The methylation-independent mismatch repair machinery in Pseudomonas aeruginosa

Abstract

Over the last 70 years, we've all gotten used to an Escherichia coli-centric view of the microbial world. However, genomics, as well as the development of improved tools for genetic manipulation in other species, is showing us that other bugs do things differently, and that we cannot simply extrapolate from E. coli to everything else. A particularly good example of this is encountered when considering the mechanism(s) involved in DNA mismatch repair by the opportunistic human pathogen, Pseudomonas aeruginosa (PA). This is a particularly relevant phenotype to examine in PA, since defects in the mismatch repair (MMR) machinery often give rise to the property of hypermutability. This, in turn, is linked with the vertical acquisition of important pathoadaptive traits in the organism, such as antimicrobial resistance. But it turns out that PA lacks some key genes associated with MMR in E. coli, and a closer inspection of what is known (or can be inferred) about the MMR enzymology reveals profound differences compared with other, well-characterized organisms. Here, we review these differences and comment on their biological implications.

Keywords: MutL; MutS; Pseudomonas aeruginosa; hypermutation; mismatch repair.

Fig. 1.

A comparison showing the canonical E. coli MMR pathway and a proposal for non-methyl-directed MMR in P. aeruginosa . In E. coli (left panel) the nascent strand is hypomethylated relative to the template strand (a). MMR begins with ATP-bound dimeric MutS binding to the site of a mismatch (b). This may be via carriage on the β subunit (‘sliding clamp’) of the replisome, or simply through continual bidirectional ‘scanning’ of the genome. Upon recognition of a mismatch, MutS changes its conformation and recruits MutL in an ATP-dependent reaction (c). This, in turn, subsequently leads to recruitment of MutH, which nicks the nascent DNA strand opposite a nearby Dam-methylated adenine base. The MutH endonuclease activity dependent on ATP hydrolysis by the MutS-MutL-MutH complex, and likely involves bending of the DNA if ‘action at a distance’ is required (d, e). Following nicking of the hypomethylated strand, the MutS-MutL complex recruits the UvrD helicase, which unwinds the duplex in the direction of the mismatch (f). The exposed single-stranded hypomethylated DNA, protected by SSB, is then digested by one of the exonucleases present, to a point beyond the original mismatch. The extent of retrograde (3′ → 5′) digestion is presumably limited by the processivity of the exonuclease (g). The resulting gap is then filled in and sealed through the combined action of DNA polymerase and DNA ligase (h, i). Some of the reactions in PA (right panel) are superficially similar, although many of the details are yet to be elucidated, so the presented model is inevitably a simplification. The key difference between PA and E. coli is that neither DNA strand is significantly methylated in the former (a). Current evidence suggests that MutS tetramers recognize DNA mismatches during DNA replication and are delivered to these sites via the β-sliding clamp, although post-replicative scanning surveillance also seems likely (b). DNA binding by MutS is accompanied by ATP hydrolysis (and concomitant release of the β-clamp) and is followed by recruitment of ATP-inhibited MutL. Again, this delivery of MutL to the MutS-bound mismatch may be via the β-clamp during replication, or via the β-clamp post-replicatively, or independent of the β-clamp (c). In the ATP-bound form, the CTD-associated endonuclease activity of MutL is inhibited. However, upon binding MutS, the ATPase activity of the NTD of MutL becomes enhanced, thereby relieving the ATP-dependent inhibition of the CTD-endonuclease. This leads to the generation of a nick (d). Quite how far away from the original mismatch this nick is, or whether the process involves DNA bending (as seems likely and as inferred for E. coli MMR) is not yet clear. Nor is it clear how strand discrimination is maintained. [We note that the model proposed here implies that MutS physically moves away from the mismatch as a complex with MutL before the nick is made. However, it is equally likely that MutS remains bound to the mismatch and that the interaction between dynamic β-bound MutL and ‘static’ MutS is achieved through DNA bending.] Following nicking of the DNA on the nascent strand, UvrD is recruited (e) and unwinds the DNA towards the mismatch. Again, how this directionality is ensured is not clear. Subsequent exonucleolytic degradation of the nascent strand and ‘fill-in/polishing’ are presumably the same as in E. coli .