Chromatographic and Calculation Methods for Analysis of the Lipophilicity of Newly Synthesized Thiosemicarbazides and their Cyclic Analogues 1 , 2 , 4-Triazol-3-thiones

This paper describes the evaluation of the lipophilicity of newly synthesized thiosemicarbazides and their cyclic analogues 1,2,4-triazol-3-thiones obtained using experimental and calculated methods. Previous studies have shown these compounds have antibacterial activity. The chromatographic behavior of analyzed compounds was studied by reversed phase high performance liquid chromatography (RP-HPLC) and reversed phase thin layer chromatography (RP-TLC). The aqueous mobile phases containing methanol were used in order to determine retention parameter (RM) and capacity factors (log k) of analyzed compounds. The lipophilicity parameters were obtained by linear extrapolation and they were compared with the calculated log P obtained using several software packages. The results indicate that both experimental chromatographic methods yielded similar results, and these methods are appropriate for determining the lipophilicity of analyzed compounds. High values of correlation coefficients between the log P values calculated using known algorithms (milogP, ALOGPs, AClogP, AlogP, MLOGP, KOWWIN, XLOGP2, XLOGP3) and the experimental data were obtained. Eight standard solutes with known log POW were analyzed under the same conditions as the tested substances in order to determine the log PHPLC and log PTLC parameters. A good correlation was obtained between log kw (or RMW) and the slope. All tested compounds were in agreement with the rule of five claims by Lipiński. The calculated log P values were experimentally confirmed (log PHPLC and log PTLC).


Introduction
Synthesis and confirmation of identity of studied thiosemicarbazide derivatives and their cyclic analogues 1,2,4-triazole-3-thiones were described early. 1 In the cited work, the antibacterial activity of several compounds is presented.
It is well known that the biological activity of some substances is related to their lipophilicity.This parameter determines the bioavailability of the chemical compound and it has been important, in prediction of crossing biological barriers of drug molecules and its interactions with receptors.3][4] For a long time, the distribution coefficient between n-octanol and water was used as an experimental lipophilicity index of the compound.However, due to some limitations and technical barriers, the other methods for determination of the lipophilicity were applied.The chromatographic techniques have proved to be an important alternative method for determination of the lipophilicity.The chromatographic techniques such as reversed phase thin layer chromatography (RP-TLC), [5][6][7][8][9][10] reversed phase high performance liquid chromatography (RP-HPLC), [11][12][13][14] microemulsion electrokinetic chromatography, 15 immobilized artificial membrane (IAM) chromatography, [16][17][18] biopartitioning micellar chromatography (BMC), 16,19 immobilized liposome chromatography (ILC) are commonly used. 16he logarithm of n-octanol-water partition coefficient log P OW is the most frequently used parameter for measuring of lipophilicity and it has been shown that this system is a good model for many biological processes. 20This parameter is also used as one of the standard properties identified by Moreno et al. 20 and Lipiński et al. 21in the "rule of five" for drug-likemolecules.
The classical shake flask method for determining lipophilicity has many disadvantages, i.e., it is time consuming, the quantitative analysis must be used, the log P OW value is limited to the range from 2 to 4. 22 Therefore, many chromatographic methods have successfully been used to determine the lipophilicity of potential drugs.1][22][23][24][25][26][27] Moreover, many calculated methods were used in order to predict the log P OW value. 24,28,29hromatographic parameter of lipophilicity, log k w , obtained by extrapolation to pure water is calculated using the linear equation: 30 log k = log k w -Sϕ (1)   where log k w is the retention coefficient for pure water, S is the slope of the regression line, ϕ is the concentration expressed as molar fraction of organic solvent and water.The R M parameter determined in TLC is an analogous with log k value and can be combined with modifier concentration: where the definition of R MW , S and ϕ are the same as in equation 1.
The aim of this work is the comparison the log P OW of thiosemicarbazide derivatives and their cyclic analogues 1,2,4-triazole-3-thiones, which can be used as potential drugs, determined by RP-HPLC and RP-TLC methods with the calibration curve technique with the log P values calculated using known algorithms (milogP, ALOGPs, AClogP, AlogP, MLOGP, KOWWIN, XLOGP2, XLOGP3). 24,25,31According to the Organisation for Economic Co-operation and Development (OECD) guidelines, in the chromatographic method selecting the appropriate reference compounds were required. 31The influence of the structure of analyzed derivatives on the retention is also discussed.The different chromatographic behavior of both groups of compounds (linear and cyclic derivatives) were compared.

