Bacteria possess amino acid export systems, and Corynebacterium glutamicum excretes L-isoleucine in a process dependent on the proton motive force. In order to identify the system responsible for L-isoleucine export, we have used transposon mutagenesis to isolate mutants of C. glutamicum sensitive to the peptide isoleucyl-isoleucine. In one such mutant, strong peptide sensitivity resulted from insertion into a gene designated brnF encoding a hydrophobic protein predicted to possess seven transmembrane spanning helices. brnE is located downstream of brnF and encodes a second hydrophobic protein with four putative membrane-spanning helices. A mutant deleted of both genes no longer exports L-isoleucine, whereas an overexpressing strain exports this amino acid at an increased rate. BrnF and BrnE together are also required for the export of L-leucine and L-valine. BrnFE is thus a two-component export permease specific for aliphatic hydrophobic amino acids. Upstream of brnFE and transcribed divergently is an Lrp-like regulatory gene required for active export. Searches for homologues of BrnFE show that this type of exporter is widespread in prokaryotes but lacking in eukaryotes and that both gene products which together comprise the members of a novel family, the LIV-E family, generally map together within a single operon. Comparisons of the BrnF and BrnE phylogenetic trees show that gene duplication events in the early bacterial lineage gave rise to multiple paralogues that have been retained in ␣-proteobacteria but not in other prokaryotes analyzed.

Twenty putative aminotransferase (AT) proteins of Corynebacterium glutamicum, or rather pyridoxal-5-phosphate (PLP)-dependent enzymes, were isolated and assayed among others with L-glutamate, L-aspartate, and L-alanine as amino donors and a number of 2-oxo-acids as amino acceptors. One outstanding AT identified is AlaT, which has a broad amino donor specificity utilizing (in the order of preference) L-glutamate > 2-aminobutyrate > L-aspartate with pyruvate as acceptor. Another AT is AvtA, which utilizes L-alanine to aminate 2-oxo-isovalerate, the L-valine precursor, and 2-oxo-butyrate. A second AT active with the L-valine precursor and that of the other two branched-chain amino acids, too, is IlvE, and both enzyme activities overlap partially in vivo, as demonstrated by the analysis of deletion mutants. Also identified was AroT, the aromatic AT, and this and IlvE were shown to have comparable activities with phenylpyruvate, thus demonstrating the relevance of both ATs for L-phenylalanine synthesis. We also assessed the activity of two PLPcontaining cysteine desulfurases, supplying a persulfide intermediate. One of them is SufS, which assists in the sulfur transfer pathway for the Fe-S cluster assembly. Together with the identification of further ATs and the additional analysis of deletion mutants, this results in an overview of the ATs within an organism that may not have been achieved thus far.

Although L-serine proceeds in just three steps from the glycolytic intermediate 3-phosphoglycerate, and as much as 8% of the carbon assimilated from glucose is directed via L-serine formation, previous attempts to obtain a strain producing L-serine from glucose have not been successful. We functionally identified the genes serC and serB from Corynebacterium glutamicum, coding for phosphoserine aminotransferase and phosphoserine phosphatase, respectively. The overexpression of these genes, together with the third biosynthetic serA gene, serA ⌬197 , encoding an L-serine-insensitive 3-phosphoglycerate dehydrogenase, yielded only traces of L-serine, as did the overexpression of these genes in a strain with the L-serine dehydratase gene sdaA deleted. However, reduced expression of the serine hydroxymethyltransferase gene glyA, in combination with the overexpression of serA ⌬197 , serC, and serB, resulted in a transient accumulation of up to 16 mM L-serine in the culture medium. When sdaA was also deleted, the resulting strain, C. glutamicum ⌬sdaA::pK18mobglyA (pEC-T18mob2serA ⌬197 CB), accumulated up to 86 mM L-serine with a maximal specific productivity of 1.2 mmol h ؊1 g (dry weight) ؊1 . This illustrates a high rate of L-serine formation and also utilization in the C. glutamicum wild type. Therefore, metabolic engineering of L-serine production from glucose can be achieved only by addressing the apparent key position of this amino acid in the central metabolism.The demand of L-serine is about 300 tons per year, and this amino acid is required for the pharmaceutical and the cosmetic industries, in addition to being a building block for chemical and biochemical purposes (6). The current production relies mainly on its enzymatic or cellular conversion from the precursor glycine plus a C 1 compound. Utilizing the condensing activity of serine hydroxymethyltransferase, an enzymatic system has been elaborated to convert glycine plus formaldehyde to L-serine (15). The cellular systems employed, among others, resting cells of methanol-utilizing bacteria such as Hyphomicrobium methylovorum where L-serine accumulation from glycine plus methanol was achieved (16). Also, a fermentative production of L-serine from glycine alone by Corynebacterium glycinophilum was described (19). However, there is not much information on the direct fermentative production of L-serine from glucose. Attempts to isolate L-serine-producing strains using different bacteria by applying undirected mutagenesis yielded mutants accumulating only traces of L-serine (38). Apparently, the direct conversion of glucose is a demanding challenge, probably due to the role of L-serine as a central intermediate for a number of cellular reactions (Fig. 1).We are interested in the amino acid-synthesizing capabilities of Corynebacterium glutamicum, which is traditionally used for the large-scale production of L-glutamate and L-lysine (9). In general, the efforts to engineer producing strains were focused on the enzymes of the biosynthesis pathways. For instance, con...

No abstract