Type Ill protein secretion systems (TTS systems), the subject of several recent reviews (Hueck, 1998; Galan and Collmer, 1999; Cheng and Schneewind, 2000; Staskawitz et al., 2000; Cornelius and Gijsegem, 2000; Plano et al., 2001) are found in many animal and plant pathogens and in at least one insect endosymbiont (Nguyen et al., 2000; Saltiel et al., 2001). With one exception, they are all proteobacteria, and that exception is Chlamydiaceae. Bacteria with TTS systems are usually host - cell associated at some time in their life histories, and the components of TTS systems are often virulence factors. In many of these systems, contact of the pathogen with host cells prompts delivery of bacterial proteins into the host-cell cytoplasm or cell membrane where they disrupt cellular functions to the benefit of the pathogen. Energy needed for delivery comes from ATP.  Proteins of the secretion mechanism are conserved, but the secreted proteins tend to be unique to the bacterium secreting them. In proteobacteria, the TTS genes are clustered in the bacterial chromosome and have GC contents at variance with the rest of the genome, which suggests that they have been horizontally transferred. In the enterobacteria, the apparatus for secreting proteins by the TTS system is arranged in a supramolecular structure, the needle complex, which spans the inner and outer bacterial membranes, contains TTS proteins, and is thought to be the instrument of bacterial protein translocation into the host cell (Kubori et al., 1998. Blocker et al., 2001). Type Ill genes are homologous to those of the flagellar export apparatus from which they probably evolved.

Chlamydial TTS genes were first isolated from C. cavaiae by Hsiu and associates (1997) and soon thereafter identified in all completed Chlamydiaceae sequences [see: Genome comparison for references]. C. trachomatis, C. muridarum, and C. pneumoniae have complete sets of TTS genes. They also have structures resembling the needle complexes of Shigella and Salmonella that were described years before the existence of Type Ill systems was even suspected (Matsumoto et al., 1976) [see: EB structure; development cycle in pictures]. On the surfaces of both EBs and RBs are patches of regularly spaced protuberances, each with a projecting central spike (Matsumoto, 1988; Chang et al., 1997). They appear to anchor RBs to the inner surface of the inclusion membrane [see: RB structure]. It has been proposed that the hollow spikes provide conduits for passage of nutrients from the host cytoplasm into RBs, but now it appears more likely that the passage consists of Type Ill secretion from the RBs into the inclusion membrane and possibly the host cytoplasm. Recent investigtions have provided evidence that chlamydial proteins are indeed secreted into the inclusion membrane by a TTS mechanism (Fields and Hackstadt, 2000; Subtil et al., 2001); [see also: inclusion proteins]. Chlamydial TTS systems have been discussed by Stephens (1999); Hatch (1999); Hackstadt (1999); Bavoil et al., (2000) and also elsewhere on this site.

In the phylogenetic distribution of TTS genes, Chlamydiaceae is the odd one out; all others are proteobacteria. It would have been more fitting for rickettsiae to have TTS systems, not chlamydiae!   If Chlamydiaceae got its Type Ill genes from a member of the proteobacteria, then the difficulties of effective contact between extant species living in modern hosts, as just described for possible horizontal transfer of ATP importer genes, would also hold true here. If Chlamydiaceae got its TTS genes by horizontal transfer, I would guess that it happened a long time ago when the Chlamydiales lineage was in unicellular hosts, and that the LCA of the four Chlamydiales familes had a TTS system.

When the genome sequence of C. trachomatis serovar D was completed, Stephens et al., (1998) noted that its Type Ill genes were scattered about the genome in several groups and that their GC content was the same as that of the genome as a whole. They concluded that the TTS system of C. trachomatis must be of ancient origin. Hueck (1998) and Kim (2001) went a step further and suggested that TTS systems actually arose in the Chlamydiales lineage. Kim has provided a detailed genetic argument for his view. Being partial to chlamydiae, I find this an attractive idea, but I have been concerned about the different ways in which proteobacterial and chlamydial TTS systems operate. Proteobacteria use them to inject proteins into host cells from the outside, but chlamydiae use their TTS systems to inject proteins into host cells from the inside (that is, from inside the inclusion). The recent discovery that the proteome of EBs, the infecting stage of the developmental cycle, contains several proteins of the TTS system (Vandahl et al., 2001) offers a way to ease my concern. Since EBs also have their own pool of ATP to supply the needed energy (Tipples and McClarty, 1993) TTS proteins could play a role in chlamydial attachment and entry. There is precedent for such a role in other bacteria. Shigella and Salmonella use TTS systems for invasion of nonphagocytic cells (Galan and Collmer, 1998) and Type III genes are essential for entry of the tsetse fly endosymbiont Sodalis glossinidius into host cells (Dale et al., 2001; Moran, 2001). All this leads to a ‘Just - So’ story of "How chlamydiae got their TTS genes". The extracellular progenitor of Chlamydiales was flagellated, and when Chlamydiales began to live in eukaryotic hosts, there was selection for individuals more efficient at getting into those cells. Thus the primitive TTS system of Chlamydiales was one that facilitated entry by an outside - to - inside injection route. Then after protracted residence in the intracellular habitat, selection for better ways to exploit the advantages of that habitat and to ameliorate its disadvantages resulted in the appearance of the unique ability of inclusion - dwelling Chlamydiales to secrete proteins into the inclusion membrane and surrounding cytoplasm. This idea obviously conflicts with my earlier suggestion that the Chlamydiales lineage was symbiotic before it was parasitic [see: endosymbiosis].

The evolutionary advantages accruing to chlamydiae by virtue of their possession of a TTS system should become obvious when the nature and function of the secreted proteins is known. Many chlamydial activities will probably turn out to be dependent on the TTS system, such as the secretion of proteins into the inclusion membrane (Fields and Hackstadt, 2000; Subtil et al., 2001) and secretion into the host cytoplasm of proteins that modify host vesicular traffic (Hatch, 1999; Hackstadt, 1999), that hasten (Ojcius et al., 1998) or retard (Fan et al., 1998) the onset of apoptosis, and that degrade host transcription factors (Zhang et al., 2000; 2001).
[JWM]