Indian Journal of Chemistry Vol. 428 , June 2003, pp. 1485-1496

QSAR of adenosine receptor antagonists I : Exploration of receptor interaction sites of 1 ,2-dihydro-2-phenyl-I ,2,4-triazolo[ 4,3-a]quinoxalin-I-one derivatives

using AMI calculations

Kun a l Roy

Drug Thcoretics and Chcminform3t ics Lab. Division of Medicinal and Pharmaccutical Chemistry. Department or Pharmaceutical Tcchnology. Jadavpur Uni versity, Kolkat<l 700 032. India

Email : kunalroy_in @yahoo.com

URL : hllp ://www.geoci ties.com/kunalroy_in

Received 5 August 2002; accepted (revised) 24 December 2002

QSAR of adenosine reccptor antagonist I ,2-dihydro-2-phenyl- 1 ,2,4-tri azoloI4,J-a lquinoxalin- I-one deri vat i ves has becn attempted to explore using AM J calculations with an objective to identify different interacti on sites on the ligands for different receptor sUbtypes [A I> A2A and A)J. Five possible interaction sites have been suggested and it is observed that the requirements fo r optimum binding for these receptor sUbtypes (AI> A2A and A)) are different to some ex tent.

Adenosine is an endogenous nucleoside that modu­lates a wide variety of physiological effects through activation of G-protein coupled AI, A2A, A2B and A3 receptors 1.2. Presence of the receptors on basically every cell s makes them an interesting target fo r the pharmacological intervention in many pathophysi­ological conditions2. The sy nthesis of agonists and antagoni sts to the adenosine receptors and their clon­ing has enabled the exploration of their physiological function s.

Adenosine shows general inhibitory effect on neu­ronal acti viti, which is med iated by AI receptors4

.

However, excitatory actions of adenosine, medi ated by acti vation of A2A receptors, have also been demon­strated5

. Evidences suggest the ro le of adenosi ne in a diverse array of neural phenomena including regula­tion of sleep and the level of arousal, neuroprotection, regul ation of seizure susceptibility, locomotor effects, analgesia. mediat ion of effects of ethanol etc4

. The somnogenic action of adenosine in the basal fo rebrain is mediated by AI receptors and its expression might be regulated by induction in the nuclear factor kappaS protein as its tran cription factor6

. Increase of adenosinergic neuromodul ation may be one of the several future therapeuti c strategies in neuroprotec­tion] Selective agoni sts of adenos ine AI receptor are shown to have neuroprotective ac ti on in brain ische­mi a6

.s, while A2A receptor antagoni sts (like KW-6002)

are being developed for treatment of neurodegenera-

ti ve diseases like parkinsoni sm9.lo. AI receptor medi­

ated mechani sm for protection against myocardi al stunning is under investigation II . AI receptor antago­nists may provide a therapeutic tool to prevent con­trast-media induced acute renal failure in patients with diabetes mellitus l2.

The A2A receptors are thought to playa ro le in a number of physiological and pathological conditions. A2A receptors show antagoni sti stic interac tions with d . (D ) 13 14 A . opamlI1e 2 receptors '. 2A receptor agonl sts may be useful for treatment of certain type of seizures or sleep disorders 13. A2A receptor has become a focus of major interest in the treatment of inflammatory re­nal injury because of its ability to broadly inacti vate inflammatory cascade.

Only recently, A2B receptors have been implicated in the regul ation of vascular smooth muscle tone, cell growth, intestinal function and neurosecreti on. The ro le of A2B receptors in mas t cell acti vati on makes it a

I h . . h 15 nove t erapeutlc target In ast ma · . Acti vati on of A3 receptors has been shown to

stimulate phosphilipase C and 0 and inhibit adenylate cyclase. It also causes release of infl ammatory media­tors like hi stamine from mast ce ll s. A3 receptors play important role in brain ischemia, immunosuppression and bronchospasm. Selective A3 agoni sts andlor an­tagoni sts have been indicated as potenti al drugs for the treatment of asthma and infl ammation while selec­ti ve agonists have been shown to have card ioprotec

1486 INDIAN J. CHEM., SEC B, JUNE 2003

live action l. A} receptors playa key role in the inhibi­tory and stimulatory growth activities of adenosine. A3 receptor agonists possess dual activity: antiprolif­erative activity towards tumor cells and stimulatory effect on bone marrow cell s. Activation of A3 recep­tors may serve as a new approach for cancer ther­apyl6

Accumulated evidences indicate a widening role of adenosine receptors in many therapeutic areas like immunological and inflammatory responses, respira­tory regulation, renal protection, parkinsonism, car-:I . . t'f I ry 16· 18 I I I ( loprotectl ve e ects, cancer etc .-. . n t 1e ast two decades, many efforts have been made to develop se­lective adenosine receptor ligands for their potential

. . 9 70 therapeutic uses '- . In this present effort, QSAR of adenosine receptor

antagonist 1 ,2-dihydro-2-phenyl-l ,2,4-triazolo[ 4.3-a]­qui noxalin-l-one deri vati ves, recently reported by Colotta et 01.

21, has been attempted to explore using

AMI calculations with an objective to identify differ­ent in teraction sites on the li gands for different recep­tor subtypes [AI , A2A and A3]. The biological data consist of the binding affinity of the compounds at bov ine A I and A2A and human cloned A, receptors (there is species difference in A3 primary amino acid sequence while for AI and A2A adenosine receptor subtypes there is a good amino acid sequence homo­log/I).

Materials and Methods The biological activity values and structural fea­

tures of the compounds are presented in Table I. Compounds without quantitative biological activity data in a particular se ries (All A2P..! A,) were exc luded from the study . One compound appearing in the origi­nal data set was not included in the present study be­cause of absence of the oxo subst ituent in the triazole ring, a feature that is present in all other compounds. Quantum mechanical calculations were done accord­ing to AM I (A ustin Model 1)22-24 method using Chem 3D Pr025 package. The general structure of the com­pounds (Figure 1) was drawn in Chel1l Draw Ultro lIe r 5.025 and it was saved as the templ ate struct ure. For every compound, the template structure was suitably changed considering its structural features, copied to Chell1 3D ver 5.025 to create the 3-D model and finally the model was 'c leaned up'. The non­hydrogen common atoms of the compounds were given a serial number so that these maintain same se­rials in all the models (Figure 1). Next, energy mini­mizati on was done under MOPAC module usi ng RHF

A

, , '- ___ __ __ _ _______ _____ .J

E

Figure li - General struclure of 1,2-dihydro-2-phcnyl-I .2,4-Iri azoloI4,3-a]quinoxalin-l -one derivatives: the common atoms have been numbered I through 20 ( it has no re lat ion to the chemi­cal nomenclature . ystcm) and important intcraction site~ A through E have been suggested.

(restricted Hartree-Fock : closed shell) wave function. The energy minimized geometry was used for calcula­tion of Wang-Ford charges (obtained from molecular electrostatic potential surface) of different atoms.

The charges (qx) of different atoms (x) were sub­jected to intercorrelation study. The biological activity data of the compounds [pK().1M)] were subjected to regression with the charges of di fferent common atoms and also their26 different combinations to obtain the best relat ions using the program A UTOREC27 de­veloped by the author. Some indica tor variables, which emerged as important descriptors from the pre­liminary analyses, were retained du ri ng finding out the best rel ati ons involving differem combinations of charge parameters in some cases (vide infra ). For the relations havin g more than one predictor variables. only those variables with less intercorrelation were cons idered (a cutoff intercorrelation val ue of 0.7 was considered only as a screening process) . Further, rela­tions having correlation coefficients higher than those of the best equations involving less number of predic­tor variables were only recorded.

