« December 2010 | Main | February 2011 »

January 28, 2011

Also this week...

Defeating Creationism in the Courtroom, But Not in the Classroom
Only 28% of high school biology teachers are doing their job. See comments by PZ.

Phylogeny and palaeoecology of Polyommatus blue butterflies show Beringia was a climate-regulated gateway to the New World Nabokov's hypothesis was apparently right, "A novel method is used to estimate ancestral temperature tolerances using the limits of distribution ranges of extant organisms", etc. See comments by Carl Zimmer.

Stress-induced evolution of Escherichia coli points to original concepts in respiratory cofactor selectivity Evolution of a new function for the key metabolite, NADPH

Leaving home ain't easy: non-local seed dispersal is only evolutionarily stable in highly unpredictable environments Spreading seeds widely won't usually get them to consistently better spots, so it's more of a bet-hedging mechanism.

Importance of single molecular determinants in the fidelity of expanded genetic codes The latest on making organisms that use a DNA stop codon to encode an amino acid not normally found in proteins

Inland post-glacial dispersal in East Asia revealed by mitochondrial haplogroup M9a'b Using DNA to track human migration history

The Newest Synthesis: Understanding the Interplay of Evolutionary and Ecological Dynamics Ecological changes can drive evolution, but what about the reverse?

Oceanic rafting by a coastal community
I still don't believe that's how camels got to Australia.

Evidence of parasitic Oomycetes (Peronosporomycetes) infecting the stem cortex of the Carboniferous seed fern Lyginopteris oldhamia

Newly identified and diverse plastid-bearing branch on the eukaryotic tree of life
We can't grow these newly discovered microbes in culture, but we can classify them as a new branch based on their DNA, and we can use their DNA sequences to make fluorescent probes to label them for microscopy.

Oxytocin promotes human ethnocentrism

January 26, 2011

Plants punish cheaters' relatives

This week's paper is the fourth from Ryoko Oono's PhD thesis. "Failure to fix nitrogen by non-reproductive symbiotic rhizobia triggers host sanctions that reduce fitness of their reproductive clonemates" was just published on-line in Proceedings of the Royal Society.
Rhizobia are bacteria that can live either in soil or in root nodules, like those shown above. Legume plants (alfalfa, soybean, the lupines loved by Monty Python and many wild species) let rhizobia in because the rhizobia (usually) convert atmospheric nitrogen into forms the plant can use.

But what if the rhizobia don't deliver? What if, once established inside a nodule, they use plant resources only for their own reproduction? In my most-cited paper, Toby Kiers showed that soybean plants impose fitness-reducing "sanctions" on rhizobia that fail to fix nitrogen. Ellen Simms' lab found similar results with wild lupines. But Ryoko had three good reasons to question whether certain legumes, including alfalfa and pea, impose sanctions similar to soybean's.

A key difference is that, in alfalfa and pea nodules, rhizobia lose the ability to reproduce when they become the nitrogen-fixing, bacteroid form. Earlier, Ryoko showed that legumes have repeatedly evolved the ability to suppress bacteroid reproduction, apparently because it increases the efficiency with which bacteroids use plant carbon to fix nitrogen from the atmosphere. If a soybean plant cuts off resources to a bacteroid that isn't fixing nitrogen, that bacteroid is less likely to survive and reproduce. But bacteroids in alfalfa nodules can't reproduce anyway.

One obvious question is, why do rhizobia infect alfalfa roots, if they'll lose the ability to reproduce? The answer is that rhizobia that can infect alfalfa don't usually have the option of infecting soybeanm where bacteroids remain reproductive, instead. They could just stay in the soil, but they reproduce much more inside nodules -- often a million-fold or more, even after subtracting those that lose the ability to reproduce by becoming bacteroids.

Given this difference, why did sanctions seem somewhat less likely in alfalfa nodules (hosting nonreproductive bacteroids) than in soybean nodules (hosting reproductive bacteroids)?

First, because reproductive bacteroids, like those in soybean nodules, might be more likely to "cheat." They can cheat simply by hoarding plant resources for their own reproduction, rather than using those resources to power the nitrogen fixation that benefits their plant host. But why would a nonreproductive bacteroid hoard resources? They can't take it with them when they die. If rhizobia that cheat alfalfa are rare, would natural selection in alfalfa populations still maintain the ability to impose sanctions? Maybe. Nonreproductive bacteroids do still have some cheating options, such as diverting plant resources to their reproductive clonemates in the same nodule, perhaps via chemicals called rhizopines. Also, some nonfixing rhizobia could just be defective mutants rather than strains whose cheating strategies have been honed by natural selection. Defective mutants might be just as common in species with nonreproductive bacteroids.

