Characterisation of a synergistic interaction between a thymidylate synthase inhibitor, ZD1694, and a novel lipophilic topoisomerase I inhibitor karenitecin, BNP1100: mechanisms and clinical implications.

We developed a combination protocol for inhibitors of thymidylate synthase (TS) and DNA topoisomerase I (Topo I) that can exert highly lethal effects in vitro against HCT-8 human colorectal cancer cells. The specific schedule was constructed so that a TS inhibitor could induce not only primary DNA damage but also cellular conditions optimal for the efficient action of a Topo I inhibitor. The initial drug treatment consisted of a brief exposure to a quinazoline-based antifolate, ZD1694. After an interval of approximately one cell-doubling time, cells were exposed for 8-24 h to BNP1100, a Karenitecin-class 7-thiomethyl-camptothecin, in the presence of 1-10 microM thymidine; the latter acted as a crucial factor to promote the collision of moving replication forks with the drug-stabilised DNA-Topo I cleavable complexes even under continuous TS inhibition. Clonogenic analyses confirmed that these mechanistically distinct drugs at clinically achievable concentrations worked in a highly synergistic manner, with a maximum effect abolishing the viability of virtually all cancer cells (> 99.9%). The pretreatment with ZD1694 increased the amount of DNA-bound Topo I by up to 4-fold and the DNA-damaging capability of BNP1100 by up to 15-fold. The possibility of at least four DNA-damaging pathways is proposed which might have resulted from the individual actions of TS and Topo I inhibitors as well as their concerted actions. Taken together, the present findings provided a logically permissible explanation as to why TS and Topo I inhibitors in concerted interactions induced a highly lethal effect which was more than a simple additive effect. Since these drugs are effective specifically on actively proliferating cancer cells, but not on non-cycling G0/G1 cells, this mechanism-based protocol may warrant consideration for clinical verification.

Modulation of platinum-induced toxicities and therapeutic index: mechanistic insights and first- and second-generation protecting agents.

Platinum-type drugs have proven to be valuable in the treatment of a variety of solid tumors, beginning with the commercial approval of cisplatin 18 years ago. There are several clinically important toxicities commonly associated with the administration of these drugs. Despite the extensive use of cisplatin and carboplatin, the fundamental chemical transformations and mechanisms that underlie their antitumor and toxic effects have not been fully characterized. Several first-generation protective thiols have been clinically studied in an attempt to reduce the toxicity of platinum-type drugs; while some of these agents appear to protect against certain toxicities, nearly all platinum-protecting drugs have their own intrinsic toxicities, which can be additive to the toxicity of platinum-type drugs. Tumor protection by platinum-protecting drugs is an additional untoward effect that is associated with certain types of agents and must be addressed with care. Recent advances in theoretical and laboratory methods and the use of supercomputers have extended our understanding of the possible major mechanisms underlying platinum drug antitumor activity and toxicity; we present strong evidence that there are two classes of chemical species of platinum drug. One class appears to predominantly account for the antitumor activity, and the other class of chemical species produces many of the toxic effects of platinum drugs. We have discovered a new nontoxic, second-generation platinum-protecting agent, known as BNP7787, which appears to selectively inactivate and eliminate toxic platinum species. BNP7787 has recently entered phase I clinical testing in cancer patients.

Topotecan lactone selectively binds to double- and single-stranded DNA in the absence of topoisomerase I.

We report the first experimental observation that a clinically important camptothecin [CPT; topotecan (TPT), a water-soluble CPT] binds directly and noncovalently to double-stranded DNA and single-stranded DNA structures in the absence of topoisomerase I, but only in the lactone form. We observed clear DNA sequence specificity of the TPT lactone binding to duplex DNA, which was comprised of alternating purine-pyrimidine sequences that contained dT. These structural studies of direct TPT lactone-DNA binding support several important considerations involving possible mechanism(s) of anticancer activity of CPT-type drugs containing a 20(S) lactone moiety.

Ab initio quantum mechanical and X-ray crystallographic studies of gemcitabine and 2'-deoxy cytosine.

