{Quantitative structure-activity relationship modelling of influenza М2 ion channels inhibitors} J. Serb. Chem. Soc. 86 (7–8) 625–637 (2021) Original scientific paper JSCS–5449 625 Quantitative structure–activity relationship modelling of influenza М2 ion channels inhibitors IVANKA G. STANKOVA1*, RADOSLAV L. CHAYROV1, MICHAELA SCHMIDTKE2, DANCHO L. DANALEV3**, LIUDMILA N. OGNICHENKO4, ANATOLY G. ARTEMENKO4, VALERY A. SHAPKIN5 and VICTOR E. KUZ'MIN4 1Department of Chemistry, South-West University “NeofitRilski”, Blagoevgrad, 2700, Bulgaria, 2Friedrich Schiller University, Department of Virology and Antiviral Therapy, Jena, 207745, Germany, 3University of Chemical Technology and Metallurgy, Biotechnology Department, 1756 Sofia, 8 blvd. Kliment Ohridski, Bulgaria, 4A.V. Bogatsky Physico- -Chemical Institute of Ukrainian National Academy of Sciences, Department of Molecular Structure and Chemoinformatics, 86, Lustdorfskaya doroga, Odessa, 65080, Ukraine and 5Department of Department of Theoretical Foundations of Chemistry, Odessa National Polytechnic University, 1, Shevchenko ave., Odessa 65044, Ukraine (Received 9 May 2020, revised 15 March, accepted 5 May 2021) Abstract: A series of adamantane derivatives (rimantadine and amantadine) incorporating amino-acid residues are investigated by simplex representation of molecular structure (SiRMS) approach in order to found correlation between chemical structures of investigated compounds and obtained data for antiviral activity and cytotoxicity. The obtained data from QSAR analysis show that adamantane derivatives containing amino acids with short aliphatic non-polar residues in the lateral chain will have good antiviral activity against the tested virus A/H3N2, strain Hong Kong/68 with low cytotoxicity. QSAR experiments and in vitro data also show good correlation and reveal that modified adaman- tine derivatives including guanidated in the lateral chain amino acid and β- -amino acids as substituents show low to none activity. Keywords: QSAR study; molecular simplex; adamantine derivatives; aman- tadine; rimantadine; amino acids. INTRODUCTION Adamantane derivatives have been used successfully for the prevention and treatment of influenza A virus infection for more than 30 years.1,2 The aminoada- mantanes (amantadine and rimantadine) block M2 proton channel and thus stop virus replication. However, they are no longer effective because of widespread drug resistance. S31N is the predominant and amantadine-resistant M2 mutant, present in almost all of the circulating influenza A strains as well as in the pan- *,** Corresponding authors. E-mail: (*)ivastankova@abv.bg; (**)ddanalev@uctm.edu https://doi.org/10.2298/JSC200509036S ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. 626 STANKOVA et al. demic 2009 H1N1 and the highly pathogenic H5N1 flu strains.3 Structural and biochemical studies of the S31N mutant showed that replacing 31Ser, which is located in the helix–helix interface, with the bulkier Asn results insubstantially weaker helix–helix packing. Since the pocket is composed of residues from two adjacent TM helices, the stability and physical properties of the pocket depend on the dynamics and conformation of helical packing.4 The M2 proton channel from influenza A virus, a prototype for a class of viral ion channels known as viro- porins, conducts protons along a chain of water molecules and ionizable side- chains, including 37His. Drugs inhibit proton conduction by binding to an aque- ous cavity adjacent to M2's proton-selective filter, thereby blocking access of proton to the filter, and altering the energetic landscape of the channel and the energetics of proton-binding to 37His. According to Gaiday et al. studied cage compounds inhibit the M2 ion channel by binding to the 37His residue. The