QSAR approach - Semantic Scholar

Indian Journal of Chemistry Vol. 49B, February 2010, pp. 224-233

Structural optimization of new class of selective carbonic anhydrase inhibitors: QSAR approach Blessy Pothen, Vineet Singh, Surendra Kumar & Meena Tiwari* Computer Aided Drug Design Lab, Department of Pharmacy, Shri Govindram Seksaria Institute of Technology and Science, 23, Park Road, Indore 452 003, India E-mail: [email protected] Received 10 December 2008; accepted (revised) 16 September 2009 Quantitative structure activity relationship studies have been conducted on a series (24 compounds) of sulfonamide derivatives with selective carbonic anhydrase inhibitory activity using ChemOffice v.8.0 software. The best predictions have been obtained for hCA-II enzyme inhibition activity (Q2 = 0.552, r2 = 0.724) and with hCA-VII enzyme inhibition activity (Q2 = 0.501, r2 = 0.704). Both equations are validated by a test set of compounds and give satisfactory predictive r2 values of 0.434 and 0.608, respectively. The equations selected emphasized the importance of LogP (octanol/water partition coefficient), Highest Occupied Molecular Orbital (HOMO) and Radius of gyration (Rgy) on biological activity i.e.; hydrophobic groups, presence of electron donating groups, size and shape of molecule might be influencing the selective carbonic anhydrase inhibitory activity. Keywords: Sulfonamide, carbonic anhydrase, antiglaucoma, anticancer, quantitative structure activity relationship

Carbonic anhydrases (CAs, EC 4.2.1.1) are zinccontaining metalloenzymes that catalyze the reversible hydration of carbon dioxide to bicarbonate with discharge of a proton, thus playing important physiological and pathophysiological functions. Till date, sixteen different carbonic anhydrase isozymes with different distribution have been described in higher vertebrates, including humans, some of them have been considered as important targets for inhibitors with therapeutic applications viz; topically acting antiglaucoma, anticancer, antiobesity, antithyroid and antihyperglycemic agents. Some sulphonamides also find applications as diagnostic tools, in positron emission tomography (PET) and magnetic resonance imaging (MRI)1. Carbonic anhydrase inhibitors (CAIs) are of two types: (i) metal complexing anions, (ii) Unsubstituted sulfonamides. Both of these bind to the Zn(II) ion of the enzyme either by substituting the non-protein zinc ligand or adding to the metal coordination sphere, generating trigonal-bipyramidal species. Clinically useful sulfonamide derivatives as carbonic anhydrase inhibitors like; acetazolamide, methazolamide, ethoxzolamide, dichlorophenamide, dorzolamide topiramate, zonisamide, and brinzolamide, bind in a tetrahedral geometry of the Zn(II) ion (Figure 1 A), in deprotonated state with the nitrogen atom of the

sulfonamide moiety coordinated to Zn(II) and an extended network of hydrogen bonds, involving residues Thr199 and Glu106, also participating to the anchoring of the inhibitor molecule to the metal ion. Anions may bind either in tetrahedral geometry of the metal ion or as trigonal-bipyramidal adducts (Figure 1B), (ref. 2). The aromatic/heteroaromatic part of the inhibitor (R) viz; sulfonamide/sulfamate/sulfamide interacts with hydrophilic and hydrophobic residues of CA-I, CA-II and CA-IV. In deprotonated state as SO2NH--; it is coordinated to Zn(II) ion of the enzyme, and its –NH- moiety participates in a hydrogen bond with the Oγ of Thr199, which in turn is engaged in another hydrogen bond to the carboxylate group of Glu106. One of the oxygen atoms of the -SO2NH- moiety also participates in a hydrogen bond with the backbone –NH- moiety of Thr199 (refs. 2-9). Carbonic Anhydrase VII isozymes are less studied and understood among the cytosolic CA’s. Montgomery et al.10, isolated it from a human genomic library in 1991; showing 50, 56, and 49% identity with hCA-I, hCA-II and hCA-III isozyme respectively. Later Tashinan’s group carried out purification, characterization and kinetic studies on the mouse isozyme, mCA VII, and concluded that this enzyme is also inhibited by sulfonamides with high