Second, cutting off resources to nonfixing bacteroids wouldn't necessarily reduce the fitness of their reproductive clonemates in the same nodule. To qualify as sanctions, by our definition, plants would also have to cut off resources to those clonemates, like a tyrannical government punishing the relatives of insurgents. Of course, the plants are just reducing their losses, with fitness effects on the rhizobia as a side-effect.

Third, two published studies of Medicago truncatula, a wild relative of alfalfa, concluded that it doesn't impose sanctions.

Nonetheless, Ryoko found that nodules prevented from fixing nitrogen -- she used a nitrogen-free, argon/oxygen atmosphere -- grew less (lower two panels in photo) and contained less than half as many rhizobia, relative to the same strain allowed to fix nitrogen. So alfalfa does impose sanctions after all, perhaps cutting off resources to an entire nodule when it fails to fix nitrogen. She got similar results with pea, which also hosts nonreproductive bacteroids.

What about those published results that didn't find sanctions? Heath and Tiffin (2009) compared three strains, but maybe none of them performed poorly enough to trigger sanctions. (Earlier, Toby Kiers found that minor cheating doesn't necessarily trigger sanctions.) Gubry-Rangin found less nodule growth with a nonfixing strain (consistent with our results, inconsistent with Heath and Tiffin) but no difference in the number of rhizobia inside (inconsistent with our results). It seems odd that a smaller nodule would have just as many rhizobia inside, but maybe there's something interesting going on. I expect there will be more to this story.

Our research on cooperation and conflict in the symbiosis between legumes and rhizobia has been supported by the National Science Foundation, most recently by grant NSF/IOS-0918986.

January 21, 2011

Modeling reproduction/longevity tradeoffs and phenotypic plasticity in fluctuating environments

A year ago, I was passing through beautiful Brisbane (in the news recently because of disastrous flooding) on my way back from the Applied Evolution Summit on Heron Island. This week, I'll discuss one figure from a paper I wrote for that meeting. An online-early version of "Past evolutionary tradeoffs represent opportunities for crop genetic improvement and increased human lifespan" is up at Evolutionary Applications, which will publish a special issue of papers from the meeting.

The basic idea is that, if there is even a small trade-off between reproduction and longevity, natural selection will sometimes sacrifice reproduction in favor of longevity. The graph shows predicted numbers for three genotypes, based on a computer model (Hormesis8.py). Fertile adults are assumed to have a 25% chance of dying in a given year, while those delaying reproduction have a 20% chance of dying. That slightly lower chance of dying isn't enough to give genotypes that always delay reproduction (dots) an advantage over those that never delay reproduction (solid line).

But what if individuals delay reproduction only when some environmental cue predicts a population decline? That's what the dashed line represents. In good years, their population grows as fast as the never-delay genotype. In bad years, 50% of the dashed-line genotype are assumed to detect the population-decline cue and delay reproduction, so that they die off a bit more slowly. In an environment where conditions fluctuate (assumed to affect the survival of juveniles), this form of phenotypic plasticity beats both of the fixed strategies.

So what's the "population-decline cue"? The model just assumes there is such a cue and that half of the dashed-line individuals detect it. In a PLoS One paper in 2009, we proposed "eating famine foods" as an example of such a cue. We hypothesized that our ancestors ate less-desired foods mainly when more-desired foods weren't available, that is, during a famine. Famines often lead to population declines. Delaying reproduction until after the population has declined means that each offspring makes a greater relative contribution to the gene pool.

Plants high in insect-repelling toxins might be an example of such "famine foods", even if some modern humans have developed a taste for kale, coffee, or hot peppers. These plant toxins might have small negative effects on our health. But, if our bodies respond to the information carried by those toxins -- famine! population decline likely! delay reproduction! -- then those negative effects may be outweighed by the health benefits of setting our hormone levels etc. to values optimized for longevity rather than reproduction.

January 14, 2011

This week's picks

Some recent papers that look interesting:

Speciation along a depth gradient in a marine adaptive radiation

Causes of lifetime fitness of Darwin's finches in a fluctuating environment

Inland post-glacial dispersal in East Asia revealed by mitochondrial haplogroup M9a'b

Mothers matter! Maternal support, dominance status and mating success in male bonobos (Pan paniscus)

Reproductive state affects reliance on public information in sticklebacks

Is evolutionary history repeatedly rewritten in light of new fossil discoveries?

