovic, K. Kuhajda, J. Hranisavljevic, J. Chem. Res. 2 (1996) 106 (https://open.uns.ac.rs/handle/123456789/12941) 37. G. M. Morris, R. Huey, W. Lindstrom, M. Sanner, R. Belew, D. Goodsell, A. Olson, J. Comput. Chem. 16 (2009) 2785 (https://doi.org/10.1002/jcc.21256) https://doi.org/10.3390/molecules24244513 https://doi.org/10.1152/ajpgi.00205.2012 https://doi.org/10.1016/j.molliq.2020.112634 https://doi.org/10.1016/0039-128X(89)90022-6 https://doi.org/10.1016/0039-128X(89)90022-6 https://doi.org/10.1016/0039-128X(83)90087-9 https://doi.org/10.1002/ardp.19612940608 https://doi.org/10.1093/oxfordjournals.jbchem.a126869 https://doi.org/10.1016/j.steroids.2017.12.002 https://doi.org/10.1002/cbic.200300606 https://doi.org/10.1093/nar/20.6.1425 https://doi.org/10.1007/978-1-4939-2269-7_19 https://doi.org/10.1186/1758-2946-4-17 https://doi.org/10.1016/s0092-8674(02)00817-6 https://doi.org/10.1093/bioinformatics/btaa774 http://www.pymol.org/pymol https://doi.org/10.1021/jo00024a036 https://doi.org/10.1016/j.bmcl.2012.08.064 https://doi.org/10.1016/j.steroids.2016.04.013 https://doi.org/10.1107/S2056989018007259 https://doi.org/10.3390/ijms18030516 https://doi.org/10.1002/poc.4133 https://doi.org/10.3762/bjoc.17.174 https://doi.org/10.1002/jcc.21256 BILE ACID DERIVATIVES FOR THE GLUCOCORTICOID RECEPTOR 17 38. T. Mitić, S. Shave, N. Semjonous, I. McNae, D. Cobice, G. Lavarey, S. Webster, P. Hadkoe, B. Walker, R. Andrew, Biochem. Pharmacol. 86 (2013) 146 (https://doi.org/10.1016/j.bcp.2013.02.002) 39. U. Lind, P. Greenidge, M. Gillner, K. F. Koehler, A. Wright, J. Carlstedt-Duke, J. Biol. Chem. 275 (2000) 19041 (https://doi.org/10.1074/jbc.M000228200). https://doi.org/10.1016/j.bcp.2013.02.002 https://doi.org/10.1074/jbc.M000228200