POTHEN et al.: STRUCTURAL OPTIMIZATION OF CARBONIC ANHYDRASE INHIBITORS

(A)

225

(B)

Figure 1 ― α-CA inhibition mechanism by sulfonamide (A) and anionic (B) inhibitors. In the case of sulfonamides, in addition to the Zn(II) coordination, an extended network of hydrogen bonds ensues, involving residues Thr199 and Glu106, whereas the organic part of the inhibitor (R) interacts with hydrophilic and hydrophobic residues of the cavity. For anionic inhibitors such as thiocynate (B) the interactions between inhibitor and enzyme are much simpler.

activity level in low nanomolar range11. Further Kinetic studies carried out by D. Vullo et al.12, suggested that CA-VII is similar physiologically to the isozyme CA-II. Carbonic anhydrase VII has been shown to be highly expressed in the brain and promoting epileptogenesis13,14. Although X-ray crystal structure of hCA-VII is not known, however, the homology of the active-site amino acid sequence in hCA-VII is high with that of the well investigated (by X-ray crystallography) isoforms hCA-I and II. Thus, in addition to zinc ligands, some important amino acid residues for the catalytic/inhibition mechanisms are also identical to CA-I, II, and VII. These are His64 involved in proton transfer processes between the active site and the environment, Thr199, Glu106 involved in a network of hydrogen bonds with the zinc ligand, and Thr200, participating in the stabilization of inhibitors bound to the zinc ion, by formation of a hydrogen bond with its -OH moiety1,15,17. However, there are several amino acid residues in the active site of hCA-VII which are characteristic only of this isozyme and which may explain the inhibition. These amino acids are Asp67 (which is histidine in CA-I and Asn in CA-II) and Asp69 (which is Asn is CA-I and Gln in CA-II) (refs. 17-20). Many carbonic anhydrases isolated from other organisms open a new therapeutic target, such as α-CAs from Plasmodium falciparum and

Helicobacter pylori, and β-CAs from Mycobacterium tuberculosis, Candida albicans etc. Research is being carried out for developing specific inhibitors targeting these enzymes that would lead to conceptually novel therapies21,22. In the present study, quantitative structure activity relationship studies were performed on aromatic/heteroaromatic sulfonamide derivatives in order to correlate the structural requirements for enzyme inhibition which may be useful in designing new molecules against hCA-II and hCA-VII enzyme. Results and Discussion The data set was divided into training set of 19 compounds and test set of 5 compounds for hCA-II (human carbonic anhydrase II) inhibitory activity. For hCA-VII inhibitory activity (human carbonic anhydrase-VII), training set of 18 compounds and test set of 6 compounds were prepared using random selection method. Among the several generated QSAR equations; the best equation was selected on the basis of observed squared correlation coefficient (r2), percent explained variance (%EV), standard deviation (S), sequential Fischer test (F), and validation parameters viz; bootstrapping r2, chance, cross validated correlation coefficient (Q2) value, SPRESS, standard deviation of error prediction (SDEP) and predictive squared correlation coefficient of the test set (r2pred).

INDIAN J. CHEM., SEC B, FEBRUARY 2010

226

QSAR equation for human carbonic anhydrase II inhibitory (pKi(hCA-II)) activity: pKi(hCA-II) = 0.034 (±0.019)*ETor + 0.140 (±0.098)* LogP – 2.706e – 005 (±1.981e –005)*EElec + 6.733 (±0.302) …(Eqn.1) Statistical parameters: n = 19; r = 0.851; r2 = 0.724; variance = 0.052; S = 0.227; F = 13.145; FIT = 140.833; ICAP < 0.27 Validation parameters: Q2 = 0.552; r2bs = 0.768; chance < 0.001; SPRESS = 0.290; SDEP = 0.258; Sbs = 0.101, r2pred = 0.434 QSAR equation for human Carbonic Anhydrase VII inhibitory (pKi(hCA-VII)) activity: pKi(hCA-VII) = 0.071 (±0.031)*ETor + 0.090 (±0.348)* EHOMO + 0.156(±0.145)*Rgy + 7.842 (±3.361) …(Eqn.2) Statistical parameters: n = 18; r = 0.839; r2 = 0.704; variance = 0.089; S = 0.299; F = 11.077; FIT = 123.076; ICAP < 0.17

