Home Archive Vol 35, No.2, 2009 Original Papers The QSPR Study of Water - Octanol Partition Coefficients for a Series of 24 Benzene Derivatives

The QSPR Study of Water - Octanol Partition Coefficients for a Series of 24 Benzene Derivatives

P.G. Anoaica(1), Emilia Amzoiu(1), L. Giubelan(2)

(1) Faculty of Pharmacy, University of Medicine and Pharmacy, Craiova, (2) Faculty of Medicine, University of Medicine and Pharmacy, Craiova

Abstract: A series of 24 benzene derivatives are commonly used as solvents or reactants in the chemical process of synthesis. For this reason, a circumstantial study of skin permeability (or biologic membranes) to such classes of substances is enforced. This paper presents a correlation between the log P partition coefficients for these compounds and the descriptors characterizing the molecular structures. These descriptors were calculated using quantum ab initio molecular methods as Polarizable Continuum Model (PCM) for water and 1-octanol solvents which feign biological phase (the interface between water medium and cellular membrane). The partition coefficients analysis depends on the energies of solvation, molecular shape descriptors (area / volume of molecular cavity and Connolly area).

Keywords: benzene derivatives, partition coefficient, transmembranar model, QSPR study

Introduction

The measurement of integument skin absorption from exogenous substance is of big interest in interconnection areas of chemical applicability, as pharmaceutics or cosmetics, for the toxicological evaluation and risk of this utility [1].

It is nearly impossible to estimate the penetrability of skin for all these new types of substance, especially the new substance or prescription, as it uses merely “in vivo” experiments. Therefore, the models “ex vivo” and “in vitro” were used for the evaluation of risk and injuriousness associated with exposure of human skin to exogenous substances [2,3].

The methodology of QSAR – type, rather QSPR [4] method is very useful for us to establish a predictive model for the biological activity of a homolog series of chemical substances used commonly solvents or mediators in a different chemical industrial process.

Although the biological activity for skin penetrability refers precisely (in an essential way) to the process of diffusion, we consider that such a diffusion succeeds at last in a partition between two immiscible phases of solvents, among which one represents the human skin.

The solvent 1-octanol feigns biologic membrane [5,6,7] in the best way due to its special structure, of hydrophilic –OH specificity and hydrophobic (lipophilic) character owing to a long catena of carbon atoms (Figure 1).

Figure 1: The biologic membrane simulation in 1-octanol model

At the interface between water and lipidic phase (the lated simulate by 1-octanol) an organized appears structure due to the polar character of water and also to the amphipathic character of the 1-octanol molecule [8].

Figure 2: Substituted derivatives of benzene shown in Tabele 1

Such arrangement of molecules at the interface water – octanol will better simulate the partition processes in the cellular membrane.

The chemical substance studied in this article was represented by 24 substituted derivatives of benzene (figure 2). From these, the partition coefficients (log P) between water and octanol were established experimentally and were known in literature [9]. In Table 1 this 24 value was shown.


Table 1: The partition coefficients for 24 substituted derivatives of benzene

Nr.

R

Molecule

log P

Nr.

R

Molecule

log P

1

H–

C6H6

2.13

13

H2C=CH–

C6H5-CHCH2

2.95

2

F–

C6H5-F

2.27

14

HC≡C–

C6H5-CCH

2.53

3

Cl–

C6H5-Cl

2.84

15

–C3H7

C6H5-C3H7

3.68

4

Br–

C6H5-Br

2.99

16

H2C=CH-CH2

C6H5-CH2-CHCH2

3.23

5

HS–

C6H5-SH

2.52

17

i-C3H7

i-C3H7-C6H5

3.66

6

CH3

C6H5-CH3

2.69

18

n-C4H9

n-C4H9-C6H5

4.26

7

CH2Br–

C6H5-CH2Br

2.92

19

t-butil

t-butil-C6H5

4.11

8

CH2Cl–

C6H5-CH2Cl

2.30

20

ciclopropil

ciclopropil-C6H5

3.27

9

CH3S–

C6H5-SCH3

2.74

21

ciclopentil

ciclopentil-C6H5

4.27

10

CF3

C6H5-CF3

3.01

22

ciclohexil

ciclohexil-C6H5

4.64

11

CF3S–

C6H5-SCF3

3.57

23

C6H5

C6H5-C6H5

4.09

12

C2H5

C6H5-C2H5

3.15

24

C6H5CH2

C6H5-CH2-C6H5

4.14


A comparative study between structure and property (QSPR) assumes in fact a statistical correlation between bio-chemical properties (log P in this case) and physical or chemical properties (descriptors) which represent in fact the chemical structure of the analyzed substance. These descriptors constitute the interface correlation between bio-chemical property and chemical structure of the substance [10].

