by Nick Lane
The centrality of mitochondria to both the main forms of apoptosis raises the possibility that it was ever thus. We have discussed the fact that bacteria and cancer cells act independently in their own interests, and so can be seen as ‘units of selection’. At one and the same time, selection can operate at the level of the cell and that of the individual. Mitochondria were once free-living bacteria, and at that time operated independently. Once incorporated into the eukaryotic cell, they presumably retained the power to operate as independent cells, at least for a while: they were independent cells living within a larger organism, and could rebel in the same way as cancer cells (also independent cells living within a larger organism).
If, today, mitochondria bring about the demise of their host cell, might it be that from the very beginning, the mitochondria killed their host cells in their own interests? In other words, the origin of apoptosis was not an altruistic act on behalf of the individual, but a selfish act on behalf of the tenants themselves. If this view is correct, then apoptosis is better seen as murder than suicide. And if so, the reason why single cells should apparently commit suicide is clear: they are sabotaged from within. So is there any evidence that the mitochondria brought with them to the eukaryotic merger the apparatus of death? Indeed there is.
Parasite wars?
We know that the gene for cytochrome c was brought to the eukaryotic merger by the ancestors of the mitochondria, rather than the host cell, and was later transferred to the nucleus (see Part 3). We know this because the gene sequence has almost identical counterparts in the α-proteobacteria, and is part of the respiratory chain that was their most important contribution to the partnership. Just how important cytochrome c was in the early evolution of apoptosis is less certain. While it seems to play a pivotal role in mammals, and perhaps in plants, it is not needed for apoptosis in fruit flies or nematode worms: certainly it is not a universal player. But it might once have played a central role in apoptosis, before being superseded in a few species, or it might have assumed its decisive role more recently, independently, in plants and mammals. We won’t know which is closer to the truth until we know more about how apoptosis works in the most primitive eukaryotes.
But cytochrome c, as we have seen, is only one of quite a number of proteins released from the mitochondria during apoptosis—proteins with strange names, like Smac/DIABLO, Omi/HtrA2, endonuclease G, and the AIF (and in fruit flies, more evocatively, Reaper, Grim, and Sickle). Their names need not concern us, but we should note that some of these proteins can at times play a more important role than cytochrome c itself. Most of them have only been identified since the turn of the millennium, but already, from the prolific genome sequencing projects around the world, we know something of their provenance. The pattern is striking. With the sole exception of AIF (apoptosis-inducing factor), all known apoptotic proteins released from the mitochondria are bacterial in origin, and are absent from the archaea. (Recall from Part 1, page 48, that the host cell was almost certainly an archaeon, whereas the mitochondria are bacterial in origin.) The bacterial origin of these proteins means that they were not contributed by the host cell, which must have had little in the way of death machinery. Not all of these proteins were necessarily brought to the merger by the mitochondria themselves—a few seem to have gained access to eukaryotic cells more recently, by lateral gene transfers from other bacteria—but it looks as if only AIF came from the archaeal host cell, and even this does not act to kill in archaea.
These mitochondrial proteins are not the only ones to have come from bacteria. If their gene sequences are to be believed, the caspase enzymes, too, almost certainly came from the bacteria, probably by way of the mitochondrial merger. It’s worth noting, though, that the bacterial caspases are tame—they slice up proteins but do not cause cell death. More intriguing is the ancestry of the bcl-2 family. Here, the gene sequences have little in common with either the bacteria or archaea—but the 3-dimensional structure of the proteins does betray a possible link with bacterial proteins, in particular a group of toxins found in infectious bacteria such as diphtheria. Like the pro-apoptosis proteins of the bcl-2 family, the bacterial toxins form pores in the host cell membranes, sometimes even inducing apoptosis, suggesting a plausible functional link.
Taken together, these findings imply that most of the machinery of death was brought to the eukaryotic merger by the ancestors of the mitochondria. It really does look like murder from within, rather than suicide, a thankless act of ingratitude on the part of the tenant. This idea was developed into a forceful hypothesis by José Frade and Theologos Michaelidis, of the Max-Planck Institute for Psychiatry in Martinsried in Germany, in 1997. Much of the evidence discussed above, accumulated since 1997, seems to lend support to their case.
