Hfq binding at RhlB-recognition region of RNase E is crucial for the rapid degradation of target mRNAs mediated by sRNAs in Escherichia coli

Authors

  • Yoshiki Ikeda,

    1. Division of Biological Science, Graduate School of Science, Nagoya University, Chikusa, Nagoya 464-8602, Japan.
    Search for more papers by this author
    • Research Fellow of the Japan Society for the Promotion of Science.

    • Yoshiki Ikeda and Mieko Yago contributed equally to this work.

  • Mieko Yagi,

    1. Division of Biological Science, Graduate School of Science, Nagoya University, Chikusa, Nagoya 464-8602, Japan.
    Search for more papers by this author
    • Yoshiki Ikeda and Mieko Yago contributed equally to this work.

  • Teppei Morita,

    1. Faculty of Pharmaceutical Sciences, Suzuka University of Medical Sciences, Suzuka, Mie 513-0816, Japan.
    Search for more papers by this author
  • Hiroji Aiba

    Corresponding author
    1. Faculty of Pharmaceutical Sciences, Suzuka University of Medical Sciences, Suzuka, Mie 513-0816, Japan.
      E-mail aiba@suzuka-u.ac.jp; Tel. (+81) 59 340 0573; Fax (+81) 59 368 1271.
    Search for more papers by this author

E-mail aiba@suzuka-u.ac.jp; Tel. (+81) 59 340 0573; Fax (+81) 59 368 1271.

Summary

An RNA chaperon Hfq along with Hfq-binding sRNAs stably binds to RNase E in Escherichia coli. The role of the Hfq–RNase E interaction is to recruit RNase E to target mRNAs of sRNAs resulting in the rapid degradation of the mRNA–sRNA hybrid. The C-terminal scaffold region of RNase E is responsible for the interaction with Hfq. Here, we demonstrate that the scaffold region can be deleted up to residue 750 without losing the ability to cause the rapid degradation of target mRNAs mediated by Hfq/sRNAs. The truncated RNase E750 can still bind to Hfq although the truncation significantly reduces the Hfq-binding ability. We conclude that the subregion between 711 and 750 is sufficient for the functional interaction with Hfq to support the rapid degradation of ptsG mRNA although additional subregions within the scaffold are also involved in Hfq binding. Deletion of the 702–750 region greatly impairs the ability of RNase E to cause the degradation of ptsG mRNA. In addition, a polypeptide corresponding to the scaffold region binds to Hfq without the help of RNA. Finally, we demonstrate that overexpression of RhlB partially inhibits the Hfq binding to RNase E and the rapid degradation of ptsG mRNA.

Introduction

RNase E is a major endoribonuclease responsible for mRNA degradation and/or RNA processing in Escherichia coli (Carpousis, 2002; 2007). It forms a multiprotein complex called the RNA degradosome associating with PNPase, the ATP-dependent RNA helicase B containing a DEAD-box motif (RhlB), and the glycolytic enzyme enolase as the principle partners (Carpousis et al., 1994; Miczak et al., 1996; Py et al., 1996). It is believed that both RNase E and PNPase within the degradosome act in a concerted fashion during the degradation of RNAs while RhlB unwinds the secondary structure of target RNAs to permit access of RNase E and/or PNPase. Indeed, PNPase and RNase E were shown to act together in the degradation of RNA I of ColE1 in vitro (Xu and Cohen, 1995). In addition, RhlB was shown to stimulate the degradation of mRNAs by RNase E and PNPase in vitro and in vivo (Vanzo et al., 1998; Coburn et al., 1999; Leroy et al., 2002; Khemici and Carpousis, 2004; Khemici et al., 2005). The role of enolase in the degradosme is still unknown although we showed previously that enolase itself is required for the induction a small regulatory RNA (sRNA), SgrS, in response to glucose-phosphate stress to downregulate of the target ptsG mRNA encoding the glucose transporter (Morita et al., 2004; 2005).

It has been known that other proteins such as polyphosphate kinase, DnaK and GroEL associate with RNase E in substoichiometric amounts relative to the major components (Miczak et al., 1996; Blum et al., 1997; Regonesi et al., 2006). Furthermore, recent studies demonstrated that several different RNase E-based multiprotein complexes (alternative forms of the RNA degradosome) exist, in addition to the canonical degradosome, in E. coli and that the composition of the complexes can undergo larger changes depending on physiological conditions (Carpousis, 2007). For example, another DEAD-box RNA helicase, CsdA, which is a member of cold shock proteins, was shown to associate with RNase E at low temperature (Prud'homme-Genereux et al., 2004). This change of the degradosome composition may be responsible for the changes in mRNA stability under cold shock response (Awano et al., 2007). The composition and function of degradosome are also modulated by proteins RraA and RraB that inhibit the activity of RNase E (Gao et al., 2006).

One of newly identified partners of RNase E is an RNA-binding protein Hfq that is involved in a variety of physiological processes including the regulatory action of base-pairing sRNAs (Valentin-Hansen et al., 2004; Aiba, 2007; Brennan and Link, 2007). Previously we found that Hfq stably associates with RNase E and at least two Hfq-binding sRNAs, SgrS and RyhB, bind to RNase E through Hfq (Morita et al., 2005). The resulting ribonucleoprotein complexes act on the target mRNAs through base-pairing guided by each sRNA to cause translational inhibition and coupled degradation of both mRNAs and sRNAs under specific stress conditions such as accumulation of glucose 6-phosphate (G6P) and Fe depletion (Masse et al., 2003; Morita et al., 2005). The role of the Hfq–RNase E interaction is to recruit RNase E to the target mRNAs of sRNAs and to cause the rapid degradation of the mRNA–sRNA hybrid although the RNase E recruitment is not necessary for the translational silencing of mRNAs mediated by Hfq/sRNAs (Morita et al., 2006). Interestingly, the RNase E–Hfq complex is distinct from the canonical RNA degradosome because it does not contain enolase and RhlB (Morita et al., 2005). The Hfq binding to RNase E was also shown in a systematic analysis of protein–protein interactions in E. coli (Butland et al., 2005).

The RNase E polypeptide is composed of a globular N-terminal catalytic region corresponding to amino acid residues 1–529 and a natively unstructured C-terminal non-catalytic region corresponding to residues 530–1061 that are responsible for physical and functional interactions with other components of the RNA degradosome (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008). An arginine-rich segment of residues 604–688 is involved in RNA binding and believed to enhance the RNase E activity in mRNA degradation. The remaining C-terminal region corresponding to residues 701–1061 forms a scaffold for interactions with partner proteins. The fine biochemical analyses revealed that the binding sites for PNPase, enolase and RhlB are mapped at residues 1021–1061, 833–850 and 696–762 respectively. We demonstrated previously that Hfq no longer binds to the truncated RNase E701 lacking the entire scaffold region and therefore the C-terminal scaffold region of RNase E is responsible for the interaction with Hfq (Morita et al., 2005). The Hfq–RNase E interaction is necessary for the rapid degradation of target mRNAs because the RNase E701 no longer causes the sRNA-mediated degradation of target mRNAs (Morita et al., 2005). However, it remains to be clarified which portion of the scaffold region is involved in the Hfq binding.