Gemcitabine 2',2'-difluoro 2'-deoxy cytosine (GEM) is a novel nucleoside which has demonstrated broad preclinical anti-cancer activity and appears promising in early stage human clinical trials. One purpose of this study was to characterize the energetically favored conformational modes of GEM by means of ab initio quantum mechanical studies with comparison to a novel X-ray crystallographic structure, and to determine the performance of ab initio quantum mechanical theory by comparison with X-ray structural data for GEM and 2'-deoxy cytosine (CYT). Another objective of this study was to attempt to determine key structural and electronic atomic interactions relating to the 2',2'-difluoro substitution in GEM by the application of ab initio quantum mechanical methods. To our knowledge, these are the first reported ab initio quantum mechanical geometry optimizations of nucleosides using large (e.g. 6-31G*) slit valence function basis sets. The development of accurate physicochemical models on a small scale enables us to extend our studies of GEM to more complex studies including DNA incorporation, deamination, ribonucleotide reductase inhibition, and triphosphorylation.

Recognition of Telugu characters using neural networks.

The aim of the present work is to recognize printed and handwritten Telugu characters using artificial neural networks (ANNs). Earlier work on recognition of Telugu characters has been done using conventional pattern recognition techniques. We make an initial attempt here of using neural networks for recognition with the aim of improving upon earlier methods which do not perform effectively in the presence of noise and distortion in the characters. The Hopfield model of neural network working as an associative memory is chosen for recognition purposes initially. Due to limitation in the capacity of the Hopfield neural network, we propose a new scheme named here as the Multiple Neural Network Associative Memory (MNNAM). The limitation in storage capacity has been overcome by combining multiple neural networks which work in parallel. It is also demonstrated that the Hopfield network is suitable for recognizing noisy printed characters as well as handwritten characters written by different "hands" in a variety of styles. Detailed experiments have been carried out using several learning strategies and results are reported. It is shown here that satisfactory recognition is possible using the proposed strategy. A detailed preprocessing scheme of the Telugu characters from digitized documents is also described.

Models of delta-hemolysin membrane channels and crystal structures.

Molecular modeling and energy calculations have been used to study how delta-hemolysin and melittin helices may aggregate on membrane surfaces and insert through membranes to form channels. In these models adjacent antiparallel amphipathic helices form planar "raft" structures, in which one surface is hydrophobic and the other hydrophilic. Models of delta-hemolysin crystal structure were developed using these "rafts." These models are based on the unit cell constants and the crystal symmetry obtained from the preliminary crystal data. Energy calculations favor channel models of delta-hemolysin with six or eight monomers per channel.

Molecular model of the action potential sodium channel.

Secondary and tertiary structural models of sodium channel transmembrane segments were developed from its recently determined primary sequence in Electrophorus electricus. The model has four homologous domains, and each domain has eight homologous transmembrane segments, S1 through S8. Each domain contains three relatively apolar segments (S1, S2 and S3) and two very apolar segments (S5 and S8), all postulated to be transmembrane alpha-helices. S4 segments have positively charged residues, mainly arginines, at every third residue. The model channel lining is formed by four S4 transmembrane alpha-helices and four negatively charged S7 segments. S7 segments are postulated to be short, partially transmembrane amphipathic alpha-helices in three domains and a beta-strand in the last domain. S7 segments are preceded by short apolar segments (S6) postulated to be alpha-helices in three domains and a beta-strand in the last domain. Positively charged side chains of S4 form salt bridges with negatively charged side chains on S7 and near the ends of S1 and S3. Putative extracellular segments that contain 5 of the 10 potential N-glycosylation sites link S5 to S6. Channel activation may involve a 'helical screw' mechanism in which S4 helices rotate around their axes as they move toward the extracellular surface.

Origins of life: conformational energy calculations on primitive tRNA nestling an amino acid.

We report conformational energy calculations on our proposal of a molecular interaction theory for the origin of the nucleic acid-directed, adaptor-mediated synthesis of proteins that links the phenomena of chemical and biological evolution. A particular conformation of a pentanucleotide turns out to be a double-sided template for a primitive decoding system. It is able to neatly nestle an amino acid via hydrogen bonds, and this complex is found to be an energetically favourable conformation. The total potential energy of the complex is calculated using semi-empirical potential energy functions. A local-minimum conformation is obtained and its features are reported. The template conformation of the pentanucleotide is found to have an energy value far lower than a regular helical conformation. When the amino acid is nestled in the cleft of the template-conformation through specific hydrogen bonds, the energy is further lowered. A D-amino acid nestled into the PIT (Primitive tRNA) is found to be less stable than its L counterpart, as revealed by energy calculations.

