Interview with Roger Butlin

«Mixing is probably a good thing if we need to adapt to a changing environment»

Professor at Department of Animal and Plant Sciences of the University of Sheffield

Roger Butlin

Professor Roger Butlin works at the Department of Animal and Plant Sciences, Univ. of Sheffield. He has been the past president of European Society of Evolutionary Biology (ESEB) for 2013-2015 and was awarded in 2015 the Darwin-Wallace Medal of the Linnean Society of London to his outstanding contribution to research on Evolution. We had the pleasure to listen to his talk on «How do species evolve?» at the last Memorial Peregrí Casanova, organized the 19th May by the Cavanilles Institute of Biodiversity and Evolutionary Biology and the Unit Chair for Scientific Innovation of the University of Valencia. Despite his tight agenda, he does not show impatience when I ask him for an interview, on the opposite, he agrees very friendly and ready to explain the complexities of Evolution. He focused his conference on the appearance of new species through evolution. Diverging new lineages need to become reproductively isolated to proceed to speciation, so that they could follow their evolutionary way independently, without interbreeding. However, in that process they might cross again, therefore remixing their genomes and hampering the way to speciation. But sometimes these hybrid forms, if fertile and sufficiently different from their progenitors, may give rise to new unexpected evolutionary lines.

Reproductive isolation is the reduction in successful interbreeding between divergent populations. Speciation requires that reproductive isolation evolves to completion, that is, to the point where either no mating occurs between populations or mating results only in unviable or sterile offspring. According to Roger Butlin, to understand how reproductive isolation is achieved to allow new species to appear, we need to take into account three main categories of driving forces: the mutation order process, which occurs when two populations evolve genetic differences due to the spread of new mutations that happen to occur only in one population or the other.  Divergent selection is where natural selection operates in different directions in two environments, such as favouring thick-shelled snails in the presence of crab predation but thin-shelled snails where crabs are absent. Reinforcement is the evolution of stronger assortative mating (the tendency of individuals in one population to reproduce with other individuals from the same population, rather than from a different population) as a result of low fitness of hybrid offspring produced by between-population matings.

I started the interview asking him to clarify these aspects.

You consider there are three main processes acting for reproductive isolation, so favouring speciation; these are: mutation order, divergent selection and reinforcement. Which one do you thing is the most important?
It is a really difficult question to answer. I think that what is critical under the divergent selection option is that the changes that are happening in one population could not happen in the other population, because the environment is different in some way. Whereas the idea behind mutation order is that the changes that happen could happen in either population. So either they happen by drift or they happen by response to natural selection, but it is a selection in the same direction, or it is some sort of internal selection if you like, like sexual conflict or genomic conflict, selfish-gene type processes. So, genotype fixation processes could happen in either population, it is just a question whether the relevant mutation arises.

So, in the mutation order process there is no local effect for each population
Yes, that is right, that is the key to the difference between those two. And traditionally people have thought that mutation order processes only happen in spatial isolation. Whereas divergent selection can definitely happen in the presence of gene flow, can overcome gene flow. But actually I think that opinion is changing as well. Because it is clear that if you have a continuous population over a large area, what happens in one part of the population is actually quite independent of what happens elsewhere. So a mutation can happen and start to spread from one part of the range, and a different mutation from another part of the range. And both sides can turn out to be incompatible when they meet somewhere in the range. So the mutational processes are not restricted to things where there is some sort of strong physical barrier. And that means that it is even more difficult to tell the two apart, really. So I think which of those two is contributing most is a completely open question. Reinforcement is more difficult. I think a lot of effort has gone into trying to demonstrate reinforcement over a very long time and there are still a rather small number of good examples. So my suspicion is that reinforcement is not a major contributor to speciation compared to the other two. At least at the early stages. Reinforcement may be important as a sort of finishing off step in speciation. But the bulk of the isolation has to come from the other processes.

Reproductive isolation is considered a central point in promoting speciation events. Yet also hybridization can result in new genotypes founding new evolutionary lines. Does not this look like a paradox?
Yes. I think it is really important to think of speciation as a very long process in which there may be phases during which there is hybridization, and phases where something prevents that, either a physical barrier or some sort of reproductive isolation. That can come and go over time. It is very unlikely that we start with no reproductive isolation and progress smoothly through increasing reproductive isolation to the point where there is no gene flow. More often there are eventually phases when there is not much gene flow and there is a chance for hybridization and things get mixed up a bit. That may provide new evolutionary opportunities. And then there is another phase of increasing isolation. Sooner or later we reach a point where complete isolation is achieved.

butlin2Still, reproductive isolation is more commonly contributing to speciation than hybridization, right? Or is it just from time to time that we have these effects of hybridization, breaking the barriers of reproductive isolation?
Yes, hybridization always tends towards breaking down reproductive isolation, apart from the way it contributes to reinforcement. You have to have hybridization in order to have reinforcement; otherwise there is no selection pressure. Otherwise hybridization is always working against reproductive isolation, but it may be working for adaptation. So, what people call adaptive introgression. It can bring new alleles or new allele combinations. And occasionally may produce these so-called hybrid species where a hybrid, or a population of hybrid origin, let’s say, can invade a new environment not available to either parent. And then sometimes that can result in a new species.