The aim of the present study is to refine the physical and functional interaction site(s) of Hfq within the scaffold region of RNase E. We demonstrate that the C-terminal scaffold region can be deleted up to residue 750 without losing the ability to support the rapid degradation of ptsG mRNA mediated by Hfq/SgrS. The truncated RNase E750 is still able to bind to Hfq although the C-terminal truncation reduces significantly the Hfq-binding ability. In addition, we show that deletion of 702–750 greatly reduces the ability of RNase E to support the degradation of ptsG mRNA while it affects only moderately the Hfq-binding ability. We conclude that the subregion between 711 and 750 of RNase E is sufficient for the interaction with Hfq to cause the rapid degradation of ptsG mRNA. In addition, we show that a polypeptide corresponding to the scaffold region binds to Hfq without the help of RNA. We also demonstrate that overexpression of RhlB reduces the Hfq binding to RNase E and the SgrS-mediated degradation of ptsG mRNA.

Results

Expression of truncated RNase E proteins

Previously we manipulated the chromosomal rne gene to express RNase E-FLAG, RNase E844-FLAG or RNase E701-FLAG (Morita et al., 2004). RNase E-FLAG possessing a C-terminally fused FLAG epitope tag carries the complete amino acid sequence (1061 residues) of RNase E while RNase E844-FLAG and RNase E701-FLAG are missing the C-terminal 217 and 360 amino acid residues respectively. As will be shown later, RNase E844-FLAG retains the Hfq-binding activity as well as the ability to cause the rapid degradation of target mRNAs mediated by sRNAs. This implies that the Hfq binding site required for sRNA-mediated rapid degradation of target mRNAs is located between residues 702 and 844 of RNase E. To map the functional Hfq binding site of RNase E more precisely, we constructed strains carrying truncated rne alleles encoding RNase E800-FLAG, RNase E750-FLAG and RNase E710-FLAG (Fig. 1A). It is expected that RNase E800-FLAG possesses the RhlB binding site but lacks the PNPase and enolase binding sites while RNase E750-FLAG and RNase E710-FLAG lacks also the RhlB binding site. These strains carrying the modified chromosomal rne alleles would be useful to analyse the effects of RNase E truncations on the Hfq–RNase E interaction and on the destabilization of target mRNAs mediated by sRNAs.

Figure 1.

Truncated FLAG-tagged RNase E proteins and their expression.
A. The organization of RNase E (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008) is shown at the top. Truncated FLAG-tagged RNase E proteins expressed from modified chromosomal rne alleles are shown below. The shaded boxes represent the FLAG polypeptide. RNase E-FLAG retains the 1061 amino acid residues of the full-length RNase E. RNase E844-FLAG, E800-FLAG, E750-FLAG, E710-FLAG and RNase E701-FLAG lack the last 217, 261, 311, 351 and 360 C-terminal amino acid residues of RNase E.
B. Western blot analysis of FLAG-tagged RNase E proteins. Strains indicated were grown in LB medium to A600 = 0.6. Cell extracts were prepared and each sample corresponding to approximately 0.01 A600 unit was subjected to Western blot analysis using anti-FLAG antibody. Lane 1, IT1568 (wild-type); lane 2, TM338 (rne-FLAG); lane 3, TM528 (rne701-FLAG); lane 4, MY12 (rne710-FLAG); lane 5, TM719 (rne750-FLAG); lane 6, MY02 (rne800-FLAG); and lane 7, TM527 (rne844-FLAG). Protein size standards are indicated on the right.

We examined first the expression of the FLAG-tagged RNase E proteins in these strains by Western blotting using anti-FLAG antibodies. As shown in Fig. 1B, all of the truncated FLAG-tagged RNase E proteins were stably expressed. We also examined the growth property of strains carrying the truncated rne alleles in LB medium. All strains exhibited growth rates that were essentially identical to the wild-type cells, indicating none of the truncations affect the cell growth at least under normal conditions (data not shown).

Effects of C-terminal truncations of RNase E on sRNA-mediated destabilization of target mRNAs

We examined the effects of a series of C-terminal truncations of RNase E on the SgrS-mediated destabilization of ptsG mRNA. Cells expressing each of the FLAG-tagged truncated RNase E proteins were grown in LB medium, exposed to either glucose (Glc) or α-methylglucoside (αMG) for 10 min, and then total RNAs were isolated and analysed by Northern blotting using sgrS and ptsG probes. The ptsG mRNA but not SgrS RNA was well expressed in all strains when exposed to Glc (Fig. 2A, odd lanes). As expected, SgrS RNA was induced at a significant level and the ptsG mRNA was dramatically destabilized when cells expressing wild-type RNase E or RNase E-FLAG were exposed to αMG (Fig. 2A, lanes 2 and 4) while the degradation of ptsG mRNA in response to αMG addition was abolished without losing the induction of SgrS in cells expressing RNase E701-FLAG (Fig. 2A, lane 14). Similar experiments were carried out using strains expressing RNase E844-FLAG, RNase E800-FLAG, RNase E750-FLAG or RNase E710-FLAG. The degradation of ptsG mRNA in response to αMG exposure occurred in cells expressing RNase E844-FLAG, RNase E800-FLAG or RNase E750-FLAG but not in cells expressing RNase E710-FLAG while SgrS was induced in all strains (Fig. 2A, lanes 6, 8, 10 and 12).

Figure 2.

A. Effects of C-terminal truncations of RNase E on the expression of ptsG mRNA. Strains indicated were grown in LB medium. At A600 = 0.6, the culture was split, 0.2% glucose (Glc) or α-methylglucoside (αMG) was added to each culture and incubation was continued for 10 min. Total RNAs were prepared and 15 µg or 5 µg of each RNA sample was subjected to Northern blot analysis using ptsG or sgrS probe respectively.
B. Effects of the C-terminal truncations of RNase E on the expression of sodB mRNA. Strains indicated were grown in LB medium. At A600 = 0.6, the culture was split and 250 µM 2,2′-didipyridyl (Dip) was added to one culture and incubation was continued for 15 min. Total RNAs were prepared and 3 µg of each RNA sample was subjected to Northern blot analysis using sodB or ryhB probes.

We also examined the effects of the C-terminal truncations of RNase E on the RyhB-mediated destabilization of sodB mRNA. RyhB RNA is induced by Fe depletion resulting in the destabilization of target mRNAs such as sodB mRNA encoding superoxide dismutase (Masse and Gottesman, 2002; Masse et al., 2003). RyhB RNA was induced upon the addition of 2, 2′-dipyridyl (Dip) in all strains expressing FLAG-tagged RNase E proteins. The RyhB-mediated destabilization of sodB mRNA occurred in cells expressing RNase E-FLAG (Fig. 2B, lane 2) but not RNase E710-FLAG and RNase E701-FLAG (Fig. 2B, lanes 6 and 8). The destabilization of sodB mRNA was also observed although less weakly in cells expressing RNase E750-FLAG (Fig. 2B, lane 4).

