Two people are dancing a waltz, and it is not going well. One is tall and the other short; one is graceful, the other flat-footed; and both are stepping to completely different rhythms. The result is chaos, and the dance falls apart. Their situation mirrors a problem faced by all complex life on Earth. Whether we’re animal or plant, fungus or alga, we all need two very different partners to dance in step with one another. A mismatch can be disastrous.
Virtually all complex cells – better known as eukaryotes – have at least two separate genomes. The main one sits in the central nucleus. There’s also a smaller one in tiny bean-shaped structures called mitochondria, little batteries that provide the cell with energy. Both sets of genes must work together. Neither functions properly without the other.
Mitochondria came from a free-living bacterium that was engulfed by a larger cell a few billion years ago. The two eventually became one. Their fateful partnership revolutionised life on this planet, giving it a surge of power that allowed it to become complex and big (see here for the full story). But the alliance between mitochondria and their host cells is a delicate one.
Both genomes evolve in very different ways. Mitochondrial genes are only passed down from mother to child, whereas the nuclear genome is a fusion of both mum’s and dad’s genes. This means that mitochondria genes evolve much faster than nuclear ones – around 10 to 30 times faster in animals and up to a hundred thousand times faster in some fungi. These dance partners are naturally drawn to different rhythms.
This is a big and underappreciated problem because the nuclear and mitochondrial genomes cannot afford to clash. In a new paper, Nick Lane, a biochemist at University College London, argues that some of the most fundamental aspects of eukaryotic life are driven by the need to keep these two genomes dancing in time. The pressure to maintain this “mitonuclear match” influences why species stay separate, why we typically have two sexes, how many offspring we produce, and how we age.
Dancing out of step
Here’s the problem: both sets of genes help to create proteins that sit in the mitochondria and carry out one of the most important of chemical reactions: respiration. The proteins strip electrons from our food and pass them along from one to another. They eventually deposit the electrons onto oxygen; this produces water and releases energy. These ‘electron transfer chains’ are the stuff of life, and they only work if the proteins involved are built correctly.
The proteins in the chain are made of different subunits. Some are built using instructions from nuclear genes, while others are built using mitochondrial genes. They different parts must fit together with nanometre precision. Even a small change in their shape will produce botched proteins that fumble their electrons. If fewer electrons make it to the end of the chain, the mitochondria produce less energy. The leaking electrons can also react with oxygen directly to produce destructive molecules called free radicals.
So, cells with mismatched nuclear and mitochondrial genes face a double whammy of less energy and leaking free radicals. This has two important consequences for the evolution of eukaryotes: it creates a barrier between different species, and it favours the evolution of two sexes.
Within a species, the nuclear and mitochondrial genomes have adapted alongside one another so that their protein components seamlessly fit together. These dancers don’t swap their partners easily. If different species mate, they destroy this exquisite co-evolution, which might explain why hybrids encounter so many problems. Respiration is difficult for them, their mitochondria can’t produce any energy, they’re bombarded by leaking free radicals, and many of their cells top themselves. With so many problems, it’s no wonder that many hybrids become sterile or weak, or fail to develop properly at all. This is the price of a mitonuclear mismatch.
Why two sexes?
Mismatches can be easily weeded out by natural selection because every individual has the same mitochondrial genome in all of its cells. Those that match up well with the nuclear genome will survive; those that match poorly will die. This weeding process breaks down if individuals have many different types of mitochondria. In this scenario, the bad matches cancel out the good ones in any individual, and everyone ends up being decidedly average. Natural selection has little to work with.
Over time, individuals with a single mitochondrial genome will do better than those with many. The fittest of them will thrive thanks to natural selection, while their peers stagnate. Lane argues that one of the easiest ways of ensuring that an individual has a uniform set of mitochondria is to have two sexes. One (usually female) hands down an identical set of mitochondria to its young, and the other (usually male) doesn’t. That’s a major difference between the two sexes; some (including Lane) would argue it’s the main difference.
There are species that do things differently, but they are exceptions that prove the rule. Some slime moulds have 13 different sexes, but after mating, they destroy all but one set of mitochondria. Some fungi, like baker’s yeast, inherit mitochondria from both parents, but they are quickly separated so that individual cells only contain one type.
