WHAT IS COMMON among a yeast, a diatom, a tobacco leaf cell, and a human neuron? At first sight, even with the help of a good microscope, not much. To be sure, all four are cells, and they have a nucleus, which puts them in the category of eukaryotes, as opposed to prokaryotes, or bacteria. But, otherwise, the differences among them are enormous. The yeast cell is small, the size of a big bacterium, and it shows few inner components. The diatom is big and complex; especially, it is surrounded by an elaborate mineral shell of exquisite design.1 The leaf cell, which inhabits a small, rigid cellulose chamber (Hooke’s original cell, see Chapter 1), is conspicuously filled with green particles. As to the neuron, its most striking feature is represented by remarkable extensions, some exceedingly slender and amazingly long—up to one meter or more—others branching into a dense bush of ramifications.
Yet, modern electron microscopy has revealed within all four cells a complex set of inner structures that are the same in all four and are not observed in bacteria. Thanks to advances in biochemistry, the functions of these structures, also the same in all four cells, have been unravelled. Finally, with the progress of molecular biology, many of the genes behind the common structures and functions have been sequenced and found to be related in all four cell types. In fact, all available evidence indicates that the four cells considered are, together with all other eukaryotes, the descendants of a single ancestral form. This form should not be confounded with the common ancestor of all life referred to in the beginning of this book (see Introduction and Chapter 3). As will be pointed outlater, we are dealing here with a new and much later evolutionary bottleneck, from which are issued all eukaryotic organisms, to wit: a collection of unicellulars grouped under the name of protists and the three great groups of pluricellulars, the plants, fungi, and animals, including human beings.
Eukaryotic cells share all the basic properties of prokaryotes. They are made of the same kinds of chemical components, use similar metabolic pathways, and depend on the same genetic mechanisms. But they are much more voluminous than prokaryotes, and their internal organization is so much more complex that their origin from bacterial ancestral forms seems almost incredible. This, nevertheless, is what must have happened, as there is abundant proof, irrefutably imprinted in their molecular char-acteristics, that the two are related.
No fossil trace of this momentous transformation has yet been uncovered. Nor, as far as is known, has any of the intermediates the transformation must have involved left a legacy that has stretched unto this day, thus allowing its characterization. We have only our knowledge of the two cell types to attempt a reconstruction of the pathway between them. Before embarking on such an attempt, which will be the subject of the next chapter, let us take a look at what is known.
Eukaryotic cells, even the simplest of them, are organisms of considerable complexity, composed of many different parts endowed with distinct functions. In the description that follows, I have restricted myself to the properties that are common to the vast majority of eukaryotic cells, keeping details to the minimum required for a meaningful discussion of the cells’ origin. For more information, readers can consult any of the many good textbooks of cell biology that are now available.
Perhaps the most important property distinguishing eukaryotic from pro- karyotic cells is the division of the former into two structurally and functionally distinct parts, a highly organized pulp, called cytoplasm, and a central kernel, the nucleus. Whereas the cytoplasm is the site of the vast majority of metabolic processes, including protein synthesis, the nucleus centralizes the bulk of genetic operations. The nucleus is notably where genes are stored and transcribed into messenger RNA molecules and where these molecules are further processed. In addition, the nucleus is the site of DNA replication when cells prepare to divide, and it subsequently participates in a complex set of processes leading to the formation of two nuclei.
This characteristic partition of functions does not exist in prokaryotes, in which ribosomes are commonly seen in the process of translating into proteins genetic messages that are still being transcribed from chromosomal DNA. It all takes place in a single phase. In contrast, in eukary- otes, transcription and translation are topographically separated.
Eukaryotic cells are much larger than prokaryotes. Their diameter is usually on the order of 20 to 30 thousandths of a millimeter, with exceptions that are sometimes large enough to be visible to the naked eye. This means that a typical eukaryotic cell has a volume equivalent to that of some ten thousand prokaryotic cells. Relatively speaking, it is enormous.
Eukaryotic cells surround themselves with a great diversity of outer coverings, which, in unicellular organisms, may vary from a thin slimy coat to structures of enormous complexity (remember the diatom mentioned in the beginning of this chapter). Pluricellular organisms all rely for their architectural support on a complex tangle of extracellular structures assembled from materials secreted by the cells. The cells themselves are just soft blobs. Deprived of their extracellular scaffoldings, trees, whales, humans, and the rest of visible life would be no more than shapeless masses of gooey stuff.
