Preparation of riboflavin 4',5'-cyclic phosphate by incubation of flavin-adenine dinucleotide with Mn2+ in the absence of riboflavin 5'-phosphate cyclase

Synthesis of Mixed Dinucleotides by Mechanochemistry

We report the synthesis of vitamin B1, B2, and B3 derived nucleotides and dinucleotides generated either through mechanochemical or solution phase chemistry. Under the explored conditions, adenosine and thiamine proved to be particularly amenable to milling conditions. Following optimization of the chemistry related to the formation pyrophosphate bonds, mixed dinucleotides of adenine and thiamine (vitamin B1), riboflavin (vitamin B2), nicotinamide riboside and 3-carboxamide 4-pyridone riboside (both vitamin B3 derivatives) were generated in good yields. Furthermore, we report an efficient synthesis of the MW+4 isotopologue of NAD⁺ for which deuterium incorporation is present on either side of the dinucleotidic linkage, poised for isotopic tracing experiments by mass spectrometry. Many of these mixed species are novel and present unexplored possibilities to simultaneously enhance or modulate cofactor transporters and enzymes of independent biosynthetic pathways.

Characterization of a non-nudix pyrophosphatase points to interplay between flavin and NAD(H) homeostasis in Saccharomyces cerevisiae

The flavin cofactors FMN and FAD are required for a wide variety of biological processes, however, little is known about their metabolism. Here, we report the cloning and biochemical characterization of the Saccharomyces cerevisiae pyrophosphatase Fpy1p. Genetic and functional studies suggest that Fpy1p may play a key role in flavin metabolism and is the first-reported non-Nudix superfamily enzyme to display FAD pyrophosphatase activity. Characterization of mutant yeast strains found that deletion of fpy1 counteracts the adverse effects that are caused by deletion of flx1, a known mitochondrial FAD transporter. We show that Fpy1p is capable of hydrolyzing FAD, NAD(H), and ADP-ribose. The enzymatic activity of Fpy1p is dependent upon the presence of K+ and divalent metal cations, with similar kinetic parameters to those that have been reported for Nudix FAD pyrophosphatases. In addition, we report that the deletion of fpy1 intensifies the FMN-dependence of null mutants of the riboflavin kinase Fmn1p, demonstrate that fpy1 mutation abolishes the decreased fitness resulting from the deletion of the flx1 ORF, and offer a possible mechanism for the genetic interplay between fpy1, flx1 and fmn1.

An unusual diphosphatase from the PhnP family cleaves reactive FAD photoproducts

Flavins are notoriously photolabile, but while the photoproducts derived from the iso-alloxazine ring are well known the other photoproducts are not. In the case of FAD, typically the main cellular flavin, the other photoproducts are predicted to include four- and five-carbon sugars linked to ADP. These FAD photoproducts were shown to be potent glycating agents, more so than ADP-ribose. Such toxic compounds would require disposal via an ADP-sugar diphosphatase or other route. Comparative analysis of bacterial genomes uncovered a candidate disposal gene that is chromosomally clustered with genes for FAD synthesis or transport and is predicted to encode a protein of the PhnP cyclic phosphodiesterase family. The representative PhnP family enzyme from Koribacter versatilis (here named Fpd, FAD photoproduct diphosphatase) was found to have high, Mn²⁺-dependent diphosphatase activity against FAD photoproducts, FAD, and ADP-ribose, but almost no phosphodiesterase activity against riboflavin 4',5'-cyclic phosphate, a chemical breakdown product of FAD. To provide a structural basis of the unique Fpd activity, the crystal structure of K. versatilis Fpd was determined. The results place Fpd in the broad metallo-β-lactamase-like family of hydrolases, a diverse family commonly using two metals for hydrolytic catalysis. The active site of Fpd contains two Mn²⁺ ions and a bound phosphate, consistent with a diphosphatase mechanism. Our results characterize the first PhnP family member that is a diphosphatase rather than a cyclic phosphodiesterase and suggest its involvement in a cellular damage-control system that efficiently hydrolyzes the reactive, ADP-ribose-like products of FAD photodegradation.