Validation parameters: Q2 = 0.501; r2bs = 0.748; chance < 0.002; SPRESS = 0.388; SDEP = 0.342; Sbs = 0.139, r2pred = 0.608 On the basis of various significant statistical and validation parameters; best equations were selected as a representation to explore the factors responsible for inhibition of carbonic anhydrases hCA-II and hCAVII, respectively. Equations for hCA-II and hCA-VII inhibitory activity show better correlation coefficient r = 0.851 (pKi(hCA-II)), r = 0.839 (pKi(hCA-VII)), which accounts for more than 70% of the variance in the activity (Table I, Figures 2 and 3). The intercorrelation among the parameters is less than (0.3), indicating orthogonality among the descriptors used for deriving the equation. In multivariable equation, the dependent variables can be predicted from a linear combination of the independent variable. The data showed an overall internal statistical significance better than 99.9% as it exceeded the tabulated F value (F3 α 0.10 = 5.20). The equations were further tested for the outlier by the Z-score method and no compound was found to be an outlier; suggesting that the equations is able to explain the structurally diverse analogs and is

Table I ― Calculated pKi and predicted pKi of training set for hCA II and hCA VII enzyme S. No. Compd.

hCA-II Calculated pKi

a

b

Predicted pKi

Compd..

hCA-VII Calculated pKi

a

1. 6.6041 6.6363 Sulf-1 7.3384 Sulf-1 2. 6.4915 6.4365 Sulf-2 7.1150 Sulf-2 3. 6.7906 6.8177 Sulf-3 7.2441 Sulf-3 4. 6.9624 7.0851 Sulf-5 7.1621 Sulf-4 5. 6.8939 6.9020 Sulf-6 7.1521 Sulf-6 6. 6.9219 6.8978 Sulf-7 7.1748 Sulf-7 7. 6.9407 6.9387 Sulf-9 7.0588 Sulf-8 8. 7.0200 6.9909 Sulf-11 7.5368 Sulf-10 9. 7.3837 7.4865 Sulf-12 7.1198 Sulf-11 10. 7.0021 6.9836 Sulf-15 8.0177 Sulf-12 11. 7.3689 7.4267 Sulf-16 7.2921 Sulf-13 12. 7.3183 7.2233 Sulf-18 8.5109 Sulf-14 13. 7.1804 7.1032 Sulf-19 7.7796 Sulf-16 14. 7.3286 7.3455 Sulf-20 7.9778 Sulf-17 15. 7.6906 7.8062 Sulf-21 7.1197 Sulf-18 16. 7.7631 7.6712 Sulf-22 7.1974 Sulf-20 17. 6.9736 6.9796 Sulf-23 7.0181 Sulf-22 18. 6.8773 6.8774 Sulf-24 7.0698 Sulf-23 19. 6.7788 6.7659 Sulf-24 a Calculated pKi: Calculated enzyme inhibition activity; b Predicted pKi: Predicted enzyme inhibition activity determined by Leave-one-out method

b

Predicted pKi 7.3368 7.1069 7.2600 7.1661 7.1589 7.1837 7.0271 7.6053 7.1814 7.9611 7.4676 8.8627 7.6900 7.8799 7.0953 7.1998 6.9053 7.0419 -

