Influence of topology on bacterial social interaction

Abstract

The environmental topology of complex structures is used by Escherichia coli to create traveling waves of high cell density, a prelude to quorum sensing. When cells are grown to a moderate density within a confining microenvironment, these traveling waves of cell density allow the cells to find and collapse into confining topologies, which are unstable to population fluctuations above a critical threshold. This was first observed in mazes designed to mimic complex environments, then more clearly in a simpler geometry consisting of a large open area surrounding a square (250 x 250 microm) with a narrow opening of 10-30 microm. Our results thus show that under nutrient-deprived conditions bacteria search out each other in a collective manner and that the bacteria can dynamically confine themselves to highly enclosed spaces.

Fig. 1.

Schematic diagrams of the fabrication process (A) and a single-square chip (B).

Fig. 2.

Epifluoresence images of GFP-expressing E. coli in a random maze. (A) Cells in the maze immediately after being loaded with a culture of a wild-type E. coli (RP437) in LB. (Magnification, ×600.) (B) Dynamical accumulation of cells at 2 h after the initial loading, showing accumulation in a “dead-end” part of the maze. (Magnification, ×400.) (C) Accumulation of cells into several different confining regions. No clustering is seen in the open region seen at the top. (Magnification, ×40.)

Fig. 3.

Epifluoresence images of RP437 E. coli grown in LB medium in a large open area with a center square. The center square is 250 μm per side and has a 30-μm-wide channel leading to the center. A culture of cells grown in LB are filled from both ends into a dry device, so the initial bacteria density is uniform. The small black rectangles are PDMS pillars that support the roof of the device. (A) Fifteen minutes after filling. (B) Thirty minutes after filling. (Inset) The pattern of E. coli around the opening at higher magnification. (C) Sixty minutes after filling. (D) Ninety minutes after filling.

Fig. 4.

Epifluorescence image of RP437 cells grown in M9 glycerol medium. Bacteria are inoculated from both ends into a dry device, so the initial bacteria density is uniform. The center square is 250 μm per side and has a 10-μm-wide channel leading to the square. The chip was photographed at 15 min (A), 30 min (B), 1 h (C), 1 h 30 min (D), 2 h (E), 2 h 30 min (F), 3 h 30min (G), and 5 h (H).

Fig. 5.

Comparison between the Δtar (RP2361) (A–D) and Δtsr (RP5700) (E–G) strains. The pictures are spaced at 15 min (A and D), 30 min (B and E), 60 min (C and F), and 90 min (D and G) after filling the chip with a culture of cells.

Fig. 6.

OD (A) and extracellular amino acid concentration (B) plotted as a function of time for RP437 cells grown in M9 glycerol medium.

Fig. 7.

Wild-type E. coli collapsing into confining squares in LB broth. The collapse of the population into the squares is shown as a function of time. Waves launch within the confining structure (A), steepen (B), and finally localize in the small squares (C). Each frame is separated by 1 h. Simulations of Eqs. 2–4, with bacterial growth set to zero, are shown below the microphotographs.

Fig. 8.

Relative density of the bacteria in the center rectangle (solid squares) shown in Fig. 3 as determined from relative optical luminance vs. time after the passage of a wave at time t = 0. The relative density was determined by measuring the average gray-scale luminance in the square and subtracting the average gray-scale luminance in an equal area on either side of the square in Fig. 3. As the wave passes, this number can become negative. The solid curve is a fit to the increase in density.