The Tangled Tree

by David Quammen

  The fact that resistance genes could move sideways so easily was a clue that genes for other traits might be moving too. And so, Anderson wrote, “the temptation is very strong to suggest that the transfer factors may have influenced bacterial evolution.” Maybe it wasn’t all just a matter of mutation and natural selection after all. Maybe horizontal gene transfer also played a big role in the long history of microbial life.

  This was a gentle statement of a revolutionary prospect. Was the history of evolution—at least among one group, the bacteria—really that different from Darwin’s theory as we have come to embrace it?

  Anderson’s suggestion was echoed in 1970 by a pair of British researchers in a different area of bacteriology. Dorothy Jones and Peter Sneath were microbial systematists—namers and classifiers—at the University of Leicester. They worked in the same long tradition as Ferdinand Cohn, the great early classifier of bacteria, but with a concerted effort to make use of new data, modern methods, and fresh thinking. Their preferred method was known as numerical taxonomy, which at that time stood adamantly opposed by another school of classification, newer, known as cladistics. The numerical taxonomists classified creatures into species and higher categories by overall similarity, regardless of evolutionary history. The cladists argued that common ancestry—and therefore evolutionary history—is the only cogent basis for classifying. This was a bitter and arcane fight about which, trust me, you don’t need the details. Suffice to say here that in 1970 Jones and Sneath coauthored an influential review paper titled “Genetic Transfer and Bacterial Taxonomy,” the main purpose of which was to use horizontal gene transfer as a cudgel for bashing cladistics on its head.

  Cladistics works poorly for classifying bacteria, they noted, because of the near-total absence of a fossil record. But even more damning for that evolution-based approach was the evidence emerging from studies in Japan and elsewhere of gene transfer between one bacterial species and another. Jones and Sneath proceeded, through the bulk of their long paper, to describe and document much of what was then known about horizontal gene transfer among bacteria. Then they speculated that one “foreign” gene, transferred into a bacterium and integrated there, might make that genome more able to accept other transfers. Barriers between species might begin to fall. “This in turn could favor extremely reticulate modes of evolution, with numerous partial fusions of phyletic lines.” Reticulate modes? That meant weblike. Fusions? That meant genes jumping sideways, from one genome to another. Tree branches never fuse, never reticulate, so how do you draw this situation as a tree of relatedness?

  “It may well be that gene exchange is so frequent,” they wrote, “that the evolutionary pattern in bacteria is much more reticulate than is commonly believed.” Weblike, not treelike. What they were saying, implicitly, was: Whew, all this gene transfer makes classifying bacteria tricky—and harder for us, but impossible for those poor obdurate cladists.

  Jones and Sneath and their allies were destined to lose this battle. Cladistics would triumph and become the reigning approach to classification, at least among evolutionary biologists. But the paper by Jones and Sneath served other purposes. It broadened awareness of how horizontal gene transfer does complicate the enterprise of classifying organisms and portraying their evolutionary history. And it seems to have been one of the first scientific sources, if not the first, to float the concept of reticulate evolution—the idea that the limbs of the tree of life are intertangled.

  56

  At this point, we come to the frangibility of another absolute. By which I mean, the ricketiness of another apparent absolute. The concept of “species” is commonly supposed to be secure. It isn’t secure. It’s especially insecure in the realm of bacteria and archaea, but it’s even a bit blurry when scientists try to distinguish one species of plant or one species of animal from another. The boundaries blur. The edges are as porous as Goretex—or, in some cases, as cheese cloth. One reason for the blur, one symptom of the porosity, is horizontal gene transfer—genes moving sideways instead of just downward from parent to offspring. If genes cross the boundary between one species of bacteria and another, then in what sense is it really a boundary?

  The conviction that bacterial species are fixed and discrete goes back, as you’ve read, to Ferdinand Cohn. Working in Breslau during the 1860s and 1870s, trying to put order into bacterial classification, contending against the notion of bacteria as shape-shifting creatures that change their forms according to conditions of environment, Cohn found ways to grow pure cultures of one bacterial strain or another on solid media. A pure culture had a continuity of identity and form that suggested giving it a species name—Bacillus anthracis, for instance—was a rational exercise. And the confidence of identity was very useful when you wanted to tell Bacillus anthracis (causing anthrax in humans) from Bacillus subtilis (causing rot in potatoes). Robert Koch helped this effort by developing the technique of streaking bacteria onto solidified gelatin surfaces, from which one tiny smidge of cells from a mixed sample could be teased out and grown again separately, yielding a pure culture. Koch’s lab assistant Julius Petri helped further by inventing the Petri dish, in which a pure strain grown on gelatin or agar could be protected from airborne contamination with a glass lid. Ferdinand Cohn won the battle of bacterial classification, and his victory held good for about fifty years.

