Understanding the response of bacteria, one bug at a time

Much like people, individuals in a population of bacteria respond differently to their environment. The experimental apparatus developed in this work allows HFSP-funded researchers to examine these responses under controlled chemical and nutrient conditions.

HFSP Young Investigator Grant holders Kevin Dorfman, Pietro Cicuta, Bianca Sclavi and Marco Cosentino Lagomarsino and colleagues
authored on Thu, 28 March 2013

The classic view of a bacterium treats it as a “bag” of molecules, such as DNA and protein, held in place by the cell wall within the envelope of the cell membrane. However, recent research is making it clear that the interior of bacteria is actually highly organized and there is a diverse response within the population. In particular, there are a host of proteins known as nucleoid-shaping transcription factors that control the shape of the DNA. The shape of the DNA is also affected by the growth rate of the bacteria, since faster growth rates build and spatially organize more DNA for the daughter cells, and access the genes on the DNA more frequently for faster synthesis of cell components. The growth rate, in turn, is affected by the external chemical environment. In order to understand how changes in the environment can affect the response of bacteria in general, we need to understand the distribution of responses within a large population of bacteria exposed to the same chemical environment. 

The standard approach used to measure individual bacterium responses is to grow the bacteria on an agarose gel pad. Starting from a single cell, the bacteria gradually divide and expand outward as a single layer colony that can be imaged under a microscope. Unfortunately, the chemical conditions of these experiments are difficult to control. First, the nutrients in the agar are typically depleted at the center of the colony with time. As a result, the youngest bacteria at the edge of the colony have the most food and may respond differently than their now-starved ancestors. Second, the bacteria colony eventually expands so far that the gel pad deforms, leading to several layers of bacteria that prevents identification of single cells. Both factors limit the duration of a single experiment, making it tedious to obtain the large-scale data required to understand the diversity in the population.

We have developed a new experimental apparatus that allows us to examine the response of bacteria to their environment one cell at a time under controlled environmental conditions. Such experiments are called chemostatic, since the chemical and nutrient conditions are constant in time, and the corresponding devices are known as chemostats. Our “microchemostat” consists of bacteria-sized channels, fabricated in the rubbery material polydimethylsiloxane, and resembles a ladder. The two sides of the ladder are feeding channels, and the rungs of the ladder contain the bacteria. The width of the rungs is slightly smaller than the size of the bacteria, typically around one micron. We can thus gently hold the bacteria in place. The chemostat keeps the bacteria in the exponential growth phase by maintaining a weak flow through the rungs and preventing the bacteria from overtaking their environment; if the bacteria grow outside the rungs, they are removed in the feeding stream. Since the bacteria grow in straight lines, the image analysis is much simpler than traditional experiments that take place on an agar pad. As a result, we can easily examine thousands of bacteria per experiment in the device.

Figure: (a) Photograph of the microchemostat. The four circles are the inlets/outlets of the device. The “rungs” of the “ladder” are too small to see at this magnification.  (b) Fluorescence microscopy image of E. coli cells growing inside the rungs of the device. Each rung contains several cells expressing Green Fluorescent Protein. (c) Snapshot of the motion of fluorescent loci inside E. coli  cells growing inside the rungs of the device. Each dot corresponds to a single locus on the E. coli chromosome.

In this paper, we demonstrated the power of our device. In one example, we used a strain of bacteria that respond with a fluorescent reporter to changes in the growth conditions. We initially grew the bacteria in the device in a slow growth medium, switched to a fast growth medium, and then returned to the slow growth medium. Our device allowed us to do these experiments while observing the fluorescence of the bacteria with high temporal resolution. Importantly, we could completely exchange the medium within a given rung in seconds, making the change essentially instantaneous on the time scale of the bacteria’s responses. We observed an overshoot in the bacteria response that would be impossible to measure using standard experimental techniques. In a second example, we measured the mobility of thousands of fluorescently labeled loci on the DNA using microscopy powered by image analysis that allows resolution of almost one millionth of a millimeter to be reached. These experiments require that the bacteria be immobilized, since any bacteria swimming would mask the subtle jiggling motion of parts of the bacterial DNA. Our microchemostat results, easily obtained with an automated microscope setup, are consistent with the data obtained on agarose pads in a much more facile experiment. We anticipate that this microchemostat will provide a key technology enabling us to unravel the connection between DNA structure and growth conditions and to better understand the mechanisms allowing bacteria’s adaptation to changes in their environment.


Microfluidic chemostat for measuring single cell dynamics in bacteria. Long Z, Nugent E, Javer A, Cicuta P, Sclavi B, Cosentino Lagomarsino M, Dorfman KD. Lab Chip 2013, 13(5):947-954. doi: 10.1039/c2lc41196b.

Videos from the Dorfman lab 

Pubmed link