Microcosm

by Carl Zimmer

  Viruses are quickly losing their reputation as insignificant parasites. They are the most abundant form of life on Earth, with a population now estimated at 1030—a billion billion trillion. Most of the diversity of life’s genetic information may reside in their genomes. Within the human gut alone there are about a thousand species of viruses. As viruses pick up host genes and insert them in other hosts, they create an evolutionary matrix through which DNA can shuttle from species to species. According to one estimate, viruses in the ocean transfer genes to new hosts 2 quadrillion times every second.

  It’s a bizarre coincidence that just as scientists were discovering the evolutionary importance of viruses, computer engineers were creating a good metaphor for their effect. In the late 1990s, a group of American engineers became frustrated by the slow pace of software development. Corporations would develop new programs but make it impossible for anyone on the outside to look at the code. Improvements could come only from within—and they came slowly, if at all. In 1998, these breakaway engineers issued a manifesto for a different way of developing programs, which they called open-source software. They began to write programs with fully accessible code. Other programmers could tinker with the program, or merge parts of different programs to create new ones. The open-source software movement predicted that this uncontrolled code swapping would make better programs faster. Studies have also shown that software can be debugged faster if it is open source than if it is private. Open-source software has now gone from manifesto to reality. Even big corporations such as Microsoft are beginning to open up some of their programs to the world’s inspection.

  In 2005, Anne O. Summers, a microbiologist at the University of Georgia, and her colleagues coined a new term for evolution driven by horizontal gene transfer: open-source evolution. Vertical gene transfer and natural selection act like an in-house team of software developers, hiding the details of their innovations from the community. Horizontal gene transfer allows E. coli to grab chunks of software and test them in its own operating system. In some cases, the combination is a disaster. Its software crashes, and it dies. But in other cases, the fine-tuning of natural selection allows the combination to work well. The improved patch may later end up in the genome of another organism, where it can be improved even more. If E. coli is any guide, the open-source movement has a bright future.

  ASSEMBLING ASSASSINS

  Among its many accomplishments, open-source evolution has produced a lot of ways for us to get sick. When Kiyoshi Shiga discovered Shigella, he believed it was a distinct species, and so did generations of scientists who followed him. But when scientists began to examine the genes of Shigella in the 1990s, they realized it was just a particularly vicious form of E. coli. More detailed comparisons revealed that Shigella is actually many separate strains. Many of them are more closely related to harmless strains of E. coli than they are to other strains of Shigella. In other words, Shigella is not a species. It is not even a single strain. It is more a state of being, one that has been achieved by several lineages of E. coli.

  Shigella strains typically evolved from less sophisticated parasites. Their ancestors sat on top of the cells of the intestinal wall, injecting molecules into host cells to make them pump out fluids. (Many strains of E. coli still make this sort of living today.) Shigella’s ancestors acquired new genes that allowed them to invade and move inside cells, to escape the immune system and manipulate it. These innovations did not happen in a single lineage of E. coli. They evolved many times over.

  Just as important as the genes Shigella gained were the ones it lost. Flagella are wonderful for swimming in the gut, but they are useless inside the crammed interior of a host cell. No Shigella strain can make flagella, although they all still carry disabled copies of the flagella-building genes. Shigella also has disabled copies of genes for eating lactose and other sugars that it no longer feeds on. And it has abandoned an enzyme called cadaverine, which other strains of E. coli make to protect themselves from acid. (Other bacteria produce this foul-smelling substance as they feed on cadavers; hence the name.) For Shigella, cadaverine is a burden because it slows down the migration of immune cells across the wall of the intestines. Shigella depends on immune cells to open up passageways that it can use to get into the intestinal tissue and invade cells. As a result, one of the genes essential for making cadaverine has been disabled in every strain of Shigella.

  Other strains of E. coli have evolved into different sorts of pathogens, and their genomes still record that transformation. Horizontal gene transfer, lost genes, and natural selection all were at play in their histories as well. Scientists who study E. coli O157:H7, the strain that can be carried in spinach or hamburgers, have done a particularly good job of reconstructing its evolution step by step. Its ancestors started out as far gentler pathogens, but about 55,000 years ago, they began to be infected with a series of viruses, each installing a new weapon in its arsenal. The devastating toxin that makes E. coli O157:H7 so dangerous, for example, is encoded on a gene that lies nestled among the genes of a virus. The virus is such a recent arrival in the E. coli O157:H7 genome that it still makes new viruses that can escape the microbe.

  Scientists who study E. coli O157:H7 face a strange paradox, however. Other disease-causing strains of E. coli, such as Shigella, are highly adapted to living in humans and are rarely found in other species. But E. coli O157:H7 is just the opposite. It rarely turns up in humans (for which we can be grateful), but it lives in many cows and other farm animals. In us it can be deadly, but in them it causes no harm at all. It has adapted to them, in other words, as a benign passenger. The fact that O157:H7’s toxins make us deathly ill is just an evolutionary accident, because we are not their normal host.

