Illuminating the Evolutionary History of Chlamydiae

See allHide authors and affiliations

Science  30 Apr 2004:
Vol. 304, Issue 5671, pp. 728-730
DOI: 10.1126/science.1096330


Chlamydiae are the major cause of preventable blindness and sexually transmitted disease. Genome analysis of a chlamydia-related symbiont of free-living amoebae revealed that it is twice as large as any of the pathogenic chlamydiae and had few signs of recent lateral gene acquisition. We showed that about 700 million years ago the last common ancestor of pathogenic and symbiotic chlamydiae was already adapted to intracellular survival in early eukaryotes and contained many virulence factors found in modern pathogenic chlamydiae, including a type III secretion system. Ancient chlamydiae appear to be the originators of mechanisms for the exploitation of eukaryotic cells.

Chlamydiae have long been considered a unique coherent group of prokaryotes comprising a few closely related pathogenic species (1), which formed a deep branch in rRNA-based phylogenetic trees (2, 3), indicating their ancient divergence about 2 billion years ago (Fig. 1 and fig. S1). The recent discovery of chlamydia-related endosymbionts in free-living amoebae (46) and the existence of an untold diversity of chlamydiae in the environment (79) was thus surprising and markedly changed our perception of chlamydial distribution and host range. Environmental chlamydiae are ubiquitous, share the unique biphasic chlamydial developmental cycle, and represent an evolutionary early-diverging sister group (about 700 million years ago) of present-day pathogenic chlamydiae (Fig. 1 and fig. S1). All attempts to cultivate chlamydiae outside their host cells or to establish a genetic system for chlamydiae have failed so far. Here, we report on comparative and phylogenetic genome analysis of a symbiotic, previously uncharacterized relative of pathogenic chlamydiae.

Fig. 1.

Phylogeny of chlamydiae. 16S rRNA-based neighbor-joining tree showing the affiliation of environmental and pathogenic chlamydiae with major bacterial phyla. Arrow, to outgroup. Scale bar, 10% estimated evolutionary distance.

The genome of the chlamydia strain analyzed in this study (10), the Acanthamoeba sp. endosymbiont UWE25 (5), is about twice as large as the genomes of any of the pathogenic chlamydia species investigated to date (1115) (Table 1). Only a few pseudogenes (n = 26) and gene remnants, mostly affecting transposase genes, were detected. This indicates that the genome of UWE25 is not in the process of becoming smaller but has stabilized at 2.4 Mb. Of the 2031 predicted coding sequences (CDSs) identified in the UWE25 genome, 938 showed significant homology to CDSs found in other chlamydial genomes. In total, 711 CDSs are shared among all chlamydial genomes, representing the core gene set of chlamydiae (fig. S2). However, pairwise comparison of genome structure based on FastA reciprocal best matches shows little conservation of gene order between UWE25 and pathogenic chlamydiae, indicating massive reorganization that probably occurred after the divergence of environmental and pathogenic chlamydiae and genome size reduction that took place during the evolution of pathogenic chlamydiae (fig. S3). This finding extends previous observations of genome evolution in other bacterial symbionts and pathogens (16). Pathogenic chlamydiae lost more genes because they thrive in a more homeostatic niche than UWE25, which is exposed to fluctuating environmental conditions in its amoebal host. This difference is also illustrated by the higher number of rRNA operons in UWE25 than in pathogenic chlamydiae (Table 1) (17). The 1093 CDSs present in the UWE25 genome but absent from all other chlamydial genomes are almost evenly distributed around the UWE25 genome, indicating that they were not acquired by only a few lateral gene-transfer events (fig. S4 and table S1). Previous analyses have shown that pathogenic chlamydiae contain an unexpectedly high number of cyanobacterial and plant gene homologs (18). These numbers are even higher in UWE25 (fig. S5 and table S2). Furthermore, UWE25 shares the unique ribosomal superoperon structure (18) with pathogenic chlamydiae, cyanobacteria, and chloroplasts. Taken together, these findings provide further evidence for an ancient relationship of chlamydiae with cyanobacteria. Consistent with this hypothesis, many of the UWE25-homologous plant gene encoded proteins are localized in the chloroplast, indicating a cyanobacterial ancestry (table S2). However, some of the homologous plant proteins are phylogenetically more closely related to chlamydiae than to cyanobacteria, suggesting complex ancestral gene transfers between plants and these bacterial groups (fig. S6). Although a notable number of remnant transposases (n = 82) is present in UWE25 (fig. S4), with one exception, little evidence for recent lateral gene-transfer events was identified by codon usage, G+C content, or phylogenetic analyses. Because of the absence of considerable gene acquisition after the divergence of environmental and pathogenic chlamydiae, the genome sequence of the amoebal symbiont UWE25 offers an insight into the genetic makeup of their last common ancestor.