The regression analyses were carried out using a GW-BASIC program RRR9827

. The stati stical quality of the eq uati ons28 was judged by the parameters like explained variance (R}, i.e., adjusted R 2), correlation coefficiel1l (r or R), standard error of estimate (s), average of absolute values of the residuals (A VRES), variance ratio (F) at specified degree of f reedom (df) and 't' values of the regression coefficienrs. Use of more than one variable in a multi variate eq uati on was justified by intercorrelation study. All the accepted

ROY : QSAR OF ADENOSINE RECEPTOR ANTAGONISTS I 1487

Table ( - Adenos ine receptor binding acti vity of I ,2-dihydro-2- phenyl- 1 ,2.4-triazolo[4,3a Jquinoxali n- l -ones

Compd R R/X

2

3

4

5

6

7

8

9

10

1\

12

13

14

15

16

17

18

19

20

21

22

23 24

25

26

27

28

29

30

3 1

32 33

34

35

36

*nM

H

H

H

3-CHJ

4-CH3

3-F

4-OCH3

4-CI

H

3-CH,

3-F

4-CI

H

3-CH3

3-F

H

H

H

H

H

H

H

H

H

H

H

3-CH,

H NHBn

H NH,

H H

H H

H H

H H

H H

H H

cyclo-hexyl H

cyclo-hexyl H

cyclo-hexyl H

cyclo-hexyl H

cyclo-pentyl H

cyclo-pentyl H

cyclo-pentyl H

CH,Ph H

(CH,),Ph H

(CH,hPh H

CH,CH(Ph), H

COCH3 H

COCH2CHJ H

CO Ph H

COCH, Ph H

CONHPh H

CONHC6H..-4- H OCH3 H H

H H

4-CH, H H

H

H

H

H

H

H

H

H

3-F H

4-OCH, H

4-CI H

H CH,

4-CH, CH3 4-CI CH3 H n-CJH7

H CH,C::CH

Bn = Benzyl, pK = - log[K(~M)]

1 - 26

Binding affinity for different receptor subtypes Bov ine A, Bovine A2A Human A3

K;* pKAI Calc Res K;* pKA2A Calc Res K;* pKA3 Calc Res

730

9.2

11.0

20.0

19.5

28.5

312.0

426.0

1.43

4.2

4.9

80. 1

0.42

1.2 1

1.1

55.0

4.8

17.9

4.3

9.3

89.6

6.3

0. 137 0.473 -0.336

2.036 1.606 0.430

1.959 1.3 16 0.643

1.699

1.71 0

1.545

0 .506

0.37 1

1.353 0.346

1.652 0.058

1.850 - 0 .305

1.2 19 -0.7 13

0.299 -0.072

6.5

18.7

49.0

14.6

85.8

72.0

376.0

2 .187

1.728

1.31 0

1.836

1.067

1.1 43

0 .425

2.074 0 .11 3

1.806 -0.078

0.794 0.51 6

1.222 0.6 14

0.870 0.197

1.784 -0.641

0.447 -0.022

2.845 2.46 1 0.384 1370.0 -0. 137 -0. 177 0.040

2.377 2.694 - 0.3 17

2.3 10 3.0 13 -0.703

1.096 1.874 -0.778

3.377 2.440 0.937

2.9 17 2.486 0.43 1

2.959 2.9 13 0.046

1.260 1.267 -0.007

2.3 19 1.557 0.762

1.747 1. 809 -0.062

2.367 2.040 0.327

66.0 1.1 8 0.742 0 .438

986.0 0.006 -0.136 0. 142

1400.0 -0. 146 0.238 -0.384

148.0 0 .830 0 .798 0.032

2.032 2.4 12 - 0.380 28 18.0 -0.450 -0.695 0.245

1.048 1.475 -0.427

2.20 1 2. 147 0.054

54.0 1.268 0.89 1 0.377

28.5

48.3

157.0

45.3

329

1.545

1.3 16

0.804

1.344

0.483

506 0.296

548 0.26 1

44.2

56. 1

55.4

1.355

1.25 1

1.256

27.5 1.56 1

173 0.762

1700.0 -0.230

1.226 0.319

1.274 0.042

1.130 -0.326

1.226 0.1 18

1.058 -0.575

1.1 78 0. 177

1.1 06 0.145

1.226 0.030

1.250 0.3 11

1.154 -0.392

20 1.0 0.697 1.1 54 -0.457

40.9 1.388 1.1 78 0.2 10

1020.0 -0.009

2.0 2.699 2.482 0.2 17

15.8 1.80 I 2.434 -0.633

1.47 2.833 2.458 0.375

3.75 2.426 2.386 0.040

50.8 1.294 1.1 52 0 .1 42 2300.0 -0.362 0 .437 -0.799 276.0 0.559 1.034 - 0.475

0.0 18 0.747 -0.729 2600.0 -0.4 15 0.776 - 1.1 9 1 2800.0 -0.447 -0.034 -0.4 13 960.0

5 15 .0

436.0

0.288 0.40 I -0.1 13

0.36 1 0.275 0.086

155.0 0.8 10 0 .5 11 0.299

200.0 0.699 0 .905 - 0.206

934.0 0 .03 0.200 -0. 170

10 15.0 -0.006 -0.236 0.230

309.0 0 .510 0.20 I 0 .309

406.0 0.39 1 0 .090 0.30 I

2200.0 -0.342 0.058 -0.400

80.0

9 1.0

25.0

63.0

16.0

11 4.0

36.6

504.0

137.0

1.097 1.130 -0.033

1.04 1 1.1 78 -0. 137

1.602 1.202 0.400

1.20 I 1.058 0. 143

1.796 1.1 54 0.642

0.943 1.034 - 0.09 1

1.437 1.1 30 0.307

0.298 0.598. - 0.300

0.863 0.334 0.529

1246.0 -0.096 -0.3 13 0.21 7

479.0 0.320 0.766 -0.446

Calc = Calculated values; Res = Observed (pC I/pC, /pC.,) - Calculated Binding affiniti es (A/ A'A / A, ) have been calculated according to Eqs. 9/ 19/27.

1488 INDIAN J. CHEM .. SEC B. JUNE 2003

equations have regression constants and F ratios sig­nificant at 95 % and 99% levels respectively , if not stated otherwise. A compound was considered as out­lier if residual is more than twice the standard error of estimate for a particular equation . In case that inter­cept of an equation was statistica lly insignificant and omission of the sa me did not affect the quality of the equation. exclusion of the intercept gave stati stically more acceptable equation.

Results and Discussion The charge parameters of different co ml1lon atoms

of the compounds are given in Table II . Stepwise development of QSAR equati ons is shown in Table III . Stati stical quality of the binding affinities with different combinations of charge parameters with or without indicator parameters is shown in Tables IV­VI. The Discussion part uses arbitrary numbering as­signed to the common atoms of the co mpounds (Figu"e 1), which has no relation to the chemi ca l no­menc la ture system. Energy minimized geometri es of the most active members in the indi vidual series [(a) : compound 13 (A I), (b) : compound 1 (A2A), (c): compound 22 (A ~ ) I are shown in Figure 2.