POTHEN et al.: STRUCTURAL OPTIMIZATION OF CARBONIC ANHYDRASE INHIBITORS

helpful in designing more potent compounds using physicochemical parameters. The predictive power of the equations was validated by leave-one out cross validation method. The cross validated square correlation coefficient (Q2) (Figures 4 and 5); predictive residual sum of square (SPRESS) and standard deviation of error of prediction (SDEP) suggested a good internal consistency as well as predictive ability of the equation. The boot strapping r2 is at par with conventional squared correlation coefficient (r2). The robustness and applicability for further optimization of the molecule was explained by significant r2pred value (Table II), (Figures 6 and 7). The selected equations fulfil the statistical validation criteria to a significant extent to be useful as theoretical basis for proposing more potent compounds. Equation 1 for pKi(hCA-II); parameters like LogP (octanol/water partition coefficient) and Torsion energy (ETor) contributed positively where as Electronic energy (EElec) contributed negatively towards biological activity. LogP is a representative

227

of hydrophobic nature of atoms in the molecules and is related to penetration, distribution and interaction with the receptor23,24. It was suggested that substitution of groups, which are highly hydrophobic in nature might increase the biological activity. After analyzing the descriptor table it was found that contributions towards LogP value is greater for compound no. 15, 20 and are potent compounds in the series supporting the hypothesis. Thus, improving the LogP characteristics of the molecule increases the carbonic anhydrase inhibitory activity. Whereas; minimizing the electronic energy of a molecule is helpful for rationalizing the interaction between molecule and the receptor surface. The study reveals that substitution on aromatic/heteroaromatic portion of molecule results in interaction with a hydrophobic pocket at receptor site and hence influences selectivity as well as activity. Supuran et al.25, in their review has shown that presence of electronic groups makes weak H-bond interaction with water molecule confined to binding pocket (Wat1199). In addition, the orientation of compounds no.7-12 within the binding pocket may

Figure 2 ― Scatter plot between the observed activity and calculated activity of training set for hCA II

Figure 3 ― Scatter plot between the observed activity and calculated activity of training set for hCA VII

INDIAN J. CHEM., SEC B, FEBRUARY 2010

228

Figure 4 ― Scatter plot between the observed activity and predicted activity of training set for hCA II

Figure 5 ― Scatter plot between the observed activity and predicted activity of training set for hCA VII Table II ― Observed pKi and Predicted pKi of test set for hCA II and hCA VII enzyme S. No.

Compd.

hCA-II Observed pKi

a

b

Predicted pKi

Compd.

hCA-VII Observed pKi

a

b

1. 6.7696 6.8157 Sulf-4 7.0555 Sulf-5 2. 7.3979 6.9515 Sulf-8 6.9208 Sulf-9 3. 8.6990 7.5411 Sulf-10 6.8239 Sulf-15 4. 7.5229 7.3913 Sulf-13 8.2840 Sulf-19 5. 8.0969 7.8661 Sulf-14 8.8367 Sulf-21 6. Sulf-17 8.1871 a Observed pKi: Experimentally determined enzyme inhibition activity; b Predicted pKi: Predicted enzyme inhibition activity determined by Leave-one-out method

be associated with less activity values for hCA-II; which further supports hypothesis for selective activity. Equation. 2 for pKi(hCA-VII); EHOMO, Radius of gyration (Rgy) and Torsional energy (ETor) contributes positively towards biological activity. EHOMO (Highest Occupied Molecular Orbital energy) is the highest energy level in the molecule that contains electrons; which is important in governing molecular reactivity. When a molecule acts as a lewis base (an electron pair

Predicted pKi 6.9567 7.1130 7.0472 8.1228 7.9828 7.2860

donor) in bond formation, the electrons are supplied from the molecule’s HOMO. Molecules with high HOMOs are more able to donate their electrons in charge transfer phenomenon and hence are relatively reactive compared to molecules with low HOMOs; thus the HOMO descriptors measure the nucleophilicity of a molecule. Substituents with electron withdrawing group will increases the activity and selectivity for hCA-VII. Due to unavailability of Xray crystallographic studies but its homology with