Maybe that is a fast process that under some circumstances can appear when a new environment allows such colonization...
Yes. It happens much more easily in plants because of the way it combines with polyploidy [a polyploid organism has more than two sets of homologous chromosomes. For instance, humans have two sets, one originating from the mother, another from the father, they are diploid. Polyploidy involves extra copies of the genome. Allopolyploid organisms are polyploids whose sets of chromosomes originate from different species i.e. by hybridization]. Allopolyploidy is a really good way in which plants can invade new environments. And take advantage of new combinations of genes. Whereas in animals (or in plants) without polyploidy it is much more difficult for that new hybrid combinations of genes to get stabilized in some way and protected from hybridization again back with the parents.

We know it is less common in vertebrates. But are we sure it is not happening more frequently in less studied animals, such as invertebrates?
Not completely sure, I mean, there is certainly evidence now for genome duplications of some sort in many animal lineages over long periods of time. And one of the main barriers to polyploidy in animals is sex determination systems (sex chromosomes). We know now that sex determination varies enormously among animal groups. Some animal groups do not have a big barrier to polyploidy.

What about asexual organisms? There are examples of very successful ones. Do we need sexual reproduction at all in the evolutionary play?
I think it is fairly clear that sexual reproduction is advantageous, that it is favoured by selection in some way. Maybe, it is not very clear exactly what it is that seems to favour sexual reproduction so widely. But there are still many paradoxes about sex and one of which is why we need so much of it. Occasional sex seems to do pretty much as good a job at creating new gene combinations and speeding up evolutionary response as having sex every generation.

Roger Butlin during his lecture

Roger Butlin during his conference at the Cavanilles Institute of Biodiversity and Evolutionary Biology. / Irene Yuste


Maybe there is no easy way to have sex just from time to time. If you have sex, then you must keep on with sexual reproduction.
There may be a case there for what people call lineage selection. The problem is that if you alternate sexual and asexual reproduction, then in the short term the asexual phase can increase in frequency and lose sex. Then the whole lineage may go extinct at some point when the environment changes. Whereas a lineage that is completely committed to sex does not risk being overtaken by asexual lineages in the short term and so can persist in the long term.

Darwin showed speciation processes related to geographic isolation as a major point in Origin of Species. However you say we must forget about discussing the importance of sympatric (in the same place) vs. allopatric (in separated areas) speciation. Why?
There was a polarization; that was the problem. They were considered two completely different process: sympatric and allopatric, or speciation with gene flow and speciation without gene flow. I think what we need is to take account of the fact that many different spatial structures are possible. Actually, most species exist with some isolation by distance. Some patchiness, so with low local connection between some populations, with incomplete barriers to gene flow in space. But also we need to think about the changes in that spatial layout over the time course of speciation. If you think about speciation as being protracted over thousands or tens or even hundreds of thousands of generations, then, during that time it is very unlikely that diverging groups of populations will maintain the same sort of spatial relationship. Stronger physical barriers between them sometimes, then other times when they are more in close contact. It is a dynamic process. I think it is very unlikely, unusual, that speciation happens completely without any gene flow.

But we still see this higher frequency of endemisms appearing on islands
Sure, it is certainly true that all the processes that underlie the origin of reproductive isolation get easier the lower the level of gene flow. So divergent selection is more effective if gene flow is low. And mutation order processes are more likely the less gene flow there is. Reinforcement needs some gene flow in order to produce the selection, but if gene flow is low, then the response is easier. So everything depends on the extent of gene flow. It is just that we cannot divide it up into these two categories of sympatric and allopatric speciation. But there were many debates about whether a particular case was really sympatric or not, which was just a waste of time.

