Members of the RNase III family are the primary cellular agents of dsRNA (double-stranded RNA) processing. Bacterial RNases III function as homodimers and contain two dsRBDs (dsRNA-binding domains) and two catalytic sites. The potential for functional cross-talk between the catalytic sites and the requirement for both dsRBDs for processing activity are not known. It is shown that an Escherichia coli RNase III heterodimer that contains a single functional wt (wild-type) catalytic site and an inactive catalytic site (RNase III[E117A/wt]) cleaves a substrate with a single scissile bond with a kcat value that is one-half that of wt RNase III, but exhibits an unaltered Km. Moreover, RNase III[E117A/wt] cleavage of a substrate containing two scissile bonds generates singly cleaved intermediates that are only slowly cleaved at the remaining phosphodiester linkage, and in a manner that is sensitive to excess unlabelled substrate. These results demonstrate the equal probability, during a single binding event, of placement of a scissile bond in a functional or nonfunctional catalytic site of the heterodimer and reveal a requirement for substrate dissociation and rebinding for cleavage of both phosphodiester linkages by the mutant heterodimer. The rate of phosphodiester hydrolysis by RNase III[E117A/wt] has the same dependence on Mg2+ ion concentration as that of the wt enzyme, and exhibits a Hill coefficient (h) of 2.0±0.1, indicating that the metal ion dependence essentially reflects a single catalytic site that employs a two-Mg2+-ion mechanism. Whereas an E. coli RNase III mutant that lacks both dsRBDs is inactive, a heterodimer that contains a single dsRBD exhibits significant catalytic activity. These findings support a reaction pathway involving the largely independent action of the dsRBDs and the catalytic sites in substrate recognition and cleavage respectively.
The enzymatic cleavage of dsRNA (double-stranded RNA) is an essential event in the maturation and decay of diverse eukaryotic and bacterial RNAs. Members of the RNase III family are the primary agents of dsRNA processing and are highly conserved in eukaryotic and bacterial cells [1–8]. Bacterial RNases III participate in rRNA and mRNA maturation and also initiate mRNA decay [1,2,5]. The eukaryotic RNase III orthologues Dicer and Drosha are involved in microRNA maturation [9–12], and RNase III polypeptides are essential components of multisubunit RNA editosomes that catalyse uridine insertion or deletion in trypanosomatid kinetoplast mRNAs [13,14]. Several viruses express RNase III orthologues that may be involved in antagonizing the RNAi (RNA interference)-based antiviral response [15,16]. RNase III family members can be grouped into three classes, according to polypeptide structure (Figure 1A). The Class 3 enzymes include Dicer, whereas Class 2 enzymes are represented by Drosha. Members of the structurally simplest Class 1 include the bacterial RNases III, which typically contain ∼220 amino acid residues. The N-terminal portion of bacterial RNase III polypeptides includes the NucD (nuclease domain), which self-associates to form a single ‘processing centre’ at the subunit interface that contains two catalytic sites [17,18]. The C-terminal region of the bacterial RNase III polypeptide consists of a dsRBD (dsRNA-binding domain) that carries a single copy of the conserved dsRBM (dsRNA-binding motif) [19–21]. As the Class 1 RNase III polypeptides form stable dimers [1,3,5], the bacterial RNase III holoenzyme contains two dsRBDs (Figures 1B and 1C). Ji and co-workers solved the structure of Aquifex aeolicus RNase III and specific mutant versions, either bound to dsRNA in catalytically non-productive modes [22,23] or bound to a cleaved minimal substrate . These studies revealed the conformational flexibility of the short linker that connects the dsRBD and NucD (Figure 1C) and demonstrated the involvement of conserved NucD residues in bivalent metal cation binding at the catalytic site, as well as in subunit association.
RNase III family members
RNase III family members are phosphodiesterases and require a bivalent metal ion to hydrolyse phosphodiester linkages, creating 5′-phosphate, 3′-hydroxyl product termini [25,26]. For bacterial RNases III, Mg2+ is presumably the physiologically relevant species. However, several other metals, including Mn2+, Co2+ and Ni2+, can support catalysis [26–28]. Ongoing biochemical and structural studies are revealing the basic features of the bacterial RNase III catalytic mechanism. Each subunit contains a catalytic site that is responsible for cleaving one of the two phosphodiesters at a dsRNA target site. Kinetic and inhibitor studies indicate a two-metal-ion catalytic mechanism for Escherichia coli RNase III . Since only one metal ion is observed in each catalytic site in a crystal structure of A. aeolicus RNase III [17,24], it has been proposed that substrate binding promotes the binding of the second metal ion . Structural comparisons of A. aeolicus RNase III with Bacillus halodurans RNase H  and a crystal structure of Giardia intestinalis Dicer with bound Europium (Eu3+) ions  have provided a tentative placement of the second metal-binding site, approx. 4 Å (1 Å=0.1 nm) from the other metal, in a manner consistent with a two metal ion mechanism. Several conserved side chains of one subunit are near the catalytic site of the other subunit. Thus the possibility arises that intersubunit interactions may be important for optimal catalytic site function and perhaps also for regulation of activity. Although the mutational inactivation of one catalytic site does not suppress the function of the other catalytic site , certain mutations may confer more subtle changes on the pathway and kinetics of dsRNA cleavage.
