J . Chem. In$ Comput. Sei. 1989, 29, 163-172
163
Atomic Physicochemical Parameters for Three Dimensional Structure Directed Quantitative Structure-Activity Relationships. 4. Additional Parameters for Hydrophobic and Dispersive Interactions and Their Application for an Automated Superposition of Certain Naturally Occurring Nucleoside Antibiotics VELLARKAD N. VISWANADHAN, ARUP K. GHOSE,* GANAPATHI R. REYANKAR, and ROLAND K. ROBINS Department of Molecular Modeling and Computers, Nucleic Acid Research Institute, 3300 Hyland Avenue, Costa Mesa, California 92626 Received June 29, 1988 We have shown previously that atomic values for physicochemical properties are an important guide for correlating the observed biological activity of the ligands to their chemical structure (Chose, A. K.; Pritchett, A.; Crippen, C.M. J. Comput. Chem. 1988,9, 80-90, and references cited therein). The objective of the present work is to (i) report the hydrophobicity and the molar refractivity for phosphorus and selenium atoms at different structural environments that are ubiquitous in biologically active systems, (ii) refine the atomic values of the various elements reported earlier to satisfy the largely extended data set, and (iii) suggest a method for selecting the best superposition of different molecules on a reference structure using these atomic physicochemical properties. The octanol-water partition coefficient was used to scale the atomic hydrophobicity. The hydrophobicity values of 120 atom types were evaluated from 893 compounds. The observed and calculated octanol-water partition coefficient showed a correlation coefficient of 0.926 and a standard deviation of 0.496. The atomic refractivity values were evaluated from the molar refractivities of 538 compounds; the corresponding correlation coefficient and standard deviation were 0.999 and 0.774, respectively. The atomic values were tested by predicting the respective properties for a large number of compounds. The superposition method has been applied to certain naturally occurring nucleoside antibiotics. The algorithm presented here shows various important superpositions of two or more molecules with minimum physical assistance to avoid any personal bias. INTRODUCTION An important step in drug action is the interaction of the drug with a biological receptor. However, the direct study of the drug (ligand)-receptor interaction by molecular mechanics and dynamics1i2 is not feasible in most cases because the receptor structure or the binding site is unknown. The QSAR3 approach deals with the situation indirectly. It correlates the biological activity of the ligands with their structural or physicochemical properties and extends the correlated properties for the prediction of new active ligands. In the linear free energy relationship (LFER)approach3 (the Hansch approach), physicochemical properties of the ligands (bioactive compounds) are used in multiparametric regression models for correlating with biological activity. Such approach was acceptable when the undertaking of the three-dimensional structure of the ligand was computationally too expensive. However, the rapid improvement of the computational facility prompted many to develop methods for three dimensional structure directed quantitative structure-activity relationships.&' Application of these techniques for the three-dimensional mapping of an unknown receptor site cavity requires not only analyses of conformational flexibilities in the ligands but also a knowledge of the physicochemical characteristics of the ligands. The most important properties related to biomolecular interactions are hydrophobicity,8 formal charge density,*" and molar refractivity.12 The basic hypotheses for receptor mapping are that (i) the interactions at different parts of the ligands with the receptor are different and (ii) the nature of interaction can be determined by correlating the local physicochemical properties with their binding free energies. To exploit the second hypothesis, we need to know the relative orientation of the ligands at the binding site (binding modes) and a method that can estimate the physicochemical properties
of the ligands at any region from the occupancy of the atoms. Previously we s h o ~ e d ~ ~how - ' ~atomic physicochemical properties can be developed and used to estimate local or overall physicochemical properties. The primary objective of the present work is to give the hydrophobicity (octanol-water partition coefficient) and the molar refractivity for phosphorus and selenium atoms in different structural environments since these atoms are found in many biologically important nucleosides and nucleotides. The atomic values of the various types of atoms that were reported earlier have been refined to satisfy the largely extended data set. The secondary objective of the work is to show a method of estimating the goodness of fit of various geometrically feasible superpositions of a molecule on a reference structure on the basis of physicochemical property matching, using these atomic parameters. Such superposition is often helpful in determining the molecular similarity and in rationalizing biological activity of compounds in diverse structure. It is not very straightforward to decide what physicochemical properties should be considered to determine the best molecular superposition. Since understanding biological activity is the ultimate objective of our work, we should consider the properties that govern the interaction of the ligand (drug) molecules with the biological receptor.16 In the absence of any information regarding the structure of the binding site, we considered three properties: hydrophobicity,8 molar refractivity,12 and formal charge density." The term hydrophobicity refers to the force or corresponding energy that operates between two or more nonpolar solutes in water and arises from dispersive and electrostatic forces and the consequent entropic factor. A hydrophobic substance is soluble in nonpolar solvents but only sparingly soluble in water. Though the hydrophobic effect plays an important role in biological systems, it is not well understood theoretically. A great deal of effort is nec-
0095-2338/89/1629-0163$01.50/0 0 1989 American Chemical Society
164 J . Chem. In& Comput. Sci., Vol. 29, No. 3, 1989
VISWANADHAN ET AL.
