P1 plasmid partition: a mutational analysis of ParB.

P1 plasmid partition to daughter cells requires plasmid proteins, ParA and ParB, and a centromere analog, parS, to which ParB binds. ParA and ParB decrease the loss frequencies of some low copy number plasmid vectors with parS inserts, as well as that of P1 itself, more than 100-fold. Unexpectedly, we find that Par proteins can actively destabilize other parS-plasmid constructs under similar conditions. Only ParB is required for this partition dysfunction. The destabilization can be dramatic; certain parS-plasmid constructs cannot even be maintained in the presence of ParB. We take advantage of this partition dysfunction to select parB mutants unable to destabilize a particularly ParB-sensitive parS-plasmid. The mutations obtained are widely distributed in the gene. Several of the mutations affected binding of ParB to parS; others did not. Amino acid substitutions that affected binding without influencing dimerization were mapped to two separate regions of ParB. C-Terminal amino acid substitutions in ParB blocked its dimerization and binding to parS, as did deletions of the entire C-terminal part of the protein. These results are consistent with ParB binding to parS as a dimer and suggest that the C-terminal part is a dimerization domain of the protein. Cell extracts of a frameshift mutant analyzed contained, in addition to the expected truncated ParB, a second species of ParB with the normal C-terminal part. Apparently, a site of ribosomal frameshifting is encoded within parB, at which some ribosomes change reading frame in the mutant. All obtained mutant ParB proteins were unable to complement the partition defect of a parB null mutant mini-P1. This suggests that the initial reactions between ParB and parS leading to partition dysfunction and to partition function are similar.