Here’s the gist: mitonuclear mismatches are easier to weed out if individuals test-drive one set of mitochondrial genes against one set of nuclear genes. And having two sexes is an easy way to do that.
The death threshold
The threat of mitonuclear mismatch can also explain the different lifestyles of different species. Mismatches cause a leak of free radicals and cells have two ways of dealing with that. If the leak is fairly minor, the cell can make more mitochondria to compensate. If the leak is severe enough, the cell commits suicide through a process called apoptosis. Lane’s idea is that there’s a threshold that determines which route a cell will take – a level of leakage where it chooses to cut its losses rather than fix the problem.
Different species set their ‘apoptotic threshold’ at different levels. For example, birds and bats need a lot of energy to fly, and their nuclear and mitochondrial genomes must match perfectly. The proteins of their mitochondria have to shunt electrons from one to another quickly and efficiently. Even slight mistakes would compromise their energy levels, and that can’t be tolerated.
So, birds and bats have very low leak thresholds. Even a slight trickle of free radicals betrays the fact that their two genomes aren’t meshing properly – time for their cells to die. Dying cells mean dying embryos, and many are eliminated before they fully develop. Only a precious few would make it through this harsh selection process. Lane thinks that this could explain why these species tend to have low fertility rates and few offspring.
By contrast, a rat has less demanding energy needs, and the electron transfer chain in its mitochondria can afford to be leakier and less efficient than that of a bird. The rat can handle a poorer mitonuclear match, so it sacrifices fewer embryos on the altar of perfection. It follows that rats are also more fertile, and produce larger litters.
Ageing apart
Even well-matched nuclear and mitochondrial genomes don’t stay that way forever. As individuals age, leaking radicals will damage and mutate the mitochondrial genome, ruining its match with the nuclear one, and causing even heavier leaks. This happens, even if the initial stream of radicals is small. As time wears on, the dancers inevitably fall out of step with each other. You can see this if you compare young and old tissues: the young cells will all have genetically identical mitochondria, while those in the old cells will be a mix of different mutants.
As more cells pass the tolerance threshold, more of them die. Tissues that use the most energy, like the muscles and brain, have the heaviest leaks and wear away faster. Meanwhile, the surviving cells experience even greater energy demands. They enter a downward spiral with sweeping consequences: they leak free radicals like sieves; their DNA becomes more fragile; their genes become switched on in different ways; they release chemicals that trigger inflammation. In short, they create the perfect set-up for cancer, heart disease, diabetes, Alzheimer’s and many of the other diseases of old age.
Almost all of the major traits of ageing can be predicted by a growing rift between two genomes, and a widening leak of free radicals. The leak worsens with time, so tissues die, especially gas-guzzling ones. Those that survive are more likely to become diseased. And the fast the leak, the faster all of this happens. This explains why species that tolerate less free radical leaks tend to enjoy longer lives. Consider pigeons and rats: both species are similar in both size and metabolic rates, but pigeons have far lower rates of leaking electrons in their mitochondria. They also live ten times longer.
A simple idea
For now, this is all a grand hypothesis, albeit one that is grounded on a lot of existing evidence. Lane now wants to explore ways of testing his idea. The most obvious first step would be to see if there actually is a lea threshold that varies between cells. It should be straightforward to measure the extent of free radical leaks in cells, and the level that makes them kill themselves.
He also wants to look at species with high energy needs like birds, to see whether a large proportion of their embryos are being lost. He’s also interested in how this applies to humans. “It would be interesting to get data from fertility clinics to see if there are any groups or populations that struggle to conceive,” he says, “and if any of this can be put down to incompatibilities between mitochondria and nuclear backgrounds. Around 40% of pregnancies end in miscarriage and we don’t know why.”
There is compelling majesty to Lane’s idea. At its heart, it is deceptively simple: we have two genomes that need to work together, and you can tell how well they’re doing this by the strength of the free radical leak. From that simple concept, you can logically derive how fitness, fertility and lifespan are linked in different species. You can also predict the process of ageing and the onset of age-related diseases within individuals.
