Royaume-Uni
Prix Balzan 1991 pour la génétique et l'évolution
Problems in Evolutionary Biology – 15.11.1991 (anglais)
The central theory in biology is Charles Darwin’s theory of evolution by natural selection: without it, nothing in biology makes sense. Yet there are certain almost universal features of living organisms that, at first sight, are not what one would expect to find in an organism that evolved by natural selection. These features have been central in evolutionary biology during the past thirty years.
The first is the phenomenon of ageing. An organism that did not deteriorate with age would leave more descendants: why, then, is senescence so widespread? The likely answer, first proposed by G.C. Williams, is that changes that confer greater fitness on an animal when young may carry the inevitable price of deterioration when old. A rather trivial example – but one that concerns all mammals – concerns the teeth. Mammalian teeth are specialized far stabbing, cutting, grinding, and so on, and the teeth in the upper and lower jaws fit neatly with one another. This would not be possible in an animal like a crocodile whose teeth were continually falling out and being replaced. In fact, mammals replace their teeth only once. This carries with it the consequence that, when the second set of teeth wear out, the animal cannot feed. An evolutionary change, single tooth replacement, increases fitness when young, but causes death when old.
Of course, if teeth were the only cause of ageing, human life could easily be prolonged. But there are other changes that cannot so easily be pre vented. Por example, the nerve cells in the brain do not divide in the adult. Presumably a brain works better if its cells are not continually dividing. But cells die and cannot be replaced. If nothing else killed us, in time this would.
This raises a question that people are still arguing about. Deteriorative changes in various organs are observed in humans after about seventy years: in mice, very similar changes occur after only two years. Does this mean that there is some single fundamental process, running faster in mice than in men, responsible for deterioration in the brain, heart, arteries, skin, liver, and so on? As an evolutionary biologist, I doubt it. Even if these changes have different causes, they would be synchronized by natural selection. Suppose, for example, that in some species the heart deteriorated much sooner than any other organ. There would be selection favouring any change in the heart that made it last longer, but no comparable selection on the brain, skin, and so on. Tue consequence of such selection would be the synchrony of ageing changes in different organs. This, of course, is bad news for those who would like to prolong human life: they may not have one problem to solve, but many.
A second puzzle that has attracted a lot of attention recently is cooperative behaviour. Natural selection acts on individuals: it has little power to increase the fitness of groups. Why, then, should worker bees sacrifice their own reproduction to the good of the colony? Why should animals often respect prior ownership of a resource? For example, in territorial animals, intruders usually give way to the territory owner without a fight, even if they have no territory of their own.
Two main approaches have been taken to these problems. The first, pioneered by W.D. Hamilton, is to take genetic relatedness into account. The members of an insect colony are related. Hence a gene that causes a worker bee to act in an apparently altruistic manner may also be present (or, more precisely, an exact copy may be present) in the reproductive members of the colony. Hence an altruistic worker bee may be ensuring the propagation of its own genes. Mathematical analysis has shown that, if interacting individuals are related, this favours the evolution of cooperation. It is not an accident that complex animal communities are composed of relatives.
The second approach has been to ask whether the apparently self-sacrificing behaviour may not, in fact, increase the reproductive success of the individual. Clearly, this cannot explain worker bees, but it does explain respect for territorial ownership. To analyse such problems, it proved useful to borrow the concept of a « game » from the social scientists. Essentially, a game is any interaction in which the best thing to do depends on what one’s opponent is doing. In the social sciences, following von Neumann and Morgenstem, games are analysed on the assumption that individuals act rationally, and can assume that their opponent will do likewise. In evolutionary game theory, the concept of rationality is replaced by Darwinian selection. The result is the concept of an « evolutionarily stable strategy », or ESS. We imagine an evolving population of individuals adopting different strategies and producing offspring adopting the same strategy as themselves. An ESS is then defined as a strategy such that, if all members of a population adopt it, no alternative, mutant, strategy can invade the population. The idea was first developed when attempting to understand the apparently ritual behaviour sometimes seen during animal fights. If an animal can inflict serious injury on its opponent, why doesn’t it? Rather surprisingly, the mathematical technique developed to analyse this problem has proved to be of wide application. Topics that have been treated by evolutionary game theory include the ratio of the number of males to females, hermaphroditism, parental care, territorial behaviour, the formation of alliances, the growth of plants, and, oddly but effectively, the evolution of viruses and transposons.
