In a recent paper, Lee and Marx (2012) test both how and why they observe large-scale patterns of gene loss in their experimentally evolved populations of Methylobacterium extorquens. They evolved these bacterial populations under different treatments of resource availability (realms of specialization) and found that all replicate populations adapted to their specific treatment over 1,500 generations. During experimental evolution, 80% of the bacterial populations exhibited nearly a 10% reduction in genome size, and many of the gene losses occurred in similar regions of the genome, some even across treatments.
It is a truth universally acknowledged in evolutionary biology, that one species interacting with another species, must be having some effect on that other species’ evolution.
Actually, that’s not really true. Biologists generally agree that predators, prey, parasites, and competitors can exert natural selection on the other species they encounter, but we’re still not sure how much those interactions matter over millions of years of evolutionary history.
On the one hand, groups of species that are engaged in tight coevolutionary relationships are also very diverse, which could mean that coevolution causes diversity. But it could be that the other way around: diversity could create coevolutionary specificity, if larger groups of closely-related species are forced into narower interactions to avoid competing with each other.
Part of the problem is that it’s hard to study a species evolving over time without interacting with any other species—how can we identify the effect of coevolution if we can’t see what happens in its absence? If only we could force some critters to evolve with and without other critters, and compare the results after many generations …
A team of evolutionary microbiologists has performed exactly the experiment I outlined above. The study’s lead author is Diane Lawrence, a Ph.D. student in the lab of Timothy Barraclough, who is listed as senior author.
For the experiment, the team isolated five bacterial species, of very different lineages, from pools of water at the bases of beech trees—ephemeral pockets of habitat for all sorts of microbes that break down woody debris, dead leaves, and other detritus. They cultured the bacteria on tea made from beech leaves, in vials containing either a single species, or all five species, and let them evolve for eight weeks—several dozens of bacterial generations. In a particularly clever twist on standard experimental evolution methods, they also used nuclear magnetic resonance (NMR) to identify the carbon compounds in sterilized tea that had been “used up” by the bacterial cultures, and compared the compounds in fresh beech tea to determine what the bacteria were eating.
And, maybe not surprisingly, the bacterial species’ evolution with company turned out to be quite a bit from their evolution alone. Left alone, most of the species evolved a faster growth rate. This is a common result in experimental evolution, because the process of transferring evolving bacteria to fresh growth medium—”serial transfers” that were performed fifteen times over the course of the experimetn—can create natural selection that favors fast-growing mutants. But, grown all together in the same tube, species that had evolved faster growth rates in the solo experiment evolved slower growth instead.
To find out what had evolved in the multi-species tubes, the team tested the growth of the bacterial species on beech tea that had been used to grow one of the other species, then sterilized. The original, ancestral strains of bacteria generally had negative effects on each others’ growth—they lived on similar compounds in the beech tea, and so their used tea wasn’t very nourishing for the other species. The same thing occurred with the strains that had evolved alone, only stronger, which makes sense in light of the increased growth rates, which would’ve depleted the growth medium faster.
But the interactions among the strains of the different bacterial species that had evolved together was strikingly different. Many of them actually made the tea more nutritious for other species in the evolved community. That is, some of the bacteria had evolved the capacity to eat the waste products of another species that was evolving with them. Using the NMR method to track changes in the presence of different carbon compounds in the tea before and after use provided confirmation that the co-evolved species were using, and producing, complementary sets of resources.
In short, the evolving community didn’t simply become more diverse—it evolved new kinds of mutually beneficial relationships between species that began as competitors.
That evolutionary shift toward mutual benefit had a significant impact on the bacterial community as a whole, too. Lawrence et al. assembled new communities of bacteria extracted from the end-point of the group evolution experiment, and compared their carbon dioxide production, a proxy for overall metabolic activity, to that of a community assembled from bacteria extracted from the end point of the solo-evolution experiments. The community of co-evolved bacteria produced significantly more carbon dioxide, suggesting they were collectively able to make more use out of the growth medium.
So that’s a pretty nifty set of results, I have to say. But I’m also left wondering what it tells us more generally. In both Lawrence et al.‘s paper, and in accompanying commentary by Martin Tucotte, Michael Corrin, and Marc Johnson, there’s a fair bit of emphasis on the unpredictability of the result. Lawrence et al. write, in their Discussion section,
The way in which species adapted to new conditions in the laboratory when in monoculture—the setting assumed for many evolutionary theories and experiments—provided little information on the outcome of evolution in the diverse community.
And, as Corrin et al. note,
These results imply that predictions constructed from single-species experiments might be of limited use given that most species interact with many others in nature.
So … evolution went differently under different conditions? That isn’t exactly a shocking revelation. The fact that this is one of the study’s major conclusions is a symptom of how little experimental work has actually tested the effects of multiple species on evolution. One experiment I’ve discussed here previously, focused on the joint effects of predators and competitors on microbes that live in pitcher plant pitfalls, similarly emphasized the fact that it wasn’t possible to predict the evolutionary effects of predators and competitors together based solely on their individual effects. Work in this line of inquiry is hanging at the point of establishing that complex conditions lead to complex results.
