« March 2008 | Main | May 2008 »

April 27, 2008

Gene networks: evolved not designed

This week's paper, "Evolvability and hierarchy in rewired bacterial gene networks", was suggested by Joel Lopez.

Randomly changing parts in a machine often breaks it. "Intelligent design" nuts claim this is also true of living things and that this is somehow evidence for design. The argument is nonsense - just because we had distant ancestors small enough that they didn't need lungs doesn't mean we can survive without them -- but is the claim even true? If a genetic change is big enough to have some effect, is it likely to be lethal? Or do many mutations preserve basic functions, just increasing or decreasing fitness (survival and reproduction) under particular circumstances?

Mark Isalan and colleagues randomly switched promoters that control production of transcription factors in bacteria. Promoters are the sections of DNA molecules that, instead of being transcribed into RNA and then translated into proteins themselves, control transcription of nearby DNA sections. Transcription factors are proteins that bind to and control promoters, so they are among the most important control signals in cells.

Linking the DNA coding for a transcription factor to a random promoter would be equivalent to changing an army codebook, so that "retreat" is translated into "destroy the bridge," for example. So it may seem surprising that 95% of the randomly reorganized bacteria, including some with changed regulation of genes that themselves regulate hundreds of other genes, appeared to grow normally.

One explanation that occurred to me after reading the abstract of the paper was that maybe the added promoter-gene combinations were somehow defective, not doing anything. But each combination also included a fluorescent protein gene after the transcription factor. The modified bacteria made the fluorescent protein, so they presumably made the transcription factor as well. Changes expected to result in negative feedback gave similar fluorescent protein levels to those expected to give positive feedback, however. This suggests that other feedback loops were limiting the effects of those added by the researchers. (Similarly, an army unit receiving a strange order due to a codebook error might ask for independent confirmation first. Or so I assume.)

This paper reminds me of an earlier experiment in which researchers made mice without myoglobin, previously thought to be essential for oxygen supply in muscles. Other aspects of their physiology apparently filled the gap.

Nonlethal genetic changes are the raw material for evolution, so the authors tested their genetically modified strains to see whether there were any conditions under which they were more successful than the original strain. Twelve strains outgrew the original in liquid culture, apparently because they didn't make flagella (propellers that are expensive to make and only useful if there's somewhere better to swim to). Some strains also did better at high temperature than the original.

Maybe our earliest single-cell ancestors were so simple that most genetic changes would be lethal. But humans, mice, and the bacteria used in these experiments are the products of billions of years of evolution. Lineages that were killed by every little mutation died out, while those with enough redundancy to tolerate most mutations survived and continued to evolve.

April 25, 2008

Readers' Choice!

Today or tomorrow I'll be reviewing a paper suggested by a reader last week, but meanwhile there are lots of other recent papers that look interesting, at least to me....

Clicking on a title below should take you to the abstract, even if you don't have access to full text. If there's a paper you would particularly like me to discuss, leave a comment to that effect. If there's a quorum (at least 10 votes total), I'll review the paper with the most votes.

Molecular Phylogenetics of Mastodon and Tyrannosaurus rex

Availability of prey resources drives evolution of predator-prey interaction

Protein robustness promotes evolutionary innovations on large evolutionary time-scales

Darwinian Evolution on a Chip

The mechanism of sex ratio adjustment in a pollinating fig wasp

Host manipulation by parasites in the world of dead-end predators: adaptation to enhance transmission?

Selection on Major Components of Angiosperm Genomes

You are what your mother eats: evidence for maternal preconception diet influencing foetal sex in humans

Social transmission of nectar-robbing behaviour in bumble-bees

Feeding, fecundity and lifespan in female Drosophila melanogaster

Is it necessary to assume an apartheid-like social structure in Early Anglo-Saxon England?

April 19, 2008

Separate vacations and other sexual differences

Three recent papers in Proceedings of the Royal Society discuss differences between males and females or, in one case, among males.

"The costs of risky male behaviour: sex differences in seasonal survival in a small sexually monomorphic primate" by Cornelia Kraus and others, is based on a 10-year study of differences between male and female behavior in grey mouse lemurs. During the breeding season, males had lower survival than females, despite any possible risks associated with pregnancy or raising young. The higher risk for males apparently resulted from their tendency to travel more, looking for females.

