Microcosm

by Carl Zimmer

  This sort of behavior in bacteria turned a lot of microbiologists into neo-Lamarckians. E. coli seemed to respond to viruses the same way shorebirds responded to mud. The challenge had caused them to acquire resistance, which they could then pass on to their descendants. Other experiments seemed to fit this pattern as well. When scientists switched E. coli’s diet from glucose to lactose, it began to produce the enzyme necessary for feeding on lactose, as did its descendants. And one other factor also made many microbiologists into neo-Lamarckians: there was little evidence that bacteria had genes. As far as many microbiologists could tell, a microbe such as E. coli was nothing but a bag of enzymes and other molecules that could react to changes in its environment.

  But some microbiologists thought otherwise. They argued that bacteria did have genes, and that, like the genes of animals, these could mutate spontaneously. In some cases, a mutation might, through pure luck, give a microbe an advantage, such as resistance to a virus. According to this rival explanation, E. coli followed Darwin’s rules, not Lamarck’s.

  No one had put the alternatives to a good test, and Luria and Delbrück spent months puzzling over how they might do so. They had failed to come up with an experiment by 1942, when they parted ways after Luria accepted a job at Indiana University, “a place I had never heard of,” he wrote later. Not long afterward Luria found himself in Bloomington sitting next to a colleague who was playing a slot machine. The professor was losing, and when Luria teased him he stalked off.

  “Right then I began giving some thought to the actual numerology of slot machines,” Luria wrote in his autobiography.

  The slot machine the professor was playing was programmed to deliver only a few big jackpots. It might have been built differently. It might have provided the same small chance of paying out a jackpot on every pull of the arm. In that case the jackpot would have given out many more prizes, but much smaller ones. Suddenly Luria realized he had figured out how to run an experiment on E. coli’s resistance that could test Darwin’s theory versus Lamarck’s.

  The next day Luria began rearing flasks of E. coli. Each flask started out with just a few hundred microbes. Since resistant E. coli are extremely rare—about one in a million—the founders of each flask were all almost certainly vulnerable. Any resistance to viruses would appear in the flask only after its population began to grow.

  After the E. coli populations had grown for a while, Luria took some bacteria from each one and spread them on petri dishes laced with viruses. He waited for epidemics to strike, and then for resistant E. coli colonies to emerge.

  According to Lamarck, living things acquire new traits as they face new challenges, then pass these traits down to their offspring. If Luria’s E. coli obeyed Lamarck, the bacteria would acquire resistance after Luria exposed them to viruses. That would mean that once Luria had inoculated his virus-laden dishes, every microbe had the same small chance of evolving resistance. Luria ought then to have discovered a few resistant colonies in every dish. The experiment would have resembled a slot machine that pays out a lot of small wins.

  If E. coli obeyed Darwin, on the other hand, the experiment would play out like a slot machine with a few big wins. According to Darwin’s followers, E. coli has a rare random chance of mutating every time it divides regardless of what it is experiencing. In other words, the bacteria in Luria’s experiment might have acquired resistance to viruses while they were growing in the flasks, long before Luria exposed them to the viruses. That head start would have produced a very different result for the experiment. If a mutation had emerged early on in one of the colonies, the mutant would have had a lot of time to produce offspring. When Luria took some of the bacteria from such a colony and placed them in a petri dish with viruses, a fair number of them would already be resistant. They would grow into many new colonies in the dish.

  In some of the other flasks, resistant mutants would arise much later. They would have had less time to produce offspring by the time Luria exposed them to viruses. As a result, they’d produce fewer colonies in the petri dishes. And in still other flasks, no mutants would arise at all. Their bacteria would all die, leaving their dishes empty. So instead of a few colonies growing in most dishes—the Lamarckian prediction—Darwinian mutations would produce a few dishes loaded with colonies and the rest with few or none.