There are large numbers of these descriptors in special literature publications of last decade [11]. The reason for large numbers of descriptors is the incapacity to explain – in many cases - of the cause – effect interaction between structures and the activity or mechanism of this interaction (for ex: interaction of medicament (ligand) and receptor – in pharmacology).

In this QSPR study we analyze the structural descriptors involve in partition coefficients for benzene derivatives. Such a study allows the better knowledge of interaction between solvents and dermatome-membranes. This QSAR way helps the labor study and was carried out with success between pollutants or nocive substances and derma interaction, with hope of artificial membrane optimization [12].

For molecular descriptor descriptors calculation involved in partition process we effectuated the chemical design of the structure that was studied with MM+ Molecular Mechanics and RHF and PM3 semi-empirical method (MOPAC 7.0). The atomic coordinate obtained in this way was used at input data in quanto-molecular ab initio calculus RHF, PCM, STO 6G (GAMESS) about Restricted Hartree Fock approximation. In this method the atomic orbital – tip Slater – compute with 6 Gaussian function (STO – 6G) [13,14,15,16] was used.

Descriptors for molecular shape and electrostatic interaction

The solvents water and 1-octanol were in PCM model (Polarizable Continuum Model) incorporate in GAMMES-soft. Thus, the solvent is described as a macroscopic continuum medium with a specified dielectric constant. The immersion of the solute molecule inside the solvent created a specific molecular cavity. In this model the interaction between solute and solvent molecules computed such a modification of electro-magnetic field in reaction with other dielectric media.

In PCM approach were calculated the surface and volume for molecular cavity with the intersection and reunion of Van der Waals sphere of atoms in molecule – Figure 3.

For the water solvent this dielectric constant er = 78.4 (relative to vide or air) was used with maximal raze of rotating sphere rw = 1.385 Å and for the 1-octanol solvent er = 10.3 and ro = 3.25 Å.

In correlation with these molecular descriptors of molecular shape in solute we evaluate the Connolly descriptors of molecular shape [17] such as CSAA – Connolly Solvent Accessible Area and CSEV - Connolly Solvent Excluded Volume expose in Figure 4.

Figure 3: Molecular cavity in solute – solvent interaction model

Figure 4: Connolly Solvent Accessible Area (CSAA) and Connolly Solvent Excluded Volume (CSEV)

Results and Discussion

In Table 2 the value of molecular descriptors of shape for these 24 of molecule involved was described.

We also calculated, apart from of shape descriptors, the energy of electrostatic interaction between solute and solvent with ab initio quanto-molecular procedure. These values depend on the partition of solute among two immiscible phases.

 


Table 2: Molecular shape descriptors value for the 24 benzene derivatives

Nr.

log P

CSAA

CSAA#

CSEV

CSEV#

CA

CV

1

2.13

232.453

481.432

70.805

74.213

109.703

83.510

2

2.27

239.083

491.119

74.263

77.986

113.953

86.385

3

2.84

255.698

515.450

84.871

89.023

127.229

98.645

4

2.99

263.831

527.309

90.550

94.879

132.924

104.640

5

2.52

259.492

520.844

87.328

91.609

132.387

102.729

6

2.69

262.983

525.375

87.374

92.608

131.075

100.111

7

2.92

293.926

569.043

108.830

116.671

155.075

121.793

8

2.30

286.710

559.165

102.966

110.094

149.397

116.090

9

2.74

291.838

568.098

105.990

111.748

152.501

119.610

10

3.01

280.295

549.957

98.054

104.410

143.827

109.182

11

3.57

310.590

595.142

117.229

124.178

166.943

129.191

12

3.15

291.963

565.854

105.544

113.464

152.769

116.574

13

2.95

276.319

546.316

94.661

99.237

141.371

110.606

14

2.53

273.975

545.588

88.029

93.076

136.931

105.864

15

3.68

322.621

610.523

122.700

132.790

174.593

133.717

16

3.23

312.620

596.975

114.351

123.486

166.249

128.337

17

3.66

315.517

597.930

122.558

133.357

172.148

133.147

18

4.26

353.247

655.312

139.821

151.548

196.316

150.412

19

4.11

329.335

616.486

138.489

148.857

186.396

148.893

20

3.27

306.745

586.314

112.864

122.774

164.598

127.017

21

4.27

346.534

641.780

145.293

156.329

194.418

156.196

22

4.64

369.906

673.817

163.646

178.173

214.055

172.604

23

4.09

339.269

637.668

131.694

138.044

184.394

153.636

24

4.14

372.197

682.178

153.250

164.204

213.642

172.219

CSAA – Connolly Solvent Access Area (Å2);