The sexes also differ in winter behavior: females hibernate, while males remain active. Is there something about female physiology that makes hibernation healthier for them than it would be for males? Maybe, but there was no difference in winter survival between the sexes, which don't differ much in size in this lemur species. The authors suggest that hibernation might have longer-term benefits in females, such as increased lifespan, whereas males need to stay active to bulk up in preparation for the breeding season.

This paper reminded me of an earlier paper on albatrosses, in which "in each pair, the male spent the winter just north of the pack ice in Antarctic waters whereas the female stayed south of Madagascar." It's not hard to understand why males and females might differ in various ways (size, color, etc.) but differences in behavior outside of the breeding season are more interesting.

The second paper addresses an old argument between Charles Darwin and Alfred Russel Wallace, who developed similar explanations of evolution by natural selection at about the same time.

Darwin thought that differences between male and female butterflies resulted from sexual selection: evolution of males, driven by female preferences. Wallace thought that female evolution was more important. They could evolve to be more cryptic (better camouflaged) or they could evolve to mimic a poisonous or distasteful species avoided by predators. In a paper titled "Mimetic butterflies support Wallace's model of sexual dimorphism," Krushnamegh Kunte analyzed the family tree of a large group of related butterfly species. Males of related species resembled each other, but females sometimes resembled other species avoided by predators, although there were some exceptions. This pattern is more consistent with Wallace's hypothesis than Darwin's. I heard a seminar by Steve Pruett Jones last week on fairy wrens; there, too, differences between the sexes have sometimes resulted from female rather than male evolution.

In the third paper, "Genetic variation in threshold reaction norms for alternative reproductive tactics in male Atlantic salmon, Salmo salar", Jacinthe Piche and others consider differences among males. Male salmon that spend years in the ocean grow to > 1 kg, whereas other males become sexually mature as much smaller "parr" (10-150 g) without leaving the river. Parr fare poorly in direct competition with much larger males, but they can sneak in and fertilize some eggs. (This reminds of "sneaker male" lizards, which resemble females enough that other males may fail to drive them away.)

The authors compared different salmon populations. In a common environment, larger males from each population were more likely to become sexually mature - smaller males would presumably migrate out to sea and return when they were much larger - but there were differences among populations. This suggests that the threshold for sexual maturity depends on genes which differed among populations. Genetic differences are the raw material for evolution. So if ocean conditions become more dangerous, e.g., due to fishing, genes for males maturing sexually in rivers at a smaller size would become more common. In some fish species, individuals can change to the opposite sex, depending on size and other factors. The threshold conditions for sex change in fish could presumably evolve similarly.

April 14, 2008

Dumbing down intelligent design

'No practical biologist interested in sexual reproduction would be led to work out the detailed consequences experienced by organisms having three or more sexes; yet what else should he do if he wishes to understand why sexes are, in fact always two?' -- R. A. Fisher (1930).

The scientific definition of "theory" is very different from its popular meaning of "wild speculation." The Theory of Evolution, like the Germ Theory of Disease, or the Atomic Theory that forms the foundation of chemistry, is solidly based in observations and experiments. The "theoretical" part, in each case, is a collection of well-tested principles that make sense of the masses of data and let us make predictions. For example, Germ Theory led to measures to limit the spread of AIDS, where the Divine Punishment Theory failed. The Theory of Evolution has been equally successful, with slowing the spread of insecticide resistance in insect pests among its recent contributions.

But is there a place for speculation in biology? I think there is, so long as we don't confuse it with fact or well-grounded theory. For example, life as we know it uses nucleic acids for heredity and makes much of its cellular machinery from proteins, but can we think of other possibilities? If so, can we design experiments that would detect such alien lifeforms, if they exist, on Mars or perhaps even on Earth?