ada- mantane cage fits into a pocket formed by 41Trp residue, while the hydrogen bond is formed between hydrogen atom of ammonium nitrogen and the nitrogen of histidine residue.6 One of the possible approaches to restore the antiviral pro- perties of this class of compounds is to incorporate more than one functional group in their molecule. Amino acids and peptides are promising alternative for such kind of modification because of their multi functionality. Herein, we report the QSAR analysis of some adamantanes modified with different substituents and the role of specific modifications on the biological act- ivity. EXPERIMENTAL Chemical synthesis Compounds aimed to this study were previously synthesized according to Scheme 1. The synthetic protocol of target compounds was as follow: 3 mmol of 2-(1H-ben- zotriazole-1-yl)-1,1,3,3-tetramethylaminium tetrafluoroborate (TBTU) was dissolved in 15 mL CH2Cl2. Further, the corresponding Boc-αN-protected amino acid (3 mmol) and DIPEA (3.1 mmol) were added to a solution of TBTU. The obtained mixture was stirred at room tem- perature for 30 min and 3 mmol of rimantadine or amantadine and 3 mmol N,N-dimethyl- aminopyridine (DMAP) were also added. This mixture was stirred at room temperature for another 3 h, and then evaporated to dryness. After evaporation the obtained oil was purified by flash chromatography in system ethyl acetate:n-hexan (50:50 volume ratio). 1 eq. of Boc-αN-Aaa-amantadine or Boc-αN-Aaa-rimantadine was dissolved in 10-fold excess of trifuoroacetic acid (TFA) at 0 °C. The reaction mixture was stirred until fully depro- tection of Boc-group under chromatographic control in systems chloroform-methanol (95:5 volume ratio). The excess of TFA evaporated and the obtained oil was dissolved in 10 mL methanol. Further 25 % ammonium hydroxide was added until the pH reached around 8. The solvent was evaporated under vacuum. Tha obtained crystals were dissolved in ethyl acetate and washed with 3×25 mL water. The organic layers were combined, dried on anhydrous Na2SO4 and solvent removed under vacuum. The yield of all compounds as well as 1H- and 13C-NMR analysis confirming their structures are given in Chayrov et al.17 For guanidated analogues synthesis the following procedure was used. ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 627 Scheme 1. Synthesis of amino acid derivatives of adamantine; R = –NH2 is amantadine (Am), R = CH3CHNH2 is rimantadine (Rim), Aaa = amino acid; for amantadine Aaa = Ala; Phe; Phe(4-F); Val; for rimantadine Aaa = Ala; Gly; Ile; Leu; Phe; D-Phe(4-F); L-Phe(4-F); Val; beta-Ala; Tyr. 0.67 mmol of Boc-deprotected Aaa-rimantadine or Aaa-amantadine was dissolved in 2 mL of acetonitrile. Further to the obtained solution 1.00 mmol of 1H-pyrazole-1-carboxami- dine hydrochloride and 2.0 mmol triethylamine were added. The reaction was run for 48 h at room temperature. At the end of the reaction time the solvent was evaporated under vacuum. The obtained crude oil was dissolved in 25 mL chloroform and washed several times with 5 % NaHSO4 (pH 3). The organic layers were combined, dried with anhydrous Na2SO4 and sol- vent was evaporated under vacuum. The residue was crystalized in methanol/diethyl ether. Biological studies The results obtained by the realized biological tests17 are presented in Table I. TABLE I. Data of biological tests of studied compounds; CC is cytotoxic concentration; HNTC is high-nontoxic concentration No. Structural formula of compound (abbreviation) CC50 / μM HNTC / μM IC50 / μM log (IC50 / μM) MDCK cells MDCK cells Against A/Hong Kong/68 Average Class numb. Average Class numb. Average – 1 Rim O NH2 F (D-Phe(4-F)-Rim) 19.91 1 6.23 1 100.00 2.00 2 Rim O NH2 OH (L-Tyr-Rim) > 100 0 9.97 1 0.52 –0.28 ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. 