by Nick Lane
By 2002, Wallace and colleagues were beginning to look into the matter more seriously, and signalled their outlook in some thoughtful discussion papers, but it was not until 2004 that they finally found proof. The idea is breathtakingly simple, and yet holds important implications for human evolution and health. The mitochondria, they said, have two main roles: to produce energy, and to produce heat. The balance between energy generation and heat production can vary, and the actual setting might be critical to our health. Here’s why.
Much of our internal heat is generated by dissipating the proton gradient across the mitochondrial membranes (see page 183). Since the proton gradient can either power ATP production or heat production, we are faced with alternatives: any protons dissipated to produce heat cannot be used to make ATP. (As we saw in Part 2, the proton gradient has other critical functions too, but if we assume that these remain constant, they don’t affect our argument.) If 30 per cent of the proton gradient is used to produce heat, then no more than 70 per cent can be used to produce ATP. Wallace and colleagues realized that this balance could plausibly shift according to the climate. People living in tropical Africa would gain from a tight coupling of protons to ATP production, so generating less internal heat in a hot climate, whereas the Inuit, say, would gain by generating more internal heat in their frigid environment, and so would necessarily generate relatively little ATP. To compensate for their lower ATP production, they would need to eat more.
Wallace set out to find any mitochondrial genes that might influence the balance between heat production and ATP generation, and found several variants that plausibly affected heat production (by uncoupling electron flow from proton pumping). The variants that produced the most heat were favoured in the Arctic, as expected, while those that produced the least were found in Africa.
While this seems no more than common sense, the connotations conceal a twist worthy of a murder mystery. Recall from Part 4 (page 183) that the rate of free-radical formation doesn’t depend on the speed of respiration, but rather on how fully the respiratory chains are packed with electrons. If electron flow is very sluggish, because there is little demand for energy, electrons build up in the chains and can escape to form free radicals. In Part 4, we saw that a fast rate of free-radical formation can be reduced if electron flow is maintained down the chains—and this can be achieved by dissipating the proton gradient to generate heat. We compared the situation with a hydroelectric dam on a river, in which the overflow channels prevent flooding. The pressing need to dissipate the proton gradient may have overridden its wastefulness, and given rise to endothermy, just as the need to prevent flooding may override the waste of water through the overflow channels. The long and short of it is that raising internal heat generation lowers free-radical formation at rest, whereas lowering internal heat production increases the risk of free-radical production when at rest.
Now think what is happening in Africans and (let’s say) Inuit. Because Africans generate less internal heat than Inuit, their free-radical production ought to be higher, especially if they overeat. According to Wallace, Africans can’t burn off excess food as heat as efficiently as do Inuit, so if they eat too much they will generate more free-radicals instead. This means that they ought to be more vulnerable to any diseases linked with free-radical damage, such as heart disease and diabetes, and indeed this is the case. Africans in the United States eating an American diet are notoriously susceptible to diseases like diabetes. Conversely, Inuit ought to burn off excess food as heat, and so should suffer much less from heart disease and diabetes, and again this is found to be true. Of course there are other reasons too (such as intake of oily fish, etc.) so these conclusions are necessarily tentative. However, if there is some truth in these ideas, then another logical connotation should also be true, and there is a hint that it is: any peoples adapted to arctic climates should be more vulnerable to male infertility.
The reasoning is exactly the same. Arctic peoples divert less of their food into energy, and more into heat. This may not matter in most circumstances (they just eat more) but it matters in one instance: sperm motility. Sperm are powered by their mitochondria as they swim towards the egg, and because there are fewer than a hundred mitochondria in each cell, sperm cells are uniquely dependent on the efficiency of the remaining few—and uniquely vulnerable to energy failures. If these mitochondria fritter away their energy as heat, the sperm are more likely to be dysfunctional, and the man to suffer from asthenozoospermia. This means we should see patterns of male fertility that depend not on male genes, but on mitochondrial genes passed down the maternal line. In other words, male infertility should be inherited at least partly from the mother, and ought to vary according to mitochondrial haplogroup. One recent study has confirmed this to be true in Europeans: asthenozoospermia is more common in people of haplogroup T (widespread in northern Sweden) than in people of haplogroup J (more widespread in southern Europe). Whether it is also true of the Inuit, I don’t know: unfortunately I can’t find any data on the incidence of asthenozoospermia among the Inuit.
Altogether, these twisted relationships show that mitochondrial genes are indeed subject to natural selection.3 The precise weighting depends on factors that include energetic efficiency, internal heat production and free-radical leakage, all of which affect our overall health and fertility, and our capacity to adapt to diverse climates and environments.
When coupled with the other findings we have discussed in this chapter, the orthodox position looks tarnished. Mitochondrial genes can be inherited from both parents, albeit rarely; they recombine, if very rarely; they mutate at variable rates according to circumstances, questioning the accuracy of imputed dates; and they are unquestionably subject to natural selection. If these unexpected findings don’t topple the edifice of human prehistory, do they at least give a more cohesive understanding of mitochondrial inheritance? More pertinently, can these findings explain why we have two sexes?