POTHEN et al.: STRUCTURAL OPTIMIZATION OF CARBONIC ANHYDRASE INHIBITORS

229

Figure 6 ― Scatter plot between the observed activity and predicted activity of test set for hCA II

Figure 7 ― Scatter plot between the observed activity and predicted activity of test set for hCA VII

isozyme CA-I and CA-II, electron withdrawing matic/heteroaromatic ring with Thr 199 and Glu106 enzyme function.

it can be suggested that substituents on aromake a H-bond network for better inhibition of

Radius of gyration (Rgy) is a descriptor related to size and shape; which reflects the distribution of atomic masses in a molecule from its center of mass, and is a measure of the compactness of molecule. As can be observed from Tables III and IV; compounds no. 13-20 have perfect size and shape to be easily accommodated in the enzyme’s active site to inhibit its function, further support for its positive contribution to pKi reflects that some bulky and branched chain compounds increase the activity. Other parameter which shows significant contribution to hCA-VII enzyme inhibition activity is torsional energy. It is the energy required to rotate carboncarbon single bond. The positive contribution implicates that the substituents/alkyl side chain prefers to be in staggered conformation for favorable interaction.

Conclusion The fundamental properties of the molecule that are overwhelmingly involved in selective activity of sulfonamide derivatives are thermodynamic, electronic and topological parameter. It is apparent from selected QSAR equations that the thermodynamic (LogP, ETor), electronic (EElec, HOMO) and Topological indices (Radius of gyration) play an important role in the selective inhibition of hCA-II and hCA-VII inhibitory activity. Based on understanding from the selected equations, it may be concluded that sulfonamide derivatives having hydrophobic groups with bulky substituents improve the activity against hCA-II whereas groups with small size and shape bearing electronegative substituents will improve the hCA-VII inhibition activity. Materials and Methods Data set The carbonic anhydrase inhibition activity data of sulfonamide derivatives was taken from the reported work of Supuran et al.26 and is given in Table III.

INDIAN J. CHEM., SEC B, FEBRUARY 2010

230

Table III ― Structure and Activity data of sulfonamide derivatives as carbonic anhydrase inhibitors

a

Compd

hCA-II Ki (nM) d(pKi)(M)

R

c

b

hCA-VII d Ki (nM) (pKi) (M)

c

NH2

Sulf-1

295

6.5302

45

7.3468

Sulf-2

NH2

240

6.6198

70

7.1549

Sulf-3

NHNH2

300

6.5229

89

7.0506

Sulf-4

CH3

320

6.4948

88

7.0555

Sulf-5

CH2NH2

170

6.7696

75

7.1249

Sulf-6

CH2CH2NH2

160

6.7959

80

7.0969

60

7.2218

75

7.1249

110

6.9586

120

6.9208

40

7.3979

61

7.2147

70

7.1549

150

6.8239

63

7.2007

100

7.0000

75

7.1249

210

6.6778

60

7.2218

5.2

8.2840

19

7.7212

4.3

8.3653

2

8.6990

7.0

8.1549

F

Sulf-7

NH2 Cl

Sulf-8

NH2

Br Sulf-9

NH2 I

Sulf-10

NH2 SO2NH2 NH2

Sulf-11 F3 C

SO2NH2 NH2

Sulf-12 Cl

NH2

S

Sulf-13

N N

S

Sulf-14

N N

HN CH3

O S N N

HN

Sulf-15

S

NH2

O ―Contd

POTHEN et al.: STRUCTURAL OPTIMIZATION OF CARBONIC ANHYDRASE INHIBITORS

231

Table III ― Structure and Activity data of sulfonamide derivatives as carbonic anhydrase inhibitors―Contd

a

Compd

hCA-II d Ki (nM) (pKi)(M)

