Benzoylformate decarboxylase encoded by ro02985 in Rhodococcus jostii RHA1 which may involve in lignin catabolism were investigated. In addition, BFD encoded by 92745 in Gram-negative bacteria Pseudomonas fluorescens pf-5 was also cloned and compared with that in RHA1. To improve the solubility of proteins, pETite N-His SUMO Kan Vector was used to overexpress the proteins. Lignin model compounds such as guaiacylglycerol, 1-(4-hydroxy-3-methoxyphenyl)propane-1,2-dione, and 2-hydroxy-1-(4-hydroxy-3-methoxyphenyl)propan-1-one were synthesised. BFD activities were tested using commercially available benzoylformate, via HPLC assay. Evaluation of the BFD activities reacted with synthesised lignin model compound were also tested in different buffer. The degraded products were then be identified by HPLC and LC/MS assay. Moreover, the reverse C-C bond formation reaction catalysed by BFD were examined using vanillin and aldehyde including acetaldehyde, glycolaldehyde, propionaldehyde and benzaldehyde. Moreover, the crystal structure of the P.f BFD has been solved at 1.8 Å resolution.
Results showed that a 1 L fermentation yielded about 9 mg purified recombinant R.jBFD and 7.6 mg P.fBFD with specific decarboxylase activities of 28 U/mg and 63 U/mg respectively. Then we tested the abilities of both the enzymes to catalyse the lignin model compounds aryl C3 molecules containing COCH(OH)CH2(OH) and analogues However, the HPLC traces of enzymatic reactions under the optimal conditions (pH=7, 25 ℃) showed that even after 12 h there were no measurable conversion of these substrates, which suggested that BFD would not accept the tested compounds as substrates under the current conditions. Meanwhile, it was found that both the BFD were able to catalyse the C-C bond formation between vanillin and acetaldehyde due to the formation of 2-hydroxy-1-(4-hydroxy-3-methoxyphenyl)propan-1-one, but they were unable to cleavage the C-C bond when using the product as substrate. The alignment between these two proteins showed that they shared about 37% sequence identity and 20% sequence similarity, which indicated that they possessed the similar structure and function. However, the decarboxylase activity and carboligase activity of P.fBFD is much stronger than R.jBFD under the same conditions. Comparison of the X-ray structures of both enzymes identified one residue in each (Ser73 in RHA1 and Ala73 in Pf-5) that were likely to be involved in determining the activity. Site-directed mutagenesis was conducted to interchange the residues in both R.jBFD and P.fBFD. LC/MS results revealed that BFD variant containing the A73S mutation was found to lose most of its activity, while BFDS73A mutant showed an enhanced activity. These findings demonstrated that steric restraint might be predominantly responsible for the differences in activity. Based on the result of this project, another lignin model compounds-3-methoxy-4-hydroxy-benzoylformate and its analogues-will be tested in future work.
Meanwhile the researcher tried to design a deletion system to remove genes encoding BFD in both RHA1 and pf-5. However, although construction of the mutagenic plasmid for ΔR.jbfd and ΔP.fbfd genetic knockout strains were successful, the researchers haven’t obtain any proof of plasmid integration into the chromosomal DNA.
The results from the Action have been presented in 3 high-level international conferences (2 poster presentation, 1 flash oral presentation), and will be written up for publication in scientific journals in the next 6 months. The biochemical characterisation of BFD catalysed pathway studied in this project may boost future projects considering the break-down of lignin.