Of all the features that are puzzling for a Darwinist, perhaps the most puzzling is sex. We are so used to the association between sex and reproduction that we forget that, at the cellular level, these two processes are the precise apposite of one another. In reproduction, one cell tums into two: in the sexual process, two cells turn into one. As Darwinists, we expect natural selection to favour reproduction: why, then, has it led to the opposite? The paradox is perhaps best illustrated by the so-called »two-fold cost of sex ». Imagine an animal – say, a marine fish – with no parental care. Suppose that there arises, in such a population, a parthenogenetic female: that is, a female that does not mate, and produces only daughters like herself. She will, on average, lay the same number of eggs as a sexual female, but all her eggs will develop into parthenogenetic females, whereas only half the eggs laid by a sexual female will develop into females – the rest will become males. Hence, other things being equal, the frequency of parthenogenesis in the population will double in every generation.
Given this twofold advantage of abandoning sex, why do the great majority of species continue to reproduce sexually? The question is easier to ask than to answer. Clearly, the answer must lie in the phrase « other things being equal ». For some reason, other things are not equal, but why not?
One answer is that sex is needed for the repair of damaged DNA. It probably is true that many of the enzymes that are today used in genetic recombination (e.g. the recombination of chromosomes during meiosis) first evolved because of their role in DNA repair. But why is sex required? The idea is that, if a DNA molecule suffers damage to both strands of the double helix, it can be repaired only if an undamaged molecule, with the same original base sequence, is present in the same cell. Otherwise, there is no source of information that can be used during repair. But this seems to be a reason why cells should be diploid (that is, have two copies of each chromosome), but not why they should be sexual.
It seems to me more likely that the evolutionary advantages of sex arise because the offspring of sexual reproduction have genes from both parents. Why should this help? There are two main theories. One answer is that a sexual population is better able to cope with harmful mutations, not by DNA repair but by producing recombinant offspring that Jack the harmful mutations present in their parents. I have called this the « engine and gear box » theory: it says that you can make one functional motor car from two non-functional ones, one of which has a broken engine and the other a broken gear box.
The other main theory – and this is the one this I would like to be true – is that sexual reproduction makes possible more rapid evolution. A popular version of this theory says that species must evolve rapidly to es cape from their enemies – in particular their parasites – which are also evolving. We need sex because of the arms race between hosts and parasites, or between predators and prey. The problem, of course, is whether selection in this context is sufficient to outweigh the twofold cost of sex.
What of the future of evolutionary biology? I do not think it possible to predict the future course of science. The important advances are made by people who set off in a direction different from that taken by the rest of us. This makes science hard to predict, and, more important, impossible to plan. All the same, I will speculate about the future.
During the past decade, evolutionary biology has been much influenced by technical progress in molecular biology, in particular the availability of DNA sequences. Taxonomy is being revolutionised by these molecular data, both on the large scale, in determining the relationships of the major groups of organisms, and on the small scale, as in recent studies of the relationships of the human races. It is now practicable to sequence a number of copies of the same gene from closely related individuals. This can tell us a lot about the genetic structure of populations. At present, I am particularly interested in the light this is shedding on the role of « parasexual » processes in bacteria. It is becoming clear that bacteria can and do exchange genes.
Ever since I graduated, people have been saying that a study of development will shed new light on evolution: I have even said it myself. So far, it has not happened. The reason is that we do not understand development. But in the last few years there have been signs of a breakthrough in developmental biology, particularly, perhaps, by the isolation of genes that have specific effects on the early development of the fruit fly Drosophila. If we really do make a breakthrough in understanding development, it will indeed have great significance for evolution, but as yet we are still waiting.
One last suggestion. With the development of fast computers, people are beginning to use programs that mimic natural selection – so-called » genetic algorithms » – to solve practical problems: for example, how to design an aircraft wing, or a power distribution system. Others are watching the evolution of « artificial life forms » within the computer. So far, computer scientists have probably got more inspiration from biologists than we have from them. But it may not stay like that. I think that biologists at least ought to watch what the computer scientists are up to.