CSEV – Connolly Solvent Excluded Volume (Å3);

CA – Surface Area (Å2); CV – Cavity Volume (Å3)

The # signifies the value of interaction energy between solvent and solute (ours) molecule, while the other values was for water.


Table 3 summarizes the calculated value of interaction for solvent – solute energies.

The molecular descriptor values in Table 2-3 were calculated using the ab initio quantomolecular procedure (RHF, STO-6G GAMESS) for water and 1-octanol solvents. For the CSAA (Connolly Accessible Surface Area) and CSEV (Connolly Solvent Excluded Volume) molecular descriptors, we used Chem 3D programs (ChemOffice 2005 Ultra).

All molecular descriptors summarized in Table 2 and 3 were imported in CODESSA file and the correlation with partition coefficients (log P) was effectuated with multiple linear regression technique.

The examination of the contribution of shape descriptors and solute – solvent interaction descriptors summarized in Table 4 is very suggestive

As it is shown in Table 4, all 4 combinations of two descriptors of shape have the same correlation coefficient, R2 = 0.88. This aspect denotes the importance of molecular shape in the process of partition between the two non-miscible phases, due to the of interaction access of solvent - solute molecules. In this direction the solvent – solute interaction must depend on the solvent medium radius of the two molecules involved. To record this aspect, the most sensible descriptor is the Connolly Solvent Excluded Volume (CSEV). This shape descriptor indirectly describes the access of the solute molecule to the solvent.

Also, we present in table 5 the results of the correlation in multiple linear regressions to mark the contribution of these shape descriptors.

In fact the electrostatic interaction (EI) alone does not have a significant contribution in the partition process. In this case the correlation coefficient is insignificant (R² = 0.3832). Therefore, the single descriptor TFES contribution is irrelevant (R2 = 0.043543).

Actually, this descriptor summarizes many interactions (TFES = IES + TI = EI + PCE + RFE + IES). Individually grouped, this specifies a very good correlation with a great coefficient (0.9161) [12].

Similarly, TI consists of many specific interaction (TI = EI + PCE + RFE). If that were individually grouped the corelational coefficient would be better (R2 = 0.906).

 


Table 3: Solute – solvent interaction energy (kcal/mol)

Nr.

