| dc.description.abstract | The DNA binding behaviour of netropsin (Nt) and distamycin-A (Dst-A) has been the subject of intense investigation in recent years. Nt and Dst-A belong to a class of oligopeptide antibiotics possessing two or three N-methylpyrrole units linked by peptide bonds and bounded by a terminal cationic propionamidine side chain. They interact strongly with DNA and are specific for (1) B-DNA, (2) the minor groove, and (3) AT base pairs, thus showing a unique triple specificity. Unlike the more common intercalators, Nt and Dst-A do not insert themselves between the DNA base pairs; instead, they are found to be deeply buried in the minor groove in a nonintercalative mode of binding.
Over the last few years there have been numerous studies attempting to elucidate the mechanisms of nonintercalator binding and to understand the nature of the forces which give rise to the exquisite specificity of their interaction. Our laboratory was also involved in a number of such studies. The results obtained from some of those studies which were carried out by the author are presented in this thesis.
Chapter I reviews the existing literature on DNA binding ligands. Both the intercalators and the nonintercalators have been described in some detail and the essential differences between their modes of interaction have been highlighted. The existing status of our knowledge regarding the origin of the interaction specificity in nonintercalators has been discussed. The different theories which seek to explain the origin of this specificity have been examined and their merits and demerits analysed.
It was earlier reported from this laboratory that the intrinsic curvature of the ligand backbone appears to play an important role in its DNA binding behaviour. Thus a compound known as the MPD derivative (bis-1,3-(4-(1-ethyl nicotinamido)benzamido)benzene 4-toluenesulphonate), synthesised in this laboratory, which has a curvature intermediate between Dst-A and the synthetic nonintercalator NSC 101327, is also intermediate in specificity. While Dst-A is specific both for AT base pairs and the B conformation, NSC 101327 is specific for neither. The MPD derivative, on the other hand, is specific for the B conformation but binds to any base sequence with similar affinity (Rao et al., 1988). Another Dst-A analogue prepared in this laboratory, in which the N-methylpyrrole groups of Dst-A had been substituted with 1,3-disubstituted benzenes and consequently having a different backbone curvature from that of Dst-A, showed drastically reduced AT specificity (Dasgupta et al., 1986). It was thought, therefore, that a systematic search for ligands with varying backbone curvatures should be made so as to accurately determine the curvature-induced effects in ligand-DNA interaction.
Chapter II describes the strategy used to search for potential DNA binding compounds with different curvatures. The method used is originally due to Goodsell and Dickerson (1986), and consists of the generation of infinite helices made up of the backbone conformation of a particular chemical type of ligand as the repeating unit. The helical parameters ‘n’ (the number of repeating units per turn of the helix) and ‘h’ (the axial rise in going from one repeating unit to the next) are determined. A conformation which gives ‘n’ and ‘h’ values close to that of the DNA duplex is judged to have suitable ligand curvature and is termed an ‘isohelical conformation’. The conformation of the repeat unit was built up from the structures of several disubstituted benzene rings linked with peptide bonds.
It was found that a monomer repeating unit consisting of a single 1,3- or 1,4-disubstituted benzene cannot take up any isohelical conformation in accordance with the results obtained by Goodsell and Dickerson (1986). The results are essentially the same even if one uses a dimer repeat made up of two 1,3- or 1,4-disubstituted benzenes. In this case, if the repeat unit consists of two 1,3-disubstituted benzene rings, then a very small number of isohelical conformations are encountered. No such conformation is found if the repeat unit consists of two 1,4-disubstituted benzenes. However, a dimer repeating unit consisting of an alternating system of 1,3- and 1,4-disubstituted benzene rings linked with peptide bonds could take up many isohelical conformations.
Thus it appeared that a compound having an alternating system of 1,3- and 1,4-disubstituted benzenes would have a suitable ligand curvature for DNA binding, whereas a compound having solely 1,3- or 1,4-disubstituted benzenes would not. Accordingly, four compounds were designed. Each had two benzene rings linked with peptide bonds. Two of them were made up of alternating 1,3- and 1,4-disubstituted benzenes and two were made up solely of either a pair of 1,3- or 1,4-disubstituted benzene rings. The former group of compounds were predicted to have a suitable ligand curvature for DNA binding, while the latter two would not have the same. All the four compounds were terminally bounded by a nitro group at one end and a propionamidine group on the other.
The nitro group, which is not normally found in DNA binding nonintercalators, was chosen because it was felt that the highly electronegative nature of the group would be able to modify the electrostatic interactions between the ligand and the DNA and thus would affect the binding specificity. As a control, a fifth compound was designed where the nitro group was replaced by an additional benzamide.
