Molecular Recognition in Biological Systems and Bioinformatics

A superposition of the α-carbon chains of five SCOR enzymes (different colors) that have steroids as substrates illustrates the overlap of the β -sheet structure, and the locations of the bound cofactors (orange).

Molecular Details of cofactor recognition in TGYK enzyme. (a) Typical Asp(D) NAD recognition usually associated with oxidation [PDB:2HSD] (b) Typical Arg(R) NADP(H) recognition usually associated with reduction [PDB:1NSD] (C) Unusual Thr(T)-NADP recognition found only in α, ® and © proteobacteria [PDB:1Q7B] (d) Y[AN] NADP recognition found in higher order bacterial [PDB:1GON].

Protein structure and function; genomics; X-ray crystal structure; bioinformatics.

We use X-ray crystallographic analysis to determine the structure of biologically active molecules and correlate structures with biological activity.

In 2000 we began combining information from the 3D structures of SCOR enzymes with sequence analysis of all putative SCOR genes in the gene bank to predict fold, function, cofactor, and substrate for 10,500 gene products. We are using this family of enzymes to trace the evolution of the amino acid composition of proteins. We discovered that 20% of all the genes in the gene bank, including 20% of all human genes have inexplicable anomalously high frequencies of full length multiple open reading frames (MORFs). We discovered a codon bias in the genes having MORFs that is so severe that more than 85% of the codons used in these genes are from the GC-rich half of the genetic code. On the basis of this and related data we are tracing the origin and evolution of the genetic code and the order of introduction of amino acids into proteins over the course of evolution. These data and their systematic analysis have significant implications concerning three billion years of evolution, phylogenetic relationships and protein folding.

In the last year we have developed a search vector that isolates at least 90% of all of the members of the TGxxxGIG SCOR family (10,500) with no false positives. Because there is not one fully conserved residue in the entire family the alignment is achieved by combining our analysis of the quasi conserved fold fingerprint and insertions and deletions between the critically important functional domains. To our knowledge no other investigators have been able to align more than 300 members of a family and most alignments required at least 25% sequence conservation in all family members. The fact that Prosite is not configured to deal with an alignment of more than 1000 members, and is limited to search vectors of a maximum of 200 characters indicates the unique character of what we have achieved.

We have improved methods for identifying the ensemble of residue in the three loops of the binding pocket that determine substrate binding in SCORs. We discovered that in addition to the traditional patterns of Asp recognition of NAD(H) and Arg recognition of NADP(H) the earliest members of the SCOR family exhibit at least four other recognition patterns that are confined to primarily bacteria.

We have studied in greater detail the evolution of the largest subset of the TGxxxGIG SCORs, the β-k-ACPR reductase enzymes. There are 1120 members of this enzyme family. The reason for this high population is that β-k-ACPR is an essential enzyme in fatty acid biosynthesis in all bacteria. The 1120 members of the family are in over 1000 different species.

The most unexpected results of the analysis were the discovery that NADP recognition in the most ancient bacteria (in particular a, β and γ proteobacteria) was via Thr interaction with the NADP phosphate (not Arg), that these primitive members of the family have the highest conservation of the 40 fold residues and the 9 substrate defining residues. They also have additional highly conserved residues that determine dimer and tetramer formation essential to function. A subset of the a-proteobacteria also have an exceptionally high incidence of multiple open reading frames, and a bias in codon use and amino composition suggesting that the oldest surviving members of the β-k-ACPR family are in a proteobacteria.

We have developed a number of unique programs including ones that (a) identify combinations of residue in superficially hypervariable region of substrate determining loops in large family of proteins and (b) tabulate codon and nucleotide triple bias in genes having multiple open reading frames. We have prepared modified versions of SWISS-PROP search engines that are not limited to capturing 1000 genes or less.

Members of the short chain oxidoreductase family and the (β/α) barrel, and the products of the two genes in which mutation are responsible for polycystic kidney disease.

The SCOR family includes specific enzymes involved in over a dozen human diseases including hypertension, cancer, Alzheimer's and polycystic kidney disease and has up to 200 different specific substrates.

Polycystic kidney disease affects one in 500 people worldwide. It affects more people than all other disease for which mutation of a specific gene or gene has been identified.

Carter, W.W. and Duax, W.L., Did tRNA Synthetase Classes Arise on Opposite Strands of the Same Gene, Mol. Cell., 10, 705-708 (2002).

Duax, W.L., Pletnev, V., Addlagatta, A., Bruenn, J., and Weeks, C.M., Rational Proteomics I. Fingerprint Identification and Cofactor Specificity in the Short-Chain Oxidoreductase (SCOR) Enzyme Family, PROTEINS - Structure, Function and Genetics, 53, 931-943 (2003).

Pletnev, V. Z., Weeks, C. M and Duax, W. L., Rational Proteomics II. Electrostatic Nature of Cofactor Preference in the Short-Chain Oxidoreductase (SCOR) Enzyme Family, Proteins, Structure, Function and Bioinformatics 57, 294-301 (2004).

Pletnev, V. and Duax, W. L., Rational Proteomics IV. Modeling the Primary Function of the Mammalian 17b-Hydroxysteroid Dehydrogenase type 8, J. Steroid Biochemistry & Molecular Biology, 94, 327-335 (2005).

Duax, W. L., Huether, R., Pletnev, V., Langs, D., Addlagatta, A., Connare, S., Gill, J. and Yu, P., Rational Genomics I: Antisense ORFs and Codon Bias in Short Chain Oxido Reductase Enzymes and the Evolution of the Genetic Code, Proteins: Structure, Function and Bioinformatics, 61, 900-906 (2005).

Pletnev, V. Z., Thomas, J. L., Rhaney, F. L., Holt, L. S., Scaccia, L. A., Umland, T. and Duax, W. L., Rational Proteomics V: Structure-Based Mutagenesis has Revealed Key Residues Responsible for Substrate Recognition and Catalysis by the Dehydrogenase/Isomerase Type 1, J. Steroid Biochem & Molecular Biol., 101, 50-60 (2006).

Pletnev, V., Huether, R., Habegger, L., Schultz, W. and Duax, W., Rational Proteomics of PKD1. I. Modeling the Three Dimensional Structure and Ligand Specificity of the C lectin Binding Domain of Polycystin-1, J. Mol. Modeling (2007), in press.

Huether, R., Duax, W. L., Weeks, C. M., Connare, S., Pletnev, V. and Umland, T. C., Open Reading Frames and Codon Bias in Streptomyces coelicolor and the Evolution of the Genetic Code in Proteobacteria, International Journal of Bioinformatics Research and Applications, (2007) in press.

Duax, W. L., Huether, R., Pletnev, V., Umland, T. C. and Weeks, C. M., Divergent Evolution of a Specific Protein Fold and Identification of Its Oldest Surviving Ancestor in Proteobacteria, International Journal of Bioinformatics Research and Applications, (2007) in press.