Similarly, what if some alien life-form -- any sufficiently advanced life-form is indistinguishable from a god -- has intervened in evolution here on earth? Could we develop quantitative methods to measure this effect, as we now do for natural selection and gene flow? Or, suppose we had an old bloodstain purported to be from a demigod; could we extract DNA, look for alleles that don't match anything in the human genome, and (if we found any) clone them into E. coli? A gene for smiting might have military applications. Perhaps others could be reverse-engineered for flood control. (Hey, mixing religion and science was their idea, not mine!)

The topics in the last paragraph may be too speculative to be competitive for tax-supported research grants -- success rates for many NSF programs are around 10% -- but private foundations could certainly fund such research, if they chose. To be taken seriously, however, researchers looking for evidence of intelligent design would need, as in all of science, to design experiments that have the potential to disprove their hypotheses, if those hypotheses turn out to be wrong. And they would need to publish their results in peer-reviewed journals, so that other scientists have a chance to catch any logical fallacies or methodological problems they may have missed.

This is what the advocates of intelligent design have failed to do. Put all the intelligent design papers ever published in a pile and you don't match the productivity of one good graduate student. Whining, as in the much-discussed film, Expelled, is no substitute for science.

April 12, 2008

Fear of flying -- in plants

"Every one is familiar with the difference between the ray and central florets of, for instance, the daisy... But with respect to the [two types of] seeds, it seems impossible that their differences in shape...can be in any way beneficial"--Charles Darwin

The theory of evolution is famously linked to the Galapagos Islands, but this week's paper "Rapid evolution of seed dispersal in an urban environment in the weed Crepis sancta," published in Proceedings of the National Academy of Science, studied much smaller "islands." In an urban environment dominated by concrete, patches of soil around sidewalk trees (below left) are among the few places where plants can grow.
Photo credits: Gilles Przetak and Eric Imbert.

Members of the daisy or sunflower family (Asteraceae) often produce two types of seeds (above right) on the same disk-shaped composite flower head. Seeds from the center of the disk are light in weight and plumed, so they are easily dispersed by wind. Those from the outer edge of the disk are heavier and not plumed, so they tend to fall near the mother plant. Although Darwin apparently failed to see the benefit of having two types of seeds, this kind of diversity acts as a form of bet-hedging. Wind dispersal of seeds over a wide area decreases the chances that all of a plant's offspring will be killed.

Then why not disperse all of the seeds? Because, given that the mother plant managed to reproduce -- many plants don't -- conditions near the mother plant may be better than where most wind-blown seeds might land. This was particularly true in the study discussed here. Earlier, Jonathan Silvertown pointed out, in an essay titled "When plants play the field," that the ratio of the two seed types changes in beneficial ways with changes in flower head diameter. The area of a disk increases four-fold as the circumference doubles, giving proportionally more of the wind-dispersed central seeds. So the plant will always drop some seeds in the same place that it managed to reproduce. But if favorable conditions lead to larger flower heads, more seeds will be dispersed by wind over a larger area, where they can compete with other plant's seedlings rather than with each other.

So, without any genetic change, this disk-size dependence adjusts the ratio of dispersing to nondispersing seeds to match current conditions. But what if conditions consistently favor more or less seed dispersal? Can this ratio also evolve, with a genetic change over generations?

Pierre-Olivier Cheptou and colleagues in Montpelier, France, hypothesized that urban environments would select for fewer dispersing seeds. They showed that most wind-blown, i.e., dispersing seeds fall far from the mother plant. In an urban environment, where the mother plant may occupy the only patch of soil for many meters, most dispersing seeds will land on concrete pavement and die. Therefore, if there are versions of genes that decrease the ratio of dispersing to nondispersing seeds, relative to alternative versions of the same genes, plants with those versions should have more of their seeds survive, making those versions (alleles) more common in each subsequent generation. (Making smaller diameter flower heads, which would generally allow making more of them, would be one way to achieve this.)

To test their hypothesis, the authors collected seeds from the soil "islands" in the urban environment and from outside of town, where dispersing seeds would be more likely to land on soil than concrete. They grew all the seeds in a greenhouse and compared the ratios of dispersing to nondispersing seeds produced there. (By growing seeds from both sources in the same environment, they mainly saw genetic rather than environmental differences, although there is a small possiblity of "maternal effects." For example, if seeds from urban environments had higher lead content, could that effect the development of flower heads on plants grown from those seeds, even apart from any genetic differences? Probably not.) As predicted, seeds from urban islands grew into plants that made relatively few dispersing seeds. Interestingly, they did not find a correlation between flower head diameter and the ratio of seed types. With additional measurements and modeling, they calculated that this genetic difference between city and country plants would only have taken about 12 years to evolve.