QSA R of A I receplOr binding qlfinity Table III shows that indicator va ri able I denoting

resence or absence of cyclopentylamino or cyclo­hex ylam ino group at RI positi on ca n exp lain 34.9% variance of the AI bindin g affinity (Eq. 1) whi ch is increased to 47.0% when another indicator variable 1m denoting presence or absence of - C H2CO group (as a part of alkylamino group) at RI position is used (Eq. 2) . These two predictor va ri ab les (I and Ico) were retained when the bi olog ica l act ivity was correlated with different combinati ons of charge parameters of different common atoms (Table IV). The atom 8 (Cs)

merged as sin gle best charge parameter (Eq. 3). The best equati ons with hi gher number of charge para me-

ters invo lved atoms C8 & C I9 (Eq. 4), CR, C I6 and C I9 (Eq. 5) and CR, N13• C I6 and C I9 CEq. 6). However, atoms C I6 and C I9 are considerabiy autocorrelated (Table VII) . Hence, a parameter q16+ 11J was defined as sum of charge parameters of atom C I6 and Cl l) and the resultant eq uation (Eq. 7) was statistically similar to Eq. 6. Replacement of the parameter q8 in Eq. 7 with q7 yields an equation (Eq. 8) with similar statisti­cal quality. Atoms Cs and N7 are highly autocorre­lated (Table VII) and the charge values are of oppo­site sign . A new parameter qS-7 was defined to denote diffe rence between the charge parameter va lues of atoms Cs and 7 and Eq. 9 was obtained. Remov ing poss ibl outliers (co mpounds 25, and 2S & 13). Equa­tions 10 and II were obtained. The ca lculated binding affi nity va lues according to Eq. 9 are given in Table I. Autocorrelation among the important predictor vari ables is presented in Table VII.

From the results it appears tha t presence of cyclopentylamino or cyclohexylamino group at RI pos ition and a CH }CO fragment as a part of al­kylami no substituent at RI is conducive to the adeno­sine AI receptor binding affinity. The atom N7 (nitro­gen) with its adj acent (o rtho-) oxo or exocy li c amino substituent pl ays important rol e in the AI binding af­finity. Futher. atom 13 (one of the triazo le nit rogens) also modul ate significantly the binding affinity. The atoms CI6 and CI9 signify the importance of the ben­zene ring and substituents on it.

QSAR of AlA reap/or hinding (!/Ji ni')' From Tables III and V it is ev ident that charge pa­

rameter of atom N IJ emerged as si ngle best parameter CEq . 12) for A2A receptor binding affinit y. The best bivariate relati ons in volved atoms NIJ & C I and IJ & C2 while the best tri va ri ate relation invo lved atoms CI, NI2 and C I6 (the qu ality of the tri variate re lations in­volving charge parameters of atoms CI, NIJ & CI7 and C2• NI 2 & C I6 and C2, NIJ & C 17 closely foll ow that or

Figure 2 - Encrgy min illli zcd gco lllc iri c, o f the most ac ti vc Illcmbcrs in the indIv idual seri es I (a) : co mpound 13 (AI), (b) : cOlllpolind I I A lA) . (cl : compoll nd 22 (A.) I arc shown

Table II - Wang-Ford charges (q,) of the common atoms (x) of 1.2-dihydro-2-phenyl-I.2.4-triazolO[4.3-a)quinoxalin- l -one derivatives

Compd Atoms (x} no. 2 3 4 5 . 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20

2 3 4 5 6 7

8 9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

0.382 -D.27 I -D.074 -D.222 0.155 -D.064 -D.393 0.513 0.308 -D.359 0.808 -D.390 -D.330 -D.534 0.361 -D.249 -D.080 -D.130 -D.I08 -D.152 0.439 -D.351 -D.030 -D.193 -D.09O 0.168 -D. 64 I 0.6920.142 -D.150 0.778 -D.422 -D.302 .-0.5310.424 -0.290 -D.06O -0.132 -D.III -D.174

-D.159 -D.I09 -D.135 -D.105 -D.Oll 0.327 -D.667 0.709 0.178 -D.330 0.835 -D.378 -D.311 -D.543 0.388 -D.280 -D.055 -D.145 -D.087 -D 183 -D.161 -D.I08 -D.l34 -D.I09 -D.005 0.327 -D.668 0.711 0.171 -D.335 0.837 -D.365 -D.307 -0.545 0.380 -D.303 0.100 -0.201 -D.Q63 -0.205 -D.158 -D. II O -D.l33 -D.112 0.001 0.322 -0.664 0.703 0.187 -D.345 0.843 -0.371 -D.3 15 -D.547 0.355 -D.275 -D.078 0.017 -D. I44 -D.153 -D. 159 -0.109 -D.135 -=D.I06 0.000 0.324 -D.662 0.694 0.207 -0.357 0.851 -0.374 -0.322 -0.541 0.4ll -0.385 0.286 -0.236 -0.072 -0.194 -D.157 -O. lll -0.131 -O.1l9 0.012 0.318 -0.664 0.705 0.180 -0.347 0.831 -0.348 -0.305 -0.545 0.251 -D.146 -D.254 0.313 -0.201 -D.l50 -0.160 -0.108 -D.141 -D.091 -0.027 0.333 -0.666 0.695 0.203 -0.329 0.835 -0.392 -0.324 -0.538 0.446 -0.351 -0.107 -0.265 0.053 -0.2 16 -0.136 -0.117 .0.126 -0.139 0.081 0.260 -0.612 0.642 0.261 -0.436 0.879 -0.432 -0.261 -0.545 0.380 -0.265 -0.054 -0.149 -0.081 -0.184 -D.137 -D.117 -0.124 -0.142 0.090 0.257 -0.610 0.640 0.264 -0.447 0.887 -0.429 -0.261 -0.545 0.389 -0.309 0.1l0 -0.190 -0.069 -0.207 -D.135 -0.1 18 -0.123 -D.I44 0.092 0.257 -O.6ll 0.645 0.270 -D.454 0.895 -D.428 -0.264 -0.543 0.400 -0.367 0.289 -0.245 -0.057 -0.202 -D.136 -D.117 -0.129 -0.130 0.067 0.268 -0.618 0.650 0.262 -D.427 0.880 -0.445 -0.265 -0.540 0.438 -0.338 0.109 -0.269 0.068 -0.228 -D.133 -D.118 -0.122 -0. 149 0.103 0.239 -0.582 0.611 0.277 -0.4510.884 -D.435 -0.263 -0.5450.390 -0.276 -0.052 -0.149 -0.080 -0.188 -0.136 -O.1l6 -0.122 -0.146 0.096 0.244 -0.585 0.619 0.257 -D.439 0.879 -D.422 -D.256 -0.546 0.382 -0.296 0.102 -0.204 -0.055 -0.213 -0.133 -D.1l8 -0.121 -D.149 0.102 0.242 -0.582 0.613 0.277 -0.453 0.888 -D.425 -0.265 -0.542 0.405 -0.368 0.289 -0.246 -0.054 -0.208 -D.122 -0.121 -0.120 -D.136 0.048 0.242 -0.600 0.673 0.161 -0.337 0.820 -D.360 -0.298 -0.540 0.377 -0.283 ·-D.050 -0.146 -0.077 -D.189 -0.1 22 -0.126 -0. 122 -D.129 0.042 . 0.268 -0.637 0.710 .. 0.168 -D.366 0.849 -0.392 -D.267 -0.542 0.363 -D.26 I -D.059 -D.I44 -0.084 -D.179 -D.141 -0.120 -D.129 -0.123 0.027 0.307 -D.67 I 0.725 0.173 -0.373 0.863 -D.412 -0.253 -D.543 0.375 -0.271 -0.055 -0.147 -D.081 -0.182 -0.1 55 -0.116 -D.137 -0.109 -D.003 0.346 -D.699 0.724 0.173 -0.3560.856 -0.414 -0.244 -D.542 0.358 -0.25 1 -0.067 -D.143 -D.09O -0.162 -D. I08 -D.129 -D.I04 -D. 158 0.097 0.228 -D.530 0.618 0.211 -D.407 0.865 -D.413 -D.260 -D.540 0.373 -D.252 -D.063 -D.I44 -D.073 -D.201 -D.119 -D.122 -D.I06 -D.153 0.080 0.239 -D.531 0.628 0.178 -D.383 0.863 -D.438 -D.213 -D.538 0.364 -D.245 -0.062 -D.156 -D.065 -0.195 -D.125 -0.122 -D.Ill -0_150 0.078 0.253 -D.535 0.605 0.194 -D.389 0.862 -D.427 -D.234 -D.539 0.371 -D.250 -0.067 -D.145 -D.074 -D.194 -D.123 -D.120 -D.11O -0.150 0.071 0.247 -D.530 0.6180.176 -D.370 0.853 -0.424 -D.224 -0.536 0.355 -D.232 -D.068 -D.157 -D.059 -D.200 -D.132 -O.1l8 -D.113 -0.136 0.04 1 0.284 -D.554 0.6070.210 -D.354 0.841 -0.404 -D.292 -0.537 0.394 -D.27 I -D.064 -D.138 -D.084 -D.193 -D.155 -0.092 -0.125 -0.122 0.013 0.284 -D.496 0.496 0.201 -0.288 0.786 -0.388 -D.256 ...:0.525 0.348 -D.216 -0.092 -D.136 -D.068 -0.199 -D.043 -0.158 -D.064 -0.218 0.347 -0.135 -D.285 0.415 0.383 -D.602 0.856 -0.297 -0.285 -0.541 0.265 -D.214 -0.089 -0.112 -0.107 -0.128 -0.044 -D.158 -D.063 -0.223 0.358 -0.137 -D.286 0.424 0.369 -D.606 0.851 -0.267 -D.281 -D.543 0.234 -0.215 0.060 -0.166 -D.082 -D.145

