by Nick Lane
There is still one step to go: we need to put together a cell with bacterial-style membranes throughout, in other words we need to replace the archaeal lipids of the cell membrane with bacterial lipids. How did this happen? Presumably, if bacterial lipids offered any advantage, such as fluidity, or adaptability to different environments, then any cell that expressed only the bacterial lipids would be at an advantage. Natural selection would ensure that the archaeal lipids were replaced, if such an advantage existed: there was little call for evolutionary ‘novelty’; it was merely a matter of playing with existing parts. It remains possible, however, that some eukaryotes did not go the whole hog. It would be interesting to know if there are still any primitive eukaryotic cells that retain vestiges of archaeal lipids in their membranes. In support of the possibility, virtually all eukaryotes, including fungi, plants, and animals like ourselves still possess all the genes for making the basic carbon building blocks of archaeal lipids, the isoprenes (see page 99). We don’t use them for building membranes any more, however, but for making an army of isoprenoids, otherwise known as terpenoids or terpenes. These include any structure composed of linked isoprene units, and together make up the single largest family of natural products known, totalling more than 23 000 catalogued structures. These include steroids, vitamins, hormones, fragrances, pigments, and some polymers. Many isoprenoids have potent biological effects, and are being used in pharmaceutical development; the anticancer drug Taxol, for example, a plant metabolite, is an isoprenoid. So we haven’t lost the machinery for making archaeal lipids at all; if anything, we have enriched it.
If his theory is correct, then Martin has derived an essentially complete eukaryotic cell via a simple succession of steps: it has a nucleus enveloped by a discontinuous double membrane; it has internal membrane structures; and it has organelles such as mitochondria. The cell is free to lose its cell wall (but not, of course, its external cell membrane), as it no longer needs a periplasm to generate energy. Being derived from a methanogen, it wraps its genes in histone proteins and has a basically eukaryotic system of transcribing its genes and building proteins (see Part 1). On the other hand, this hypothetical progenitor eukaryotic cell probably did not engulf its food whole by phagocytosis—despite having a cytoskeleton (inherited from the archaea or the bacteria), it has not yet derived the dynamic cytoskeleton characteristic of mobile protozoa like amoeba. Rather, the first eukaryotes may have resembled unicellular fungi, which secrete various digestive enzymes into their surroundings, to break down food externally. This conclusion is corroborated by some recent genetic studies, but we won’t look into these here, for too many uncertainties remain.
Why did mitochondria retain any genes at all?
So the transfer of genes from the mitochondria to the host cell is capable of explaining the origin of the eukaryotic cell, without requiring any evolutionary innovations (new genes with different functions) whatsoever. Yet the sheer ease of gene transfer raises another suspicious question. Why are there any genes left in the mitochondria at all? Why were they not all transferred to the nucleus?
There are big disadvantages to retaining genes in the mitochondria. First, there are hundreds, even thousands of copies of the mitochondrial genome in each cell (usually 5 to 10 copies in every mitochondrion). This enormous copy-number is one of the reasons that mitochondrial DNA is so important in forensics, and in identifying ancient remains—from such an embarrassment of riches, it is usually possible to isolate at least a few mitochondrial genes. But by the same token, it also means that whenever the cell divides a vast number of ostensibly superfluous genes must be copied. Not only that, but every single mitochondrion is obliged to maintain its own genetic apparatus, enabling it to transcribe its genes and build its own proteins. By thrifty bacterial standards (which, as we have seen, eliminate any unnecessary DNA post haste) the existence of these supernumerary genetic outposts seems a costly extravagance. Second, as we shall see in Part 6, there are potentially destructive consequences of competition between different genomes within the same cell—natural selection can pit mitochondria against each other, or against the host cell, with no consideration of the long-term cost, merely the short-term gain for the individual genes. Third, storing genes, vulnerable informational systems, in the immediate vicinity of the mitochondrial respiratory chains, which leak destructive free radicals, is equivalent to storing a valuable library in the wooden shack of a registered pyromaniac. The vulnerability of mitochondrial genes to damage is reflected in their high evolution rate—in mammals, some twentyfold greater than the nuclear genes.