TFES

TFES#

RFE

RFE#

PCE

PCE#

EI

EI#

TI

TI#

IES

1

-144392.68

-144396.59

1.96

1.96

14.79

10.75

-0.94

-0.79

15.81

11.92

-144408.49

2

-206127.92

-206131.97

1.74

1.74

15.32

11.12

-1.37

-1.18

15.69

11.68

-206143.61

3

-431417.48

-431422.03

1.78

1.78

16.70

11.96

-1.64

-1.41

16.84

12.32

-431434.31

4

-1749668.2

-1749671.9

1.70

1.70

17.26

12.29

-10.32

-8.64

8.64

5.34

-1749676.8

5

-393011.82

-393016.34

2.10

2.08

17.51

12.59

-2.98

-2.48

16.63

12.19

-393028.45

6

-168836.15

-168840.84

2.35

2.32

17.82

13.01

-1.02

-0.85

19.15

14.49

-168855.30

7

-1774076.7

-1774081.7

2.09

2.05

20.40

14.63

-5.70

-4.84

16.79

11.83

-1774093.4

8

-455859.70

-455864.9

2.16

2.13

19.83

14.29

-2.78

-2.34

19.21

14.08

-455878.91

9

-417456.73

-417462.01

2.51

2.48

20.37

14.74

-2.94

-2.47

19.94

14.75

-417476.67

10

-354043.11

-354048.14

1.71

1.69

19.43

14.13

-2.36

-2.02

18.78

13.80

-354061.89

11

-602655.71

-602661.66

1.93

1.90

22.20

16.01

-2.07

-1.75

22.06

16.16

-602677.78

12

-193277.02

-193282.5

2.75

2.69

20.90

15.31

-1.04

-0.85

22.61

17.16

-193299.64

13

-192509.12

-192514.17

2.39

2.35

19.03

13.82

-1.27

-1.05

20.14

15.13

-192529.26

14

-191748.54

-191753.40

2.12

2.13

18.18

13.10

-1.77

-1.50

18.54

13.72

-191767.08

15

-217718.18

-217724.49

3.12

3.00

24.00

17.63

-1.08

-0.88

26.03

19.75

-217744.22

16

-216947.27

-216953.23

2.84

2.78

22.58

16.47

-1.24

-1.02

24.18

18.24

-216971.44

17

-217717.99

-217724.15

3.13

3.06

23.66

17.38

-1.11

-0.90

25.68

19.55

-217743.67

18

-242159.30

-242166.34

3.52

3.37

27.08

19.93

-1.23

-0.93

29.37

22.36

-242188.67

19

-242154.38

-242161.07

3.32

3.23

25.67

18.87

-1.23

-1.00

27.76

21.10

-242182.14

20

-216952.35

-216958.24

2.89

2.87

22.43

16.40

-1.17

-0.98

24.15

18.29

-216976.00

21

-265873.59

-265880.56

3.61

3.57

26.80

19.72

-1.07

-0.89

29.34

22.40

-265902.93

22

-290322.14

-290329.83

3.88

3.78

29.63

21.85

-1.14

-0.92

32.38

24.71

-290354.52

23

-288069.37

-288075.92

3.05

3.06

24.67

17.85

-1.83

-1.52

25.89

19.38

-288095.26

24

-312510.50

-312518.12

3.40

3.33

28.86

21.00

-1.82

-1.46

30.44

22.87

-312540.94

TFES – Total Free Energy in Solvent (TFES = IES + TI);

RFE – Repulsion Free Energy;

PCE – Pierotti Cavity Energy;

TI – Total Energy (TI = RFE + PCE + EI);

IES – Internal Energy in Solvent


Table 4: Multiple linear regression: log P = a0 + a1·Desc.1 + a2·Desc.2

 

Desc.

R2

CSAA

CSAA#

CSEV

CSEV#

CA

CA#

CV

CV#

0.882

×

×

 

 

 

 

 

 

0.896

 

 

×

×

 

 

 

 

0.878

 

 

 

 

×

×

 

 

0.876

 

 

 

 

 

 

×

×

“Desc.” is one of the descriptors: CSAA - Connolly Solvent Access Area (Å2), CSEV - Connolly Solvent Excluded Volume (Å3), CA –Surface Area (A2) and The sign # simbolise values computed for 1-octanol solvent (non mark – for water solvent);

R2 is correlational coefficient.


 

Moreover, if TFES descriptor (which represents the total energy of salvation in the interaction between solute and solvent) was associated with shape descriptors, an optimization in previous correlations would result with an increase of regressional coefficients S and R2 = 0.88÷0.90.

1. TFES + CA:

log P = - 0.361 + 0.0236·CA - 0.00600·TFES + 0.00600·TFES#

S = 0.2646; R2 = 0.883

2. TFES + CV:

log P = - 0.025 + 0.0276·CV - 0.00619·TFES + 0.00619·TFES#

S = 0.2663      R2 = 0.881

3. TFES + CSEV:

log P = 0.146 + 0.0866·CSEV - 0.0528·CSEV# - 0.00951·TFES

S = 0.2433; R2 = 0.906

4. TFES + CSAA:

log P = 4.18 - 0.00351·TFES + 0.00351·TFES# + 0.0730·CSAA - 0.0393·CSAA#

S = 0.2698; R2 = 0.884

This result is in concordance with Tanaka results [17] for shape descriptors CSAA and salvation energies in their case.


Table 5: Multiple linear regression: log P = a0 + S (ai·Desc.i)

 

Desc.

R2

TI

TI#

EI

EI#

PCE

PCE#

RFE

RFE#

TFES

TFES#

IES

0.383

 

 

×

×

 

 

 

 

 

 

 

0.868

×

×

 

 

 

 

 

 

 

 

 

0.043

 

 

 

 

 

 

 

 

×

×

 

0.862

×

×

 

 

 

 

 

 

 

 

×

0.906

 

 

×

×

×

×

×

×

 

 

 

0.916

 

 

×

×

×

×

×

×

 

 

×

Now “Desc.” is one of the descriptors: EI – Electrostatic Interaction; PCE – Pierroti Cavity Energy; RFE - Repulsion Free Energy; TFES – Total Free Energy in Solvent; IES – Internal Energy in Solvent, TI = EI + PCE + RFE; TFES = IES + TI = EI + PCE + RFE + IES

The # signifies the value of interaction energy between solvent and solute (ours) molecule, while the other values were for water.


