Why have sex?

Obvious as the answer may be, from an evolutionary perspective the existence of sexual reproduction is a genuine puzzle: sexually reproducing species partition half their genetic material into eggs and sperm (or ovules and pollen in plants), therefore producing offspring that are only 50% related to them.

But some plant species can self-fertilise (AKA self-crossing), producing offspring which are 100% related to themselves.  If such a self-fertilising gene arises, the individual carrying it enjoys two advantages: it will transmit twice as many of its genes (i.e. all of them rather than half) to its offspring (known as the automatic selection hypothesis); and it is guaranteed a mate by fertilising its own ovules (known as the reproductive assurance hypothesis).

A recent paper in Molecular Biology and Evolution set out to discover which of these two reasons is the most important: does self-crossing evolve because there are few mates available, as in the reproductive assurance hypothesis, or does it evolve according to the automatic selection hypothesis, because self-crossing produces such closely related offspring?

The authors looked at the mustard plant (Leavenworthia alabamica), which has self-compatible (SC) and self-incompatible (SI) (and therefore sexually reproducing) races that have only recently evolved, making them ideal for studying the genetic basis of self-compatibility.  The evolution of self-compatibility in one of these races has produced a spectacular derived flower shape, making entry exclusive to pollen from that race.

They sequenced the Lal2 gene from four SC and two SI races, and inferred the gene’s function in each race, as well as how different the alleles (different forms of the gene) are in each race.  They found that these alleles segregated (separated into offspring) concurrently with the ability to self fertilize: SC races only received one of four forms, and SI races received one of another two forms.  From this they inferred that the gene is indeed governing whether the plants self-fertilized or not.

Next they looked at the relationships between the alleles.  The Hillis plot [Fig. 2] shows the Lal2 alleles in different races (in different colours) clustering with each other, with each cluster assumed to represent a different allele of the Lal2 gene.

This, combined with the high level of divergence between the genes from different races, shows that the different races have different forms of the Lal2 gene.  More interestingly, that self-incompatibility has evolved more than once (and thus from multiple genetic causes) is supported by three results: one allele is fixed in each different race (i.e. no other forms of that gene are found in the population); there is no close relationship between the two self-compatible races’ alleles; and the SC races split from the SIs 150,000 (for three races) and 48,000 (for the one remaining race) years ago, with very little migration between them since (and hence little opportunity for different races to mix genes again).

They then used these sequences to estimate the genetic diversity and past population structure of each race.  They found that “all populations except population TNC17 showed evidence of a strong bottleneck at roughly the same time in the past,” and that genetic diversity is extremely low (0 or close to it) in the SC races, as would be expected if they do not mix genes by sexual reproduction.

The most interesting result came when they simulated the calculation of genetic diversity for the Landersville population, an SC race, to account for the effects of inbreeding: basically, they modelled what would have happened if it had been sexually reproducing rather than self-fertilising.  They found that this produced the same genetic diversity as the SI races: the population history of bottlenecking was still the same, but removing inbreeding caused an increase in genetic diversity.

From these results the authors concluded the following (my brackets):

“the fact that the a4 race [an SC race] underwent a large population bottleneck in population size and is restricted to the periphery of the species range appears to collectively suggest that reproductive assurance facilitated the evolution of selfing in this lineage.”

In other words, the mustard plants jumped at the chance to spread more of their genes, rather than being pushed, by a lack of mates, into mating with themselves.

However, in the line previously the authors admit (my brackets again):

“Declines in genetic diversity in selfing [self-compatible] populations may reflect a bottleneck in population size that triggered the evolution of selfing, yet the evolution of selfing may also facilitate bottlenecks once it has evolved, since a single selfing seed can found a new population.”

This is a fundamental flaw with this kind of approach: if you are trying to infer the cause (population bottlenecks) of a trait (the evolution of self-compatibility) from the presence of the cause, you can’t use a cause which could have itself have been caused by the trait!

This paper produces valuable results by pinning down the genetic basis of self-compatibility, and is thorough in its investigation of the evolutionary history of those genes.  However, I can’t say I agree with their conclusion: the arrow of causality between population bottlenecks and self-compatibility could still point either way.

Returning to the evolutionary logic behind self-compatibility, I’m afraid a sexless life isn’t all good news.  Self-fertilising species have a nasty habit of quickly going extinct, as seen in self-fertilising species of the genus Arabidopsis.  This is caused by a process known as Muller’s Ratchet: put simply, in a self-fertilising species any mutation arising in the parent will be inherited by the offspring, which will then contribute their own mutations to the load and pass that onto their offspring, ad infinitum.  This increases the chance that a lethal (or at least very harmful) mutation, or combination of mutations, will be inherited.  Offspring of sexual species receive a mixture of their parents’ genes, so some may receive lots of their parents’ mutations, and some will receive very few.  Natural selection will cause the mutation-heavy offspring to die out, but the mutation-light survive.  So sex is good for something after all.


Busch, J.W., Joly, S. and Schoen, D.J. (2011) Demographic signatures accompanying the evolution of selfing in Leavenworthia alabamica.  Molecular Biology and Evolution 28: 1717-1729


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s