Plasmid partition: sisters drifting apart.

Remarkably little is known about how bacterial chromosomes and their plasmids are partitioned to daughter cells prior to cell division, despite extensive experimentation and bioinformatic analysis. An in vitro study described in the current issue of The EMBO Journal recapitulates many features of plasmid partition in vivo and provides insight into how protein patterning on the nucleoid, determined by a diffusion-ratchet mechanism, directs plasmid positioning and partition. Related protein complexes likely use the same mechanism to position and partition newly replicated sister chromosomes and large protein complexes that act in processes other than chromosome segregation. The fundamentals of chromosome segregation in eukaryotes have been known for decades, with mitosis first being described cytologically at the end of the nineteenth century long before the chemical and genetic nature of chromosomes was known (reviewed in Nasmyth (2001)). A key feature of eukaryotic chromosome segregation is the use of a tubulinbased mitotic spindle to attach to the centromeres of sister chromosomes and to pull them apart during anaphase, a machinery clearly absent in bacteria. Ironically, genetic analyses aimed at revealing the mechanisms of bacterial chromosome, and plasmid segregation have revealed most insight into how plasmids partition, because mutations that lead to severe defects in plasmid partition are readily obtained. Although such studies have shown that some plasmids specify homologues of tubulin or actin to push apart newly replicated sister plasmids, most low copy plasmids encode ParA ATPases featuring a characteristic deviant Walker A motif to ensure their effective partition (reviewed in Howard and Gerdes, 2010; Reyes-Lamothe et al, 2012; Vecchiarelli et al, 2012). Plasmid-specified ParA ATPases act as part of a threecomponent system in which ParA-ATP interacts with DNA nonspecifically, whereas multiple dimers of the second protein, ParB, bind plasmid parS-sites. ParA–ATP–ParB– parS interactions tether plasmid partition complexes to the nucleoid matrix (Figure 1). Related systems can facilitate replication origin positioning as a consequence of adjacent parS sites and the consequent segregation of bacterial chromosomes; direct cell division away from the cell poles to midcell; and position large protein complexes regularly over the bacterial nucleoid, or at cell poles (reviewed in Vecchiarelli et al, 2012). Previous in vivo and in vitro experiments have led to conflicting views as to whether ParA ATPases function as cytoskeletal filaments whose retraction pulls sister plasmids and chromosomes apart, or whether their action occurs in the absence of filaments or extensive multimerization, with a diffusion-ratchet mechanism underlying their action (Ringgaard et al, 2009; Howard and Gerdes, 2010; Vecchiarelli et al, 2012). The work of Mizuuchi and colleagues presented in this issue (Hwang et al, 2013), along with a complementary study from the same laboratory (Vecchiarelli et al, 2013), provides convincing evidence for a diffusion-ratchet mechanism. The experiments employed a ‘carpet’ of sonicated nonspecific DNA molecules on the surface of a 25-mm-deep flow cell to mimic the surface of the nucleoid. Fluorescent proteins and DNA molecules were passed into the flowcell and two-colour total internal reflectance fluorescence microscopy (TIRFM) imaging was used to visualize ParAB-parS dynamics in the vicinity of curtain DNA. The results were qualitatively the same irrespective of whether ParAB-parS encoded by plasmid P1, or the related F plasmid machinery (SopAB-sopC, respectively) was used. Binding and turnover experiments at physiologically relevant concentrations (o1% saturation of carpet DNA, conditions that seem unlikely to favour filament formation), using fluorescence recovery after photobleaching (FRAP), showed that ParA-ATP binds DNA nonspecifically as dimers or small oligomers after a slow conformational change that occurs on ATP binding, and dissociates from DNA slowly in an ATP hydrolysisindependent manner. Reaction of appropriate concentrations of ParA, ParB, parS-containing plasmid and ATP led to the formation of DNA carpet-bound plasmid partition complexes with 5–20 plasmid copies associated with 500–1500 ParA monomers and 2-3-fold fewer dimers of ParB. Stimulation of ParA-ATP hydrolysis by the high local concentration of parS-bound ParB led to release of ParA and the associated plasmid partition complexes from the carpet. In order to facilitate observation and analysis, complexes containing many more plasmid molecules were formed by preincubating plasmid with ParAB. These larger complexes behaved similarly to the smaller complexes, but now partition complex ‘jumping’ and repositioning could be observed directly. The ParA concentration associated with plasmid clusters determined the extent of anchoring of the complexes to the carpet; tightly anchored complexes had many ParA-ParB-parS tethers, The EMBO Journal (2013) 32, 1208–1210 www.embojournal.org