In 1988, Richard Lenski, an evolutionary biologist now at Michigan State University, launched the longest running experiment on natural selection. It started with a single microbe–E. coli–which Lenski used to seed twelve genetically identical lines of bacteria. He placed each line in a separate flask, which he provisioned with a scant supply of glucose. The bacteria ate up the sugar in a few hours. The next day, he took a droplet of microbial broth from each flask and let it tumble into a new one, complete with a fresh supply of food. The bacteria boomed again, then starved again, and then were transferred again to a new home. Lenski and his colleagues have repeated this procedure every day for the past 24 years, rearing over 55,000 generations of bacteria.
I first reported on Lenski’s experiment 12 years ago, and since then I’ve revisited it every few years. The bacteria have been evolving in all sorts of interesting ways, and Lenski has been able to reconstruct the history of that evolution in great detail, thanks to a frozen fossil record. Every 500 generations Lenski and his students sock away some bacteria from each flask in a freezer. They can thaw out these ancestors whenever they wish and compare them to their youngest descendants. Biotechnology has improved drastically since 1988, giving Lenski an increasingly powerful evolutionary microscope. When he started out, it could take months to identify just one of the many mutations that arose in each lineage. These days, he and his colleagues can sequence an entire E. coli genome for a few hundred dollars and find every single new mutation in its DNA.
Four years ago, I wrote here about one particularly fascinating episode in the evolution going on in Lenski’s lab. It started in 2003, when the scientists there noticed something odd in one of the 12 flasks. It had become much more cloudy than the others. In a microbiology lab, that’s a sure-fire sign that the bacteria in a flask have experienced a population explosion.
At first the team suspected that some other species of bacteria had slipped into the flask and was breeding quickly. But they found that the flask was packed with E. coli—descendants of the original ancestor that Lenski had used to start the entire experiment. Somehow the bacteria in this one flask had evolved a way to grow much faster than the other bacteria.
The scientists determined that the bacteria had made a drastic switch: from feeding on the glucose to another compound, called citrate. Citrate is an ingredient in the broth where Lenski’s E. coli grows. It’s not food; instead, it helps keep the minerals in the broth in the right balance for E. coli to grow.
To a microbiologist, the emergence of E. coli that can eat citrate in a lab is deeply weird. E. coli typically can’t feed on citrate in the presence of oxygen. Some strains of E. coli can draw in citrate, but only if there’s no oxygen around. To make the reaction work, they have to pump out another compound called succinate at the same time. The ability of E. coli to feed on citrate in the presence of oxygen is extremely rare; it occurs when E. coli picks up the necessary genes from other species. In its normal environment (inside us), natural selection must not favor these mutants. Scientists have been studying E. coli in labs for over a century, making it the most intensely studied species, on Earth (as I explain in my book Microcosm). But in all that time, there has been only a single report of a citrate-feeding E. coli in a lab, back in 1982.
The inability of E. coli to grow on citrate is so stark, in fact, that microbiologists use it as a way to tell whether bacteria they come across are E. coli or not. It was thus a surprise to Lenski and his students to find a flask of E. coli suddenly feeding on citrate. The bacteria had not picked up the genes from another species. Instead, their new ability must have evolved after Lenski started his experiment with a single E. coli. This was not simply a case of natural selection enabling a species to do something better. This was a case of doing something new.
Zachary Blount–then a graduate student in Lenski’s lab and now a post-doctoral researcher there–led the investigation into this strange new development. Blount and his colleagues took away the glucose and found that the E. coli could thrive on citrate alone. They then defrosted the bacteria’s ancestors and fed them citrate to figure out when they acquired the ability. They found that a tiny fraction of the bacteria around generation 31,000 were able to grow very slowly on the citrate. Over a couple thousand generations, they got better at growing on citrate–so good, eventually, that they took over the flask and turned it cloudy.
Blount and Lenski first reported the evolution of the citrate-eaters in 2008. Now, after another four years of painstaking research, they’re back with a new paper that details what happened down at the molecular level. It’s a fascinating look at how new traits evolve by duplicating and recycling old parts.
The scientists found that the evolution took place in three chapters. In the first chapter, the stage was set for the transformation. In the second chapter, the bacteria became citrate feeders. And in the third chapter, they became much better citrate feeders.
It usually makes sense to start a story with Chapter One. But in this case, it’s better to start with Chapter Two: The Birth of the Citrate Feeders.
