QUANTITATIVE STRUCTURE ACTIVITY RELATIONSHIPS FOR GUANIDINOTHIAZOLECARBOXAMIDES USING THEORETICALLY CALCULATED MOLECULAR DESCRIPTORS

: Guanidinothiazolecarboxamides (GTCs) are a class of anti-tumor agents supposed to exert their anti-tumor activity through intercalation. The intercalation of GTCs with DNA duplex tetramer d(GCTA) was investigated using molecular mechanics methods. The results showed that GTCs intercalate favourably at the CGlTA step. Two quantitative structure activity relationships (QSAR) for 5 and 6 substituted GTCs were obtained using theoretically ialculated molecular descriptors. results showed that the GTCs intercalate favourably at the CGmAstep of the duplex DNA tetramer. The significant distortions in the double strand structure of the DNA were observed upon intercalation of GTCs. The results also revealed that the planarity of the GTC was also distorted upon the intercalation. We have demonstrated that the molecular descriptors derived from the quantum chemical and molecular mechanics methods can be used to develop quantitative structure-activity relationships for anti-tumor active GTCs. Two independent QSARs were obtained for 5 and 6 substituted GTC analogues. However, further investigations are required to improve theoretical QSARs obtained in this study and to develop theoretical QSAR for substituted GTC analogues regardless of their substitution position.


INTRODUCTION
Cancer is one of the most formidable diseases in the world. Lung cancer, in particular, has increased rapidly in incidence, due in part to the long-term effects of air pollution, especially that from tobacco. All structure-activity relationships made so far on anti-tumor drugs have primarily established that the main cellular target of these drugs is DNA. The anti-tumor drugs that directly bind with DNA and stop its proliferation mainly involve three types of binding1: 1) covalent bond formation 2) intercalation and 3) non-intercalation groove binding.
Guanidinothiazolecarboxamides (GTCs), (Figure I), are a novel class of anti-tumor agents found to be systemically active against experimental pulmonary metastases 3LL Lewis Lung Car~inoma.~.~ The GTCs are supposed to exert their anti-tumor effects through interca1ation.l It has been generally observed that fused ring aromatic compounds bind most strongly to DNA as an intercalation complex . 4,5 The widely studied intercalators are the derivatives of 9-anilinoacridine, 9aminoacridine and the analogues of anth~acycline.~ There are no detailed experimental or theoretical results available in literature on the favourable binding sites of GTCs to DNA. The major objectives of this research project are 1) to find favourable binding sites of GTC to DNA and 2) to explore whether there exists a correlation bet wee^ biologisal activity of GTCs and their calculated electronic and structural properties. The concept of quantitative structure activity relationship (QSAR) is to transform searches for compounds with desired properties using chemical intuition and experience into a mathematically quantified and ' Corresponding author computerised form. The importance of this later objective is that once a correlation between structure and activity is found, any number of compounds, including those not yet synthesized, can be readily screened on the computer in order to select . structures with the desired properties. It is then possible to select the most promising , compounds to synthesize and test in the laboratory.
To obtain a significant correlation, it is crucial that appropriate descriptors be employed, whether they are theoretical, empirical or derived from readily available experimental characteristics of the structure. Many descriptors reflect simple molecular properties and can thus provide insight into the physicochemical nature of the activity under consideration. Recent progress in computer hardware and the development of efficient algorithms have assisted the routine development of molecular quantum mechanical calculations. New semi-empirical methods supply realistic quantum chemical molecular properties in a relatively short computational time frame using even a personal computer. Quantum chemical calculations are thus an attractive source of new molecular descriptors, which can, in principle, express all of the electronic and geometric properties of molecules and their interactions. The quantum chemical descriptors have been coGelated with biological activities such as enzyme inhibition activity, hallucinogenic activity, et~.~-'O In using theoretically based descriptors with a series of related compounds, the computational error is considered to be approximately constant throughout the series. A basic weakness of theoretical descriptors is the failure to directly address the bulk effects." In this study, semi-empirical quantum chemical and molecular mechanics methods were used to obtain molecular descriptors of free drug molecules. The favourable intercalating sites for drug molecules were investigated by using the molecular mechanics methods.

Calculation of Malecular Descriptors:
Initial atomic coordinates of all GTC molecules investigated in this study were obtained from the X-ray crystal structure of 5F-GTC.3 All structures were first energy minimised in the gas phase within molecular mechanics (MM) using the MMX force field" and the resulting structures were further optimised by the semi-empirical quantum chemical method PM3. The calculations were performed using PCMODEL12 and MOPAC version 513 software packages. Finally, the gas phase optimised structures were re-optimised in the aqueous phase with the PM3 method using the AMSOL" package. LogP (partition coefficient between octanol and water) was calculated by the Crippen's fragrnentation16 using the CHEMDRAWlB program.

Intercalation of GTC with DNA:
In order to determine the favourable intercalating site for GTC, three DNA base steps CGITA, GCICG and TAIAT'were considered. GTC was intercalated using the PCMODEL software into DNA tetramer d(GCTA), having the above base steps. The DNA tetramer was constructed using the X-ray crystal structure of d(CGCTAGCG),. 15 The intercalated GTC-DNAcomplexes, DNA tetramer and GTCs were energy minimised using the MMX force field. The binding energy (BE) of the GTC-DNA complex was calculated using the following equation.
where EDNA is the energy of the DNA fragment, E, is the energy of GTC and E,,,.DNA is the energy of the GTC-DNA complex.

