ACCORDING to the model outlined in preceding chapter, the history of eukaryotic cells is divided into two parts: an anaerobic phase, in the course of which were developed all the main eukaryotic properties, with the exception of the organelles of oxygen metabolism, and an aerobic phase dominated by the acquisition of those organelles.
We have seen in Chapter 8 that the eukaryotic line probably detached from the two prokaryotic lines early after the tree of life first emerged from its root. By what kind of cell was this process inaugurated? This is a much-debated question.
Early sequencing results first suggested an archaebacterial origin of the eukaryotic line. Later, however, a number of eukaryotic genes were found to be more closely related to eubacterial than to archaebacterial genes. In the opinion of many investigators, this genetic blend could be the outcome of horizontal gene transfers, such as are believed to have occurred on a large scale in the early days of life (see Chapter 8).
Some workers, however, point out that the mixture is apparently not random. To put a complicated matter in simple terms, the eubacterial genes mostly code for “housekeeping” enzymes, whereas the archaebacterial genes tend to define components of genetic information-transfer systems. Or, put in even more simplistic terms, the cytosol of the ancestral cell appears to be of eubacterial origin, its nucleus of archaebacterial origin. On the strength of this dichotomy, the suggestion has been made that the eukaryotic line was initiated by the fusion (the first of a number of fateful encounters) between a eubacterium, which ended up providing most of the cytoplasm to the chimera, and an archaebacterium, which furnished the chimera’s genetic machinery. Other models have also been proposed, including that of an endosymbiotic origin of the eukaryotic nucleus.
Leaving this question to the experts, I shall take as a working hypothesis that the founder of the eukaryotic line, whatever its origin, was a relatively simple, anaerobic, heterotrophic (subsisting on organic food) organism resembling present-day prokaryotes in its main properties. I shall further assume that this prokaryote did not have a cell wall. This hypothesis is derived from the fact, noted in the preceding chapter, that no eukaryotic cell is known that possesses a covering made of the same chemical substances as bacterial cell walls. This fact suggests that the ability to build a typical prokaryotic cell wall was lost in the eukaryotic line.1 It is possible, as will be mentioned later, that this loss may have played a significant role in the prokaryote-eukaryote transition.
Barring evidence to the contrary, it seems reasonable to assume that the ancestral prokaryote converted into a primitive eukaryote by a long succession of largely conserved intermediates each of which was only just a little less prokaryotic and a little more eukaryotic than its predecessor.
A first modification, which can be safely postulated in the framework of this hypothesis, is that the cells slowly grew bigger. They did not fundamentally change; they went on using the same enzymes and following the same metabolic pathways as before; they just made more of everything per cell, simply by delaying DNA replication and the onset of division. Interestingly, something of the kind happens to bacterial cells that have been stripped of their external wall (by an enzyme called lysozyme) or prevented from building a wall (by penicillin) in a medium where they are protected against osmotic bursting. The resulting naked cells, known as protoplasts, increase in size.
A second predictable modification is that the cell membrane expanded, perhaps even overexpanded, making numerous convolutions around an increasingly contorted cell body. Such a change, which, incidentally, may have been facilitated by the loss of a rigid external wall, was an almost obligatory concomitant of cellular enlargement, as more surface area was needed to allow the increasing exchanges of matter with the environment required to support the growing cell mass.2
The next step, which is seen as the crucial event in the proposed scenario, is assumed to be the occasional formation of closed, membrane- bounded, intracellular vesicles, splitting off from deepening invaginations of the cell membrane. Such phenomena, which, initially, could have depended on no more than the natural self-sealing property of lipid bilayers (see Chapter 6) helped by surface tension, are taken to have initiated the genesis of the cytomembrane system, leading, in the course of a protracted and complex evolutionary history, which also involved cytoskeletal and motor elements (see below), to a progressive differentiation of the vesicles into distinct parts, functionally specialized in a variety of import-export exchanges with the outside, such as are known today.
