Acquiring Genomes: A Theory of the Origins of Species. Lynn Margulis and Dorion Sagan. Basic Books, New York, 2002. 240 pp., illus. $28.00 (ISBN 0465043917 cloth).
Over 40 years ago DNA was discovered in chloroplasts and mitochondria. This led Hans Ris, a former supervisor of Lynn Margulis, to revive the 1905 theory of the Russian Konstantin Mereschkowsky that chloroplasts originated by enslavement of formerly free-living cyanobacteria by a nonphotosynthetic protozoan host. Mereschkowsky invented the term symbiogenesis for such permanent mergers of phylogenetically disparate organisms into a single chimeric one. His idea of chloroplast symbiogenesis and Ivan Wallin's later advocacy of a similar bacterial origin of mitochondria are accepted by all serious scientists, making mainstream the idea that symbiogenesis occasionally plays a key role in evolution.
When Ris, an excellent electron microscopist, revived the symbiogenetic theory, the close resemblance in membrane organization of chloroplasts and cyanobacteria impressed him as much as that of their DNA. He realized that symbiogenesis was a way of acquiring not only foreign genomes but also foreign membranes. My own contributions to the understanding of symbiogenesis over the past 23 years have emphasized the importance of both genomes and novel membranes, which I call genetic membranes, because, like DNA, they never arise de novo and they do have key genetic roles.
Lynn Margulis seems never to have appreciated or accurately discussed the central roles of membranes in symbiogenesis in her numerous popular writings on the subject. Now a Distinguished Professor at the University of Massachusetts–Amherst, she and her science journalist son Dorion Sagan have written another popular book overemphasizing genome acquisition and ignoring that of membranes. The authors incorrectly claim that only characters encoded by genes can be transmitted transgenerationally, which ignores membrane heredity. Their book purports to be a new theory of the origin of species. The authors rightly stress that the biological species concept does not apply to bacteria, but they wrongly call the concept new—this was recognized explicitly all along by its originator, Ernst Mayr, who wrote a tactful foreword to Acquiring Genomes, gently indicating that their central thesis about eukaryote speciation is overstated. I consider it totally mistaken. Mayr notes that average readers will learn much they did not know before about the fascination of microbes, a fascination I share with the authors. Sadly, however, they will also be misled about the role of symbiogenesis in evolution and what is now understood about the mechanisms and history of cell evolution.
The authors assert that all eukaryotes (nucleated organisms, e.g., protozoa, plants, animals, fungi) form new species only by symbiogenesis and that random mutation is relatively unimportant. They even present a new definition of species that necessarily would make their claim true. They say that species are different only if they are chimeras of a different set of symbiogenetic partners. As I interpret symbiogenesis and the established picture of cell evolution, this would mean that all animals belong to the same species—the same species as all fungi and most nonphotosynthetic protozoa—because their cells were formed by the symbiogenetic merger of the same protoeukaryote host and α-proteobacterium to form the first true eukaryote, no symbiogenesis having occurred in them subsequently. Plants belong to a different species because they also enslaved a cyanobacterium, but are all just one species, according to the authors' curious definition.
Apart from the origins of mitochondria and chloroplasts, symbiogenesis has occurred on only about four other occasions in the history of life, none even mentioned by the authors. The two most important are (1) the enslavement of a red alga to make chromalveolates (e.g., brown seaweeds, diatoms, dinoflagellates, a host of other chromophyte algae, malaria parasites and their sporozoan relatives, and pseudofungi); and (2) the enslavement of a green alga to form euglenoid and chlorarachnean algae (Cavalier-Smith 2003). The other, more trivial symbiogenetic events were replacements by a few dinoflagellates of their own chloroplasts by foreign ones. The authors devote only about two sentences to the symbiogenetic origins of mitochondria and chloroplasts, all the book has to offer on the well-established cases of symbiogenesis.
What constitutes the rest of the book? Four things: repeated diatribes against evolutionary biologists and the established idea that mutation is fundamentally important in evolution; readable descriptions for the layman of many fascinating cases of symbiosis involving microbes; one-sided summaries of Margulis's current idiosyncratic view of cell evolution, which ignores most phylogenetic evidence and anyone else's criticisms or sounder interpretations; and peremptory attacks on many standard biological concepts (e.g., genes, competition, mutualism). The authors imply that 10 million to 30 million species of eukaryotes evolved by symbiogenesis, yet evidence exists for only about six symbiogenetic events in the whole history of life.
They do this by ignoring the careful distinction I made in 1985 (Cavalier-Smith and Lee 1985) between an obligate intracellular symbiont and a true organelle of symbiogenetic origin. I pointed out that integration of foreign genomes into the host nucleus can occur only after the host evolves novel, generalized protein-targeting machinery that can place many products of transferred genes back into the former symbiont. I used the presence or absence of such machinery— e.g., the protein-import machinery of mitochondria and chloroplasts—to establish a clear-cut boundary between a symbiont that lacks it and a symbiogenetic organelle that has it. This provides objective demarcation between symbiosis, which is very common, and symbiogenesis, which is exceedingly rare. Nobody disputes that intracellular symbionts are widespread in eukaryotes and often of great physiological and evolutionary importance, as are other types of symbiosis. But acquiring a symbiont is not symbiogenesis. It is simply symbiosis. The authors never so much as mention the central role of novel protein-targeting in symbiogenesis.