Materials
Thiosemicarbazide derivatives and their cyclic analogues 1,2,4-triazole-3-thiones (Table 1) were synthesized in the laboratory at the Department of Organic Chemistry, Medical University of Lublin. 1 Methanol LiChrosolv (Merck, Darmstadt, Germany) for liquid chromatography grade and bidistilled water were used as mobile phase components.

High performance liquid chromatography
All HPLC experiments were performed using a chromatograph equipped with Elite LaChrom L-2130 gradient pump (Hitachi-Merck, Darmstadt, Germany), SPD-10AVP UV-VIS detector (Shimadzu, Kyoto, Japan) and Rheodyne 7725i valve with 20 µL loop.20 µL of each sample (0.1% solution) was applied into the chromatographic column (RP-18 Waters Symmetry, 15 cm length, 4.6 mm i.d., 5 µm particle size) using a Hamilton syringe (Hamilton, Bonaduz, Switzerland).Mobile phases were degassed by use of built-in membrane degasser.Chromatograms were developed at flow rate of 1.0 mL min −1 in isocratic mode using various concentrations of modifier in binary polar mobile phases: methanol ranges were 40-65% (v/v) changed by 5% per step (Table 2).
Chromatograms were detected at 254 nm.All experiments were repeated in triplicate and the final results were their arithmetic mean.Dead time was measured by use of uracil (Calbiochem-Merck, Darmstadt, Germany).All the experiments were performed at ambient temperature.

Thin layer chromatography
Thin layer chromatography was performed on 10 × 10 cm TLC plates coated with RP-18 254 using methanol-water mixtures as mobile phases (Table 2).0.1% of the methanolic solutions were applied on the plates and they were developed to a distance of 9 cm at room temperature in horizontal chambers (Chromdes, Lublin, Poland).The plates were not evaporated before the development.After drying in air, the chromatograms were visualized at a wavelength of 254 nm.Each experiment was performed three times.

Standard solutes
According to the OECD guideline, in order to correlate the measured capacity factor log k of a standard compound with its log P OW , a calibration graph using at least six points has to be established.It is preferable that the appropriate reference compounds should be structurally related to the test substances.Eight compounds were selected as standard solutes with optimal range of log P OW units (0.9 to 4.9).The following standard substances were selected (the log P OW values in brackets): aniline (0.9), 2-hydroxyquinoline (1.26), bromobenzene (3.0), naphthalene (3.6), propylbenzene (3.7), biphenyl (4.0), butylbenzene (4.6), pentylbenzene (4.9).
The standard compounds with known log P OW were analyzed under the same chromatographic conditions as the tested substances (RP-HPLC and RP-TLC) in order to determine the lipophilicity parameter (log P HPLC and log P TLC ).
All experiments were repeated in triplicate and the final results were their arithmetic mean.