A conformational rationale for the replication of nucleic acids under prebiotic conditions.

The Watson-Crick model at once gave an explanation for the mechanism of replication of DNA. But the hydrogen-bonding forces between the bases alone are not enough for the specificity of base-pairing mechanisms, since any pair of bases can be positioned to have at least two hydrogen bonds. In the present-day biological organisms, sophisticated enzymatic machinery is supposed to constrain the ribose-phosphate backbone to have regular structure, aiding the self-templating duplication. For the prebiotic stage, whence sophisticated enzymes would not have been evolved, we propose a novel double helical conformation of DNA wherein the two sugar-phosphate backbones are pulled towards each other by (C-H - -O) hydrogen bonds conferring stereospecificity for the formation of (A:T)- and (G:C)-pairs, in the self-templating chains of DNA. Our model-building efforts and computer calculations endorse the stereochemical feasibility of the conformation. The pairing of homologous sequences of two double helices of DNA is explained by direct hydrogen-bonding interactions in our model and it is thus relevant to the present-day biological functions also, at least in some stages of the cell-cycle.

A conformational rationale for the wobble behaviour of the first base of the anticodon triplet in tRNA.

We present a conformational rationale for wobble behaviour of the first base in the anticodon triplet of tRNA and hence for the well-known degeneracy of the genetic code. The U-turn hydrogen bond plays an important role in the structure of the anticodon arm and particularly for the anticodon triplet to be in a geometry suitable for the process of recognition in the adaptor-mediated synthesis of proteins. This hydrogen bond in turn precludes a hydrogen bond between the first two sugars of the anticodon triplet, allowing the first base to wobble, while it facilitates one between the second and third sugars of the triplet, positioning these bases for the standard base-pairing with the codon. This neatly explains why there is a degeneracy in the code and why a RNA happens to be the adaptor for protein synthesis. Relevent conformational calculations are presented in support of the theory.

Possible role of RNA-dependent DNA-polymerase in early stages of evolution.

The tDNA cistrons have permuted sequences of triplets corresponding to anti-codons in tRNA at specific regions in their sequences. We invoke reverse transcription for the generation of such sequences in the genome during early stages of evolution. Making the assumption that a single tDNA cistron, in a genome might have come into existence by an 'accident', after transcription, tRNA is expected to fold into a three-dimensional shape analogous to the contemporary tRNA, where the anti-codon triplet bases are sticking out well-exposed for chemical mutagens. The mutated tRNAs would have been reverse-transcribed into the genome by crude analogs of now-known reverse-transcriptases. The back and forth process of transcription and reverse transcription would give rise to all the tDNA cistrons with the required anti-codons. This process may act as an important feedback mechanism for the efficient progress of evolution.

A conformational rationale for the origin of the mechanism of nucleicacid-directed protein synthesis of 'living' organisms.

The physical basis for the natural evolution of a primitive decoding system is presented using the concepts of molecular interactions. Oligoribonucleotides of five residues having U at the 5'-end, a purine at the 3'-end and any combination of three bases in the middle is taken as a primitive tRNA (PIT). From conformational considerations PIT is expected to have U-turn conformation wherein, N3-H3 of base U hydrogen-bonds with phosphate, three residues ahead leaving triplet bases called primitive anticodons (PAC) into a helical conformation, and this creates a cleft between U and PAC. An amino acid can be comfortably nestled into the cleft with the amide hydrogens and carboxyl oxygen hydrogen-bonded to the last purine and the first uridine, while the side-chain can interact with the cleft side of PAC. The other side of PAC is free to base-pair with triplet codons on a longer RNA. Also two PACs can 'recognize' consecutive triplet codons, and this leads to a dynamic interaction in which the amino and carboxyl ends are brought into proximity, making the formation of peptide bond feasible. The cleft formed by different anticodon triplets, broadly speaking, shows preferences for the corresponding amino acids of the presently known codon assignment. Thus the nucleicacid-directed protein synthesis, which is a unique feature of all 'living' organisms is shown to be a natural consequence of a particular way of favourable interaction between nucleic acids and amino acids, and our model provides the missing link between the chemical evolution of small organic molecules and biological evolution through the process of mutations in nucleicacids and nucleicacid-directed protein synthesis.