R

c

b

hCA-VII d Ki (nM) (pKi) (M)

c

O HN

Sulf-16

S

NH2

46

7.3372

5.6

7.2518

50

7.3010

6.5

8.1871

33

7.4815

6.8

8.1675

30

7.5229

4.0

8.3979

12

7.9208

5.4

8.2676

O O HN

Sulf-17

S

NH2

O NH2 N

Sulf-18

N

HN S

Sulf-19

OH

N

Cl

Sulf-20

S N N

Sulf-21

CH2OH

8

8.0969

60

7.2218

Sulf-22

CH2CH2OH

125

6.9031

66

7.1805

Sulf-23

COOH

133

6.8761

52

7.2840

125

6.9030

68

7.1675

H2NHN Sulf-24 a

hCA-II and bhCA-VII: human carbonic anhydrase II and VII enzyme respectively determined by esterase method; Ki(nM): representing the binding affinity of substrate to enzyme in nanomole; d pKi(M): negative logarithm of Ki in mole. c

The biological activity data (Ki in nM) was converted to negative logarithmic inhibition constants (M) for quantitative structural activity relationship analysis. Energy minimization and Geometry optimization The molecular structures of all 24 compounds were sketched using ChemDraw Ultra (Version 8.0) software27 of ChemOffice and subjected to energy minimization technique using Allinger’s Molecular Mechanics (MM2) force field followed by geometry optimization using semi-empirical Quantum Mechanics based on AM-1 (Austin Model-1). Hamiltonian approximations method and closed shell (restricted) wave function available in MOPAC module by

fixing root mean square (RMS) gradient as 0.1 and 0.001 kcal/mol-Å respectively was used for calculating partial atomic charges and electron density on various atoms. Charges were kept as mulliken and maximum number of iterations was set to 1000. The energy minimized structures were added to a molecular database to compute various physicochemical properties using ‘compute properties’ module of software. The descriptor values used in the equation generation are shown in Table IV. Molecular descriptors and statistical methods The physicochemical descriptors were taken as independent variable and biological activity (in terms

INDIAN J. CHEM., SEC B, FEBRUARY 2010

232

Table IV ― Descriptors used in QSAR equations generation Comp. No.

a

ETor

b

LogP

c

EElec

d

EHOMO

-2.1674 -2.3700 -10211.9 -9.1142 Sulf-1 -5.2721 -2.3700 -9954.1 -9.1578 Sulf-2 -5.3174 -0.5079 -11443.2 -9.4226 Sulf-3 -6.2882 1.2662 -9846.5 -10.1215 Sulf-4 -5.8482 -0.1760 -11322.3 -9.9189 Sulf-5 -5.9212 0.1037 -12858.4 -9.9724 Sulf-6 -4.3819 0.1344 -11800.0 -9.1932 Sulf-7 -5.2813 0.5345 -11557.1 -9.1738 Sulf-8 -6.0143 0.8052 -11479.4 -9.2004 Sulf-9 -6.1299 1.3337 -11429.1 -9.2378 Sulf-10 -0.5490 -0.3575 -26586.6 -0.9165 Sulf-11 -4.3739 -0.7204 -19169.2 -9.8118 Sulf-12 9.6523 0.4331 -9119.31 -9.6832 Sulf-13 6.8654 0.4639 -10597.4 -9.0479 Sulf-14 1.4189 0.8400 -23738.6 -9.5855 Sulf-15 -11.6554 0.6898 -27631.2 -9.1118 Sulf-16 -11.5222 1.6010 -28232.6 -9.2851 Sulf-17 9.9196 0.7332 -19119.6 -9.0467 Sulf-18 2.1444 1.5858 -13435.6 -9.3305 Sulf-19 5.4905 2.8662 -16350.6 -9.7595 Sulf-20 -6.0177 0.2047 -11432.9 -10.2573 Sulf-21 -5.1845 0.4844 -12909.8 -10.0476 Sulf-22 -7.1746 0.3361 -12622.0 -10.4777 Sulf-23 -6.0841 -0.5079 -11970.7 -9.0227 Sulf-24 a ETor:Torsion energy; b LogP: Octanol/water partition coefficient; c EElec: Electronic energy; d EHOMO: Highest Occupied Molecular Orbital Energy; e Rad: Radius of gyration

