Competence regulons, genomics and streptococci

Sir, Recently, Campbell et al. (1998, Mol Microbiol27: 929–939) reported a very elegant study combining a genetic screen with genomic analysis, which led to the identification of six loci belonging to the competence regulon of Streptococcus pneumoniae. Starting from the first identified competence-induced (cin) operon, initially described as containing two (exp10-recA ; Pearce et al., 1995, J Bacteriol177: 86–93) or three genes (cinA-recA-dinF ; Martin et al., 1995, Mol Microbiol15: 367–379) and recently shown to also contain the lytA gene (Mortier-Barrière et al., 1998, Mol Microbiol27: 159–170) (Fig. 1), the authors used a exp10 ::lacZ transcriptional fusion to establish a genetic screen for cin promoters on solid media. Then, out of 1029 random lacZ chromosomal insertion mutants screened, Campbell et al. (1998, ibid.) retained one mutant, TYG, that showed a fivefold induction during competence in liquid medium. The tyg transcription start site was mapped. In parallel, a 52 bp (−38 to +14) minimal cin promoter sequence was defined by deletion analysis of the exp10 upstream region. The −14 to −7 region of tyg was found to be identical to the corresponding region of exp10 (Fig. 1). The authors hypothesized that this sequence (TACGAATA), referred to hereafter as the cin box, was critical for cin expression. A search of the pneumococcal genome uncovered six contigs containing the cin box. Further analysis revealed that five contained loci essential for genetic transformation. These loci encoded proteins with strong similarity to proteins involved in DNA processing during transformation in other bacteria (Campbell et al., 1998, ibid.). Despite its key role in this study, the TYG mutant was apparently not characterized further. . Location of the tyg promoter (pr-tyg ) within the cinA–recA operon (bottom) and alignment of regions containing putative or proven cin promoter sequences (top). The cin-box identified by Campbell et al. (1998, ibid.) as well as the transcription start of the cinA-recA operon (Pearce et al., 1995, ibid.) and that of tyg are boxed. Sequences corresponding to 17 putative competence loci (nine from S. pneumoniae, i.e. those identified by Campbell et al. (1998, ibid.), including tyg and cinA [cinA], plus comF and addAB ; seven from S. pyogenes, namely homologues of S. pneumoniae cinA [Spy-cinA], dprA and ssb, and homologues of the four B. subtilis late com operons; and comY from S. gordonii [Sgo-comY]) were aligned using Multalin (Corpet, 1988, Nucleic Acids Res 16: 10881–10890). The consensus sequence (top line) was based on the frequency of a nucleotide in each position being greater than 60% (lower case) or 75% (capital); invariant positions are underlined. The translation start codon of the S. pneumoniae and S. pyogenes cinA genes is underlined. We were curious to see whether the tyg promoter controlled a genetic transformation locus and whether other specific transformation loci (e.g. the homologue of the B. subtilis comF operon) might be present in this regulon. We first searched the TIGR S. pneumoniae genome database (http://www.ncbi.nlm.nih.gov/BLAST/tigrbl.html) with the published tyg promoter sequence. Much to our surprise, the tyg promoter was found to be located within the previously identified dinF gene and oriented so as to produce an antisense RNA (Fig. 1). As this raised the question of the function of this promoter, we searched for potential orf(s) expressed under its control. This search did not reveal any obvious candidate. Even the longest intergenic region (306 bp) between recA and dinF was unlikely to encode a protein because it was found to contain two copies of a repeated 108-bp-long element that we tentatively called rup (for repeated unit of pneumococcus; J. P. Claverys and B. Martin unpublished). An alternative possibility would be that a tyg promoter-driven antisense RNA plays a role in the processing of cinA–recA operon transcripts. There are several examples of RNA–RNA interactions governing mRNA stability (e.g. the hok/sok maintenance system of plasmid R1; Gerdes et al., 1997, Annu Rev Genet31: 1–31). The putative antisense RNA may encompass the rup elements. It could therefore have a more general effect on mRNA stability by affecting the processing of any rup-containing transcript. These possibilities deserve examination. It should, however, be pointed out that, among cin promoters, the lowest β-Gal activity was obtained with the tyg promoter fusion (Campbell et al., 1998, ibid.). We noticed that the tyg promoter was atypical in having a consensus −10 promoter sequence (TATAAT) partially overlapping the cin box (Fig. 1). Moreover, the tyg promoter was also atypical with respect to a highly conserved T residue located seven nucleotides downstream of the cin box (Fig. 1). Interestingly, this position corresponds to the transcription start site of cinA and tyg (Fig. 1). The tyg promoter is unique among 17 putative cin promoters (see below and legend to Fig. 1) in having an A instead of a T at this position (13 cases) or a G (three cases) (data not shown). Altogether, these observations raise the question of the biological relevance of the tyg promoter. A search of the pneumococcal genome with the cin box then revealed two new candidate late com genes. A cin box was found 7 bp upstream of a putative operon encoding homologues of the ComF proteins of B. subtilis that are implicated in DNA processing (Dubnau, 1997, Gene192: 191–198). A cin box was also found 430 bp upstream of a putative operon encoding homologues of the B. subtilis AddAB proteins. The addAB genes have been shown to be induced at competence (Haijema et al., 1995, Mol Microbiol15: 203–211) and the corresponding proteins are required for recombination of transforming DNA in B. subtilis (Haijema et al., 1996, Mol Microbiol21: 989–999). That these promoters are induced during competence in S. pneumoniae remains to be established, particularly for the addAB locus, because the distance between the putative cin box and the first orf is atypical (see below). In addition, we observed that the addAB putative cin box is located within a variant copy of the rup element, unique in the genome of S. pneumoniae (i.e. other copies lack the cin box). It may correspond to the sixth cin box for which no induction during competence could be detected (Campbell et al., 1998, ibid.). We then decided to look for the cin box in another transformable streptococcal species, Streptococcus gordonii. A cin box was found 59 bp upstream of the comY operon (GenBank accession number U81957) (Fig. 1). This locus encodes homologues of the B. subtilis ComGA-D proteins, and comYA is essential for transformation of S. gordonii (Lunsford and Roble, 1997, J Bacteriol179: 3122–3126). Finally, we extended our search to the Streptococcus pyogenes genome database (http://www.ncbi.nlm. nih.gov/BLAST/ouacgtbl.html). A CinA orthologue (67.4% and 62.7% identity at the protein and DNA level respectively) was found in this bacterium. We observed that not only is the cinA-recA gene arrangement conserved in S. pyogenes, but that sequences upstream of cinA are strikingly similar to those in S. pneumoniae. The cin box is entirely conserved and adjacent nucleotides are identical in both species (Fig. 1). cin boxes were also detected in front of S. pyogenes open reading frames encoding homologues of the B. subtilis comCEFG genes (L. S. Håvarstein, personal communication). Distances between the putative or proven cin boxes and translation starts vary from 7 to 22 bp, except for the S. pneumoniae addAB and for the S. gordonii comY loci. Interestingly, these distances appear highly conserved for orthologous pairs (see for example the cinA pair; Fig. 1). Altogether, these observations suggest that expression of these putative operons in S. pneumoniae, S. gordonii and S. pyogenes could be under the control of a transcriptional regulator conserved during evolution. They open the intriguing possibility that although S. pyogenes has not been reported to be transformable, it possesses a competence regulon, and could be transformed provided the appropriate environmental stimuli leading to activation of this putative transcriptional regulator are applied.