suggests that electrostatic effects are not a major contributor to the driving force of interaction, since ligand–DNA interactions, which have large contributions from hydrophobic and electro- static forces, are largely driven by entropy due to the release of water and counter ions from the polyanion DNA duplex upon ligand binding. 30,37 The published literature to date has shown that minorgroove recognition by small molecules can be enthalpically or entropically driven or both and that the ther- modynamic signature of MGBs is highly dependent on ligand structure and the sequence of the binding site, which all of our studies with different sequences and related ligands confirm. Further detailed studies along similar lines to those reported here are required on related systems in order to widen our under- standing of the role played by thermodynamics for alternative types of DNAminorgroove binding ligands, such as those represented by 1.
By understanding the rules that govern the tight, side-by-side binding of ligands with the DNAminorgroove it has become possible to develop tailored approaches to drug design. These developments have precipitated the design of sequence reading molecules with DNA specificities and affinities comparable to DNA-binding proteins and have allowed convincing demonstrations of an ability to interrupt gene expression in vitro. 1-2 The development of minorgroove binders (MGBs) proceeded from the observation that netropsin and distamycin bound selectively in the DNAminorgroove by a combination of hydrogen bonding with the bases on the groove floor allowed by their natural curvature. 3-6 A significant breakthrough in the field came with the observation that a number of MGBs could bind in the minorgroove as a side-by-side 2:1 complex. 7 Whilst hydrogen bonding to the groove floor endowed specificity for particular sequences, lipophilic interactions with the groove walls were also highly relevant. 8-9 The balance between enthalpic and entropic contributions to MGB binding is the subject of extensive research, and appears to vary with both MGB structure and the binding sequence of the DNA. 10
Molecular recognition between the DNAminorgroove and classical small molecule DNA binders including Hoechst 33258 [1-234567] berenil [8,9] netropsin  distamycin A [11,12] and their derivatives ( Fig. 1) typically occurs as either 1:1 or 2:1 ligand:DNA complexes [13-,14,15]. For aromatic peptide minorgroove binders (MGBs) based on heterocyclic monomers such as N- methylpyrrole (Py) and N-methylimidazole (Im), binding in either 1:1 or 2:1 modes is governed by the DNAminorgroove width, which is DNA sequence dependent [16,17]. Non-equivalent monomers are capable of binding to non-complementary DNA recognition sites when arranged anti-parallel to one another in a face-to-face manner [16,18-1920] and when joined either mid-molecule via polycarbon linker to create a sandwich dimer  or through intervening -aminobutyric acid residues to create hairpin [22,23] or cyclic [24-25,26] polyamide units. Hairpin and cyclic polyamides in particular display DNA binding association constants suitable for a programmed development of gene-targeting drugs [27,28], have cell penetration characteristics and promising biological potential [29,30]. Notwithstanding these advantages, the cyclic and hairpin MGBs are relatively large molecules supporting a case for the continued development of unlinked MGBs as smaller, simpler synthetic molecules. With scope arguably for more accessible commercial scale production together with other advantages associated with the smaller size of unlinked MGBs, such molecules remain attractive alternatives as potential therapeutic molecules, for example to treat infectious disease, or biotechnological tools.
UV-Vis absorption spectral was carried out to investigating the binding mode of the compound 8 with DNA. The CT-DNA solution (10 mM) was titrated against compound 8 in 0.1 M Tris-HCl buffer pH 7.4 (Figure2a). The absorbance of CT-DNA at 258 nm progressively increased when the concentration of compound 8solution was increased from 0 to 20 mM. There was a distinct blue shift of DNA-compound 8 complex in the 258 nm region. Hoechst 33258 has the similar increase in absorbance of CT-DNA at 254 nm associated with blue shift.So, compound 8 may have the same binding ability with DNAminorgroove.
Based on these structural studies, molecular modelling has suggested to us that compound 5 could effectively bind into the DNAminorgroove (Figure 1). Then, a series of derivatives with a head-to-head bis-benzothiazole structure were synthesized on the basis of modeling results. The interaction of compound 5 with CT (calf thymus)-DNA has been investigated using absorption spectroscopy, ﬂuorescence spectroscopy. All of these compounds were screened for anti-tumor activity in vitro and showed significant antitumor activity.
In the past, protein expression has been analyzed through mRNA studies. However, it was later shown that mRNA content does not correlate with protein con- tent as mRNA is not always translated into protein [16-19]. On the other hand, proteomics is a systematic analysis that measures protein expression directly and not via gene expression, yet serving as a complementary approach to genomics. 2D gel electrophoresis is still the most useful way to separate proteins in complex samples in proteomics profiling and allows simultaneous analyses of vast amount of protein data, making it suitable for comparative analysis of a reference cell protein profile with a profile after drug treatment in the search of new drug or drug target. At present there are minimal re- searches on the consequences of DNAminorgroove binding agents on the proteomic profile in cells.
