RecA and glnA sequences separate the bacteroides fragilis population into two genetic divisions associated with the antibiotic resistance genotypes cepA and cfiA.

The sequences of part of the glutamine synthetase-encoding gene (glnA) and of the RecA-encoding gene (recA) were determined and aligned for 45 Bacteroides fragilis isolates from different clinical and geographical origin. The patterns of sequence divergence of glnA and recA were very similar. The sequences of a 303-bp fraction of recA showed 45 nucleotide substitutions, 40 of which allowed the separation of B. fragilis into two major divisions, which were not found when the deduced amino acid sequences were considered. The 687-bp sequences analysed for the glnA gene showed 112 nucleotide substitutions, 96 of which separated the population into the same two divisions as those described for recA. In this case, the deduced amino acid sequences showed this subdivision as well: three of the six observed amino acid substitutions were division-specific. Within the two divisions, both genes presented a high degree of sequence conservation. Each B. fragilis division was associated with the presence of a different antibiotic resistance gene: cepA encoding a serine-beta-lactamase (division I) and cfiA encoding a metallo-beta-lactamase (division II). No particular clusters associated with geographical or clinical origin, or with the production of an enterotoxin were observed. Sequencing of the cfiA gene allowed identification of two different alleles in division II. However, no association of these different cfiA alleles with the expression of imipenem resistance was observed. In conclusion, the phylogenetic patterns observed by sequencing recA and glnA are in agreement with those obtained previously by MLEE (multilocus enzyme electrophoresis). Thus, it appears that the evolution of recA and glnA genes is similar to that of the whole chromosome of B. fragilis. Horizontal gene transfer between divisions I and II seems to be low, at best. However, the results of the present study could not clarify definitively whether divisions I and II should be considered as two different B. fragilis genospecies.