The upper surface of an Escherichia coli swarm is stationary

Abstract

When grown in a rich medium on agar, many bacteria elongate, produce more flagella, and swim in a thin film of fluid over the agar surface in swirling packs. Cells that spread in this way are said to swarm. The agar is a solid gel, with pores smaller than the bacteria, so the swarm/agar interface is fixed. Here we show, in experiments with Escherichia coli, that the swarm/air interface also is fixed. We deposited MgO smoke particles on the top surface of an E. coli swarm near its advancing edge, where cells move in a single layer, and then followed the motion of the particles by dark-field microscopy and the motion of the underlying cells by phase-contrast microscopy. Remarkably, the smoke particles remained fixed (diffusing only a few micrometers) while the swarming cells streamed past underneath. The diffusion coefficients of the smoke particles were smaller over the virgin agar ahead of the swarm than over the swarm itself. Changes between these two modes of behavior were evident within 10-20 microm of the swarm edge, indicating an increase in depth of the fluid in advance of the swarm. The only plausible way that the swarm/air interface can be fixed is that it is covered by a surfactant monolayer pinned at its edges. When a swarm is exposed to air, such a monolayer can markedly reduce water loss. When cells invade tissue, the ability to move rapidly between closely opposed fixed surfaces is a useful trait.

Fig. 1.

Two MgO smoke particles on the surface of an agar plate supporting an E. coli swarm, shown advancing from right to left (large arrow). One particle (at left) is over virgin agar well ahead of the swarm; the other particle (at right) is over the swarm. Two cells are shown (one truncated). They are rod-shaped, ≈1 μm in diameter by ≈5 μm long, and move in a thin layer of fluid (growth medium) at speeds of order 40 μm/s, more slowly at the edge of the swarm than farther behind. The fluid under the smoke particle at the left is shallower than the fluid under the smoke particle at the right. The fluid within the agar is more than 1,000 times deeper (≈1.6 mm).

Fig. 2.

A smoke particle in a stationary air/water interface. The particle is shown at the center of each panel before (A), during (B), or after (C and D) the arrival of a swarm. The cells moved with speeds of ≈20 μm/s in random directions near the edge of the swarm, whereas the swarm front drifted more slowly to the left at a speed of ≈1.7 μm/s. The smoke particle remained at nearly the same place but was free to diffuse, as shown by the red track in D, which followed the centroid of the particle for 6.7 s. The field of view is 21.4 μm × 13.6 μm, and elapsed time is shown at the lower left-hand corner of each panel. For video movies of this and other smoke experiments, see http://www.rowland.harvard.edu/labs/bacteria/movies_swarmecoli.html.

Fig. 3.

Diffusion of smoke particles. The mean-square displacements (MSD) of two particles are shown as a function of time (every third data point). The slopes of these curves were larger for a given particle over a swarm than over virgin agar.

Fig. 4.

Diffusion coefficient (D) of a particle as a function of its distance from a swarm front. (A) As the swarm approached the particle. (B) As the swarm moved beneath the particle. The depth of the fluid above the agar increased near the swarm's advancing edge. Changes in the diffusive behavior of particles were evident over a span of roughly 10–20 μm in front of advancing swarms.