Chemically, the outer coverings of eukaryotic cells are made of a wide variety of substances, mostly proteins, carbohydrate polymers, or combinations of both. In a number of cases, as in diatoms, bones, or mollusc shells, the structures are reinforced by minerals. Interestingly, no eukaryotic cell has an outer covering made of murein or pseudomurein, the substances that make up the cell walls of eubacteria and archaebacteria, respectively. There seems to be a distinct gap at this level between eukaryotes and prokaryotes.
Eukaryotic cells are, like their prokaryotic relatives, surrounded by a typical membrane, called plasma membrane. Contrary to what happens in prokaryotes, where the cell membrane may harbor important metabolic processes, such as, for example, ATP-generating oxidations, eukaryotic plasma membranes are almost exclusively specialized to serve as actively discriminating boundaries. They are richly fitted with transporters and receptors involved in exchanges and communications with the outside world (see Chapter 6). Similar phenomena take place in prokaryotes, but they are much more numerous and diversified in eukaryotes, as befits the greater complexity of the cells.
The cytoplasm of eukaryotic cells is occupied by a thick, semi-fluid material, termed cell sap or cytosol, closely related to the milieu that fills bacterial cells, except that it is occupied, in addition, by many structural components that do not exist in prokaryotes (see below). Like the inner medium of prokaryotes, the eukaryotic cytosol houses numerous metabolic systems of major importance. It is a concentrate of protein enzymes, coenzymes, and chemical intermediates involved in a variety of syntheses, breakdown processes, and transforming reactions, which may number up to several thousands. Interestingly, most of these reactions occur without the participation of oxygen. As will be seen, the utilization of this gas is largely the prerogative of special particulate entities embedded in the cytosol. This separation between anaerobic and aerobic processes is an important clue to our understanding of eukaryotic history.
The cytosol also contains, like its bacterial counterpart, the bulk of the ribosomes. These particles are made in the nucleus, and they also receive from the nucleus the messenger RNA molecules that provide them with the blueprints of the protein molecules they assemble. Protein synthesis itself takes place in the cytosol, from which the proteins are directed toward their final location within the cells by special targeting mechanisms (see below).
All eukaryotic cells contain a number of membrane-enclosed compartments of varying shapes and contents, organized into several distinct subsystems endowed with characteristic structural and functional specializations and forming together what is known as the cytomembrane system.2 Each of these compartments is completely surrounded by a membranous envelope. They are nevertheless able to share and exchange contents by means of transient intermembrane connections, so that materials can circulate from compartment to compartment by vesicular transport,always separated from the cytosol by a closed membrane.
Processes also exist whereby new intracellular vesicles may detach from invaginations (infoldings) of the plasma membrane, enclosing within their insides whatever extracellular materials have been trapped in the invagination. This mechanism, called endocytosis (Greek for “into the cell”), introduces extracellular materials into the cytomembrane system. Conversely, existing vesicles may join with the plasma membrane, thereby discharging their contents outside the cells, while their envelope is added to the plasma membrane. Known as exocytosis (Greek for “out of the cell”), this process allows materials present within the cytomembrane system to be discharged out of the cell.
Thanks to the transient connections that link its various compartments with each other and with the extracellular milieu, the cytomembrane system is the site of a dual traffic concerned with what may be designated import-export. In the import direction,3 all sorts of complex molecules and objects are engulfed from the outside by endocytosis and introduced into the system, most often to be broken down in specialized pockets, called lysosomes (from the Greek lyein, to dissolve, and sôma, body), within which the engulfed materials are exposed to an acid juice containing powerful digestive enzymes capable of fragmenting all major biological constituents, as do the digestive juices of our stomach and intestinal tract. The products of this chemical breakage are able, thanks to their small molecular size, to traverse the lysosomal membrane and reach the cytosol, where they enter general metabolism. Initially serving for the capture and utilization of food, this mechanism has been adapted to a variety of functions, including, as will be seen in the next chapter, the fight against pathogenic bacteria.
