by Nick Lane
Among cells, it is interesting that predation and parasitism tend to pull in opposite directions. As a rule of thumb, parasites are regressive in character, and in this regard the eukaryotic parasites are no exception. The very word ‘parasite’ conveys something contemptible. Conversely the term ‘predator’ can send shivers up and down the spine. Predation tends to drive evolutionary arms races, in which the predator and prey compete to grow ever larger: the red queen effect, whereby both sides must run to stay in the same place, relative to each other. I know of no bacterial cells that are predatory in the eukaryotic fashion of physically engulfing their prey. Perhaps this should not be surprising. A predatory lifestyle requires a very substantial energetic investment before anything is caught and eaten. At the cellular level, engulfing food by phagocytosis, in particular, demands a dynamic cytoskeleton and an ability to change shape vigorously, both of which consume copious quantities of ATP. So phagocytosis is made possible by three factors: the ability to change shape (which requires losing the cell wall, then developing a far more dynamic cytoskeleton); sufficiently large size to physically engulf prey; and a plentiful supply of energy.
Bacteria can lose their cell wall but have never developed phagocytosis. Vellai and Vida, whom we met earlier, argue that the additional requirements of phagocytosis for large size and plentiful ATP may have prevented bacteria from ever becoming effective predators in the eukaryotic style. Respiring over the outer membrane means that bacteria are obliged to generate less energy, relative to their size, as they become bigger. When they become large enough to physically engulf other bacteria they are less likely to have the energy needed to do so. Worse, if the cell membrane is specialized for energy generation, then phagocytosis would also be detrimental, for it would disrupt the proton gradient. It is possible that bacteria could circumvent such problems by relying on fermentation, rather than respiration, as this does not require a membrane. But fermentation also generates substantially less energy than respiration, and this may limit the ability of cells to survive by phagocytosis. Vellai and Vida note that all the eukaryotic cells that live by the combination of fermentation and phagocytosis are parasites, and so might be able to make energy savings in other areas (for example, not synthesizing their own nucleotides and amino acids, the building blocks of DNA and proteins).3 By sacrificing their energetic expenses in some areas they might be able to justify the energetic costs of phagocytosis. But I’m not aware of any research that looks into this hypothesis systematically, and unfortunately Tibor Vellai has moved on from this field of research.
10 Internal bioenergetic membranes of the bacterium Nitrosomonas, giving it a ‘eukaryotic’ look.
These ideas are interesting and may go some way towards explaining the dichotomy between bacteria and eukaryotes, but they leave a suspicion at the back of my mind. Why are bacteria invariably penalized if they get bigger? Bacteria are so inventive that it is remarkable that none of them have ever solved the challenge of simultaneously increasing their size and their energy status. It doesn’t sound so difficult: all they needed to do was grow some internal membranes for generating energy. If internalization of energy production inside the cell enabled eukaryotes to make their quantum leap in size and behaviour, what was to stop bacteria from having internal membranes themselves? Some bacteria, such as Nitrosomonas and Nitrosococcus do in fact have quite complex internal membrane systems, devoted to generating energy (Figure 10). They have a eukaryotic ‘look’ about them. The cell membranes are extensively infolded, creating a large periplasmic compartment. It seems to be a small step from here to a fully compartmentalized eukaryotic cell; so why did it never happen?
In the next chapter, we’ll take up the story of the first chimeric eukaryote that we abandoned without a nucleus at the end of Part 1, and look into what may have become of it next. Guided by the principles of energy generation, which we explored in Part 2, we’ll see why a symbiosis between two cells was successful, and why, by the same token, it was not possible for bacteria to compartmentalize themselves in the same way as eukaryotes, by natural selection alone. We’ll see why only eukaryotes could become giant predators in a bacterial world—indeed, why they overturned the bacterial world forever.
8
Why Mitochondria Make Complexity Possible
In the last chapter, we considered why bacteria have remained small and unsophisticated, at least in terms of their morphology. The reasons relate mostly to the selection pressures that face bacteria. These are different from eukaryotic cells because bacteria, for the most part, do not eat each other. Their success in a population therefore depends largely on the speed of their replication. This in turn depends on two critical factors: first, copying the bacterial genome is the slowest step of replication, so the larger the genome, the slower is replication; and second, cell division costs energy, so the least energetically efficient bacteria replicate the slowest. Bacteria with large genomes will always tend to lose out in a race against those with smaller genomes, because bacteria swap genes, by way of lateral gene transfer, and so can keep loading up cassettes of useful genes, and throwing them away again as soon as they become burdensome. Bacteria are therefore faster and more competitive if genetically unburdened.
