Professional Documents
Culture Documents
Abstract: A new drug design method, the multiple eld three-dimensional quantitative structureactivity relationship (MF-3D-QSAR), is proposed. It is a combination and development of classical 2D-QSAR and traditional 3DQSAR. In addition to the electrostatic and van der Waals potentials, more potential elds (such as lipophilic potential, hydrogen bonding potential, and nonthermodynamic factors) are integrated in the MF-3D-QSAR. Meanwhile, a
principal component analysis (PCA) and iterative double least square (IDLS) technique is developed for predicting
the bioactivity of query drug candidates. As an example, the MF-3D-QSAR is applied to the design of neuraminidase inhibitor and to prove its predictive power, and some useful ndings are obtained for developing drugs against
inuenza virus.
q 2007 Wiley Periodicals, Inc.
Key words: 3D-QSAR; potential elds; drug design; neuraminidase inhibitor; H5N1 inuenza
Introduction
The MF-3D-QSAR is actually a direct extension and development of the traditional 2D-QSAR. Typically, the 2D-QSAR
approach is represented by a linear free energy equation13
logAb ab fhydr Xh ae felec Xe ast fst Xs
bj ui rj pKi
(2)
j1
(1)
Mi
X
qk
r
k1 j;k
Mi
X
k1
"
ek
k
rj;k
12
6 #
k
2
(3)
rj;k
In eq. (3) only the electrostatic interaction and van der Waals
interaction are calculated using Coulomb equation and Lennard-
212
i 1; 2; . . . ; M
(4)
i i
i 1; 2; . . . ; M
(5)
i
i 1; 2; . . . ; M
(6)
bj
K
X
!
ak ei;j;k
pKi
i 1; 2; . . . ; M
(7)
k1
DOI 10.1002/jcc
K
X
ei;j;k ak
213
v
u
M
u1 X
n1
n1 2
Q
Qn t
pKi pKi
M i1
v
u M
u1 X
n 2
pKi pKi
t
e
M i1
16
(8)
k1
(9)
bj
,
1
(10)
Then the values of fbj g are used to reduce the 3D data matrix
EMLK to a 2D data matrix HMK with the elements given by
hi;k
K
X
ei;j;k bj
(11)
j1
HMK ak pKM
(12)
The above equation can be solved by using the least square ap1
proach, leading to the solution of sensitive coefcients fal g,
1
t
t
HMK 1 HMK
pKM
aK HMK
(13)
Then the values of {ak } are used for the new solutions of the
2
sensitive coefcients {bj } of the principal components of space
points. The above procedure is performed iteratively for n steps,
n
i.e., until reach the converged solutions as denoted by fak g and
n
n
n
{bj }. Now, the values of {ak } and {bj } can be used to predict the bioactivities of the ith drug candidate through the following equations:
it
pKipred
K
X
k1
n
ak
L
X
(14)
!
n
ei;k;j bj
(15)
j1
DOI 10.1002/jcc
214
Table 1. The Structures and Experimental pIC50 of the 32 Molecules Used in the Benchmark Dataset.
M03 (9.52)
M04 (7.92)
M05 (7.66)
M06 (9.00)
M07 (9.00)
M08 (6.21)
M09 (6.28)
M10 (8.00)
M11 (9.00)
M12 (9.00)
(continued)
DOI 10.1002/jcc
215
Table 1. (Continued)
M13 (8.52)
M14 (6.22)
M15 (6.74)
M16 (6.68)
M17 (6.70)
M18 (6.57)
M19 (5.20)
M20 (8.05)
M21 (9.00)
M22 (5.43)
M23 (7.22)
M24 (5.70)
(continued)
DOI 10.1002/jcc
216
Table 1. (Continued)
M25 (6.89)
M26 (6.74)
M27 (6.52)
M28 (6.70)
M29 (7.05)
M30 (6.28)
M31 (5.66)
M32 (5.70)
Mi
X
lk
r
1
j;k
k1
(17)
DOI 10.1002/jcc
Conclusion
The MF-3D-QSAR is a direct extension of the classical 2D-QSAR
and retains the advantages of traditional 3D-QSAR. The MF-3D-
217
Coefcients
20
fak g
Potential
aelec
avdw
alip
ahbd
ahba
Initial
Converged
1.0
1.0
1.0
1.0
1.0
1.5365
8.6323
0.4440
2.2793
0.8046
DOI 10.1002/jcc
218
Molecule
M01
M02
M03
M04
M05
M06
M07
M08
M09
M10
M11
M12
M13
M14
M15
M16
M17
M18
M19
M20
M21
M22
M23
M24
M25
M26
M27
M28
M29
M30
M31
M32
Rcalc 0.9627
Rpred 0.8194
a
pIC50_expta
6.650
7.220
9.520
7.920
7.660
9.000
9.000
6.210
6.280
8.000
9.000
9.000
8.520
6.220
6.740
6.680
6.700
6.570
5.200
8.050
9.000
5.430
7.220
5.700
6.890
6.740
6.520
6.700
7.050
6.280
5.660
5.700
pIC50_calcb
pIC50_predc
6.357
7.970
7.348
8.232
9.623
10.496
7.914
7.146
7.825
7.241
8.975
8.292
8.971
6.557
6.259
6.534
6.060
5.874
7.620
7.620
8.968
8.666
8.848
9.780
8.596
Unconverged
6.498
6.736
6.498
6.244
6.735
Unconverged
7.199
5.786
6.536
6.321
4.988
4.930
7.791
8.107
8.332
9.772
6.169
5.841
7.268
7.686
6.542
6.040
6.697
6.355
6.758
7.249
6.817
6.563
6.854
7.040
6.854
6.694
5.967
5.187
6.067
7.218
5.147
3.972
Qcalc 60.3226
Qpred 60.8578
ments are from the better theoretical model and the advanced
mathematical technique of MF-3D-QSAR.