They vaguely define symbiogenesis as “symbiosis that leads to evolutionary change.” Probably all symbiosis leads to evolutionary change. The authors' failure to distinguish symbiogenesis from symbiosis ignores the fact that Mereschkowsky's symbiogenesis meant permanent merger of two organisms into one. Lichens comprise two separate organisms; the fungus temporarily enslaves the alga/cyanobacterium—neither cells nor genomes are merged. The same is true of all the other examples Margulis and Sagan cite, despite their frequent, unsubstantiated claims to the contrary. Consider two examples of tendentious misrepresentation: the sea slugs that harbor chloroplasts temporarily for photosynthesis or cnidarian nematocysts for defense. The former lack even the nuclear genomes of the algae from which they came and have to be replaced periodically. Nematocysts are organelles without genomes, not cells (contrary to the authors' assertion); they also lack the nucleus and the rest of the cell from which they were stolen (Greenwood and Mariscal 1984). Yet the authors say the slugs “flaunt their stolen genomes” even though the slugs have no stolen genomes to flaunt! Algal or cnidarian genomes are not integrated into the slug.
The authors assert that symbionts “often fuse their genomes” and “many such fusions have been documented in all five kingdoms of life” (p. 90). This is false. The latter claim is peculiar, as the authors spend much space arguing that bacteria never undergo cellular symbiogenesis (there is one putative case, not mentioned). At present, there is clear evidence for fusion of cellular genomes for only one of their five kingdoms, the so-called Protoctista, which no serious biologists accept as a sensible group, because it is undoubtedly polyphyletic. In the next paragraph, the authors imply, without actually asserting it, that corals, giant clams, tubeworms, termites, and cows are all examples of such genome fusion. Again, these are simply symbioses. By juxtaposing wild claims with irrelevant examples that fail to support them but nonetheless lose the reader in fascinating symbiotic detail—chattily written in the fashionable mode of pop-science journalism—the authors may make the nonexpert think there is something in their central thesis, despite its lack of empirical evidence or rationally argued theory. They repeatedly confuse a symbiotic consortium of different species of organisms with a single organism by misusing the term “individual” for it.
The authors' viewpoint is illogical and superficial. In one place they argue that bacteria have no species (reasonable), in another that they constitute only one species (unreasonable). That would make cyanobacteria and proteobacteria the same species and talking about the separate symbiogenetic origin of mitochondria and chloroplasts problematic. If there were only one bacterial species, how could you make 10 million to 30 million different eukaryote species merely by mixing and matching that one species in the absence of mutation? Even if we were to equate symbiogenesis with the acquisition of a novel symbiont, as the authors often seem to do, there would be immensely fewer such acquisitions than recognized morphological species. Except for the six established cases of symbiogenesis, all differences between eukaryotic species or bacterial strains have arisen by mutation, plus occasional lateral transfer of individual genes or small gene clusters, not by the symbiogenetic merger of genomes. Mutation is the greatest innovator by far. Even lateral gene transfer is less innovative and less frequent than widely supposed. Unsurprisingly, the authors uncritically repeat early claims that the human genome has laterally transferred genes of bacterial origin, though that idea has been refuted by W. Ford Doolittle and others (Andersson et al. 2001).
The chapter on eukaryogenesis adopts as true, without acknowledgment, my 1983 theory (Cavalier-Smith 1983a, 1983b) that the amitochondrial Archamoebae were the first eukaryotes, which I gave up over five years ago because it is thoroughly refuted by phylogenetic evidence that the authors ignore (Roger 1999). Their so-called phylum Archaeprotista is polyphyletic: Archamoebae and metamonad flagellates independently lost mitochondria and are no closer phylogenetically than are animals and plants. They ignore compelling evidence that the first eukaryote was aerobic and had mitochondria. Instead, they espouse the unwarranted theory of a second symbiogenesis prior to the origin of mitochondria and link it to Margulis's long-standing, but long-refuted, idea of the symbiogenetic origin of cilia from spirochaetes.
Some minor irritations: Charles Darwin did not “consistently fail to credit” his grandfather; it is untrue that ctenophores have “stinging cells”; the “40-volume” work on 50 phyla of animals, published in 1940 by “Libby” [properly Libbie] Hyman actually had only six volumes (the first recognised only 20 nonprotozoan animal phyla, and other volumes were published decades later); and statements about Cnidaria or Coelenterata are garbled.
Acquiring Genomes is an impressively undiscriminating collage of interesting fact, occasional error, sloppy reasoning, didactic pop biology, and sensible and stupid ideas (among which are the suggestions that nematocysts evolved from microsporidia and that animal larvae evolve by hybridization between phyla). Overall, this book is too confused to recommend to other biologists or even to general readers, for whom it seems intended.