There is much to be said for this model. Membranes, we have seen, always arise from membranes—remember Blobel’s aphorism, omnis membrana e membrana (see Chapter 2). It is therefore likely that the inner membranes of eukaryotes arose from the surrounding membrane of their prokaryotic ancestor. In agreement with this hypothesis, the membranes of the two cell types have in common a number of genetically related functional systems. Most impressive are the systems involved in secretion. We have seen that, in eukaryotic cells, the ribosomes that synthesize secretory proteins stick to certain membranes of the cytomembrane system (rough endoplasmic reticulum3) and directly inject the proteins they manufacture into the cavities bounded by these membranes. In certain bacteria, ribosomes attached to the inner face of the cell membrane similarly deliver secretory proteins into the extracellular medium. Remarkably, the two systems depend on closely similar targeting sequences for directing the ribosomes to their membrane receptors, to the point of being able to obey each other’s signals. This similarity is clear evidence of an evolutionary kinship between the two systems and strongly supports the view that intracellular vesicles studded with ribosomes originally arose from similarly studded invaginations of the cell membrane, eventually to become what is now known as the rough endoplasmic reticulum (see below).
Another development, which, like membrane expansion, was to some extent mandated by the growing cell size, was the acquisition of cytoskeletal and motor elements. Little is known concerning the origin of these elements, which, as we have seen, are made essentially of certain specific proteins. Attempts at identifying in bacteria genes possibly ancestral to the eukaryotic genes coding for these proteins have gleaned limited results so far.4 It may be that fairly radical genetic innovations lie behind the eukaryotic development. It could also be, however, that the relevant innovations were fairly commonplace but of little use to prokaryotes and therefore not retained by natural selection.
The biologist Lynn Margulis, known for her early defense of, against considerable opposition, the endosymbiont theory, has proposed that eukaryotic flagella and cilia, motor organelles that she groups under the name “undulipodia,” are also derived from endosymbiotic bacteria, which she believes to be related to present-day spirochetes.5 This, according to her hypothesis, is how eukaryotic cells acquired microtubules, which are the main constituents of the motor organelles but have many other functions as well. The evidence put forward in support of this theory has, however, failed to convince the majority of experts. As we have seen in the preceding chapter, there is no evidence of either structural or functional kinship between eukaryotic flagella and the motor appendages of spirochetes.
It seems more likely that tubulins and the resulting microtubules arose first, in relation with one of the various functions carried out by these entities in present-day cells (mitotic division?) and that the characteristic structures of eukaryotic flagella and cilia, which share a highly complex arrangement of microtubules and associated proteins,6 developed later, though perhaps early enough for the common ancestor of eukaryotes to be motile, propelled, as are a number of protists, by beating cilia or undulating flagella.
Interestingly, the characteristic segregation of the genetic machinery within a central nucleus, which is the hallmark of eukaryotes, could possibly have been initiated by the membrane expansion process to have led to the cytomembrane system. In prokaryotes, the chromosome is anchored to the cell membrane. Invagination of this part of the cell membrane, in the general framework of cytomembrane development, would have dragged the chromosome into the interior of the cell. The resulting vesicle, perhaps with the participation of other vesicles, could conceivably have folded around the attached chromosome and enclosed it within a double-membranous envelope, related to the cytomembrane system as is the nuclear envelope in present-day eukaryotes. Cytoskeletal elements would then have been added later to reinforce this structure. This, incidentally, is how new nuclear envelopes form, at the end of mitosis, from parts of the cytomembrane system that fuse together around the segregated chromosomes.
Thus, the basic separation between nucleus and cytoplasm could have been initiated as part of the formation of the cytomembrane system. After that, a large number of innovations, about which virtually nothing is known, must have taken place to produce the highly complex organization of eukaryotic nuclei and the elaborate machinery involved in mitotic division.
In summary, the pathway proposed for the prokaryote-eukaryote transition is centered on a process of membrane expansion and vesiculation, associated with cellular enlargement and supported by the coevolutionary development of cytoskeletal and motor systems of increasing complexity. A process of this sort could, indeed, as postulated earlier, have involved a very large number of viable intermediates with progressively modified properties. Furthermore, plausible mechanisms can be suggested for some of its main steps. But for these to take place and for their acquisitions to become genetically transmissible, they must have enhanced the cells’ ability to survive and produce progeny under the prevailing environmental conditions. What could those advantages have been?