  Luria let his slot machines spin, and then he began to count spots. When he was done, the verdict was clear: a few dishes were packed with colonies while many were empty. Life’s slot machine had paid out a handful of big jackpots. Darwin had won.

  In 1943, Luria and Delbrück published these results, which would earn them shares in a Nobel Prize in 1969. Later generations of biologists would look back at Luria’s experiment as one of the greatest of the twentieth century. It provided compelling evidence that bacteria, like animals and plants, pass down their traits to their offspring through genes. It showed that those genes change spontaneously, and they can become more common in a population through natural selection. And the experiment became a powerful scientific tool: simply by counting colonies of bacteria, scientists can calculate how often mutations arise.

  But when Luria and Delbrück first published the experiment, they did not bowl over the skeptics. Neo-Lamarckians remained unconvinced, pointing out that the researchers had had to rely on a lot of indirect clues to draw their conclusions. It was possible that the test tubes had not all been alike. Some might have had traces of soap or some other contamination that might have altered the bacteria. For another decade, microbiologists went on debating how bacteria adapted.

  The controversy did not die until Joshua Lederberg, the scientist who discovered E. coli sex, tested the jackpot hypothesis with a new experiment. Lederberg and his wife, Esther, wrapped sheets of velvet around the ends of wooden cylinders that were as wide as a petri dish. The Lederbergs then stamped the velvet into a dish of E. coli, coating the material with the microbes, and then pressed it into a dish stocked with viruses. The Lederbergs repeated the procedure, stamping three virus-laden dishes with E. coli from the same original dish.

  Within a few hours almost all the bacteria the Lederbergs had put in the dishes were dead from infections. A few mutants survived, however, and began to produce colonies visible to the eye. The Lederbergs photographed each dish and then looked at the pictures side by side. The constellation of mutant colonies was the same in each dish.

  The Lederbergs concluded that the bacteria must have acquired mutations in the original dish. When they stamped it, the Lederbergs picked up mutants from the same spots. They transferred the bacteria to the same spot in the dishes laden with viruses. If E. coli had obeyed Lamarck, it would have acquired resistance only after the Lederbergs had exposed it to the viruses. There would be no reason to expect resistant bacteria to emerge in precisely the same spots in different dishes.

  The Lederbergs recognized that they were seeing resistant bacteria only after they had been exposed to the viruses, so they took the experiment one step further to prove that the mutants were resistant before they encountered viruses. They pressed a velvet stamp into a dish that contained just a few colonies and then pressed it into a dish full of viruses. They waited for resistant bacteria to produce new colonies in the virus-laden dish. Each new colony corresponded to a colony in the original dish. The Lederbergs took some bacteria from the original colonies and put them in flasks, where they could breed into huge numbers. They then repeated the experiment on the new bacteria, growing a few colonies in a dish and pressing them with the velvet stamp. Now all the colonies were resistant. The Lederbergs seeded a second flask of bacteria from the dish and repeated the experiment yet again.

  No matter how many times they repeated the procedure, the bacteria remained resistant to viruses even though none of them had been exposed to viruses over the course of the experiment. In 1952, the Lederbergs published their results, arguing that a few resistant bacteria had acquired mutations before the experiment began. Those bacteria had passed
down the resistance gene to their descendants. To cling to Lamarck now became absurd.

  These experiments on E. coli helped fuse evolution and genetics into a new synthesis. And as scientists continued to learn more about genes and proteins, the workings of natural selection became more clear. A mutation may change the sequence of a gene and thus the structure of its protein. In some cases, a lethal mutation might disable an essential protein. Others make no difference. And a few actually increase reproductive success. The advantage or disadvantage of a mutation depends on the environment. A mutation that confers resistance to viruses will give E. coli a reproductive advantage if viruses are menacing it. If not, the mutation makes no difference. It may even be a burden.

  Over the past fifty years, evolutionary biologists have heaped up a mountain of evidence demonstrating that evolution does indeed take place this way. In most cases, though, they have had to study evolution indirectly, by comparing the genes of different organisms to see how natural selection has driven them apart from a common ancestor. But in a few species scientists have observed evolution as it happens, generation by generation, mutation by mutation. Among the most generous of these species is E. coli.