-D.04O -0.160 -D.06O -D.226 -D.038 -0.161 -D.06O -D.226 -D.034 -0.163 -0.057 -D.235 -0.044 -0.157 -D.069 -0.209

0.032 -0.192 -0.051 -D.227 0.D35 -D.193 -0.048 -D.230 O.oz8 -0.186 -0.062 -0.209 0.002 -0. 184 -0.065 -0.204 0.067 -0.207 -0.043 -0.249

0.362 -D.I44 -0.280 0.413 0.389 -D.617 0.861 -D.283 -D.288 -0.544 0.223 -D.189 -D.125 0.050 -D.152 -D. 107 0.367 -D.149 -D.275 0.405 0.404 -D.625 0.865 -0.286 -D.296 -D.538 0.285 -D.316 0.258 -0.210 -0.09 1 -D.137 0.376 -D.157 -0.274 0.410 0.393 -D.6 18 0.844 -0.273 -D.292 -D.542 0.203 -D.129 -D.257 0.304 -0.177 -D.152 0.336 -D.131 -0.288 0.410 0.394 -D.595 0.850 -0.299 -D.295 -0.537 0.317 -D.283 0.074 -0.240 0.034 -D. I64 0.386 -D.225 -0.093 0.352 0.447 -D.650 0.871 -0.326 -D.301 -0.54 1 0.344 -D.27 I -D.052 -D.143 -D.087 -D.168 0.397 -D.234 -0.086 0.342 0.461 -0.666 0.878 -0.317 , -D.305 -0.543 0.306 -D.256 -D.087 0.020 -D.137 -D.142 0.366 -0.2 13 -0.100 0.356 0.448 -0.633 0.856 -0.330 -0.305 -0.532 0.388 -0.330 0.104 -0.259 0.058 -0.206 0.365 -0.161 -0.198 0.354 0.519 -0.668 0.838 -0.260 -0.364 -0.505 0.239 -0.277 -0.016 -0.1 36 -0.133 -0.074 0.441 -0.263 -0.032 0.335 0.503 -0.744 0.935 -0.361 -0.305 -0.550 0.35 1 -0.273 -0.042 -0.153 -0.084 -0.163

;:<:l o -< 10 C/J » ;:<:l

o 'T]

» o tTl Z o C/J

Z tTl ;:<:l tTl n tTl

~ o ;:<:l

» ~ » o o z C/J -l C/J

+>-00 \0

1490 INDIAN J. CHEM ., SEC B, JUNE 2003

Table 111 - Stepwise development of QSAR equations Model equat ion pKj = I~i Xi + ex

Key Eq. Regress ion Constants (r. c.) Statistics activity no. (pK) (DC) a p, p, ~, p, ps P6 R.,' F AVRES

(s.e.) (s.c.) (s.e.) (s.e.) (s.e.) (s.e .) (s. e.) (ror R) (5) (n)

+1.021 + 1.533 I 0.349 18.1 0.7 16 (0. 166) (0.360) (0.607) (0.846) (33)

2 +0.868 +1.687 I +1 .332 Ico 0.470 15.2 0.605 (0. 159) (0.329) (0.468) 0.709 (0.763) (3)

3 - 1.343 +1.38 1 I + 1.0661co +3.986 q, 0.69 1 24 .9 0.4 17

(0.500) (0.259) (0.362) (0.840) (0. 49) (0.582) (33)

4 - 1.802 +1.548 I +1.1 58 leo +4. 142 qs -4. 12 1 q,. 0.726 22.2 0.387

(0.496) (0.256) (0343) (0.794) ( 1.900) (0. 72) (0548) (33)

5 -3.20 1 +1.422 1 + 1.370 leo +3.595 q, - 5.450 q ,. -7 .285 q l" 0.766 2 1.9 0.377

(0.790) (0.243) (0 .330) (0.770) (2 .288) (2.204) (0.896) (0.507) (33)

A, 6 +1.0161 +0.735 leo +2.884 q, +1 1.254 qu -7.200 q ,o -8. 195 q ,., 0.76'1 58.0 0.369

(0.27 1) (0.329) (0.73 1) (2 .579) (2.579) (2.32 1) (0. 97) (0.503) (33)

7 +0.962 I +0.725 leo +2.795 q, +1 1.555 q" -7.85 1 q' fH '. 0 .776 7 1.5 0.373

(0.244) (0.324) (0698) (2 .47 1 ) (2. 183) (0.896) (0.496) (33)

8 +0.945 I +0.893 leo - 1.974 q7 + 10.047 qu -8.406 q 'h.'~ 0.771 fl9 .9 0.369

(0.248) (0.3 16) (0.507) (2 .465) (2. 164) (0.894) (0.50 1) (33)

9 +0.952 I +0.808 leo + 1.233 q.-7 +9.989 q u -7.428 q' 6.,~ 0.776 7 1.6 0.363

(0245) (03 18) (0.289) (2.273) ( 1.961) (0.897) (0.496) (33)

10 +0.913 I +0.775 Ico + 1.245 q.-7 +10. 128 qu - 7.594 q '6.,'i 0.80 1 87.2 0.344

(25) (0.224) (0.290) (0.269) (2.240) ( 1.980) (0.909) (0.45 3) (32)

II +0.746 I +0.768 leo + 1.253 q.-7 +10.330 q'J - 7.74 1 q".,., 0.:514 9 1.2 0.325

(13,25) (0.2 17) (0266) (0.248) (2.06 1) ( 1.82 1) (0.9 16) (0.4 16) (3 1)

12 - 5.076 - 20.250 q u 0.5 10 15.6 0.50K

(1.466) (5. 129) (0.73 ) (0.627) (15)

13 -4. 111 + 1.502 q, * - 17 .229 tt " 0.5 I 10.7 0 .391

( 1.459) (0.840) (5.035) (0. 0 1) (0.5g0) ( 15)

