Chemical riboflavin production, successfully used for decades, is in the course of being replaced by microbial processes. These promise to save half the costs, reduce waste and energy requirements, and use renewable resources like sugar or plant oil. Three microorganisms are currently in use for industrial riboflavin production. The hemiascomycetes Ashbya gossypii, a filamentous fungus, and Candida famata, a yeast, are naturally occurring overproducers of this vitamin. To obtain riboflavin production with the gram-positive bacterium Bacillus subtilis requires at least the deregulation of purine synthesis and a mutation in a flavokinase/FAD-synthetase. It is common to all three organisms that riboflavin production is recognizable by the yellow color of the colonies. This is an important tool for the screening of improved mutants. Antimetabolites like itaconate, which inhibits the isocitrate lyase in A. gossypii, tubercidin, which inhibits purine biosynthesis in C. famata, or roseoflavin, a structural analog of riboflavin used for B. subtilis, have been applied successfully for mutant selections. The production of riboflavin by the two fungi seems to be limited by precursor supply, as was concluded from feeding and gene-overexpression experiments. Although flux studies in B. subtilis revealed an increase both in maintenance metabolism and in the oxidative part of the pentose phosphate pathway, the major limitation there seems to be the riboflavin pathway. Multiple copies of the rib genes and promoter replacements are necessary to achieve competitive productivity.

SummaryWe describe the cloning, sequence and expression of gabA, encoding the ␥-amino-n-butyrate (GABA) permease of the fungus Aspergillus nidulans. Sequence changes were determined for three up-promoter (gabI ) and six gabA loss-of-function mutations. The predicted protein contains 517 residues and shows 30.3% overall identity with a putative GABA permease of Arabidopsis thaliana, 29.6% identity with the yeast choline transporter and 23.4% identity with the yeast UGA4 GABA permease. Structural predictions favour 11-12 transmembrane domains. Comparison of the genomic and cDNA sequences shows the presence of 19 introns, an unusually large number of introns for, we believe, any fungal gene. In agreement with the wealth of genetic data available, transcript level analyses demonstrate that gabA is subject to carbon catabolite and nitrogen metabolite repression, -amino acid induction and regulation in response to ambient pH (being acid-expressed). In agreement with this, we report consensus binding sites 5Ј to the coding region, six each for CreA and AREA and one for PacC, the transcription factors mediating carbon catabolite and nitrogen metabolite repression and response to ambient pH respectively.

Determination of enzyme-specific activities revealed that GLY1 encodes a threonine aldolase (TA) in Saccharomyces cerevisiae. A knock-out mutant auxotrophic for glycine lacked detectable activity. After transformation with YEp24GLY1 glycine prototrophy was restored and TA-specific activity was 16-fold higher than in the wild type. Growth experiments using glucose as the sole carbon source showed that GLY1 is more important for glycine biosynthesis than SHM1 and SHM2 encoding alternative serine hydroxymethyltransferases. On ethanol as carbon source simultaneous disruption of GLY1, SHM1 and SHM2 did not lead to glycine auxotrophy because glycine biosynthesis proceeds via alanine glyoxylate aminotransferase.

lsocitrate lyase (ICL) was assayed during batch cultivations of Ashbya gossypii on soybean oil or glucose as carbon source. On soybean oil, a correlation between enzyme activity and riboflavin synthesis was observed. On glucose as carbon source, riboflavin overproduction started in the late growth phase when glucose was exhausted. ICL activity appeared in parallel and reached a maximum of 041 U (mg protein)-'. This suggested synthesis of vitamin B, from the intracellular reserve fat. ICL specific activity correlated with the enzyme concentration detected by specific antibodies. Itaconate, an efficient inhibitor of ICL, was used as an antimetabolite to screen mutants with enhanced ICL activity. Cultivations of an itaconate-resistant mutant on soybean oil revealed a 15 O/ o increase in enzyme specific activity and a 25-fold increase in riboflavin yield compared to the wild-type. On the other hand, growth experiments on glucose resulted in an eightfold increase in riboflavin yield but showed a 33% reduction in ICL specific activity compared to the wild-type grown on the same medium. These results support the idea of an ICL bottleneck in the riboflavin overproducer A. gossypii when plant oil is used as the substrate.

Riboflavin (vitamin B2), essential in tiny amounts as a precursor for oxidoreductase coenzymes, is a yellow pigment. Although it causes cytotoxicity via photoinduced damage of macromolecules, several microorganisms are striking overproducers. A question, unanswered for decades, is whether riboflavin overproducers can benefit from this property. Here, we report an ultraviolet (UV) protective effect of riboflavin. The spores of Ashbya gossypii, a riboflavin-overproducing fungus, are more sensitive to UV than those of Aspergillus nidulans. The addition of riboflavin to suspensions improves the UV resistance of both spore types. Interestingly, we show that regulation of sporulation and riboflavin overproduction in A. gossypii are linked. In batch culture, both were elevated when growth ceased. At constant growth rates, obtained in a chemostat culture, neither was elevated. Supplementation of cultures by cAMP, a known stress signal, negatively affected sporulation as well as riboflavin overproduction, establishing a second, independent argument for the linkage.