“A lot of this has to be true on logical grounds,” says Lane. “We know that there is co-adaptation between these two genomes and many predictions emerge seamlessly from some simple reflections on that process. The big question is whether it’s important in the greater scheme of things.”
Reference: Lane, N. (2011). Mitonuclear match: Optimizing fitness and fertility over generations drives ageing within generations BioEssays DOI: 10.1002/bies.201100051
Lane, N. (2011). The Costs of Breathing Science, 334 (6053), 184-185 DOI: 10.1126/science.1214012
Image: by Ticipico
More on mitochondria
October 17th, 2011 at 10:42 am
What I don’t understand is, why does a person have _A_ mitochondrial genome? Why not a set of a thousand mitochondrial genomes?
There are about a thousand(?) mitochondria per cell. Why isn’t a person’s mitochondrial DNA that of a population of ~1000 that has had intracellular evolution through many mitochondrial generations (many more than the 20 year human generations)? Yes, the populations are passed down only maternally, but that’s like the case of isolated villages of a thousand people who never intermarry into the neighboring village. Within a village there will still be genetic diversity.
October 17th, 2011 at 11:49 am
This is a great piece. I saw Nick Lane speak on this topic in Oxford a few months ago – his ideas are extremely interesting, and they’re explained very clearly here.
I’m a bit confused about one point, where you imply that mitochondrial genomes evolve faster than nuclear ones because mitochondria are maternally inherited. Why should maternal inheritance cause them to evolve more quickly?
October 17th, 2011 at 11:53 am
Interesting idea, but I see lots of problems with it. One, rats don’t have low energy demands. Comparing it to a hummingbird is rather unfair because everything has lower energy demands than a hummingbird. Birds are extraordinarily diverse and while they do in general have higher metabolisms than mammals of the same size, these sorts of statements cannot be applied across the board. I won’t even get into all the flightless birds that completely throw this sort of statement out the window. Secondly, the leaky free radicals by mitochondria is one of the most important ways endotherms create body heat. All that inefficiency is a feature, not a bug. So birds have even leakier mitochondria than mammals. Reptilians on the other hand have far less leakage because they don’t need to create the excess heat. Else and Hurlburt have done a lot of work looking at this sort of thing. Lane knows this, so I would like to know how he explains this in his hypothesis. I expect it is thus not the amount of leakage per se, but the control over exactly how leaky that is key.
October 17th, 2011 at 12:15 pm
Really fascinating stuff. I am just starting to read ‘Power, Sex, Suicide’, Lane’s book on mitochondria. That was published in 2005, and I’m wondering how much in this paper represents new work since then, since the book seems to promise to cover much of the same ground. (Lane also talks about the free-radical-leak->diseases-of-aging theory in his great book last year, ‘Life Ascending’).
I’m also wondering whether Lane is on his own here, or whether other researchers agree with some of this picture (and which bits). The mitochondrial free-radical theory of aging has been around for decades, hasn’t it? And no-one else has bothered to measure free-radical leak yet?
Like the commenter above, I’m confused about why mitochondrial genomes evolve faster than nuclear ones.
October 17th, 2011 at 12:31 pm
Really interesting piece Ed but I too wonder why mitochondial evolution is faster than nuclear?
Also, I’d say that the molecular fit has to be sub-nanomolar – addition of a single atom (or methyl group), about 0.1 nm, could be enough to disrupt an interaction between molecules. But that’s just a bit of scientific pedantry.
October 17th, 2011 at 1:11 pm
Hi guys, I’m going to break the habit of a lifetime and chip in with a few answers, as Ed has written such a great piece and there are some very good questions here. I don’t normally because I try to avoid getting waylaid into answering questions all the time instead of pushing ahead with work… but anyway here goes:
Why not a set of 1000 genomes? Because what we inherit is close to a clonal population, and the reason for that is that there is a maternal bottleneck during oocyte development, which cuts right down to as few as 10 copies of mtDNA (though this is disputed) and then ramps up to around 100,000 copies. So with a village of 1000 people, it’s as if we just sample from one person each generation. The bottleneck is coupled to uniparental inheritance (mother only) and the effect is a clonal population of mitochondria in most cases. As we age, though, we begin to lose that clonal population as some mutations accumulate.