628 STANKOVA et al. TABLE I. Continued No. Structural formula of compound (abbreviation) CC50 / μM HNTC / μM IC50 / μM log (IC50 / μM) MDCK cells MDCK cells Against A/Hong Kong/68 Average Class numb. Average Class numb. Average – 3 Rim O NH2 (L-Phe-Rim) 21.93 1 4.75 1 100.00 2.00 4 Rim O NH2 (L-Val-Rim) 70.33 1 10.20 1 100.00 2.00 5 Rim O NH2 (L-Leu-Rim) > 100 0 11.66 1 0.75 –0.12 6 Rim O NH2 (L-Ile-Rim) 58.35 1 5.33 1 100.00 2.00 7 Rim O NH2 (L-Gly-Rim) > 100 0 13.89 1 0.11 –0.96 8 Rim O NH2 (L-Ala-Rim) > 100 0 14.65 1 1.53 0.18 9 Rim O NH2 (β-Ala-Rim) > 100 0 20.78 1 15.72 1.20 10 Rim O NH OH NH2 NH (Gua-L-Tyr-Rim) > 100 0 36.69 1 8.19 0.91 11 Rim O NH N NH H H (Gua-L-Ala-Rim) > 100 0 > 100 0 41.90 1.62 ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 629 TABLE I. Continued No. Structural formula of compound (abbreviation) CC50 / μM HNTC / μM IC50 / μM log (IC50 / μM) MDCK cells MDCK cells Against A/Hong Kong/68 Average Class numb. Average Class numb. Average – 12 Rim O NH NH NH2 (Gua-L-β-Ala-Rim) > 100 0 > 100 0 100.00 2.00 13 Am O NH2 F (L-Phe(4-F)-Am) 83.61 1 15.35 1 0.81 –0.09 14 Am O NH2 OH (L-Tyr-Am) > 100 0 > 100 0 3.93 0.59 15 Am O NH2 (L-Phe-Am) 89.40 1 15.10 1 0.75 –0.12 16 Am O NH2 (L-Val-Am) > 100 0 > 100 0 1.32 0.12 17 Am O NH2 (L-Ala-Am) > 100 0 > 100 0 1.41 0.15 18 Am O NH OH NH2 NH (Gua-L-Tyr-Am) > 100 0 > 100 0 59.71 1.78 19 Am O NH N NH H H (Gua-L-Val-Am) > 100 0 > 100 0 17.13 1.23 ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. 630 STANKOVA et al. TABLE I. Continued No. Structural formula of compound (abbreviation) CC50 / μM HNTC / μM IC50 / μM log (IC50 / μM) MDCK cells MDCK cells Against A/Hong Kong/68 Average Class numb. Average Class numb. Average – 20 Am O NH N NH H H (Gua-L-Ala-Am) > 100 0 > 100 0 18.37 1.26 21 Am-H > 100 0 n/a 0 0.39 –0.41 22 Rim-H – – – – 0.06 –1.25 QSAR calculations In the present study the simplex representation of molecular structure (SiRMS) approach7,8 was used for calculation of structural descriptors for all investigated compounds. The simplex approach is based on isolating and counting the number of molecular fragments (pairs, triples, quadruples of atoms). Any molecule in the framework of SiRMS can be represented as a system of different specific fragments (simplexes) of fixed composition and topology. Various atomic charac- teristics can be used for the vertex differentiation in the simplex, such as the uniqueness of the atom (atom nature or a more detailed type), electronegativity, partial charge, lipophilicity, electronic polarizability (refraction), H-bond donor/acceptor potential, van der Waals interact- ions, etc. For atomic characteristics having real values (electronegativity, lipophilicity etc.) at the preliminary stage the range of values is divided into a certain number of groups G (G is a tuning parameter of models and can vary, as a rule from 3 to 7; see the example showing how using vertexes differentiation by atomic charges in Supplementary materials). In addition, electronegativity, refraction, molecular weight and octan-1-ol–water partition coefficient (Log P) are calculated as integral characteristics that describe the whole molecule. The calculation of descriptors was carried out at the 2D level of molecular structure rep- resentation. The 2D-QSAR models are the most popular in structure-property studies.7,8 In this case only molecular topology is taken into account, i.e. all information is extracted from the structural formula. The relationships between the calculated molecular descriptors and the studied properties of investigated molecules were established by methods of partial least squares (PLS)9 and random forest (RF).10 The removal of highly correlated and constant descriptors, the trend vector method,11 genetic algorithm (GA)12 and the automatic variable selection (AVS) strategy7 have all been used to select the descriptors in PLS. The removal of highly correlated descriptors is not necessary for PLS analysis since descriptors are reduced to a series of uncorrelated latent vari- ables. This procedure frequently helps to obtain more adequate models. During this procedure one descriptor from each pair having a pair correlation coefficient r satisfying |r|>0.90 is eli- minated. It was previously discovered13 that descriptors involved in the best Trend Vector models (several decades of models with approximately identical quality) form a good subset for their subsequent usage in PLS. The noise elimination can be one of the more probable ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 631 explanations of the success of the trend vector procedure. GA is used as a tool for the select- ion of adequate PLS models. We used the small set of following algorithm parameters as mut- ation rate = 0.3, crossingover rate = 0.7, type of crossingover is double. Descriptors, from the best model obtained by a preliminary AVS procedure, are normally used as the starting “population”. In this work we run GA only once. The GA is definitely not a tool for the elu- cidation of the global maximum or minimum, and very often subsequent AVS procedures and different enumerative techniques allow one to increase the quality of the PLS models obtained. According to the QSAR/QSPR OECD principles,14 applicability domain (AD) of dev- eloped models was estimated. A ellipsoid model of structural space,7 Williams plot15 and tree AD approach16 were used for AD estimation. Within the framework of SiRMS approach it is possible to define the relative influence of the different physical and chemical factors on the character of the molecules interaction with the biological target.7,8 For this purpose it is necessary to sum and compare absolute values of PLS regression coefficients of descriptors for all groups used for atom differentiation. The clear interpretation is one of the advantages of SiRMS approach.8 On the basis of developed QSAR models the influence of each atom over a particular property can be calcul- ated. The contribution of each atom in the molecule can be defined as the ratio of the sum of PLS regression coefficients for all simplexes containing this atom to the number of atoms in the simplex. The atomic contribution depends on the number of simplexes that include this atom. The number of simplexes is not constant. It varies in different molecules depending on other constituents. Thus, this contribution is non-additive. The analysis of such information allows selecting different fragments which have negative or positive influence on a considered property. The descriptors calculation and data analysis were performed using HiTQSAR software,7 which was developed in the department of molecular structure and chemoinformatics of the A. V. Bogatsky Physico-chemical institute of National Academy of Sciences of Ukraine. RESULTS AND DISCUSSION During the first step of synthetic scheme Boc-αN-protected amino acid are bonded to rimantadine or amantadine. Next step is reaction of Boc-group depro- tection with treatment with 10 fold excess of TFA. Finally target compounds are obtained as a free bases on their amino functions by treatment with 25 % ammo- nium hydroxide till pH around 8. Thus using structures and results from biological investigations of target rimantadine and amantadine analogues QSAR analysis was carried out and the effects of the substituent on the in vitro cytotoxicity (CC50), HNTC and antiviral activity (against human influenza virus A/H3N2 strain/Hong Kong/68) were also investigated.17 In this study we used a dataset consisting of 22 derivatives (13 are riman- tadine derivatives and 9 – amantadine derivatives, Table I). In addition, the derivatives Tyr-rimantadine, Ala-rimantadine, β-ala-Riman- tadine, and Ala-amantadine, Val-amantadine, which exhibit the lowest cytotoxic- ity were guanidated using 1H-pyrazole-1-carboxamidine reagent (Scheme 1).18 All compounds were tested against human influenza virus A/H3N2 strain Hong Kong/68 and their activity is described by Chayrov et al.17 The antiviral ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. 