Production of ansamycin polyketide precursors in Escherichia coli

For the heterologous production of ansamycin polyketides such as rifamycin and geldanamycin in Escherichia coli, a number of unusual but important tools must be engineered into the bacterium. Here we demonstrate efficient production of the starter unit 3-amino-5-hydroxybenzoic acid (AHBA) and the methoxymalonyl extender unit in E. coli. Previous work has demonstrated the production of the ansamycin starter unit AHBA in E. coli in low yield. It was shown that the low yield is primarily due to acetylation of AHBA into N-acetyl-AHBA. Three methods for minimizing this side reaction were evaluated. First, a putative N-arylamine-acetyltransferase (NAT) was deleted from the E. coli chromosome, although this did not alter N-acetyl-AHBA production. Next, E. coli grown in media devoid of glucose yielded a nearly equal mixture of AHBA and N-acetyl-AHBA. Lastly, the NAT inhibitor glycyrrhizic acid was shown to further inhibit the acetylation reaction. The entire set of genes for synthesizing the methoxymalonyl extender unit was transferred from the geldanamycin producer Streptomyces hygroscopicus into E. coli. The pathway specific ACP isolated from the resulting recombinant strain was found to predominantly occur as methyoxymalonyl-ACP. Together, these findings set the stage for engineered biosynthesis of ansamycin polyketides in E. coli.