Why do mitochondrial genes evolve faster? It has nothing to do with maternal inheritance. Nobody knows the answer and it is true of animals and fungi but not of plants. The reason might relate to the proximity of mitochondrial genes to the respiratory chains, where the free radicals leak from in the first place; or it could relate to poor repair systems. But more likely, it reflects some form of adaptation, as the mutation rate appears to be genetically determined and varies between species. Doug Wallace thinks that the fast mutation rate of mitochondrial genes helps animals adapt to different foods and climates. The elimination of damaging mutants in the bottleneck then means that the cost is low. There’s little evidence for that but it’s a neat idea. Plants, incidentally, have a much low nucleotide mutation rate, but have more recombination, so overall there are still problems.
Re energy demands in birds vs mammals, these comments are true, and it is problematic comparing groups across groups. But flightless birds have higher free-radical leak and lower aerobic capacity exactly as predicted. And bats have lower leak and higher aerobic capacity as predicted. S0 the patterns make sense within groups too.
The patterns of uncoupling across groups are not always clear and depend on temperature. Birds do not have greater uncoupling than equivalent mammals; it is in fact very similar (around 20-25% of proton flux). That’s not surprising as flighted birds require high ATP yields for flight. Flightless birds might be expected to have greater uncoupling.
Much the same is true of lizards and cold-blooded creatures. My recollection of Else and Hurlburt’s work (especially their earlier work) is that the degree of uncoupling is quite similar, but the number of mitochondria is very different – mammals have around 5-fold more in visceral organs, and that makes up the biggest distinction in terms of heat.
Am I on my own out there? Of course not. But I’m unusual in thinking that the free-radical theory of ageing is correct, but antioxidants really don’t work. Why not? Because free radicals act as signals. But why use ‘toxic’ signals? Why not something less damaging? I used to think the answer lay in immunology and the stress response, but the demonstration that the oxidative burst in neutrophils does not kill bacteria implies that free radicals are not strictly necessary for that either. So why are they used? This is where my reflections on the role of free radicals in mitonuclear matching come in. I would say that I have pieced together a tapestry of other peoples work, along with some ideas of my own, to come up with a perspective that is at least slightly new, and maybe helpful. Now I need to test it!
Hope that’s helpful, and please forgive me if I don’t get too actively involved in answering questions. To give them the time they deserve is quite time consuming.
October 17th, 2011 at 3:05 pm
I would imagine that there’s probably a level of purifying selection going on within individual cells as well. Individual mitochrondria that are too leaky due to gene mismatch could well get tagged for destruction.
October 17th, 2011 at 9:49 pm
Dr. Lane, thanks for your response. I had forgotten about the large difference between mitochondrial numbers between endotherms and ectotherms. Thank you for the reminder. I retract my objection and look forward to reading more about your hypothesis.
October 17th, 2011 at 9:55 pm
Do free radicals cause a specific type of mutation and has the rate of this mutation been compared between the mitochondrial and nuclear genomes?
October 17th, 2011 at 11:41 pm
Excellent post and great responses from Dr. Lane in the comments section.
The lower mutation rate plants seems most intriguing and I’m guessing that explains why plant species have such a (relatively) easy time hybridizing with each other. But between the role of chloroplast DNA, asexual reproduction, and the uncanny ability of some species to live forever, I can’t even imagine where to begin attacking the problem of how plants fit into the whole aging puzzle.
I do find it curious though why something as vital to life as the electron transport chain is even split between two separate genomes. It seems that life would be spared a lot of trouble had eukaryotic cells just incorporated all the important genes into the nuclear genome(or vice versa) and proceeded with respiration from there. More specifically what purpose does it serve to have certain proteins in the electron transport chain mutate faster than others or has anyone studied if there is even a significant difference in the rates between genes in the chain coded in mtDNA and those in the nuclear DNA?
October 18th, 2011 at 12:40 am
A leading theory (described in Lane’s books) is that a small local genome retained in the mitochondria is necessary for fine-tuned local control of respiration, within that mitochondrion.