632 STANKOVA et al. activity against influenza virus strain A/H3N2 strain Hong Kong/68 in vitro of new analogs of amantadine and rimantadine conjugated with amino acids reveal that the highest antiviral activity combined with low cytotoxicity was demon- strated by the rimantadine derivative conjugated with the simplest in structure glycine. Moreover, glycyl-rimantadine presented a high stability profile after incubation in human plasma for 24 h. Interestingly, the analogues of amantadine with the amino acids L-phenylalanine and L-(4-F)-phenylalanine exhibited high activity although lower than those of the amantadine in the same concentration. In addition, amantadine and rimantadine analogues conjugated with guanidine showed low toxicity but they also exhibited low activity. The various aliphatic and aromatic amino acids as substitutes in the rimantadine molecule have no sig- nificant effect on antiviral activity. The use of β-alanine reduces antiviral act- ivity. Guanidation of the rimantadine and amantadine analogues do not increase antiviral activity unlike guanidated oseltamivir analogues. Using SiRMS approach for calculation of structural descriptors a total of about 2000 different structural characteristics were calculated for the investigated compounds. The ranges of dividing the atom properties into intervals are, as already men- tioned above, tuning characteristics and the following schemes were used in the calculation of simplex descriptors: electronegativity: A < 2.19 ≤ B < 2.5 ≤ C < 3 ≤ ≤ D, refraction: A < 1.5 ≤ B < 3 ≤ C < 8 ≤ D, atomic charge: A < –0.16 ≤ B < < –0.10 ≤ C < –0.04 ≤ D < 0.01 ≤ E < 0.07 ≤ F < 0.13 ≤ G, lipophilicity: A < –1.51 ≤ B < –0.96 ≤ C < –0.42 ≤ D < 0.13 ≤ E < 0.68 ≤ F < 1.23 ≤ G, VDW attraction: A < 50 ≤ B < 100 ≤ C < 250 ≤ D < 400 ≤ E < 650 ≤ F < 2000 ≤ G, VDW repulsion: A < 20.000 ≤ B < 32.000 ≤ C < 50.000 ≤ D < 100.000. All atoms corresponding to the simplex vertices were also divided into three groups: A – acceptors of potential H-bond, D – donors of potential H-bond and I – indifferent ones. On the first stage of this study, 2D QSAR model reflecting the structural inf- luence of investigated compounds on their antiviral activity against human inf- luenza (virus A/H3N2 strain Hong Kong/68) was developed. This model (Model A1) was used for interpretation, i.e., for estimation of influence of different struc- tural factors on investigated activity (on value of log IC50), but this model was not intended for prediction of activity. Model A1 was built using the PLS method with two latent variables (based on 9 descriptors in the final) and with the follow- ing statistical characteristics: determination coefficient for the training set R2 = = 0.88, coefficient of determination for cross-validation (leave-one-out) Q2 = 0.77, standard error S(ws) = 0.35, S(cv) = 0.52 for training set and cross-validation, respectively. A “randomization” procedure (Y-Scrambling) was used to confirm the “non-randomness” of the developed QSAR model.7 The statistical character- istics obtained using the Y-Scrambling procedure were lower in indices than in ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 633 the final model: R2 (Y-scr) = 0.26 ± 0.03, Q2 (Y-scr) = 0.10 ± 0.03. Thus, the non- -randomness of the established relationship between the structure of the inves- tigated compounds and their antiviral activity can