  EVOLUTION UNFOLDING

  When Salvador Luria ran his slot machine experiment, he captured a single round of evolution. A population of E. coli faced a challenge—an attack of viruses—and natural selection favored resistant mutants. But in every generation, natural selection can shape a species. New mutations arise, genes mix to form new combinations as they pass from parent to offspring, and the shifting environment creates new challenges. On this grander scale, evolution can be far harder to observe. Life has had millions of years to change, whereas scientists are on this earth for only a few decades. Darwin had resigned himself to studying evolution from a distance, and a century later most evolutionary biologists were following suit. They would compare genes in different species to learn how they diverged or search for new versions of genes that had arisen in response to new challenges. They would look for the effects of natural selection in the past. But in the 1980s a number of scientists decided to watch evolution in the present. They set out to observe E. coli and other bacteria undergo natural selection in their laboratories.

  One of those scientists was Richard Lenski. Lenski started his scientific career hiking the Blue Ridge Mountains in search of beetles. He wanted to learn how beetles help hold together the Southern Appalachian food web. Lenski focused his work on a handful of species of Carabus ground beetles. He hoped to determine what controlled their population—cold snaps and heat waves perhaps, or maybe the competition for prey. The question was not just academic. The ground beetles might well be protecting the forests by keeping tree-destroying pests in check. Understanding the ecology of ground beetles might make it possible to predict outbreaks of pests and perhaps even prevent them.

  Each spring, Lenski climbed the slopes and dug holes. He put plastic cups in them, covering the cups with funnels. Beetles tumbled down the funnels into the cups, and Lenski returned each day to count them. He marked the beetles and set them free. He tracked how much weight they gained each summer. He compared how many Carabus sylvosus he caught with how many Carabus limbatus. He compared how many beetles lived in dense forests with how many inhabited clear-cuts.

  Lenski looked for patterns. In science, patterns become stronger the more times an experiment can be repeated. Doctors put thousands of people on an experimental drug. Physicists fire a laser millions of times to discover the ways of the photon. Ecologists also replicate their experiments when they can, but each datum demands far more labor. For his clear-cutting study, Lenski built a grand total of four enclosures, two in the clear-cut and two in the forest, each holding sixteen traps. With so few trials he could catch sight of only fleeting shadows, hazy signs of the forces governing the beetles.

  Lenski came down from the mountains. He decided he would have to find another creature he could study to get some answers to the big questions on his mind. He found E. coli. When Lenski looked at a flask of E. coli, he saw a mountain. It was an ecosystem filled with billions of individual organisms. Like his beetles, E. coli searched for food and reproduced. They were preyed upon by viruses rather than by salamanders. E. coli’s ecosystem might be simpler than the Blue Ridge Mountains, but simplicity can be a virtue in science. A researcher can precisely control every variable in an experiment to see the effect of each one.

  Best of all, E. coli is the sort of creature that can, in theory, evolve very fast. Mutations may occur only rarely, but with millions of microbes in a single flask a few mutations will arise in every generation. And because E. coli can reproduce in as little as twenty minutes, a beneficial mutation may let a mutant overtake a colony in a matter of days.

  Lenski set up an experiment that was simple yet powerful. He gave his bacteria a limited supply of glucose and thus created a huge evolutionary pressure. Their ancestors had been fed endless meals of sugar, and they had adapted to that diet. The microbes that could convert the food to offspring fastest took over the population. In Lenski’s experiment, genes that sped up breeding were no longer beneficial. His bacteria grew slowly if at all. Any new mutation that allowed the microbes to survive the conditions better, Lenski reasoned, would be strongly favored by natural selection.