What I’d really like to know—and I think all the authors of both the paper and the commentary would agree with me on this—is how we can begin to make general predictions about community evolution beyond, “it depends what we put in at the start.” It may be that we’ll need a lot more studies like this current one before we can start to identify common processes, and more interesting trends.◼
Turcotte, M., Corrin, M., & Johnson, M. (2012). Adaptive evolution in ecological communities. PLoS Biology, 10 (5) DOI: 10.1371/journal.pbio.1001332
Lawrence, D., Fiegna, F., Behrends, V., Bundy, J., Phillimore, A., Bell, T., & Barraclough, T. (2012). Species interactions alter evolutionary responses to a novel environment. PLoS Biology, 10 (5) DOI: 10.1371/journal.pbio.1001330
Recently the open-access PLoS Biology published a really cool study in experimental evolution, in which a disease-causing bacterium was converted to something very like an important plant symbiont. The details of the process are particularly interesting, because the authors actually used natural selection to identify the evolutionary change that makes a pathogen into a mutualist.
Life as we know it needs nitrogen – it’s a key element in amino acids, which mean proteins, which mean structural and metabolic molecules in every living cell. Conveniently for life as we know it, Earth’s atmosphere is 78% nitrogen by weight. Inconveniently, that nitrogen is mostly in a biologically inactive form. Converting that inactive form to biologically useful ammonia is therefore extremely important. This process is nitrogen fixation, and it is best known as the reason for one of the most widespread mutualistic interactions, between bacteria capable of fixing nitrogen and select plant species that can host them.
Clover roots, with nodules visible (click through to the original for a nice, close view. Photo by oceandesetoile.
In this interaction, nitrogen-fixing bacteria infect the roots of a host plant. In response to the infection, the host roots form specialized structures called nodules, which provide the bacteria with sugars produced by the plant. The bacteria produce excess ammonia, which the plant takes up and puts to its own uses. The biggest group of host plants are probably the legumes, which include the clover pictured to the right, as well as beans – this nitrogen fixation relationship is the reason that beans are the best source of vegetarian protein, and why crop rotation schemes include beans or alfalfa to replenish nitrogen in the soil.
For the nitrogen-fixation mutualism to work, free-living bacteria must successfully infect newly forming roots in a host plant, and then induce them to form nodules. The chemical interactions between bacteria and host plant necessary for establishing the mutualism are pretty well understood, and in fact genes for many of the bacterial traits, including nitrogen-fixation and nodule-formation proteins thought to be necessary to make it work are conveniently packaged on a plasmid, a self-contained ring of DNA separate from the rest of the bacterial genome, which is easily transferred to other bacteria.
This is exactly what the new study’s authors did. They transplanted the symbiosis plasmid from the nitrogen-fixing bacteria Cupriavidus taiwanensis into Ralstonia solanacearum, a similar, but disease-causing, bacterium. With the plasmid, Ralstonia fixed nitrogen and produced the protein necessary to induce nodule formation – but host plant roots infected with the engineered Ralstonia didn’t form nodules. Clearly there was more to setting up the mutualism than the genes encoded on the plasmid.
Wild-type colonies of Ralstonia (tagged with fluorescent green) are unable to enter root hairs (A), but colonies with inactivated hrcV genes are able to enter and form “infection threads,” like symbiotic bacteria (B). Detail of Marchetti et al. (2010), figure 2.
This is where the authors turned to natural selection to do the work for them. They generated a genetically variable line of plasmid-carrying Ralstonia, and used this population to infect host plant roots. If any of the bacteria in the variable population bore a mutation (or mutations) necessary for establishing mutualism, they would be able to form nodules in the host roots where others couldn’t. And that is what happened: three strains out of the variable population successfully formed nodules. The authors then sequenced the entire genomes of these strains to find regions of DNA that differed from the ancestral, non-nodule-forming strain.
This procedure identified one particular region of the genome associated with virulence – the disease-causing ability to infect and damage a host – that was inactivated in the nodule-forming mutant strains. As seen in the figure I’ve excerpted above, plasmid-bearing Ralstonia with this mutation were able to form infection threads, an intermediate step to nodule-formation, where plasmid-bearing Ralstonia without the mutation could not. Clever use of experimental evolution helped to identify a critical step in the evolution from pathogenic bacterium to nitrogen-fixing mutualist.
Amadou, C., Pascal, G., Mangenot, S., Glew, M., Bontemps, C., Capela, D., Carrere, S., Cruveiller, S., Dossat, C., Lajus, A., Marchetti, M., Poinsot, V., Rouy, Z., Servin, B., Saad, M., Schenowitz, C., Barbe, V., Batut, J., Medigue, C., & Masson-Boivin, C. (2008). Genome sequence of the beta-rhizobium Cupriavidus taiwanensis and comparative genomics of rhizobia. Genome Research, 18 (9), 1472-83 DOI: 10.1101/gr.076448.108