The export part of the system4 serves in the manufacture and discharge of secretion products. These are mostly proteins made by membrane- bound ribosomes (see protein targeting, below) that inject their products directly into a cavity of the system. After their synthesis, the proteins travel through various specialized parts of the cytomembrane system, undergoing a number of chemical changes, as well as additions, notably of carbohydrate components, to be finally unloaded outside the cell by ex- ocytosis. The export system also makes the enzymes involved in intracellular digestion. But these, thanks to a special targeting mechanism, are discharged into the lysosomes instead of into the outside medium.
Another characteristic feature of eukaryotic cells is the presence of various intracellular structural elements that prop up the cells and consolidate them internally, preventing them from collapsing under their own weight and helping them to conserve their shape. Grouped under the general name of cytoskeleton, these elements are made of special protein molecules.
In most instances, the proteins involved conserve the linear conformation of their peptide chains and join, like the strands of a string, to form various filaments, which are sometimes twisted, woven, or interlinked into three-dimensional structures of diverse shapes. In two particularly important groups of structural elements, the protein chains are, like most proteins, folded into globular units. These spontaneously assemble into elongated structures thanks to special mutually binding sites—of the mortise-tenon kind (see Chapter 1)—with which they are equipped. Actin fibers5 and microtubules6 are the two main structures formed in this way.
Often associated with the cytoskeleton are motor elements that allow the cells to shift their inner parts with respect to each other and, in some cases, to move around as whole cells. The active components of these systems are special transducing proteins that convert to mechanical work the energy released by the splitting of ATP.7
Such associations are sometimes organized, with a number of additional proteins, into elaborate propulsion organelles. Among these, a particularly important group is represented by cilia and flagella, which, from protists to higher animals, are involved in a wide variety of cell movements. The cilia are short and act by beating. The flagella are long and have an undulating movement. Both organelles depend on the same characteristic structure, a highly complex arrangement of microtubules and proteins.8 Note that many bacteria, spirochetes, for example (see Chapter 8), are also equipped with flagella, but these are totally different from eukaryotic flagella. Instead of undulating, bacterial flagella rotate, at the astonishing speed of several thousand revolutions per minute. Their shafts are rigid, corkscrew-shaped structures made of proteins that bear no relationship to the proteins of eukaryotic flagella.
Another machinery made of microtubules and motor systems is the mitotic spindle, the structure involved in the separation of daughter chromosomes in cell division, as will be described later.
Animal locomotion depends on molecular arrangements very different from those involving microtubules. Myofibrils, the functional units of animal muscles, have actin as their cytoskeletal element and operate by shortening (contraction).9
The cytoplasm of most eukaryotic cells contains a variable number—up to several thousand in some cases—of discrete, particle-shaped, membrane-bounded organelles (small organs) typically connected with oxygen metabolism. Several distinct types of such organelles are known.
Peroxisomes These are small granules, about one half of one thousandth of a millimeter in diameter, bounded by a single membrane, often described in the morphological literature by the name of microbodies. Present in one form or another in the vast majority of eukaryotic cells, peroxisomes are involved in a number of oxidative reactions in which oxygen is utilized by way of hydrogen peroxide—hence their name—and without the coupled assembly of ATP. Their role in the supply of energy is therefore limited. They nevertheless accomplish important functions in the metabolism of a number of substances, prominently lipids, as revealed by human pathology.10 Some peroxisomes are known under a different name recalling a special metabolic property.11
Mitochondria These organelles, likewise utilizing oxygen and almost universally distributed among eukaryotic cells, are particles about the size of bacteria—we shall see that this is more than a fortuitous coincidence—surrounded by two membranes of which the inner one expands into numerous foldings, or cristae (ridges). Their shapes vary from almost spherical to slenderly filamentous, explaining their name, which comes from the Greek mitos (filament) and chondros (grain). Often referred to as the cells’ “power houses” for this reason, mitochondria are, in the whole eukaryotic world, the central sites of oxidative phosphorylations, that is, cellular oxidations coupled with the assembly of ATP. The systems involved in these all-important processes are composed of a complex set of electron carriers,12 or respiratory chain, embedded in the inner mitochondrial membrane.
Hydrogenosomes These are membrane-bounded organelles, present only in a few selected protists and fungi occupying habitats that are poor in oxygen but not necessarily totally devoid of this gas. Probably related to mitochondria (see next chapter), hydrogenosomes have as a remarkable property, which accounts for their name, the ability to produce molecular hydrogen in the absence of oxygen. They can also utilize oxygen.
