The Tangled Tree

by David Quammen

  In the 1980s, his work focused more intensely on the application of an ingenious insight: that a variant of Woese’s approach for identification of microbes and placing them on trees of relatedness could perhaps be used for detecting organisms never grown in a lab. Those unknown creatures, whatever they were, had 16S rRNA too, after all. If he could find it, extract it, sequence it, and sort one sample from another, he wouldn’t need to grow the bugs themselves. He wouldn’t need pure cultures. Or, better still, he could extract DNA instead of RNA from the samples, amplify that into workable quantities, and look at the genes coding for different versions of the 16S rRNA molecule. This might illuminate some wild, mysterious little creatures living in extreme but species-rich environments, as well as a lot of others, in places not so extreme but previously overlooked. If it worked, it would open a broad new vista on Earth’s biological diversity, since most of the planet’s life-forms are microbes, and most of those kinds of microbes had never been—possibly never could be—enticed to live and replicate in captivity. They were biological dark matter, vast and weighty, but unseen.

  It did work. Pace and his young collaborators proved the principle with samples from three very different extreme environments. The first was a hydrothermal vent thousands of feet deep beneath the Pacific Ocean. The second was a leaching pond at a copper mine near Hurley, New Mexico. The third was a hot spring in Yellowstone National Park. From the hydrothermal vent, Pace’s team got tissues of a giant tube worm, provided by colleagues at the Woods Hole Oceanographic Institution, who had presumably grabbed their worm samples using Alvin, the deep-ocean research sub. From the leaching pond at the mine, they scooped two pounds of toxic mud. Yellowstone was a different story—slightly more of a lark for the boys in Denver.

  “I was sitting in my office, reading this book,” Norm Pace told me, and he pulled a thick volume off his shelf: Thomas Brock’s treatise on heat-loving microbes. Brock is the man, as I’ve mentioned, who discovered Thermus aquaticus in a Yellowstone hot spring, thereby opening the way to the polymerase chain reaction technique (using a Thermus aquaticus enzyme) for amplifying DNA and all the work in molecular biology that has followed from it. “There is an article in there about Octopus Spring, Yellowstone,” Pace said, “with a whole bunch of these so-called pink filaments in it.” The filaments had been “so-called” by Brock himself, who recognized the pink strands as bacterial agglomerations but never succeeded in culturing them in his lab. They were more untamable than Thermus aquaticus, possibly in part because they required very high temperatures. The upper heat limit for any form of life had been thought, by most microbiologists, to be about 73 degrees Celsius (163 Fahrenheit), until Brock went to Yellowstone in the mid-1960s. He found Thermus aquaticus, for instance, not at the very source of a hot spring but downstream along its outflow channel, at a relatively balmy 69 degrees C. But these pink filaments grew in the hottest part of Octopus Spring, at upward of 92 degrees C. “That’s fucking hot,” Pace reminded me. Water boils at 100. Reading the account in Brock’s book, Pace had found himself wondering about the identity of those heat-loving microbes and how they made their living. This all occurred in 1981, but he remembered it clearly because it played a large part in setting the course of his career.

  He had dashed from his office into the lab, where three of his grad students stood. “Hey, you guys, look at this,” Pace recalls crowing. “This spring in Yellowstone, Octopus Hot Spring—supposedly kilogram quantities of high-temperature biomass!” Kilogram quantities, biomass: that’s what “Eureka!” sounds like from the mouth of a microbiologist. He meant: Whole shitloads of strange bugs that grow happily at near-impossible heat!

  Let’s go get some, he said. He envisioned extracting ribosomal RNA from the pink filaments, sequencing it, and identifying entirely new creatures—creatures not just unknown but unknowable by the classic methods of microbiology. “My God!” he recalls thinking. “We can find out who is out there in the real world.” Another quirk of microbiologyspeak: a bacterium or an archaeon is a “who.”