Imagine a heterotrophic bacterium of the kind that could have been ancestral to eukaryotic cells. It subsists on external food, which, like all present-day heterotrophic bacteria, it digests with the help of secreted exoenzymes. For obvious reasons, such an organism is forced to reside inside its food supply or, at least, in close juxtaposition with it, so that the secreted enzymes can accomplish their digestive functions without being immediately lost in the surroundings.
Now visualize such a cell beginning to undergo the membrane expansion and internalization process postulated in our model. All extracellular materials passively dragged into an invagination of the cell membrane will become segregated inside the resulting vesicle, as in endocytosis. On the other hand, exoenzymes that were secreted outside the cell by the piece of membrane involved in vesiculation will henceforth be discharged into the vesicle, where they will be able to act on the trapped material, as occurs in lysosomes. Extracellular digestion has become intracellular digestion.
The benefit of this transition, even in its primitive and random form, is enormous. Thanks to a seemingly trivial modification, the cell concerned and any of its progeny endowed with the same vesiculation ability have ceased to depend for their survival on intimate contact with a nutritive substratum. They are now free to roam around, to invade ponds and oceans, subsisting on food captured by infoldings of their membrane and digested within the resulting intracellular vesicles. This ability clearly entailed an invaluable evolutionary asset—I have referred to it as the beginning of cellular emancipation—so that any genetic modification likely to favor membrane expansion and the associated folding and fusion phenomena would have been strongly advantaged by natural selection.
Note that the first primitive vesicles proposed in this hypothesis combine the properties of three major components of the cytomembrane system: endocytic vesicles, containing material captured from the outside; lysosomes, sites of the digestion of this material; and parts of the rough endoplasmic reticulum, providers of the necessary digestive enzymes. The subsequent evolutionary history of the system may be readily pictured as a progressive separation of these functions into distinct, differentiated parts, finally leading to the complex cytomembrane system of extant cells. As we have seen, cytoskeletal and motor elements presumably participated in this evolutionary process.
According to the model sketched earlier, which resembles in many respects that long defended by the British biologist Thomas Cavalier-Smith, the first eukaryotes were large, anaerobic, heterotrophic, possibly motile, unicellular organisms with, at least in primitive form, all the main features of extant eukaryotic cells except cytoplasmic, oxygen-related organelles. The organisms caught their food by active endocytosis and digested it intracellularly within lysosomes. For all we know, such organisms could have arisen long before the 2.3-billion-year limit set by the oxygen divide; and they could have covered Earth with a wealth of thriving varieties, of which all but one disappeared without leaving any fossil trace or lasting progeny.
Some years ago, investigators thought they had found descendants of ancient witnesses to this history. Two groups of unicellular organisms devoid of mitochondria, diplomonads and microsporidia, were found by sequencing to go back to particularly remote times. The organisms possess all the other characteristics of eukaryotic cells; they thus looked for all the world like descendants of a line that had detached from the hypothetical primitive eukaryotes before the acquisition of endosymbiotic organelles. Alas! The same techniques of molecular sequencing have dashed the hopes they had raised. Not only has the great antiquity of the organisms been questioned by some investigators, but genes of mitochondrial origin have been identified in their nuclei. If they lack mitochondria, this is apparently not because they never had such organelles but because they have lost them. At present, no eukaryotic cell derived from an ancestor that never possessed mitochondria (or mitochondria-related organelles, such as hydrogenosomes, see below) is known.
This, however, hardly justifies the conclusion, sometimes drawn, that primitive eukaryotes of the type envisaged never existed. Absence of evidence is not evidence of absence. It is an incontrovertible fact that the main characteristics of eukaryotic cells were acquired at some stage in the history of life. How this happened is not known. The hypothetical scenario proposed above for this transformation may be totally wrong. But, at least, it is consistent with the meager clues available and has the merit of providing a possible selective driving force for the process. The truth is that no alternative model has yet been put forward. As to the timing of the transformation, there is a strong reason for putting it before rather than after the acquisition of endosymbionts: the transformation offers the most likely mechanism by which this phenomenon could have occurred.