Evolution is widely misunderstood. First, some species may have changed little in millions of years, at least in overall body shape, but significant changes can occur in only a few generations (hours for bacteria, years for annual plants), especially in a new environment. Second, just as a river flowing downhill may not always reach an ocean, let alone a particular ocean, evolution responds to current conditions, rather than pursuing some long-term goal. In this example, the descendants of seeds that blow from city to country, or vice versa, will start evolving in the opposite direction from their recent ancestors. And, finally, evolution does not necessarily have anything to do with monkeys! On the other hand, the principles that apply to plants often apply to animals as well, as the quotation below shows.

"Madeira, like many oceanic islands... is much exposed to sudden gales of wind... insects which flew much would be very liable to be blown out to sea and lost. Year after year, therefore, those individuals which had shorter wings, or which used them least, were preserved...." - Alfred Russell Wallace

April 9, 2008

Welcome, fellow Dr. Tatiana fans!

Olivia Judson's latest column includes a good summary of work in my lab on cooperation between soybean plants and the rhizobium bacteria that (typically) provide them with nitrogen. As she points out, "cheating" is less likely to evolve in symbiont populations if they are transmitted in eggs or seeds, relative to symbionts that are acquired from the environment. In the former, if the host dies before reproducing, the symbiont dies, too. Symbionts without brains (bacteria, say) can't anticipate the effects of their actions; it's just that those whose genetically programmed behavior increases host survival become more common over generations.

Similarly, low symbiont diversity within an individual host may favor symbiont investment in costly activities that benefit the host. If each host has many different symbionts, on the other hand, then helping the host indirectly benefits competing symbionts sharing that host.

Rhizobium bacteria reach new host plants through soil, not via seeds, and they can do so even if the host dies without reproducing. Furthermore, each individual plant has multiple strains of rhizobia, which should undermine cooperation. Why then, do most rhizobia use their limited energy supply to fix nitrogen, giving most of it to the host plant? Why not use that energy for their own reproduction, instead?
Although there are several rhizobium strains per plant, they are typically segregated into individual root nodules. So, Toby Kiers and I reasoned, if plants monitor individual nodules and do something nasty to those that provide less nitrogen, that would act as a form of natural selection against cheating rhizobia. A computer model by Stuart West came to similar conclusions. To test this hypothesis, we forced some nodules to cheat, by surrounding them with an argon-oxygen atmosphere lacking nitrogen gas. Control nodules on the same plant got normal air, which is 80% nitrogen. Would rhizobia freed from the burden of fixing nitrogen redirect resources into their own reproduction? Would the plant impose sanctions on nonfixing nodules? If the answers to these questions are yes and yes, what would be the overall effect of cheating on rhizobium reproductive success?

After 10 days, Toby found that the nodules that fixed nitrogen grew bigger and contained up to three times as many rhizobia per nodule. If the number of rhizobia that escape back into the soil (when the plant dies at the end of the season) is proportional to the number of rhizobia inside, then rhizobia that try to cheat their hosts, by diverting resources to their own reproduction at the expense of nitrogen fixation, will actually end up less common in the next generation.

We think this result explains the otherwise surprising (see introduction) level of cooperation between rhizobia and legumes, but it also leads to new questions.

How, physically, do soybeans "punish" nodules that fix little or no nitrogen? As Olivia Judson mentioned, the plant decreased the oxygen supply to the interior of nonfixing nodules. We think this low oxygen may have limited reproduction by rhizobia inside the nodule, either directly or indirectly. As a commentator noted, high levels of oxygen can actually destroy the enzyme that fixes nitrogen. But nodules have a gas diffusion barrier (like insulation, only for oxygen rather than heat), that keeps oxygen inside nodules way below toxic levels. In fact, it's typically down in the range where oxygen is too low for maximum respiration (the oxygen-requiring process by which rhizobia get energy), and it was even lower in our nonfixing nodules. Nodules are red inside because they make hemoglobin to transport oxygen within their rhizobium-infected cells, similar to its role in our blood. (The idea that hemoglobin magically makes oxygen vanish has been obsolete for over 20 years, although you may still see it in textbooks.) But could the relation between low oxygen and low rhizobium reproduction be coincidence? I will be starting experiments soon to find out.