So there are serious costs to retaining mitochondrial genes. I repeat: if gene transfer is easy, why on earth are there any mitochondrial genes left? The first and most obvious reason is that the genes are not the problem: it is the products of the mitochondrial genes, the proteins, that need to function in the mitochondria. These are mostly involved in cellular respiration and so are vitally important to the life of the cell. If the genes are transported to the nucleus, then somehow their protein products need to be routed back to the mitochondria, and if they fail to get there the cell may well die. Even so, many proteins encoded in the nucleus do get back to the mitochondria: they are ‘tagged’ with a short chain of amino acids—an ‘address’ tag, pinpointing the final destination, as we discussed when considering lipids a few pages ago. The address tag is recognized by protein complexes in the mitochondrial membranes that act as customs posts, controlling import and export across the membranes. Many hundreds of proteins destined for the mitochondria are tagged in this way. But the simplicity of this system raises a question of its own—why can’t all proteins that are destined for the mitochondria be tagged in this way?
The textbook answer is that they can—it just takes a long time to arrange, long even in terms of the vast stretches of evolutionary time. A number of chance events must be negotiated before a protein can successfully be targeted back to the mitochondria. First of all, the gene must be incorporated properly into the nucleus, which is to say that the entire gene (rather than a bit of it) must be transferred to the nucleus, and then integrated into the nuclear DNA. Once incorporated, it has to work: it must be switched on and transcribed to produce a protein. This may be difficult, as genes are inserted more or less randomly into the nuclear DNA, and can make a mess of the other genes already there, as well as regulatory sequences that govern genetic activity. Second, the protein must acquire the correct address tag, which again appears to be a chance event; otherwise it will not be targeted back to the mitochondria. Instead, it will be constructed in the cytoplasm and remain there, like a woebegone Trojan horse that failed to gain entrance to Troy. Acquiring the right address tag takes time, time that is measured in aeons. Thus, say theorists, the few remaining mitochondrial genes are just a shrinking residual. One day, perhaps a few hundred million years hence, no mitochondrial genes will be left at all. And the fact that different species have different numbers of genes that remain in their mitochondria lends support to the slow, random nature of this process.
The nucleus is not enough
But this answer is not quite convincing. All species have lost almost all their mitochondrial genomes but not one species has lost them all. None has more than a hundred genes left, having started out with probably several thousand some two billion years ago, so the process has run very nearly to completion in all species. This gene loss has occurred in parallel: different species have lost their mitochondrial genes independently. As a proportion of the genes lost, all species have now lost between 95 and 99.9 per cent of their mitochondrial genes. If chance alone were the dominating factor, we might expect that at least a few species would have gone the whole hog by now, and transferred all mitochondrial genes to the nucleus. Not one has done so. All known mitochondria have retained at least a few genes. What’s more, mitochondria isolated from different species have invariably retained the same core of genes: they have independently lost the great majority of their genes but kept esse
ntially the same handful, again implying that chance is not to blame. Interestingly, exactly the same applies to chloroplasts, which, as we have seen, are in a similar position: no chloroplast has lost all of its genes, and again, the same core of genes always figures among them. In contrast, other organelles related to mitochondria, such as hydrogenosomes and mitosomes, have almost invariably lost all their genes.