Table 1.

General features of chlamydial genomes. bp, base pairs; seq., sequence.

Environmental chlamydia UWE25 Chlamydophila caviae GPIC Chlamydophila pneumoniae CWL029 Chlamydia trachomatis sv D Chlamydia muridarum MoPn
Chromosome (bp) 2,414,465View inline 1,173,390 1,230,230 1,042,519 1,072,950
Plasmid (bp) - 7,966 - 7,493 7,501
G + C content (%) 35.8 39.2 40.6 41.3 40.3
Total CDSs 2,031 1,009 1,073 894 921
    With functional assignment 784 (38%) 605 (60%) 636 (59%) 604 (67%) 563 (61%)
    (Conserved) hypothetical 623 (31%) 320 (32%) 251 (23%) 35 (4%) 281 (31%)
    Unknown (no seq. homology) 624 (31%) 84 (8%) 186 (17%) 255 (29%) 77 (8%)
Coding density (%) 83 88 88 90 90
Average gene length (bp) 962 1,030 1,031 1,049 1,064
rRNA operons 3 1 1 2 2
tRNAs 35 38 38 37 37
  • View inline* Sequence contains one gap of about 3 kb (10).

  • As obligate intracellular bacteria, all pathogenic chlamydiae exhibit reduced central metabolic and biosynthetic pathways and are auxotrophic for most amino acids and nucleotides (19). This tendency was also observed in UWE25 [supporting online material (SOM) text]. A notable exception is the tricarboxylic acid (TCA) cycle, which is incomplete in pathogenic chlamydiae (19, 20). In contrast, all TCA cycle genes were found in the UWE25 genome. Phylogenetic analysis suggested that chlamydial TCA cycle genes were not subject to recent lateral gene transfer (fig. S7). Thus, the last common ancestor of environmental and pathogenic chlamydiae most likely still possessed a complete TCA cycle.

    Both UWE25 and pathogenic chlamydiae have a respiratory chain most similar to the respiratory chain of Escherichia coli expressed under microaerophilic conditions (19, 20). UWE25 possesses additional respiratory chain components to generate a H+ gradient, which could be used by its additional F1F0–adenosine triphosphate synthase (F1F0–ATPase) (pc1667 to pc1674; SOM text). The amount of energy gained by oxidative phosphorylation in UWE25 should thus be higher than in pathogenic chlamydiae, and the UWE25 respiratory chain should be more versatile in response to changing environmental conditions. Nevertheless, UWE25, like its pathogenic counterparts, exploits its host's ATP pool by importing ATP from the cytosol in an exchange with adenosine diphosphate (ADP), with an ATP/ADP translocase, as demonstrated by the recent characterization of this protein after heterologous expression in E. coli (21). Such energy-parasite transport proteins have so far only been identified in rickettsiae, chlamydiae, and plant plastids. In contrast to pathogenic chlamydiae, which have two nucleotide transporters (22), UWE25 encodes five nucleotide transporter isoforms (pc0240, pc0241, pc0250, pc0485, and pc1343). Phylogenetic inference suggests that this unique transporter family is an ancient chlamydial trait (21). The presence of these nucleotide transporters might explain the highly reduced nucleotide biosynthesis capabilities of UWE25, which, like pathogenic chlamydiae, lacks many key enzymes of both purine and pyrimidine metabolic pathways. Similarly, UWE25 is incapable of synthesizing most amino acids and thus acquires them from its host by means of 18–amino acid or oligopeptide transporters (SOM text). Taken together, reconstruction of metabolic pathways indicated that the last common ancestor of environmental and pathogenic chlamydiae was already adapted to intracellular life but was less dependent on host metabolites than are modern pathogenic chlamydiae.

    The unusual cell envelope of pathogenic chlamydiae contains no detectable peptidoglycan. Instead, the chlamydial cell wall gains stability from the outer membrane complex, which consists of several proteins cross-linked by disulfide bridges (23). Essential components of the outer membrane complex of pathogenic chlamydiae (OmcA and OmcB), as well as a cysteine bond isomerase, are present in the UWE25 genome (pc0617, pc0616, and pc1600), whereas no homolog of the major outer membrane protein (OmpA) could be identified. In addition, no homologs (or respective motifs) of polymorphic outer membrane proteins or the porin PorB were found. Like pathogenic chlamydiae (19, 20), UWE25 encodes proteins for biosynthesis of a truncated LPS, but lacks the genes necessary for biosynthesis of O-specific polysaccharides. Thus, the cell envelope of UWE25 not only exhibits many characteristics of the pathogenic chlamydia cell envelope but also shows distinct differences in composition, which might indicate niche-specific adaptations of pathogenic chlamydiae to their higher eukaryotic hosts, which are needed for evasion of the hosts' immune responses.

