ParM filaments (green) drive plasmid segregation.


Polymerization of an actin-related protein may drive segregation of a bacterial plasmid, according to Jakob Møller-Jensen, Kenn Gerdes (University of Southern Denmark, Odense, Denmark), and colleagues.

Segregation in bacteria was thought to be a rather passive affair, with DNA attaching to the membrane and getting dragged along as cell growth led to membrane expansion. That model was challenged when DNA replication was found to be localized to one area at the cell center, followed by rapid movement of both plasmids and chromosomal DNA from midcell to more polar sites.

In the meantime, Gerdes had been working on the segregation requirements for the R1 plasmid. He had found that ParR binds to the parC centromere-like site on the plasmid, and that ParM binds to ParR. Then, Møller-Jensen tried detecting native ParM directly, using immunofluorescence, and saw axial filaments in almost half the cells. “We were very excited,” he says, “because this explained a lot of open questions to us.”

Filament formation in vivo (and, at lower ParM concentrations, in vitro) required both parC and ParR, suggesting that a parC–ParR complex nucleates the formation of a filament that then drives separation of replicated plasmids. The active nucleator may be dimers of parC–ParR, which form only after replication thanks to the cis-restricted activity of ParR.

ParM filaments can go through multiple cycles of polymerization (which requires ATP) and depolymerization (which requires hydrolysis of the bound ATP). After the majority of the cell's ParM has polymerized into filaments, a gradual conversion into the ADP form may trigger the fragmentation that the authors saw both in vivo and in vitro.

This very dynamic behavior of ParM makes it very different from MreB, a similar filament-forming bacterial protein that helps maintain cell shape. But the two may not be so different after all. “Being a plasmid molecule, [ParM] may have to make it on its own,” says Møller-Jensen. But MreB, he says, may coopt other proteins to help make it the bacterial equivalent of a mitotic spindle. ▪


Møller-Jensen, J., et al. 2002. EMBO J. 21:3119–3127.