August 2009 featured article
An unusual reaction in the production of thymine offers hope for anti-tuberculosis drugs.
Crystal structures of the FDTS–FAD–dUMP complex. a, Wild-type TmFDTS (PDB 1O26); b, S88A mutant (PDB 3G4A); c, S88C mutant (PDB 3G4C).
Several antibiotic drugs target the synthesis of the DNA base thymine by bacteria to prevent them replicating. These drugs aim to block classical thymidylate synthases, which add a methyl to the uracil of deoxyuridine monophosphate (dUMP) to produce deoxythymidine monophosphate (dTMP). But classical thymidylate synthases are present in humans as well prokaryotes, so attacking this enzyme risks potential side effects.
Recently, a new class of thymidylate synthases was discovered, called the flavin-dependent thymidylate synthases (FDTSs). They are encoded by a thyX gene, whereas the better studied version is encoded by thyA and TYMS genes. This group of enzymes is present mainly in prokaryotes and viruses and is not found in humans. In addition, they have no structural or sequence similarity to classical thymidylate synthases, and so could be a promising new drug target. Of particular interest is that they are found in several devastating pathogens, such a tuberculosis, and in agents of biological warfare.
Classical thymidylate synthases use a cysteine residue in their active site to activate the uracil ring of dUMP, and it has been suggested that FDTS might use a similar mechanism. Sequence analysis suggested that a conserved serine in the active site acts as the nucleophile.
Several structures of various FDTSs from different organisms indicated, however, that this serine was located 4 Å from the C6 position of dUMP – too far away for it be part of the catalytic mechanism. Koehn et al. mutated this serine to an inactive alanine and found that this alteration did not prevent thymidylate synthase activity. Thus, serine is not the catalytic nucleophile in the FDTS reaction.
To understand how this unusual enzyme works, Koehn et al. crystallized thymidylate synthase from Thermotoga maritima with the serine mutated either to alanine (S88A) or to cysteine (S88C) and compared their structures to the wild type. There was no covalent bond between the cysteine and dUMP in the latter crystal structure, despite previous reports of a bond between C6 of the uracil and the cysteine. It appears that this complex is not part of the pathway and instead forms an inhibitory complex.
To establish the nature of the FDTS-catalysed reaction, Koehn et al. followed the flow of hydrogens during the reaction by introducing hydrogen isotopes that could be tracked using mass spectroscopy and NMR spectroscopy. Their results indicate that an intermediate hydride (a proton and two electrons) is transferred from the reduced flavin cofactor directly to the uracil ring. The intermediate then isomerizes to form the product dTMP. Formation of such an intermediate has never been reported before for FDTSs, but it appears to be chemically feasible and stable in solution.
This elegant combination of crystallography and detailed mechanistic studies has revealed a new nucleotide methylation pathway that will be of great interest to biochemists wishing to design drugs to thwart bacterial replication.