These results imply that the region between residues 751–1061 of RNase E is essentially dispensable for the destabilization of target mRNAs mediated by sRNAs. We conclude that the region between residues 711 and 750 of RNase E is responsible for the functional interaction with Hfq to cause the sRNA-mediated destabilization of target mRNAs.

Physical interaction of truncated RNase E proteins with Hfq

The results mentioned above strongly suggest that Hfq interacts physically with the region between residues 711 and 750 of RNase E. Then, we tested for the binding of Hfq to the truncated RNase E proteins by RNase E pull-down assay. The cell extracts were prepared first from cells expressing the FLAG-tagged truncated RNase E proteins grown in LB medium. Each extract was incubated with anti-FLAG M2-agarose beads and proteins bound were eluted from the beads. Proteins in the bound fraction were analysed by Western blotting using anti-FLAG antibodies. All of the FLAG-tagged RNase E proteins but not native RNase E were efficiently recovered in the bound fraction (Fig. 3A). Then, the proteins bound to the anti-FLAG M2-agarose beads were analysed by Western blotting using anti-Hfq antibodies. We confirmed the previous observation (Morita et al., 2005) that Hfq co-purifies with RNase E-FLAG but not with RNase E701-FLAG (Fig. 3A, lanes 2 and 7). We newly demonstrate that other truncated RNase E proteins except RNase E710-FLAG retain the Hfq-binding ability although the Hfq binding is reduced significantly with truncations (Fig. 3A, lanes 3–6). These results suggest that RNase E interacts with Hfq at several different sites within the scaffold region of RNase E either directly or indirectly. A moderate reduction in Hfq binding in RNase E844-FLAG suggests that the region between residues 845 and 1061 of RNase E is somehow involved in Hfq binding (Fig. 3A, lane 3). In addition, it is apparent that the region between residues 801 and 844 of RNase E is also involved in Hfq binding because truncation up to residue 800 markedly reduces the Hfq-binding ability while further truncation to residue 750 causes no additional reduction in Hfq binding (Fig. 3A, lanes 4 and 5). Finally, the results indicate that Hfq also binds to the region between residues 711 and 750 because truncation up to residue 710 completely eliminates the Hfq binding (Fig. 3A, lane 6). Importantly, the Hfq binding at residues 711–750, which is about 10% of total Hfq binding by the entire scaffold region, is sufficient to recruit RNase E to target mRNAs because RNase E750-FLAG retains the activity to cause the sRNA-mediated rapid degradation of target mRNAs (Fig. 2A and B). In other words, the Hfq–RNase E interactions at residues 845–1061 and at residues 801–844 are not needed for the sRNA-mediated destabilization of target mRNAs.

Figure 3.

A. Physical interaction of FLAG-tagged RNase E proteins with Hfq and other partner proteins under non-stress condition. Strains indicated were grown in 200 ml of LB medium to A600 = 0.6. Cell extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose as described in Experimental procedures. Proteins bound to the beads were analysed by Western blotting. Anti-FLAG, anti-Hfq, anti-RhlB, anti-enolase and anti-PNPase antibodies were used to detect FLAG-tagged RNase E proteins, Hfq, RhlB, enolase and PNPase respectively. Following amounts of bound fractions were subjected to the gel electrophoresis: RNase E (3 µl), Hfq (5 µl) and other proteins (2 µl).
B. Physical interaction of FLAG-tagged RNase E proteins with Hfq and SgrS under glucose-phosphate stress. Strains indicated were grown in 200 ml of LB medium. At A600 = 0.6, 0.1% αMG was added to each culture and incubation was continued for 10 min. Cell extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose. Proteins bound to the beads were analysed by Western blotting as in (A). For analysis of SgrS RNA associated with RNase E, 3 µl of deprotenized bound fraction (see Experimental procedures) was subjected to Northern blotting using the sgrS probe.
C and D. Effects of glucose-phosphate stress or Fe depletion on RNase E–Hfq interaction. TM338 (rne-FLAG) (C) and TM719 (rne750-FLAG) (D) were grown in LB medium. At A600 = 0.6, the culture was divided into three parts with the following additions: none, 0.1% αMG and 250 µM Dip. Incubation was continued for 10 min. Cell extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose. Proteins bound to the beads were analysed by Western blotting as in (A).

We also carried out pull-down assays by using extracts prepared from cells exposed to αMG for 10 min. Again, RNase E710-FLAG and RNase E701-FLAG no longer bind Hfq while RNase E750-FLAG, RNase E800-FLAG and RNase E844-FLAG retain a significant Hfq binding (Fig. 3B). To examine whether the stress-induced SgrS is associated with truncated RNase E proteins, protein samples were treated with phenol and subjected to Northern blotting. A significant amount of SgrS RNA was detected in the affinity-purified RNase E-FLAG, RNase E844-FLAG, RNase E800-FLAG and RNase E750-FLAG samples while no SgrS RNA was detected in the RNase E710-FLAG and RNase E701-FLAG samples (Fig. 3B). These results are consistent with the conclusion that residues 711–750 of RNase E is responsible for the functional interaction with Hfq to recruit RNase E to target mRNAs. To examine whether glucose-phosphate stress or Fe depletion affects the Hfq–RNase E interaction, cells expressing RNase E-FLAG or RNase E750-FLAG were grown in LB medium and treated with either αMG or Dip for 10 min. Cell extracts were prepared and subjected to the pull-down assay. The addition of αMG or Dip did not affect the amount of Hfq associated with RNase E-FLAG (Fig. 3C) and RNase E750-FLAG (Fig. 3D). Thus, neither glucose-phosphate stress nor Fe depletion seems to affect the RNase E–Hfq interaction.

Interactions of truncated RNase E proteins with PNPase, enolase and RhlB

The residues 1021–1061, 833–850 and 696–762 of RNase E are reported to correspond to micro-domains required for the interaction with PNPase, enolase and RhlB respectively (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008). In order to examine the ability of FLAG-tagged RNase E proteins to interact with these three major components of the RNA degradosome, the proteins bound to the anti-FLAG M2-agarose beads were also analysed by Western blotting using anti-PNPase, anti-RhlB and anti-enolase antibodies. We confirmed the previous observation (Morita et al., 2005) that PNPase, RhlB and enolase were co-eluted with RNase E-FLAG while none of three other major degradosome components were co-eluted with RNase E701-FLAG (Fig. 3A, lanes 2 and 7). All of the FLAG-tagged truncated RNase E proteins including RNase E844-FLAG no longer bind to PNPase (Fig. 3A, lanes 3–7). This is as expected because the PNPase binding site is located at the very C-terminal end. The pull-down assay also demonstrated that truncated RNase E proteins except RNase E844-FLAG no longer bind to enolase (Fig. 3A, lanes 3–7). On the other hand, RNase E800-FLAG and RNase E750-FLAG still retain the RhlB-binding ability (Fig. 3A, lanes 4 and 5). These results are consistent with the previous studies regarding the subdomains of RNase E involved in the interaction with partner proteins (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008).