  If E. coli O157:H7 doesn’t make toxins to exploit us, then, why do they carry the toxin genes around? Some researchers have suggested that the bacteria make the toxin to help their animal hosts. At the University of Idaho, scientists have found that sheep infected with E. coli O157:H7 do a better job of withstanding a cancer-causing virus than sheep without that strain. They speculate that E. coli O157:H7’s toxins stimulate the ovine immune system, or perhaps even trigger cells infected with the cancer-causing virus to commit suicide before they can form tumors. But it’s also possible that the toxins are a defense for the bacteria themselves. When protozoans attack E. coli colonies, the ones that make the toxin can fend off the predators.

  While E. coli O157:H7 may not have evolved to adapt to our bodies, we have still played a part in its rise. Studies on its genome show that it is a very young lineage; all of its most common forms are less than a thousand years old. Scientists suspect that by domesticating animals, humans created the conditions in which E. coli O157:H7 could thrive. Its hosts now spent much of their time penned together on farms, where the E. coli O157:H7 that they shed with their manure had a much better chance of infecting a new host than if the cows were off in the wild. E. coli O157:H7 exploded with the growth of cattle herding in recent centuries, first with the arrival of cows in the New World and more recently as cows have been packed together into feedlots. The bacteria haven’t just become more common thanks to us; they may also have been evolving faster, because viruses have been able to move from microbe to microbe, producing new strains of E. coli O157:H7.

  While some harmless strains of E. coli have evolved into deadly parasites, evolution has flowed the other way as well. Some of the most benign strains of E. coli descend from pathogens. One strain of E. coli, known as A0 34/86, shields its hosts from invasions of diarrhea-causing bacteria. Doctors sometimes administer it to premature babies to protect their underdeveloped intestines from attack. In 2005, scientists published the genome of A0 34/86. They found genes for cell-killing factors, bloodletting proteins, and other weapons used by O157:H7 and other lethal strains. A0 34/86 uses these dark powers for our good, by aggressively establishing colonies in the guts of babies, thereby preventing disease-causing strains from finding a place to settle. We may try
to draw sharp lines through nature’s diversity, to split E. coli up between its killers and its protectors. But evolution does not deal in sharp lines. It blurs.

  ONE LIFE, MANY MASTERS

  Another blurred line is the one that divides E. coli from its viruses. It may seem sharp if you are looking at E. coli ripped open by viruses streaming out to infect a new host. These seem like two different beasts. But E. coli has many different kinds of relationships with its viruses. Prophages can, at least for a time, seamlessly blend themselves into their host genomes. They do not necessarily surrender their sense of identity, though. They can sense when their host begins to suffer and at that point they turn back into familiar, host-killing viruses again. And then there are the viruses that lug around bundles of genes that can be very helpful to a microbe but offer no immediate benefit to themselves. When they slip into the genome of E. coli, it becomes much harder to say where the virus leaves off and the host begins. The viruses may then become trapped for good inside E. coli’s genome, thanks to mutations that destroy their ability to make new copies. Over time, new mutations may chop out much of the virus’s original DNA, leaving behind only those genes that are useful to the host. They are viral genes in name and origin only.

  To make sense of this confusing relationship between E. coli and its viruses, it helps to set aside the “us versus them” view of life and to think of life as a braiding stream of genes. The genes carried by a virus at any moment form a coalition of evolutionary partners that have more success working together than any one of them could have on its own. Some coalitions thrive simply by invading a host and using it to replicate more copies of themselves. But in other cases, the interests of the virus’s genes and E. coli align. They may enjoy more reproductive success if they spare their host rather than kill it. Some viruses end up as itinerant Samaritans, bringing with them many genes that benefit their host—and, by extension, themselves. They are constantly trying out new combinations of genes as they travel, and the combinations that bring the most success to their hosts are the ones that survive.

  These relationships can get complicated, as most relationships do. A virus can be simultaneously benign and malignant toward its E. coli host. E. coli O157:H7, for example, carries the genes of a virus that include the toxin gene. It’s possible that the bacteria benefit from making the toxin, perhaps because it keeps predators at bay; but for the individual microbes that actually produce it, the experience is not so pleasant. The virus forces the microbe to make both toxin molecules and new copies of itself until it bursts.

  The decision to make the toxin lies with the virus, not with E. coli. It produces the toxin in times of stress—which is one reason why doctors generally don’t prescribe antibiotics for an infection of E. coli O157:H7. The drugs trigger the viruses to escape their hosts, turning what might have been an unpleasant bout of bloody diarrhea into a potentially lethal case of organ failure. The virus’s habit of killing its E. coli O157:H7 host is almost enough to inspire pity for the microbe. It is as much a victim of the virus as we are. Even after the viruses have killed their original host, they continue to make things worse for E. coli. They infect the harmless strains of E. coli in our gut, transforming them into factories that turn out more viruses and more toxins. These hapless bystanders boost E. coli O157:H7’s production of toxins a thousandfold.

  Other viruses use a different strategy to survive, but one that’s no less cruel to E. coli. Rather than destroying their host when times are bad, they hold it hostage. One of these viruses, known as P1, carries a gene that makes a protein called a restriction enzyme. Restriction enzymes are able to grab DNA at specific sites and slice it apart. Yet P1 normally does not kill E. coli. That’s because the virus also makes a second protein that protects the microbe from the restriction enzymes. Known as a modification enzyme, it builds shields around E. coli’s DNA at exactly the sites where the restriction enzyme can grab it.

