THIAMIN BIOSYNTHESIS
We have made substantial progress in understanding the complex thiamin pyrophosphate (10) biosynthetic pathway over the past decade: we have cloned and characterized 12 of the 15 required proteins and successfully reconstituted the entire 11 step biosynthetic pathway using purified enzymes. In collaboration with Steve Ealick and Linda Nicholson, we have determined the structures of 6 of these proteins. The mechanisms of all of the later steps on the pathway are well understood and the complex mechanism of the thiazole phosphate (4) formation is now becoming clear. The mechanism of the fascinating rearrangement, resulting in the formation of the pyrimidine phosphate (7), is still a mystery. Some of the highlights of our thiamin research are as follows:

· The discovery of a new biosynthetic strategy for sulfur transfer involving the formation of peptide thiocarboxylate and a peptide acyldisulfide intermediates.
· The determination of the structure of the first enzyme carbocation complex (thiamin phosphate synthase/pyrimidine carbocation/thiazole phosphate/pyrophosphate complex).
· The determination of the structure of the first bacterial ancestor of the eukaryotic ubiquitinating system.
· The determination of the mechanism of toxicity of the antibiotic bacimethrin. This compound is converted to a toxic thiamin analog six times faster than the natural substrate by the last three enzymes on the thiamin biosynthetic pathway.
· The determination of the mechanism of thiaminase I, a thiamin degrading enzyme.

COENZYME A BIOSYNTHESIS
The decarboxylation of phosphopantothenoylcysteine (12), an intermediate on the Coenzyme A biosynthetic pathway, presents an interesting mechanistic puzzle because there is no obvious way to stabilize the carbanion intermediate (13). We have cloned and overexpressed this gene from E.coli, determined that the phosphopantothenoylcysteine synthetase and decarboxylase genes are fused, that the decarboxylase is a flavoenzyme and characterized the mechanism of the reaction. Some of the highlights of our research on this enzyme are:

· The demonstration, using a cyclopropyl substituted substrate analog, that the decarboxylation is likely to occur via a sulfur centered radical intermediate.
· The demonstration that the Coenzyme A biosynthetic pathway is highly selective for the incorporation of cysteine over serine (selectivity>5x105) and that this selectivity is mediated by the synthase and decarboxylase.
· The determination of the mechanism of action of pentyl pantothenamide. This antibiotic is a better substrate for the Coenzyme A biosynthetic enzymes than the natural substrate and is converted to ethyldethio-Coenzyme A, a toxic acetyl Coenzyme A analog.

Our research on phosphopantothenoylcysteine decarboxylase is part of a larger project being carried out in collaboration with Steve Ealick on the enzymatic stabilization of high-energy carbanions. Two other systems, on which we have made significant progress, are the enzymes catalyzing the decarboxylation of orotidine monophosphate (OMP, 15) and oxalate (18). OMP decarboxylase is likely to use active site super acid chemistry (i.e. C-C bond protonation) to avoid the formation of the high-energy vinyl carbanion (16) and oxalate decarboxylase utilizes radical chemistry to avoid the formation of the acyl carbanion (19).

ADENOSYLCOBALAMIN BIOSYNTHESIS
Dimethylbenzimidazole (23) is one of the axial ligands of adenosyl cobalamin. The biosynthesis of this ligand from riboflavin (21) is now the major unsolved problem in vitamin B12 biosynthesis. We propose that this occurs by the hydrolytic degradation of riboflavin (21) to an electron rich diaminobenzene (22). We have recently demonstrated (in collaboration with Jorge Escalante-Semerena) that this compound is converted to dimethylbenzimidazole (23) by a facile non-enzymatic oxidative cascade in which the electron rich diamine (22) undergoes iterative 2-electron oxidations first activating the system for cyclization and then for sidechain cleavage. We believe that such oxidative cascades constitute a general biosynthetic motif allowing the assembly of complex structures from simple aromatic compounds. We have uncovered two other examples of this chemistry: the oxidation of aminophenols (24) to generate the phenoxazinone chromophore (25) of the antibiotic actinomycin and the oxidation of a tyrosine residue in the pyoverdin precursor peptide (26) to generate the fluorescent iron binding fluorophore (27) of this siderophore.

NICOTINAMIDE ADENINE DINUCLEOTIDE BIOSYNTHESIS
The mechanistic enzymology of the NAD pyridine ring biosynthesis has not been worked out for any system. The eukaryotic pathway involves a five-step conversion of tryptophan (28) to quinolinic acid (33) and is of interest because two of the products generated (31 and 33) are neurotoxins. However, mechanistic studies have been limited because of severe unsolved overexpression problems. We have identified the complete tryptophan to quinolinic acid pathway in a few bacteria using comparative genome analysis and gene clustering (collaboration with Andrei Osterman). We have now overexpressed all five enzymes, at a high level as soluble proteins, opening this system for the first time to comprehensive mechanistic and structural analysis.

COMPLETING THE FUNCTIONAL ASSIGNMENT OF THE
ESCHERICHIA COLI PROTEOME
While more than 400 bacterial genomes have been sequenced, the complete functional assignment of the proteome has not been achieved for any living system. In E.coli, the best studied cell in biology, 726 of the 4286 genes are still without functional assignment, and for a new genome sequence, only half of the genes can typically be assigned. Over the past five years, the Begley group has assigned function to 16 bacterial genes. We are intrigued by the experimental challenge of determining the function of unassigned genes and we have begun a multidisciplinary project involving geneticists, molecular biologists, biochemists, and chemists to complete the functional assignment of the E.coli proteome. This is a critically important goal in proteomics and must be achieved as an essential step to understanding cellular life at a molecular level.