ABSTRACT
The above in silico prediction for a folate-dependent tRNA:m5U methyltransferase was next validated by demonstrating that purified recombinant B. subtilis TrmFO contains tightly bound flavin and has the expected U54-tRNA methyltransferase activity when tested in vitro with CH2-H4fblate as methyl donor and NADH/NADPH as reductant. Moreover, genetic experiments demonstrated that a viable E. coli strain defective for the SAM-dependent TrmA activity (carry ing a point mutation trmAS) was successfully complemented after transformation with a plasmid carrying the gene coding for B. subtilis trmFO {gid)}%;2& This indicates that E. coli is readily able to switch from using its normal SAM-dependent TrmA to a heterologous folate-dependent TrmFO enzyme. Unfortunately, further in vitro biochemical studies on the J?. subtilis TrmFO enzyme have been hampered by the low efficiency of the methylation reaction (rate and yield) observed under in vitro conditions.18,28 An explanation for these results could be that B. subtilis TrmFO protein binds CH2-H4folate poorly, as observed in earlier works,27 or that CH2-H4folate and/or FADH2 are unexpectedly unstable under our experimental conditions. However, this latter possibility ap pears unlikely, as the experimental conditions used with the B. subtilis TrmFO enzyme were very similar to those used with several different ThyX proteins. More likely, the experimental difficulties could be explained by very different types of substrates receiving methyl groups in the reaction catalyzed by the two types of flavoproteins.
