[~EVIEWS
Many prokaryotic cells contain plasmids 1, nonessential fragments of DNA that replicate extrachromosomally and are stably inherited. They differ widely in size (from a few thousand to several hundred thousand base pairs), copy number (from one to several hundred) and in the mechanism that triggers replication initiation. Although some plasmids confer on the host bacterium phenotypes that may be advantageous in natural or laboratory conditions, most of them are cryptic and constitute a burden for the host metabolic machinery. Nevertheless, they are stably maintained even in the absence of apparent selection and only rarely are cells without plasmids segregated from a population of plasmid-containing bacteria. This high degree of stability implies that low copy number and oligocopy plasmids have efficient partitioning mechanisms that ensure that both daughter cells acquire at least one plasmid copy at cell division. By contrast, partitioning of multicopy plasmids is essentially random and stability relies on the maintenance of a relatively high copy number. Since the stochastic process of plasmid segregation can occasionally lead to cells with fewer plasmids than normal, the plasmid must be able to sense this drop in concentration and to over-replicate to correct it 2. This flexibility is achieved by mechanisms that permit violation of the normal 'one doubling per cell cycle' rule that governs replication of prokaryotic (and eukaryotic) chromosomes. However, multiple initiation is not the only possible strategy to counteract occasional variations in plasmid copy number. Among eukaryotic plasmids, for instance, the 21a plasmid of Saccharomyces cerevisiae uses an elegant strategy, based on sitespecific recombination, that permits the amplification of copy number, although each molecule can initiate replication only once in the cell cycle3. Down-regulation of plasmid copy number is also required to avoid the deleterious effect of high plasmid copy number on cell growth. To achieve this, plasmids have evolved efficient control mechanisms that prevent plasmid copy number increasing above a certain threshold, which is determined by the affinity of a plasmid-encoded inhibitor for its target.
Replication initiation in the plasmid ColE1 Currently, the best-understood copy number control system is that of the ColE1 plasmid family of E. coli (reviewed in Refs 2, 4, 5). Combined genetical, biochemical and structural approaches have allowed the tbrmulation of a model for initiation of plasmid replication and its control that is approaching atomic resolution, thanks primarily to the work of Tomizawa's group. ColE1 replication is unidirectional and relies on host proteins - the plasmid does not encode any enzymatic functions required for replication. A 600 nucleotide DNA fragment containing the replication origin is necessary in cis tor initiation of replication. This fragment also contains two promoters (see upper part of Fig. 3). One promoter is responsible for initiation of the replication primer RNA (called RNA II). Tran~s'ription initiating at the RNA II promoter proceeds beyond the origin and terminates at different positions depending on the
Control of C01E1plasmid replication by antisense RNA G. CESARENI~M. HELMER-CITrERICH AND L. CASTAGNOLI One of the two major classes of regulatory strategies that control plasmid copy number involves recognition via base pairing between two plasmid-encoded complementary RNAs. The detailed analysis of this control circuitry has revealed some features of regulatory mechanisms based on RNA-RNA interaction that distinguish them from those based on protein-nacleic acid interaction, These features provide a framework with which to understand other regulatory mechanisms based on RNA-RNA interaction, and will aid in the design of efficient artificial antisense RNA systems. downstream sequence of the specific plasmid. The RNA II transcript is processed by RNaseH to yield a 550 nucleotide molecule that is efficiently used as a primer by DNA polymerase I to initiate leading strand synthesis(' (mode I in Fig. 1). In the absence of RNaseH, the 3' end of the unprocessed RNA II can still be used as a primer by DNA polymerase I, albeit at lower efficiency (mode III). RNA II transcription can also trigger initiation of DNA replication in cells defective in both RNaseH and DNA polymerase I, by displacing the nontranscribed strand, on which lagging strand DNA synthesis (by DNA polymerase lid can then occurv (mode 1I). Thus, plasmid replication initiation is a versatile process, capable of adjusting to different physiological milieux.
RNA H forms a DNA-RNA hybrid at the origin Each of these three distinctive mechanisms relies on a peculiar property of the nascent RNA II transcript its ability to form a persistent hybrid with the DNA template strand 7 near the replication origin. This property is not shared with other transcripts. It is generally believed that the separation of the nascent transcript from the template is a general property of transcription, probably associated with some activity of the RNA polymerase itself. How does RNA II resist removal from the template? Analysis of replication-defective mutants and the selection of second-site revertants led Masukata and Tomizawa ~ to propose that the formation of the persistent hybrid requires the specific interaction between a G-rich loop in a region of the RNA II transcript 265 nucleotides upstream of the origin and a Crich stretch on the DNA template strand 20 nucleotides upstream of the origin (Fig. 2). Interestingly both sequences are conserved in the related plasmids pMB1, pl5A and RSF1030. The case is strengthened by the fact that in the related plasmid CIoDF13, a C to T transition in the -20 region is compensated by an equivalent mutation in the -265 G-rich stretch. The interaction between -265 RNA and -20 DNA probably occurs while the transcription bubble is in
rl¢; .lH.v 1 9 9 1 VOL, 7 NO. 7 /
c lnand ~t> l The structure of the two inhibitors of plasmid replication, RNA 1 and Rop, is also sdlcmaticallv shown. The lower part of Hlc tigurc shov,-s the two functionally important alternative conformations that influence the ability of RNA II to form the l)c>islc,'~t RN \-l)X \ hybrid at the origin. critical in determining the effectiveness of the regulation mechanism. The detailed description of the molecular events that lead to inhibition of primer formation is complicated by the predicted continuous reshuffling of interacting sequences that characterizes the process of RNA 11 folding during elongation by RNA polymerase. The interaction between RNA I and RNA II represents a type of dynamic control circuit in v