15
Why There Are Two Sexes
In Chapter 13, we saw that the deepest biological difference between the two sexes relates to the inheritance of mitochondria. The female sex specializes to provide the mitochondria (100 000 of them in humans) in the large, immobile egg cells, while the male sex specializes to eliminate mitochondria from tiny, motile sperm cells. We looked into the reasons for this strange behaviour, and found that it often seems to boil down to conflict between genetically different populations of mitochondria. To restrict the opportunity for conflict, mitochondria are usually inherited from only one of two parents. But we also met a number of exceptions to this simple rule, including fungi, trees, bats, and even ourselves. In Chapter 14, we took a close look at ourselves to see how well the copious human data support the conflict argument. These data are controversial and arouse high passions, for they concern our own prehistory, but the coherent picture that is slowly emerging from these disputes gives fascinating insights into the deeper reasons for the difference between the two sexes. In this chapter, we’ll try to draw these insights together to come up with a more satisfying answer to the puzzle of two sexes.
The essential facet of the conflict argument is that dissimilar populations of mitochondria can compete with each other for succession, and the only way to prevent such conflict is to ensure that all the mitochondria inherited in the egg are genetically identical. The only way to guarantee they are all the same is to make sure they all come from the same source—the same parent. Mixing is said to be fatal. The belief that mitochondrial mixing (heteroplasmy) is simply not tolerated underpins the mantra of human mitochondrial population genetics. According to this mantra, the male mitochondria are swiftly eliminated from the egg, and not passed on to the next generation. This means that mitochondria are passed down the maternal line by asexual replication only. Thus, mitochondrial DNA remains basically unchanged, as there is no possibility of recombination. Even so, there is a gradual divergence in the mitochondrial DNA sequence of different populations and races, as o
ccasional neutral mutations accumulate over thousands and tens of thousands of years. These accumulated differences supposedly sit faithfully in the genome, as natural selection is said not to apply to mitochondrial genes, or at least to the ‘control region’ that doesn’t code for proteins. Without purifying selection, the mutations are not weeded out of the genome, and so remain there forever after, mute witnesses to the flow of history.
The lessons from human evolution muddy all these tenets, and suggest there is a deeper mechanism at work. That is not to say that genomic conflict is wrong, but merely part of a larger picture. Let’s rake the mud. We have seen that mitochondrial recombination does indeed occur, probably very rarely in human mitochondria, but more commonly in other species, such as yeast and mussels. It is not the taboo we had once thought. Moreover, the condition for recombination, heteroplasmy (a mixture of dissimilar mitochondria), is far more common than is assumed by the selfish conflict model. Some degree of heteroplasmy is found in 10 to 20 per cent of humans, and it is also common in many other species. Next, we have seen that there is a discrepancy in the rate of change of mitochondrial genes. The mutation rate of mitochondrial DNA in families suggests one mutation every 800 to 1200 years, whereas the long-term divergence of races suggests a mutation rate of one every 6000 to 12 000 years. The disparity can be explained if many of the variants are eliminated by natural selection. Counter to the usual mantra, there is now good evidence that natural selection does indeed act on mitochondrial genes, in ways that are subtle and pervasive.
So why are there two sexes? Think about the mitochondria. They are not independent entities but part of the larger system of the cell. Mitochondria contain proteins that are encoded by two different genomes. Genes in the nucleus encode the vast majority, some 800 proteins, whereas the handful of mitochondrial genes encodes the rest, a mere 13 proteins, all of which are critical subunits of the large protein complexes of the respiratory chains. The mitochondrial-encoded proteins are essential for respiration. It’s this necessary interaction between two genomes, the mitochondrial and nuclear, that explains the need for two sexes. Let’s see why.
The function of the mitochondria depends critically on the interaction of the proteins encoded in the nucleus with those encoded in the mitochondria. This dual control system is no frozen accident: it evolved that way, and is continuously optimized, because this is the most effective way of meeting the needs of the cell. As we saw in Part 3, the mitochondria retained a handful of genes for a positive reason: a rapid-reaction unit of genes in the mitochondria is necessary to maintain efficient respiration. In contrast, the genes that could be transferred successfully to the nucleus generally have been; there are many advantages to them being there, not least to quell the independence of the troublesome mitochondrial guests.
Any misalignment between proteins encoded in the nucleus and proteins encoded in the mitochondria holds potentially catastrophic consequences. The fine control of mitochondrial function influences not just energy availability, but other matters of life and death, such as apoptosis, fertility, sex, endothermy, disease, and ageing. But how well does dual genomic control work? Babies are a miraculous proof of the marvellous harmony that nature has attained, but perfection comes at a price. Many couples strive for years to have children, and infertility is common. Even for fertile couples, early (usually subclinical) miscarriage is the rule rather than the exception: some 70 to 80 per cent of embryos spontaneously abort in the first weeks of pregnancy, and the would-be parents may never notice. Many of these early losses occur for reasons that are still obscure.