be stated (see Supplementary material to this paper). The analysis of obtained Model A1 testifies that, electrostatic (atomic charges) factors and lipophilicity have the largest influence on changing of anti- viral activity (41 and 37 %, respectively). The relative influence of atom nature is 22 %. As already mentioned above, on the basis of developed QSAR models the contribution of each atom in the molecule can be calculated. The analysis of such information allows selecting different fragments which have negative or positive influence on a considered property. The sequence of mean relative influence of various substituents (R) for amantadine and rimantadine derivatives on antiviral activity (log IC50) is shown on Fig. 1. N RH N R H O NH NH NH2 O NH N NH H H O NH2 O NH OH NH2 NH (R1) < (R2) < (R3) < (R4) < O NH2 O NH2 O NH2 F O NH2 < (R5)< (R6)< (R7)< (R8) < O NH2 OH O NH2 O NH2 O NH2 < (R9) < (R10) < (R11) < (R12) Fig. 1. Averaged relative influence of substituents (R) for amantadine and rimantadine derivatives on antiinfluenza activity (log IC50) by the Model A1. In addition, the influence of identical substituents for amantadine and riman- tadine derivatives was estimated separately: a) for amantadine: R2 < R4 < R5 < R6 100). In these cases RF method was used for decision of classification tasks. For the first case since the dataset is unbalanced, i.e., the count of active and inactive molecules is significantly different, a special procedure for balance was used. The count of inactive molecules was constant (15 molecules) and the count of active ones was duplicated. In the first series, 6 active molecules (all of active and inactive – 21 molecules) were used for developed Model B1 (trees count = 100, randomly selected descriptors count = 5); in the second series the count of active molecules was increased twofold, i.e., 12 active molecules (a total of 27 ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 635 molecules) were used for developed Model B2 (trees count = 150, randomly sel- ected descriptors count = 13). The resulting QSAR models for the training set showed an unmistakable classification. The predictive ability of the QSAR models was evaluated using the “out-of-bag” (OOB) procedure. The quality of the classification models was assessed according to the following statistical char- acteristics: Matthew’s correlation coefficient (MCC), specify (SPC), accuracy (ACC) and sensitivity (SEN): ( ) ( )( )( )( ) TP TN FP FN MCC TP FP TP FN TN FP TN FN ⋅ + ⋅ = + + + + ( ) TN SPC TN FP = + ( ) ( ) TP TN ACC TP TN FP FN + = + + + ; ( ) TP SEN TP FN = + where TP – true positive, TN – true negative, FP – false positive, FN – false neg- ative. The obtained data are presented in Table II. TABLE II. Statistical parameters for classification models for CC50 (Models B1 and B2) and HNTC (Model B3) for “out-of-bag” set Model Set MCC ACC SPC SEN B1 6 + 15 = 21 0.77 0.90 0.88 1 B2 12 + 15 = 27 0.86 0.93 0.93 0.92 B3 12 + 9 = 21 0.81 0.90 0.89 0.91 As it can be seen from Table II, the balancing of models leads to a signific- ant quality improvement. Model B2 could be considered as the most appropriate. For HNTC data the duplication procedure was not used, since the dataset is more balanced (active molecules 12 and inactive 9). Model B3 (trees count = = 150, randomly selected descriptors count = 20) was also built for HNTC data- set, statistical parameters are shown in Table II. As can be seen from this table, the statistical characteristics of models B1– –B3 are quite acceptable. It should be noted that all molecules are in domain applicability (DA) of Models A1, A2 and B1–B3. CONCLUSION The realized QSAR studies have found good compliance between theoretical predictions and in vitro revealed activity and cytotoxicity. Compounds 2, 5 and 7 have both good antiviral activity against tested virus A/H3N2, strain Hong Kong/68 combined with low cytotoxicity which is in a perfect correlation with QSAR pre- ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. 