14 -4.826 -4. 134 q2' - 17.34fl q" 0.584 10.8 0.405

( 1.358) (2.273) (4.989) (0.802) (0.578) ( 15)

A 2A 15 6.369 +3.050 q , +19.623 q" - 8.500 q ", 0.737 14. 1 0 .326

( 1.687) (0.637) (4.43 1 ) (2. 140) (0.89 1) (0.459) ( 15)

16 -4.170 + 1.832 q , - 17.410 q u +2.043 q l7 0.712 12.5 0.3 19

( 1.1 95) (0.709) (4. 178) (0.805) (0. 79) (0.48 1) ( 15)

17 5.255 - 8.236 q, +19.997 q' 2 - 8. 158 q'h 0.723 13.2 0.338

( 1.756) ( 1.787) (4.562) (2. 188) (0. 8 ) (0.47 1 ) ( 15)

18 - 5.042 -4.909 q, - 17 .627 q 'J 1.995 q l7 0.70 12.3 0.329

( 1.224) ( 1.929) (4. 179) (0.807) (0.87 °) (0.484) (15)

19 1. 173 + 1.49 1 q, .3-2 .... +10.154q l2_'J - 1.873 q "~'7 0.735 13.9 0.3 12

(0.307) (0.388) (2. 16 1) (0.573) (0.890) (0.461) ( 15)

(s.e.) (s.e.) (s.e.) (s.e.) (s.e.) (s.e .) (s.c.) (I' or R) (5) (n)

20 1.11 2 +1.327 1'co 0.553 33.2 0.334

(0.089) (0.230) (0.755) (0.425) (27)

2 1 -27.205 +1.437 I 'eo - 52.407 q l4 0.710 32.8 0.267

(7.435) (0 .188) (I 3.759) (0.856) (0 .343) (27)

22 -37.256 + 1.555 I'eo +1.846 q , -7 1.268 q,. 0.07 37.1 0.2 18

(6.701) (0. 157) (0.5 12) ( 12.397) (0.910) (0.280) (27)

A, 23 -36.387 +1.554 1'eo -4.347 q, - 68.254 q '4 O. 07 37.2 0.2 17

(6.5 15) (0. 157) ( 1.202) ( 12.050) (0.91 1) (0280) (27)

24 - 37.004 +1.554 1'eo + 1.297 q '-2 - 70.383 q" 0.807 37.2 0.2 17

(6.64 1) (0 .157) (0.359) ( 12.286) (0.9 11) (0.280) (27)

25 - 36.773 + 1.760 I'cu + 1.290 q' -2 - 69.956 q ,. 0.85 5 1.2 0. 189

(21 ) (5.744) (0 . 152) (0.3 11 ) ( 10.629) (0.93. ) (0.242) (26)

(- Coilld)

Key Eq.

ROY : QSAR OF ADENOSINE RECEPTOR ANTAGONISTS I

Table III - Stepwise development of QSAR equations (- Collld)

Model equation pKj = L~i Xi + a

Regress ion Constant s (r. c.)

1491

Statistics activi ty no. (pK) (DC) a

(s.c.)

26 - 35.494 (2 1,30) (5.304)

27 - 11.843 (5057)

13, (s.c.)

+1.773 I'co (0 140)

+1 .375 I'eo (02 13)

132 (s.c.)

+1.216q '_2 (0.287)

- 0.580 INR ,

(0225)

13, (s.e.)

- 67.556 q" (9 .8 15)

- 23.98 1 q,. (9.375)

Eqs. I - II , 12- 19 and 20-27 explore QSAR of A I> A2A and A, binding affinities respectively .

136 (s.e.)

F (s)

0.882 60.6 (0.947) (0.222)

0.689 (0.849)

22.4 (0.389)

AYRES (n)

0.177 (25)

0.306 (30)

Keys: 95% Cl = 95% confidence intervals of r. c . (df = n- np-i, i = I if intercept is present , i = 0 otherwise); np = no. of predi ctor va ri ­ables; r.c . are significant at 95% level unless superscripted with * (significant at 90% level ); F ratios are significant at 99% leve l rdf = (np, n - np - i), i = I if inte rcept is present. i = 0 otherwi se]; AC = autocorrelation, DC = Deleted outlier compound

Table IV - Statistical quality of selected* relations of the adenosine AI receptor binding affinity of 1,2-dihydro-2-phenyl ­I ,2.4-triazolo[4,3-aJquinoxal in- l-one derivatives with Wang-Ford charges of different non-hydrogen common atoms

(alongwith indicator parameters)

Atom no(s)+ (x)

I

3

5

7

9

II

13

15

17

19

1,8

7, 19

8, 13

8, 15

8. 17

8,19 9, 20

1, 8, 19

6,16, 19

7, 16, 19

8, 11,20

8, 13,20

8, 14,20

8, 16, 19

8, 17, 19

8, 18, 20

9, 11 ,20

9, 16, 20

1,8, 16, 19

0.459

0.526

0.62 1

0.660

0.644

0.466

0.454

0.529

0.456

0.479

0.683

0.694

0.680

0.683

0.688

0.726 0.679

0.7 17

0.717

0.749

0.721

0.716

0.720

0.766 0.748

0.729

0.719

0.717

0.757

Model equation pKAI = ~I leo + ~2 1 + L~i+2 q .. + a

r or R

0.7 14

0.756

0.8 10

0.832

0.823

0.7 18

0.711

0.757

0.712

0.726

0.850

0.856

0.849

0.850

0.853

0.872 0.848

0.872

0.872

0.888

0.874

0.872

0.874

0.896 0.887

0.878

0.874

0.872

0.896

10.1 0.770

12.9 0.721

18.5 0.645

2 1.7 0.61 1

20.3 0.625

10.3 0.766

9.9 0.774

13.0 0.71 9

10.0 0.772

10.8 0.756

18.2 0.590

19.2 0.579

18.0 0.592

18.2 0.590

18.6 0.585

22.2 0.548 18.0 0.593

17.2 0.557

17.2 0.558

20. 1 0.525

17.5 0.554

17.2 0.558

17.5 0.554

21.9 0.507 20.0 0.525

18.3 0.545

17.4 0.555

17.2 0.558

17.6 0.5 17

Atom no(s).+ (x)

2 4

6

8 10

12

14

16

18

20

2, 8

8, II

8, 14

8, 16

8. 18

8,20

2,8,19

7, 15, 19

8, II , 19

8, 13 , 19

8, 14.19

8, 15, 19

8, 16,20

8. 17.20

8, 19, 20

9, 16.19

2,8, 16, 19

0.456

0.573

0.618

0.691 0.597

0.543

0.480

0.486

0.455

0.472

0.690

0.688

0.683

0.682

0.682

0.720

0.72 1

0.720

0.726

D.7 17

0.7 17

0.732

0.722

0.726

0.723

0.756

0.757

r or R

0.7 12

0.783

0.809

0.849 0.796

0.766

0.727

0.73 1

0.711

0.722

0.854

0.852

0.850

0.850

0.850

0.869

0.874

0.874

0. 877

0.872

0.873

0.880

0.875

0.877

0.875

0.891

0.896

9.9 0.773

15.3 0.685

18.3 0.647

24.9 0.582 16.8 0.665

13.7 0.708

10.8 0.756

11.1 0.751

9.9 0.773

10.5 0.761

18.8 0.584

18.6 0.586

18.2 0.590

18.2 0.590

18.2 0.59 1

2 1.6 0.554

17.5 0.553

17 .5 0.554

18.0 0.548

17.2 0.558

17.2 0.557

18.5 0.542

17 .6 0.552

18.0 0.548

17.7 0.552

20.9 0.5 17

17.6 0.5 16 (- Coli/d)

