Combining structural genomics and enzymology: completing the picture in metabolic pathways and enzyme active sites.

An important goal of structural genomics is to complete the structural analysis of all the enzymes in metabolic pathways and to understand the structural similarities and differences. A preliminary glimpse of this type of analysis was achieved before structural genomics efforts with the glycolytic pathway and efforts are underway for many other pathways, including that of catecholamine metabolism. Structural enzymology necessitates a complete structural characterization, even for highly homologous proteins (greater than 80% sequence homology), as every active site has distinct structural features and it is these active site differences that distinguish one enzyme from another. Short cuts with homology modeling cannot be taken with our current knowledge base. Each enzyme structure in a pathway needs to be determined, including structures containing bound substrates, cofactors, products and transition state analogs, in order to obtain a complete structural and functional understanding of pathway-related enzymes.

AutoDep: a web-based system for deposition and validation of macromolecular structural -information.

This paper describes the design and full implementation of a new concept in data deposition and validation: AutoDep (copyright Brookhaven Science Associates LLC). AutoDep changes the traditional procedure for data acceptance and validation of the primary databases into an interactive depositor-driven operation which almost eliminates the delay between the acceptance of the data and its public release. The system takes full advantage of the knowledge and expertise of the experimenters, rather than relying on the database curators for the complete and accurate description of the structural experiment and its results. AutoDep, developed by the Protein Data Bank at Brookhaven National Laboratory (BNL) as a flexible and portable system, has already been adopted by other primary databases and implemented on different platforms/operating systems. AutoDep was introduced at BNL in 1996 [see Manning (1996), Protein Data Bank Quart. Newslett. 77, 2 (ftp://ftp.rcsb. org/pub/pdb/doc/newsletters/bnl/newsletter96jul/newslttr+ ++.txt); Manning (1996), Protein Data Bank Quart. Newslett. 78, 2 (ftp://ftp. rcsb.org/pub/pdb/doc/newsletters/bnl/newsletter96oct/+ ++newslttr.txt)].

Automated analysis of interatomic contacts in proteins.

MOTIVATION: New software has been designed to assist the molecular biologist in understanding the structural consequences of modifying a ligand and/or protein. RESULTS: Tools are described for the analysis of ligand-protein contacts (LPC software) and contacts of structural units (CSU software) such as helices, sheets, strands and residues. Our approach is based on a detailed analysis of interatomic contacts and interface complementarity. For any ligand or structural unit, these software automatically: (i) calculate the solvent-accessible surface of every atom; (ii) determine the contacting residues and type of interaction they undergo (hydrophobic-hydrophobic, aromatic-aromatic, etc.); (iii) indicate all putative hydrogen bonds. LPC software further predicts changes in binding strength following chemical modification of the ligand. AVAILABILITY: Both LPC and CSU can be accessed through the PDB and are integrated in the 3DB Atlas page of all PDB files. For any given file, the tools can also be accessed at http://www.pdb.bnl. gov/pdb-bin/lpc?PDB_ID= and http://www.pdb.bnl. gov/pdb-bin/csu?PDB_ID= with the four-letter PDB code added at the end in each case. Finally, LPC and CSU can be accessed at: http://sgedg.weizmann.ac.il/lpc and http://sgedg.weizmann.ac.il/csu.

The protein data bank. Bridging the gap between the sequence and 3D structure world.

The protein data bank (PDB), at Brookhaven National Laboratory, is a database containing information on experimentally determined three-dimensional structures of proteins, nucleic acids, and other biological macromolecules, with approximately 9000 entries. The PDB has a 27-year history of service to a global community of researchers, educators, and students in a wide variety of scientific disciplines. Data are easily submitted via PDB's WWW-based tool AutoDep, in either PDB or mmCIF format, and are most conveniently examined via PDB's WWW-based tool 3DB Browser. Collaborative centers have been, and continue to be, established worldwide to assist in data deposition, archiving, and distribution.

Protein Data Bank (PDB): database of three-dimensional structural information of biological macromolecules.

The Protein Data Bank (PDB) at Brookhaven National Laboratory, is a database containing experimentally determined three-dimensional structures of proteins, nucleic acids and other biological macromolecules, with approximately 8000 entries. Data are easily submitted via PDB's WWW-based tool AutoDep, in either mmCIF or PDB format, and are most conveniently examined via PDB's WWW-based tool 3DB Browser.

Designing medical informatics research and library--resource projects to increase what is learned.

Careful study of medical informatics research and library-resource projects is necessary to increase the productivity of the research and development enterprise. Medical informatics research projects can present unique problems with respect to evaluation. It is not always possible to adapt directly the evaluation methods that are commonly employed in the natural and social sciences. Problems in evaluating medical informatics projects may be overcome by formulating system development work in terms of a testable hypothesis; subdividing complex projects into modules, each of which can be developed, tested and evaluated rigorously; and utilizing qualitative studies in situations where more definitive quantitative studies are impractical.