When E. coli finds itself in the absence of oxygen, it switches on a gene called citT. Like other species (including us), E. coli turns genes on and off by attaching proteins to short stretches of DNA nearby. When E. coli senses a lack of oxygen, proteins clamp onto one of these genetic switches near citT. Once they turn the gene on, it produces proteins that gets delivered to the surface of the cell. There they poke one end out into the environment and pull in citrate, while also pumping out succinate. After the citrate gets inside the microbe, the bacteria can chop it up to harvest its energy.
When E. coli is growing in oxygen, however, no proteins land on the genetic switch near the citT gene, and so it remains silent. The microbe wastes no energy making a protein that won’t help it grow.
Evolution has rewritten this little algorithm in the citrate eaters. As one cell in Lenski’s flask divided, it duplicated its DNA with one fateful mistake. It accidentally copied the citT twice. The new copy ended up near a different genetic switch–a switch that turns on neighboring genes in the presence of oxygen, not the absence.
While citT remained silent in the other bacteria in the flask, in the mutant cell, it switched on. The microbe began sticking citrate transporters on its surface and started drawing in the molecule. This mutation must have occurred by generation 31,500, when Blount found the earliest citrate eaters. The mutation was a crude hack; it produced a microbe that could draw in a little citrate, but not a lot. It still had to feed on glucose to get by.
Thus endeth Chapter Two.
In Chapter Three, life got better for the feeble citrate eaters. They copied the citT gene, along with its oxygen-switch promoter. Now the bacteria could make even more CitT channels, and thus pull in even more citrate. The bacteria made a third copy, and could pull in even more. Blount and his co-authors proved that the extra copies helped the bacteria this way by defrosting bacteria from Chapter Two and inserting copies of citT into them. Those early citrate eaters immediately got much better at feeding.
The scientists also found other mutations that arose during Chapter Three. While they have yet to figure out what those mutations did, the evidence they’ve gathered so far suggests the mutations allowed the bacteria to break down citrate more efficiently so they could get more energy from their food.
The most intriguing part of the story, however, is the first– Chapter One: Setting the Stage.
When Lenski and Blount first began to study the citrate eaters, they wondered what would happen if they wound back the evolutionary tape and let the bacteria re-evolve. Would the citrate feeding evolve again?
Blount thawed out ancestors from various moments in the history of the bacteria and started putting them through the same evolutionary experiment again. In some trials, the bacteria did indeed evolve into citrate eaters–but only if they came from after generation 20,000. This discovery suggested only after 20,000 generations were the bacteria prepared to evolve into citrate eaters. They must have already acquired other mutations that set the stage.
To test this idea, Blount and his colleagues thawed out some of the “prepared” bacteria: late-generation E. coli that had not yet gained mutations to citT. They created a miniature ring of DNA loaded with many copies of CitT and the oxygen-sensitive switch, and inserted it into the prepared bacteria. As they predicted, the bacteria now could suddenly feast magnificently on citrate.
But if Blount and his colleagues inserted the DNA ring into the original ancestor of the line, it grew poorly on citrate. That failure suggested that the early-generation bacteria were not ready to receive this evolutionary gift.
And thus a history takes shape:
Chapter One (from generaton zeo to at least generation 20,000): Our hero, E. coli, picks up mutations that don’t seem to have anything to do with feeding on citrate. They might have helped the bacteria grow better on their stingy rations of glucose. At least one of those mutations set the stage for feeding on citrate.
Chapter Two (around generation 31,500): The bacteria accidentally rewire their genome, so that a new copy of citT switches on in the presence of oxygen. Thanks to the mutations of Chapter One, this rewiring yields a modest but important improvement. Now the bacteria can feed a little on citrate, as well as on glucose.
Chapter Three (from about generation 31,500 to 33,000–and beyond): The bacteria make extra copies of the new and improved citT. They can pull in more citrate; new mutations fine-tune their metabolism to grow quickly on the molecule. World domination soon follows.
It’s remarkable how this experiment contains many elements of evolution that scientists have noted in other species. It’s common for genes to get duplicated, and for the new copy to be rewired for a new job. Snake venom, to pick one example, also evolved when genes were accidentally copied and then rewired. A gene that originally produced a digestive enzyme in the pancreas, for instance, now started making that enzyme in a snake’s mouth. It turned out to be a crude but effective venom. Later mutations fine-tuned the new venom gene until it became wickedly good.
The only important difference is that it took millions of years for snakes to evolve their arsenal of venoms, and scientists can only reconstruct their evolution by comparing living species. But in the case of E. coli, the transition unfolded fast enough for someone to track it from start to finish–and restart it when necessary.
Reference: Blount et al, “Genomic analysis of a key innovation in an experimental Escherichia coli population.” Nature, September 19 2012.