1492 INDIAN J. CHEM., SEC B, JUNE 2003

Table IV - Statistica l quality of selected* relations of the adenosine A I receptor binding affinity of I ,2-dihydro-2-phenyl­I ,2,4-triazolo[4,3-aJquinoxa lin- l-one deri vati ves wi th Wang-Ford charges of di fferent non-hydrogen common atoms

(alongwith indi ca tor parameters) (-Colllell

A tomno(s):

(x)

7, 13, 16, 19

8, 13, 16, 19 8, 16, 19,20

9, 14, 16, 19

0.76 1

0.777 0.759

0.757

M odel equation pKAI = ~I leo + ~2 I + L~;+2 qx; + a

r or R F'" Atom n (5).+ R}

0.898

0.905 0.897

0.896

18.0 0.5 12

20.0 0.495 17.R 0.5 14

17.6 0.5 16

(x)

8, II, 16, 19

8,14, 16, 19

9. I I , 16, 19

9,16, 19,20

0.759

0.762

0.759

0.757

r or R

0.897

0.898

0.897

0.896

17.8 0.5 14

18. 1 0.5 11

17 0.5 14

17.6 0.5 16

* Se lected relations involving charges of the atoms that are not much intercorrelated (r < 0.7) and having corrc lati on coeffi­cients higher than those of the best equati ons involving less number of charge parameters (predictor variables) are shown. +Wang-Ford charges of the atoms or combination of atoms shown are used to deri ve relations.

'" df = np, n - np - I ; n = no. of data points (=33); np = no. of predictor variables

Table V - Stat istica l quali ty of selected* relations of the adenos ine AlA receptor binding affini ty of 1,2-dihydro-2-phenyl­I ,2,4- triazolo[4,3-a]qu inoxalin- I -one derivati ves with Wang-Ford charges of different non-hydrogen common atoms

A tom no(s).+

(x)

I

3

5

7

9

II

13 15

17

19

1,13 3, 13

5, 13

7, 13

9, 13

I I , 13

13, 15

13, 17

13,1 9

1,7, 13

1, 10, 13

I , II , 13

1, 12, 16 1, 12, 18

1, 13, 15

1, 13, 17 I , 13, 19

2,5, 13

2, 8, 13

2, 10, 13

R 2 a

0.236

0.042

- 0.024

- 0.05 1

- 0.074

0.02 1

0.510 -0.031

-0.0 16

-0.049

0.581 0.548

0.474

0.470

0.476

0.47 1

0.545

0.575

0.532

0.578

0.566

0.608

0.737 0.547

0.586

0.712 0.6 15

0.555

0.574

0.58 1

r or R

0.539

0.333

0.222

0. 154

0.048

0.30 1

0.738 0.207

0.239

0. 162

0.801 0.783

0.74 1

0.739

0.742

0.740

0.78 1

0.797

0.774

0.8 17

0.8 12

0.832

0.891 0.802

0.822

0.879 0.835

0.807

0.8 16

0.8 19

Model equation pK A2A = L~; qx; + a

5.3 0.783

1.6 0.877

0.7 0.907

0.3 0.9 19

0.0 0.929

1.3 0.887

15.6 0.627 0.6 0.9 10

0.8 0.903

0.4 0.9 18

10.7 0.580 9.5 0.603

7.3 0.650

7.2 0.652

7.4 0.649

7.2 0.652

9.4 0.605

10.5 0.584

8.9 0.6 13

7.4 0.582

7.1 0.590

8.2 0.561

14.1 0.459 6.6 0.603

7.6 0.576

12.5 0.481 8.5 0.556

6.8 0.598

7.3 0.585

7.5 0.580

Atom nO(5).+

(x)

2

4

6 8

10

12

14

16

18

20

2, 13 4, 13

6, 13

8, 13

10,13

13, J4

13, 16

13, . 8 13,20

1,8, 13

I , 11 , 12

1, 12, 15

1, 12, 17

1, 13, 14

1, 13, 16

I , 13, 18

1, 13, 20

2,7, 13

2,9, 13

2, 11 , 12

0.229

- 0.008

0.055

0.029

0.047

0.077

- 0.076

0.040

- 0.070

0.045

0.584 0.56 1

0.55 1

0.476

0.47 1

0.470

0.575

0.568

0.482

0.596

0.600

0.566

0.692

0.596

0.68 1

0.624

0.596

0.560

0.554

0.6 16

r or R

0.533

0.252

0.350

0.3 14

0.339

0.378

0.035

0.329

0.080

0.336

0.802 0.790

0.784

0.742

0.739

0.738

0.797

0.794

0.745

0.826

0.828

0.8 12

0.87 1

0.826

0.866

0.839

0.826

0.809

0.806

0.836

s

5.2 0.787

0.9 0.900

1.8 0.87 1

1.4 0.883

1.7 0.875

2.2 0.86 1

0.0 0.929

1.6 0.878

0. 1 0.927

1.7 0.876

10.8 0.578 10.0 0.593

9.6 0.600

7.4 0.649

7.2 0.652

7. 2 0.653

10.5 0.584

10.2 0.589

7.5 0.645

7. 9 0.570

8.0 0.567

7. 1 0.590

11.5 0.497

7.9 0.569

10.9 0.506

8.8 0.549

7.9 0.569

6.9 0.595

6.8 0.598

8.5 0.555 (- Colllel)

ROY : QSAR OF ADENOSINE RECEPTOR ANTAGONISTS r 1493

Table V - Statistical quality of selecled* relations of the adenosine A2A receptor binding affinity of 1,2-dihydro-2-phenyl-I ,2,4-triazoloI4.3-a]quinoxa lin- I-one deri vat ivcs Wilh Wang-Ford charges of different non-hydrogen common atoms

Model equati on pK A2A = L~i q .. + a (-Collld)

Atom no(s)+ R} r or R FV Atom noes): R/ r or R FV

(x) (x)