The problem might often relate to the interaction of the two genomes—the need for nuclear gene products to work in cohorts with mitochondrial gene products. In mammals, the mutation rate in mitochondria is fast, on average 20 times faster than the nucleus and in places 50 times faster, owing to the proximity of mitochondrial DNA to mutagenic free radicals leaking from the respiratory chains. That’s not all. In the nucleus, genes are shuffled by sex every generation. Because the genes encoding mitochondrial proteins sit on different chromosomes, they are re-dealt as a different hand each generation. The outcome is a serious mix-and-match problem. In the respiratory chains, proteins dock on to each other with nanoscopic precision. To give a single example, cytochrome c (encoded in the nucleus) must bind to a critical subunit of cytochrome oxidase (encoded in the mitochondria) to pass on its electron. If the binding is not exact, the electron is not transferred and respiration grinds to a halt. When electrons aren’t passed on down the chains they form free radicals instead. These oxidize the membrane lipids and release cytochrome c to induce apoptosis. From this perspective, the unanticipated role of cytochrome c in apoptosis begins to look not like an oddity but a necessity. It puts a quick end to cells with inefficient respiration, due to a mismatch between the nuclear and mitochondrial genes.
The requirement for a close match means it is critical for mitochondrial and nuclear genes to co-adapt to each other in synchrony, otherwise respiration cannot work. In principle, a failure to co-adapt leads straight to an early death by apoptosis. Direct evidence of co-adaptation is mounting. If the DNA from mouse mitochondria is replaced with DNA from rat mitochondria, protein transcription proceeds normally, but respiration ceases because the rat mitochondrial proteins can’t interact properly with the mouse proteins encoded in the nucleus. In other words, control of respiration is more stringent than control of DNA transcription and translation into new proteins. Slighter differences occur within species, but even small mismatches between mitochondrial and nuclear genes affect the speed and efficiency of respiration. Importantly, the evolution rate of cytochrome c mirrors that of cytochrome oxidase over evolutionary time, even though the underlying rates of change are more than 20-fold different. Presumably, any new variants that lower the efficiency of respiration are eliminated by natural selection. The imprint of selection is betrayed by the fact that many of the sequence changes that do persist are so-called neutral substitutions, which is to say they don’t alter the sequence of the protein. The ratio of neutral substitutions to ‘meaningful’ substitutions is much higher than normal in mitochondrial genes, which implies that mutations that alter meaning are eliminated by natural selection. There are other hints that meaning is preserved at all costs. For example some protozoa such as Trypanosoma, actually edit their RNA sequences to retain the original meaning, despite changes in DNA sequence. Similarly, the fact that mitochondria harbour exceptions to the universal genetic code can be explained as an attempt to retain the original meaning despite changes in DNA sequence.
Taking all this into consideration, we can say that two sexes are needed because the dual genome system demands a close match between mitochondrial and nuclear genes. If the match is not good, respiration is impaired, and there is a much greater risk of apoptosis and developmental abnormalities. The precision of the match is continuously strained by two factors—the far higher mutation rate of mitochondrial DNA, and the randomization of new nuclear genes by sex in every generation. To ensure the match is as perfect as it can be in each generation, it is necessary to test a single set of mitochondrial genes against a single set of nuclear genes. This explains why the mitochondria need to come from just one parent. If they came from two parents, then there would be two sets of mitochondrial genes paired with one set of nuclear genes. This is like pairing two women of dissimilar build with the same man in a three-way ballroom dance. However accomplished they might be as individual dancers, the flailing threesome is likely to fall over. To dance a true metabolic waltz requires two partners—one type of mitochondria with one set of nuclear genes.
There are two important connotations to this answer. First of all, it easily encompasses existing models, while explaining the apparent aberrations noted in the studies of human evolution. To match a single mitochondrial genome with a single nuclear genome requires that (in general) the mitochondrial genome should be inherited from a single parent, hence the propensity for uniparental in
heritance. If mitochondria are inherited from both parents, then the efficiency of respiration is likely to be impaired, as the two populations are obliged to dance with the same nuclear partner. This situation is exacerbated if dissimilar mitochondrial genomes compete, as in the selfish-conflict theory. Notice, however, that some degree of heteroplasmy and recombination, is feasible as it might at times provide the best genomic match. The unexpected findings from human evolution—heteroplasmy, recombination, and selection—can be explained in this way. The most important aspect is not actually the ‘pureness’ of the population, but rather how effectively the mitochondrial genes work against the nuclear background.
Secondly, the dual-control hypothesis gives a positive basis for natural selection. One difficulty with the selfish-conflict theory is that selection can only act to eliminate the negative consequences of genome conflict. However, we’ve seen that heteroplasmy exists in many circumstances without obvious competition between the two genomes—for example in angiosperms, some fungi, and bats. If the detrimental effects of genomic competition are limited, then why does natural selection generally favour uniparental inheritance? It would do so if uniparental inheritance were positively beneficial most of the time, rather than merely mildly detrimental part of the time. The dual-control theory gives a good reason why this would be the case: the fittest individuals generally inherit the mitochondrial DNA only from the mother, as this enables the best match of nuclear and mitochondrial genomes. And if the fittest offspring tend to inherit their mitochondrial genes from only one of two parents, we have satisfied the condition for two sexes: the female sex supplies the mitochondria, the male sex generally does not.