On the Hunt: How Bacteria Find Food

This post is by MSU graduate student Joshua Franklin

Imagine you are half-starved, blindfolded, then placed into a large gymnasium with a plate full of freshly-baked cookies. How do you find the cookies? You could try to randomly walk around until you step on them. Interestingly, you are mathematically guaranteed to reach the cookies this way, even if the gym is infinitely large[1], but that could take a very long time, and you’re hungry now. Luckily, you can use your senses to speed things up.

It turns out that smell is the only useful sense here, as your vision is blocked and your other senses are either too short-ranged (taste, touch) or functionally unable (hearing) to help you find the cookies. But, generally, you cannot use smell to turn towards distant objects, because the change in odor intensity is too weak to detect when standing in place and turning around. So how can you find the cookies?

You could start by sniffing, walking for a while, then sniffing again. If the odor got stronger, then you must be moving towards the cookies. Great! You should keep moving in that direction. If the smell got weaker, that’s okay; just spin around and try again in a new direction. By following this procedure you will reach the cookies much more quickly than before (Figure 1).

Figure 1. A computer simulation of the paths taken by a person walking randomly (red) and a person using the cookie-finding procedure (blue). The black spot is the plate of cookies. Both strategies were allowed to walk for the same amount of time and both walkers move at the same speed. Even though the random walker starts closer to the cookies, they make far less progress than the cookie-finder.

It turns out that many bacteria use this cookie-finding procedure to help them move towards food, and in this context the procedure is called chemotaxis. Why do bacteria need to do this? Well, many kinds of bacteria can move using a flagella, which is a long filament that sticks off of the cell surface and rotates, working like a propeller to push the cell forward. But just like our noses are too weak to directly help us turn toward the cookies, bacteria are too small to directly sense the direction food is in. So bacteria use chemotaxis to move towards their food by swimming, sensing whether it has moved towards or away from the food, then deciding whether to keep swimming in the same direction or to change directions.

But there’s a hidden assumption here. Think back to the cookie-finding procedure. You smelled, walked, then smelled again and compared the current smell to the previous smell. That means that you need to remember what the smell was previously. Can bacteria actually do this? It turns out the answer is yes. A series of chemical reactions inside of the cell stores information about how strong the “smell” of the food was previously, so that the bacteria can tell if they have moved toward or away from the food source.

The reason chemotaxis works so well for bacteria is that, at the size of a bacterium, diffusion spreads chemicals into gradients very quickly. In fact, the bacterial world is dominated by diffusion-generated chemical gradients. This is hugely different from the world we normally see, where diffusion plays a only minor role[2]. From a bacterial point of view, the world is a series of chemical gradients that can lead them toward food or away from predators, and chemotaxis enables the cells to navigate these gradients effectively.

Figure 2. A diagram of the chemotaxis system in the E. coli. The network that controls the cell decision-making process is composed of only a handful of different proteins. Stars are molecules cells can “smell”, and the rectangular white bars are the sensory proteins. Letters in shapes are chemotaxis-associated proteins. Filaments coming off of the cell surface are flagella, which rotate to push the cell forward.

Our lab seeks to identify traits that affect how cells perform chemotaxis. Bacteria carry out chemotaxis using a set of proteins that detect food molecules outside of the cell and control the cell’s movement (Figure 2). Biochemical processes in this system control how well the cell performs chemotaxis in different environments. Cells which swim towards food more effectively will reproduce more often than cells which can’t, so natural selection will tend to optimize the chemotaxis system for a given environment.

While we have a pretty good understanding of how the chemotaxis system works, it is still difficult to predict how the strategy should be optimized to suit different environments. For example, how long should a cell swim before it is confident that it is going in the right or wrong direction? What if there are obstacles along the way? Which of the many chemicals that can be sensed should be followed? There are still many questions with unintuitive answers that need to be explored to understand why we see so much diversity in motile behavior and morphology in the microbial world.

Our lab has created a web app to explore the evolution of chemotaxis. The virtual environment consists of a rectangular world with a food gradient that increases from left to right. The simulation places cells into the world where they can perform chemotaxis to move towards the food. Each cell is given a number of traits, which control the chemotaxis system as described above and their fitness in the virtual world.

In this simulation cells that are better at chemotaxis reproduce more often than cells that are not as good. Eventually, cells with traits that are optimal for the specified environment will dominate the population. There are a number of environmental details you can change, such as the strength of the chemoattractant, the shape of the chemical gradient, and the presence of obstacles that the cell must navigate to reach the target. Different environments will select for different chemotaxis traits. For more details, see the guide on our website.

While the simulation is a simplified version of the E. coli chemotaxis system, it reproduces the behavior of real bacterial cells really well. Modeling and simulations allow us to explore the behavior of bacteria to generate hypotheses that can be tested in the laboratory. Part of my research involves using models and simulations to understand the performance tradeoffs that are imposed on bacterial motility.

The combination of biological and computational work in my research is a fantastic opportunity for me, as I have enjoyed programming since my early teens, and basically living in a forest as I was growing up nurtured my interests in biology. Being able to combine both of my main interests helps keep me engaged in my research despite the challenges it poses. When experiments are testing my patience I can generate news ideas by doing computer work, and when computer work wears thin I can refresh by returning to the bench.

Figure 3. Me in action at MSU Science Fest 2018.

One of my favorite features of my research is that, as this blog post shows, it provides an opportunity to design and write programs that allow students to study non-intuitive aspects of biology using interactive tools, without the constraints of setting up experiments. I think it can also help students appreciate the power of modeling and simulations in exploring the complexity of biological systems. Using educational programs for public outreach and education is something that I feel strongly about (Figure 3), and hope to expand upon as my graduate career continues.

[1]Technically, this assumes the floor is two-dimensional and space is discrete at some level. Perhaps even more interestingly, a bird flying around randomly in an infinite gymnasium would not be guaranteed to ever reach the cookies, even if given an infinite amount of time. For an approachable explanation of both phenomena, see: https://www.youtube.com/watch?v=stgYW6M5o4k

[2]When in school, I remember our teacher explaining diffusion by asking us what happens when when someone breaks a bottle of perfume in a store. We said “Everyone starts to smell it”. The teacher explained this as diffusion, but in reality it’s due to air currents. For example, in perfectly still air it would take about three days for oxygen (D = 0.176cm2/s, according to Wikipedia) to diffuse a distance of 10 feet. See: http://www.physiologyweb.com/calculators/diffusion_time_calculator.html


This entry was posted in BEACON Researchers at Work and tagged , , . Bookmark the permalink.

Comments are closed.