  The expedition took shape, and Pace invited Woese to join it. So the sage of Urbana rode with the Pace team to Yellowstone, camping amid the geysers, helping collect samples from Octopus Spring, a bemused participant in this “swashbuckling microbiology.” What they found in the hot spring, thriving near the boiling point and characterized solely from their sequencing of ribosomal RNA, was a microbial community dominated by three forms. Two were bacterial and one was an archaeon. All three were new to science. Woese returned home to Urbana happy, but, he told Pace later, it took months for his back to recover from sleeping on the ground.

  There is a grainy photograph, from almost precisely that time, snapped by someone in the lab and preserved in the memorial literature on Carl Woese. It shows a bearded Woese and a lean, young Norm Pace, both in T-shirts, Pace with longish hair and aviator glasses, left hand on his hip, the two toasting each other triumphantly with unidentified libations held aloft in laboratory flasks. The occasion is unspecified. Their return from Yellowstone with a harvest of strange microbial gunk? Maybe. The caption, as I’ve seen this photo reprinted, is simply “Friendship in the Kingdom.” The kingdom referenced is really three kingdoms, or more accurately three domains: Bacteria, Eukarya, Archaea. The snapshot is a reminder that it was a lonely but thrilling time, and that Pace occupied a special position. None of Woese’s long list of contacts—his departmental colleagues, his German admirers, his grad students, his postdocs, his lab assistants, his other coauthors and collaborators—ever did more than Norm Pace to confirm, and to defend for years after, the Woesean view of life’s deepest history.

  Pace was never a member of Carl Woese’s lab. Woese had served on his PhD committee, but Spiegelman was the advisor of record. His relationship with Woese had begun casually, on the corridor, and grown deep over time because of shared interests and convictions. He had his own simple way of describing it, voiced during our conversation in Boulder: “I was his scientific son.”

  72

  Microbes inhabiting people were never directly the focus of Norm Pace’s work. He was more interested in the broader subject of microbial diversity on Earth. But the illumination of the human microbiome has followed in large part from what he and Carl Woese did. Woese pioneered the technique of using 16S rRNA to characterize and compare microbes. The microbes he worked on, such as Ralph Wolfe’s methanogens, were isolated and cultured in the lab. They grew in captivity (not always readily, but Wolfe and Bill Balch solved that problem with the methanogens), and from those pure cultures Woese took his ribosomal RNA. Pace, on the other hand, with his own vision and skills, captured evidence of microbes that live only in the wild. He extended Woese’s technique to raw samples from various environments, including extreme ones such as Octopus Spring. He used 16S rRNA and the genes that code for it, extracted wholesale from those samples, to detect and characterize new creatures that never had been—maybe never could be—grown in a lab.

  Testimony to the importance of those two contributions lies in the literature of the microbiome, such as Ulf Göbel’s paper on gum disease and many others. It lies also in the foundational history of what became a whole new branch of biology: metagenomics, the study of living creatures and communities that can be known only from their genetic material.

  Research on the human microbiome has meanwhile led back to some new insights on horizontal gene transfer. That phenomenon, it turns out, occurs in our own bellies and noses and mouths. Bacteria are always evolving, and to evolve toward greater virulence, or antibiotic resistance, or transmissibility between human hosts, allows them to increase their Darwinian success. Fast evolution is generally better than slow evolution, and HGT represents the fastest way of adding major new possibilities to the heritable variation upon which evolution depends. So the acquisition of entirely new genes, by horizontal transfer from other kinds of bacteria with whom they happen to share habitat in a human colon or nostril, is an effective way for bacteria to improve their p
rospects of survival and abundance. This has been demonstrated in work led by Eric J. Alm at MIT.

  Alm’s group looked at 2,235 complete bacterial genomes, all available from an integrated, open database maintained by a federal institute. In other words, they downloaded this massive amount of sequence data gathered by other researchers the way you might download a season of Stranger Things. More than half of those 2,235 genomes came from “human-associated bacteria”—denizens of the human body, members of the human microbiome. Following Woese’s method, Alm’s team used the 16S rRNA gene as a standard of how closely or distantly those bacteria were related. They knew the geographical origin—at least by continent—of each of the bacteria represented. And they knew the ecological environment from which each came. With the human-associated bacteria, that meant which site within the human body. Stomach, mouth, vagina, armpit, skin? These three parameters, relatedness and geography and ecology, outlined the question they wanted to answer: Which factor is most significant in conducing horizontal gene transfer among bacteria?