Results and Discussion
The structures of analyzed compounds are presented in Table 1 and they were divided into two groups.The linear thiosemicarbazide derivatives (first nine compounds) and their cyclic analogues (remaining compounds) have the same substituents and they differ the lack of one molecule of water for cyclic analogues 1,2,4-triazole-3-thiones.The retention parameters were determined using the RP-HPLC and RP-TLC chromatographic systems.Mobile phases compositions for both chromatographic methods are presented in Table 2.Both the log k and the R M values decreased linearly with the increasing of methanol concentration in the mobile phase.The parameters of the linear equation for HPLC and TLC methods are presented in Table 3.The high correlation coefficients (r > 0.98) and small values of the standard errors of estimate (< 0.1) were indicated that all equations obtained were highly significant.The chromatographic lipophilicity parameters (log k W , R MW ) were obtained from equations 1 and 2 by extrapolation to pure water.In all cases, the value of log k W is always higher than R MW (Table 3).Probably, the differences in the log k W and R MW values are associated with "thin-layer effect" and the presence of apparent effluent front.
The correlation chart between the log k W and the R MW values was prepared and this relationship is described by the following equation: R MW = (1.1098± 0.064) log k w -(0.8757 ± 0.222) r 2 = 0.9488, n = 18, F = 296.7,s e = 0.158 High value of correlation coefficient confirms the similarity of both experimental methods (RP-HPLC and RP-TLC).
The differences between the chromatographic lipophilicity parameters, log k W obtained for second (cyclic analogues) and first (linear derivatives) groups (Dlog k WB-A ) were calculated (Table 4) and they are in the range from 0.021 to 0.508 (average value = 0.272).In the case of the TLC method, these differences are in the range from 0.045 to 0.597 (average value = 0.272).Negative values of were obtained for p-chlorophenyl and p-bromophenyl in HPLC and TLC methods.High value of chromatographic lipophilicity parameters log k W and R MW for compounds 3 and 12 (with cyclohexyl substituent) were observed (Table 3).Low differences between the values of R MW for 3 and 12 were noted (Table 4).
The structure of the first group is different from the second group of the lack of water molecule.The elimination of water from a molecule reduces its lipophilicity as well as absolute values of the specific hydrophobic surface, and the ratio of the intercept (log k W ) to the slope (-S) of the compound is constant in both groups. 35These results are in accordance with the fragmental method used for log P calculations. 36he lowest value of log k W and R MW was obtained for compounds 2, 8, 11 and 17.There are substances containing the ethyl group (2, 11) and but-1-ene group (8, 17) in their structure.Compounds 5 and 9 differ in chain length (one methylene group), similarly the substances 14 and 18.This small difference in structure slightly affects the value of the parameters of lipophilicity, which are higher for compounds with longer carbon chain (Table 3).
A significant influence of the structural differences was observed for substances 1, 10 and 3, 12, which contain the phenyl and the cyclohexyl groups, respectively.Higher value of lipophilicity parameters (Table 3) was obtained for compounds 3 and 12 (the average of differ for log k W is 0.8861 and for R MW is 1.2462).Moreover, the change of halogen group for compounds 6, 7, 15 and 16 did not significantly affect the change in value of lipophilicity.
Comparing the log k W and R MW values of two groups of analyzed compounds (the linear thiosemicarbazide derivatives and their cyclic analogues), some differences have been observed.Generally, the slightly higher values of these factors were obtained for second group of compounds (10-18) in most of the cases (Table 3).The increase of the lipophilicity is probably due to the presence of the additional triazole ring, which changes the position of the whole molecule in space.The exceptions are the substances 6 and 7, for which the log k W and R MW values are slightly higher than for compounds 15 and 16.The p-chlorophenyl substituent is presented in the structure of compounds 6 and 15 and the p-bromophenyl substituent is presented in the structure of compounds 7 and 16.The presence of halogen substituents affects the chromatographic behavior of the whole molecule.The proximity of chlorine and bromine (free electron pairs) can cause changes in stereochemistry and different interactions of the molecule with the stationary and mobile phases.A linear relationship between the intercept and slope from equations 1 and 2 for the used mobile phase, is one of the basis features of chromatographic determination of the lipophilicity of closely related compounds. 37In this study, the good correlation obtained between the intercept (log k W , R MW ) and slope (S) confirms the suitability of these systems for estimation of the lipophilicity of thiosemicarbazide derivatives and their cyclic analogues.The linear correlation is described by the following equations: The rule of 5" developed by Lipiński et al. 38 predicts that poor absorption or permeation is more likely when there are molecules (drug-like) that have more than 5 H-bond donors, 10 H-bond acceptors in their structure, the molecular weight (MWT) is greater than 500 and the calculated log P (C log P) is greater than 5 (or M log P > 4.15). 38In our work, newly synthesized thiosemicarbazides and their cyclic analogues 1,2,4-triazol-3-thiones were in agreement with the rule of five claimed by Lipiński et al. 38 (Table 5).
Reversed phase high performance liquid chromatography and reversed phase thin layer chromatography were also used in order to determine experimentally octanol-water partition coefficients (log P HPLC and log P TLC parameters).The measurements were conducted according with the OECD guidelines. 315][26][27] In this study, eight compounds have been selected (see Experimental section) as reference compounds from the Recommended Reference Compounds list published by the OECD. 31The determination of log P OW by HPLC and TLC methods is based on the linear relationship between the chromatographic retention parameters (log k and R M ) and the octanol-water partition coefficient determined by shake-flask method for selected standard solutes.
In the case of RP-HPLC method, the best selectivity was obtained with methanol-water (60:40, v/v) and this mobile phase was chosen for the determination of log P OW .Linear calibration equation (Figure 1) between log k values and their literature log P OW for standard compounds looks as follows: y = (2.3137± 0.07)x + (2.2013 ± 0.09) r = 0.9946 ± 0.04, n = 8, F = 552.5,s e = 0.16 (6)   Next, the log k values of the studied substances were substituted into equation 6 to calculate the partition coefficient obtained for HPLC method (log P HPLC ).Similar procedures were carried out for RP-TLC method.In this case, the mobile phase containing 70% methanol in water (%, v/v) was proved to be the best selectivity system.Linear relationships between the R M and the log P OW was obtained (Figure 2): y = (2.8458± 0.12)x + (1.6296 ± 0.09) r = 0.9950 ± 0.04, n = 8, F = 599.9,s e = 0.16 (7)   The determination of linear relationships between experimental lipophilicity parameters (log P HPLC and log P TLC ) and calculated log P values is a necessary step for QSAR analysis. 25In our work, these correlations were performed separately for two groups of tested compounds and the extrapolated log k W and R MW values and experimentally established log P HPLC and log P TLC values were compared with calculated log P (log P calc ).Generally, in the case of the RP-HPLC method, high values of correlation coefficient were obtained for first group of analyzed compounds (thiosemicarbazide derivatives) (0.8244 < r < 0.9808) in comparison with their cyclic analogues (0.6615 < r < 0.9456).The weaker correlations were obtained for RP-TLC method, where for the first group of compounds, the partition coefficient was in the range: 0.7770 < r < 0.9795 and for second group: 0.6183 < r < 0.9030.The best results for correlations between the experimental and calculated partition coefficients were obtained for relationships between extrapolated (log k W and R MW ), experimental (log P HPLC and log P TLC ) parameters and calculated partition coefficient.
In order to better illustrate these correlations, the unscaled principal component analysis (PCA) with loadings interpretation was used.The experimental data (log k W , R MW , log P HPLC and log P TLC ) from Tables 3 and 5 and the calculated log P parameters (from Table 6) were grouped as data matrix and they were analyzed using PCA, based on covariance matrix (unscaled PCA) using the Statistica 8 (StatSoft Inc. 2007) and results are presented in Figure 3.The experimental data was used as supplementary data  (in Figure 3 they are marked with squares).The strongest correlations between the experimental log P factors and parameters were confirmed.
Obtained results confirm that the chromatographic methods used to measure the lipophilicity of the thiosemicarbazides and their cyclic analogues 1,2,4-triazol-3-thiones are valid and suitable.