Frade and Michaelidis drew a parallel between the behaviour of the modern bacterium Neisseria gonorrhoeae, the cause of the sexually transmitted disease gonorrhoea, and the possible behaviour of the proto-mitochondria in the early days of the eukaryotic merger. N. gonorrhoeae infects the cells of the urethra and cervix, along with white blood cells. Once inside these cells, N. gonorrhoeae brings to bear a diabolical cunning. The bacteria produce a pore-forming protein known as PorB (which is similar to the mitochondrial bcl-2 proteins). The PorB protein is inserted into the host cell membrane, as well as the vacuole membranes that enwrap the bacteria inside the cell. These pores are kept firmly closed by their interaction with the host cell’s ATP—again, some of the bcl-2 proteins behave in a similar way—but when the host’s ATP is depleted, the pores open. Opening of the pores triggers the host cell’s apoptosis machinery, leading to cell death. The gonorrhoeae themselves survive the experience. They take the opportunity to escape from the freshly disintegrated host cell, making use of the neatly packaged remnants of the cell for fuel. Thus, the bacteria subsist within their host cell for as long as the cell is healthy, by monitoring its ability to maintain good stocks of ATP (implying plentiful fuel); but as soon as the host cell begins to run down, outlasting its usefulness, it is summarily executed and the bacteria move on to pastures new. Bastards!
Frade and Michaelidis note that N. gonorrhoea is not the only bacterium to make use of such an insidious trick—the deadly bacterial predator Bdellovibrio, which we met in Part 1, employs similar tactics when inside other bacteria. It, too, monitors their metabolic health for a period before devouring its prey from within. Bdellovibrio has been cited by Lynn Margulis as a possible ancestor of the mitochondria themselves. Another contender, which we discussed in Parts 1 and 3, is Rickettsia prowazekii, also a parasite that lives inside other cells. These examples have in common a parasitical relationship with the host. Reconstructing such biochemical archaeology suggests that, in the first eukaryotes, the relationship between the mitochondria and their host cells was parasitical. The proto-mitochondria presumably got into an archaeon and monitored its health for a period, before triggering the death of this host, devouring its packaged remains, and moving on to the next.
If apoptosis grew out of an armed struggle between the cells that were later to be united as the eukaryotic cell, then the eukaryotic merger grew out of a relationship in which the parasite initially killed its host, and moved on to another. This is of course exactly what Lynn Margulis and others propose. The relationship ultimately bequeathed the eukaryotic cell with the machinery of death, which was only later employed for the more ‘altruistic’ purpose of programmed cell suicide in multicellular organisms. But a parasite war is not the story that we told in Part 1, when we considered the origin of the eukaryotic cell; there we talked about a collaboration between two peaceful prokaryotes which lived side by side, in what amounted to metabolic wedlock. When we considered the evidence, we dismissed the possibility that the relationship between the two cells was parasitic. But now, from a different perspective, there is a challenge to that view. Nothing is certain in this kind of science—it is all about weighting the bits and pieces of evidence that have a bearing on the matter; and this ev
idence most certainly bears on the matter. So does it overturn our already unstable craft? Should I, horror of horrors, go back and start rewriting Part 1?
12
Foundations of the Individual
The multicellular individual is made up of cells that collaborate together for the greater good. Nonetheless, this collaboration is not a cellular love fest—it is enforced by the death penalty for any cells that try to abscond and return to their ancestral way of life. Occasionally, selfish cells escape detection and evade the death penalty, and when they do the result is cancer. Cancer cells replicate furiously, in their own interests rather than those of the body, undermining the integrity of the body. Ultimately, having evaded death temporarily themselves, they bring about the death of their erstwhile master, and with it their own.