e

Rad 3 3 4 3 4 4 3 3 3 3 4 3 3 3 6 7 7 5 4 4 4 4 4 3

of negative logarithms of inhibition constant) values as dependent variable. The sequential multiple linear regression analysis method was employed for generation of equations using VALSTAT program28. In sequential multiple linear regression, the program searches were made for all permutations and combinations sequentially for the data set. In this case it searched for more than 10,000 combinations and gave 10 multivariate equations based on squared correlation coefficient. The auto-correlated parameters were eliminated depending on their individual correlation with the biological activity in order to avoid simple collinearity problem. All possible combinations of descriptors were considered for the QSAR study. In the present study, an attempt has been made to find structural requirement for inhibition of carbonic anhydrase using QSAR approach with electronic, thermodynamic, steric and topological descriptors.

The predictive power of the equations were validated by Leave-one-out (LOO) cross-validation equations29,30. Predicted residual sum of square (PRESS), cross-validated correlation coefficient r2 (Q2) and standard deviation error of prediction (SDEP) were considered for the validation of these equations. The results from cross-validated analysis were expressed as the cross-validated squared correlation coefficient (Q2); which is defined as: Q2 = 1 – Σ(Ypred – Yact)2 / Σ (Yact – Ymean)2 Where Ypred, Yact, and Ymean are predicted, actual and mean values of the target property (pKi) respectively. Σ(Ypred – Yactual)2 is the Predictive Residual Error Sum of Squares (PRESS). PRESS is an important cross-validation parameter as it is a good approximation of the real predictive error of the equations. To further assess the robustness and statistical confidence of the derived equations, bootstrapping analysis was performed. The r2bs is an average squared correlation coefficient calculated during validation, which is computed from a subset of variables used one at a time for validation. The statistical parameters considered to compare and select the generated QSAR equations were correlation coefficient (r), standard deviation (s), sequential Fischer (F) test, Cross-validated correlation coefficient (Q2). A data point is considered as an outlier if it has a large magnitude (when the residual value exceeds twice the standard error of estimate of the equation). To validate the derived equation, the overall predictive ability of that analysis was evaluated by the terms r2pred, and calculated using the formula; r2pred = SD – PRESS/SD; Where SD is the sum of the squared deviation between the biological activities of the test set molecules and the mean activity value of the training set molecules. PRESS is the predictive error sum of squares derived from the leave-one-out method. Acknowledgement Authors are gratefully acknowledging the financial assistance provided by AICTE and CSIR, New Delhi, India for this work and also thankful to Director, SGSITS for providing computational facility. References 1 Supuran C T, Scozzafava A & Conway J, Carbonic Anhydrase-Its Inhibitors and Activators (CRC Press, Boca Raton (FL, USA) 2004, pp. 1–363.