Conclusions
Values of the relative lipophilicity parameters log k W and R MW were converted into log P HPLC and log P TLC values by use of the a calibration graph obtained by use of ten standard solutes.This study shows that aniline, 2-hydroxyquinoline, bromobenzene, naphthalene, propylbenzene, biphenyl, butylbenzene and pentylbenzene are useful as reference substances for the determination of partition coefficient octanol-water using HPLC and TLC methods.
The influence of the structure of the thiosemicarbazides and their cyclic analogues 1,2,4-triazol-3-thiones on the value of lipophilicity was observed.
Moreover, the good correlation between the intercept (log k W , R MW ) and slope (S) confirms the suitability of these systems for estimation of the lipophilicity of thiosemicarbazide derivatives and their cyclic analogues.
The best correlations between the experimental (or extrapolated) partition coefficients and milogP parameters were obtained.Generally, higher values of partition coefficient for these relationships were obtained for RP-HPLC method.

Figure 1 .
Figure 1.Calibration graph for standard solutes for RP-HPLC method.

Figure 2 .
Figure 2. Calibration graph for standard solutes for RP-TLC method.

Table 1 .
List of compounds investigated

Table 2 .
The concentrations of used eluents, n-number of points

Table 3 .
Parameters of the equations 1 and 2 for methanol-water system w : retention coefficient for pure water; S: slope of the regression line; r: correlation coefficient; n: number of points; s e : standard error of estimation; F: statistica F.

Table 5 .
Data applied for the rule of 5" for tested compounds

Table 6 .
Values of log P calc parameter calculated by computer programs.The values of log P HPLC and log P TLC are presented in Table5