Two new reports of bacterial genome sequences provide revealing insights into the evolution of virulence mechanisms in an important group of human pathogens. Interestingly, the new work also demonstrates there is substantial heterogeneity among the genomes of this group, even among those of the same serotype that provoke a similar immune response in humans.

Streptococcus agalactiae, also known as group B streptococcus (GBS), is responsible for invasive, and sometimes life-threatening, infections in humans, particularly newborn babies. Neonatal infections of these bacteria have decreased since 1996 after the Centers for Disease Control issued guidelines recommending antibiotic treatment before birth for babies at high risk (Centers for Disease Control, 2002). However, GBS still remains a leading cause of newborn sepsis and meningitis, as well as of severe invasive diseases in adults.

GBS is traditionally classified into serotypes according to the degree to which its polysaccharide capsule stimulates an immune response (ie its antigenicity). Nine serotypes have been identified to date.

Now complete genome sequences for representatives of two of these serotypes, III and V, have been reported in September issues of, respectively, Molecular Microbiology and Proceedings of the National Academy of Science. Philippe Glaser and his co-authors report the sequence of the circular chromosome of serotype III GBS, which is highly associated with early onset disease in the newborn (Glaser et al, 2002). Hervé Tettelin and colleagues have sequenced serotype V GBS, associated with severe invasive diseases of nonpregnant adults (Tettelin et al, 2002).

In the PNAS study, Tettelin and colleagues used DNA microarrays to do comparative genomic hybridization between the serotype V strain that was sequenced and 19 other strains of various serotypes. Interestingly, at least 18% of the GBS-specific genes were either absent or divergent from these 19 other strains, including those of the same serotype. It seems there is an enormous amount of genetic heterogeneity within these bacterial serotypes.

One of the most intriguing findings of the two reports was the discovery that most genes apparently unique to specific strains of the same serotype were clustered together in regions (islands). These islands consisted of 7 to 81 kb and encoded at least five contiguous genes. Of the 15 islands identified in the serotype V genome, 10 contained atypical nucleotide compositions differing from the 35.7% G + C content of the entire genome. One possibility is that these divergent islands correspond to horizontal gene transfer events.

The serotype III strain genome had 14 islands that contained the majority of known or putative GBS virulence factors. Importantly, all of these islands also contained sequences known to be associated with mobile genetic elements, eg insertion sequences, proteins of phages, plasmids and transposons. However, the role that these mobile genetic elements had played in the acquisition and spread of these islands remains uncertain because of substantial gene rearrangements among the islands. Such evidence of genetic deterioration may indicate that at least some of these elements may have been present in GBS for a long period of the organism’s evolutionary history. Regardless, the strong association of genetic transfer elements and virulence factors in these chromosomal islands suggests that they may be pathogenicity islands and thus have an important role in virulence acquisition and genetic diversity.

Although a large number of phage and plasmid-related genes were found in the chromosome of the serotype III strain, no complete temperate phage genomes were identified. Interestingly, three copies of a 47 kb sequence were present that had the characteristics of an integrative plasmid. These elements may also be important for horizontal gene transfer in GBS.

Tettelin and colleagues estimated that 650 proteins encoded by the serotype V genome were surface exposed and therefore were potential virulence factors. They then used a proteomic approach to identify which of these proteins was expressed.

Of 291 recombinant proteins made from this group and used to immunize mice, 139 sera were found to recognize a major GBS protein, 55 of which were shown by fluorescence-activated cell sorter analysis to be expressed on the cell surface. These data are important for understanding the physiology of GBS, as well as identifying new candidate antigens for vaccine development.

A remarkably large number of two-component regulatory systems were found in both new GBS genomes, compared with other bacteria in the Gram-positive group sequenced to date. The type III genome contained 20 histidine kinases and 21 response regulators and the type V genome had 17 of each. By comparison 14 are found in S. pneumoniae, 13 in S. pyogenes, and only eight in Lactococcus lactis.

This finding together with discovery of 107 transcriptional regulators suggests that the GBS have greater flexibility than the other Gram-positive cocci to react and survive in response to fluctuations in the external environment. Such versatility may explain why GBS are able to infect animals such as cattle in addition to humans.

Comparative genome analyses revealed that S. agalactiae is more closely related to S. pyogenes (group A streptococci, GAS) than S. pneumoniae. S. agalactiae also shares more orthologous genes with S. pyogenes than S. pneumoniae. S. galactiae or S. pyogenes also differ from S. pneumoniae in that they are not generally known for their ability to uptake exogenous DNA (transformation). In contrast, S. pneumoniae has a highly developed transformation system.

These new genomes allow important insights into the sequence heterogeneity found among and within GBS serotype strains. Future drug development and identification of candidate antigens for a universal vaccine will critically depend on these data. Ultimately these studies, and similar future ones, will lead to improved prevention and treatment of GBS diseases.