January 12, 2011

Making symbiotic rhizobia more efficient

This week I'll discuss a recent paper by Ryoko Oono, one of five from her PhD work. (If you want to hire her as a postdoc, better hurry!) Comparing Symbiotic Efficiency between Swollen versus Nonswollen Rhizobial Bacteroids, published in Plant Physiology, is available for anyone to read on-line.

Here's a little background. Rhizobia are soil bacteria, best known for infecting the roots of legume plants (clover, lupine, bean, pea, alfalfa, and many others) and multiplying inside root nodules, where they "fix" nitrogen gas from the atmosphere, converting it into forms their plant hosts can use, instead of relying on fertilizer.

Thumbnail image for AlfalfaNodules2.jpg

Left: Alfalfa nodules; copyright Inga Spence, used by permission. Below left: nonswollen bacteroids. Below right: swollen bacteroids.
Thumbnail image for Bacteroids.jpg

Apparently, rhizobia in the soil never fix nitrogen. Inside root nodules, however, some of them develop into bacteroids, which use some of the plant-supplied carbon they consume to power nitrogen fixation. They may also hoard some carbon for their own future survival and reproduction, a possible source of conflict with their host.

In nodules of some legume hosts, including pea and alfalfa, bacteroids are swollen (above right) and have lost the ability to reproduce. In other hosts, bacteroids don't look that different from the free-living rhizobia, and retain the ability to reproduce. Why this difference?

"Why" questions in biology can have two different kinds of answers. A "proximate" (as in "nearby") explanation for bacteroid swelling is that it is caused by certain peptides (like proteins, only smaller), which are produced by some legume species but not others, as shown by recent research by Van de Velde, Kondorosi, Mergaert, and collaborators.

"Ultimate" answers (aside from "42") are evolutionary explanations. Why did some legumes evolve the ability to make peptides that cause bacteroids to swell? Why have rhizobia not evolved resistance to this manipulation, particularly since swollen bacteroids have apparently lost the ability to reproduce?

Earlier, Dr. Oono showed that this plant trait (causing bacteroid swelling) has evolved at least five times. This repeated evolution made us suspect that swollen bacteroids might somehow be more beneficial to their plant host, relative to nonswollen ones.

What would "more beneficial" mean? Fixing twice as much nitrogen would be good, but not if that cost the plant three times as much carbon. In that case, it would have been cheaper for the plant to make twice as many nodules. So "more beneficial" has to mean "more efficient", that is, "fixing more nitrogen, relative to their carbon cost."

How could we test the hypothesis that, by making bacteroids swell up, plants make the bacteroids more efficient? Ideally, we wanted two things. First, some way of controlling bacteroid swelling, without changing anything else. Second, some method to measure the ratio of nitrogen fixation by bacteroids to their carbon consumption.

We didn't quite have either of those, but we came close. We could have measured the efficiency of swollen bacteroids of species X in nodules of host species A, and compared that with nonswollen bacteroids of species Y in host B. But how much of that difference would be due to bacteroid swelling, and how much due to other differences between species? Fortunately, there are a few rhizobia what make swollen bacteroids in one host and nonswollen bacteroids in another. We used two rhizobial strains with this useful property. One made swollen bacteroids in pea and nonswollen ones in bean. The other made swollen bacteroids in peanut and nonswollen ones in cowpea.

To estimate nitrogen fixation, we made use of the fact that, in the absence of nitrogen gas -- we used 20% oxygen and 80% argon -- the nitrogen-fixing enzyme produces hydrogen gas, at a rate proportional to what its nitrogen fixation would have been, if nitrogen were available. To measure total carbon consumption, we would have to measure both the carbon stored by bacteroids and the carbon they respired away as carbon dioxide. We assumed that, once a bacteroid has been around for a while, it's probably not accumulating too much carbon, at least relative to the amount it respires. So we measured nitrogenase activity (as hydrogen production) as a function of respiration rate (carbon dioxide production).

One problem is that not all of the carbon dioxide released by a nodule is directly linked to nitrogen fixation. One could argue that the ratio of total hydrogen production to total carbon dioxide production is an overall measure of efficiency at the nodule level, but differences between swollen bacteroids in one host and nonswollen bacteroids in another host could easily reflect differences in nodule structure between legume species. To tie efficiency differences more closely to bacteroid swelling, we measured the increase in respiration with an increase in hydrogen production by the nitrogen-fixing enzyme. That way, baseline differences in nodule respiration between species were left out of the ratio. To increase these rates, we simply increased oxygen concentration slightly, which had no effect on respiration in the oxygen-saturated parts of the roots and nodules, but increased activity by the oxygen-limited bacteroids. This approach was developed by the late John Witty, with whom I spent an enjoyable and productive month at the Welsh Plant Breeding Station, in 1989.


