Survival of the weakest – when doing poorly does best
"Survival of the fittest" is a phrase coined by Herbert Spencer upon his reading of Darwin's On the Origin of Species to describe the process of natural selection. In common parlance, the fittest member of a population is the strongest, fastest, biggest and, in general, the best. However, the word "fitness," as used by biologists, is not identical to its lay usage. A small, unarmored fish may be more fit than more armored ones in the absence of predators. Pairs of wild turkey brothers who cooperate in wooing mates are more fit than aggressive loners. Fitness is a description of how many descendants an organism will leave, be it through strength, resistance to disease, ability to acquire resources or some other mechanism. Here, I will describe some work I've participated in where these two definitions of fitness disagree.
First, I should introduce myself. My name is Joshua Nahum, and I am a graduate student in Dr. Ben Kerr's lab at the University of Washington, Seattle. The work I will describe below was done by Brittany Harding, Ben Kerr and myself. The work we do in the Kerr Lab is centered around performing evolution experiments in real time, where we can watch evolution as it happens. To do this, we use rapidly evolving systems (as experiments which last thousands of years are generally not awarded grants). These include microbial systems, such as E. coli and Pseudomonas fluorescens (which have a generation time of less than an hour) and digital systems, like Avida (that have a generation time of microseconds). We pick an organism to be the founding member, or ancestor, of a population. When put in a suitable environment, the ancestor replicates into a large population that with time evolves. At the conclusion of the experiment, we take isolates from the ending population (called descendants) and perform measurements to see how the population changed. We often perform competitions between the ancestor and descendants in different environments to see what traits are needed to succeed (to be fit in an environment). And now, on to this experiment.
E. coli and most other bacteria produce a number of compounds that are meant to inhibit the growth of competitors. One class of such toxins produced by E. coli are called colicins (we scientists are very inventive when we come up with names). For those who've been reading this blog regularly, this topic was covered in another BEACON Researchers at Work post, "Colicin and Immunity Binding: A Love Story," by Carrie Glenney, a fellow graduate student in the Kerr Lab. Being a colicin producer is costly, as each cell needs to make immunity proteins to not be killed by its relatives. And, a proportion of the Producer population needs to die every generation to release the colicin into the environment. But these costs are borne because the colicin can kill other sensitive E. coli. If you mix a flask of Producer with Sensitive, no Sensitive cell will survive. However, there is a small chance that a Sensitive cell will evolve to be Resistant. You have probably heard of antibiotic-resistant bacteria; this is similar. The mechanism of resistance varies, but often involves the costly loss of a nutrient-uptake transporter (the colicin binds to the transporter to infiltrate the Sensitive cell). This loss of a transporter mildly cripples the cell, making it less efficient at taking up nutrients, and slows its growth rate. Unfortunately for pharmaceutical companies, Resistant populations can often evolve to reduce or eliminate this cost though compensatory mutations. So now we have three types of bacteria: Sensitive (fastest growers), Resistant (slower growing, but not killed by colicin), and Producers (slowest growers, but able to kill Sensitive).
male mating behavior of side-blotched lizards. No single type dominates the population, because as one comes to prominence (say Resistant), its enemy (Sensitive) is quick to rise as well. If you distribute the three types on a surface (like an agar petri dish, or in our case, the grid of hundreds of small containers of liquid growth media), they form many patches of each type, whose boundaries move according to the competitions described above (Sensitive invades Resistant, etc…). The patches appear to chase each other, if they were to be viewed with time-lapse photography.
But the most interesting things happen when you consider the evolution of the system. The Resistant types is most able to evolve (change its growth rate), because it can acquire compensatory mutations which reduce the cost of resistance, allowing it to grow faster. At a glance, faster growth would be highly advantageous, as it would help repel invasion from Sensitive types, and allow the Resistant population to better invade both Producer and slower growing Resistant populations. And indeed, the evolution of faster growth is what we find in most of our experiments. However, when we limit migration (which is how populations move across the world) to small distances (only allowed the bacteria to move slowly), the Resistant cells did not evolve faster growth rates.
What we observe has been called “Survival of the Weakest,” a situation where selection favors the “weaker” competitor. Here's why: In our experiment there are many patches of each of the three types chasing each other. Let's say one of the Resistant cells gains a compensatory mutation and can now grow faster (I'll call it Superbug). Superbug will take over its patch of Resistant cells, until the patch of consists solely of Superbugs. The Superbug patch will quickly invade every adjacent patch of its victim, Producer. After all accessible Producer patches are consumed, the Superbug will be in a bad place, as it will be completely surrounded by its enemy, Sensitive. Superbug will then be invaded by Sensitive and go extinct. Meanwhile, the normal Resistant cells will continue to thrive by only invading Producer at a sustainable rate. At the end of the experiment, most Resistant cells will be the slower growers. We found restraint (slower growth) didn't evolve when migration occurred over long distances (as patches couldn't form). With long migrations distances, the three types mixed together more freely, and faster growth is favored because there is no need to conserve resources (in this case the resource is Resistant's victim, Producer).
This result, the evolution of a worse (slower growing) competitor goes against conventional wisdom that the better competitor will also be the most fit. If I could impart a takeaway message, it would be that interesting ecologies (like RPS) can generate interesting evolutionary outcomes that aren't necessarily intuitive.
For more information regarding this work, please see our recently published paper in PNAS.
For more about Josh's work, you can contact him at nahumj at uw dot edu.