Multiple linear regression analysis
A QSAR generally takes the form of a linear equation Biological Activity = C,P, + C,P, Where the parameters P, through Pn are computed for each molecule in the series and coefficients C, through Cn are calculated by fitting variations in the parameters and the biological activity using multiple linear regressions.

RESULTS
The wireframe model of the energy minimised structure of 5F-GTC in the gas phase using the PM3 method is shown in Figure 1. The calculated gas phase and aqueous phase molecular properties of 5 and 6 substituted GTCs using the PM3 and MM methods are given in Tables 1 and 2 respectively. Energies of the HOMO (highest occupied molecular orbital) and LUMO (lowest unoccupied molecular orbital) are very popular quantum chemical descriptors. In Tables1 and 2, the ionisation energy (IE) is calculated as the negative value of the energy of the HOMO. The polarity of a molecule is well known to be important for various physicochemical properties. The most obvious and most often used quantity to describe the polarity is the dipole moment of the molecule. The total dipole moment given in Tables 1 and 2, however, reflects only the global polarity of a molecule. Formation of a n intra-molecular 1 hydrogen bond between N19 and H29 atoms was observed in the PM3 optimised structures. The intra-molecular hydrogen bond distance, N19-H29 for 5 and 6 substituted GTCs is given in Table 1. Since the intra-molecular hydrogen bond stabilises the molecule, we have considered this hydrogen bond distance as a descriptor in QSAR analysis.   a heat of formation in kcallmol; ionisation energy in e~; energy of lowest unoccupied molecular orbital in eV; dipole moment in deby; ' polarisation free energy of sol*ation in kcallmol; cavity-dispersionsolvent structure free energy in kcaumol; 8 total free energy of solvation in kcallmol The binding energies for some 5 and 6 substituted GTCs intercalated into DNA tetramer d(GCTA)2 are tabulated in Table 3. Among the three base pair steps (GCICG, CGEA and TAIAT) investigated in this study, the results showed that GTC intercalates more favourably at the CGEA step. The energy minimised structures of the DNA tetramer d(GCTA), and the 6F-GTC-d(GCTA), complex at the CGPl"A step using the force field MMX are shown in Figure 2A and 2B respectively. As suggested by the hydrogen bond distances reported in Table 4, the Watson-Crick base pairing was maintained upon intercalation. However, because base pairs must separate vertically to allow intercalation, the sugar-phosphate is destroyed. As seen in Figure 2B, the average separation in the CGPTA base step was increased from 3.5-3.8 Hi to 6-8 A. However, the separation of the other two base steps remained unchanged. The present results showed significant structural changes in the DNA helical structure upon intercalation of GTCs. The results given in Tables 5 and 6 showed that the torsional angles of the phosphate backbone (a,P,S,&, < and X) and sugar (v,,v2,v3,v, and v,) change significantly. The definition of these torsion angles is shown in Figure 3. In addition to the observed structural changes in DNA upon intercalation, a small change in the planarity of the GTC molecule has also been observed. For example, the dihedral angles C12CllNlOC2, NlOC2SlC5 and C2SlC5C6 have been changed from 180" to 175", 172" and 175" respectively.        a The sugar torsion angles are defined belaw using the atom labels and numbering scheme shown in  Initially, numerous sets of regression analyses were performed on the singly substituted GTC analogues using theoretically calculated parameters. In all regression analyses, the probit transformati~n~~"~ of biological response (ILS) defined as log [100/(100-ILS)] has been used as the dependent variable. For example, an analysis using thirteen 6-substituted GTC analogues, see Table 7, resulted in equation (1) where 57% of the variance (R2=0.57) in the survival data was accounted for by an equation incorporating the energy of the lowest unoccupied molecular orbital E, , , , and the partition coefficient between octanol and water Log P. Figure 4 shows the correlation graphically of the observed versus calculated log [100/(100 -ILS)I19 values for 6-substituted analogues.
A regression analysis on the set of 5-substituted GTC analogues able 8) incorporating UEHoMo and Log P led to equation 2 that accounted for less of the variance (i.e., R2=0.52) in the observed efficacy data. Figure 5 shows the correlation between the observed and calculated log [100/(100-ILS)119 values.

DISCUSSION
In the present study, the interaction of GTCs with the double strand DNA sequence d(GCTAY has been investigated by using the molecular mechanics method. The results showed that the GTCs intercalate favourably at the CGmAstep of the duplex DNA tetramer. The significant distortions in the double strand structure of the DNA were observed upon intercalation of GTCs. The results also revealed that the planarity of the GTC was also distorted upon the intercalation.
We have demonstrated that the molecular descriptors derived from the quantum chemical and molecular mechanics methods can be used to develop quantitative structure-activity relationships for anti-tumor active GTCs. Two independent QSARs were obtained for 5 and 6 substituted GTC analogues. However, further investigations are required to improve theoretical QSARs obtained in this study and to develop theoretical QSAR for substituted GTC analogues regardless of their substitution position.