Cancer can persist because it is rare in younger individuals: if the body were to tear itself asunder by internal squabbles before the community of cells had engineered their own reproduction, though the germ-line, then the individual as a whole would fail to pass on its genes, and the selfish genes would be lost from the population. In the early days of multicellular organisms, however, the selfish cells that make up a multicellular body had a far better chance of independent survival—unlike cancer cells they could survive alone, and they had retained the potential to found a new colony of cells. This kind of independence still occurs in sponges and other simple animals today, but their laissez-faire laws of coexistence prevent them from scaling the heights of multicellular complexity. True commitment to the multicellular way of life demands the ultimate sacrifice—death for the greater good. But if cells could survive alone, how was the death penalty imposed upon them in the first place?
Today, the cellular death penalty, known as apoptosis, is executed by the mitochondria. The mitochondria integrate signals coming from different sources, and if the balance of signals indicates that the cell is damaged, and so prone to act in its own interests, then they activate the cell’s silent machinery of death. Swift and smooth, almost unnoticed, some 10 billion cells die by apoptosis every day in the human body, to be replaced by fresh, undamaged cells. The death apparatus consists of a number of proteins that are released from the mitochondria into the cell, which then activate the latent death enzymes, the caspases. These enzymes dismember the cell from within, and package its contents for reuse later by other cells. Nothing goes to waste.
Virtually all the death proteins that are released from the mitochondria, along with the caspase enzymes themselves, were brought to the eukaryotic cell by the bacterial ancestors of the mitochondria, back in the mists of evolutionary time. They still have close analogues among free-living, and especially parasitic, bacteria today. In modern bacteria, many of the death proteins are used for other purposes and are relatively ‘tame’—they don’t bring about the death of bacteria or anything else. On the other hand, one family of proteins, the bacterial porins, are actually targeted at other cells, instruments of war and murder rather than fruitful collaboration. This raises the prospect that it was ever thus: once upon a time, the bacterial ancestors of the mitochondria were parasites, and used proteins like the porins to attack and dismember their host cell from within, feeding on its remains before moving on to another cell.
Whether or not this case is reasonable hinges on the true identity of the bacterial porins. In modern parasites, they are plugged into the membranes of the host cell, and ruthlessly execute its demise as soon as it shows signs of flagging, of being unable to keep pace with the metabolic demands of its parasite. These bacterial porins bear an uncanny physical—but not genetic—resemblance to the mitochondrial porins: the bcl-2 proteins that activate the cell’s machinery of death by forming pores in the mitochondrial membranes. The larger implication is that the eukaryotic cell itself was forged in the crucible of war between an intracellular parasite, which was later tamed and went on to become the mitochondria, and the host cells, which learnt to survive the infection.
This sounds simple enough, but it presents a conundrum. In Part 1, we looked into some of the theories of the origin of the eukaryotic cell, in particular the ‘parasite model’, in which the mitochondria are derived from a Rickettsia-like bacterium, and the hydrogen hypothesis, which contends that the initial alliance grew out of mutual metabolic benefits—both partners lived from the metabolic waste products of the other. There I argued that the evidence at present supports the hydrogen hypothesis, rather than the parasite model. But the parasite story developed above doesn’t sit comfortably with the hydrogen hypothesis, which involves a peaceful metabolic union. A parasite may well benefit from killing its host and moving on to another, but a chemical addict can gain little from killing its supplier, and especially if it has no means of finding another one. So either the parasite story undermines the validity of the hydrogen hypothesis, or else it can’t be correct itself, despite its apparent explanatory power. I don’t see how both of the theories can be right. So which view is correct?
To make a stab at answering this question, we first need to distinguish between the evidence that is known to be true, or at least is not disputed, and ingenious surmise. This is not hard. It’s plain that the mitochondria supplied most of the death machinery: they are central to apoptosis today, and almost certainly were instrumental in its evolution. But the link between the bcl-2 proteins, and the bacterial porins, such as those of Neisseria gonorrhoeae, which we discussed in the final pages of the last chapter, must be classed as ingenious surmise. Certainly there are intriguing structural similarities, but this does not constitute proof of an evolutionary relationship.
