SummaryPrkC was shown to be a eukaryotic-like (Hanks-type) protein kinase from Bacillus subtilis with a structural organization similar to that of the eukaryotic sensor Ser/Thr or Tyr kinases (e.g. the TGF b b b b or PDGF receptors). The molecule consists of a catalytic domain located in the cytoplasm, joined by a single transmembrane-spanning region (TMD) to a large extracellular domain. Using a genetic reporter system, involving the cI repressor of lambda, evidence was obtained indicating that PrkC forms a dimer, involving both the TMD and the external domain in dimerization. The purified catalytic domain of PrkC was shown to autophosphorylate and to phosphorylate an external target, MBP, in both cases on threonine. These two functions require the completely conserved K40 residue in subdomain II, which is essential for enzymatic activity. Importantly, both the mutant deleted for prkC and a K40R mutant exhibit decreased efficiency of sporulation and a significant reduction in biofilm formation, demonstrating that the catalytic activity of PrkC is necessary for these two developmental processes. In addition, we showed that the product of prpC , a PPM phosphatase encoded by the adjacent gene, co-transcribed with prkC , is also required for normal biofilm and spore formation.

The natural wild-type Bacillus subtilis strain 3610 swarms rapidly on the synthetic B medium in symmetrical concentric waves of branched dendritic patterns. In a comparison of the behavior of the laboratory strain 168 (trp) on different media with that of 3610, strain 168 (trp), which does not produce surfactin, displayed less swarming activity, both qualitatively (pattern formation) and in speed of colonization. On E and B media, 168 failed to swarm; however, with the latter, swarming was arrested at an early stage of development, with filamentous cells and rafts of cells (characteristic of dendrites of 3610) associated with bud-like structures surrounding the central inoculum. In contrast, strain 168 apparently swarmed efficiently on Luria-Bertani (LB) agar, colonizing the entire plate in 24 h. However, analysis of the intermediate stages of development of swarms on LB medium demonstrated that, in comparison with strain 3610, initiation of swarming of 168 (trp) was delayed and the greatly reduced rate of expansion of the swarm was uncoordinated, with some regions advancing faster than others. Moreover, while early stages of swarming in 3610 are accompanied by the formation of large numbers of dendrites whose rapid advance involves packs of cells at the tips, strain 168 advanced more slowly as a continuous front. When sfp ؉ was inserted into the chromosome of 168 (trp) to reestablish surfactin production, many features observed with 3610 on LB medium were now visible with 168. However, swarming of 168 (sfp ؉ ) still showed some reduced speed and a distinctive pattern compared to swarming of 3610. The results are discussed in terms of the possible role of surfactin in the swarming process and the different modes of swarming on LB medium.We have recently described a wide range of swarming patterns, including successive waves of dendritic branching, followed by consolidation, of the wild-type strain 3610 on the surface of a synthetic agar medium (9). A simpler, apparently nondendritic swarming pattern of Bacillus subtilis on LuriaBertani (LB) agar has also been described (9, 10), in particular, with the wild-type strain 3610, although the details of the intermediate stages of these swarm communities during development have not so far been reported. In contrast, we showed that the formation of the complex branching communities on the synthetic B medium, as detected by an in situ microscope analysis, is unexpectedly complex, with a precise developmental program accompanied at different stages of the swarming process, for example, by the appearance of long filamentous cells and rafts of tightly aligned cells. There was no evidence, however, that these types of cells were directly involved in swarming per se (9). Moreover, in our previous study we also detected the apparent migration of large numbers of individual cells prior to their aggregation into the tips of nascent dendrites. This suggested that, in at least some stages of the swarming process, migration does indeed proceed independently of the morphologically d...

After optimizing the conditions, including nutrients and temperature, swarming of Bacillus subtilis 3610 was obtained on a synthetic, fully defined medium. The swarms formed highly branched (dendritic) patterns, generated by successive waves of moving cells. A detailed microscopic in situ analysis of swarms 1 and 2 revealed varied cell morphologies and a remarkable series of events, with cells assembling into different ‘structures’, as the architecture of the swarm developed. Long filamentous cells begin to form before the onset of the first swarming (11 h) and are again observed at later stages in the interior of individual mature dendrites. Swarm 2, detected at 18–22 h, is accompanied by the rapid movement of a wave of dispersed (non-filamentous) cells. Subsequently at the forward edge of this swarm, individual cells begin to cluster together, gradually forming de novo the shape of a dendrite tip with progressive lengthening of this new structure ‘backwards' towards the swarm centre. In both swarms 1 and 2, after the initial clustering of cells, there is the progressive appearance of a spreading monolayer of rafts (4–5 non-filamented cells, neatly aligned). The alternative possible roles of the rafts in the development of the swarm are discussed.