The formation of ICLs in vivo was determined using the single cell gel electrophoresis (comet) assay at doses of SG2000 shown in the above efficacy experiments to give dose-dependent tumour growth inhibition. LMeC tu- mours were taken after 2 h and 24 h following treatment with 0.15 or 0.3 mg/kg SG2000 and the level of cross- linking determined as the % decrease in tail moment (% DTM, Fig. 6a). A high level of DNA cross-linking was observed in tumour cells at 2 h, consistent with the rapid formation of SG2000-induced cross-links observed in the LMeC cell line in vitro . The level of cross-linking in tumour was significantly greater at the higher dose of 0.3 mg/kg, consistent with the increased antitumour effect observed at this dose (Fig. 5). At 24 h cross-links were still evident in tumours at both dose levels (Fig. 6a), consistent with the persistence of the cross-linking observed in vitro . A similar experiment was performed in mice bearing CMeC-1 tumours (Fig. 6b), however, additionally in these experiments lymphocytes were also examined for DNA ICL at the same time points as the tumour samples. Dose
contains the recognition sequence for the restriction enzyme B a n II. Interestingly, results obtained with this system did not produce significant cleavage inhibition at low doses, but instead caused a marked stimulation of DNA cleavage over a wide range of ligand concentrations. Although most of the studies involving minorgroove binders and DNA processing enzymes demonstrate inhibition of the proteins, there are some reports of an increase in enzyme activity in the presence of minorgroove binding ligands. These include a stimulation of isolated topoisomerase I activity by low doses of distamycin (McHugh et. a/., 1989), increased topoisom erase II catalysed DNA decatenation by low doses of netropsin (Beerman et. al., 1991), and distamycin (W oynarowski et. a!., 1989), and stimulation of DNA polymerase I and bacteriophage 14 DNA polymerase by distamycin (Levy et. a/., 1989). However, it should be noted that all of the proteins given in these examples are proteins that interact with the substrate DNA in a sequence non-specific manner. Conversely, the DNA cleavage reaction of Ban II is highly DNA sequence selective. There is evidence that a minorgroove-binding peptide and a major groove-binding protein (the DNA binding domain of the yeast GGN4 transactivating factor) can simultaneously occupy a common site on a DNA molecule, with no cooperativity, either negative or positive, between the two molecules (Oakley et. a!., 1992). Thus, it is not difficult to envisage a situation in which the binding of a ligand in the minor-groove of DNA alters the conformation of the adjacent major-groove to make it more accessible for binding by a major-groove recognising protein. Alternatively, the DNA conformational change induced by the ligands may somehow aid the DNA cleaving reaction rather than the DNA binding component of the reaction. For example, the binding of the compounds may facilitate more efficient association of enzymes with their cofactors.
Deoxyribonucleic acid (DNA) is the fundamental repository of genetic information used in the majority of organisms in nature. 1 A structural hallmark of DNA is the anti-parallel helix where the major and minor grooves provide sites for sequence-selective binding of proteins and small molecules. Molecular recognition of double-stranded DNA (dsDNA) sequences by transcription factors for example is essential for the initiation of transcription. Furthermore, DNA-binding small molecules such as minorgroove binders (MGBs) can perturb various processes associated with gene expression, 2 making them excellent candidates for the design of sequence-selective probes of DNA function in cells and potentially novel therapeutics. 3-5 A comprehensive set of guiding principles for the rational design of MGBs to target DNA sequences in a highly sequence-selective manner has yet to emerge however, as a consequence of the complex combination of competing enthalpic and entropic contributions from H- bonding, van der Waals forces and changes in hydration of the DNA and ligand.