  As E. coli passed through thousands of generations in his laboratory, evolution’s mark began to emerge. When Lenski pitted the ancestral bacteria against their descendants on their new diet, the new microbes reproduced faster. The more time passed, the better adapted the bacteria became. After a decade, the bacteria could grow far faster than their ancestors. The course of their evolution was not smooth: the bacteria might spend several hundred generations without any observable change, only to go through a rapid evolutionary burst. And as E. coli evolved to grow faster, Lenski detected other changes.

  Lenski’s students continue to nurture his dynasty of E. coli from one generation to the next, and other scientists have used similar methods to run experiments of their own. Some have watched E. coli adapt to life at the feverish temperature of 107 degrees Fahrenheit. Others have unleashed viruses on the bacteria and observed them become resistant, only to have the viruses evolve ways to overcome their resistance, starting the cycle all over again. While Lenski’s experiment remains the longest running by far, much shorter experiments have been able to yield striking results. Bernhard Palsson and his colleagues, for example, fed five populations of E. coli glycerol, a carbon compound used in soaps and face creams. Ordinary E. coli does a lousy job of feeding on glycerol, but Palsson drove the evolution of glycerol gourmets. After only forty-four days (660 generations of E. coli), the bacteria could grow twice as fast as the founders of the population.

  Whether it battles viruses, adapts to a diet of glycerol, or copes with heat, E. coli unmistakably evolves. Its swift pace of evolution in these experiments may reflect rapid evolution in the wild. After all, each time the microbe finds itself in a new environment, its evolutionary pressures suddenly shift. Genes that allow E. coli to thrive in a gut may mutate into forms better suited to life in the soil.

  These experiments have allowed scientists to put natural selection under a microscope, teasing apart the individual mutations that benefit E. coli. Each time the microbe divides, it has a roughly 1-in-100,000 chance of mutating in a way that lets its descendants grow faster. The boost is often small, but it can allow a mutant’s descendants to outbreed their cousins. And those mutants in turn have a small chance of picking up a second mutation that makes them even faster growers. In Palsson’s 660-generation experiment, he and his colleagues confirmed two or three mutations in each population. Lenski estimated that over the course of 40,000 generations his lines have picked up as many as 100 beneficial mutations.

  Beneficial mutations can take several forms. Some involve the change of a single base in a gene, something equivalent to changing LIFT to LIFE. These mutations can change the structure of a protein and thus change the way
it works. It may slice a molecule more effectively than before, or start responding to a new signal. Other mutations accidentally create an extra copy of a stretch of DNA. In Palsson’s experiment these duplicated segments ranged from 9 bases long to 1.3 million. Accidental duplications can create new copies of old genes. Natural selection may favor them because they produce extra proteins, which E. coli can use to grow and reproduce. But over time one of the copies may acquire new mutations, allowing it to take on a new function. Mutations can also snip out chunks of DNA, and microbes that lose genetic material are sometimes favored by natural selection. It’s possible that proteins that were originally useful become a burden to E. coli.

  Experiments such as these show that mutations arise randomly. And the effects of the mutations depend on how the mutations allow an organism to thrive in its own peculiar set of conditions. But does that mean evolution plays out purely by chance? The late paleontologist Stephen Jay Gould dreamed of an experiment to answer the question, which he called replaying life’s tape. “You press the rewind button and, making sure you thoroughly erase everything that actually happened, go back to any time and place in the past…,” he wrote in his 1989 book Wonderful Life. “Then let the tape run again and see if the repetition looks at all like the original.”

  Short of time travel, Gould thought the best way a scientist could answer that question was by examining the fossil record, documenting the emergence and extinction of species. But experiments on E. coli can also address the question, at least on a scale of years. What makes experiments such as Lenski’s particularly powerful is that evolution unfolds many times over, not just once. From an identical ancestor, Lenski produced twelve lines, each of which went through its own natural selection. Lenski and his colleagues may not be able to rewind the tape of E. coli’s evolution, but they can create twelve identical copies of the same tape and watch what happens when they all play at the same time.