    Pathogenic chlamydiae and many (pathogenic) Proteobacteria possess a type III secretion system (TTSS), which is considered essential for their virulence (20) because it is used as a molecular syringe to deliver effector proteins into eukaryotic cells (24). UWE25 also encodes a complete TTSS. The distribution of TTSS genes among different regions on the chromosome (similar to their localization in pathogenic chlamydiae) and phylogenetic analysis of TTSS key components indicated that the common ancestor of UWE25 and pathogenic chlamydiae already encoded a TTSS, suggesting that the TTSS was invented or acquired more than 700 million years ago in the chlamydial evolutionary lineage as an ancient mechanism for interaction with early eukaryotes (Fig. 2).

    Fig. 2.

    Phylogeny of the TTSS. Unrooted consensus tree based on a concatenated amino acid sequence alignment comprising SctT (periplasm/inner membrane component), SctV (inner membrane proton conducting channel), and SctN (cytoplasmic ATPase). The monophyletic grouping of pathogenic and environmental chlamydiae, which is well separated from all other bacteria, demonstrates that the last common chlamydial ancestor possessed a TTSS. Polytomic branching points indicate those parts of the tree for which different treeing methods produced different topologies. Maximum parsimony bootstrap values (upper numbers) and TREE-PUZZLE support values (lower numbers) are indicated at each node of the tree. Scale bar, 10% estimated evolutionary distance. The conservation of the TTSS operon structure among environmental and pathogenic chlamydiae is shown below the tree (gene lengths of C. trachomatis sv D are shown). The three TTSS operons of pathogenic chlamydiae are represented by different shades of gray. Capital letters refer to sct gene names according to the unified nomenclature suggested by Hueck et al. (24) (except E, which stands for sycE).

    UWE25 might in addition be able to interact with its host by using the type IV secretion system (TFSS) identified in its genome. TFSSs, which are absent in pathogenic chlamydiae, are used by a variety of bacterial symbionts and pathogens for protein export or DNA transfer into eukaryotic cells (25). The absence of TFSS genes necessary for DNA transfer in the UWE25 genome suggests that in this organism TFSS is responsible for secreting effector proteins into the amoebal host and thus fulfills a function similar to that of the TTSS. The arrangement of UWE25 TFSS genes adjacent to each other in a single region on the chromosome, their higher G+C content (41.9%, 37.9% in the third codon position) compared with the genomic G+C content (35.8%, 26.9% in the third codon position), and the presence of several transposases in proximity to TFSS genes indicates that the TFSS was recently acquired by UWE25 from a donor with a genomic G+C content greater than 42% [estimated according to (26)]. The TFSS genes thus represent the only recognized example for a possible recent lateral transfer of genes (coding for proteins with known function) to UWE25 after the split of the chlamydial lineage into its two sister groups.

    A recently recognized virulence factor of pathogenic chlamydiae, the protease-like activity factor (CPAF) (27), is also encoded in the UWE25 genome (pc0916). CPAF is one of the few proteins that have been shown to be secreted (possibly by means of the TTSS) by pathogenic chlamydiae into the host cytoplasm. Its function is still unknown, but CPAF is able to degrade the human transcription factors required for major histocompatibility complex (MHC) expression (27). The presence of CPAF in UWE25, which resides in a host that does not possess a MHC system, indicates that CPAF has evolved from a protease with a different specific function in lower eukaryotic hosts.

    Chlamydiae are among the most successful bacterial pathogens of humans and have recently also been associated with chronic diseases (1, 9). Comparative and phylogenetic genome analysis of a chlamydia-related symbiont of amoebae showed that it has retained several key features of the last common chlamydial ancestor and provided evidence that major virulence mechanisms of present-day pathogenic chlamydiae have evolved from the interaction of ancestral chlamydiae with early eukaryotes. Subsequent adaptation of pathogenic chlamydiae to animal and human host cells was most likely mediated by proteins found in modern pathogenic chlamydiae but not detectable by sequence homology in the UWE25 genome (table S3). However, the presence of key virulence factors in environmental chlamydiae, together with their documented ability to multiply in human macrophages (28), suggests that these protozoan symbionts have the genetic tools to allow them to infect mammalian cells.

    Supporting Online Material

    Materials and Methods

    SOM Text

    Figs. S1 to S9

    Tables S1 to S3

    References and Notes

    References and Notes

    View Abstract

    Navigate This Article