Rethinking plastid evolution

How easy is it to acquire an organelle? How easy is it to lose one? Michael Gray considers the latest evidence in this regard concerning the chromalveolates.How easy is it to acquire an organelle? How easy is it to lose one? These questions underpin the current debate about the evolution of the plastid—that is, chloroplast—the organelle of photosynthesis in eukaryotic cells.The origin of the plastid has been traced to an endosymbiosis between a eukaryotic host cell and a cyanobacterial symbiont, the latter gradually ceding genetic control to the former through endosymbiotic gene transfer (EGT). The resulting organelle now relies for its biogenesis and function on the expression of a small set of genes retained in the shrunken plastid genome, as well as a much larger set of transferred nuclear genes encoding proteins synthesized in the cytosol and imported into the organelle.This scenario accounts for the so-called primary plastids in green algae and their land plant relatives, in red algae and in glaucophytes, which together comprise Plantae (or Archaeplastida)—one of five or six recognized eukaryotic supergroups (Adl et al, 2005). In other algal types, plastids are ‘second-hand''—they have been acquired not by taking up a cyanobacterium, but by taking up a primary-plastid-containing eukaryote (sometimes a green alga, sometimes a red alga) to produce secondary plastids. In most of these cases, all that remains of the eukaryotic symbiont is its plastid; the genes coding for plastid proteins have moved from the endosymbiont to the host nucleus. A eukaryotic host—which may or may not itself have a plastid—might also take up a secondary-plastid symbiont (generating tertiary plastids), or a secondary-plastid host might take up a primary-plastid symbiont. You get the picture: plastid evolution is complicated!Several excellent recent reviews present expanded accounts of plastid evolution (Reyes-Prieto et al, 2007; Gould et al, 2008; Archibald, 2009; Keeling, 2009). Here, I focus on one particular aspect of plastid evolutionary theory, the ‘chromalveolate hypothesis'', proposed in 1999 by Tom Cavalier-Smith (1999).The chromalveolate hypothesis seeks to explain the origin of chlorophyll c-containing plastids in several eukaryotic groups, notably cryptophytes, alveolates (ciliates, dinoflagellates and apicomplexans), stramenopiles (heterokonts) and haptophytes—together dubbed the ‘chromalveolates''. The plastid-containing members of this assemblage are mainly eukaryotic algae with secondary plastids that were acquired through endosymbiosis with a red alga. The question is: how many times did such an endosymbiosis occur within the chromalveolate grouping?A basic tenet of the chromalveolate hypothesis is that the evolutionary conversion of an endosymbiont to an organelle should be an exceedingly rare event, and a hard task for a biological system to accomplish, because the organism has to ‘learn'' how to target a large number of nucleus-encoded proteins—the genes of many of which were acquired by EGT—back into the organelle. Our current understanding of this targeting process is detailed in the reviews cited earlier. Suffice it to say that the evolutionary requirements appear numerous and complex—sufficiently so that the chromalveolate hypothesis posits that secondary endosymbiosis involving a red alga happened only once, in a common ancestor of the various groups comprising the chromalveolates.Considerable molecular and phylogenetic data have been marshalled over the past decade in support of the chromalveolate hypothesis; however, no single data set specifically unites all chromalveolates, even though there is compelling evidence for various subgroup relationships (Keeling, 2009). Moreover, within the proposed chromalveolate assemblage, plastid-containing lineages are interspersed with plastid-lacking ones—for example, ciliates in the alveolates, and oomycetes such as Phytophthora in the stramenopiles. The chromalveolate hypothesis rationalizes such interspersion by assuming that the plastid was lost at some point during the evolution of the aplastidic lineages. The discovery in such aplastidic lineages of genes of putatively red algal origin, and in some cases suggestive evidence of a non-photosynthetic plastid remnant, would seem to be consistent with this thesis, although these instances are still few and far between.In this context, two recent papers are notable in that the authors seek to falsify, through rigorous testing, several explicit predictions of the chromalveolate hypothesis—and in both cases they succeed in doing so. Because molecular phylogenies have failed to either robustly support or robustly disprove the chromalveolate hypothesis, Baurain et al (2010) devised a phylogenomic falsification of the chromalveolate hypothesis that does not depend on full resolution of the eukaryotic tree. They argued that if the chlorophyll c-containing chromalveolate lineages all derive from a single red algal ancestor, then similar amounts of sequence from the three compartments should allow them to recover chromalveolate monophyly in all cases. The statistical support levels in their analysis refuted this prediction, leading them to “reject the chromalveolate hypothesis as falsified in favour of more complex evolutionary scenarios involving multiple higher order eukaryote–eukaryote endosymbioses”.In another study, Stiller et al (2009) applied statistical tests to several a priori assumptions relating to the finding of genes of supposed algal origin in the aplastidic chromalveolate taxon Phytophthora. These authors determined that the signal from these genes “is inconsistent with the chromalveolate hypothesis, and better explained by alternative models of sequence and genome evolution”.So, is the chromalveolate hypothesis dead? These new studies are certainly the most serious challenge yet. Additional data, including genome sequences of poorly characterized chromalveolate lineages, will no doubt augment comparative phylogenomic studies aimed at evaluating the chromalveolate hypothesis—which these days is looking decidedly shaky.