2, II , 13 0.607 0.831 8.2 0.562 2, 12, 15 0.558 0.808 6.9 0.596

2, 12, 16 0.723 0.885 13.2 0.471 2, 12, 17 0.697 0.873 11.7 0.493

2, 12, 18 0.580 0.8 18 7.4 0.58 1 2, 13, 14 0.587 0.822 7.6 0.576

2. 13, 15 0.583 0.820 7.5 0.578 2, 13.16 0.672 0.862 10.6 0.5 13

2, 13,17 0.708 0.878 12.3 0.484 2. 13, 18 0.630 0.842 8.9 0.545

2. 13, 19 0.624 0.839 8.7 0.550 2. 13.20 0.595 0.826 7.9 0.570

3, II. 13 0.565 0.8 11 7. 1 0.59 1 3, 12, 16 0.6 14 0.835 8.4 0.556

3, 12, 17 0.553 0.806 6.8 0.599 3. 13, 14 0.546 0.802 6.6 0.599

3, 13, 15 0.551 0.804 6.7 0.600 3, 13, 16 0.636 0.845 9.2 0.540

3, 13 17 0.670 0.861 10.5 0.515 3, 13. 18 0.597 0.826 7.9 0.569

3. 13, 19 0.584 0.820 7.5 0.578 4.7.13 0.59 1 0.824 7.7 0.573

4,8. 13 0.604 0.830 8. 1 0.564 4.12.1 5 0.584 0.821 7.6 0.578

4. 12. 16 0.683 0.866 11 .0 0.505 4. 13, 15 0.574 0.8 16 7.3 0.585

4. 13. 16 0.648 0.850 9.6 n.532 4. 13. 17 0.664 0.858 10.2 0.5 19

4, 13. 18 0.603 0.829 8. 1 0.565 4. 13. 19 0.589 0.823 7.7 0.574

4. 13.20 0.572 0.8 15 7.2 0.586 6,8, 12 0.58 1 0.819 7.5 0.580

6.8. 13 0.653 0.853 9.8 0.528 6, 12, 16 0.606 0.831 8.2 0.562

6. 13, 15 0.575 0.8 16 7.3 0.584 6, 13, 16 0.644 0.849 9.4 0.535

6. 13. 17 0.650 0.852 9.7 0.530 6, 13, 18 0.588 0.822 7.7 n.575

6, 13. 19 0.568 (J.8 13 7. 1 0.589 6. 13. 20 (J .557 0.808 6.9 0.596

7. 13,16 0.546 0.802 6.6 0.603 8.9. II 0.565 0.8 11 7. 1 0.59 1

8. 13. 18 0.550 0.804 6.7 0.60 1 10.1 3. 16 0.560 0.809 6.9 0.594

10. 13. 17 0.574 0.8 15 7.3 0.585 10.17.20 0.554 0.806 6.8 0.599

11.I3.16 0.562 0.8 10 7.0 0.593 II. 13, 17 0.565 0.811 7.1 0.59 1

11 . 17.20 0.559 0.808 6.9 0.595 13, 16,20 0.567 0.8 12 7. 1 0.589

13, 17.20 0.58 1 0.8 19 7.5 0.580

*Selcctcd relations in volving charges of the atom, that arc not much intcrcorrclated (r < 0.7) and hav ing correlation coefTi-cien ts hi ghcr than thosc of thc best equati ons involving less num ber of charge paramcters (pred ictor variablcs) are shown. +Wang-Ford charges of the atoms or combination of atoms shown arc used to derive relat ions. v df = np, n - np - I; n = no. of data points (= 15); np = no. of predictor variables

Table VI - Statistical quality of selectcd'" relation s of thc adcnosine A3 receptor binding affin ity of 1.2-dihydro-2-phenyl-1.2.4-triazolof4.3-a Jquinoxalin-l-onc der ivativcs wi th Wang-Ford charges of diffcrcnt non- hydrogen common atoms

(a longwi th indicator paramcters)

Model equati on pKAJ = ~I I'eo + L~i+1 qXI + a

Atom noes): R/ I' or R FV Atom no(s)+ R/ I' or R FV

(x) (x)

I 0.548 0.764 16.8 0.428 2 0.557 0.769 17.3 0.423

3 0.563 0.772 17.7 0.42 1 4 0.582 0.784 19 .1 OA II

5 0.565 0.774 17.9 OA I9 6 0.56 1 0.77 1 17.6 OA22

7 0.540 0.759 16.3 0.43 1 8 0.536 0.756 16.0 0.433

9 0.559 0.770 17.5 0.423 10 0.574 0.779 18.5 0.4 15

II 0.585 0.786 19.4 0.4 10 12 (J .556 0.768 17.3 OA24

13 0.539 (U58 16.2 0.432 14 0.7 J() n.856 32.8 0.343 15 0.584 0.785 19.2 0.411 16 0.562 0.772 17.7 0.42 1

(- Collld)

1494 INDI AN 1. C HEM., SEC B, JUNE 2003

Table VI - Stati stical qua lity of selected* relations of the adenosine AJ receptor binding affinity of L2-dihydro-2-phenyl-I ,2,4-triazolo[4,3-a]quinoxalin-l-one deri vatives with Wang-Ford charges of different non-hydrogen common atoms

(alongwith indicator parameters)

Model equation pKAJ = ~I I'eo + L~i+1 qxi + (l (- Col/ rd)

Atom no(s).+ R} r or R F\7 s Atom no(s).+ R} r or R F\7

(x) (x)

15 0 .584 0.785 19.2 0.411 16 0.562 0.772 17.7 0.421

17 0.540 0 .758 16.3 0.432 18 0 .606 0.798 21.0 0.399

19 0 .604 0.797 20.9 0.400 20 0.582 0 .784 19. 1 0.4 11

1,14 0,/:I()7 0,910 37,1 0,280 2, 14 0,/:107 0,911 37,2 0,279

3, 14 0.794 0.905 34.5 0.288 4,14 0.774 0.894 30.7 0 .303

5, 14 0.725 0.870 23.9 0.333 6,14 0.749 0.882 26 .8 0.3 19

7,14 0.730 0.872 24.4 0.33 1 8,14 0.725 0.870 23.8 0.334

9, 14 0.727 0 .87 1 24.1 0 .332 10, 14- 0.7 13 0.864 22.6 0.34 1

11 , 14 0 .699 0.857 2 1.2 0.349 12, 14 0.7 16 0.865 22.9 0.339

13, 14 0.702 0.858 2 1.4 0.347 14,15 0.7217 0 .866 22.9 0 .338

14,16 0.715 0.865 22.8 0 .339 14,1 7 0.703 0.859 2 1.5 0.346

14, 18 0.724 0.869 23 .7 0.334 14,1 9 0.71 3 0.864 22.5 0.341

14, 20 0 .71 6 0.865 22 .9 0.339

*Se lected rclations in volving charges of the atoms that are not much intercorrelated (r < 0.7) and having correlation coeffi-cients highcr than those of the best equations in volving less number of charge parameters (predictor va ri ables) are shown. +Wang-Ford charges of the atoms or combination of atoms shown are used to deri ve re lations.

\7 df = np, n - np - I ; n = no. of data points (=27); np = no. of predictor variablcs

Table VII - Autocorrelation (r) among important predictor variables (AI binding affinity) (n = 33)

q7 qg qlJ q l6

q7 1.000 0.976 0 .254 0.332

qR 1.000 0 .233 0.335

q lJ 1.000 0 .105

qll> 1.000

ql 9

I

leo

qX-7

q16+19

the best trivariate relation). Again atoms C I, C2, C) and C4 are highly intercorrelated (Table VIII). Con­sidering sign of the charge parameters, a new parame­ter ql +3-2-4 was defined. Similarl y two more parame­ters q1 2- 13 and q1 6-17 were defined and Equation 19 was obtained. The calcul ated binding affinity values ac­cording to Eq. 19 are g iven in Table I . Autocorrela­tion among the important predictor variables is pre­sented in Table VIII.

From the studies it appears that the benzene ring of the quioxaline fragment (as evidenced from importance of the atoms C I, C 2, C3 and C4), the phenyl ring present as substituent on the triazole ring

ql 9 leo qH-7 q l6+ 19

0.195 0 .30 1 0.062 0.996 0 .174

0.172 0 .225 0.112 0.992 0.203

0 .226 0.350 0 .5 11 0.247 0 .38 1

0 .629 0.424 0 .158 0.336 0.474

1.000 0.31 1 0 .078 0. 187 0 .385

1.000 0 .164 0.27 1 0 . 152

1.000 0.083 0 .277

1.000 0. 186

1.000

(as evidenced from importance of the atoms C 16, C J7 )

and atoms NI 2 and N I3 (two nitrogens of the triazole ring) are important for adenosine A lA binding affinity.

