Monday, April 02, 2007

"It's been a long time (since I rock and rolled)" - Robert Plant

A former member of our lab found that adding adenosine monophosphate (AMP) or guanosine monophosphate (GMP) to MIV or sBHI culture media greatly reduces transformation frequency in H. influenzae (Macfadyen et al., 2001, Mol. Micro.). How do nucleotides down-regulate competence in H. influenzae? An earlier candidate for mediating this response was the PurR protein that represses genes for purine biosynthesis when intracelleluar pools of guanine or hypoxanthine (a precursor of AMP and GMP) are high. Thus, in this early model, PurR blocked the expression of purine biosynthesis genes AND competence genes; however, knocking out purR did not result in elevated competence. Bioinformatic analysis still implicates PurR in the regulation of competence because the rec2 gene has a near perfect PurR binding site in its promoter. This leads to a model in which PurR plays only a limited role in regulating competence, which may be tricky to detect in the laboratory conditions used to study competence.

In collaboration with a former member of the lab, we found that transcription of the sxy gene is reduced in MIV+AMP, resulting in a dramatic decrease in Sxy protein levels. However, cells also suffer general malaise in the presence of these high concentrations of nucleotides, thus we need to repeat these experiments with controls for deleterious effects on the expression of non-CRP-S regulated genes. In other words, we can’t say for sure that nucleotides repress competence due to a specific effect on sxy expression. I recently measured the levels of Sxy protein in hypercompetent strains and found that even though cells have very high levels of Sxy protein, they experience the same repressing effect as wildtype cells. This raises three possibilities: (1) Nucleotide repression arises simply because cells are sickly in nucleotide-supplemented media, (2) An as yet unrecognized factor specifically represses competence in the presence of nucleotides, or (3) Sxy has a nucleotide allosteric effector. In (3), nucleotide binding to Sxy is expected to prevent the protein from acting at CRP-S promoters, in contrast to how purine bases cause PurR to bind DNA.

Possibilities number (1) and (2) can begin to be addressed by measuring the effect of nucleotide supplementation on the expression of genes that do not require Sxy for expression (possibility number (3) requires the same controls to know that we are testing effects on Sxy). For (3), we can now test the effects of nucleotide supplementation on competence gene expression in an E. coli strain that overexpresses sxy; this is much like the H. influenzae hypercompetence mutants, so we might see that competence gene expression is repressed even when sxy expression is high. If so, it may be worth testing if Sxy binds to nucleotide-coated resins.

What got me thinking about whether Sxy has an effector molecule was my latest BLAST search for recognizable protein domains. The result has been the same for the last five years: the top hit that isn’t a Sxy ortholog is the weak alignment of Sxy’s C-terminal half to a cytidine deaminase from N. meningitidis (e value 0.006). Cytidine deaminases hydrolyze cytidine into uridine. Unfortunately, this alignment does not appear to be informative because it is questionable whether the N. meningitidis protein is in fact a cytidine deaminase (as annotated). TIGR calls this gene “cytidine deaminase”, but TIGR’s “identification alignment” is with a hypothetical ORF of unknown function in B. subtilis, and warns that it is only an automatic annotation. The N. mening ORF does align with a portion of the zinc binding domain of the cytidine deaminase family of proteins, but this is not the same region that aligns with Sxy. Unfortunately, there is nothing published about the N. meningitidis ORF.