  But this certitude of discrete identities began to weaken with Griffith’s discovery in 1928 that one type of Pneumococcus pneumoniae could transform into another type. It weakened still further when Watanabe announced that Shigella dysenteriae could receive genes from Escherichia coli. Nowadays bacterial taxonomists recognize that Shigella genomes are very similar to Escherichia genomes—so similar that they should probably be lumped in a single genus. In fact, some strains of Shigella are more closely related to E. coli than they are to one another. And the crumbling certitude about how to sort bacteria into species, despite horizontal gene transfer, has progressed far beyond that little confusion.

  Sorin Sonea, a Romanian-born microbiologist at the University of Montreal, took the conundrum of blurring bacterial boundaries to its logical extreme. First in a French edition, then in English, in 1983, he and his coauthor, Maurice Panisset, published a book titled A New Bacteriology, making the case that all bacteria on Earth constitute a single interconnected entity, a single species—no, wait, maybe even a single individual creature—through which genes from all the variously named “species” flow relatively freely, by horizontal gene transfer, for use where needed. This freedom of transfer, this universal interchangeability of parts, gives the bacterial entity “a huge available gene pool,” Sonea and Panisset wrote, and thereby allows bacteria to adapt so well, and so quickly, to so many different environments and situations. The genome within an individual bacterium is typically small, much smaller than genomes of most eukaryotes, and contains relatively few genes—only the bare necessities for bacterial life and replication. There’s little excess, redundancy, or emergency provisioning of intermittently useful genes for special circumstances. The advantage in such parsimony is that it allows bacteria to reproduce quickly. The disadvantage is a lack of versatility for special circumstances—but horizontal gene transfer, bringing in new genes from other strains or species as needed, compensates nicely for the lack, complementing the bare-bones endowment. As a result, bacteria get by with few genes, some of which (especially those on plasmids, unattached to the bacterial chromosome) are continually being lost or gained.

  This is radically different, Sonea and Panisset claimed, from evolution as described by Darwin, who focused on animals and plants. Animal and plant species, as well as other eukaryotic species, arise mainly by genetic isolation. Bacteria are never so isolated. Instead of tortoises and mockingbirds marooned on islands, mutating and adapting, diverging slowly into distinct subspecies and eventually new species, finally reaching a point where they can’t or won’t mate with other populations—instead of that, you have relentless bacterial togethe
rness. You have genes oozing sideways all over the planet, from one bacterial strain to another, like pulsing juices inside a gigantic, invisible version of the Blob.

  Okay, they didn’t call it the Blob. They called it a “superorganism,” which is almost as spooky. Also, to be clear: this superorganism concept of Sonea and Panisset was quite distinct from the superorganism concept that James Lovelock and Lynn Margulis framed, under the name Gaia, and applied to planet Earth itself. Earthly Gaia was a superorganism comprising all physical and living constituents of the planet, according to Lovelock and Margulis. Sonea and Panisset’s superorganism was “just” the total global population of bacteria. The two ideas are related in spirit—they’re grandiose and daring and woozy—but very different in their particulars and their purposes. Sonea and Panisset intended to depict the fluidity with which bacterial “species” exchange genes; but their superorganism did not encompass all other forms of life. It was not the Earth Mother. It was the world’s biggest germ. What made it vivid, and more useful than the notion of Gaia, was that it starkly contrasted two things rather than uniting everything: the way bacterial genes move sideways versus the way tortoise genes and mockingbird genes generally don’t.

  After the death of Panisset, Sorin Sonea continued to argue their big idea, the bacterial superorganism, in English publications, with mixed response from the scientific community. Lynn Margulis liked it, not surprisingly. Ford Doolittle called it “bold if inchoate”—a fair judgment—and he recalled that Sonea’s theory and others like it “were widely dismissed during the 1970s and 1980s—they were so hopelessly radical!” Doolittle enjoys offering a bit of radical provocation himself, so this was a nostalgic and friendly comment when he wrote it, in 2004, by which time horizontal gene transfer was a hot topic throughout molecular biology and had destroyed all the old notions of sorting bacteria into species with neat boundaries and placing them like fruit on a tree.

  57

  Other scientists besides Sonea and Panisset began noticing, back in the 1980s, that this odd phenomenon might have broad implications. Slowly at first, it became a favored research theme in more than a few labs. The phrase “horizontal gene transfer” had just been coined (“lateral gene transfer” became a variant, meaning the same) and “reticulate evolution” was also in the air. Journal papers and review articles appeared, still based mostly on tenuous data, raising questions in the same spirit as Ephraim Anderson had, about the significance of HGT and whether it demanded a new theory of evolution—a major supplement to Darwin’s.

  It did seem to be widespread and common among bacteria. Some researchers even saw, or thought they saw, evidence of it in other creatures. Eukaryotes. Animals, plants. A species of fish, which carried a bacterial symbiont, appeared to have passed one of its fish genes into the bacterial genome. How was that possible? Another bacterium had sent bits of its DNA into the nuclear genomes of infected plants. Bacterium to plant? A species of sea urchin seemed to have shared one of its genes with a very different species of sea urchin, from which its lineage had diverged sixty-five million years earlier. That was a stretch. Still another bacterium, the familiar E. coli, was found to be transferring DNA on plasmids into brewer’s yeast, which is a fungus. Brewer’s yeast is microbial, a relatively simple little creature, but nonetheless eukaryotic. This mixing of fungal host and bacterial genes happened via a smooching process that looked much like bacterial conjugation, the researchers reported, and “could be evolutionarily significant in promoting trans-kingdom genetic exchange.” Trans-kingdom is a long way for a gene to go.