636 STANKOVA et al. diction studies. All guanidated derivatives have lower activity according to both QSAR investigation and in vitro assay. Based on the obtained results we can con- clude that more bulky side chain residues are not tolerated by the mechanism of interaction with M2 ion channels of influenza virus. Unfortunately, the same con- clusion is revealed for structures creating less conformational freedom including β-amino acids as moiety in the lateral chain of amantadine and rimantadine. SUPPLEMENTARY MATERIAL Additional data are available electronically at the pages of journal website: https:// //www.shd-pub.org.rs/index.php/JSCS/index, or from the corresponding author on request. И З В О Д МОДЕЛОВАЊЕ КВАНТИТАТИВНОГ ОДНОСА СТРУКТУРА–АКТИВНОСТ ИНХИБИТОРА M2 ЈОНСКОГ КАНАЛА ВИРУСА ГРИПА IVANKA G. STANKOVA1, RADOSLAV L. CHAYROV1, MICHAELA SCHMIDTKE2, DANCHO L. DANALEV3, LIUDMILA N. OGNICHENKO4, ANATOLY G. ARTEMENKO4, VALERY A. SHAPKIN5 и VICTOR E. KUZ'MIN4 1 Department of Chemistry, South-West University “NeofitRilski”, Blagoevgrad, 2700, Bulgaria, 2 Friedrich Schiller University, Department of Virology and Antiviral Therapy, Jena, 207745, Germany, 3 University of Chemical Technology and Metallurgy, Bio-technology Department, 1756 Sofia, 8 blvd. Kliment Ohridski, Bul- garia, 4 A.V. Bogatsky Physico-Chemical Institute of Ukrainian National Academy of Sciences, Department of Molecular Structure and Chemoinformatics, 86, Lustdorfskaya doroga, Odessa, 65080, Ukraine и 4 Depart- ment of Department of Theoretical Foundations of Chemistry, Odessa National Polytechnic University, 1, Shevchenko ave., Odessa 65044, Ukraine Серија нових деривата адамантана који садрже фрагменте аминокиселина испи- тана је користећи метод симплекс представљања молекулских структура (енг. simplex representation of molecular structure, SiRMS) са циљем да се утврди однос између хемиј- ске структуре испитиваних једињења и резултата антивиралне и цитотоксичне актив- ности. Добијени резултати QSAR анализе показују да ће деривати адамантана који са- држе аминокиселине кратких неполарних алифатичних група у бочном низу имати добре антивиралне активности према A/H3N2 соју Hong Kong/68 вируса као и малу цитотоксичност. QSAR експерименти и резултати in vitro активности показују добру корелацију и указују да деривати који садрже гуанидо фрагмент у бочном низу и β-ами- нокиселину као супституент, имају малу или не показују активност. (Примљено 9. маја 2020, ревидирано 15. марта, прихваћено 6. маја 2021) REFERENCES 1. R. L. Tominack, F. G. Hayden, Infect. Dis. Clin. North. Am. 1 (1987) 459. 2. R. B. Belshe, M. H. Smith, C. B. Hall, R. Betts, A. J. Hay, J. Virol. 62 (1988) 1508 (https://jvi.asm.org/content/62/5/1508) 3. J. Wang, C. Ma, J. Wang, H. Jo, B. Canturk, G. Fiorin, W. F. DeGrado, J. Med. Chem. 56 (2013) 2804 (https://doi.org/10.1021/jm301538e) 4. R. M. Pielak, J. J. Chou, Biochem. Biophys. Res. Commun. 401 (2010) 58 (https://doi.org/10.1016/j.bbrc.2010.09.008) 5. J. Wang, J. X. Qiu, C. Soto, W. F. DeGrado, Curr. Opin. Struct. Biol. 21 (2011) 68 (https://doi.org/10.1016/j.sbi.2010.12.002) ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. M2 ION CHANNELS INHIBITORS – QSAR STUDY 637 6. A. V. Gaiday, I. A. Levandovskiy, K. G. Byler, T. E. Shubina, in Proceedings of International Conference on Computational Science, 2008, Berlin, Germany, pp. 360–368 (https://doi.org/10.1007/978-3-540-69387-1_40) 7. V. E. Kuz’min, A. G. Artemenko, E. N. Muratov, J. Computer-Aided Molec. Des. 22 (2008) 403 (https://doi.org/10.1007/s10822-008-9211-x) 