Table 3.1A Observed equilibrium association constants for 1-6……………...……63 Table 3.1B Normalized equilibrium association constants for 1-6…………………63 Table 3.2 Normalized fluorescent intensities for 1-6 with DNA duplexes……….66 Table 3.3 Physical properties of different classes of fluorophore-
The DNAminorgroove is one of the most successful and attractive targets for drug design and development (1-5). Aromatic cationic diamidines binding in the minorgroove of DNA present successful and effective classes of anti-prozotoal agents (2, 3, 5). Pentamidine is the first diamidine compound that has been used in the clinical treatment for human African trypanosomiasis (HAT). But pentamidine has its limitations due to side-effects and lack of oral bioavailability (5). Furamidine, DB75, is a new exciting diamidine that exhibits excellent activities against a broad spectrum of protozoa diseases (5-8). Its orally active prodrug DB289, the diamidoxime derivative of DB75, is currently in Phase III clinical trials against human African trypanomiasis (HAT) and Phase II clinical trial against malaria (9). All these biologically active diamidines have shown strong binding preferences and high binding affinities to AT-rich sequences in the DNAminorgroove (2, 10, 11). The anti-protozoal activities of these cationic diamidines have been hypothesized to be related to inhibition of the function of DNA-targeted
In order to investigate whether these small differences in inter-atomic distance can explain the differences in frequencies of one-atom motifs from Table 4, energies of the averaged motifs have been compared with energies of a hydrogen bonding system in the optimal geometry. Both dinucleotide and mononucleotide motifs were modeled as one hydrogen donor, ligand D and two acceptors, DNAminorgroove atoms A1 and A2 in an Acceptor . . . Donor . . . Acceptor (A1 . . . D . . . A2) system. An example of such a system is shown in Figure 3A. The influence of deformation in the A1 . . . D . . . A2 system was estimated by assuming that one hydrogen bond, A1 . . . D, was optimal. Deformation energies of the other bond, D . . . A2, were calculated around four starting geometries corresponding to the respective geometries of ligand binding to the base i –O4 ′ i , base i –O4′ i + 1 , base i –base i + 1 and O4′ i –O4′ i + 1 motifs. The position of ligand D relative to the DNA acceptor A2 was estimated by the binding site position in each motif (Table 3) and the system D . . . A2 was modeled as a water dimer (Fig. 3B).
Small molecule recognition of nucleic acids has been the subject of extensive research since the elucidation of the DNA double-helix structure in 1953 by Watson and Crick. It is currently recognized as a key mechanism responsible for medico-biological properties of some drugs, in particular, those exerting antitumor properties [1,2]. One of the main difficulties in drug-based chemotherapy is the toxicity associated with nucleic acid targeted drug intervention. These toxic side effects associated with DNA binding drugs, at least in part, are directly linked to their relatively weak DNA sequence targeting specificity. The essential strategy in DNA-targeted drug design has been to find or create drugs with high specificity and cooperativity of binding to particular DNA sequences [2,3]. One of the rare examples when this strategy has been known to succeed was in the discovery of the lexitropsin class of DNAminorgroove binders (to be referred to here as MGB ligands). These exert a characteristic homo-dimeric type of complexation by which two drug molecules simultaneously occupy a particular binding site [4,5]. The discovery of this DNA recognition mode opened up an important page in the history of DNA-targeted drug design and to date remains an active field of research (e.g. see  for review).
The most conspicuous means by which selectivity might be obtained is to target a specific base sequence. Such a sequence might belong to a specific gene or its promoter or enhancer regions, with the subsequent binding resulting in some degree of modulation of the gene itself. Selectively binding a unique DNA sequence in the human genome would require an incredibly complex drug able to recognise a sequence of some 15-16 bases † . 407-409 Molecules possessed of such a binding footprint present a significant synthetic challenge and potential candidates have, to date, have had limited success. 407, 409 Even so, a smaller target sequence of 6-10 bases (for example) would still occur relatively infrequently within the genome and therefore still represent a viable objective. The lower end of this range corresponds to the footprint size of some of the more specific nucleic acid binders in use today. These ligands are typically polyamide minorgroove binders such as netropsin and derivatives thereof. Selectivity is achieved via specific hydrogen-bonding patterns between amide functionalities on the drug molecules and polar minorgroove atoms on the target bases. Species featuring pyrrole subunits favourably bind A•T and T•A base pairs, whereas those with imidazole subunits preferably bind to G•C and C•G base pairs. 410-412 Unfortunately, a mismatch in the geometries between the DNAminorgroove and polyamide geometries means that the hydrogen-bonding functionalities on each eventually become out of phase (within approximately 10 base pairs). Multiple