Because of the high mutation rate within mitochondria (or perhaps for other reasons) selection pressured favored transferring as many mitochondrial genes as possible to the nucleus (the mechanism for this would simply be lateral gene transfer between mitochondria and nucleus, say, whenever a mitochondrion died and spilled its DNA into the cytoplasm, followed by deletion of the redundant copies of the transferred genes within the other mitochondria).
But a few core genes required for regulating respiration had to remain, along with some of the apparatus needed to translate those genes.
And indeed, if it were possible for all the genes for respiration to be in the nucleus, then, when you think about it, it should have been possible for the respiring ancestor of the mitochondria to have evolved into eukaryotes themselves, with no need for endosymbiosis at all.
October 19th, 2011 at 9:51 pm
Years ago I remember reading (I believe it was a dead tree publication, so no link) precisely the opposite argument–that mitochondria were so vital to the functioning of the cell that the observed rate of mutation was very low. You can tinker with the car’s interior more than its brakes, as it were.
October 19th, 2011 at 11:17 pm
Can anyone on the thread explain how foods rich in “anti-oxidants” address free radicals? Do they give the free radicals an efficient alternative binding option, or do they in some way help prevent the formation of free radicals in the first place? Also, has there been work to correlate the creation of the free radicals to the intensity of cellular respiration in a given species? Earlier, there were some premises about different animals and tolerances to leakage; my interest is in comparing sedentary vs. athletic members of a given species (eg humans). Thanlks in advance. I’ve really enjoyed this article and the commentary has been thoughtful.
October 20th, 2011 at 8:18 am
Both Ed’s post and Nick’s work are amazing. The claim that this explains sex does not however seem to me to work. The first alternative to sexual reproduction would appear to be asexual reproduction and it is at least not obvious why asexual reproduction should involve inheriting several mitochondrial genomes. Indeed, given that there is only one sex in asexual reproduction, it would appear to reduce the risk — all other things being equal. At most, this work would seem to me to explain why mitochondiral DNA is inherited from only one sex.
October 21st, 2011 at 11:44 pm
A possible answer to this question may lie in the mechanics of sexual reproduction.
Sexual reproduction actually consists of 2 steps. The first is meiosis, in which the haploid gametes are generated from diploid precursors. And the second is fertilization, the merger of the two gametes.
In asexual reproduction, at least among unicellular eukaryotes (in which presumably sex first evolved), cell division occurs via mitosis. The nuclear genome is copied, and the cytoplasm is divided. The mitochondria are simply split, randomly, by the cytoplasmic division. Thus if there has been any divergence of mitochondrial genomes from mutants arising from regular mitochondrial divisions, both daughter cells have an equal likelihood of inheriting equally discordant mitochondrial populations.
An interesting and at first counterintuitive aspect of meiosis, is that, although the goal is to divide the diploid genome into two, the first step is actually a duplication of the genome, just as in mitosis (except that there is recombination in meiosis), which turns the nuclear genome tetraploid. And then the parent cell undergoes two rounds of division to produce four haploid progeny.
In males, all four of these progeny go on to become gametes, and presumably would have the same problem with mitochondrial hetergeneity, since their mitochondrial were simply divided among them evenly when the cytoplasm was divided.
However, in females, three out of four of these haploid cells actually degenerate, and only one matures into the gamete. This step suggests a selective bottleneck. Perhaps one reason for the bottleneck is a “trying out” of mitonuclear matching, with only the best match going on to become the gamete.
This, combined with the mitochondria in the male gamete contributing minimally if at all to the fertilized zygote, may be the mechanism for ensuring a good mitonuclear match.
(Note that asexual metazoan species are a little different – typically they still go through meiosis, rather than mitosis, for reproduction. But this could be because asexual metazoans are typically descended from sexual ancestors.)
(I’m not sure what the bdelloid rotifers do in this regard).
An asexually reproducing population could achieve the same effect via purifying selection in large populations. But, critically, it is generally believed that the early eukaryotes, in which sex first evolved, had small populations in which purifying selection would not apply.