As we saw in the preceding chapter, endosymbiont adoption is often visualized as the outcome, largely unexplained, of a fateful encounter between two bacterial species. However, if there is any truth in my proposed model, the encounter was not between two bacteria, but between a bacterium and a primitive eukaryote of the kind just described. In fact, it was not an encounter in the usual sense of the word, but rather the capture of a passive victim by an active hunter. How such events may be pictured will be briefly considered in the following pages.
The main organelles known to be derived from endosymbiotic bacteria are the mitochondria, which are the sites of the principal oxidation reactions linked to the assembly of ATP (oxidative phosphorylations), and the chloroplasts, which are the agents of photosynthesis in unicellular algae and plants.
Among the many pieces of evidence supporting the bacterial origin of these organelles, the most convincing is the presence, in mitochondria and chloroplasts, of still-functional vestiges of an ancient genetic apparatus of prokaryotic character. This apparatus consists of a small number of genes, rarely exceeding a few tens, and all that is needed to replicate, transcribe, and translate these genes. Other clues include metabolic similarities and genetic kinships between the organelles and certain extant bacteria.
In the opinion of the vast majority of investigators, these proofs are conclusive. Even the organelles’ bacterial ancestors or, to be more precise, their closest relatives among present-day bacteria have been identified. They are, for the mitochondria, aerobic organisms known under the name of α-proteobacteria and, for the chloroplasts, cyanobacteria, those photosynthetic bacteria believed to be responsible for the first generation of atmospheric oxygen (see Chapter 8).
Possibly also derived from endosymbionts—but this is far from certain—are the peroxisomes, which accomplish oxidative reactions of primitive character that, contrary to those that occur in mitochondria, are not coupled to the assembly of ATP. The origin of peroxisomes is not known. It has been suggested that they also may have arisen from endosymbiotic bacteria. This possibility is consistent with the metabolic activities of the organelles. Added together from the properties of all known members of the peroxisome family,7 these activities cover a very wide range, as would be expected of an autonomous ancestral organism. However, peroxisomes contain no trace of a genetic system, and the molecular data obtained so far are ambiguous.
The common ancestor of eukaryotes almost certainly contained both mitochondria and peroxisomes. The two kinds of organelles are present, in one form or another, in the vast majority of present-day eukaryotic cells. The rare exceptions, of which several are already known to be the outcome of evolutionary regressions, do not suffice to invalidate the generalization. As to the chloroplasts, they were probably not present in the common ancestor but were acquired later in the branch leading to the photosynthetic eukaryotes. The inverse hypothesis, a loss in the non- photosynthetic branches, seems less plausible. It is noteworthy, in this connection, that chloroplasts generally contain a larger number of genes than mitochondria, possibly indicating a more recent adoption.
Granted the nature of the postulated host cells, it seems most likely that the endosymbiont ancestors were originally caught by phagocytosis (from the Greek phagein, to eat, and kytos, cell), a term coined by the Russian- French zoologist and immunologist Ilya Metchnikoff, famous for the discovery that white blood cells protect against bacterial infections by engulfing the disease-causing bacteria and destroying them intracellularly. As already noted by Metchnikoff, this function is just a particular specialization of a more general process used by heterotrophic protists—and by our hypothetical primitive eukaryotes, which, for this reason, are sometimes referred to as primitive phagocytes—for the capture of food. As we have seen, the term endocytosis, actually derived from phagocytosis,8 now designates the more general process, which, from its original relationship to food uptake, has become adapted in higher eukaryotes to a wide variety of functions. Lysosomes are the sites in which the captured materials are digested (and where the caught bacteria are killed and broken down).
Rare cases are known in which phagocytic capture is not followed by the death and digestion of the engulfed bacteria. The agents of tuberculosis and leprosy, for example, are not killed by the cells that catch them, but, instead, settle within those cells and proliferate. The colonized cells react to this proliferation by growing into giant cells, characteristic of the diseases, but eventually succumb. Exceptionally, the eating cells and their prey both survive and establish a relationship of mutual tolerance. It may happen that the partners of such associations, having lost some essential property, become dependent on each other for their survival. The relationship then becomes authentically symbiotic. Although rare, such phenomena are sufficiently frequent to have prompted the creation of a new discipline, endocytobiology, whose specialists meet regularly to compare their findings.