The bioactivities of drugs are affected by many factors,
including thermodynamic and nonthermodynamic factors. In
additional to the electrostatic and van der Waals interactions
(described by Coulomb law and Lennard-Jones equation) there
are many other factors. In MF-3D-QSAR, the potential elds are
not limited in the thermodynamic potentials; they may also
include other physical and chemical factors that affect either receptor recognition or activation of drug candidates. For further
improvement of MF-3D-QSAR, we need better linear free
energy equation, including more potential elds, and advanced
parametrization for the potential elds.
Acknowledgments
The authors wish to express their gratitude to the anonymous
reviewers whose constructive comments are very helpful for
strengthening the presentation of this paper. The work is supported by the Chinese National Basic Research Program (973)
under the project 2004CB719606, and by the Chinese National
Science Foundation (NSFC).
References
1. Hansch, C. Drug Design; Academic Press: New York, 1971; pp.
271275.
2. Hansch, C.; Leo, A. J. Substitute Constants for Correlation Analysis
in Chemistry and Biology; Wiley: New York, 1979.
DOI 10.1002/jcc
3. Hansch, C.; Streich, M.; Geiger, F.; Muir, R. M.; Maloney, P. P.;
Fujita, T. J Amer Soc 1963, 85, 2817.
4. Klopman, G.; Wang, S.; Jacobs, M. R.; Ellner, J. J. Antimicrob
Agents Chemother 1993, 37, 1807.
5. Klopman, G.; Wang, S.; Jacobs, M. R.; Bajaksouzian, S.; Edmonds,
K.; Ellner, J. J. Antimicrob Agents Chemother 1993, 37, 1799.
6. Klopman, G.; Fercu, D.; Renau, T. E.; Jacobs, M. R. Antimicrob
Agents Chemother 1996, 40, 2637.
7. Klopman, G.; Fercu, D.; Li, J. Y.; Rosenkranz, H. S.; Jacobs, M. R.
Res Microbiol 1996, 147, 86.
8. Cramer, R. D.; Patterson, D. E.; Bunce, J. D. J Am Chem Soc 1988,
110, 5959.
9. Cramer, R. D., III; Patterson, D. E.; Bunce, J. D. Prog Clin Biol Res
1989, 291, 161.
10. Du, Q.; Arteca, G. A. J Comput Aided Mol Des 1996, 10, 133.
11. Du, Q.; Arteca, G. A.; Mezey, P. G. J Comput Aided Mol Des
1997, 11, 503.
12. Du, Q.; Mezey, P. G. J Comput Aided Mol Des 1998, 12, 451.
13. Du, Q.; Liu, P.; Mezey, P. J Chem Inf Model 2005, 45, 347.
14. Du, Q.; Mezey, P.; Chou, K. C. J Comput Chem, 2005, 26, 461.
15. Du, Q.; Li, D.; He, W. Z.; Chou, K. C. J Comput Chem 2006, 27, 685.
16. Viswanadhan, V. N.; Ghose, A. K.; Hanna, N. B.; Matsumoto, S. S.;
Avery, T. L.; Revankar, G. R.; Robins, R. K. J Med Chem 1991, 34,
526.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
219
Klebe G.; Abraham U.; Mietzner T. J Med Chem 1994, 37, 4130.
Klebe, G.; Abraham, U. J Comput Aided Mol Des 1999, 13, 1.
Waller, C. L.; Marshall G. R. J Med Chem 1993, 36, 2390.
Jain, A. N.; Koile, K.; Chapman, D. J Med Chem 1994, 37, 2315.
Loew, G. H.; Villar, H. O.; Alkorta, I. Pharm Res 1993, 10, 475.
Kellogg, G. E.; Abraham, D. J. J Mol Graphics 1992, 10, 212.
Kim, K. H. Quant Struct-Act Relat 1993, 12, 232.
Maring, C. J.; Stoll, V. S.; Zhao, C.; Sun, M.; Krueger, A. C.; Stewart, K. D.; Madigan, D. L.; Kati, W. M.; Xu, Y.; Carrick, R. J.;
Montgomery, D. A.; Kempf-Grote, A.; Marsh, K. C.; Molla, A.;
Steffy, K. R.; Sham, H. L.; Laver, W. G.; Gu, Y. G.; Kempf, D. J.;
Kohlbrenner, W. E. J Med Chem 2005, 48, 3980.
Smith, B. J.; Colman, P. M.; Von Itzstein, M.; Danylec, B.; Varghese, J. N. Protein Sci 2001, 10, 689.
Wei, D. Q.; Du, Q. S.; Sun, H.; Chou, K. C. Biochem Biophys Res
Commun 2006, 344, 1048.
Wang, S. Q.; Du, Q. S.; Chou, K. C. Biochem Biophys Res Commun 2007, 354, 634.
Dunn, C. J.; Goa, K. L. Drugs 1999, 58, 761.
Lew, W.; Chen, X.; Kim, C. U. Curr Med Chem 2000, 7, 663.
Chou, K. C. Curr Med Chem 2004, 11, 2105.
Chou, K. C. Biochem Biophys Res Commun 2004, 316, 636.
Chou, K. C.; Zhang, C. T. Crit Rev Biochem Mol Biol 1995, 30, 275.
Chou, K. C.; Shen, H. B. J Prot Res 2007, 6, 1728.
DOI 10.1002/jcc