The syntheses of all the five compounds were carried out by conventional solution-phase techniques. The synthetic routes were optimised to reduce the number of intermediate steps so as to increase the overall yield of the compounds. All the intermediate compounds were completely characterised by IR and NMR spectroscopy as well as elemental analysis. The structures of the final compounds were determined by complete spectroscopic and chemical characterisation, and their purity was checked by HPLC.
Chapter III describes the physicochemical characterisation of the DNA binding activity of the compounds. The methodology used consisted mainly of CD spectroscopy in conjunction with UV spectroscopy and thermal melting studies.
It is observed that the compounds which do not have an alternating system of 1,3- and 1,4-disubstituted benzenes have little or no DNA binding activity. Those which have the alternating 1,3- and 1,4-disubstituted benzene ring system, however, show reasonably strong DNA binding activity. This implies that optimum ligand curvature is an essential prerequisite for nonintercalator-DNA interaction.
The compounds which possess just two benzene rings with a nitro group at one end are strongly AT-specific. They do not bind at all to poly d(GC).poly d(GC) and also to the mixed A/T, G/C containing polymer poly d(AC).poly d(GT). Calf thymus DNA, which has about 58% AT base pair content, does not elicit any binding effect. Among the 100% AT-containing polymers, poly d(AT).poly d(AT) and the homopolymer poly d(A).poly d(T), the binding activity is much stronger for the latter polymer.
Comparatively weak but nevertheless easily detectable binding is found for the guanine-lacking polymers poly d(IC).poly d(IC) as well as poly d(AU).poly d(AU). Replacement of the terminal nitro group with a benzamide causes a slight reduction in the AT specificity. Thus the compound now binds to both poly d(AC).poly d(GT) as well as calf thymus DNA, but not to poly d(GC).poly d(GC).
All the compounds show a strong preference for the minor groove as judged by comparative binding titrations with Clostridium perfringens DNA and T4 phage DNA. Both the DNA polymers have similar AT base pair contents (>85%), but the major groove of the latter polymer is blocked by glycosylation. Essentially the same results are also obtained from competitive titrations with the well-established minor groove binding agents ethidium bromide and distamycin. None of the compounds show any binding activity for dsRNA and ssDNA as shown by lack of any binding effect with either poly(A), poly(U) or with poly d(AT) at high temperatures.
Chapter IV describes the crystal structure analysis of three compounds, viz., methyl 4-(3-nitrobenzamido) benzoate, methyl 3-(4-nitrobenzamido) benzoate and methyl 4-(benzamido) benzoate. These compounds are all parts of the aromatic portion of the ligands whose synthesis and DNA binding activity are described in Chapters II and III. The crystal structure analysis was carried out with a view to performing further molecular modelling studies on the ligand-DNA complexes.
It was of interest to know the orientation of the two benzene rings with respect to each other, as recent studies seem to indicate that this orientation plays a significant role in the DNA binding behaviour. It was observed that even though the compounds are expected to have a delocalised ?-electron system and therefore should tend towards planarity, there is large variation in the angle between the two rings, which shows that the rings can deviate significantly from a common plane. The exact orientation of the two rings in each of the three compounds seems to depend upon intermolecular interactions in the crystal.
Chapter V describes the results obtained from detailed molecular mechanics calculations using one of the three compounds which was found to bind with DNA (described in Chapter III) as a representative example. Molecular mechanics calculations using the monotonic sequence decamers (A)10.(T)10, (AT)5 and (GC)5 clearly reproduced the AT specificity and minor groove binding preference which was obtained experimentally.
Detailed analysis of the DNA portion of the complexes showed that the base pair orientations, and not the backbone torsion angles, are most susceptible to changes consequent to ligand binding. It was also found that the AT base pair as well as the minor groove preference arose primarily due to better electrostatic interactions followed by better van der Waals contacts, and not because of hydrogen bonding and salt bridge formation with the anionic phosphates.
The calculations were then extended to the mixed sequence dodecamer d(CGCGAATTCGCG)2. It was found that the structural peculiarities of the dodecamer sequence as determined from crystal structure analysis (Dickerson and Drew, 1981) had a significant effect on ligand binding, making the dodecamer predisposed towards ligand binding in the AATT region.
The results clearly indicated that while electrostatic interaction is primarily responsible for the AT base pair preference, it can be modulated strongly by the precise sequence-dependent geometry of the DNA molecule.
In summary, it can be stated that in order to bind with DNA, a nonintercalating ligand should have an optimum curvature which mimics the curved DNA backbone. The base sequence preference of the ligand seems to be primarily controlled by better electrostatic interactions, but the latter in turn depends on the precise sequence-dependent geometry of the binding site. | |