Added to the fact that the endocytic way of taking up extracellular material is a general eukaryotic property, which is not shared by any prokaryote and therefore must have developed in the course of the prokaryote-eukaryote transformation, all this evidence builds a compelling case in favor of the view that the ancestors of the organelles were indeed caught by phagocytosis, as surmised, rather than as the result of some fateful encounter between two kinds of bacteria.
The most striking feature of the endosymbionts, as compared to their bacterial ancestors, is that they have lost the greater part of their genes. Some of these genes probably turned out to be redundant, given the extensive support provided by the host cell, and just disappeared, in a kind of genetic “streamlining.” A number of the genes, however, were transferred to the nucleus of the host cell, there to continue their function.
How this may have taken place is not too difficult to visualize. It no doubt happened from time to time that injured bacterial guests spilled out their DNA into the cytoplasm of the host cell. On the other hand, it is known from present-day technology (see Chapter 15) that DNA molecules introduced into the cytoplasm may occasionally enter the nucleus and become integrated within the genome, henceforth to be replicated and transcribed like the cell’s own genes. Finally, plenty of time was available for experimentation, since bacterial cells with an intact genome remained present and could survive indefinitely with the help of their own copy of the transferred gene. Only after this copy had become redundant could it fall victim to the streamlining process mentioned above.
So far so good. But there is a hitch. Once integrated into the nucleus, the transferred gene behaves like a nuclear gene. It is transcribed locally, and the resulting messenger RNA is, like all messenger RNAs formed in the nucleus, delivered into the cytosol of the host cell, where it instructs ribosomes to synthesize the corresponding protein. This protein thus lands in the cytosol of the host cell, not inside the bacterial guest where it is needed. If the bacterium cannot do without the protein, the gene transfer cannot be completed until some mechanism has arisen whereby the protein (or the messenger RNA) can be transferred into its erstwhile owner.
As we now know, the transfer involves the proteins, which are directed to their site by specific mechanisms dependent on targeting sequences. The development of these sequences and that of the appropriate receptors and machineries on the surface of the captive bacteria probably represented the greatest challenge to endosymbiont adoption. Possible mechanisms involving the bacterium’s secretory machineries have been considered but are too specialized for the present account. Let it simply be stated that gene transfer has occurred on a very large scale, to the point that only a small number of the original genes, rarely exceeding a few tens, has remained in the endosymbionts. This fact calls for several comments.
First, the finding that such massive gene transfer has happened allows the hypothesis, evoked earlier, that peroxisomes also have an endosymbiotic origin, even though they contain no trace of a genetic apparatus. If, as seems likely, peroxisomes were acquired before mitochondria, they could have lost all their genes, whereas mitochondria still have retained a few. Consistent with this possibility is the fact that chloroplasts, which were probably adopted after mitochondria and peroxisomes, have conserved a larger number of their original genes.
The occurrence of gene transfer from endosymbiont to nucleus has also provided a valuable tool for probing the past. Genetic vestiges of a vanished endosymbiont may still be left in the nucleus of a cell and may reveal the endosymbiont’s erstwhile presence. This is how it was found that the ancestors of organisms lacking mitochondria did once possess such organelles.
Finally, the fact that gene transfer actually took place on such a large scale in spite of the obstacles it encountered is clear proof that this phenomenon must have been crucially important for the successful adoption of endosymbionts. Why this should be so is readily understood. Bacteria multiply much faster than eukaryotic cells. Unless the multiplication of the captured bacteria could somehow be curbed, they would inevitably overwhelm and stifle their host cells. We have seen that this happens to cells invaded by the bacterial agents of tuberculosis and leprosy, despite the fact that the cells respond to the invasion by greatly increasing their size. Removing an essential gene from the captives and transferring it to the captor’s nucleus offers a particularly simple way of adapting the multiplication of captive cells to that of their captors. In the nucleus, replication of the transferred gene becomes synchronized with that of the nuclear genes, so that the captives are forced to adopt the captor’s multiplication rate.
