Abstract
Extracytoplasmic formation of disulfide bonds (DSB) is often required for proper folding of proteins, many of which are important for the assembly of virulence factors such as the Dot/Icm type IVb secretion system (T4SS) of Legionella pneumophila (Lpn), the major virulence system that delivers multiple effector proteins into the cytoplasm of host cells. In Escherichia coli, DsbA catalyzes the formation of disulfide bonds, while DsbC repairs inappropriate disulfides through protein disulfide isomerase (PDI) activity. DsbA is a monomer that is maintained in a fully oxidized state by the cytoplasmic membrane protein DsbB, whereas DsbC is a homodimer and kept reduced by a second cytoplasmic membrane protein, DsbD.
In contrast, Lpn expresses two DsbA-like proteins (DsbA1 and DsbA2), but no orthologue of DsbC. Both DsbA proteins are capable of catalyzing DSB formation; however, dsbA1 is dispensable for viability and virulence, while dsbA2 is essential and required for disulfide bond formation in core proteins of the Dot/Icm type IVb secretion system. DsbA2 is a homodimer that is phylogenetically distinct from DsbA and DsbC lineages. We hypothesized that in a dsbA1 mutant of Lpn, DsbA2 must be responsible for both oxidase and PDI activities; i.e., a single player system as compared to the two player DsbA/DsbC system of E. coli.
In this study we confirmed that DsbA2 possesses PDI activity using a PDI detector assay in which DsbA2 replaces DsbC in E. coli. Consistent with a single player bifunctional system, we showed that DsbA2 exists in the periplasm of Lpn as a mixture of disulfides (S-S) and free thiols (SH). This equilibrium can only occur if the two DsbB oxidases (DsbB1 and DsbB2) cooperate with the two DsbD reductases (DsbD1 and DsbD2). To test this hypothesis, we reconstituted the Lpn Dsb system in various dsb mutants of E. coli, including a dsb null mutant strain. Our studies showed that DsbA2 did not restore motility to a dsbA mutant of E. coli, whereas a monomeric dimerization domain mutant (DsbA2∆N) restored motility. We showed that both DsbBs of Lpn were able to oxidize DsbA2ΔN, but not DsbA2 and that the E. coli DsbD reduced DsbA2 to the thiol. Motility was restored to a dsb null mutant of E. coli expressing Lpn dsbA2, dsbB1 or dsbB2 and dsbD1 or dsbD2 and this was possible because the reconstituted system maintained DsbA2 as a mixture of S-S/SH forms. This specificity of Lpn DsbB and DsbD proteins for DsbA2 bifunctional activity suggests a distinct evolutionary tract for the DsbA2 system.
In addition, we identified biological roles for nonessential DsbA1 in Lpn, by showing that a dsbA1 mutant had lost the ability to produce melanin-like pigment and was more infectious of HeLa cells. This mutant was also more virulent by the contact hemolysis assay, and all of these phenotypes were restored to WT through complementation. Finally, overexpression of DsbA1P150T mutant protein in Lpn did not result in a dominant negative effect on DsbA2 function as was seen with either the DsbA2P198T or DsbA2ΔNP198T mutant proteins. Taken together, these studies suggest that DsbA1 might compete for substrate proteins with DsbA2 and perhaps modulate virulence.
These findings indicate that DsbA2 is a bifunctional oxidoreductase in Lpn, which is maintained by the DsbB and DsbD proteins that cooperate in maintaining this single player system. In addition, this protein, which is found in many other pathogens expressing type IV secretion systems, likely provides a selective advantage by economizing disulfide bond management in these slow growing bacteria.