Table 1. Classification of Atoms and Their Contributions to Octanol-Water Partition Coefficient Which Is a Measure of Hydrophobicity and Molar Refractivity type 1 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 37 38 39 40 41 42 43 44 45
46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61' 62-63 64 65 66 67 68 69
description" C in :CH3R, CH4 :CH2R2 :CHR3
:CH;RX :CH,X, :CHR2X :CHRX2 :CHX3 :CR,X :CR2X2 :CRX3
:cx4
:=CH2 :=CHR :=CR2 :=CHX :=CRX
:=cx2 :=CH :=CR, R=C=R
:=cx
:R- -CH- -R :R- -CR- -R :R- -CX- -R :R- -CH- -X :R- -CR- -X :R- -CX- -X :X- -CH- -X :X- -CR- -X
:x--cx--x :R- -CH***X :R- -CR**.X :R- -CX.**X :AI-CH=X :Ar-CH=X :AI-C(=X)-AI :Ar-C(=X)-R :R-C(=X)-X, x-C(=X)-x :X- -CH*.*X :X- -CR***X
R-CFX,
X=C=X
:x--cx...x
unused H attached toC :CoSp3, having no X attached to next C :c;,,3, COP2 : c sp3, c sp2, co : ~ 3 , ~~ ~ 2, c3:i2, ~ ~3~~ ~ ~ , :heteroatom :CY-Cd having 1 X attached to next carbon having 2 X attached to next carbon having 3 X attached to next carbon having 4 or more X attached to next carbon 0 in :alcohol :phenol, enol, carboxyl OH .-.-0 :AI-0-A1 :AI-0-Ar, Ar20 R...O...R, R-O-C=X :- -0 unused Se in :Any-Se-Any ..-Se N in :AI-N H2 :AI2" :A13N :Ar-NH2, X-NH2
hydrop*
no. of compounds
freq of use
atomic refrac
no. of compounds
freq of use
-0.6771 -0.48 7 3 -0.363 3 -0.1366 -1.0824 -0.8370 -0.6015 -0.5210 -0.4042 0.365 1 -0.5399 0.401 1 0.2263 0.8282 -0.1053 -0.0681 -0.2287 -0.3665 -0.9188 -0.0082 -0.1047 0.1513
385 245 46 24 196 302 6 87 36 4 15 3 36 6 25 48 9 23 13 5 3 4
589 506 51 24 299 485 6 152 36 4 15 3 38 6 31 70 10 24 13 6 4 5
0.0068 0.1600 -0.1033 0.0598 0.1290 0.1652 0.2975 0.9421 0.2074 -0.1774 -0.2782 -0.3630 -0.032 1 0.3568 0.8255 -0.1116 0.0709 0.457 1 -0.1 3 16 0.0498 0.1847
584 307 432 92 66 70 23 7 14 16 39 7 5 5 27 23 289 107 22 33 10 -
2222 371 737 142 73 85 23 7 14 19 44 8 5 5 31 29 356 115 22 33 13 -
2.9680 2.91 16 2.8028 2.6205 3.0150 2.9244 2.6329 2.5040 2.3770 2.5559 2.3030 2.3006 2.9627 2.3038 3.2001 4.2654 3.9392 3.6005 4.4870 3.2001 3.4825 4.2817 3.9556 3.4491 3.8821 3.7593 2.5009 2.5000 3.0627 2.5009
283 169 33 11 76 199 14 42 15 5 15 16 12 4 38 44 7 11 14 8 12 18 7 187 107 102 14 14 4 1
495 389 37 11 114 320 15 51 19 5 16 25 15 4 44 56 7 11 16 12 13 23 7 896 144 156 16 15 4 1
2.6632 3.4671 3.6842 2.9372 4.0190 4.7770 3.9031 3.9964 3.4986 3.4997 2.7784 2.6267 2.5000 -
1 19 13 4 15 15 15 8 111 13 6 4 2
-
3 22 14 5 15 15 16 8 136 13 6 4 2 -
0.4418 0.3343 0.3161 -0.1 48 8 -0.3260 0.2099 0.3695 0.2697 0.3647
280 799 74 138 603 214 218 11 3
1582 4252 87 209 1084 556 790 25 7
0.8447 0.8939 0.8005 0.8320 0.8000 0.8 188 0.9215 0.9769 0.7701
185 422 68 55 107 123 197 24 1