Chloroplasts This name, derived from the Greek chlôros (green) and plastos(fashioned), designates the characteristic photosynthetic organelles of unicellular algae and green plant cells. Larger than mitochondria and similarly surrounded by two membranes, chloroplasts are filled with stacks of flattened membranous sacs, the thylakoids, which contain the light-utilizing systems. It will be remembered that the reactions catalyzed by these systems are associated with the production of molecular oxygen, as are those that occur in cyanobacteria (see preceding chapter). The name plastids is given to colorless chloroplast relatives present in plant cells that do not carry out photosynthetic processes.
The eukaryotic nucleus is separated from the cytoplasm by a double- membranous envelope belonging to the cytomembrane system, reinforced by an inner framework, the lamina, made of cytoskeletal elements. This assemblage is pierced with pores that mediate all the exchanges between nucleus and cytoplasm. The genetic material is confined within this envelope, organized with proteins into a number of discrete entities, the chromosomes. The nucleus is basically an RNA factory. In it, all the different kinds of RNAs that take part in protein synthesis and other functions are synthesized by transcription of the corresponding DNA genes and further processed by splicing and other modifications. With the exception of a few RNA molecules involved locally in this processing, all these RNAs are conveyed to the cytoplasm, where they carry out their various functions. This transport takes place through the pores of the nuclear envelope, with the help of special protein carriers that are allowed inside in naked state and return to the cytoplasm in combination with the RNAs.
A special intranuclear structure, the nucleolus, is the site of a particularly important activity, the manufacture of ribosomes, which, being relatively short-lived, have to be continually provided to the cytoplasm in large quantities. The nucleolus houses the production and processing of ribosomal RNAs and their association with proteins imported from the cytoplasm. The ribosomes thus formed are delivered into the cytoplasm, ready to carry out protein synthesis. The nucleus itself does not make proteins and is virtually devoid of metabolic and energy-yielding systems. All the NTP building blocks used for the synthesis of RNA (and DNA, see below) come from the cytoplasm through the nuclear pores. All nucleo-cytoplasmic exchanges take place through these openings. This complex, two-way molecular traffic is strictly controlled by highly elaborate mechanisms.
Contrary to prokaryotes, which tend to multiply exponentially whenever conditions allow, eukaryotic cells can remain without dividing for a long time, even indefinitely, subject to complex controls regulating what is known as the cell cycle. Special triggers—of burning interest to all those who study cancer, which is essentially due to unchecked cell multiplication—awaken the cells from this stationary phase and stimulate DNA replication in the nucleus, leading to doubling of the chromosomes. The nuclear envelope subsequently breaks into pieces and is replaced by an impressive scaffolding, called the spindle because of its shape. This structure, made largely of microtubules combined with special motor elements (see above), acts mechanically to separate the two chromosome sets and drag one set toward one spindle pole and the other to the opposite pole. There, a remarkable self-assembly process surrounds each chromosome set by a newly formed envelope, with, as a result, the genesis of two identical nuclei. This mode of cell division is called mitotic division, or mitosis.
With very rare exceptions,13 all the proteins of a eukaryotic cell are made in the cytosol and are conveyed to their final location within the cell by a variety of mechanisms that all rely on a specific interaction (“recognition”) between a short amino acid sequence (targeting sequence) of the protein molecule and a receptor present on the target’s surface. Illustratively described as postal addresses utilized by the cellular mailing systems, these targeting sequences represent an additional piece of information written into the protein sequences, which, thus, not only determine the structural and functional properties of the molecules (see Chapter 1), but also specify the molecules’ location inside the cell.
Targeting sequences may work co-translationally or post-translationally,that is, while the protein is being synthesized (while the RNA message is being translated), or afterwards. In the former case, the protein-making ribosome is “pinned,” so to speak, to a membrane surface by the end of the protein chain it is in the process of synthesizing. This end carries a targeting sequence that interacts with a specific membrane receptor. As a result of this interaction, the membrane rearranges locally into a tunnel through which the growing peptide chain is directly injected into the lumen of the compartment delimited by the interacting membrane. This mode of transfer is characteristically involved in the synthesis of secretory proteins (see above). Membranes with ribosomes attached in this manner have a rough appearance in cross-section.14
Post-translational transfer takes place in most other instances; it conveys the finished proteins to their intracellular location by means of sophisticated mechanisms that often involve the participation of special proteins known as “chaperones.” Peroxisomes, mitochondria, chloro-plasts, and the nucleus all receive their proteins by post-translational transfer, dependent in each case on different targeting sequences. Vesicular transport may also be directed in certain cases by targeting sequences. Lysosomal enzymes, for example, are diverted from the secretion machinery to their destination with the help of such sequences (see above).