At the end of the first half of the 20th century, with the merge of the genetic mechanisms discovered by Mendel with the idea of natural selection by Darwin, biologists built the so-called New Synthesis of evolutionary theory. What do you think is the most important advancement in our knowledge of how evolution works since the New Synthesis?
That is tough, to come up with the single most important… One that comes to mind is kin selection in all its forms. In a sense, that is not entirely new. It was suggested before the New Synthesis by Haldane but it was not really understood. 1964 was really the basis of kin selection. That is probably the most fundamental advance in understanding beyond Darwinian natural selection. Since the 1940s, if you are taking that as the New Synthesis. Because it opened up understanding of a whole range of processes or observations which are difficult to understand in straight natural selection terms, such as cooperative behaviour and altruism. It led on to new ways of thinking about how evolution happens where there are individuals interacting. So the whole game theory view of behavioural evolution was dependent on Hamilton’s insight, which was taken forward by people like Maynard-Smith. So I think kin selection meant a whole area of understanding of evolution, which was not in the New Synthesis.

butlin4Moving on now to some general interest related to human impacts… Are we strongly modifying evolutionary patterns through our transport of species (invasive organisms, pests, viruses) and impacts worldwide (e.g. antibiotics, pesticides)? Do you think this will make a major effect on evolution in the whole planet?
There is a good argument that the change that is happening now is faster than any change that has ever been in the past. Well, catastrophic events such as the meteorite impact at the end of the Cretaceous produced very dramatic changes in a short evolutionary time scale. But we are doing some sort of things that are comparable to that now. Traditionally, the slowness of evolution has been emphasized more than it should have been. Actually, adaptation can occur very rapidly. So, in a sense, I think the natural system, if you like, probably can respond to these very rapid changes that we have induced. I do not think we can expect things to stay the same in biodiversity terms. But I am not a catastrophist in a sense that I do not think the whole system is going to collapse because we are pushing too hard. There is a lot of research at the moment on what people call the limits to adaptation, i.e. how fast can environment change and populations keep up there without going extinct. And that can be both in time and in space. Many organisms exist on environmental gradients, and they seem to have boundaries somewhere on those gradients. We only have a beginning of an understanding of why they do not evolve beyond their current limits. That is very relevant now because those gradients are moving. We are pushing them along, and we need to know whether the response to that is going to be that the species go extinct or moved to different parts, or they would be able to adapt in situ.

On the other side, we are promoting gene flow among different genetic backgrounds, in particular in human populations. Does this mean that we are acting against speciation, homogenising our genetic pool?
In general, globally, there is far more movement now and far more contact and interbreeding, if you like, between people, between groups that were ethnically separated in the past. So I think the chances of there being new subdivisions of human populations or forming new races or species is very unlikely, since there is so much movement. However, it is very clear in some parts of the world that, despite groups of people living very close to one another, they still do not interbreed freely. We think of it, you know, from a purely biological perspective. You can see these groups remaining remarkably distinct despite the long-term contact. For instance in Botswana we have the San (Bushmen) people, and they are still a very distinct group of people, compared with Tswana, which is the dominant ethnic group in Botswana, despite hundreds, maybe thousands, of years of contact. Sometimes these differences evolved in the past because of spatial separation. Historically mostly people did not move around very much, so it was possible for quite strong divisions to appear. We of course do not know if that was mutation order or divergent adaptation, probably a bit of both. Those barriers apparently break down remarkably slowly when people get in contact. On the other side, the human population is depauperate in genetic variation compared with many species, and so mixing is probably a good thing if we need to adapt to a changing environment.

So many different ways to proceed with speciation… Is it not possible to see a general picture of which are the main mechanisms of evolution, or rather, that is the picture, i.e. that evolution of life has itself many ways to carry on?
Yes, and that is a criticism you can make of many types of evolutionary biology. That actually every speciation event is going to be unique. But there are many generalizations, and that is partly why I am keen to promote the idea that for reproductive isolation, there are many specific ways in which it can happen, but they actually fall into a small number of general possibilities. So we can hope to get some idea of broad patterns. I think it is about the individual processes, and the way they fit into those categories I spoke about yesterday. Because otherwise speciation research, like some other topics of biology, would descend into just collecting different examples and saying they happen in different ways. That is partly why I did not like the old sympatric speciation and allopatric speciation idea, and I do not like the idea of «ecological» speciation as a category (this is the way some people call divergent selection; I think the ecological part of that is really only a part of divergent selection, as it can include divergent sexual selection, not just ecological speciation). Because you just get an increasing number of those badges; then you have chromosomal speciation, hybrid speciation… It is not helpful to try to slot events into those categories. It is more helpful to say we have these basic underlying processes, probably each of them contributes to most individual speciation events, but you need to know how they interact, in which way one is more important than another.

© Mètode 2016