Genomes of multiple independently transitioned obligate symbionts reveal lineage-specific gene loss and unexpected gene retentions.
The organisms with the smallest genomes on Earth all have something in common: they live inside other organisms. Especially small are the genomes of bacteria that cannot live outside a host cell and where the host depends on the bacteria for survival, the so-called obligate endosymbionts. Since obligate endosymbionts are not all related, small genome size has clearly evolved several times and is a result of this particular lifestyle. The process itself is referred to as genome reduction, and involves the loss of hundreds to thousands of genes as a bacterium transitions from a free-living to an obligate endosymbiotic lifestyle1,2. Which genes are thrown out along the way is not random, but rather reflects those functions that are no longer necessary in the specific environment inside a eukaryotic host cell3. Such convergence suggests that the process is predictable, but do endosymbionts follow the same road to genome reduction?
There are two reasons why this question is hard to answer. First, the majority of obligate endosymbiont lineages are ancient, making it hard to see the starting point of the journey. Second, most of them transitioned only once, leaving no steps along the way (Fig. 1). Writing in this issue of Nature Ecology & Evolution, Boscaro et al.4 are able to circumvent these issues when investigating the roads taken by the obligate endosymbiotic bacterium Polynucleobacter associated with ciliates of the genus Euplotes. In this unusual symbiotic system, several closely related Polynucleobacter independently transitioned to an obligate endosymbiotic lifestyle from free-living Polynucleobacter that still exist, thus providing a unique opportunity to increase the resolution of the road to genome reduction.
By building an evolutionary tree using a set of sixteen genomes from seven free-living and nine obligate endosymbiotic Polynucleobacter, Boscaro et al. show that eight out of their nine endosymbionts transitioned independently. As expected, all endosymbiont genomes are small and have lost hundreds of genes. Similar to genomes of other obligate endosymbionts, genes associated with basic cellular functions such as translation and replication are often retained and genes that are no longer needed for survival inside the host, such as those involved in regulation and transport, are commonly lost in Polynucleobacter endosymbionts. But when Boscaro et al. investigated exactly which genes had been lost, they found that it wasn’t the same ones that were lost in the different endosymbionts, suggesting that the road to genome reduction is unpredictable and unique for every symbiont (Fig. 1).
Most similar in content are the endosymbiont genomes with the fewest genes, possibly because they have reached the limit of how many genes can be lost. Interestingly, a few of the genes retained in all symbiont genomes are components of incomplete pathways or protein complexes, indicating that they could be involved in other, as yet unknown functions. Given that genes not necessary for survival are lost, these retained genes might be important for the symbioses between Polynucleobacter and Euplotes.
Faster protein evolution is another common feature of obligate endosymbiont genomes. It is often explained by one of two hypotheses: an increased mutation rate5 or a stronger effect of genetic drift due to the small population sizes of obligate endosymbionts6,7,8. Boscaro et al. find that although the mutation rate is sometimes higher in the endosymbiotic Polynucleobacter compared to free-living Polynucleobacter, the genetic drift effect is consistently higher in endosymbionts, suggesting that genetic drift is a more general cause of faster protein evolution in obligate endosymbionts. Hence, in agreement with previous studies of endosymbiont genomes, both the faster rate of protein evolution and the loss of genes, as seen in the stochasticity of individual gene losses, are largely explained by genetic drift in the genome of endosymbiotic Polynucleobacter.
Since the Polynucleobacter–Euplotes symbiosis is obligate, every independent transition also entails an independent replacement of an already existing Polynucleobacter endosymbiont. Symbiont replacements have been seen in other systems9,10, but an unusual aspect of the system studied by Boscaro et al. is the high frequency of replacement of obligate endosymbionts by close relatives. Two questions that arise are thus why some obligate symbionts, including Polynucleobacter, are more frequently replaced than others, and why some free-living bacteria more easily transition to a symbiotic lifestyle. Is it a character of the bacteria or the hosts, or something particular about the symbiosis itself?
In order to properly investigate these questions it is important to know when a transition is independent. The conclusion that multiple independent transitions have occurred is often drawn by inferring events on an evolutionary tree under the assumption that an endosymbiont can never revert back to a free-living state. This assumption is likely valid in general, since it is hard to regain lost genes, and for the set of genomes investigated by Boscaro et al. in particular, since very few gene gains were identified in the free-living Polynucleobacter. Even so, a more detailed investigation of free-living Polynucleobacter and other bacteria that are nested inside symbiotic clades would be of interest to see if reversion is possible. Perhaps being better at transitioning from free-living to endosymbiont also makes you better at reverting from endosymbiont to free-living.
In conclusion, the results by Boscaro et al. add more details about what happens on the road to genome reduction in obligate symbiotic bacteria. The authors highlight the stochasticity of the steps leading to reduction, but the relatively high predictability of the end result and provide new possibilities for investigating the functional basis of the symbiotic relationship between Polynucleobacter and Euplotes.
McCutcheon, J. P. & Moran, N. A. Nat. Rev. Microbiol. 10, 13–26 (2012).
Toft, C. & Andersson, S. G. Nat. Rev. Genet. 11, 465–475 (2010).
Merhej, V., Royer-Carenzi, M., Pontarotti, P. & Raoult, D. Biol. Direct 4, 13 (2009).
Boscaro, V. et al. Nat. Ecol. Evol. http://dx.doi.org/10.1038/s41559-017-0237-0 (2017).
Itoh, T., Martin, W. & Nei, M. Proc. Natl Acad. Sci. USA 99, 12944–12948 (2002).
Woolfit, M. & Bromham, L. Mol. Biol. Evol. 20, 1545–1555 (2003).
Moran, N. A. Proc. Natl Acad. Sci. USA 93, 2873–2878 (1996).
Wernegreen, J. J. & Moran, N. A. Mol. Biol. Evol. 16, 83–97 (1999).
Husnik, F. & McCutcheon, J. P. Proc. Natl Acad. Sci. USA 113, E5416–E5424 (2016).
Koga, R. & Moran, N. A. ISME J. 8, 1237–1246 (2014).
The author declares no competing financial interests.
About this article
Cite this article
Klasson, L. The unpredictable road to reduction. Nat Ecol Evol 1, 1062–1063 (2017). https://doi.org/10.1038/s41559-017-0263-y
A simple stochastic model describing genomic evolution over time of GC content in microbial symbionts
Journal of Theoretical Biology (2020)