Analysis of the c-di-GMP mediated cell fate determination in Caulobacter crescentus

Cyclic-di-GMP (c-di-GMP) is a ubiquitous second messenger in bacteria, which has been
recognized as a key regulator, antagonistically controlling the transition between motile,
planktonic cells and surface attached, multicellular communities. The biosynthesis and
degradation of c-di-GMP are mediated by the opposing enzymatic activities of di-guanylate
cyclases (DGCs) and phosphodiesterases (PDEs), generally in response to internal and
environmental signals. These activities reside in GGDEF and EAL domains respectively,
which represent two large families of output domains often found in bacterial one- and twocomponent
systems. In this work, the cell cycle-embedded differentiation from a free-living,
motile swarmer cell into a sessile stalked cell in the model organism Caulobacter crescentus,
and the role of c-di-GMP in this process was investigated. A systematic analysis was used to
identify key regulatory enzymes involved in c-di-GMP metabolism that influence this
developmental process. The function and regulation of these genes was then examined. One
component that has already been implicated in this process, the DGC PleD, was investigated
in more detail, with special emphasis on the mechanisms underlying its timed activation and
cell cycle specific subcellular localisation. In the first part of this work, a systematic functional analysis of all GGDEF, EAL and
GGDEF/EAL composite proteins from C. crescentus with a focus on motility and surface
attachment is described. In this screen, the phosphodiesterase PdeA was identified as a
gatekeeper that prevents premature paralysis of the flagellum and holdfast synthesis in the
C. crescentus swarmer cell. It is shown that PdeA, together with its antagonistic DGCs DgcB
and PleD, are components of converging pathways and orchestrate polar development during
the swarmer-to-stalked cell transition. Furthermore, evidence is presented for a proteolytic
regulation mechanism for PdeA.
Secondly, the PleD localisation factor CC1064 is analysed. This transmembrane
protein has pleiotropic effects on motility, surface attachment and polar localisation of PleD.
It is shown that the motility and PleD localisation phenotypes of a Δcc1064 strain are
conditional and depend on environmental factors such as oxygen and temperature stress.
Moreover, evidence is presented that the impaired motility of a Δcc1064 mutant is caused by
an assembly defect of the motor proteins MotA and MotB, leading to paralysis of the
flagellum. A model is suggested that links altered membrane composition under
environmental stress conditions to the Δcc1064 phenotypes. In Paul, Abel et al. (2007), insights were gained into the regulation of PleD. In
addition to the well characterised non-competitive feedback inhibition, a second independent
layer of activity control via dimerisation was investigated. The response regulator PleD is
activated by phosphorylation of the N-terminal receiver domain. Here we show that the
phospho-mimetic chemical beryllium fluoride specifically activates the enzymatic activity of
PleD in vitro and in addition leads to dimerisation. Fractionation experiments showed that the
DGC activity exclusively resides within the dimer fraction. Finally, evidence is provided that
dimerisation of PleD is not only required for catalytic activity, but also leads to sequestration
to the differentiating stalked pole of the C. crescentus cell, thereby providing an elegant way
of restricting PleD activity to a subcellular region of the cell.
In Paul, Jaeger & Abel et al. (2008), a network of proteins belonging to the two
component system that regulates PleD activation and thereby leads to its localisation were
investigated in detail. The single domain response regulator DivK is controlled by the
phosphatase activity of PleC and the kinase DivJ. It is shown that DivK allosterically
activates the kinase activities of PleC and DivJ and thereby switches PleC from a phosphatase
into a kinase state. Increased DivJ activity further activates DivK in a feed-forward loop,
while PleC and DivJ together stimulate PleD activity and localisation. Evidence is provided
that DivJ, PleC, and DivK colocalise in a short time window during the cell cycle, directly
prior to PleD activation, suggesting a role for the spatial distribution of these proteins. At last,
the wider role of single domain response regulators in the interconnection of two-component
signal transduction circuits is discussed. Finally, in Dürig, Folcher, Abel et al. (2008), a role for c-di-GMP in the cell cycle of
C. crescentus via regulation of targeted proteolysis of the regulator CtrA is shown. During the
swarmer-to-stalked cell transition CtrA is recruited to the incipient stalked pole, where it is
degraded by its protease ClpXP. This recruitment and subsequent degradation is dependent on
the enzymatically inactive GGDEF domain protein PopA. PopA itself localises to the cell
pole and can bind c-di-GMP. It is shown that mutants in the c-di-GMP binding site fail to
localise to the developing stalked pole and consequently fail to promote CtrA degradation.
Finally, evidence is provided that interconnects PopA with the pathway responsible for
substrate inactivation and protease localisation in a cell cycle dependent manner.