Many prokaryotic cells contain plasmids 1, nonessential fragments of DNA that replicate extrachromosomally and are stably inherited. They differ widely in size (from a few thousand to several hundred thousand base pairs), copy number (from one to several hundred) and in the mechanism that triggers replication initiation. Although some plasmids confer on the host bacterium phenotypes that may be advantageous in natural or laboratory conditions, most of them are cryptic and constitute a burden for the host metabolic machinery. Nevertheless, they are stably maintained even in the absence of apparent selection and only rarely are cells without plasmids segregated from a population of plasmid-containing bacteria. This high degree of stability implies that low copy number and oligocopy plasmids have efficient partitioning mechanisms that ensure that both daughter cells acquire at least one plasmid copy at cell division. By contrast, partitioning of multicopy plasmids is essentially random and stability relies on the maintenance of a relatively high copy number. Since the stochastic process of plasmid segregation can occasionally lead to cells with fewer plasmids than normal, the plasmid must be able to sense this drop in concentration and to over-replicate to correct it 2. This flexibility is achieved by mechanisms that permit violation of the normal 'one doubling per cell cycle' rule that governs replication of prokaryotic (and eukaryotic) chromosomes. However, multiple initiation is not the only possible strategy to counteract occasional variations in plasmid copy number. Among eukaryotic plasmids, for instance, the 21a plasmid of Saccharomyces cerevisiae uses an elegant strategy, based on sitespecific recombination, that permits the amplification of copy number, although each molecule can initiate replication only once in the cell cycle3. Down-regulation of plasmid copy number is also required to avoid the deleterious effect of high plasmid copy number on cell growth. To achieve this, plasmids have evolved efficient control mechanisms that prevent plasmid copy number increasing above a certain threshold, which is determined by the affinity of a plasmid-encoded inhibitor for its target.
Replication initiation in the plasmid ColE1 Currently, the best-understood copy number control system is that of the ColE1 plasmid family of E. coli (reviewed in Refs 2, 4, 5). Combined genetical, biochemical and structural approaches have allowed the tbrmulation of a model for initiation of plasmid replication and its control that is approaching atomic resolution, thanks primarily to the work of Tomizawa's group. ColE1 replication is unidirectional and relies on host proteins - the plasmid does not encode any enzymatic functions required for replication. A 600 nucleotide DNA fragment containing the replication origin is necessary in cis tor initiation of replication. This fragment also contains two promoters (see upper part of Fig. 3). One promoter is responsible for initiation of the replication primer RNA (called RNA II). Tran~s'ription initiating at the RNA II promoter proceeds beyond the origin and terminates at different positions depending on the
Control of C01E1plasmid replication by antisense RNA G. CESARENI~M. HELMER-CITrERICH AND L. CASTAGNOLI One of the two major classes of regulatory strategies that control plasmid copy number involves recognition via base pairing between two plasmid-encoded complementary RNAs. The detailed analysis of this control circuitry has revealed some features of regulatory mechanisms based on RNA-RNA interaction that distinguish them from those based on protein-nacleic acid interaction, These features provide a framework with which to understand other regulatory mechanisms based on RNA-RNA interaction, and will aid in the design of efficient artificial antisense RNA systems. downstream sequence of the specific plasmid. The RNA II transcript is processed by RNaseH to yield a 550 nucleotide molecule that is efficiently used as a primer by DNA polymerase I to initiate leading strand synthesis(' (mode I in Fig. 1). In the absence of RNaseH, the 3' end of the unprocessed RNA II can still be used as a primer by DNA polymerase I, albeit at lower efficiency (mode III). RNA II transcription can also trigger initiation of DNA replication in cells defective in both RNaseH and DNA polymerase I, by displacing the nontranscribed strand, on which lagging strand DNA synthesis (by DNA polymerase lid can then occurv (mode 1I). Thus, plasmid replication initiation is a versatile process, capable of adjusting to different physiological milieux.
RNA H forms a DNA-RNA hybrid at the origin Each of these three distinctive mechanisms relies on a peculiar property of the nascent RNA II transcript its ability to form a persistent hybrid with the DNA template strand 7 near the replication origin. This property is not shared with other transcripts. It is generally believed that the separation of the nascent transcript from the template is a general property of transcription, probably associated with some activity of the RNA polymerase itself. How does RNA II resist removal from the template? Analysis of replication-defective mutants and the selection of second-site revertants led Masukata and Tomizawa ~ to propose that the formation of the persistent hybrid requires the specific interaction between a G-rich loop in a region of the RNA II transcript 265 nucleotides upstream of the origin and a Crich stretch on the DNA template strand 20 nucleotides upstream of the origin (Fig. 2). Interestingly both sequences are conserved in the related plasmids pMB1, pl5A and RSF1030. The case is strengthened by the fact that in the related plasmid CIoDF13, a C to T transition in the -20 region is compensated by an equivalent mutation in the -265 G-rich stretch. The interaction between -265 RNA and -20 DNA probably occurs while the transcription bubble is in
rl¢; .lH.v 1 9 9 1 VOL, 7 NO. 7 /
c lnand ~t> l The structure of the two inhibitors of plasmid replication, RNA 1 and Rop, is also sdlcmaticallv shown. The lower part of Hlc tigurc shov,-s the two functionally important alternative conformations that influence the ability of RNA II to form the l)c>islc,'~t RN \-l)X \ hybrid at the origin. critical in determining the effectiveness of the regulation mechanism. The detailed description of the molecular events that lead to inhibition of primer formation is complicated by the predicted continuous reshuffling of interacting sequences that characterizes the process of RNA 11 folding during elongation by RNA polymerase. The interaction between RNA I and RNA II represents a type of dynamic control circuit in v
