Abstract
BACKGROUND: Fumonisin B(1) is a cancerogenic mycotoxin produced by Fusarium verticillioides and other fungi. Sphingopyxis sp. MTA144 can degrade fumonisin B(1), and a key enzyme in the catabolic pathway is an aminotransferase which removes the C2-amino group from hydrolyzed fumonisin B(1). In order to study this aminotransferase with respect to a possible future application in enzymatic fumonisin detoxification, we attempted expression of the corresponding fumI gene in E. coli and purification of the enzyme. Since the aminotransferase initially accumulated in inclusion bodies, we compared the effects of induction level, host strain, expression temperature, solubility enhancers and a fusion partner on enzyme solubility and activity. RESULTS: When expressed from a T7 promoter at 30 degrees C, the aminotransferase accumulated invariably in inclusion bodies in DE3 lysogens of the E. coli strains BL21, HMS174, Rosetta 2, Origami 2, or Rosetta-gami. Omission of the isopropyl-beta-D-thiogalactopyranoside (IPTG) used for induction caused a reduction of expression level, but no enhancement of solubility. Likewise, protein production but not solubility correlated with the IPTG concentration in E. coli Tuner(DE3). Addition of the solubility enhancers betaine and sorbitol or the co-enzyme pyridoxal phosphate showed no effect. Maltose-binding protein, used as an N-terminal fusion partner, promoted solubility at 30 degrees C or less, but not at 37 degrees C. Low enzyme activity and subsequent aggregation in the course of purification and cleavage indicated that the soluble fusion protein contained incorrectly folded aminotransferase. Expression in E. coli ArcticExpress(DE3), which co-expresses two cold-adapted chaperonins, at 11 degrees C finally resulted in production of appreciable amounts of active enzyme. Since His tag-mediated affinity purification from this strain was hindered by co-elution of chaperonin, two steps of chromatography with optimized imidazole concentration in the binding buffer were performed to obtain 1.45 mg of apparently homogeneous aminotransferase per liter of expression culture. CONCLUSIONS: We found that only reduction of temperature, but not reduction of expression level or fusion to maltose-binding protein helped to produce correctly folded, active aminotransferase FumI in E. coli. Our results may provide a starting point for soluble expression of related aminotransferases or other aggregation-prone proteins in E. coli.