QSAR of A3 receptor binding affinity No acceptable equation could be generated

considering a ll compounds, quantitati ve A3 binding affi nity data of which were availab le. Some compounds (9, 10, 16, 19, 25~ 33, 35, 36) with low binding affinity were deleted fro m the set for the QSAR study . A dummy parameter {co indicating presence of O=C-C fragment in the a lky lamino group at RI was fo und to singularly explai n 55.3% of the

ROY : QSAR OF ADENOSINE RECEPTOR ANTAGONISTS I 1495

Table VIII - Autocorre lation (r) among important predictor variables (A2A binding affinity) (n = 15)

q l q 2 q 3 q 4 q l2

ql 1.000 0.985 0.942 0.860 0. 135

q2 1.000 0.964 0.8 15 0. 166

q ) 1.000 0.839 0.325

q 4 1.000 0.477

q l2 1.000

q u

q l6

q l7

q1 2·13

Ql b· 17

q l+3.2-4

variance (Table III) . Thi s parameter was retained when correlating di fferent charge parameters and their di fferent combinati ons to the biological acti vity (Table VI). Atoms 0 ,4, C, and and C2 emerged as mos t important atoms. Again , atoms C, and C2 were highl y intercorrelated (Table IX) and hence, a new parameter q' .2 was defined. When a dummy parameter f NRI indi cating the presence o f a lkyl substitutio n on atom N7 (nitrogen) was included, the resulting equation (Eq . 27) could accommodate three more compounds (33, 35, 36). The calculated acti vity va lues accordi ng to Eq. 27 are given in Table I. Autocorre lation among the important predictor variables is presented in Table IX.

From the analys is, it appears that the benzo fused moiety of the quinoxalin ring (as evidenced from im­pOlt ance of the atoms C, and C2) and oxo group (atom 0 ,4) on the tri azole ring are esssentia l for the A3 bind­ing affinity. Alkyl substitutions o n the nitrogen (N7)

atom of quinoxalin ring are de trimental to the acti vity while presence of acy lamino substituent at R, is con­ducive to the activity.

Overview of QSA R Based on the above anal yes, the general structure

of the compounds was partitio ned into so me frag­ments as menti oned below:

(a) Fragmenl A : composed of the quinoxalin nitro­gen (N7) with the exocyclic ortho-amino substituent or orth.o-oxo group.

(b) Fragment B : composed o f atoms N ' 2 and N I3

(two nitrogens o f the tri azo le ring) which represents the e lectron rich area.

(c) Fragment C : composed o f phenyl substitutio n on the triazole ring which may be accommodated in the lipophilic si te in the receptor.

q l3 q l6 q l7 q1 2· 13 q 16· 17 q 1+3·2-4

0.336 0.069 0.200 0. 120 0.1 67 0.997

0.320 0.029 0. 180 0.096 0. 14 1 0.989

0. 102 0.065 0.20 1 0.108 0. 166 0.960

0.038 0.051 0. 109 0.265 0.095 0.879

0.732 0.3 11 0.259 0.923 0.276 0. 196

1.000 0.039 0.083 0.938 0.05 1 0.280

1.000 0.938 0.1 37 0.967 0.059

1.000 0. 179 0.995 0. 190

1.000 0. 170 0.058

1.000 0.1 57

1.000

Table IX - Autocorre lation (r) among important predictor vari ­ab les (A) binding affi nity) (n = 27)

ql q 2 q l4 I'eo INRI q l·2

ql 1.00 0.996 0.394 0. 128 0. 186 1.000

q 2 . 1.000 0.334 0.1 40 0. 188 0.998

q l4 1.000 0.154 0.325 0.377

I'eo 1.000 0.082 0. 131

INRI 1.000 0. 187

ql .2 1.000

(d) Fragment D: representing the oxo group o n the tri azo le ring, which may interact with some hydrogen bond donor group in the receptor.

(e) Fragment E: composed of benzo fu sed moiety of the quinoxalin fragment, which may inte ract wi th planar anchoring site in the receptor.

The above study suggests that fragments A, B and C may be involved in adenosine A, binding while fragments B , C and E may pl ay impo rtant role in adenosine A2A binding. Fragments A, 0 and E may be important for A3 binding. Thus it appears that the re­quirements fo r optimum binding for these receptor SUbtypes (A" A2A and A3) were di fferent to some ex­tent. However, because of non-ava ilability of all three binding affinity data for all the compounds, quantita­ti ve se lecti vity re lations could not be generated .

References I Beraldi P G, Cacciari B, Romagnoli R, Merighi S , Varani K,

Borea P A & Spalluto G, Med Res Rev, 20, 2000, 103. 2 Klo tz K N, Na wlYIl Schmiedebergs Arch Pharmacol, 362,

2000,382. 3 Deckert J & Gleiter C H, J Neural Translll Stlppl, 43. 1994, 23. 4 Dunwiddie T V & Masino S A, Annu Rev Neurosci. 24, 2001 ,

31. 5 Latini S, Pazzag li M, Pepeu G & Pedata F. Cen Phan nacol,

27, 1996, 925 .

1496 INDIAN J. CHEM ., SEC B, JUNE 2003

6 Basheer R, Porkka- Heiskanen T, Strecker R E, Thakkar M M & McCarley R W, Bioi Sigllals Recept , 9, 2000, 3 19.

7 Kulinskii V I. Vesl/1 Ross Akad Med Nauk, 9, 2000, 39. 8 Rudolphi K A, Schubert P, Parkinson F E & Fredholm B B,

Cerebrovasc Brain Merab Rev, 4, 1992,346. 9 Ongini E, Monopoli A, Cace iari B & Bara ldi P G, Farmaco.

56. 2001 , 87. 10 Knutsen L J & Weiss S M, CLirr Opin Ill vestig Drugs, 2,

2001 , 668. II Randhawa (Jr) M P. Lasley R 0 & Mentzer (J r) R M, J Card

SLirg. 8, 1993,332. 12 Pnueger A. Larson T S, Nath K A, King B F, Gross J M &

Knox F G. Ma\'() Ciill Pro(', 75, 2000. 1275. 13 Moreau J L & Hu ber G. Brain Res Brain Res Rev. 3 1, 1999,

65. 14 Fuxe K, Ferre S, Zo li M & Agnati L F, Braill Res Brain Res

RCI'. 26. 1998. 258. 15 Feokkti stov I, Polosa R, Holgate S T & Biaggioni I. Trellds

Phar/1/acol Sci, 19, 1998, 148. 16 Dunwiddie T V & Masino S A, Ail/ill Rev Nellrosci. 24, 2001 ,

3 1.

17 Kaiser S M & Quinn R J, Drug Discov Today , 4,1999. 542. 18 Linden J, Annu Rev Pharmacal Toxicol . 41, 2001 , 775 . 19 Muller C E & Stcin B, CLirr Pharm Des. 2, 1996.0 1 20 Poul sen S-A & Quinn R J. Bioorg Med Chem. 6, 1998.619. 21 Colotta V, Catarzi 0, Varano F, Cecchi L. Filacchi oni G.

Martini C, Trincavelli L & Lueacehini A. J Med Chem, 43. 2000, 1158.

22 Dewar M J S, Zoebiseh E G, Healey E F & Stewart J J p, J Ani Chem Soc, 107. 1985,3902.

23 Dewar M J S, Hwang C H & Kuhn 0 R, J Alii Chem Soc, 113, 1991 , 735.

24 Civcir P U, J Mol Struct: THEOCHEM, 535,2001 , 121. 25 Chem 3D Pro I'ersion 5.0 and Chem Druw Ultra I'ersioll 5.0

are programs of Chambridgcsoft Corporation. U.S .A. 26 Roy K, De A U & Sengupta C, Quant Slruct-Ac{ Relat. 20.

2001. 3 19. 27 The GW-BASIC programs AUTOREG and RRR98 were de­

ve loped by Kunal Roy ( 1998) and standardi zed on known datase(s.

28 Snedecor G W & Cochran W G, Swtistical Methods . (Oxford and IBH Publishing Co. PVI. Ltd. , New Delhi), 1967.38 1.