POTHEN et al.: STRUCTURAL OPTIMIZATION OF CARBONIC ANHYDRASE INHIBITORS 2 Stams T & Christianson D W, The Carbonic Anhydrases-New Horizons, (Birkhauser Verlag: Basel) 2000, pp. 159–174. 3 Abbate F, Casini A, Scozzafava A & Supuran C T J, Enzyme Inhib Med Chem, 18, 2003, 303. 4 Casini A, Antel J, Abbate F, Scozzafava A, David S, Waldeck H, Schafer S & Supuran C T, Bioorg Med Chem Lett, 13, 2003, 841. 5 De Simone G, Di Fiore A, Menchise V, Pedone C, Antel J, Casini A, Scozzafava A, Wurl M & Supuran C T, Bioorg Med Chem Lett, 15, 2005, 2315. 6 Menchise V, De Simone G, Alterio V, Di Fiore A, Pedone C, Scozzafava A & Supuran C T, J Med Chem, 48, 2005, 5721. 7 Alterio V, Vitale R M, Monti S M, Pedone C, Scozzafava A, Cecchi A, De Simone G & Supuran C T, J Am Chem Soc, 128, 2006, 8329. 8 Abbate F, Winum J Y, Potter B V, Casini A, Montero J L, Scozzafava A & Supuran C T, Bioorg Med Chem Lett, 14, 2004, 231. 9 Winum J Y, Temperini C E, Cheikh K, Innocenti A, Vullo D, Ciattini S, Montero, J L, Scozzafava A & Supuran C T, J Med Chem, 49, 2006, 7024. 10 Montgomery J C, Venta P J, Eddy R L, Fukushima Y S, Shows T B & Tashian R E, Genomics, 11, 1991, 835. 11 Earnhardt J N, Qian M, Tu C, Lakkis M M, Bergenhem N C, Laipis P J, Tashian R E & Silverman D N, Biochemistry, 37, 1998, 10837. 12 Vullo D, Voipio J, Innocenti A, Rivera C, Ranki H, Scozzafava A, Kailab K & Supuran C T, Bioorg Med Chem Lett, 2005,15, 971. 13 Lakkis M M, Shea K S & Tashian R E, J Histochem Cytochem, 45, 1997, 657. 14 Ruusuvuori E, Li H, Huttu K, Palva J M, Smirnov S, Rivera C, Kaila K & Voipio J J, Neurosci, 24, 2004, 2699. 15 Scozzafava A, Mastrolorenzo A & Supuran C T, Expert Opin Ther Patent, 14, 2004, 667. 16 (a) Christianson D W & Fierke C A, Acc Chem Res, 29, 1996, 331; (b) Whittington D A, Grubb J H, Waheed A,

17 18 19 20 21

22

23 24 25 26 27

28 29 30

233

Shah G N, Sly W S & Christianson D W, J Biol Chem, 279, 2004, 7223. Abbate F, Supuran C T, Scozzafava A, Orioli P, Stubbs M T & Klebe G, J Med Chem, 45, 2002, 3583. Casini A, Antel J, Abbate F, Scozzafava A, David S, Waldeck H, Schafer S & Supuran C T, Bioorg Med Chem Lett, 13, 2003, 841. Weber A, Casini A, Heine A, Kuhn D, Supuran C T, Scozzafava A & Klebe G, J Med Chem, 47, 2004, 550. Menchise V, De Simone G, Alterio V, Di Fiore A, Pedone C, Scozzafava A & Supuran C T, J Med Chem, 48, 2005, 5721. a) Supuran C T, Vullo D, Manole G, Casini A & Scozzafava A, Curr Med Chem, 2, 2004, 49; b) Supuran C T, Scozzafava A & Conway J, Carbonic anhydrase-Its inhibitors and activators, (CRC Press, Boca Raton (FL), USA) 2004, pp. 1-363. a) Krungkrai J, Scozzafava A, Reungprapavut S, Krungkrai S R, Rattanajak R, Kamchonwongpaisan S & Supuran C T, Bioorg Med Chem, 13, 2005, 483; b) Marcus E A, Moshfegh A P, Sachs G & Scott D R, J Bacteriol, 187, 2005, 729. Gupta S P, Chem Rev, 89, 1989, 1765. Huo G, Katzenellenbogen J A, Garg R & Hansch C, Chem Rev, 99, 1999, 723. Supuran C T & Scozzafava A, Bioorg Med Chem, 15, 2007, 4336. Vullo D, Voipio J, Innocenti A, Rivera C, Ranki H & Scozzafava A, Bioorg Med Chem Lett, 15, 2005, 971. CS Chem office molecular modeling software, version 8.0. Cambridge Softcorporation, Software Publishers association, 1730 M street, NW, suite 700, Washington D.C. 20036 (202), 452-1600, USA. Gupta A K, Babu M A & Kaskhedikar S G, Indian J Pharm Sci, 66, 2004, 396. Schaper K J, Quant Struct Act Relat, 18, 1999, 354. Wold S & Eriksson L, QSAR: Chemometric Methods in Molecular Design, (VCH, Weinheim) 1995, Vol. 2, 321.