α-Methyl-L-serine

By screening microorganisms that are capable of assimilating alpha-methyl-DL-serine, we detected alpha-methylserine aldolase in Ralstonia sp. strain AJ110405, Variovorax paradoxus AJ110406, and Bosea sp. strain AJ110407. A homogeneous form of this enzyme was purified from Ralstonia sp. strain AJ110405, and the gene encoding the enzyme was cloned and expressed in Escherichia coli. The enzyme appeared to be a homodimer consisting of identical subunits, and its molecular mass was found to be 47 kDa. It contained 0.7 to 0.8 mol of pyridoxal 5'-phosphate per mol of subunit and could catalyze the interconversion of alpha-methyl-L-serine to L-alanine and formaldehyde in the absence of tetrahydrofolate. Formaldehyde was generated from alpha-methyl-L-serine but not from alpha-methyl-D-serine, L-serine, or D-serine. Alpha-methyl-L-serine synthesis activity was detected when L-alanine was used as the substrate. In contrast, no activity was detected when D-alanine was used as the substrate. In the alpha-methyl-L-serine synthesis reaction, the enzymatic activity was inhibited by an excess amount of formaldehyde, which was one of the substrates. We used cells of E. coli as a whole-cell catalyst to express the gene encoding alpha-methylserine aldolase and effectively obtained a high yield of optically pure alpha-methyl-L-serine using L-alanine and formaldehyde.

Nonribosomal peptide synthetases (NRPSs) are attractive targets for bioengineering to generate useful peptides. FmoA3 is a single modular NRPS composed of heterocyclization (Cy), adenylation (A), and peptidyl carrier protein (PCP) domains. It uses α-methyl-l-serine to synthesize a 4-methyloxazoline ring, probably with another Cy domain in the preceding module FmoA2. Here, we determined the head-to-tail homodimeric structures of FmoA3 by X-ray crystallography (apo-form, with adenylyl-imidodiphosphate and α-methyl-l-seryl-AMP) and cryogenic electron microscopy single particle analysis, and performed site-directed mutagenesis experiments. The data revealed that α-methyl-l-serine can be accommodated in the active site because of the extra space around Ala688. The Cy domains of FmoA2 and FmoA3 catalyze peptide bond formation and heterocyclization, respectively. FmoA3's Cy domain seems to lose its donor PCP binding activity. The collective data support a proposed catalytic cycle of FmoA3.

JBIR-34 and -35 produced by Streptomyces sp. Sp080513GE-23 are nonribosomal peptides that possess an unusual 4-methyloxazoline moiety. Through draft genome sequencing, cosmid cloning, and gene disruption, the JBIR-34 and -35 biosynthesis gene cluster (fmo cluster) was identified; it encodes 20 proteins including five nonribosomal peptide synthetases (NRPSs). Disruption of one of these NRPS genes (fmoA3) resulted in no JBIR-34 and -35 production and accumulation of 6-chloro-4-hydroxyindole-3-carboxylic acid. Stable isotope-feeding experiments indicated that the methyl group of the methyloxazoline ring is derived from alanine rather than methionine. A recombinant FmoH protein, a glycine/serine hydroxymethyltransferase homolog, catalyzed conversion of α-methyl-l-serine into d-alanine (the reverse reaction of α-methyl-l-serine synthesis catalyzed by FmoH in vivo). Taken together, we concluded that α-methyl-l-serine synthesized from d-alanine is incorporated into JBIR-34 and -35 to form the 4-methyloxazoline moiety. We also propose the biosynthesis pathway of JBIR-34 and -35.