Characterization of RNase E variants lacking Hfq binding sites

To examine further the role of Hfq binding sites, we constructed strains carrying rne alleles encoding RNase EΔ1-FLAG and RNase EΔ2-FLAG in which residues 702–750 and 702–844 are missing respectively (Fig. 4A). The expression of the FLAG-tagged RNase E variants in these strains was confirmed by Western blotting (data not shown). Then, we examined the effects of the internal deletions of RNase E on the SgrS-mediated destabilization of ptsG mRNA. Cells expressing RNase E-FLAG, RNase E701-FLAG, RNase EΔ1-FLAG or RNase EΔ2-FLAG were grown in LB medium, exposed to either Glc or αMG for 10 min, and then total RNAs were analysed by Northern blotting using sgrS and ptsG probes. SgrS was induced in all strains when cells were exposed to αMG (Fig. 4B). We confirmed again that the ptsG mRNA was dramatically destabilized in RNase E-FLAG- but not RNase E701-FLAG-expressing cells in response to αMG (Fig. 4B, lanes 2 and 4). We newly found that the degradation of ptsG mRNA in response to αMG exposure occurred only partially in cells expressing RNase EΔ1-FLAG (Fig. 4B, lane 6). The degradation of ptsG mRNA no longer occurred at all in cells expressing RNase EΔ2-FLAG (Fig. 4B, lane 8).

Figure 4.

A. Schematic representation of RNase E variants possessing internal deletions. RNase EΔ1-FLAG and RNase EΔ2-FLAG are expressed from modified chromosomal rne alleles. The residues 702–750 and 702–844 are missing in RNase EΔ1-FLAG and RNase EΔ2-FLAG respectively. The shaded boxes represent the FLAG polypeptide.
B. Effects of internal deletions of RNase E on the expression of ptsG mRNA. Strains indicated were grown in LB medium. At A600 = 0.6, the culture was split, 0.2% Glc or αMG was added to each culture and incubation was continued for 10 min. Total RNAs were prepared and 15 µg or 5 µg of each RNA sample was subjected to Northern blot analysis using ptsG or sgrS probe respectively.
C. Physical interaction of RNase E variants with Hfq and other partner proteins. Strains indicated were grown in 200 ml of LB medium to A600 = 0.6. Cell extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose as described in Experimental procedures. Proteins bound to the beads were analysed by Western blotting. Anti-FLAG, anti-Hfq, anti-RhlB, anti-enolase and anti-PNPase antibodies were used to detect FLAG-tagged RNase E proteins, Hfq, RhlB, enolase and PNPase respectively. Following amounts of bound fractions were subjected to the gel electrophoresis: RNase E (3 µl), Hfq (5 µl) and other proteins (3 µl).

We also tested for the binding of Hfq to RNase EΔ1-FLAG and RNase EΔ2-FLAG by pull-down assay. The crude extract was prepared from cells expressing RNase EΔ1-FLAG or RNase EΔ2-FLAG. Each extract was incubated with anti-FLAG M2-agarose beads and proteins bound were eluted from the beads. Proteins in the bound fraction were analysed by Western blotting. RNase EΔ1-FLAG but not RNase EΔ2-FLAG retains the ability to bind to Hfq (Fig. 4C, lanes 3 and 4). These results are consistent with the conclusions derived from the analysis of the C-terminal deletion mutants. Namely, the subregion between 711 and 750 of RNase E is crucial for the functional interaction with Hfq to cause the sRNA-mediated rapid degradation of target mRNAs although Hfq binds to the scaffold region through several different sites. The data also indicate that the region between 845 and 1061 alone is not sufficient to bind to Hfq. As expected, RNase EΔ1-FLAG binds to enolase and PNPase but not to RhlB while RNase EΔ2-FLAG binds to PNPase but not to RhlB and enolase (Fig. 4C, lanes 3 and 4).

A polypeptide corresponding to the scaffold region can bind to Hfq

The pull-down assays along with analyses of sRNA-mediated destabilization of mRNAs showed convincingly that the Hfq binding to the subregion between residues 711–750 is responsible for the RNase E recruitment to target mRNAs. However, these experiments do not exclude the possibility that the N-terminal catalytic region and/or the central RNA-binding domain of RNase E are also required for the effective interaction of RNase E with Hfq. To examine this possibility, a DNA fragment encoding a polypeptide corresponding to residues 702–844 of RNase E possessing a C-terminally fused FLAG tag was prepared by PCR and cloned in pQE80L to construct pQE-His-702-844-FLAG. The resulting plasmid was introduced in the wild-type cells and the 702-844 polypeptide possessing FLAG tag at its C-terminus and His6 tag at its N-terminus was expressed by adding IPTG (Fig. 5A, lane 3). A distinct band of smaller molecular weight that is apparently a degradation product of His6-702-844-FLAG was also detected by anti-FLAG probed Western analysis (shown by an asterisk in Fig. 5A). Similarly, plasmid pQE-His-711-844-FLAG was constructed and His6-711-844-FLAG polypeptide was expressed (Fig. 5A, lane 2). The cell extracts were prepared and subjected to the pull-down assay using anti-FLAG M2-agarose beads. Proteins in the bound fractions along with those in the crude extracts were analysed by Western blotting (Fig. 5A, lanes 4–6). The anti-FLAG probed Western blot revealed that the double-tagged polypeptides were efficiently recovered in the bound fraction (Fig. 5A, lanes 5 and 6). The anti-Hfq probed Western blot revealed that Hfq co-purified with His6-702-844-FLAG and His6-711-844-FLAG polypeptides (Fig. 5A, lanes 5 and 6). Thus, Hfq is able to bind the C-terminal scaffold region of RNase E without the N-terminal region including the central RNA-binding domain. As expected, RhlB also co-purified with the double-tagged polypeptides (Fig. 5A, lanes 5 and 6). The interaction between the scaffold polypeptide and Hfq was not affected when the extracts were treated with micrococcal nuclease prior to the pull-down assay (Fig. 5B).

Figure 5.

Physical interaction of scaffold polypeptide with Hfq.
A. IT1568 cells harbouring pQE80L, pQE-His711-844-FLAG or pQE-His702-844-FLAG were grown in 200 ml of LB medium containing 1 mM IPTG to A600 = 0.6. Crude extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose as described in Experimental procedures. Crude extracts (CE) and bound fractions (B) were analysed by Western blotting. Following amounts of crude extracts and bound fractions were subjected to the gel electrophoresis: scaffold polypeptides (3 µl), Hfq (5 µl) and RhlB (5 µl). Anti-FLAG, anti-Hfq and anti-RhlB antibodies were used to detect the double-tagged scaffold polypeptides, Hfq and RhlB respectively. The asterisk indicates a degradation product derived from the C-terminal portion of the scaffold polypeptides.
B. The crude extracts were treated with micrococal nuclease prior to the pull-down assay. The asterisk indicates a degradation product derived from the C-terminal portion of His6-711-844-FLAG.

Purified scaffold polypeptide binds to purified Hfq in vitro

The results mentioned above along with our previous study suggest that Hfq interacts with RNase E without the help of RNA. On the other hand, a recent in vitro study suggested that the interaction between RNase E and Hfq is mediated by RNA (Worrall et al., 2008). Then, we tried to examine the interaction between purified scaffold polypeptide and purified Hfq in vitro. For this, pQE-His-711-844-FLAG and pQE-Hfq-His (Kawamoto et al., 2006) were introduced separately in ΔhfqΔrhlB cells and crude extracts were prepared. As expected, the double-tagged 711–844 polypeptide and C-terminally His-tagged Hfq were detected in the extracts by Western blotting (Fig. 6A, lanes 2 and 3). To degrade RNAs, the extracts were treated with RNase A and then heated. It is known that the heat treatment is useful for the purification of heat-stable Hfq (Brescia et al., 2003) because most proteins become insoluble without affecting Hfq activity. We also expected that the scaffold polypeptide would be tolerant of heat because it is essentially unstructured (Carpousis, 2007). In fact, the heat treatment only slightly affected the stability of His6-711-844-FLAG polypeptide while it did not affect the stability of Hfq-His6 at all (Fig. 6A, lanes 5 and 6). In addition, the heating may permit efficient digestion of bulk RNA by melting RNA secondary structure. Then, His6-711-844-FLAG polypeptide and Hfq-His6 were purified by using Ni2+-NTA agarose resin (Fig. 6A, lanes 8 and 9). The purity of Hfq-His6 was estimated to be about 80% by SDS-PAGE and Coomassie Brilliant Blue staining while it was difficult to estimate the purity of His6-702-844-FLAG due to the presence of a non-specific band of the same mobility on the gel. In addition, it should be noted that the affinity purification by Ni2+-NTA agarose removes the degradation product, suggesting that it is derived from the C-terminal portion of His6-711-844-FLAG.

Figure 6.

Binding between scaffold polypeptide and Hfq in vitro.
A. Western blotting of crude extracts and affinity purified His-tagged proteins. Crude extracts and RNase A/heat-treated crude extracts were prepared from TM777 cells harbouring pQE80L, pQE-His711-844-FLAG or pQE-Hfq-His as described in Experimental procedures. Affinity purification of His6-711-844-FLAG and Hfq-His6 are also described in Experimental procedures. Following amounts of crude extracts and affinity purified proteins were subjected to the gel electrophoresis: His6-711-844-FLAG (0.5 µl) and Hfq-His6 (0.1 µl). Anti-FLAG and anti-Hfq antibodies were used to detect His6-711-844-FLAG and Hfq-His6. The asterisk indicates a degradation product derived from the C-terminal portion of His6-711-844-FLAG.
B. Physical interaction between Hfq and His6-711-844-FLAG in vitro. Binding between the affinity-purified His6-711-844-FLAG and Hfq-His6 was tested by the pull-down assay using anti-FLAG M2-agarose as described in Experimental procedures. Following amounts of the bound fractions were subjected to the gel electrophoresis: His6-711-844-FLAG (2.5 µl) and Hfq-His6 (5 µl).

The purified Hfq-His6 was mixed with the purified His6-711-844-FLAG polypeptide and the mixture was incubated with anti-FLAG M2-agarose beads. Proteins bound to anti-FLAG M2-agarose were eluted and analysed by Western blotting. A significant amount of Hfq-His6 was recovered in the bound fraction along with His6-711-844-FLAG polypeptide (Fig. 6B, lane 2). When the purified Hfq-His alone was subjected to anti-FLAG M2-agarose, no Hfq-His6 was recovered in the bound fraction (Fig. 6B, lane 1). Thus, the affinity purified His6-711-844-FLAG polypeptide is able to bind to the affinity purified Hfq-His6in vitro. These results strongly suggest that Hfq interacts with the scaffold region of RNase E without the help of RNA.

Overexpression of RhlB inhibits the Hfq binding to RNase E

The functional Hfq binding site (from residues 711 to 750) responsible for the recruitment of RNase E to the target mRNAs is within the known RhlB recognition region (from residues 696 to 762). This raises the possibility that Hfq and RhlB bind to RNase E in a mutually exclusive manner. To test this, we examined the effect of overexpression of RhlB on the Hfq–RNase E interaction. For this, a plasmid pQE-His-RhlB was constructed in which the expression of His6-tagged RhlB protein is under the control of IPTG-inducible promoter. The pQE-His-RhlB was introduced in the strain TM522 carrying the rne-FLAG allele. As expected, the His6-tagged RhlB was overexpressed in the presence of IPTG without affecting the expression of other proteins (Fig. 7A, lanes 1 and 2). The extent of overexpression was estimated to be about 30-fold compared with the wild-type level. In addition, the RhlB overexpression caused a moderate inhibitory effect on cell growth (data not shown). The cell extracts were prepared and incubated with anti-FLAG M2-agarose beads. Proteins in the bound fraction were analysed by Western blotting. The anti-FLAG probed Western blot revealed that the FLAG-tagged RNase E was efficiently recovered in the bound fraction (Fig. 7A, lanes 3 and 4). Then, the proteins bound to the anti-FLAG M2-agarose beads were analysed by Western blotting using anti-RhlB and anti-Hfq antibodies. The amount of Hfq in the bound fraction was clearly reduced when RhlB was overexpressed while the amount of RhlB in the bound fraction slightly increased (Fig. 7A, lane 4). Thus, RhlB overexpression significantly inhibits the Hfq–RNase E interaction, suggesting that Hfq and RhlB bind to RNase E in a mutually exclusive manner. However, it is also possible that RhlB overexpression blocks the Hfq–RNase E interaction by other ways because the amount of RhlB in the bound fraction increased only slightly.

Figure 7.

Effects of RhlB overexpression.
A. Effect on RNase E–Hfq interaction. TM522 cells harbouring pQE-His-RhlB or pQE80L were grown in LB medium containing 1 mM IPTG to A600 = 0.6. Crude extracts were prepared and subjected to the pull-down assay using anti-FLAG agarose. Following amounts of crude extracts (CE) and bound fractions (B) were subjected to the gel electrophoresis and Western blotting: RNase E (3 µl), Hfq (5 µl) and other proteins (2 µl). Anti-FLAG, anti-Hfq, anti-RhlB, anti-enolase and anti-PNPase antibodies were used to detect FLAG-tagged RNase E, Hfq, RhlB, enolase and PNPase respectively.
B. Effect on RNase E-dependent destabilization of ptsG mRNA mediated by SgrS. IT1568 cells harbouring pQE-His-RhlB were grown in LB medium in the presence and absence of 1 mM IPTG. At A600 = 0.6, the culture was split, 0.1% Glc or αMG was added to each culture and incubation was continued for 10 min. Total RNAs were prepared and 15 µg or 5 µg of each RNA sample was subjected to Northern blot analysis using ptsG or sgrS probe respectively.

Effect of RhlB overexpression on the RNase E-dependent destabilization of target mRNA mediated by SgrS

It is expected that overexpression of RhlB affects the rapid degradation of target mRNAs mediated by sRNAs due to the reduced Hfq binding to RNase E. To test whether this is the case, we analysed first the effect of RhlB overexpression on the rapid degradation of ptsG mRNA mediated by SgrS. Wild-type cells harbouring pQE-His-RhlB were grown in the presence and absence of IPTG in LB medium, exposed to either Glc or αMG for 10 min, and then total RNAs were analysed by Northern blotting by using the sgrS and ptsG probes. SgrS is induced resulting in the rapid degradation of ptsG mRNA in the absence of IPTG when exposed to αMG as expected (Fig. 7B, lane 2). On the other hand, the degradation of ptsG mRNA in response to αMG was significantly inhibited in the presence of IPTG although SgrS was induced normally (Fig. 7B, lane 4). Thus, overexpression of RhlB partially suppresses the RNase E-dependent rapid degradation of ptsG mRNA. This is consistent with the finding that RhlB overexpression significantly reduces Hfq binding to RNase E.

Discussion

An abundant RNA-binding protein Hfq and Hfq-binding sRNAs are newly identified partners of RNase E (Morita et al., 2005). The ribonucleoprotein complexes consisting of RNase E, Hfq and each specific sRNA such as SgrS and RyhB act on the target mRNAs through base pairing to cause translational inhibition and RNase E-dependent coupled degradation of mRNAs and sRNAs under specific stress conditions (Masse et al., 2003; Morita et al., 2005). The role of the Hfq–RNase E interaction is to recruit RNase E to target mRNAs of sRNAs and to support the rapid degradation of the mRNA–sRNA hybrid by RNase E (Morita et al., 2006). It is well established that the C-terminal scaffold region of RNase E provides the recruitment sites for degradosome components through small recognition motifs (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008). Our previous study demonstrated that Hfq also binds to the C-terminal scaffold region of RNase E (Morita et al., 2005). In the present study, we explored further the physical and functional interaction sites of Hfq within the scaffold region of RNase E. We constructed strains in which the chromosomal rne gene was manipulated to encode a series of C-terminally truncated FLAG-tagged RNase E proteins. Using these strains, we first found that C-terminal truncation of RNase E up to residue 750 does not affect the ability to cause the rapid degradation of ptsG and sodB mRNAs mediated by sRNAs (Fig. 2). Then, we showed that RNase E844, RNase E800, RNase E750 but not RNase E710 retains the ability to bind Hfq in both non-stress and stress conditions (Fig. 3). We also constructed strains carrying rne alleles encoding RNase EΔ1-FLAG and RNase EΔ2-FLAG in which 711–750 and 711–844 regions are deleted respectively. These internal deletions eliminated markedly (RNase EΔ1-FLAG) or completely (RNase EΔ2-FLAG) the ability to cause SgrS-mediated destabilization of ptsG mRNA (Fig. 4). It should be noted that RNase EΔ1-FLAG loses greatly the ability to degrade the ptsG mRNA while it retains fairly well the Hfq-binding activity. Taken together, we conclude that the subregion between residues 711 and 750 of RNase E is responsible for the functional Hfq binding site that is necessary for RNase E recruitment to target mRNAs mediated by sRNAs. The residual activity of RNase EΔ1-FLAG to degrade the ptsG mRNA suggests that the Hfq binding at the region 801–844 partially contributes to the functional RNase E recruitment at least when the 711–750 region is deleted. The identified Hfq binding sites within the C-terminal scaffold region of RNase E are schematically shown in Fig. 8 along with binding sites of other proteins. The functional Hfq binding site is within or near the RhlB binding site (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008). The fact that overexpression of RhlB eliminates the Hfq binding to RNase E resulting in partial suppression of the rapid degradation of target mRNAs is consistent with this conclusion (Fig. 7). This is also consistent with our previous finding that Hfq and RhlB are not able to bind RNase E simultaneously (Morita et al., 2005). Although our data suggest that Hfq and RhlB bind to RNase E in a mutually exclusive manner, exactly how RhlB overexpression eliminates Hfq binding to RNase E and how partner proteins interact with the scaffold region of RNase E to generate various forms of degradosome are important unsettled questions.

Figure 8.

Hfq binding sites and summary of organization of RNase E. Several different sites within the scaffold region of RNase E are involved in the binding to Hfq. The Hfq binding at residues 711–750, which falls on the RhlB binding site, is responsible for the sRNA-mediated rapid degradation of target mRNAs. Additional Hfq-binding regions (residues 801–844 and 845–1021) are shown by a dashed line. The possible role of Hfq bindings at residues 801–844 and at residues 845–1021 is to stabilize the functional interaction between residues 711–750 and Hfq.

Another finding is that the regions between 845 and 1061 and 801–844 of RNase E are also involved in additional physical interactions with Hfq (Figs 2 and 3) although the internal deletion analysis indicates that the region between 845 and 1061 alone does not bind to Hfq stably. It should be noted that the 801–844 region contains the second arginine-rich RNA-binding domain AR2 (Marcaida et al., 2006; Carpousis, 2007; Worrall et al., 2008). It is possible that the Hfq binding in the 801–844 region is mediated by RNA that binds to AR2 as proposed recently (Worrall et al., 2008). However, we would like to emphasize that the AR2-mediated Hfq binding does not contribute to the functional RNase E–Hfq interaction that is responsible for the sRNA-mediated degradation of target mRNAs. We believe that the role of Hfq bindings at the regions 801–844 and 845–1061 of RNase E is to stabilize the functional interaction between residues 711–750 and Hfq. In addition, we showed that RhlB overexpression inhibits the Hfq binding to RNase E not only at the 711–750 region but also at the 801–844 and 845–1061 regions (Fig. 7A). It would be interesting to investigate how the RhlB–RNase E interaction affects the Hfq binding to RNase E at the region outside of the RhlB binding site.

We showed previously (Morita et al., 2005) and in this study (Fig. 5B) that Hfq remains bound to RNase E or the scaffold polypeptide even when the cell extracts containing two proteins are treated with micrococcal nuclease, suggesting that Hfq may interact with RNase E without the help of RNA. On the other hand, Worrall et al. examined the interaction between purified Hfq and purified segments of RNase E containing the C-terminal scaffold region in vitro by cross-linking, non-denaturing PAGE and isothermal titration calorimetry (Worrall et al., 2008). They detected the interaction of two proteins in the presence but not absence of RNA. Based on these results, they proposed that the interaction between RNase E and Hfq may be mediated by RNAs presumably through two RNA-binding domains of RNase E, the central RNA-binding domain (RBD) corresponding to residues 604–688 and the second arginine-rich domain (AR2) corresponding to residues 798–819.

How do the results of two studies reconcile each other? One possible explanation would be that micrococcal nuclease is unable to digest a tightly bound RNA that maintained an indirect RNase E–Hfq interaction as proposed (Worrall et al., 2008). However, we observed that treatment of the cell extracts not only by micrococcal nuclease but also by RNase A, a more ‘efficient’ ribonuclease, did not affect the interaction between Hfq and RNase E (data not shown). Thus, the nuclease-treatment experiments did not produce any evidence that RNA stimulates and/or mediates the Hfq–RNase E interaction. The finding that the truncated RNase E800 and RNase E750 lacking AR2 still retains the ability to bind Hfq resulting in the sRNA-mediated degradation of target mRNAs implies that AR2, presumed to be responsible for RNA-mediated Hfq–RNase E interaction, is not needed for the functional interaction between RNase E and Hfq. In addition, it is less likely that the central RNA-binding domain is involved in Hfq–RNase E interaction because we showed that a polypeptide corresponding to the scaffold region without this region can bind efficiently to Hfq (Fig. 5). Finally, we showed that the purified scaffold polypeptide stably binds to purified Hfq in vitro (Fig. 6). Note that the extracts were treated with RNase A and then with heat prior to the affinity purification. These treatments remove RNAs at least to undetectable level judging from ethidium bromide staining. Together, it is highly possible that Hfq directly interacts with the scaffold region of RNase E without the help of RNA. Nevertheless, our experiments still do not exclude completely the possibility that tiny amounts of RNA fragments remain to bind tightly either the purified Hfq or scaffold polypeptide to mediate the Hfq–RNase E interaction.

Experimental procedures

Media and growth conditions

Cells were grown aerobically at 37°C in LB medium supplemented with indicated compounds. Antibiotics were used at the following concentrations when needed: ampicillin (50 µg ml−1) and chloramphenicol (15 µg ml−1). Bacterial growth was monitored by determining the optical density at 600 nm.

Bacterial strains and plasmids

The E. coli K12 strains and plasmids used in this study are listed in Table 1. IT1568 (W3110 mlc) was used as a parent wild-type strain. TM338 (rne-FLAG-cat), TM527 (rne844-FLAG-cat) and TM528 (rne701-FLAG-cat) were described previously (Morita et al., 2004). MY02 (rne800-FLAG-cat), TM719 (rne750-FLAG-cat) and MY12 (rne710-FLAG-cat) were constructed from IT1568 according to the modified Datsenko–Wanner protocol using pSU313 harbouring the FLAG-tag sequence (Uzzau et al., 2001). TM782 (rneΔ1-FLAG-cat) and TM783 (rneΔ2-FLAG-cat) encoding RNase E variants lacking the residue 702–750 and 702–844, respectively, were constructed as follows. First, TM774 (rne701-HA) was constructed by removing the cat gene flanked by two FRT sequences from TM642 (rne701-HA-cat). Then, a DNA fragment 1 corresponding to residue 751–1061 of RNase E followed by FLAG-cat was amplified from chromosomal DNA of TM338 (nre-FLAG-cat) by PCR with primers 1219 (AAACGTCAGGCGCAACAAGAAGCGAAGGCGCTGAATGTTGAAGAGGCCGAACCAATTGTTCAGGA) and 1223 (ATGTTTTGTCTGCCTGCTCTGGGATCGCTGGGGCGGGCATTTTTTTGCCT). The 5′ portion (underlined) of primer 1219 matches exactly to the 3′ end of rne701 while primer 1223 is derived from the sequence downstream of FRT-cat-FRT in TM338. The fragment 1 was introduced into TM774 harbouring pKD46 and colonies resistant to chloramphenicol were selected. A strain harbouring rneΔ1-FLAG-cat allele was generated by homologous recombination between the fragment 1 and TM774. This was confirmed by DNA sequencing and the resulting rneΔ1-FLAG-cat allele was moved to TM338 to construct TM782. Similarly, DNA fragment 2 corresponding to the region from residue 845–1061 of RNase E followed by FLAG-cat was amplified from chromosomal DNA of TM338 (rne-FLAG-cat) by PCR with primers 1220 (AAACGTCAGGCGCAACAAGAAGCGAAGGCGCTGAATGTTGAAGAGCGCTATCCAATTGTACGTCC) and 1223. The fragment 2 was introduced into TM774 harbouring pKD46 (Datsenko and Wanner, 2000) to obtain a strain harbouring rneΔ2-FLAG-cat allele. The rneΔ2-FLAG-cat allele was moved to TM338 to construct TM783. TM675 (rne844-FLAG) was constructed by removing the cat gene from TM527 (rne844-FLAG-cat). TM777 (ΔrhlBΔhfq::cat) was constructed as follows. First, TM726 (ΔrhlB) was constructed by removing the cat gene from TM390 (ΔrhlB::cat) (Morita et al., 2004). Then, the Δhfq::cat allele of TM587 (Morita et al., 2005) was moved to TM726.

Table 1.  Bacterial strains and plasmids used in this study.
Strain/plasmidRelevant genotype and propertySource
Strain  
 IT1568W3110mlcLaboratory stock
 TM338W3110mlc rne-FLAG-catMorita et al. (2004)
 TM528W3110mlc rne701-FLAG-catMorita et al. (2004)
 TM527W3110mlc rne844-FLAG-catMorita et al. (2004)
 YM02W3110mlc rne800-FLAG-catThis study
 TM719W3110mlc rne750-FLAG-catThis study
 YM12W3110mlc rne710-FLAG-catThis study
 TM522W3110mlc rne-FLAGMorita et al. (2005)
 TM642W3110mlc rne701-HA-catMorita et al. (2005)
 TM774W3110mlc rne701-HAThis study
 TM782W3110mlc rneΔ1(Δ702-750)-FLAG-catThis study
 TM783W3110mlc rneΔ2(Δ702-844)-FLAG-catThis study
 TM675W3110mlc rne844-FLAGThis study
 TM390W3110mlcΔrhlB::catMorita et al. (2004)
 TM726W3110mlcΔrhlBThis study
 TM587W3110mlcΔhfq::catMorita et al. (2005)
 TM777W3110mlcΔrhlBΔhfq::catThis study
Plasmid  
 pQE80LVectorQiagen
 pQE-Hfq-HisDerivative of pQE80L carrying hfq-His6Kawamoto et al. (2006)
 pQE-His-711-844-FLAGDerivative of pQE80L carrying His6-rne711-844-FLAGThis study
 pQE-His-702-844-FLAGDerivative of pQE80L carrying His6-rne702-844-FLAGThis study
 pQE-His-RhlBDerivative of pQE80L carrying His6-rhlBThis study
 pKD46Red recombinase expression vectorDatsenko and Wanner (2000)