This is not the only advantage. For the captives to lose their independence and become increasingly integrated within the host cell’s economy, nothing could have been more efficient than the transfer of their genes to their host cell’s nucleus. It is evidently very advantageous for a host cell to have endosymbiont genes, of which there previously existed up to thousands of copies, housed in as many semi-independent entities, reduced to single copies present in the nucleus, where their replication can be coordinated with that of the host’s genes and their transcription subjected to centralized controls.
These facts leave one important question unanswered. What is it that made the endosymbionts so vitally important to their host cells that all eukaryotes lacking endosymbionts seem to be extinct?
We have seen that, according to the latest evidence, no eukaryotic cell is known that does not have in its ancestry cells that contained mitochondria. This fact strongly suggests that mitochondria offered an enormous selective advantage, perhaps even a vitally important one, to their possessors, so that all the primitive eukaryotes that did not acquire these organelles were eliminated by natural selection. It has long been assumed that protection against oxygen toxicity made up this advantage. This explanation, which was already favored by Margulis in her early advocacy of the endosymbiont theory, is consistent with the hypothesis, evoked earlier, that oxygen poisoning wiped out all the primitive eukaryotes except those that had acquired endosymbionts.
Applied to mitochondria, however, the explanation does not hold water. Mitochondria, together with the a-proteobacteria with which they share the nearest common ancestor (see above), contain the most sophisticated oxygen-utilizing systems found in nature. True marvels of molecular organization, with an ATP yield near the maximum authorized by the laws of thermodynamics, these systems can be but the products of a very long evolution. This makes it very unlikely that the mitochondria could have saved the primitive anaerobic eukaryotes from the deadly oxygen attack. By the time the bacterial ancestors of these organelles had developed their sophisticated systems, the cells they are assumed to have saved would long have succumbed to the oxygen holocaust.
This does not necessarily invalidate the oxygen bottleneck hypothesis. But we must look for more primitive rescuers. The peroxisomes appear as excellent candidates for this function. Indeed, their properties are very much what would be expected of a primitive system of protection against the toxic gas. Their enzymes do nothing but convert oxygen and its noxious products into harmless water molecules, doing this by means of simple reactions that, unlike those that take place in mitochondria, are not coupled to the assembly of ATP. Peroxisomes or their close relatives are, like mitochondria, present in the vast majority of eukaryotic cells. It is thus perfectly possible that they were acquired before mitochondria. We have seen that the possible endosymbiotic origin of peroxisomes is at present a moot question. But this does not fundamentally change the proposed hypothesis. Even if peroxisomes were acquired in a different way, they could still have protected their owners against oxygen toxicity.
Granted this possibility, the fact remains that mitochondria must have provided a sufficiently powerful advantage to the cells that acquired them that natural selection eliminated all the cell types that did not enjoy this benefit, as seems to be the case. It is tempting to assume that mitochondria owed their selective value to their remarkable energetic efficiency. Peroxisomes, remember, contain no ATP-retrieval system. Their sole advantage, in terms of energy, would have been to provide the cytoplasm of their host cells with additional fuel arising from the fatty acids and other materials that only they are able to metabolize. For the actual generation of ATP, the cells endowed with peroxisomes remained entirely dependent on the coupled ATP-generating systems that support anaerobic metabolism. In such a context, the kind of oxidative machineries provided by the mitochondria represented a tremendous asset, possibly sufficient to explain why they would be retained by natural selection.
If this theory is correct, we may well ask why the acquisition of mitochondria did not drive out the more primitive peroxisomes. And, especially, why did no cell fitted only with peroxisomes survive? The answer to the first question is simple. By the time mitochondria were adopted, peroxisomes may have become indispensable because they were carrying out reactions that the newcomers could not perform, in particular in lipid metabolism, where peroxisomes are known from human pathology to accomplish vitally important functions (see Chapter 9). The fact that peroxisomes did not disappear after the adoption of mitochondria could thereby be explained.