A number of reasons have been put forward to account for the fact that all known mitochondria have retained at least a few genes. Most are not terribly convincing. One idea once popular, for example, is that some proteins can’t be targeted to the mitochondria because they are too large or too hydrophobic—but most of these proteins have in fact been successfully targeted to the mitochondria, either in one species or another, or by means of genetic engineering. Clearly the physical properties of proteins are not insurmountable obstacles to their parcelling and delivery to the mitochondria. Another idea is that the mitochondrial genetic systems harbour exceptions to the universal genetic code, and so mitochondrial genes are no longer strictly analogous to nuclear genes. If these genes were moved to the nucleus and read off according to the standard genetic code, the resulting protein would not be quite the same as that produced by the mitochondrial genetic system, and might not function correctly. But this can’t be the full answer either, as in many species the mitochondrial genes do conform to the universal genetic code. There is no discrepancy in these cases, and therefore no reason why all the mitochondrial genes could not be transferred to the nucleus—and yet they remain stubbornly in the mitochondria. Likewise, there are no variations in the universal genetic code in chloroplast genes, and yet, like mitochondria, they always retain a core contingent of genes on site.
The answer that I believe to be correct is only now gaining credence among evolutionary biologists, despite being put forward by John Allen, then at the University of Lund in Sweden, as long ago as 1993. Allen argues that there are many good reasons why all the mitochondrial genes should have moved to the nucleus, and no clear ‘technical’ reasons why any should have stayed. Therefore, he says, there must be a very strong positive reason for their retention. They have not remained there by chance, but because natural selection has favoured their retention despite the manifold disadvantages. In the balance of pros and cons, the pros prevailed, at least in the case of the small number of genes that remain. But if the cons are so obvious and important, it is remarkable that we have overlooked the pros—they must be even weightier.
The reason, says Allen, is no less than the raison d’être of mitochondria: respiration. The speed of respiration is very sensitive to changing circumstances—whether we’re awake or asleep, or doing aerobics, sitting around, writing books, or chasing a ball. These sudden shifts demand that mitochondria adapt their activity at a molecular level—a requirement that is too important, and abruptly swinging, to be controlled at a distance by the bureaucratic confederation of genes far away in the nucleus. Similar sudden shifts in requirements occur not just in animals but also in plants, fungi, and microbes, which are even more subject to the vicissitudes of the environment (such as changing oxygen levels, heat, or cold) at the molecular level. To respond effectively to these abrupt changes, Allen argues, mitochondria need to maintain a genetic outpost on site, as the redox reactions that take place in the mitochondrial membranes must be tightly regulated by genes on a local basis. Notice that I’m referring to the genes themselves here, and not to the proteins that they encode; we’ll look into why the genes are important in a moment. But before we move on, let’s note that the need for local genetic rapid-response units not only explains why mitochondria must retain a contingent of genes, but also, I believe, why the bacteria could not evolve into more complex eukaryotic cells by natural selection alone.
The problem of poise
Let’s think back again to how respiration works. Electrons and protons are stripped from food, and react with oxygen to provide the energy that we need to live. The energy is released a bit at a time, by breaking the reaction into a series of small steps. These steps take place in the respiratory chain, down which electrons flow, as if down a tiny wire. At several points the energy released is used to pump protons across a membrane, trapping them on the other side, like water behind the dam of a reservoir. The flow of protons back from this reservoir, through special channels in the dam (the drive shafts of the ATPase motor) powers the formation of ATP, the energy ‘currency’ of the cell.
Let’s consider briefly the speed of respiration. Everything is coupled like cogs, so the speed of one cog controls the speed of the rest. So what controls the overall speed of the cogs? The answer is demand, but let’s think this through. If electrons flow quickly down the chain, then the protons are pumped quickly (for proton pumping depends on electron flow) and the proton reservoir ‘fills up’. A full reservoir, in turn, provides a high pressure to form ATP quickly, as protons flow back through the dedicated drive shaft of the ATPase. Now think what happens if there is no demand for ATP. In Chapter 4, we saw that ATP is formed from ADP and phosphate, and when it is broken down again, to provide energy, it reverts to ADP and phosphate. When demand is low, ATP is not consumed by the cell. Respiration converts all the ADP and phosphate into ATP, and that’s that: the raw materials are exhausted, and the ATPase must grind to a halt. If the ATPase motor is not turning, then protons can no longer pass through the drive shaft. The proton reservoir brims full. As a result, protons can no longer be pumped against the high pressure of the reservoir. And without proton pumping, electrons can’t flow down the chain. In other words, if demand is low, everything backs up and the speed of respiration slows right down until fresh demand starts all the wheels turning again. So the speed of respiration ultimately depends on demand.
