Bacterial Chemotaxis: How a Cell Navigates Without a Compass

A random walk with a bias

A bacterium like E. coli is too small to sense a concentration gradient across the length of its own body — the difference in nutrient concentration between its front and back is swamped by thermal noise. Instead it navigates with a strategy called run-and-tumble: swim roughly straight (a "run", driven by its flagella rotating counter-clockwise and bundling into a coherent propeller) for about a second, then briefly stop and reorient in a random new direction (a "tumble", caused by one or more flagella switching to clockwise rotation and the bundle flying apart), then run again in the new direction. With no gradient present, this produces an unbiased random walk that explores space without heading anywhere in particular.

How a temporal comparison creates a spatial decision

The bias that turns a random walk into directed motion is remarkably simple: the cell does not measure the gradient in space, it compares concentration now to concentration a few seconds ago, using a short-term chemical memory encoded in the phosphorylation state of an internal signalling protein (CheY-P) that controls the flagellar motor. If things are getting better while running — concentration increasing — the cell suppresses tumbling and extends the run; if things are getting worse, it tumbles sooner. Averaged over many runs, this produces a population that drifts up the gradient even though any single run points in a random direction: better-aimed runs simply last longer, so paths trending toward the food source dominate the statistics.

running "up" the gradient  (dC/dt > 0)  → suppress tumbling, run continues
running "down" the gradient (dC/dt < 0) → tumble sooner, pick a new random heading

net effect over many runs: a biased random walk that drifts toward higher concentration

Adaptation: why the response resets instead of saturating

A purely proportional response to concentration would saturate almost immediately, since real nutrient gradients span many orders of magnitude. E. coli instead implements near-perfect adaptation: the receptor cluster methylates itself over roughly tens of seconds to a minute, gradually resetting its baseline sensitivity so that a cell exposed to a step increase in concentration briefly suppresses tumbling and then returns almost exactly to its original tumbling frequency, ready to detect the next change. This adaptation is what makes the sensing genuinely a gradient detector — sensitive to the rate of change, not the absolute level — and it lets a single signalling pathway operate over roughly five orders of magnitude of background concentration.

From single cells to colony-scale patterns

Individual biased random walks aggregate into visible, population-scale structure. On a plate with a point source of nutrient, chemotactic bacteria form sharp travelling bands that migrate outward together, because cells that fall behind the band sit in a locally depleted, shallower part of the gradient and tumble more, which slows them relative to the band's leading edge — a self-organizing effect first analysed quantitatively by Keller and Segel in 1971, whose continuum PDE model (coupling a bacterial density field to a diffusing, consumed chemoattractant field) is still the standard mathematical description of chemotactic aggregation. Under nutrient stress, some species instead build multicellular biofilms — colonies embedded in a self-produced extracellular matrix that confers protection from antibiotics and desiccation — and within a biofilm or a dense colony, cells also communicate by quorum sensing: secreting and detecting a small diffusible signal molecule whose concentration reports local cell density, letting the population switch collective behaviours (like biofilm formation, bioluminescence or virulence factor release) only once a density threshold is reached.

Why microfluidics is the modern lab for this

Because the relevant length scales are microns and the relevant gradients are set up over minutes, microfluidic devices — channels and chambers etched at the scale of the cells themselves — have become the standard tool for chemotaxis research: they can hold a stable, precisely shaped concentration gradient in place (often by continuously diffusing a source and sink across a thin channel) for long enough to track thousands of individual cell trajectories by video microscopy, which is exactly how the run-length statistics behind the biased random walk were first measured directly rather than inferred.

Frequently asked questions

Try it live

Everything above runs in your browser — open Bacterial Colony and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.