In shifting from extracellular to intracellular existence, the Chlamydiales ancestor had three hurdles to clear. It had to get inside a compatible host cell, it had to survive long enough to produce progeny, and it had to get some of those progeny into new host cells. Over deep geological time, the opportunities for establishing endocytobiotic relationships must have been almost infinite. They were established in a sea of chance internalizations, and this is probably how the Chlamydiales ancestor got inside its first host cell. I have always thought of the inside of a host cell as an extreme environment colonized only by organisms with long-evolved fitness for that habitat (Moulder, 1974). However, intracellular habitats are ephemeral, host cells die, and the hypothetical founder of the Chlamydiales lineage had no time for long evolutionary adaptation. It had to quickly produce progeny that reached new host cells. The solution to this dilemma may be preadaptation. The extracellular Chlamydiales ancestor must have already been a complex organism with a gene repertoire that enabled it to survive in different habitats and under different stresses. It may already have been capable of minimal propagation in a host cell, particularly when that host cell was a single-celled eukaryote without the efficient combinatorial immune mechanisms of vertebrates. Compared to the highly evolved extant representatives of the order, it probably did a poor job of it. The first intracellular Chlamydiales was not necessarily obligately dependent on host cells for its existence. All endocytobiotic lineages were probably started by facultative progenitors that became obligate only when a vital biosynthetic function was lost [see: Part II. Genome degradation].

The first association between the hypothetical ancestor and a eukaryotic cell could have been either intracellular parasitism or endosymbiosis, living intracellularly with (horizontal transmission) or without (vertical transmission) an extracellular phase intervening between successive host cells. My reason for guessing that it was endosymbiotic is that in the world outside of medical and veterinary microbiology, endosymbionts are much more frequently encountered than are intracellular parasites. It may be that the probability of an endosymbiont being transferred to a daughter cell is greater than the probability of an intracellular parasite leaving its host cell and finding a new one somewhere out there in the extracellular jungle. This is a heretical view. Others have come to the opposite conclusion, that endosymbionts are descended from intracellular parasites (Corsaro et al, 1999; Moran, 2001). In the Parachlamydiaceae, some organisms are vertically transmitted endosymbionts, others are horizontally transmitted intracellular parasites, and some follow either way of life depending on the circumstances (Fritsche et al, 2000). All characterized members of the other families are intracellular parasites. What selective forces favored emergence of mechanisms for horizontal transmission of the hypothetical ancestral Chlamydiales endosymbiont? This was a major evolutionary event, one that determined the lifestyle of present-day Chlamydiales. The elaborate and efficient way of getting from one host cell to another employed by Chlamydiaceae suggests that the shift to intracellular parasitism had a long and complicated evolutionary history. Elucidation of transmission mechanisms in the other families may throw light on that history.
[JWM]