  They screened all the genomes for suspicious anomalies—cases where extremely similar versions of a given gene were shared by two very different kinds of bacteria. This was bioinformatics on a vast scale, but vast bioinformatics is now relatively quick and cheap. Their supposition, entirely reasonable, was that each close match of genes shared in two distant lineages signaled a relatively recent horizontal transfer event for that gene. Alm was hoping, in advance, that they might find a handful of such instances. “I was thinking five to ten,” he told me by telephone from his lab. “If we could find five to ten genes, that would make it interesting.” Instead they found 10,770. To find so much transfer, Alm said, was in itself “pretty shocking.”

  Of the three parameters, ecology was clearly important. They found twenty-five times more HGT among bacteria inhabiting the human body, for instance, than among bacteria inhabiting various other environments. But that wasn’t all. Dicing their data at another angle, Alm’s group saw that HGT between bacteria inhabiting the same site on a human body occurs more often than transfer between sites. To be explicit: bacteria that live in our guts tend to trade genes with other gut bugs. Bacteria of the gums, likewise. Of the vagina, likewise. Skin bacteria, likewise. These gene transfers mainly happen at short spatial distances, even when the phylogenetic distance (between two very different bacterial lineages) is great. The ecological circumstance of being resident in someone’s vagina offers a proximity, a shared adaptedness to a very particular environment, and therefore an especially good opportunity, it seems, for two kinds of bacteria to share a gene by HGT.

  So ecology matters more than phylogeny, in this transfer of genes within us. But there was still another message in the data. Alm’s team also found that ecology matters more than geography. The microbes in their group of human-associated genomes came from people on several different continents. Not surprisingly, gene transfer between bacteria on the same continent occurred more often than gene transfer between continents. But the home-or-away difference at that scale wasn’t as great as the home-or-away difference involving body sites. The shared ecology of the human gut, or the vagina, or the nasal passages, or the skin, was most conducive to horizontal transfer. The shared phylogeny of membership in the same bacterial lineage came second. The shared geography of the same continent was a weak third.

  Alm’s group wrote: “Taken together, these analyses indicate that recent HGT frequently crosses continents and the Tree of Life to connect the human microbiome globally in an ecologically structured network.” In simpler language: genes go sideways, from limb to limb, even within our bodies.

  73

  Carl Woese followed the new discoveries by reading the journal reports, though he no longer stood at the forefront of lab work and discovery himself. He was now primarily a thinker.

  In the early 2000s, as Ford Doolittle and other senior scientists published their influential alerts about horizontal gene transfer, as Eric Alm and other young researchers absorbed those alerts with high interest, Woese wrote his “millennial series” of papers on cell evolution and “the universal phylogenetic tree.” He acknowledged in those papers that HGT must have been an extremely important process during the early stages of life on Earth, while making the rearguard argument, based on little offered data, that it had been far less significant since. This allowed him to continue insisting on the universality of his “universal” tree—the one drawn solely from the evidence of 16S rRNA, with three major limbs labeled Bacteria, Eukarya, Archaea. His pronouncement of the Archaea as the third great domain of life, revolutionary in 1977, had become an orthodoxy that he wanted to defend. This is what happens in science. Paradigms shift, but seldom twice in the life and mind of one scientist.

  Woese wasn’t alone (Norm Pace always stood with him, as did some others), but his adamant advocacy of a One True Tree left him increasingly in the minority, and he knew it. Many molecular biologists now seemed to feel, Woese admitted, that HGT over the ages had “erased the deep ancestral trace” of any uniquely valid, tree-shaped phylogeny. He disagreed. “These hasty conclusions are wrong,” he wrote.

  His own work life, meanwhile, changed in two ways. The first involved a new colleague. One day in September 2002 he reached out by email to a theoretical physicist in another corner of the University of Illinois campus. Nigel Goldenfeld was an Englishman, almost thirty years younger, who had arrived in Urbana as an assistant professor, risen to full professor, and spent his middle career studying the dynamics of complex interactive systems. That included topics such as crystal growth, the turbulent flow of fluids, structural transitions in materials, and how snowflakes take shape. The common element was patterns evolving over time. Goldenfeld had never met Woese but knew him by reputation. Later, he called that first ping “the most important email of my life.”

