« October 2011 | Main | December 2011 »

November 28, 2011

Experimental evolution of metabolism, sex, and multicellularity

Update: links to our open-access Proceedings of the National Academy of Sciences paper on experimental evolution of multicellularity, including PDF and great videos, can be found at the Microbial Population Biology (Micropop) website.

The Nov. 18 issue of Science has a news feature by Elizabeth Pennisi on recent research using experimental evolution, including some work on the evolution of multicellularity, led by Mike Travisano and Will Ratcliff, in which I've been involved.

Two interesting experimental evolution projects are underway in Canada. In Montreal, Graham Bell has been evolving algae that get their energy from a simple organic molecule, acetate, instead of from light. At first, the algae could barely survive without light, but after five years (still a fraction of the time that Richard Lenski has been evolving E. coli) he has hundreds of independent lines that have evolved a variety of ways to grow on acetate in the dark.

In Toronto, Aneil Agrawal is subjecting the sex life of rotifers to experimental evolution. Like aphids, Daphnia, and some other species, rotifers normally reproduce asexually, resorting to sex only under stress. Populations consisting of females, producing other females asexually, grow twice as fast as populations that are half male. (In my forthcoming book, Darwinian Agriculture, I discuss how reindeer herders increase production by harvesting mostly male calves for meat, so that most adults are females producing more calves, rather than males fighting over females.) But sexual reproduction shuffles genomes in ways that may be beneficial under different conditions. Agrawal and his postdoc Lutz Becks found that the balance between sexual and asexual reproduction evolved in response to environmental conditions. In stable environments, sex eventually disappeared. Once you've evolved the perfect genotype for some particular stable environment, why scramble that genotype through sex?

Meanwhile, we've been exploring the transition to multicellularity.
Cellular differentiation in multicellular clusters evolved from unicellular yeast (photo by Will Ratcliff).

Unicellular life apparently had the earth to itself for over a billion years before even simple multicellular life evolved. So you might think that this major evolutionary transition requires some complicated series of genetic changes that would only happen rarely. Alternatively, maybe the first simple multicellular organisms weren't that different, genetically, from their unicellular ancestors -- they just couldn't out-compete their unicellular parents until conditions were right.

Individual cells would have greater access to nutrients in their environment than cells in the middle of a cluster, but what advantages might clusters have, under what conditions?

Clusters create opportunities for different cells to specialize in different tasks, but was that benefit key to early multicellularity? A genetically diverse group of cells that happened to stick together, as in a biofilm, might instantly benefit from such complementarity. But genetic differences among cells can also lead to conflicts of interest and the evolution of "free-riders", which benefit from the activities of other cells in a cluster without contributing in turn.

Another way to get get a cluster of cells is for cells to stick together after dividing. This solves the conflict-of-interest problem, but you wouldn't expect any beneficial specialization, at least at first. Possible benefits of clusters, even before division of labor evolves, include economies of scale and larger size itself.

In PLoS Biology, Koschwanez et al. recently explored economies of scale. Microbes release various extracellular enzymes, often to break large food molecules into pieces small enough to absorb. If a unicellular organism does this, other cells nearby may get much of the benefit. But if each cell in a cluster releases a little bit of enzyme, they can make enough to do the job, without any individual cell paying too high a cost. Koschwanez at al. showed that clusters can survive under some conditions where unicells can't, and they can compete better against "free-riders."

What about the benefits of size itself? In 1998, Boraas et al. published a paper in Evolutionary Ecology showing that, when subject to predation, unicellular algae quickly evolve an 8-cell phenotype, too big for the predators to eat. Maybe bigger predators would have selected for still larger and possible more-complex clusters, but nobody seems to have continued this work.

Larger clusters might be more resistant to some physical stresses as well. For example, individual cells in a cluster tumbling in shallow water would be exposed to UV light from above less continuously than individual cells in the same environment. And clusters tend to settle faster through liquid. This isn't always an advantage, but it can be. Deeper water tends to be colder, which slows metabolism and can thereby increase survival when resources are limited. Settling to the bottom could also help a cluster stay in a favorable location, while unicells are washed downstream.

We may never know what factors selected for multicellularity the few times it evolved in nature. But lab experiments may still be able to answer some key questions. Does multicellularity evolve more easily through aggregation of genetically distinct cells, or through adhesion after division? Once simple clusters evolve, what happens next? And, how repeatable are key events?

To answer such questions, my colleague Mike Travisano and our postdoc Will Ratcliff chose rapid settling as a simple, flexible, and repeatable method to give clusters of cells a competitive advantage over single cells. When this screen was applied once per day to unicellular yeast, faster settling evolved quickly. Microscopic examination showed that clusters evolved in each replicate population. When we isolated single cells from these clusters, they grew into new clusters by adhesion after division, not by aggregation of previously separate cells.

And then what? One nice thing about rapid settling as a selection screen is that we can change the intensity of selection without changing its qualitative nature. This would be trickier with predation. I'll have more to say about our results once they're published, but a key finding is that clusters start evolving as clusters, sacrificing some individual cells for the good of the whole.

November 14, 2011

A cure for aging? Not yet.

"Clearance of p16Ink4a -positive senescent cells delays ageing-associated disorders" is an accurate statement of the results presented in a recent paper in Nature. I guess the title would have been too long if they'd added "...in mice genetically engineered to have short-life spans, but the treated mice didn't live longer."

The authors argue that "removal of senescent cells can prevent or delay tissue dysfunction and extend healthspan." They hypothesize that senescent cells release chemicals that cause excessive aging of other cells nearby. So they genetically engineered mice, which were already genetically engineered to age faster than normal, so that their senescent cells could be killed by a drug. Giving the mice the drug eliminated senescing cells faster and had various health benefits, such as better performance on treadmills, relative to the same fast-aging mice not given the drug.

But our ancestors already evolved mechanisms, long ago, to eliminate cells that are no longer needed or are causing problems, like cancer. Are those evolved mechanisms too slow? If so, why? That is, why haven't mutants that clear these cells faster out-competed those that clear them slower? Speeding up an existing natural process is well within the capabilities of natural selection.

Sometimes, there can be trade-offs between longevity (or late-life vigor) and reproduction, especially early reproduction. In such cases, natural selection will often -- but not always -- favor early reproduction over longevity. Could keeping senescing cells around longer increase fertility, or something? If so, then a drug that would speed the elimination of senescing cells could still be useful. We just wouldn't take it until we were finished having children.

But speeding the clearance of senescent cells in older but not younger individuals doesn't seem too difficult for natural selection to have managed either, given millions of years to work on the problem. I'm assuming that older individuals have been contributing to the survival of their children and grandchildren for at least a few million years.

I'm betting that faster-than-natural clearance of senescent cells, which didn't extend actual life-span even in mice engineered to age faster than normal, won't do much for the health-span of normal mice.

November 11, 2011

This week's picks

Bayesian phylogenetic analysis supports an agricultural origin of Japonic languages
The Hadropithecus conundrum reconsidered, with implications for interpreting diet in fossil hominins

The perception of self-agency in chimpanzees (Pan troglodytes)

Explaining rapid reinfections in multiple-wave influenza outbreaks: Tristan da Cunha 1971 epidemic as a case study
The causes of epistasis
Recent Synchronous Radiation of a Living Fossil

Evolution of olfaction in non-avian theropod dinosaurs and birds
Intercontinental dispersal of giant thermophilic ants across the Arctic during early Eocene hyperthermals
Postglacial migration supplements climate in determining plant species ranges in Europe