RpoS, an RNA polymerase s factor, controls the response of Escherichia coli and related bacteria to multiple stress responses. During nonstress conditions, RpoS is rapidly degraded by ClpXP, mediated by the adaptor protein RssB, a member of the response regulator family. In response to stress, RpoS degradation ceases. Small anti-adaptor proteins-IraP, IraM, and IraD, each made under a different stress condition-block RpoS degradation. RssB mutants resistant to either IraP or IraM were isolated and analyzed in vivo and in vitro. Each of the anti-adaptors is unique in its interaction with RssB and sensitivity to RssB mutants. One class of mutants defined an RssB N-terminal region close to the phosphorylation site and critical for interaction with IraP but unnecessary for IraM and IraD function. A second class, in the RssB C-terminal PP2C-like domain, led to activation of RssB function. These mutants allowed the response regulator to act in the absence of phosphorylation but did not abolish interaction with anti-adaptors. This class of mutants is broadly resistant to the anti-adaptors and bears similarity to constitutively activated mutants found in a very different PP2C protein. The mutants provide insight into how the anti-adaptors perturb RssB response regulator function and activation.

The genome of the soil bacterium Pseudomonas putida KT2440 encodes singular orthologues of genes crp (encoding the catabolite repression protein, Crp) and cyaA (adenylate cyclase) of Escherichia coli. The levels of cAMP formed by P. putida cells were below detection with a Dictyostelium biosensor in vivo. The cyaA(P. putida) gene was transcribed in vivo but failed to complement the lack of maltose consumption of a cyaA mutant of E. coli, thereby indicating that cyaA(P. putida) was poorly translated or rendered non-functional in the heterologous host. Yet, generation of cAMP by CyaA(P. putida) could be verified by expressing the cyaA(P. putida) gene in a hypersensitive E. coli strain. On the other hand, the crp(P. putida) gene restored the metabolic capacities of an equivalent crp mutant of E. coli, but not in a double crp/cyaA strain, suggesting that the ability to regulate such functions required cAMP. In order to clarify the breadth of the Crp/cAMP system in P. putida, crp and cyaA mutants were generated and passed through a battery of phenotypic tests for recognition of gross metabolic properties and stress-endurance abilities. These assays revealed that the loss of each gene led in most (but not all) cases to the same phenotypic behaviour, indicating a concerted functionality. Unexpectedly, none of the mutations affected the panel of carbon compounds that can be used by P. putida as growth substrates, the mutants being impaired only in the use of various dipeptides as N sources. Furthermore, the lack of crp or cyaA had little influence on the gross growth fingerprinting of the cells. The poor physiological profile of the Crp-cAMP system of P. putida when compared with E. coli exposes a case of regulatory exaptation, i.e. the process through which a property evolved for a particular function is co-opted for a new use.

In exponential phase, the stationary phase sigma factor σS is maintained at low levels by proteolysis; RssB binds directly to σS and targets it to the ClpXP protease. Under stress conditions this process is inhibited, in part by sequestration of RssB by anti‐adaptor proteins. IraP is an anti‐adaptor made after phosphate starvation. We screened for RssB mutants resistant to IraP. The mutants identify a critical and conserved region between amino acids 214–221 in the C‐terminal domain of RssB. In vivo, cells carrying plasmids encoding RssBL214H or RssBA216T were able to degrade σS even in the presence of high levels of IraP. In a bacterial two‐hybrid assay, RssBL214H and RssBA216AT do not interact with IraP, while RssB+ does. In vitro, purified RssB mutant proteins were fully functional for RpoS degradation; RssBwt but not RssBL214H and RssBA216T were inhibited by IraP. Pull‐down analysis confirmed the loss of interaction of RssBL214H and RssBA216T with IraP. Thus, RssBL214H and RssBA216T have lost the interaction with IraP. Because other data suggests that IraP interacts with the N‐terminal domain of RssB, and these mutations are in the C‐terminal domain, we believe these mutations lead to an allosteric change in RssB, blocking the IraP interaction site. This allosteric change may normally be part of the activation cycle allowing RssB to deliver σS to the protease.