Search

Bioinformatics involves both the automatic processing of large amounts of existing data and the creation of new types of information resource. Both will be required if the data are to be transformed into information and used to help in the discovery of drugs.

In the development of a treatment for AIDS, the HIV-1 protease has been identified as a good target enzyme for inhibitor design. We previously reported a series of dimeric penicillin-derived C2-symmetric HIV-1 protease inhibitors [Humber, D., et al. (1993) J. Med. Chem. 36, 3120-3128]. In an attempt to reduce the size and optimize the binding of these C2-symmetric inhibitors, molecular modeling studies led to a novel series of monomeric penicillin-derived inhibitors of HIV-1 protease. The binding modes of these monomeric inhibitors have been characterized by X-ray crystallographic and NMR studies. Crystal structures of HIV-1 protease complexed to three inhibitors (GR123976, GR126045, and GR137615) from this series identify the molecular details of the interactions. The binding of GR123976 (IC50 = 2.3 microM) exhibits good hydrophobic contacts but few electrostatic interactions. A strategy of structure-based design and chemical synthesis led to the elaboration of GR123976 to optimize interactions with the protein. Crystallographic analysis of HIV-1 protease complexed to GR126045 and GR137615 identified these interactions with the catalytic aspartates and the protein binding pockets. The crystal structures of the three complexes confirm the presence of the major interactions modeled in order to optimize potency and reveal details of the molecular recognition by HIV-1 protease of this novel series of nonpeptidic inhibitors.

A comparison has been made between the homology and hydrophobicity profiles of six interleukin amino acid sequences and that of the human interleukin 1 beta (IL-1 beta) for which a crystal structure exists. The resulting sequence alignment was used to build model structures for the sequences for three IL-1 alpha, two IL-1 beta and an interleukin receptor antagonist. Analysis of these structures demonstrates that the interleukin molecule has a strong electric dipole which is generated by the topological position of the amino acids in the sequence. Electrostatic surface calculations implicate a particular residues (Lys145) as being fundamental to interleukin activity and this supports site-directed mutation evidence that this residue is required for activity.

The binding modes of a series of penicillin-derived C2 symmetric dimer inhibitors of HIV-1 proteinase were investigated by NMR, protein crystallography, and molecular modeling. The compounds were found to bind in a symmetrical fashion, tracing and S-shaped course through the active site, with good hydrophobic interactions in the S1/S1' and S2/S2' pockets and hydrogen bonding of inhibitor amide groups. Interactions with the catalytic aspartates appeared poor and the protein conformation was very similar to that seen in complexes with peptidomimetics, in spite of the major differences in ligand structure.

An analysis of the stereochemistry of hydrogen bonding and metal binding to some nitrogen-containing heterocycles found in crystal structure determinations has shown that the interacting atom will generally lie in the plane of the heterocyclic ring system in a direction that approximately bisects the C-N-C angle of the heterocycle. The Cambridge Structural Database (CSD) of crystal structures of small molecules was used for this analysis because stereochemical data are available at high resolution and are amendable to comparative analysis. It was found that, for hydrogen bonding, a slight out-of-plane deviation of the binding atom is marginally more likely than an in-plane deviation. Metal ions appear to bind in a manner that is similar to that of hydrogen bonding to a protonated heterocycle, no matter what the chemical identity of the metal. The binding is more rigid, with less in-plane or out-of-plane deviation of the metal ion compared to the interaction with a hydrogen-bonding group. Some ab initio molecular orbital energy calculations give a measure of the energies involved when metal ions or hydrogen-bonding groups deviate from the plane of the ring system or from the line bisecting the C-N-C angle of the heterocycle. These results are compared with reported structural data (at lower resolution) for some acridine-oligonucleotide complexes and the surroundings of histidine rings in some protein crystal structures.

The solution structure of the fibrinogen antagonist, echistatin, has been determined by a combination of NMR and simulated annealing methods. While the structure of the disulphide-linked core is well-defined by the NMR data, the N- and C-termini and the loop bearing the RGD sequence (which is responsible for the fibrinogen antagonist properties) are poorly defined. The pattern of disulphide bridges, which could not be determined by classical methods, was predicted by a statistical analysis of the simulated annealing structures. This pattern is distinct from that for the homologous protein kistrin, leading to the novel suggestion that homologous proteins possess non-conserved patterns of disulphide bridges.

The 1H-NMR spectrum of the snake toxin echistatin has been assigned using homonuclear two-dimensional methods. Consideration of the NOE patterns, coupling constants and putative hydrogen bonds enabled two regular features of secondary structure to be deduced: a beta-sheet/turn between residues 8 and 13 and a small anti-parallel beta-sheet and bulge linking residues 16-20 with residues 30-33. The recognition region of the protein containing the residues RGD lies in a loop joining the two strands of the beta-sheet. The beta-bulge and the loop containing the RGD sequence undergo pH-dependent conformational interconversion, modulated by the side chain of Asp29.

Rapid advances in site-directed mutagenesis and total gene synthesis combined with new expression systems in prokaryotic and eukaryotic cells have provided the molecular biologist with tools for modification of existing proteins to improve catalytic activity, stability and selectivity, for construction of chimeric molecules and for synthesis of completely novel molecules that may be endowed with some useful activity. Such protein engineering can be seen as a cycle in which the structures of engineered molecules are studied by X-ray analysis and two-dimensional nuclear magnetic resonance. The results are used in the improvement of the design by using knowledge-based procedures that exploit facts, rules and observations about proteins of known three-dimensional structure.