Conclusion

In this paper a possible correlation between partition coefficient (log P) and 24 structural descriptors for benzene derivatives was studied.

1. The shape molecular descriptors for benzene derivatives (CSAA, CSEV, CA, CV) have the same correlation coefficient R2 = 0.88. These results have shown the role played by the shape of the molecules in the partition process between the two non-miscible phases. This fact is correct because the interaction solute-solvent depends on the access of the molecular solvent to the molecular solute.

2. The single electrostatic interaction (EI) does not contribute significantly to the partition process. Now the corelational coefficient has only the value R2 = 0.3832. Also, the contribution of TFES descriptor is irrelevant (R2 = 0.043543). The electrostatic interaction in detail TFES = TI + IES = EI + PCE + RFE + IES) shows a correlation coefficient very close to unity (0.9161). Similarly, the TI descriptor consists of many specific interactions (TI = EI + PCE + RFE) very well correlated in separate mode (R2 = 0.906).

3. The TFES descriptor (it represents a total energy of salvation in solute- solvent interaction)  in association with shape descriptors calculate a very relevant corelational coefficient 0.88 ÷ 0.90.

References

1.     E. Merică, Tehnologia produselor cosmetice”, Editura Kolüs, Iaşi, p. 225-226

2.     Understanding variation in partition coefficient”, Kd, values, USA Environmental Protection Agency, Office of Air and Radiation, I, 1999, 212

3.     G.S.D. Ayton, S. Bardenhagen, P. McMurtry, D. Sulsky and G.A. Voth, IBM J. Res. & Dev., 2001, 45, 3/4, 417-420

4.     C. Amovilli and B.Mennucci, J. Phys. Chem. B, 1997, 101, 1051-1054

5.     B. Mennucci and J. Tomasi, J. Chem. Phys., 1997, 106, 5151-5155

6.     C.W. Wimley, T.P. Creamer and S.H. White, Biochemistry, 1996, 35, 5109-5111

7.     C.W. Wimley, S.H. White, Nat. Struct. Biol., 1996, 10 / 3, 842-845

8.     H. Chuman, A. Mori and H. Tanaka, Analytical Sciences, Japan Society for Analytical Chemistry, 2002, 18 / 9, 1015-1018

9.     L. Tarko, Rev. Chim. (Bucharest), 2007, 58 / 2, 191

10. M. Karelson, “Molecular Descriptors in QSAR/QSPR studies”, Wiley Intersc., New York, 2000

11. C.T. Zawodzinski, M. Eikerling, L. Pratt, A. Redondo, T. Rockward, M. Hickner and J. McGrath, “Membranes for Operation above 100, Materials Science and Technology Division”, Los Alamos National Laboratory, Report, 2002

12. M.W. Schmidt, K.K. Baldrige, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen, S. Koseki, N. Matsunaga, K.A. Nguyen, S.J. Su, T.L. Windus, M. Dupuis and J.A. Montgomery, “The General Atomic and Molecular Electronic Structure System.”, J. Comput. Chem., 1993, 14, 1347.

13. R.G. Parr and W. Yang, “Density functional Theory of Atoms and Molecules”, New York, Oxford University Press, 1989

14. R.G. Pearson, Absolute electronegativity and hardness: applications to organic chemistry, J. Org. Chem., 1989, 54, 1423.

15. R.G. Parr and Z. Zhou, “Absolute Hardness: Unifying concept for identifying shells and subshells in nuclei, atoms, molecules, and metallic clusters”, Chem.Res., 1993, 26, 256

16. Z. Zhou and R.G. Parr, “Activation hardness: new index for describing the orientation of electrophilic aromatic substitution”, J. Am. Chem. Soc., 1990, 112, 5720

17. M.L. Connolly, “Solvent - accessible surfaces of proteins and nucleic acids”, Science, 1983, 221, 709-713


 

Corresponding Adress: P.G. Anoaica PhD, Faculty of Pharmacy, University of Medicine and Pharmacy, Craiova gabriel@umfcv.ro

 


All articles in this issue