ing-base interference. The J-strand of the VSG-G substrate was pre- modified and analyzed for effects on JBP binding as described under “Experimental Procedures.” B and F, refer to the “bound” versus “free” DNA population, respectively. The sequence of the oligo and position of residues (i.e. ⫹ 8 to ⫺ 8) are indicated on the left. Arrowheads indicate residues that are over-represented in the free population and under- represented or absent in the bound population. Notice that base J is hypersensitive to hydrazine relative to T (34), explaining the relative increase in cleavage at this position. B, methylation interference-pro- tection. Methylation-interference (left panel). The J-strand was pre- treated with DMS (see “Experimental Procedures”) and analyzed for effects on JBP binding as in A. Arrowhead indicates the guanine at position J-1 where a methyl group in the major groove resulted in the greatest interference with JBP binding. Thus, the guanine at this position is over-represented in the free population as described in A. Methylation-protection (right panel); the effect of JBP binding on meth- ylating the guanine residues of the J-strand in the major groove was determined as described in “Experimental Procedures.” Notice that here B or U refers to methylation profiles of DNA that is either bound (B) or unbound (U) by JBP. The arrowhead indicates the increase in DMS modification at J-1 upon JBP binding. C, summary of modifica- tion-interference data for the VSG-G substrate. Interference data from A and B, and data not shown, were quantified by PhosphorImager analysis as described under “Experimental Procedures.” The distance from the line indicates the degree of interference that modification of that particular residue has on the ability of the DNA substrate to bind JBP. The degree of interference is represented as percent inhibition relative to 100% inhibition due to the removal of base J. Open boxes and
‘local’ region immediately around the Pt center. The AMBER94 force field (in MOE format) and the LFMM parameter file are provided in the Supporting Information (files SF1 and SF2). For Sim1 and Sim2, the cisplatin/DNA system was frozen and the solvent molecule positions were energy-optimized. With cisplatin/DNA still fixed, the system was then heated to 300 K in two rounds of 10 ps by coupling to a heat bath using the Berendsen algorithm followed by 100 ps simulation to equilibrate the water at a temperature of 300 K maintained by a Nosé– Hoover thermostat. All atoms were then optimized to give the starting point for LFMD simulations followed by 5 ns of production dynamics under NVT conditions. All the atoms were propagated according to Newton’s equations of motions with a time step of 2 fs at a mean temperature of 300 K using a Nosé–Hoover thermostat. Bonds involving hydrogen atoms were constrained during the simulations using the SHAKE algorithm. 60 Nonbonded cutoffs (r 1 ) of 10 Å and an onset (r 0 ) of 8 Å (the MOE default) were applied to the
Data Bank (Wenskowsky et al., 2018; Vandevenne et al., 2013). Preparation of HSA and DNA for docking simulations was done by removing the crystallized ligand, water molecules, and cofactors in the Discovery Studio 4.0 (BIOVIA Discovery Studio 2016) (BIOVIA, Dassault Systèmes, 2017). The cal- culations of Kollman charges and adding of the polar hydrogen were performed using graphical user interface AutoDockTools (ADT). The optimization of the investigated molecules was performed by B3LYP- D3BJ/6-311+G(d,p) level of theory using the Gaussian09 software package (Frisch et al., 2013). The HSA and DNA retained the rigid structure during molecular docking simula- tions in the ADT, while the investigated mol- ecules were flexible. For calculations of par- tial charges, the Geistenger method was se- lected. Lamarckian Genetic Algorithm (LGA) was used in all calculations. A docking box with a grid consisting of 60 x 60 x 60 points with 0.375 Å spacing was placed into the ac- tive side of the receptor. All binding sites of the HSA and DNA were thus covered, which enabled free movement of examined com- pounds.
Although the trans isomer of cisplatin, transplatin has no anti-tumour activity, it also binds to DNA but forms a spectrum of DNA-adducts distinct from that of cisplatin. Therefore, many studies have addressed the origin of the biological differences between cisplatin and transplatin, in an attempt to elucidate the mechanism of action of cisplatin. Using alkaline elution, Zwelling et a l, (1979) compared the level of interstrand cross links produced by these two agents. Their results showed that a much higher concentration of transplatin than cisplatin was required to produce the same level of ICLs and cytotoxicity in L1210 cells. These studies lead to the suggestion that the ICL is the major cytotoxic lesion. However, the results presented in this study do not favour this suggestion. Both the highly sensitive XPF and ERCCl mutants, and the slightly sensitive XPB and XPG mutants were unable to unhook ICLs, indicating that all these components are required for the removal of ICLs. No correlation was observed between cytotoxicity and either induction or subsequent repair of ICLs. It was surprising that, although the XPB and XPG mutants were completely defective in the unhooking step of ICL repair, they were only 1.4 and 3.1 fold more sensitive to cisplatin than the parent cell line, compared to 37-40 fold increase in sensitivity observed in XPF and ERCCI mutants. A similar study by Meyn et al., (1982) also showed that another ERCCI mutant CHO cell line, UV20, was extremely sensitive to cisplatin and defective in the repair of ICLs, suggesting a direct correlation between cytotoxicity and ICL repair. However, the ICL repair ability of the less sensitive XPB and XPG cell lines was not investigated in this study. Results shown here demonstrate for the first time that the XPB and XPG components of the NER machinery are also essential for the unhooking of cisplatin ICLs in vivo.