As to the second question, the intensity of the selective pressure may provide the answer. If competition for available resources was fierce enough, only the better-equipped cells would be expected to survive. Note, however, that our knowledge of unicellular eukaryotes is still far from exhaustive. Perhaps representatives of the missing intermediates are still waiting to be found. Such a discovery would be most revealing.
As will be mentioned later, a new, startling theory, based on the production of molecular hydrogen by the ancestors of mitochondria, has been proposed to explain the adoption of these organelles. Before we consider this new theory, a brief look at the chloroplasts is in order.
We have seen that chloroplasts are derived from cyanobacteria, the oxygen-generating photosynthetic organisms believed to be responsible for the oxygen holocaust. According to all available evidence, the mechanisms involved in the uptake of these organisms and in their integration, including the massive transfer of genes to the nucleus and the development of specific protein-targeting mechanisms, must have been very similar to those involved in the adoption of mitochondria. There are good reasons to believe that the cells that did the acquisition already possessed peroxisomes and mitochondria. First, all the cell types that contain chloroplasts also contain the other two kinds of organelles. In addition, it is difficult to see how cells not properly protected against oxygen toxicity could possibly have come to harbor guests that actually produce the toxic gas.
The cells that adopted chloroplasts became the first unicellular algae, which, in turn, are ancestral to the pluricellular plants (see following chapter). Considered from an evolutionary point of view, the adoption of chloroplasts poses no special problem. The advantages the cells derived from their new acquisition are obvious. Henceforth freed from the obligation to find food, they housed photochemical factories that, in the presence of light, allowed them to live on water, carbon dioxide, and a few mineral salts. The benefits were immense, but not to the extent of creating a necessity. Cells devoid of chloroplasts continued to thrive, supported by their photosynthetic relatives, which became their food supply. Thus were born the main groups of unicellular eukaryotes out of which the whole visible part of the living world was to emerge.
This question has been posed in recent years as a result of startling findings indicating that hydrogenosomes, those hydrogen-generating organelles already briefly mentioned in the preceding chapter, may be genetically related to mitochondria. The metabolic properties of these organelles hardly would have suggested such a possibility. Present in a small number of protists and fungi devoid of mitochondria, hydrogenosomes lack all the characteristic oxidative machineries of mitochondria. Their most typical property, which is absent in mitochondria, is the ability to generate molecular hydrogen anaerobically by a reaction linked to the assembly of ATP. In the presence of oxygen, this hydrogen is diverted toward the formation of water by an oxidizing system of primitive character. Thus, organisms endowed with hydrogenosomes can develop under anaerobic conditions, their usual habitat, but are also able to tolerate oxygen, if necessary, and, even, to take advantage of it. They are facultative anaerobes.
Hydrogenosomes do have some properties in common with mitochondria: they are surrounded by two membranes and they have been found in one case (see below) to contain a vestigial genetic machinery; especially, they share some genes with mitochondria. This is the discovery that has led to the conclusion that the two organelles have a common ancestry.
If such is the case, the question arises as to which metabolic properties characterized their common ancestor. In view of the kinship of mitochondria with a-proteobacteria, revealed by molecular sequencing data, there can be little doubt that their ancestor already possessed the sophisticated ATP-generating oxidizing systems they share with these organisms. In any case, it is hardly conceivable that mitochondria could have developed such elaborate systems independently, after their adoption as endosymbionts. On the other hand, the fact that hydrogenosomes have been found in several distantly related protists and, even, in some fungi indicates that the ability to produce molecular hydrogen must likewise be of ancient origin and may also have belonged to the putative bacterial ancestor hydrogenosomes have in common with mitochondria. Thus, the ancestor seems to have combined the main properties of both of these organelles.
One is thus faced with a strange case of evolutionary divergence. Starting from an ancestor simultaneously endowed with highly efficient oxidizing systems and with an anaerobic hydrogen-generating machinery, the vast majority of organelles would have kept only the former and lost the latter, becoming mitochondria. A small minority would have done the opposite and given rise to hydrogenosomes. None would have retained both machineries. A divergent adaptation to aerobic and anaerobic milieus could conceivably explain this occurrence, which does, however, remain puzzling.
