Multicopy plasmid stability: revisiting the dimer catastrophe

In this study, we have constructed a stochastic simulation of the replication and distribution of the bacterial multicopy plasmid ColE1 in a population of exponentially growing cells. It is assumed that ColE1 is randomly distributed between daughter cells at division such that copy number is a critical determinant of plasmid loss. High copy number is threatened by plasmid dimers, which arises initially by homologous recombination and accumulate by replication in a process known as the 'dimer catastrophe'. Summers et al. (1993) modelled this process and demonstrated that the accumulation of dimers is limited by the metabolic load that they exert on their hosts. ColE1 also encodes the cer site, at which host-encoded proteins act to convert dimers to monomers by site-specific recombination. The cer site also encodes a regulatory RNA, Rcd, whose synthesis from plasmid dimers triggers a checkpoint that delays cell division, presumably allowing sufficient time for dimer resolution. Here we have developed the original dimer catastrophe model by incorporating copy number variance with a stochastic model of plasmid replication. We demonstrate that the Rcd checkpoint is necessary when the rate of dimer resolution is slow. Our results indicate that dimers over-replicate compared to monomers, suggesting a mechanism for their increased metabolic load. We find that the effect of dimers on plasmid stability is significantly less severe than suggested by the original model. Consequently, we propose that the primary role of dimer resolution and the Rcd checkpoint is to reduce the metabolic burden imposed by the plasmid in a recombinogenic host, rather than to ensure plasmid stability.