If two cells have the same number of genes, and have equally efficient energy-generating systems, then the cell that can replicate the fastest will be the smaller of the two. This is because bacteria depend on their outer cell membrane to generate energy, as well as absorbing food. As bacteria become larger in size, their surface area rises more slowly than their internal volume, so their energetic efficiency tails away. Larger bacteria are energetically less efficient, and always likely to lose out in competition with smaller bacteria. Such an energetic penalty against large size precludes phagocytosis, for physically engulfing prey demands both large size and plenty of energy to change shape. Eukaryotic-style predation—catching and physically eating prey—is therefore absent among bacteria. It seems that eukaryotes escape this problem because they generate their energy internally, which makes them relatively independent of their surface area, and enables them to become many thousands of times larger without losing energetic efficiency.
As a distinction between the bacteria and eukaryotes, this reason sounds flimsy. Some bacteria have quite complex internal membrane systems and could be released from the surface-area constraint, yet still don’t approach eukaryotes in size and complexity. Why not? We’ll look into a possible answer in this chapter, and it is this: mitochondria need genes to control respiration over a large area of internal membranes. All known mitochondria have retained a contingent of their own genes. The genes that mitochondria retain are specific, and the mitochondria were able to retain them because of the nature of their symbiotic relationship with their host cell. Bacteria do not have this advantage. Their tendency to throw away any superfluous genes has prevented them from ever harnessing the correct core contingent of genes to govern energy generation, and this has always prevented them from developing the size and complexity of the eukaryotes.
To understand the reasons why mitochondrial genes are important, and why bacteria can’t acquire the correct set of genes for themselves, we’ll need to penetrate further into the intimate relationship between the cells that took part in the original eukaryotic union, two billion years ago. We’ll take up the story where we left off in Part 1. There, we parked the chimeric eukaryote as a cell that had mitochondria but had not yet developed a nucleus. Because a eukaryotic cell is, by definition, a cell that has a ‘true’ nucleus, we can’t really refer to our chimera as a eukaryote. So let’s think now about the selection pressures that turned our strange chimeric cell into a proper eukaryotic cell. These pressures hold the key not just to the origin of the eukaryotic cell, but also to the origin of real complexity, for they explain why bacteria have always remained bacteria: why they could never evolve into complex eukaryotes by natural selection alone, but required symbiosis.
Recall from Part 1 that the key to
the hydrogen hypothesis is the transfer of genes from the symbiont to the host cell. No evolutionary novelties were called for, beyond those that already existed in the two collaborating cells entered in an intimate partnership. We know that genes were transferred from the mitochondria to the nucleus, because today mitochondria have few remaining genes, and there are many genes in the nucleus that undoubtedly have a mitochondrial origin, for they can be found in the mitochondria of other species that lost a different selection of genes. In all species, mitochondria lost the overwhelming majority of their genes—probably several thousand. Exactly how many of these genes made it to the nucleus, and how many were just lost, is a moot point among researchers, but it seems likely that many hundreds did make it to the nucleus.
For those not familiar with the ‘stickiness’ and resilience of DNA, it may seem akin to a conjuring trick for genes from the mitochondria to suddenly appear in the nucleus, like a rabbit produced from a top hat. How on earth did they do that? In fact such gene hopping is commonplace among bacteria. We have already noted that lateral gene transfer is widespread, and that bacteria routinely take up genes from their environment. Although we normally think of the ‘environment’ as outside the cell, acquiring spare genes from inside the cell is even easier.
Let’s assume that the first mitochondria were able to divide within their host cell. Today, we have tens or hundreds of mitochondria in a single cell, and even after two billion years of adaptation to living within another cell they still divide more or less independently. At the beginning, then, it’s not hard to picture the host cell as having two or more mitochondria. Now imagine that one dies, perhaps because it can’t get access to enough food. As it dies, it releases its genes into the cytoplasm of the host cell. Some of these genes will be lost altogether, but a handful might be incorporated into the nucleus, by means of normal gene transfer. This process could, in principle, be repeated every time a mitochondrion dies, each time potentially transferring a few more genes to the host cell.
Such transfer of genes might sound a little tenuous or theoretical, but it is not. Just how rapid and continuous the process can be in evolutionary terms was demonstrated by Jeremy Timmis and his colleagues at the University of Adelaide in Australia, in a Nature paper of 2003. The researchers were interested in chloroplasts (the plant organelles responsible for photosynthesis), rather than mitochondria, but in many respects chloroplasts and mitochondria are similar: both are semi-autonomous energy-producing organelles, which were once free-living bacteria, and both have retained their own genome, albeit dwindling in size. Timmis and colleagues found that chloroplast genes are transferred to the nucleus at a rate of about 1 transfer in every 16 000 seeds in the tobacco plant Nicotiana tabacum. This may not sound impressive, but a single tobacco plant produces as many as a million seeds in a single year, which adds up to more than 60 seeds in which at least one chloroplast gene has been transferred to the nucleus—in every plant, in every generation.