The new findings also raise another intriguing question: which of the two properties offered the selective advantage host cells derived from adopting the ancestors of the organelles? All earlier theories have invoked the possession of oxidizing systems with a high ATP yield as the main benefit. This is what was suggested earlier. But there is now the alternative possibility that it was the ability to produce hydrogen that made the endosymbionts useful to their host cells.
A theory based on this second eventuality has been proposed by the discoverer of hydrogenosomes, my erstwhile collaborator and present colleague at the Rockefeller University in New York, Miklos Müller, together with an American investigator stationed in Germany, William Martin. As suggested by these workers, the host would have been an organism related to present-day methanogens. These microbes (see Chapter 8) are strictly anaerobic, autotrophic archaebacteria that use molecular hydrogen to convert carbon dioxide into methane by a reaction coupled to the assembly of the ATP they need to satisfy their energy requirements. According to the proposed theory, the benefit host cells derived from the endosymbionts was the hydrogen they needed as fuel for making ATP, not ATP itself.
A detailed discussion of the two competing theories does not belong in this book. Let it simply be pointed out that the model based on hydrogen supposes an encounter between two typical bacteria. Like other fateful encounter models, it does not include the participation of a primitive phagocytic host cell and says nothing about the manner in which all the main properties of eukaryotic cells could have arisen. The model thus needs at least to be completed. The two theories could be reconciled if the primitive eukaryote envisaged in this chapter happened to derive some advantage from a hydrogen-producing endosymbiotic partner, as is supposed by the new model. Unfortunately, no eukaryotic organism answering this description is known. This does not mean that none ever existed.
Another possibility that deserves to be considered is that the postulated symbiotic association did occur between two kinds of bacteria, as assumed, but took place inside a primitive eukaryote, which somehow benefited from hosting the two partners. Interestingly, an association of this kind actually exists. Some cockroaches harbor in their hindgut a parasitic protist that contains hydrogenosomes and, in close contact with them, endosymbiotic methane-producing bacteria that obviously take advantage of the hydrogen produced by the neighboring organelles. The hydrogenosomes involved in this suggestive threesome have the additional distinction of possessing a vestigial genome.9
The birth of eukaryotic cells, with all their extraordinary, finely tuned attributes, so different from their “simple”—tout est relatif—prokaryotic relatives, is often depicted as the outcome of highly improbable events, one of the major hurdles on the way to humankind, one, perhaps, if the defenders of intelligent design are to be believed, that could not have been overcome without the help of “something else.”
This view is understandable; but it is unfounded. Whatever value may be attached to the evolutionary models offered in this chapter, they have at least the merit of showing that the development of eukaryotic cells can be explained in terms of natural processes likely to occur when and where they did and to lead to acquisitions retained by natural selection. No doubt, the speculations offered will need to be amended, perhaps abandoned. But their very plausibility should encourage further search for natural explanations. It is far too early to call on “something else.”
Also worth recalling is that the single ancestry of the eukaryotic world may not be, as is often claimed, a reflection of its rarity and improbability but could be the simple consequence of a bottleneck, a term that has come up several times already in this book. For all we know, the pre-oxygen world may have harbored a wealth of different eukaryotic organisms. Granted that there is no proof of this possibility, it still deserves to be kept in mind. Perhaps some day, molecular sleuthing may become incisive enough to test its reality.
Another point to be remembered from this chapter is that eukaryotic cells are not mere associations of bacteria, contrary to an affirmation sometimes made by the unconditional admirers of these organisms (see Chapter 8) and implying, or even stating explicitly, that, since we ourselves are associations of eukaryotic cells, we are “nothing but” associations of associations of bacteria. This view is misleading. No bacteria are known that possess the main features of eukaryotic cells. Any theory claiming to account for the origin of these cells cannot just invoke fateful encounters between prokaryotes. It must provide a plausible explanation for the development of those main features in ancestral cells derived, whether chimerically or otherwise, from prokaryotic cells.