There are three possible relationships between the bcl-2 proteins and the bacterial porins, on the basis of what is known today. First, the similarities may result from convergent evolution, such that the mitochondria and N. gonorrhoeae both independently invented similar-looking proteins with similar purposes. Nothing in the known gene sequences rules out this possibility—and anyone who doubts the power of convergent evolution at a molecular level should read Life’s Solution by Simon Conway Morris. If this possibility were true then we wouldn’t expect to find any genetic relationship between the bcl-2 proteins and bacterial porins, since they evolved from different starting places; but we would expect to find structural similarities, as their functional purpose is similar. As there are only a few possible ways to form a large pore in a lipid membrane, these must place functional constraints on possible 3-dimensional structures. If two different cells both need large pores, they are obliged to come up with something similar.
The second possibility is that the mitochondria really did inherit the bcl-2 proteins from their bacterial ancestors, as suggested by Frade and Michaelidis, and discussed in the previous chapter. This can only be proved by similarities in the gene sequences, which have not been found so far. What’s more, such similarities would need to be found among representatives of the α-proteo-bacteria, the known ancestors of the mitochondria, or else lateral gene transfer at a later stage could not be ruled out. Clearly, if lateral gene transfers took place later on, that would say nothing about the initial relationship between the mitochondria and the host cells. So a more systematic sampling of genes across the α-proteobacteria might lend support to this hypothesis, but in the meantime the structural similarities are suggestive at best.
Finally, it’s feasible that N. gonorrhoea and other parasitic bacteria acquired their porins from the mitochondria, rather than the other way around. Such transfers of genes from host to parasite are common. If this were the case, we might expect to find similarities in the gene sequences between the mitochondria and the parasites. The lack of such genetic similarities might only be for want of trying—they’ll turn up when we have sequenced more genes—or it might be that sequence similarities have simply been lost over time, smothering any evidence of common ancestry. This is not unlikely, as the unceasing evolutionary war waged between parasite and host means that parasite genes are notoriously volatile. Furthermore, the bacterial
porins do not themselves bring about the whole of apoptosis—they merely plug into the host’s existing death apparatus. Effectively, they bring with them a portable ‘on’ switch, which triggers the host’s own machinery. The behaviour of parasites that cause cell death today is therefore not comparable with the inferred role of the protomitochondria, for they would have had to bring the entire death apparatus with them, and implement it in the host cell without killing themselves in the process. (Today, of course, the mitochondria die along with their host cell.)
On the basis of evidence to date, it isn’t possible to resolve between these three possibilities. Nonetheless, the picture of parasite wars painted by Frade and Michaelidis does seem at least coherent and plausible. Or does it? There are a few other, rather knotty, problems with the story. First and foremost, mitochondria are no longer independently replicating cells, and in all probability they would have lost their independence soon after their genes began to be transferred to the host cell nucleus. Once a few critical genes were held hostage in the nucleus, then the mitochondria could gain nothing from killing their host, as they could no longer survive independently out in the wild. Their future was tied to that of their host. That’s not to say they could gain nothing from manipulating their host, but surely they could not gain from actually killing it. In contrast, none of the parasites that we’ve discussed, even tiny Rickettsia, has ever lost its independence. They all maintain complete control over their life cycle and resources. They can get away with murder in a way that mitochondria cannot.
Exactly when mitochondria lost power over their own future is unknown, but it is likely to have happened quite early in the evolution of the eukaryotic cell. Consider, for example, the evolution of the ATP carrier, the membrane pump that exports ATP from the mitochondria (see page 145). For the first time, this enabled eukaryotic cells to extract energy in the form of ATP from mitochondria (which could hardly even be called mitochondria until then). It was a symbolic moment, for the symbionts no longer had control of their own energy resources—they had suffered a loss of sovereignty. For the mitochondria, it marked the transition from a symbiotic relationship to a captive state. We can date the transition reasonably accurately by comparing the sequences of the ATP-carrier gene in the various different groups of eukaryotes. In particular, the fact that the carrier is found in all groups of eukaryotes, including plants, animals, fungi, algae, and protozoa, implies that it evolved before the divergence of these groups, placing it very early in the history of the eukaryotic cell. I need hardly say that this places it well before the evolution of multicellular organisms; from fossil evidence, probably by a few hundred million years.