At this moment, complete genome sequences have been published for Chlamydia trachomatis serovar D (Stephens et al., 1998), Chlamydia muridarum (Read et al., 2000), and 3 strains of Chlamydophila pneumoniae (Kalman et al., 1999; Read et al., 2000; Shirai et al., 2000). Sequencing of Chlamydophila caviae, Chlamydophila felis, Chlamydophila abortus and Parachlamydia UWE 25 is also completed but not yet [Jan 2003] published. For description, analysis, comparison, and interpretation of the results, consult the original publications as well as reviews by Stephens (1999b), Tanner et al., (1999), and Rockey et al., (2000). Here I will concentrate on how far knowledge of the initial five complete chlamydial genomes goes in answering some important evolutionary questions. What was the genome of the LCA of Chlamydiaceae and what evolutionary forces drove the genetic changes responsible for radiation from this LCA into extant genera and species? How do the genomes of present species differ from each other and how do these differences account for the phenotypic individuality of each species? That only incomplete answers will be forthcoming for these questions does not diminish their importance.

The distribution of phenotypic characters in a group of related organisms has long been used to establish evolutionary relationships. Now genes are used for the same purpose. Similar genes suggest common ancestry, while genes that differ in otherwise related populations suggest divergence. Comparison of whole genomes is a powerful tool for establishing evolutionary relationships, but its application is often limited by the number of available genomes. The published genome sequences for Chlamydiaceae spp. have much in common, indicating relatively limited radiation from a common ancestor. However embedded in this matrix of similarity are enough differences to explain the individuality of each chlamydial species - if we only knew how.

C. trachomatis is a widely disseminated human pathogen that is transmitted sexually to produce infections of mucosal cells (Schachter, 1999). In nature, C. muridarum latently infects mice, but it causes acute pulmonary infection after serial passage of apparently normal mouse lung (Hilleman, 1945). It is often used as an experimental surrogate for C. trachomatis. Despite differences in natural host and mode of transmission, the two genomes are remarkably alike in gene content and order. C. pneumoniae is, like C. trachomatis, a pathogen of human mucosal cells but, unlike C. trachomatis, it is transmitted by the respiratory route to cause a variety of respiratory ailments (Schachter, 1999). It also colonizes the vascular system where it is suspected of playing a role in development of atherosclerosis [see: C. pneumoniae index]. The almost identical genomes of the different C. pneumoniae strains closely resemble the genomes of C. trachomatis and C. muridarum. One cannot be sure that the remarkable similarities among the genomes of the 3 species so far sequenced will also be found among the 6 unsequenced species, but there is a mass of phenotypic and genotypic evidence suggesting that it will. For the purposes of further discussion, I am going to assume that the LCA of the family Chlamydiaceae was not too different from the chlamydiae we have with us today.

The genomes of C. trachomatis serovar D and C. muridarum are very close to the same size, 1.07 and 1.04 X 10⁶ base pairs respectively [see: Chlamydial genome]. C. trachomatis has genes for tryptophan biosynthesis and C. muridarum does not. Both of these species have genes that express cytotoxins related to the large clostridial toxin (Belland et al., 2001). They will be discussed elsewhere [see: Genome degradation]. The genes for tryptophan biosynthesis and the cytotoxin, as well as several other genes that differ in the 2 species, are located in a single region, the plasticity zone. It has been suggested that only a handful of genes may determine the pathogenic signature of each species.

The three C. pneumoniae genomes known have a mean genome size of 1.23 X 10⁶ base pairs; about 200,000 more than either C. trachomatis or C. muridarum. By pulsed field electrophoresis (PFGE), the other Chlamydophila spp. have genomes larger than those of C. trachomatis or C. muridarum and about the same size as C. pneumoniae (Everett et al, 1999; Bush and Everett, 2001) C. pneumoniae has about 200 more protein-coding genes than C. trachomatis. About 85% of these extra genes code for unknown functions, a much higher proportion than in the genome as a whole. Conversely, C. trachomatis has about 70 genes not found in C. pneumoniae, and again 85% of these distinct genes have no predicted function. These differences could have developed during divergence from a common ancestor, either by loss of genes or acquisition of new ones. The genes unique to either genome are good candidates for determinants of species individuality, including their unique disease-producing potential. The significance of the high proportion of genes of unknown function in the unique fraction of each genome remains to be determined. Of the functionally identifiable unique genes in C. pneumoniae, many are accounted for by expansion of the pmp family (Rockey et al., 2000) from 9 genes in C. trachomatis to 21 in C. pneumoniae. Other extra genes code for enzymes of biotin synthesis, phospholipase-D-like proteins, and enzymes of purine and pyrimidine salvage. As in the other genomes, the unique genes tend to lie in a plasticity zone. Like C. muridarum, C. pneumoniae lacks genes for biosynthesis of tryptophan. The cytotoxin gene is absent. All the genomes have inc genes (Rockey et al., 2000) but in different assortment. Roughly half of the inc genes in one species have no counterpart in the others. The pmp and Inc proteins are especially good prospects for determinants of species individuality. Both have potential for nearly infinite reassortment, and both are surface proteins.

One can hope that in the not too distant future, complete genome sequences will be available, not only for all the Chlamydiaceae species, but also for representatives of the other 3 families¹ of the order Chlamydiales. If only genomes could somehow be retrieved from the ancient past! Think of how instructive it would be to have the genome of the LCA of Chlamydiaceae or the genome of the LCA of all four Chlamydiales families.

[JWM]

[ ¹ MEW comment: Details of the Parachlamydia sequencing are posted on the website of the environmental chlamydiae genome project, at the Technical University, Munich. Click on the following hyperlinks for the Chlamydophila abortus sequence [Sanger Institute] and for information on the Chlamydophila felis sequence [Yamaguchi University].