What about legumes other than soybean? Nodules containing nonfixing rhizobia typically grow faster in other species, as Ellen Simms and colleagues have shown with wild lupines. Ryoko Oono, in my lab, started a preliminary experiment today, similar to our soybean experiment, only with alfalfa.

Given the severe effects of sanctions, why are cheating rhizobia still common in some soils? Will Ratcliff, also in my lab, has some ideas and some data on this. Rhizobia can accumulate more resources per cell inside nodules, so numbers aren't the whole story. Also, some rhizobia can manipulate their hosts to give them more resources.

Some cheating rhizobia may prosper by sharing nodules with more altruistic strains. Work in my lab by Alain Chapon, Bob Rousseau, and Lysistrata Munson found that mixed nodules may be more common than we thought.

Thanks to the National Science Foundation for supporting this work; be sure to fund at least one of my submitted proposals so we can continue! For example, the rhizobia that fix nitrogen in some nodules, including those of alfalfa, have lost the ability to reproduce. So why cheat? On the other hand, why fix nitrogen? We think the answer involves kin selection, but it's complicated.

April 7, 2008

Evolutionary trees

This week's paper is "Rapid evolution towards heavy metal resistance by mountain birch around two subarctic copper–nickel smelters", published in the Journal of Evolutionary BIology by J.K. Eranen.

Evolution is a change over generations, so evolution is typically faster (more change per year) in species with short generation times. Signficant evolutionary change in bacterial populations, therefore, can take only a day or two, under ideal conditions. Long-lived species like humans and trees evolve, too, but it takes much longer. So, for example, are trees likely to evolve fast enough to survive climate change?

Eranen studied birch trees that have been exposed to pollution from smelters in Russia. Near the smelter, these trees are typically much shorter (<2 m) than in less-polluted areas (>6 m). This difference is presumably due to some combination of the harmful effects of pollution, beneficial acclimation (nongenetic responses of individuals to conditions), and evolution (a change in the genetic composition of the population). An example of acclimation would be if, in polluted areas, shorter trees actually produce more seeds than tall ones.

To separate these effects, he did what is known, for historical reasons, as a "common garden" experiment, although the "garden" in this case was a greenhouse. Tree seeds from severely polluted areas near two smelters, and control seeds from less-polluted areas, were grown with and without addition of toxic heavy metals. With the toxic metals, trees from polluted areas grew taller, with bigger leaves. This is presumably good, although it would have been interesting to see data on seed production. Because seeds from both environments were grown under the same experimental conditions in the greenhouse, this represents a genetic difference, presumably due to evolutionary change over the 70 years or so that the smelters have been in operation. This could represent a few tree generations, depending on the age at which birches reproduce.

As is often the case, there was a cost to genetic stress resistance: in soil without toxic metals, the trees from less-polluted areas grew better. All of these differences were statistically significant -- that is, we can be 95% sure that they are not due to chance differences among trees -- but they were also fairly small. For example, in the pollution treatment, trees from unpolluted areas grew about 80% as tall as those from polluted areas. This may be because soil concentrations of nickel and copper in the experiments were about one-third what the trees had been exposed to in the polluted areas.

So evolution is detectable in trees, but will it be fast enough to allow trees to survive in their present locations? Or, as boreal forests warm, will they suffer some combination of devestation and invasion by trees from warmer climates? The latter seems more likely to me.

Also this week:

Bacteria Subsisting on Antibiotics

...discussed by Ed Yong.

Social networks in the lek-mating wire-tailed manakin (Pipra filicauda)

Chromosomal Gene Movements Reflect the Recent Origin and Biology of Therian Sex Chromosomes

Climate Change, Humans, and the Extinction of the Woolly Mammoth

...dIscussed by Ed Yong.