hence to inhibit the synthesis of new plasmid copies. Antisense RNAs also play important roles in modulating plasmid replication through their ability to interfere with the expression of plasmid‐encoded replication factors (Brantl 2014).
RK2 and its plasmid relatives use a theta model of DNA replication (Figure 1.12). Plasmid ColE1, the backbone for many cloning vectors used in recombinant DNA technology, also uses theta replication, but differs from RK2‐like plasmids in relying on host factors to open the double‐stranded origin and to prime synthesis (Lilly and Camps 2015; Wang et al. 2004). DNA duplex unwinding is driven by transcription of a stable RNA pre‐primer that forms an R‐loop in the ori region of ColE1. This process is driven by negative supercoiling of the plasmid DNA. RNase H then processes the bound RNA to generate the primer RNA that is then extended by Pol I. This marks the beginning of leading strand DNA synthesis. As the newly synthesised DNA strand makes progress through the plasmid DNA duplex, it base pairs with the template to create a D‐loop that recruits PriA. Pol III takes over leading strand synthesis and initiates the synthesis of the lagging strand; the converging replisomes continue moving until they are at or near the termination site terH (Nakasu and Tomizawa 1992). Gaps between the strands are then filled in by Pol I (Troll et al. 2014).
Rolling circle replication allows plasmids to replicate independently of chromosomal DNA (Khan 2005). The process relies on a nick made by a plasmid‐encoded initiator protein in one plasmid DNA strand, providing a primer for leading strand initiation and a lagging strand origin (Figure 1.13). No RNA primer is required. Rolling circle replication is chiefly found in plasmids from Gram‐positive bacteria, although it does occur in replicons from Gram‐negatives and archaea (del Solar et al. 1993; Ruiz‐Masó et al. 2015). The process replicates the leading strand and the lagging strand in two separate steps. Leading strand replication begins with the nicking of the double‐strand origin (dso) by a plasmid‐encoded replication protein, Rep. This is a member of the HuH superfamily of DNA endonucleases (Chandler et al. 2013) and it has a binding site located adjacent to the dso that positions it appropriately to cut the DNA. The DNA to be cleaved is presented to Rep in a single‐stranded form within a stem‐loop structure that extrudes from the negatively supercoiled plasmid. This extrusion event is Rep‐binding‐dependent (Ruiz‐Masó et al. 2007). Rep forms a covalent bond with the cleaved DNA through an active site tyrosine (Noirot‐Gros et al. 1994; Thomas et al. 1990). Host DNA polymerases use the intact template strand to guide DNA synthesis while simultaneously displacing the non‐template strand. The displaced strand is coated with SSB and is ejected as a covalently closed, single‐stranded circle at the end of leading strand synthesis. This single‐stranded circle is then used as the template for lagging strand synthesis, a process that involves only host‐encoded proteins (especially RNA polymerase and DNA polymerase I) and initiates at a structured region in the circle known as the single‐strand origin, sso (del Solar et al. 1987; Gruss et al. 1987; Kramer et al. 1997). Control of rolling circle replication is achieved principally through the control of Rep protein production. For this reason, the expression of the rep gene is strictly regulated, typically via mechanisms that employ an antisense RNA or an antisense RNA working with a DNA‐binding regulatory protein. The first type operates through transcriptional attenuation while the second involves protein‐mediated transcriptional repression backed up by translational inhibition using a trans‐acting RNA (Brantl 2014; del Solar and Espinosa 2000; Novick et al. 1989).