  Why should P1 bother building both a poison and its antidote? Like many viruses, P1 lives on a plasmid. Each time E. coli divides, it usually makes new copies of both its own DNA and the P1 plasmid. Sometimes E. coli makes a mistake, however, and all the plasmids end up in one offspring with none in the other. Those plasmid-free bacteria might be able to outcompete the ones that still carry the P1 virus, because they don’t have to use extra energy to make virus proteins and copy their DNA. So the P1 virus kills them—even though it’s not actually in the bacteria. The deadly beauty of restriction and modification enzymes is that restriction enzymes are durable, whereas modification enzymes are short-lived. If E. coli loses the P1 virus, it quickly loses its shields and cannot make new ones. Eventually its DNA becomes vulnerable, and the restriction enzymes move in for the kill. Once E. coli is infected with P1, in other words, it can’t live without the virus.

  Genes for restriction and modification enzymes aren’t unique to P1. E. coli carries many of them on its chromosome. Ichizo Kobayashi, a geneticist at the University of Tokyo, has argued that they also got their start as selfish genes holding their host hostage. He points out, too, that restriction and modification enzymes could have allowed viruses to battle other viruses trying to take over their host. A new virus invading E. coli does not have the shields made by the resident virus, leaving it open to attack by restriction enzymes. While restriction and modification enzymes may have gotten their start as ways to let a parasite thrive, some of them appear to have been harnessed by their E. coli hosts. By killing incoming viruses, they have become a primitive sort of immune system for the bacteria.

  Genes come into similar conflict in all species. Many insects are infected with a microbe called Wolbachia, for example, that can only live inside their cells. It relies for survival almost entirely on being passed down from one generation to the next. This strategy has one major shortcoming: Wolbachia cannot infect sperm, and so males are a dead end for its posterity. In other words, the success of Wolbachia’s genes and those of its male hosts are in conflict.

  Wolbachia has evolved many ways to win this struggle. In some species of wasps, for example, Wolbachia manipulates infected females so that they give birth only to females, and it alters their offspring so that they have no need to mate with males to reproduce. In other species, Wolbachia kills an infected mother’s male eggs. The bacteria in the male eggs die as well, but the strategy ensures the overall success of Wolbachia genes: the Wolbachia-infected female eggs survive, and when they hatch the female larvae don’t face competition for food from their brothers. In fact, their brothers become their food. Wolbachia, in other words, has hit on some of the same strategies that viruses use to thrive in E. coli.

  These murky struggles between parasite and host, these blurrings of species, may seem profoundly alien. Yet we are not above the shaping forces of viruses. Most viruses simply invade our cells, which produce new viruses that move on to the next host. But some viruses insert their genetic material in a cell’s genome. If they manage to infect a sperm or an egg, these viruses will be passed down from one generation of humans to the next. Over many generations, mutations cause the viruses to lose their ability to escape their host cells. Many lose most of their genes. What remains are instructions for making copies of their DNA and pasting that DNA back on their host’s genome. These genomic parasites now make up about 8 percent of the human genome. Recent research suggests that some of them have been harnessed by their hosts. A number of essential human genes, which help build things as different as antibodies and placentas, evolved from virus genes. Without our resident viruses we would not be able to survive. Once again, what is true for E. coli is true for the elephant: Where do our own viruses stop, and where do we begin?

  Nine

  PALIMPSEST

  BURIED MESSAGES

  WHEN SCIENTISTS PUBLISHED THE FIRST genome of E. coli in 1997, they titled it “The Complete Genome Sequence of E. coli K-12.” Strictly speaking, the title was a piece of false advertising. Nowhere in the paper can you find the raw sequence of 4,639,22
1 bases. The omission was simply a matter of space: E. coli K-12’s genome would fill about a thousand journal pages. Those who crave a direct confrontation with its genetic code must visit the Internet.

  One of the sites that houses its code is the Encyclopedia of Escherichia coli K-12 Genes and Metabolism, EcoCyc for short. EcoCyc displays the K-12 genome as a horizontal line stretching across the screen, scored with a hash mark every 50,000 bases. If you click on the mark labeled “1,000,000,” you will zoom in on the 20,000 bases that straddle that point in the genome. Bars run above the line to show the location of individual genes. Click on the bar for the gene pyrD and you can read its sequence. If you seek something more meaningful, you can also read about pyrD’s function (creating some of the building blocks of RNA). On EcoCyc you can learn about the network of genes that controls when pyrD switches on and off.

  If you browse EcoCyc for very long, you may fall under a peculiar spell. You may begin to imagine its genome as an instruction manual for an exquisite piece of nanotechnology crafted by some alien civilization. Its genome holds all the information required to assemble and run a sophisticated machine that can break down sugar like a miniature chemical factory, swim with proton-driven motors, and rewire its networks to withstand stomach acids and cold Minnesota winters.