Unexpected similarities in the crystal structures of the Mcg light-chain dimer and its hybrid with the Weir protein.

The covalently linked hybrid of two human lambda-type light chains (Mcg and Weir) crystallizes as trigonal bipyramids in ammonium sulfate [Ely et al., Molec. Immun. 22, 85-92 (1985)]. While markedly different in appearance from the barrel-shaped crystals of the parental Mcg dimer, the bipyramids of the hybrid have the same space group: trigonal P3(1)21. Moreover, the unit cell dimensions are practically identical: a = 72.3 A in both proteins; c = 188.1 A in the hybrid and 185.9 A in the Mcg dimer. These observations imply that the crystal packing and the main features of the three-dimensional structures are closely similar in the Mcg X Weir hybrid and the Mcg dimer. The "constant" domains of the Mcg and Weir proteins belong to the same genetic subclass and were expected to interact in comparable ways in hybrids and parental dimers. However, the overall similarities in the "variable" domain pairs in the hybrid and Mcg dimer were completely unpredicted, since the amino acid sequences of the heterologous variable domains differ by 36 residues. By difference Fourier analysis the Weir light chain has been tentatively identified as monomer 1 (heavy-chain analogue) and the Mcg protein as monomer 2 (light-chain analogue) in the hybrid dimer. Substitutions in key positions in the hypervariable loops explain the differences in binding activity of the Mcg and Weir dimers. In the Mcg dimer bis(dinitrophenyl)lysine spans two relatively spacious subsites (A and B), with primary contacts involving tyrosines 34 and 38 of monomer 2. The Weir dimer, which does not bind dinitrophenyl ligands, has serine and phenylalanine in homologous positions. Moreover, the bilateral replacement of valine 48 and serine 91 in Mcg by leucine and methionine in the Weir dimer should effectively block access to subsite B. In the hybrid binding activity for bis(dinitrophenyl)lysine is restored because the Mcg light chain is present as the monomer 2 subunit.

Three-dimensional structure of the Mcg IgG1 immunoglobulin.

The three-dimensional structure of an IgG1(lambda) immunoglobulin from a patient (Mcg) with amyloidosis was determined at 6.5-A resolution with X-ray diffraction techniques. The protein crystallized from water in the space group C2221, with a = 87.8, b = 111.3 and c = 186.3 A; the crystallographic asymmetric unit was a half-molecule consisting of one light and one heavy chain. The structure was solved by the multiple isomorphous replacement method with five heavy-atom derivatives. Electron density maps were interpreted with the aid of a protein modeling system used in conjunction with an Evans and Sutherland Picture System II graphics station. IgG1 molecules were tightly packed in the crystal lattice, with numerous intermolecular contacts. The two-fold axis relating identical halves of each molecule was found to be parallel to the y crystallographic axis. Electron density modules collectively representing one molecule were identified as three lobes representing the two antigen-binding (Fab) arms and the Fc region. An interchain disulfide bond connecting the two CL domains was located on the molecular diad and used as a landmark in the interpretation of the electron density map. A computer graphics method was developed to produce a solid image model of the IgG1 molecule in any prescribed orientation.

Conformational isomerism, rotational allomerism, and divergent evolution in immunoglobulin light chains.

Immunoglobulin light chains are examples of single polypeptide chains synthesized under the control of two genes. The three-dimensional structure of a human (Mcg) lambda-type chain (Bence-Jones) dimer supports the hypothesis of a common primordial gene for the amino ("variable" or V) and carboxyl ("constant" or C) halves of each monomer. However, sequence homologies have been obscurred by divergent evolution of the V and C regions ("domains"). The types of evolutionary changes that have occurred in the domain can be surmised by a comparison of the sequences, using the three-dimensional structures as a basis for alignment. Despite substantial differences in sequences, the hydrophobic character of key internal sites has been maintained in each domain. Regions present in only one domain are situated in position appropriate for their functions, but not deleterious to the general structural integrity of a common fold. The divergence of the V and C domains can be interpreted in terms of rotational allomerism. The cylinders of beta-pleated sheets have rotated in such a way that homologous regions in the two domains perform different functions in their interactions with a second molecule of light or heavy chain. These regions include complementarity-determining sites for antigen binding in the V domains and crossover sites stabilizing dimer formation in the C domains. Differences in surface properties between the V1-V2 and C1-C2 dimeric modules may partially explain why the V regions have been implicated in the formation of amyloid fibrils and in the characteristic thermal behavior of Bence-Jones proteins.