  A 1982 essay in Science offered an overview, titled “Can Genes Jump Between Eukaryotic Species?,” with the implicit answer: probably. Some of these cases of distant transfer later turned out to be illusory—disproved when better data became available—but the basic premise was correct, and the research agenda took hold. Genes were moving sideways, across boundaries, between very different kinds of creatures, to a degree previously unimagined. And the new recognition did raise a challenge to Darwin and Darwin’s tree.

  The notion that genes might be transferred sideways among complex eukaryotic organisms, as the Science essay noted, was a radical step beyond the established reality of horizontal gene transfer in bacteria. It was an “apparently fanciful and certainly unorthodox” idea, a perplexing anomaly that violated some axiomatic principles, and its investigation would need to progress through two stages. First, does this weird thing actually happen? Second, if it does, how common and how important is it?

  New investigations, as time passed, showed that the weird thing does happen. For instance: there’s a peculiar group of tiny animals known as rotifers, once studied only by invertebrate zoologists, including Leeuwenhoek, but now notable throughout molecular biology for their “massive” uploads of alien genes.

  Rotifers are homely beyond imagining—so homely you almost (but not quite) have to love them. They live in water, mainly freshwater, and in moist environments such as soils and mosses. They live in rain gutters and sewage treatment tanks. Look at one through a microscope, and you’ll see what resembles a maggot, with a lamprey mouth and a long, thin tail, but the tail isn’t really a tail. Rotifer biologists call it a foot. Some kinds of rotifer can suck their foot back into their body when it’s not in use. An extrudable and retractable foot. At the end of the foot is a toe, or maybe two toes, or four toes, depending on the species. Among rotifers that attach to surfaces or inch around, the toes have cement glands for getting a grip. Very handy; if you had only one foot, one shoe, you’d want a Vibram sole or cleats. But some of them float free as plankton. The lamprey mouth is circled by cilia, little hairs, that move quickly and set up a swirling flow to bring bits of food down the gullet. That circular swirling is what gives them their name, rotifer, derived from Latin syllables meaning “wheel bearer.” They eat detritus, bacteria, algae, and other minuscule forms of digestible mulch. Some fish hobbyists put them in aquariums to help clean the glass. Captive rotifers will reproduce, if your tank is hospitable, and, as a side benefit, your tetras and swordtails can eat them. A bottle of pet rotifers, starter rotifers, can be had for $17 from a company in Nashville.

  Invisible teeming maggots. A big rotifer might be a millimeter long, barely big enough to see, but small as they are, these are not single-celled creatures. They’re multicellular animals.

  The rotifers of one particular group are especially peculiar and interesting. They’re called the bdelloids, a name I can type easily but not pronounce. Bdelloid rotifers tend to live in harsh, changeable environments that sometimes go dry. The bdelloids cope with such crises by hunkering into a dehydrated, dormant state, like instant coffee, in which they can survive for as long as nine years. When the water returns, they rehydrate and come alive. Another oddity of the bdelloids is that they reproduce without sex. Females give birth to females, no fertilization necessary. The fancy term for that is parthenogenesis. A male bdelloid has never been seen. Genetic evidence suggests that bdelloids have gone without sex for twenty-five million years—quite a period of celibacy by anyone’s measure. Despite the absence of sexual recombination, which shuffles the genetic deck in a population and offers up new combinations of genes, bdelloids have managed to find newness somehow. They have diversified into more than 450 species.

  That diversification without sex might be partly explained by the other bdelloid anomaly deduced recently from genetic evidence: their strong propensity for horizontal gene transfer. This was noted in 2008 by three Harvard researchers who sequenced sections of genome in one bdelloid species and found all sorts of craziness that shouldn’t have been there. More specifically, they found at least twenty-two genes from non-bdelloid creatures, genes that must have arrived by horizontal transfer. Some of those were bacterial genes, some were fungal. One gene had come from a plant. At least a few of those genes were still functional, producing enzymes or other products useful to the animal. Later work on the same rotifer suggested that 8 percent of its genes had been ac
quired by horizontal transfer from bacteria or other dissimilar creatures. A team of researchers based mostly in England looked at four other species of bdelloids and also found “many hundreds” of foreign genes. Some of the imports had been ensconced in bdelloid genomes for a long time, since before the group diversified, while some were unique to each individual species, and therefore more recently acquired. This implied that horizontal gene transfer is an ancient phenomenon among bdelloid rotifers, and that it’s still occurring. Genes going sideways among animals? That was definitely supposed to be impossible. It wasn’t.