8. V. E. Kuz’min, A. G. Artemenko, E. N. Muratov, P. G. Polischuk, L. N. Ognichenko, A. V. Liahovsky, A. I. Hromov, E. V. Varlamova, Recent Advances in QSAR Studies: Methods and Applications,Springer, Dordrecht, 2010, p.127 (ISBN 978-1-4020-9783-6) 9. S. Rännar, F. Lindgren, P. Geladi, S. Wold, J. Chemometrics 8 (1994) 111 (https://doi.org/10.1002/cem.1180080204) 10. L. Breiman, Machine Learning 45 (2001) 5 (https://doi.org/10.1023/A:1010933404324) 11. R. E. Carhart, D. H. Smith, R. Venkataraghavan, J. Chem. Inform. Comp. Sci. 25 (1985) 64 (https://doi.org/10.1021/ci00046a002) 12. K. Hasegawa, Y. Miyashita, K. Funatsu, J. Chem. Inform. Comp. Sci. 37 (1997) 306 (https://doi.org/10.1021/ci960047x) 13. V. E. Kuz’min, A. G. Artemenko, P. G. Polischuk, E. N. Muratov, A. I. Hromov, A. V. Liahovskiy, S. A. Andronati, S. Y. Makan, J. Mol. Model 11 (2005) 457 (http://doi.org/10.1007/s00894-005-0237-x) 14. OECD (2014) Guidance Document on the Validation of (Quantitative) Structure-Activity Relationship [(Q)SAR] Models, OECD Series on Testing and Assessment, OECD Publishing, Paris, 2014, p.154 (http://doi.org/10.1787/9789264085442-en) 15. M. Meloun, J. Militku, M. Hill, Analyst 127 (2002) 433 (http://doi.org/10.1039/B110779H) 16. P. G. Polischuk, E. N. Muratov, A. G. Artemenko, O. G. Kolumbin, N. N. Muratov, V. E. Kuz'min, J. Chem. Inf. Mod. 49 (2009) 2481 (http://doi.org/10.1021/ci900203n) 17. R. Chayrov, N. A. Parisis, M. V. Chatziathanasiadou, E. Vrontaki, K. Moschovou, G. Melagraki, H. Sbirkova-Dimitrova, B. Shivachev, M. Schmidtke, Y. Mitrev, M. Sticha, T. Mavromoustakos, A. G. Tzakos, I. Stankova, Molecules 25 (2020) 3989 (https://doi.org/10.3390/molecules25173989) 18. A. Chintakrindi, Ch. D'souza, M. Kanyalkar, Mini Rev. Med. Chem. 12 (2012) 1273 (https://doi.org/10.2174/138955712802761997). ________________________________________________________________________________________________________________________Available on line at www.shd.org.rs/JSCS/ (CC) 2021 SCS. << /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.4 /CompressObjects /Tags /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /CMYK /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams false /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo false /PreserveOPIComments true /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 300 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 300 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile () /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false /Description << /ARA /BGR /CHS /CHT /CZE /DAN /DEU /ESP /ETI /FRA /GRE /HEB /HRV (Za stvaranje Adobe PDF dokumenata najpogodnijih za visokokvalitetni ispis prije tiskanja koristite ove postavke. Stvoreni PDF dokumenti mogu se otvoriti Acrobat i Adobe Reader 5.0 i kasnijim verzijama.) /HUN /ITA /JPN /KOR /LTH /LVI /NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken die zijn geoptimaliseerd voor prepress-afdrukken van hoge kwaliteit. De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 5.0 en hoger.) /NOR /POL /PTB /RUM /RUS /SKY /SLV /SUO /SVE /TUR /UKR /ENU (Use these settings to create Adobe PDF documents best suited for high-quality prepress printing. Created PDF documents can be opened with Acrobat and Adobe Reader 5.0 and later.) >> /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ << /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >> << /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /ConvertColors /ConvertToCMYK /DestinationProfileName () /DestinationProfileSelector /DocumentCMYK /Downsample16BitImages true /FlattenerPreset << /PresetSelector /MediumResolution >> /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /DocumentCMYK /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /UseDocumentProfile /UseDocumentBleed false >> ] >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice