Antisense RNA Control of Plasmid R1 Replication

Abstract

The replication frequency of plasmid R1 is controlled by an antisense RNA (CopA) that binds to its target site (CopT) in the leader region of repA mRNA and inhibits the synthesis of the replication initiator protein RepA. Previous studies on CopA-CopT pairing in vitro revealed the existence of a primary loop-loop interaction (kissing complex) that is subsequently converted to an almost irreversible duplex. However, the structure of more stable binding intermediates that lead to the formation of a complete duplex was speculative. Here, we investigated the interaction between CopA and CopT by using Pb(II)-induced cleavages. The kissing complex was studied using a truncated antisense RNA (CopI) that is unable to form a full duplex with CopT. Furthermore, RNase III, which is known to process the CopA-CopT complex in vivo, was used to detect the existence of a full duplex. Our data indicate that the formation of a full CopA-CopT duplex appears to be a very slow process in vitro. Unexpectedly, we found that the loop-loop interaction persists in the predominant CopA-CopT complex and is stabilized by intermolecular base pairing involving the 5′-proximal 30 nucleotides of CopA and the complementary region of CopT. This almost irreversible complex suffices to inhibit ribosome binding at the tap ribosome binding site and may be the inhibitory complex in vivo. The replication frequency of plasmid R1 is controlled by an antisense RNA (CopA) that binds to its target site (CopT) in the leader region of repA mRNA and inhibits the synthesis of the replication initiator protein RepA. Previous studies on CopA-CopT pairing in vitro revealed the existence of a primary loop-loop interaction (kissing complex) that is subsequently converted to an almost irreversible duplex. However, the structure of more stable binding intermediates that lead to the formation of a complete duplex was speculative. Here, we investigated the interaction between CopA and CopT by using Pb(II)-induced cleavages. The kissing complex was studied using a truncated antisense RNA (CopI) that is unable to form a full duplex with CopT. Furthermore, RNase III, which is known to process the CopA-CopT complex in vivo, was used to detect the existence of a full duplex. Our data indicate that the formation of a full CopA-CopT duplex appears to be a very slow process in vitro. Unexpectedly, we found that the loop-loop interaction persists in the predominant CopA-CopT complex and is stabilized by intermolecular base pairing involving the 5′-proximal 30 nucleotides of CopA and the complementary region of CopT. This almost irreversible complex suffices to inhibit ribosome binding at the tap ribosome binding site and may be the inhibitory complex in vivo. Natural antisense RNAs control a variety of biological functions, mainly in prokaryotic accessory genetic elements such as phages, transposons, and plasmids (1Wagner E.G.H. Simons R. Annu. Rev. Microbiol. 1994; 48: 713-742Crossref PubMed Scopus (365) Google Scholar). Most antisense RNAs found regulate the plasmid copy number. The efficiency of antisense RNA control depends on the rate constant for the binding between antisense and target RNA and on the intracellular concentration of the inhibitor RNA. The binding pathways were extensively studied in ColE1 (2Tomizawa J. Cell. 1984; 38: 861-870Abstract Full Text PDF PubMed Scopus (235) Google Scholar, 3Tomizawa J. Cell. 1986; 47: 89-97Abstract Full Text PDF PubMed Scopus (103) Google Scholar, 4Tomizawa J. J. Mol. Biol. 1990; 212: 683-694Crossref PubMed Scopus (84) Google Scholar) and R1 (5Persson C. Wagner E.G.H. Nordström K. EMBO J. 1988; 7: 3279-3288Crossref PubMed Scopus (97) Google Scholar, 6Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3767-3775Crossref PubMed Scopus (79) Google Scholar, 7Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3777-3785Crossref PubMed Scopus (63) Google Scholar) plasmids. In both systems, the binding process is seen as a series of reactions leading to a stable duplex. The initial, rate-limiting step involves the formation of a transient loop-loop interaction (kissing complex) that is followed by a subsequent nucleation step via the single-stranded 5′ segment of the antisense RNA. This interaction is believed to result in the formation of a double-stranded RNA along the entire length of the target site. However, the full duplex may not be required for inhibition since the transient kissing complex appears to be sufficient for control in ColE1 (8Lin-Chao S. Cohen S.N. Cell. 1991; 65: 1233-1242Abstract Full Text PDF PubMed Scopus (178) Google Scholar), R1 (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar), and IncB (10Siemering K.R. Praszkier J. Pittard A.J. J. Bacteriol. 1994; 176: 2677-2688Crossref PubMed Google Scholar) plasmids. Plasmid R1 belongs to the IncFII group whose initiation frequency is controlled by an antisense RNA, CopA. CopA inhibits RepA synthesis by binding to the leader region of the repA mRNA (CopT) 80 nucleotides upstream of the repA translational start codon (11Nordström K. Molin S. Light J. Plasmid. 1984; 12: 71-90Crossref PubMed Scopus (124) Google Scholar). RepA synthesis requires the translation of a short leader peptide (Tap) encoded immediately downstream of the CopA target site, overlapping the repA reading frame by two nucleotides (12Blomberg P. Nordström K. Wagner E.G.H. EMBO J. 1992; 11: 2675-2683Crossref PubMed Scopus (87) Google Scholar). This implies that CopA inhibits repA translation indirectly by preventing tap translation in vivo (13Blomberg P. Engdahl H.M. Malmgren C. Romby P. Wagner E.G.H. Mol. Microbiol. 1994; 12: 49-60Crossref PubMed Scopus (38) Google Scholar). Extensive mutational analyses revealed that the structures of both the antisense and the target RNAs are critical for obtaining an optimal binding rate and in vivo control (1Wagner E.G.H. Simons R. Annu. Rev. Microbiol. 1994; 48: 713-742Crossref PubMed Scopus (365) Google Scholar). However, although the role of loop-loop interactions in the binding pathway of the antisense RNA and its target has been characterized (7Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3777-3785Crossref PubMed Scopus (63) Google Scholar, 8Lin-Chao S. Cohen S.N. Cell. 1991; 65: 1233-1242Abstract Full Text PDF PubMed Scopus (178) Google Scholar, 9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar), the structure(s) of more stable binding intermediate(s) that lead to a complete duplex remained speculative. In the present study, the interaction between CopA and CopT was followed by using Pb(II)-induced cleavages. The loop-loop interaction was also probed using a truncated antisense RNA (CopI) that is unable to form a full duplex with CopT (6Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3767-3775Crossref PubMed Scopus (79) Google Scholar). Furthermore, RNase III that processes the CopA-CopT complex (14Blomberg P. Nordström K. Wagner E.G.H. EMBO J. 1990; 9: 2331-2340Crossref PubMed Scopus (150) Google Scholar) was used to distinguish between antisense-target RNA complexes differing in the extent of duplex formation. Our results indicate that the stable CopA-CopT complex is not fully base-paired in the region of complementarity. Remarkably, an extended loop-loop interaction persists in the CopA-CopT complex, maintaining the structures of the major CopA and CopT stem loops. This kissing complex is additionally stabilized by an intermolecular helix involving the 5′-proximal 30 nucleotides of the antisense RNA and its counterpart in CopT. The significance of this CopA-CopT complex for the regulatory mechanism is discussed. CopT, CopA, and CopI were synthesized with T7 RNA polymerase from polymerase chain reaction-generated DNA fragments as described by Hjalt and Wagner (15Hjalt T. Wagner E.G.H. Nucleic Acids Res. 1995; 23: 571-579Crossref PubMed Scopus (38) Google Scholar). Transcription of CopT yields a run-off product of 302 nucleotides that carries two G residues at its 5′ end instead of the GU sequence of the wild type repA mRNA. The CopA RNA contains a 5′-terminal G instead of an A residue. 5′-End labeling of dephosphorylated RNA was performed with T4 polynucleotide kinase and [γ-32P]ATP (16Maniatis T. Fritsch E.F. Sambrook J. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1982Google Scholar). Labeled RNAs were purified by polyacrylamide-urea gel electrophoresis, eluted, and precipitated twice with ethanol. Before use, the different RNAs were dissolved in RNase-free water and renatured by incubation at 90 °C for 2 min followed by slow cooling at room temperature in TMN buffer (20 mm Tris acetate, pH 7.5, 10 mm magnesium acetate, 100 mm sodium acetate). Experimental conditions were adapted from Hjalt and Wagner (15Hjalt T. Wagner E.G.H. Nucleic Acids Res. 1995; 23: 571-579Crossref PubMed Scopus (38) Google Scholar). Complex formation was carried out at 37 °C for 5 or 15 min in TMN buffer with 5′-end-labeled CopT (3 × 10−8m) and a 5-fold excess of unlabeled CopA or CopI (1.5 × 10−7m) or with 5′-end-labeled CopA or CopI (4 × 10−8m) and a 5-fold excess of unlabeled CopT (2 × 10−7m). Full duplexes between 5′-end-labeled CopT and CopA or CopI were formed by incubation at 90 °C for 2 min followed by slow cooling to 37 °C in TMN buffer. Reaction mixtures (10 μl) contained end-labeled RNA (20,000 cpm) either free or in complex, carrier tRNA (2.5 μg), and lead(II) acetate at a final concentration of 8 or 16 mm in TMN buffer. Incubation was done at 37 °C for 5 min, and the reactions were stopped by the addition of 5 μl of 0.1 m EDTA followed by ethanol precipitation. Pellets were dissolved in 10 μl of stop buffer (93% formamide, 30 mm EDTA, 0.05% xylene cyanol, 0.05% bromphenol blue, 0.5% SDS). Before electrophoresis on 15% polyacrylamide-urea 8 m gels, the samples were heated at 90 °C for 3 min. 5′-End-labeled CopT was incubated with a 5-fold excess of CopA or CopI (10−7m) in 10 μl of TMN buffer containing 1 mm dithiothreitol at 37 °C for 5, 15, or 30 min. Full RNA duplexes were formed as described above. Cleavage of free or complexed CopT was at 37 °C from 25 s to 10 min with 0.002 μg of RNase III or from 1 to 10 min with 0.25 μg of RNase III (a gift from D. Court). All reactions contained 10 μg of carrier tRNA. Reactions were stopped by phenol extraction followed by RNA precipitation. Pellets were dissolved in 10 μl of stop buffer and analyzed as described above. The primary binding intermediate involving the loop-loop interaction was studied by using the truncated antisense RNA, CopI, which is unable to promote a full duplex with CopT (6Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3767-3775Crossref PubMed Scopus (79) Google Scholar). In both free RNAs, the major cleavages occur in single-stranded regions, mainly in the hairpin loops and the interhelical regions (Fig. 1). The cleavage positions are shown schematically on the secondary structures of the RNAs (Fig. 2). The cleavages in loop II of CopI (but also CopA) and in loop II′ of CopT extend into the helical regions, which is probably due to the presence of several bulged nucleotides (15Hjalt T. Wagner E.G.H. Nucleic Acids Res. 1995; 23: 571-579Crossref PubMed Scopus (38) Google Scholar). Under native conditions (kissing complex), CopI binding induces significant changes in the Pb(II) cleavage pattern of CopT (Fig. 1 A, lanes 5 and 6). Strong protection is observed on the 3′ side of the upper stem II′ at U117 to A123 and to a lesser extent at U102 to C107. Furthermore, increased reactivity induced by CopI binding occurs at U109 and G112. The Pb(II) sensitivity of the non-complementary regions is not altered, indicating that CopI binding induces structural changes restricted to the upper part of stem loop II′ of CopT (Fig.1 A). When full duplexes are formed, complete resistance to Pb(II) hydrolysis is observed in the region encompassing U94 to C138 (Fig. 1 A, lanes 8 and 9). Therefore, Pb(II) is a sensitive probe that allows discrimination between a full duplex and the kissing complex. Conversely, binding of CopT induces enhanced Pb(II) cleavages at A29 and, to a lesser extent, at C28, A30, and A31 in CopI (Fig. 1 A). Most of the nucleotides located in loop II of CopI are still cleaved by Pb(II) upon CopT binding, indicating that the upper part of helix II participates in the formation of the intermolecular helix that characterizes the kissing complex. Significant protection on both sides of the upper helix of CopI is observed, especially on the 5′ side of helix II at U33 to A39 (Fig.1 B).Figure 2Pb(II) probing and effect of binding of antisense RNA or CopT. The results are summarized on the secondary structure of CopA (left), CopI, and CopT (right). Positions of Pb(II) cleavages in the free RNAs are indicated byarrows. Large filled arrows, strong; small shaded arrows, moderate; and open head arrows, weak cleavages. The effect of antisense RNA or CopT binding is indicated as follows: strong (•) and moderate (○) protection. Enhancements are represented by stars proportional to the intensity. The labeled RNA used is indicated by an asterisk. The effects of CopT binding on CopI or CopI binding on CopT are shown in theinsets, using the same symbols.View Large Image Figure ViewerDownload (PPT) A CopA-CopT complex in which either CopA or CopT was labeled at the 5′ end was formed under native conditions at 37 °C and subjected to Pb(II) probing. Under these conditions, stable complexes were formed within 5 min at 37 °C (data not shown). Unexpectedly, CopA binding induces the same reactivity changes in stem loop II′ of CopT as CopI (cf. lanes 5 and 6 withlanes 11 and 12 in Fig. 1 A). In particular, the bulged cytosine 101 and part of the external loop II′ of CopT remain accessible to Pb(II). Additional protections occur in the 140–162 region (Fig. 2). No major structural rearrangement of CopT RNA occurs in the non-complementary region since the Pb(II) sensitivity remained unchanged (Fig. 1 A). In contrast, the formation of a full CopA-CopT duplex entails complete protection of the CopA binding site from nucleotide 81 to 162 of CopT (Fig. 1 A, lanes 14 and 15). Conversely, CopT binding induces the same reactivity changes in stem loop II of CopI and CopA (see Fig. 1,B and C). This is clearly illustrated by a strong enhancement at A60 and weaker increase at C59, A61, and A62 in CopA. In addition, the 5′-proximal 30 nucleotides of CopA are protected from Pb(II) in the presence of CopT (Fig. 2). Altogether, these results suggest that a full duplex is not the major product of the CopA-CopT complex in vitro. The extended loop-loop contact, as found in the CopI-CopT complex, persists in the CopA-CopT complex, but this complex is additionally stabilized by the formation of an intermolecular helix that involves the 5′-proximal 30 nucleotides of CopA and the complementary region of CopT. RNase III was shown to process the CopA-CopT complex in vivo (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar). Therefore, we reinvestigated RNase III-dependent cleavages under conditions of full duplex formation and native CopA-CopT complex formation to define the extent of duplex formation. The cleavage positions were mapped using 5′-end-labeled CopT RNA (Fig.3 A). The cleavage sites are indicated on a schematic drawing (Fig. 3 B). When CopA or CopI was heated and slowly cooled in the presence of labeled CopT, allowing the formation of full RNA duplexes, the same concentration of RNase III was able to fully digest both duplexes. The strongest cleavages were located at A93, U94, U99, and U100 (Fig. 3 A). Conversely, with the native CopA-CopT complex, RNase III produced two major cleavages in CopT at U149 and C154 (Fig. 3 A). This result is in agreement with the primary RNase III cleavage identified at C154 in vivo and in vitro (14Blomberg P. Nordström K. Wagner E.G.H. EMBO J. 1990; 9: 2331-2340Crossref PubMed Scopus (150) Google Scholar). However, weak cleavages corresponding to the presence of the full duplex were also found in CopT bound to CopA, especially after a prolonged time of incubation (at least up to 15 min, Fig.1 A). Under the same conditions, the CopI-CopT complex formed at 37 °C was completely resistant to RNase III hydrolysis (data not shown). Therefore, these experiments indicate that the full duplex is not the major product of CopA-CopT complex in vitro and that its formation appears to be a slow process in vitro. Previous studies suggested that the binding of CopA to CopT proceeds in at least two experimentally distinguishable steps: kissing and hybridization duplex formation (5Persson C. Wagner E.G.H. Nordström K. EMBO J. 1988; 7: 3279-3288Crossref PubMed Scopus (97) Google Scholar, 6Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3767-3775Crossref PubMed Scopus (79) Google Scholar, 7Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3777-3785Crossref PubMed Scopus (63) Google Scholar). The first, rate-limiting step is the formation of a transient loop-loop interaction (a so-called kissing complex) that was supposed to permit the subsequent formation of a complete RNA duplex. The present structural probing performed on the kissing complex shows that most of the nucleotides in loop II of the truncated antisense RNA CopI, as well as in the complementary loop, are accessible to Pb(II). However, since point mutations within the CGCC sequence in loop II of CopA strongly affect the pairing rate and result in copy number-up phenotypes (5Persson C. Wagner E.G.H. Nordström K. EMBO J. 1988; 7: 3279-3288Crossref PubMed Scopus (97) Google Scholar, 11Nordström K. Molin S. Light J. Plasmid. 1984; 12: 71-90Crossref PubMed Scopus (124) Google Scholar, 17Hjalt T. Wagner E.G.H. Nucleic Acids Res. 1992; 20: 6723-6732Crossref PubMed Scopus (65) Google Scholar), these nucleotides are inferred to be part of the primary kissing complex (Fig.4 B). This intermediate, not detectable by the present probing approach, rapidly rearranges to an extended kissing complex in which the 5′ side of the upper part of stem II interacts in an interstrand pairing with CopT (Fig. 4 C). This is further supported by RNase V1 cuts in the upper stem II of CopI and increased accessibility of the 3′ side of loop II toward RNase T2 (6Persson C. Wagner E.G.H. Nordström K. EMBO J. 1990; 9: 3767-3775Crossref PubMed Scopus (79) Google Scholar). Notably, the formation of the extended kissing complex induces a strong Pb(II) cleavage in the loop II of the antisense RNA. This cleavage may be due to the presence of a magnesium binding site. In line with this view, Gregorian and J. Mol. Biol. 