Very similar transfers take place from the mitochondria to the nucleus. The reality of such gene transfers in nature is attested by the discovery of duplications of chloroplast and mitochondrial genes in the nuclear genomes of many species—in other words the same gene is found in both the mitochondria or chloroplast and in the nucleus. The human genome project has revealed that there have been at least 354 separate, independent transfers of mitochondrial DNA to the nucleus in humans. These DNA sequences are called numts, or nuclear-mitochondrial sequences. They represent the entire mitochondrial genome, in bits and pieces: some bits repeatedly, others not. In primates and other mammals, such numts have been transferred regularly over the last 58 million years, and presumably the process goes back further, as far as we care to look. Because DNA in mitochondria evolves faster than DNA in the nucleus, the sequence of letters in numts can act as a time capsule, giving an impression of what mitochondrial DNA might have looked like in the distant past. Such alien sequences can cause serious confusion, however, and were once mistaken for dinosaur DNA, leaving one team of researchers with red faces.
Gene transfer continues today, occasionally making itself noticed. For example, in 2003, Clesson Turner, then at the Walter Reed Army Medical Center in Washington, and collaborators, showed that a spontaneous transfer of mitochondrial DNA to the nucleus was responsible for causing the rare genetic disease Pallister-Hall syndrome in one unfortunate patient. How common such genetic transfers are in the pantheon of inherited disease is unknown.
Gene transfers occur predominantly in one direction. Think back again to the first chimeric eukaryote. If the host cell were to die, it would release its symbionts, the proto-mitochondria, back into the environment, where they may or may not perish—but regardless of their fate, the experiment in chimeric co-existence would certainly have perished. On the other hand, if a single mitochondrion were to die, but a second viable mitochondrion survived in the host cell, then the chimera as a whole would still be viable. To get back to square one, the surviving mitochondrion would just have to divide. Each time a mitochondrion died, the genes released into the host cell could potentially be integrated into its chromosome by normal genetic recombination. This means there is a gene ratchet, favouring the transfer of genes from the mitochondria to the host cell, but not the other way around.
The origin of the nucleus
What happens to the genes that are transferred? According to Bill Martin, whom we met in both Parts 1 and 2, such a process might account for the origin of the eukaryotic nucleus. To understand how, we need to recall two points that we have discussed in earlier chapters. First, recall that Martin’s hydrogen hypothesis argues that the eukaryotic cell was first forged from the union of an archaeon and a bacterium. And second, recall from Chapter 6 (page 98) that archaea and bacteria have different types of lipid in their cell membranes. The details don’t matter here, but consider the kind of membranes we would expect to find in that first, chimeric eukaryote. The host cell, being an archaeon, should have had archaeal membranes. The mitochondria, being bacterial, should have had bacterial membranes. So what do we actually see today? Eukaryotic membranes are uniformly bacterial in nature—both in their lipid structure and in many details of their embedded proteins (like the proteins that make up the respiratory chain, and similar proteins found in the nuclear membrane). The bacterial-style membranes of the eukaryotes include the cell membrane, the mitochondrial membranes, other internal membrane structures, and the double nuclear membrane. In fact there is no trace of the original archaeal membranes in the eukaryotes, despite the fact that other features make it virtually certain that the original host cell was indeed an archaeon.
Such basic consistency, when we would expect to find disparity, has led some researchers to question the hydrogen hypothesis, but Martin considers the apparent anomaly to be a strength. He suggests that the genes for making bacterial lipids were transferred to the host cell, along with many other genes. Presumably, if functional, the genes went ahead with their normal tasks, such as making lipids; there is no reason why they should not function normally as before. But there may have been one difference—the host cell may have lost the ability to target protein products to particular locations in the cell (protein targeting relies on an ‘address’ sequence that differs in different species). The host cell may therefore have been able to make bacterial products, such as lipids, but not known exactly what to do with them; in particular, where to send them. Lipids, of course, don’t dissolve in water, and so if not targeted to an existing membrane would simply precipitate as lipid vesicles—spherical droplets enclosing a hollow watery space. Such droplets fuse as easily as soap bubbles, extending into vacuoles, tubes, or flattened vesicles. In the first eukaryote, these vesicles might simply have coalesced where they were formed, around the chromosome, to form loose, baggy membrane structures. Now this is exactly the structure of the nuclear membrane today—it is not a continuous double membrane structure like the mitochondria or chloroplasts, but is composed of a series of flattened v
esicles, and these are continuous with the other membrane systems within the cell. What’s more, when modern eukaryotic cells divide, they dissolve the nuclear membrane, to separate the chromosomes destined for each of the daughter cells; and a fresh nuclear membrane forms around the chromosomes in each of these daughter cells. It does so by coalescing in a manner reminiscent of Martin’s proposal, and remains continuous with the other membrane systems of the cell. Thus, in Martin’s scenario, gene transfer accounts for the formation of the nuclear membrane, as well as all the other membrane systems of eukaryotic cells. All that was needed was a degree of orientational confusion, a map-reading hiatus.