But this is what happens when everything is working well and the cogs are well greased. There are other reasons for respiration to slow down, and these are not related to demand but to supply. We have noted one instance: the supply of ADP and phosphate. Normally, the concentration of these raw materials reflects the consumption of ATP, but it is always possible that there is simply a shortage of ADP and phosphate. Then there is the supply of oxygen or glucose. If there is not enough oxygen around—if we are suffocating—electron flow down the chain must slow down because there is nothing to remove the electrons at the end. They are forced to back up in the chain, and everything else slows down just as if there were a shortage of ADP. What about glucose? Now the number of electrons and protons that enter the chain is restricted—as if we were starving—so the flow of electrons is forced to slow down, which is to say the volume of electrons flowing down the chain per second falls.
So, the overall speed of respiration should ideally reflect demand, which is to say consumption of ATP, but under difficult conditions, such as starvation or suffocation, or perhaps a metabolic shortage of raw materials, then the speed of respiration reflects the supply rather than the demand. In both cases, however, the overall speed of respiration is reflected in the speed that electrons flow down the respiratory chain. If electrons flow quickly, glucose and oxygen are consumed quickly, and by definition, respiration is fast. Now, after this little detour, we can return to the point. There is a third factor that causes respiration to slow down, and this relates neither to supply nor demand, but rather to the quality of the wiring: it relates to the components of the respiratory chain themselves.
The components of the electron-transport chains have a choice of two possible states: they can either be oxidized (they don’t have an electron) or they can be reduced (they do have an electron). Obviously they can’t be both at once—they either have an electron or they don’t. If a carrier already has an electron, it can’t receive another one until it has passed on the first to the next carrier in the chain. Respiration will be held up until it has passed on this electron. Conversely, if a carrier doesn’t have an electron, it can’t pass on anything to the next carrier until it has received an electron from an earlier carrier. Resp
iration will be held up until it receives one. The overall speed of respiration therefore depends on the dynamic equilibrium between oxidation and reduction. There are thousands of respiratory chains in a single mitochondrion. Respiration will proceed most rapidly when 50 per cent of the carriers within these chains are oxidized (ready to receive electrons from an earlier carrier), and 50 per cent are reduced (ready to pass on electrons to the next carrier). If the rate of respiration is plotted out mathematically, it fits the equation of a bell curve. Respiration is fast at the top of the bell curve and slows precipitously on either side, as the carriers become more oxidized or reduced. The point of optimal balance, the top of the bell curve at which respiration is fastest, is known as ‘redox poise’. Straying from redox poise slows down energy production, and such inefficiency, as we have seen, is strongly selected against in bacteria.
But the penalty for straying from redox poise is worse than inefficiency: there is the devil to pay. All the carriers of the respiratory chain are potentially reactive—they ‘want’ to pass their electrons to a neighbour (they have a chemical propensity to do so). If respiration is progressing normally, each carrier is most likely to pass on its electrons to the next carrier in the chain, each one of which ‘wants’ the electron a bit more than did its predecessor; but if the next carrier is already full then the chain becomes blocked. There is now a greater risk that the reactive carriers will pass on their electrons to something else instead. The most likely candidate is oxygen itself, which easily forms toxic free radicals such as the superoxide radical. I discussed the damage caused by free radicals in Oxygen; here, the important point is that free radicals react indiscriminately to damage all kinds of biological molecules. Formation of free radicals by the respiratory chain has influenced life in profound and unexpected ways, including the evolution of warm-bloodedness, cell suicide, and ageing, as we’ll see in later chapters. For now, though, let’s just note that if the chain becomes blocked, it is more likely to leak free radicals, just as a blocked drainpipe is more likely to spring water from small cracks.