1995; PubMed Scopus Google Scholar) that the of at least two magnesium stabilized the loop-loop interaction in the stem loop complex of plasmid The Pb(II) probing also that the extended kissing complex persists in the CopA-CopT complex. Furthermore, protection in the 5′ segment of CopA and the complementary region of CopT the formation of interstrand base also that RNase III, which processes the CopA-CopT vivo (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar), the native CopA-CopT complex and the full in completely different The major RNase III cleavage site in CopT bound to CopA under native conditions is within the intermolecular helix (Fig.3 which is with known site of RNase III EMBO J. PubMed Scopus (65) Google Scholar). The extended kissing complex that is not in vivo (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar) is also resistant to RNase III hydrolysis in vitro. The RNase III experiments indicate that a of the CopA-CopT complex proceeds to a full duplex in vitro. This that the extended kissing complex induces a structural that the two complementary single-stranded the II and II′ to pairing (Fig. 4 studies that stable CopA-CopT complexes were formed very rapidly (5Persson C. Wagner E.G.H. Nordström K. EMBO J. 1988; 7: 3279-3288Crossref PubMed Scopus (97) Google Scholar), the of a stable and not fully base-paired CopA-CopT complex is The present that this intermediate (Fig. 4 for at least 15 min under conditions in which kissing complex formation occurs within a and that to a full duplex is very as was in the ColE1 J. J. Mol. Biol. 1990; 212: 683-694Crossref PubMed Scopus (84) Google Scholar). for the binding pathway of the antisense RNA in translational control of repA in IncB plasmids been (10Siemering K.R. Praszkier J. Pittard A.J. J. Bacteriol. 1994; 176: 2677-2688Crossref PubMed Google Scholar). The antisense inhibits the formation of a structure that Praszkier J. Pittard A.J. J. Bacteriol. 1994; Google Scholar). CopA and stable with within the loops. However, has a short 5′ and a 5′ stem loop probing indicated that and the target of plasmid not rapidly form a full duplex in form an extended kissing complex stabilized by base involving the 5′ of In experiments that a kissing complex between two complementary stem loops of involves base pairing of loop residues Gregorian 1995; PubMed Scopus Google Scholar). In this the of both are short base and the different structures of the stem loop complex with that of CopI and is to that a loop permit an induced rearrangement to a transient but stable extended kissing complex. of the bulged residues in stem II of CopA and stem II′ of CopT strongly both binding rate and kissing complex T. Wagner E.G.H. Nucleic Acids Res. 1995; 23: PubMed Scopus Google Scholar). The same major RNase III cleavage site was identified in vitro and in vivo at C154 in CopT RNA (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar). the of the CopA-CopT complexes that in vivo may not be fully The extent of complete pairing in vivo may be due to RNase III followed by very of the we that the formation of the CopA-CopT complex in vitro is sufficient to for the observed translational shown that this complex initiation complex formation at ribosome binding site C. Engdahl H.M. Romby P. Wagner E.G.H. RNA. Google Scholar). the extended kissing complex was able to with ribosome binding that well with in vivo results (9Wagner E.G.H. Blomberg P. Nordström K. EMBO J. 1992; 11: 1195-1203Crossref PubMed Scopus (31) Google Scholar). In agreement with Wagner E.G.H. Nucleic Acids Res. PubMed Scopus Google Scholar), the present data also that the binding of CopI or CopA not major changes in the non-complementary of CopT, in the tap and repA ribosome binding site Pb(II) is very sensitive to changes that the structure of the kissing complex an induced RNA structure ribosome binding In the wild type the presence of the 5′ segment of CopA the transient kissing complex, helical of double-stranded RNA immediately 5′ of the tap which is sufficient to completely and tap The that complete pairing between an antisense and a target RNA is slow and be for inhibition may a an complex stem loops of antisense and target RNAs to be to complex and biological control may to binding intermediates in the inhibitory of such a mechanism is that inhibition be pairing S. Wagner E.G.H. EMBO J. 1994; PubMed Scopus Google Scholar) since the stem loops for binding rate T. J. Annu. Rev. 1991; PubMed Scopus Google Scholar, K. Wagner E.G.H. 1994; Full Text PDF PubMed Scopus Google Scholar), may be to are suggest that such may be for the of antisense RNA and for and for critical reading of the