1058 1982 93 61 148 366 883 83 3
0.1402 0.4860 -0.3 5 14 0.1720 0.2712
98 165 464 39 212
160 192 638 43 289
1.7646 1.4118 1.4429 1.6191 1.3502
20 35 187 30 146
22 40 220 44 217
1.5810
81
178
1.9450
21
45
0.1473
12
12
11.1366 13.1 149
2 6
2 6
0.1187 0.2805 0.3954 0.3 132
24 23 18 84
24 25 20 90
2.6221 2.5000 2.8980 3.6841
10 9 8 9
11 9 8 12
-
J. Chem. In$ Comput. Sci., Vol. 29, No. 3, 1989 165
HYDROPHOBIC AND DISPERSIVE INTERACTIONS Table I (Continued) type 70 71 72 13 74 75 76 71 78 79-80 81 82 83 84 85 86 81 88 89 90 91 92 93 94 95 96 97 98 99 100 101-105 106 107 108 109 110 111-1 14 115 116 117 118 119 120
description' :Ar-NH-AI :Ar-NAI2 : R C O - N < , >N-X=X :Ar2NH, Ar3N Ar2N-A1, R-N-Rf :R=N, R=N:R--N--R,g R--N--X :Ar-N02, R- -N(- -R)- -0" RO-NO2 :AI-NO2 :Ar-N=X, X-N=X unused F attached to :Cisp3 :cJIp3 :clsp3 : c sp2 :c2-4sp2, ClSP c4sp, CI attached to
hydropb 0.4238 0.8678 -0.0528
no. of compounds 10 17 297
freq of use 10 17 393
atomic refrac 4.2808 3.6189 2.5000
no. of compounds 7 10 18
freq of use 7 11 19
0.4198 0.1461 -0.1 106 -2.7640
87 62 170 75
89 90 25 1 87
2.7956 2.7000 4.2063 4.0184
7 27 24 15
7 29 27 17
-2.7919 0.5721
6 40
6 53
3.0009 4.7142
6 10
6 12
0.4174 0.2167 0.2792 0.5839 0.3425
5 8 34 17 1
5 14 103 26 2
0.8725 1.1837 1.1573 0.8001 1.5013
8 7 7 21 8
8 32 28 34 14
0.9609 0.5594 0.4656 0.9624 0.6345
20 8 15 100 20
27 14 37 148 36
5.6156 6.1022 5.9921 5.3885 6.1363
25 16 10 25 32
30 28 29 28 42
1.0242 0.4374 0.4332 1.2362 0.9351
12 3 2 39 4
13 4 4 49 8
8.5991 8.9188 8.8006 8.2065 8.7352
21 10 3 14 9
25 21 9 14 9
1.4350
4
4
1.7018 0.9336
14 1
14 3
13.9462 14.0792 14.0730 12.9918 13.3408
7 4 3 5 1
8 7 3 5 1
0.7268 0.6145 0.3828 -0.1708 0.3717
10 39 25 2 57
10 42 25 2 61
7.8916 7.7935 9.4338 7.7223 5.7558
9 19 7 5 8
10 22 9 5 8
-1.6251 0.3308
1 49
1 52
0.0236
3
3
5.5306 5.5152 6.8360 10.0101 5.2806
5 13 10 2 7
14 10 2 7
x
q s p 3
:cJIp3 :clsp3 : c sp2 :c2-4sp2,CIsp
c4sp, x
Br attached to :Cisp3 :CJIp3 :C1,p3 : c sp2 :c2-4sp2, ClSP c4,, x I attached to :Cisp3 : c sp3 :c3v3 :C'sp2 :C2-4sp2rCIsp c4sp,x unused halogens S in :R-SH :R2S, RS-SR :R=S :R-SO-R :R-SOI-R unused P in :ylids :Rj--P=X :X3-P=X (phosphate) :PX3 (phosphite) :PR, (phosphine) :C-P(X),=X (phosphonate)
5
'R represents any group linked through carbon; X represents any heteroatom (0, N, S, P, Se, and halogens); AI and Ar represent aliphatic and aromatic groups, respectively; = represents double bond; represents triple bond; - represents aromatic bond as in benzene or delocalized bonds such as the N-0 bond in nitro group; represents aromatic single bonds as the C-N bond in pyrrole. bAtomic hydrophobicity in the unit of log P(octano1-water).