Plasmid pQE-His-702-844-FLAG was constructed as follows. A DNA fragment corresponding to residue 702–844 of RNase E was amplified from chromosomal DNA of TM675 (rne844-FLAG) by PCR with the primers 1225 (GCGCGGATCCCAATCTGTTCAGGAAACCGA) and 796 (GCGCGTCGACGCTCTGGGATCGCTGGGGCGGGCAT). The amplified fragment was digested with BamHI and SalI and cloned into pQE80L (Qiagen) to construct pQE-His-702-844-FLAG. Similarly, a DNA fragment corresponding to residue 711–844 of RNase E was amplified from chromosomal DNA of TM675 (rne844-FLAG) by PCR with the primers 1224 (GCGCGGATCCGAACGTGTACGTCCGGTTCA) and 796 (GCGCGTCGACGCTCTGGGATCGCTGGGGCGGGCAT). The amplified fragment was digested with BamHI and SalI and cloned into pQE80L to construct plasmid pQE-His-711-844-FLAG. Plasmid pQE-His-RhlB was constructed as follows. The fragment containing the coding region of rhlB was amplified from chromosomal DNA of IT1568 by PCR with primers 923 (CGCGCGGATCCAGCAAAACACATTTAAC) and 924 (CCCAAGCTTACCAGCATATGAAAAC). The amplified fragment was digested BamHI and HindIII and cloned into pQE80L.

Pull-down assay

Cells were grown in 200 ml of LB medium to A600 of 0.6, harvested and washed with 10 ml of STE buffer (100 mM NaCl, 10 mM Tris-HCl pH 8.0, 1 mM EDTA). The cell pellets were suspended in ice-cold 10 ml of IP buffer (20 mM Tris-HCl pH 8.0, 0.25 M KCl, 5 mM MgCl2, 10% glycerol, 0.1% Tween20). The cell suspension was sonicated and centrifuged at 10 000 r.p.m. for 30 min at 4°C. The supernatant (crude extract) was incubated with 50 µl of anti-FLAG M2-agarose suspension (Sigma) for 30 min at 4°C. When indicated, the supernatant was incubated with 40 unit of micrococal nuclease (Takara) in the presence of 2.5 mM CaCl2 for 10 min at 37°C prior to the incubation with the M2-agarose beads. The mixture was filtered by using a mini chromatography column (Bio-Rad). The agarose beads were washed by 10 ml of IP buffer two times. The proteins bound to the beads were eluted with 50 µl of IP buffer containing 0.4 mg ml−1 FLAG peptide (Sigma) and used as bound fraction (B). The samples were analysed by Western blotting. To analyse RNAs, 10 µl of the bound fraction was treated with phenol, precipitated and washed with ethanol. The precipitate was dissolved in 10 µl of RNA buffer (0.02 M sodium acetate, pH 5.5, 0.5% SDS and 1 mM EDTA). Three microlitres of RNA sample was subjected to Northern blotting.

Western blotting

The sample was heated at 100°C for 5 min and subjected to a polyacrylamide-0.1% SDS gel electrophoresis and transferred to Immobilon membrane (Millipore). The 8% and 15% polyacrylamide gels were used to detect RNase E and Hfq, respectively, while the 12% gel was used for other proteins. The membranes were treated with anti-FLAG monoclonal antibody (Sigma), anti-Hfq, anti-RhlB, anti-PNPase and anti-enolase, polyclonal antibodies. Signals were visualized by the Lumi-Light Western Blotting Substrate (Roche). Polyclonal anti-Hfq and anti-enolase antibodies were described previously (Morita et al., 2004). Polyclonal anti-RhlB and anti-PNPase prepared by immunizing rabbits with synthetic polypeptides corresponding to RhlB and PNPase, respectively, were obtained from Operon biotechnologies.

Northern blotting

For Northern blot analysis, RNA samples were resolved by 1.2% agarose gel electrophoresis in the presence of formaldehyde and blotted on to Hybond-N+ membrane (Amersham Biosciences). The mRNAs were visualized using digoxigenin (DIG) reagents and kits for non-radioactive nucleic acid labelling and detection system (Roche Molecular Biochemicals) according to the procedure specified by the manufacturer. The following DIG-labelled DNA probes were prepared by PCR using DIG-dUTP: 305 bp fragment corresponding to the 5′ region of ptsG; 500 bp fragment corresponding to the 5′ region of sodB; 150 bp fragment corresponding to the sgrS; 87 bp fragment corresponding to the ryhB.

Analysis of binding between Hfq and scaffold polypeptide in vitro

TM777 cells harbouring pQE-Hfq-His (Kawamoto et al., 2006) or pHis-702-844-FLAG were grown in 200 ml of LB medium at 37°C. At OD600 = 0.2–0.3, 1 mM IPTG was added to the culture and incubation was continued for 60 min. Cells were harvested and washed with 10 ml of STE (100 mM NaCl, 10 mM Tris-HCl, pH 8.0, 1 mM EDTA), and suspended in 0.3 ml of buffer A (50 mM Na2H PO4-NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0). The cell suspension was treated with lysozyme (10 mg ml−1) for 10 min on ice, sonicated and centrifuged at 14 000 r.p.m. for 10 min at 4°C. The supernatant was treated with RNase A (10 µg ml−1) for 10 min on ice and heated for 10 min at 80°C. The sample was centrifuged again at 14 000 r.p.m. for 10 min at 4°C to remove aggregated proteins. The supernatant was incubated with 60 µl of Ni2+-NTA agarose resin (Qiagen) for 20 min at 4°C. The agarose resin was collected by centrifugation (14 000 r.p.m., 1 min at 4°C) and washed by 500 µl of buffer B (50 mM Na2H PO4-NaH2PO4, 300 mM NaCl, 20 mM imidazole, pH 8.0) three times. The proteins bound to the beads were eluted with 60 µl of buffer C (50 mM Na2H PO4-NaH2PO4, 300 mM NaCl, 250 mM imidazole, pH 8.0) to obtain affinity purified Hfq-His6 and His6-702-844-FLAG.

Forty microlitres of His6-702-844-FLAG polypeptide was mixed with 20 µl of Hfq-His6 or buffer C (control), and 800 µl of IP buffer 2 (20 mM Tris-HCl pH 8.0, 0.1 M KCl, 5 mM MgCl2, 10% glycerol, 0.1% Tween20) was added. The mixture was incubated with 50 µl of anti-FLAG M2-agarose suspension (Sigma) for 30 min at 4°C. The agarose beads were collected by the centrifugation (14 000 r.p.m., 1 min at 4°C). The beads were washed by 500 µl of IP buffer 2 three times. The proteins bound to the beads were eluted with 50 µl of IP buffer 2 containing 0.4 mg ml−1 FLAG peptide (Sigma) and analysed by Western blotting.

Acknowledgements

This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to H.A., T.M. and Y.I.) and Uehara Memorial Foundation (to T.M.).

Ancillary