Large size, a cytosol richly endowed with enzymes involved mostly in anaerobic metabolism and, in addition, housing the protein-synthesizing ribosomes, compartmentation by membranes into multiple pockets specialized in import-export exchanges with the external milieu, internal shoring by cytoskeletal structures often associated with motor elements, cytoplasmic organelles involved in the utilization and (only in photosynthetic cells) production of oxygen, organization of the genetic material into structured chromosomes confined within an envelope of both membranous and cytoskeletal nature, strict separation between gene transcription (nucleus) and expression (cytoplasm), mitotic cell division subject to elaborate control: such are the essential characteristics of virtually all eukaryotic cells, be they unicellular organisms or the components of plants, fungi, or animals. No doubt, these characteristics were all present already in the common ancestor of all eukaryotic cells. The genesis of these characteristics from some prokaryote ancestor is what needs to be explained.
Faced with the problem posed in this way, one may well wonder whether its solution will ever be found or, even, whether a “natural” solution exists. Could not the number and complexity of the mutually complementary innovations that have to be explained exemplify the “irreducible complexity” claimed by some to demand the intervention of “something else?”15 Fortunately, we are nowhere near such a stage. As I shall show in the next chapter, modern biology has provided a number of revealing clues that illuminate certain aspects of the problem. Possibly the most significant of these concerns the role of oxygen in the genesis of eukaryotes, a role so important that it warrants special treatment now.
We have seen in the preceding chapter how the development of oxygen- generating photosynthesis must have upset the physical conditions to which all early living forms were adapted, creating one of the most stringent and devastating bottlenecks in the whole history of life. There are good reasons to believe that the genesis of eukaryotes was also drastically influenced by the rise of oxygen in the atmosphere. To begin with, there is the matter of timing.
Two important landmarks flank the history of eukaryotes. On one side are Woese’s sequencing results16 indicating that the line leading to eukaryotes separated from the two prokaryotic lines very early after the tree of life first branched out from its root, that is, if the evidence in Chapter 3 is to be trusted,17 no later than some 3.5—perhaps even 3.8—billion years ago. On the other side, there is strong evidence that the common ancestor of eukaryotes must have been aerobic, as it almost certainly contained mitochondria as well as peroxisomes. If, as geochemical findings seem to indicate, the level of atmospheric oxygen was too low to support aerobic organisms before about 2.3 billion years ago, the common ancestor of eukaryotes cannot have lived earlier than that date.
There thus seems to be, between what may be called the “founder” of the eukaryotic line and the common ancestor of all present-day eukaryotic life, a gap of at least 1.2 billion years, not counting the time it took for oxygen to rise to a level allowing aerobic life. This estimate may need revision in view of the uncertainties surrounding the time of the first appearance of life on Earth. Doubts have also been expressed with respect to the validity of sequencing phylogenies extended, as done by Woese, to extremely ancient times. But the gap is certain to remain huge, at least several hundred million years. During this immense amount of time, one of the most extraordinary and momentous developments in the history of life took place: the transformation of what was most likely a primitive bacterium (see Chapter 3) into a much more complex cell type that, in turn, was to give rise to the rich array of protists and to the whole visible panoply of plants, fungi, and animals, including ourselves. Had this transformation not taken place, the living world of today would still contain only bacteria.
The time limits just mentioned do not necessarily mean that the transformation of a prokaryotic into a eukaryotic cell took the whole of the time allowed by the record, only that this amount of time was available for the transformation. Nor do these limits imply that the transformation followed a single line, determined, as is often assumed, by a number of improbable “quantum jumps” that guarantee its unique character. It is quite possible—I would even say probable—that the ancestral eukaryote emerged from an evolutionary bottleneck preceded by a long, complex history. I shall come back several times to this important question.