Figure 1.13 Rolling circle plasmid replication. A circular plasmid using the rolling circle mechanism of replication is shown at top left. The double‐stranded circle is shown in a topologically relaxed state, but it would be negatively supercoiled in the bacterium, a state that encourages extrusion of a cruciform that contains the double‐strand origin (dso), the site of replication initiation. The dso is represented by a slightly thicker line in the drawing. Extrusion of the cruciform presents the dso in single‐stranded form to the plasmid‐encoded replication protein, Rep. The Rep protein is positioned appropriately by binding to a recognition site on the plasmid adjacent to the dso. The bacterial DNA polymerase use the 3′‐OH at the nick to prime DNA synthesis; no RNA primer is required. A dotted line represents the newly synthesised DNA and an arrow next to this line shows the direction of DNA synthesis. The plasmid duplex unwinds as DNA synthesis proceeds, displacing the non‐template DNA strand, which is then coated by the single‐stranded DNA‐binding protein, SSB. A full round of replication displaces the non‐template strand completely, producing a double‐stranded plasmid (with one newly synthesised strand) and a single‐stranded circle. This circle is used as the template for the synthesis of the lagging strand. Host proteins exclusively conduct lagging strand synthesis (especially RNA polymerase and DNA polymerase I), a process that begins with priming by RNA polymerase via RNA synthesis at the single‐strand origin, sso.
1.15 Plasmid Segregation
Plasmids employ two strategies to ensure segregation of their copies at cell division: active partitioning mechanisms (low copy number plasmids) and reliance on dispersal through the cytosol of the mother cell to ensure that some copies end up in each daughter cell (high copy number plasmids). A third strategy acts post‐segregationally. It is based on toxin–antitoxin systems and eliminates those bacterial cells that do not acquire a plasmid copy at cell division (Hayes 2003) (see also Sections 2.30 and 2.35). The plasmid encodes both a stable toxin and an unstable antitoxin in an operon known as an addiction module: maintenance of the supply of the antitoxin requires the continued presence of the plasmid and the antitoxin‐encoding gene. Bacteria that become liberated from the burden of plasmid‐carriage may outgrow their plasmid‐carrying counterparts. Eliminating the plasmidless bacteria helps to prevent extinction of the plasmid carriers by the fitter, plasmid‐free, segregants. The CcdA/B antitoxin/toxin pair produced by the F plasmid provides an example of this post‐segregational killing strategy. CcdB inhibits DNA gyrase by trapping it in a cleavage complex with DNA, and CcdB activity is neutralised when it binds the unstable CcdA antidote. A bacterium that loses the F plasmid will retain the toxic CcdB molecule after the unstable CcdA molecule has been broken down by the ATP‐dependent Lon protease (van Melderen et al. 1994). The resulting poisoning of DNA gyrase by CcdB kills the plasmidless cell.
High copy number plasmids, i.e. those with 10 or more copies per cell, lack genes that are capable of encoding active partitioning machinery (Million‐Weaver and Camps 2014). Random distribution of the multicopy plasmid through the cytosol seems to account for the faithful inheritance of plasmids such as ColE1 (Durkacz and Sherratt 1983). The formation of plasmid multimers poses a risk because this process reduces the copy number of independently segregating units, but multimer resolution systems such as cer/XerCD in ColE1 provide a potent antidote (Summers and Sherratt 1984). This resolution mechanism is a close relative of the chromosomal dif/XerCD system, albeit with additional co‐factors (Section 1.8). Plasmid distribution in the cytosol is likely to be influenced by the presence of other molecules and structures, not least the nucleoid, and nucleoid exclusion does seem to be a factor in confining plasmids to the space just inside the cytoplasmic membrane (Reyes‐Lamothe et al. 2014; Wang et al. 2016; Yao et al. 2007). The plasmids occur in clusters and frequently these clusters are seen at the poles of the cell; clusters are dynamic, they can divide, with some sub‐clusters relocating to the mid‐cell (Yao et al. 2007). The introduction of another type of plasmid produces an even more complex clustering pattern (Diaz et al. 2015; Yao et al. 2007). It appears that multicopy plasmids move between existing in clusters and being alone, and that these forms diffuse randomly within the confines imposed by the nucleoid and other cell structures (Wang 2017).
Low‐copy number plasmids cannot rely on strategies based on random spatial