  Woese had contacted the Physics Department chairman and said: “I need a physicist to help. Who should I call?” The chairman had suggested Goldenfeld. In the email, Woese now explained that he wanted to discuss—with someone—the subject of complex dynamic systems. He felt that molecular biology had exhausted its vision, he wrote, and that it needed refocusing around drastic new insights. Those might come from recognizing that the cell itself is a complex dynamic system, and that cell evolution might have occurred through stages understandable only in that way. Properties sometimes emerge, when complex systems interact randomly, that go beyond what exists among unconnected elements. “My telephone is 3-9369,” Woese wrote, “if you care to discuss the matter with me.”

  Goldenfeld replied fast. He was flattered, he was interested, but he didn’t know much about biology, he confessed.

  Woese emailed back: “You may not feel too much at home with biology as it now stands, but if I am any judge, the field is decidedly moving to meet you.” What he had in mind, it turned out, was not just a conversation but a collaboration. Woese wanted a partner who understood complex interactive systems and could quantify their dynamics with brilliant math. Whether that partner knew a bacterium from an archaeon, or Darwin from Dawkins, mattered to him less.

  “So began a scientific partnership and friendship that lasted more than a decade until his death,” Goldenfeld wrote later. “During that time, we met nearly every day and talked on the phone or via email otherwise.” Woese himself had trained as a physicist: math and physics as an undergraduate at Amherst, biophysics for his doctorate at Yale. Not only that, but also Woese prided himself on an academic lineage that went back through his PhD advisor, Ernest Pollard, to the exalted Cavendish Laboratory at Cambridge. Pollard’s own PhD advisor at the Cavendish had been James Chadwick, who discovered the neutron; the lab leader at that time was Ernest Rutherford, who solved other mysteries of radioactivity and the atom; before Rutherford’s tenure, William Thomson (Lord Kelvin) had led the Cavendish, and before Thomson, James Clerk Maxwell. These are some of the giants of modern physics, and in addition to Chadwick, Rutherford, and Thomson, twenty-six other m
embers of the Cavendish Laboratory (including two biologists named Watson and Crick) had won Nobel Prizes. The fact that Nigel Goldenfeld had done his own physics doctorate at the Cavendish may have added to his appeal as a youngish collaborator for the last phase of Woese’s career. That career was now circling back toward biophysics and math, as Woese groped for deeper understanding of early evolution, the RNA-world that preceded his Darwinian Threshold, and the mystery of how so much complexity arose so quickly after the origin of life.

  Woese and Goldenfeld wrote a number of papers during the decade of their collaboration, of which the most important and bold concerned the origin of the genetic code. This paper, coauthored with a young physicist named Kalin Vetsigian, appeared in 2006, bringing forward the subject with which Woese had grappled in his 1967 book. It offered the radical proposition that the universality of the code—all creatures using the same three-letter DNA combinations to call in the same amino acids—reflects a dynamic evolutionary process during the early history of life, not a “frozen accident” traceable to happenstance in one small population of universal ancestors, as Francis Crick had suggested. That proposition was supported in the paper by a computer model and a tonic dose of math. The dynamic process involved “non-Darwinian” mechanisms that made innovation sharing between different lineages possible—and not only possible but also so advantageous that a capacity for such sharing became mandatory. Among those “non-Darwinian” mechanisms, they argued, was horizontal gene transfer, at a level even “more rampant and pervasive” than in later times. Horizontally transferred genes could be valuable to the recipient, by this view, even if the donor organism was still using a slightly different code. And the result of such rampant transferal, among other factors, was consensus on a single mode of coding and translation. Hence the universal code, by which a gene from one lineage of creatures can be optimally applied within another lineage. HGT is not just an ancient and widespread phenomenon, according to Vetsigian, Woese, and Goldenfeld; it was one of the preeminent factors in early evolution, shaping life as we know it and the informational system from which all life is built.