In both cases, swollen bacteroids had more nitrogen-fixing enzyme activity, relative to their respiration cost, when compared to the nonswollen bacteroids of the same rhizobial strain. Consistent with this apparent difference in intrinsic efficiency, each rhizobial strain gave more plant growth, per gram of root nodule, in the host where its bacteroids became swollen.

We think that's why legumes have repeatedly evolved the ability to make rhizobial bacteroids swell up inside their nodules: it costs them less carbon to get the nitrogen they need.

That's fine for the plants, but what about the rhizobia? I mentioned that swollen bacteroids have apparently lost the ability to reproduce. Failure to reproduce is unlikely to fare well under natural selection. So I wouldn't be surprised if further research uncovers some attempts by rhizobia to resist this manipulation by their legume-plant hosts.

But should these rhizobia give up on symbiosis altogether? After all, rhizobia can survive and reproduce in the soil, without nodulating plants. However, each nodule containing swollen, nonreproductive bacteroids also contains many (a million or so) genetically identical rhizobia that haven't yet turned into bacteroids. These nonswollen clonemates of the swollen nonreproductive bacteroids can still reproduce. Many of them are thought to escape alive back into the soil, when the nodule senesces. That's apparently enough of an inclusive-fitness benefit to keep rhizobia coming back to legume roots, through a process similar to kin selection.

I will be discussing some practical implications of our research on legume-rhizobium coevolution next month, at the Soybean Breeders' Workshop in Saint Louis. For example, would legume crops that don't make bacteroids swell up have higher yield if they did? Or does this benefit depend on the environment or on other aspects of legume physiology, so that it's already evolved in every case where it would be beneficial?

This material is based upon work supported by the National Science Foundation under Grant No. NSF/IOS-0918986.

January 3, 2011

How inevitable was the origin of life on Earth?

MAJIKTHISE: Seven-and-a-half million years? What are you talking about?

I said I'd have to think about it didn't I? And it occurs to me, that running a program like this is bound to cause sensational public interest.

Evolutionary biologists mainly study how life has diversified and changed or how it is diversifying and changing, rather than how it originated. But I thought this article was really interesting. "Chance or necessity? Bioenergetics and the probability of life", published in the Journal of Cosmology by Nick Lane (University College, London), doesn't have any original data, but cites lots of other papers that do.

This journal publishes some strange stuff, but journal review processes are never more than an somewhat-useful filter. If you can't get your hypothesis or data published in any peer-reviewed journal, you're probably a crackpot. On the other hand, there's no guarantee that a paper published in a top journal will stand up to subsequent criticism. Anyway...

First, Lane argues that "the emergence of life is probable on any wet, rocky planet." Serpentization, a chemical reaction between water and rock, leads to complex mineral structures on the sea floor, with moderate temperatures, cell-size pores, and gradients in hydrogen-ion concentration similar to those which, when they occur across cell membranes, power life (ATP synthesis, rotation of bacterial flagella, transport across membranes, etc.). Given these conditions, he suggests, life is virtually certain to arise eventually.

On the other hand, he argues that complex life is far from inevitable. If cellular processes depend on energy stored as hydrogen-ion gradients across membranes, then cells can't get very big. As a cell doubles in diameter, its volume goes up by a factor of 8, while its surface membrane area only goes up by a factor of 4. To get more energy per unit volume, cells need more membranes-with-gradients per unit volume. We humans have solved this problem. We just pack our cells with mitochondria, each a little power plant with its own membranes. It's now generally accepted that mitochondria, which have their own DNA, are descended from symbiotic bacteria. Lane provides various lines of evidence that plants, animals, and fungi are all descended from the same mitochondrion-containing ancestorm i.e., that this key step only happened once. Various single-celled organisms that look like intermediate steps, he argues, are instead examples of secondary losses of complexity. For example, cells with real nuclei, like ours, but without mitochondria, are all apparently descended from ancestors that had mitochondria and lost them.

For more detail, read the paper. But I do have a couple of comments.

First, if cell-size pores with hydrogen-ion gradients and a few other characteristics (simple organic molecules that can arise from nonbiological processes, minerals that act as simple catalysts, etc.) have a high probability of producing life, can we repeat the process in the laboratory? I'm thinking lots of ceramic sheets, each with billions of cell-sized pores, with different inorganic chemical solutions on the two sides to generate the right gradients. Of course, it might take a few million years to get results. Then again, it might not. Seems like it might be worth trying.

Second, is there some good reason why bacteria couldn't get bigger and more complex, without increasing their volume:membrane ratio, by growing as branching filaments or nets?