An Fe2+-dependent cyclic phosphodiesterase catalyzes the hydrolysis of 7,8-dihydro-D-neopterin 2',3'-cyclic phosphate in methanopterin biosynthesis

7,8-Dihydro-D-neopterin 2',3'-cyclic phosphate (H(2)N-cP) is the first intermediate in biosynthesis of the pterin portion of tetrahydromethanopterin (H(4)MPT), a C(1) carrier coenzyme first identified in the methanogenic archaea. This intermediate is produced from GTP by MptA (MJ0775 gene product), a new class of GTP cyclohydrolase I [Grochowski, L. L., Xu, H., Leung, K., and White, R. H. (2007) Biochemistry 46, 6658-6667]. Here we report the identification of a cyclic phosphodiesterase that hydrolyzes the cyclic phosphate of H(2)N-cP and converts it to a mixture of 7,8-dihydro-D-neopterin 2'-monophosphate and 7,8-dihydro-d-neopterin 3'-monophosphate. The enzyme from Methanocaldococcus jannachii is designated MptB (MJ0837 gene product) to indicate that it catalyzes the second step of the biosynthesis of methanopterin. MptB is a member of the HD domain superfamily of enzymes, which require divalent metals for activity. Direct metal analysis of the recombinant enzyme demonstrated that MptB contained 1.0 mol of zinc and 0.8 mol of iron per protomer. MptB requires Fe(2+) for activity, the same as observed for MptA. Thus the first two enzymes involved in H(4)MPT biosynthesis in the archaea are Fe(2+) dependent.

Identification of human and rat FAD-AMP lyase (cyclic FMN forming) as ATP-dependent dihydroxyacetone kinases

Rat liver FAD-AMP lyase or FMN cyclase is the only known enzymatic source of the unusual flavin nucleotide riboflavin 4',5'-cyclic phosphate. To determine its molecular identity, a peptide-mass fingerprint of the purified rat enzyme was obtained. It pointed to highly related, mammalian hypothetical proteins putatively classified as dihydroxyacetone (Dha) kinases due to weaker homologies to biochemically proven Dha kinases of plants, yeasts, and bacteria. The human protein LOC26007 cDNA was used to design PCR primers. The product amplified from human brain cDNA was cloned, sequenced (GenBank Accession No. ), and found to differ from protein LOC26007 cDNA by three SNPs. Its heterologous expression yielded a protein active both as FMN cyclase and ATP-dependent Dha kinase, each activity being inhibited by the substrate(s) of the other. Cyclase and kinase activities copurified from rat liver extracts. Evidence supports that a single protein sustains both activities, probably in a single active center. Putative Dha kinases from other mammals are likely to be FMN cyclases too. Future work will profit from the availability of the structure of Citrobacter freundii Dha kinase, which contains substrate-interacting residues conserved in human Dha kinase/FMN cyclase.

Fluorimetric HPLC detection of endogenous riboflavin 4',5'-cyclic phosphate in rat liver at nanomolar concentrations

We have investigated the biological occurrence of riboflavin 4',5'-cyclic phosphate (cyclic FMN or cFMN), the flavin product known to be formed in vitro from FAD by the rat liver enzyme FAD-AMP lyase (cyclizing) or FMN cyclase (EC 4.6.1.15). The expected difficulties were the low concentration of the compound, the tendency of the more abundant FAD to decompose chemically to cFMN, and the acid lability of cFMN itself. A protocol was devised to prepare deproteinized rat liver extracts, avoiding conditions prone to the chemical formation of cFMN and making exposure to conditions of cFMN degradation as short as possible. In these extracts, cFMN was assayed by HPLC with fluorimetric detection. The identity of liver cFMN was confirmed by its HPLC separation from other known flavins, its coelution with authentic cFMN, and its susceptibility to acid degradation, yielding a mixture of 5'-FMN and 4'-FMN. The amount of total cFMN recovered in the liver extracts was 22+/-11 pmol/g fresh tissue. Careful control experiments were performed to rule out the possibility that this could be a chemical product of FAD degradation during extract preparation. These controls indicated that, on average, 97% of the measured extract concentration of cFMN, i.e., about 21+/-10 pmol/g, was already present in the liver at the beginning of the process and was extracted from the tissue. This figure is likely to be an underestimation of the hepatic content, as indicated by control experiments.