Whatever the events that took place during the gap mentioned above, they must, if the evidence referred to is correct, have occurred within the framework of anaerobic life. It is striking in this respect that all the main features of eukaryotic cells that have just been surveyed are, with the exception of the organelles of oxygen metabolism, associated in major part with biochemical mechanisms that do not involve oxygen. The rare exceptions to this generalization—biochemists may think, for example, of the oxygenations and hydroxylations catalyzed by components of the cytomembrane system—could have been late additions. In view of these facts, it is tempting to assume that the main eukaryotic features were acquired before oxygen started rising in the atmosphere.
Strong support for this hypothesis comes from a discovery of capital importance that has projected a most revealing light on the genesis of eukaryotic cells. It is now established beyond reasonable doubt that at least two major organelles of oxygen metabolism, namely mitochondria and chloroplasts, were once free-living bacteria, which were adopted at a remote time by host cells within which they became integrated as endosymbionts, a Greek-derived term that literally means: living (biont) together (sym) within (endo)18
I shall come back at length to this epoch-making event. For the time being, let me simply emphasize an important implication of the endo- symbiotic origin of the organelles concerned. It means that the actual development of these organelles was not part of the eukaryotic transformation. It belongs to prokaryotic history, which still includes surviving descendants of the transformations actors, as will be mentioned later. The organelles were acquired ready-made and only their subsequent integration within the host cells’ economy is part of eukaryotic history.
The origin of peroxisomes is unfortunately not known. It is possible that these organelles also originate from endosymbiotic bacteria. But this cannot be affirmed, as the evidence available so far on this topic is ambiguous. Whatever their origin, peroxisomes must obviously go back, like the mitochondria and the chloroplasts, to a time in the history of life, when oxygen was already present in significant amounts in the atmosphere, that is, less than 2.3 billion years ago.
The apparent division of eukaryotic cells into an anaerobic and an aerobic part suggests a development in two phases. The first, anaerobic phase is assumed to have taken place before the rise in atmospheric oxygen and to have led to the development of anaerobic, heterotrophic cells possessing all the main properties of eukaryotic cells, with the exception of oxygen-linked organelles. This phase may have taken a greater or lesser part of the 1.2 billion-year stretch left open for it. Quite possibly, a wide variety of such cells may have existed, subsisting largely on bacteria.
The second phase of this hypothetical scenario is pictured as being initiated by the emergence of oxygen in the atmosphere. As we saw in the preceding chapter, this event may have signalled a widespread extinction of living forms, sometimes called the oxygen holocaust. It is quite possible that most of the primitive eukaryotes supposed to have arisen in the first phase fell victim to this catastrophe, leaving only those that had acquired the necessary defenses (or had found refuge in an oxygen- free niche). In the end, only one form would have made it safely through the bottleneck and survived the resulting competition, thus accounting for the single ancestry of extant eukaryotes (see above).
Surprisingly, few theories proposed for the origin of eukaryotes postulate a long anaerobic phase, followed by endosymbiont adoption. In popular accounts, the establishment of an endosymbiotic relationship is often presented as the starting point of eukaryote development, or even its triggering event. All is taken to have started with a “fateful encounter” between two different bacteria that established some kind of mutually beneficial association in which, eventually, one became dominant and the other submissive.
If this is what happened, then we are faced with the question as to what took place during the huge span of time that has elapsed between the first branching out of the eukaryotic line and the adoption of endosymbionts. Another difficulty is that the alternative theories that are proposed usually offer no explanation for the acquisition of the many characteristic components of eukaryotic cells other than the organelles of oxygen metabolism. The genesis of these features is rarely addressed by the defenders of the fateful encounter model. Finally, the model throws little light on the manner in which the encounter ended with the enslavement of one participant by the other.
It could be argued that the above objections rest on questionable data. The methods of dating by molecular sequencing results, for example, are the object of vigorous debates. There is also some disagreement on the actual level of atmospheric oxygen before the critical date of 2.3 billion years ago. Some authors believe this level to have been high enough to support some forms of aerobic life. This book hardly lends itself to a detailed discussion of these highly specialized issues. What I believe to be important is that an alternative model, not open to the formulated objections, can be proposed. This is the model I have adopted in this book. Its main features will be described in the next chapter.
Note that even if the proposed model should turn out to be incorrect, the possibility that most of the eukaryotic properties were acquired before the prokaryotic ancestors of present-day organelles were adopted as endosymbionts seems sufficiently plausible, if not likely, to deserve being seriously entertained. Furthermore, whatever model is adopted, many of the mechanisms envisaged could still be relevant.