Periodic Growth of Bacterial Colonies


Department of Physics, Chuo University,

Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan

The growth of bacterial colonies is a ``Treasure Island'' of pattern formation. Apparently, the patter-forming members (bacterial cells) are strongly interesting, self-reproducible and self-driven ``particles'', very different from conventional physical systems. Some colony patterns grown on the surface of agar plates can be understood in terms of well-defined models of pattern formation such as DLA, but not in call cases.

Here we used a bacterial species {\it Proteus mirabilis}. This species has been well-known for more than a century to form a macroscopically almost perfect concentric-ring like colony with approximately equal spacing of terraces on the surface of semi-solid substrate such as an agar plate. The colony grows cyclically with the interface repeating an advance (migration) and a step (consolidation) alternately. We distinguish three phases (initial lag phase, the following migration and consolidation phases that appear alternately) for the colony growth. Even when we decreased the inoculum cell density to some threshold value, the interval of the lag phase was almost constant. However, if the inoculum density was decreased below th threshold, the lag phase became prolonged. This clearly suggests that it is necessary for the cell density to reach some threshold value for the onset of migration. When we cut a colony just behind a migrating front shortly after, say, the fourth migration started, the first migration after the cut ended earlier and the following consolidation lasted longer. However, the following cycles were not influenced by the cut at all. We also checked phase entrainment by letting two colonies collide with each other and found that it does not take place in macroscopic scales.

All these experimental results suggest that
(1) no global chemical signals from the center exist for the colony formation,
(2) the periodic growth cannot be explained by the idea of interacting nonlinear
oscillators, and
(3) there may be two threshold values of cell density, the higher threshold for the
onset of migration and the lower one for the onset of consolidation.
In order to form periodic colonies {it Proteus mirabilis} seems to use ``quorum sensing'' mechanism which is intrinsically multicellular.