The importance of the dsRBD in bacterial RNase III function was demonstrated by the in vitro catalytic inactivity of an E. coli RNase III mutant that lacks both copies of the domain . The absence of activity reflected an inability to bind the substrate . Examination of the A. aeolicus RNase III–dsRNA crystal structures reveals a primary role of the dsRBD in substrate recognition, in that the domain engages in multiple contacts with both strands of dsRNA, and with most of the buried surface area of the substrate involving both dsRBDs . However, it is not known whether both dsRBDs are needed for catalytic activity or whether the dsRBDs function in a co-operative manner in substrate recognition. The present study addresses these questions, as well as examining subunit cross-talk through analysis of the biochemical behaviours of specific mutant heterodimers of E. coli RNase III.
Water was deionized and distilled. Chemicals and reagents were molecular biology grade and were purchased from Sigma–Aldrich or Fisher Scientific. Standardized 1 M solutions of MgCl2 and MnCl2 were obtained from Sigma–Aldrich. Ribonucleoside 5′-triphosphates were obtained from Amersham Biosciences. [γ-32P]ATP (3000 Ci/mmol) and [α-32P]UTP (3000 Ci/mmol) were purchased from PerkinElmer. E. coli bulk stripped tRNA was purchased from Sigma–Aldrich and was further purified by repeated phenol extraction followed by ethanol precipitation. T4 polynucleotide kinase was purchased from New England Biolabs, whereas calf intestine alkaline phosphatase was obtained from Roche Molecular Biochemicals. Bacteriophage T7 RNA polymerase was purified in-house as described previously . Oligodeoxynucleotide transcription templates and mutagenic oligonucleotides were synthesized by Invitrogen, and the deprotected DNAs were purified by denaturing gel electrophoresis . Purified DNAs were stored at −80 °C in 10 mM Tris/HCl and 1 mM EDTA (pH 8).
RNase III heterodimer production
RNase III heterodimers were purified from an E. coli strain harbouring separate compatible plasmids that contained RNase III (rnc) genes under control of a T7 promoter. Plasmid pET-15b(rnc) was a recombinant form of pET-15b (Novagen) with a pBR322 replication origin and a copy number of ∼30, whereas pACYC(rnc) was a recombinant form of pACYC184 (New England Biolabs), with a p15A replication origin and a copy number of ∼12. Mutant rnc genes were cloned into pET-15b as described previously , using the NdeI and BamHI sites. To create the pACYC184-rnc plasmid, the E. coli rnc gene was transferred from plasmid pET-15b(rnc) into plasmid pCAL-n (Stratagene) by using the NcoI and HindIII restriction sites. This provided a recombinant plasmid [pCAL-n(rnc)], in which the rnc gene is fused to an N-terminal CBP (calmodulin-binding peptide) coding sequence , directly downstream of an IPTG (isopropyl β-D-thiogalactoside)-inducible T7 promoter (T7p-lacO). pCAL-n(rnc) was cleaved with SphI and EcoRV, and the T7p-CBP-rnc-containing DNA fragment ligated into SphI and EcoRV-cleaved pACYC184, providing plasmid pACYC184-rnc. The construct was verified by DNA sequencing. Mutations were introduced into the pACYC184 plasmid by a two-step mutagenic PCR protocol employing a single mutagenic primer [35–37]. To create the E117A mutation, the mutagenic oligonucleotide used was 5′-ATTAATGCTGCGACGGTGTCG-3′, with the mutated base shown in bold and underlined. Correct introduction of the mutation was verified by DNA sequencing.
His6-tagged RNase III was purified as described previously . The catalytic properties of His6–RNase III are essentially the same as those of the native enzyme . RNase III heterodimers were overproduced in E. coli BLR(DE3), cells carry (Novagen) containing pET-15b(rnc) and pACYC184(rnc). BLR(DE3) cells carry the DE3 gene encoding T7 RNA polymerase under control of the IPTG-inducible lac promoter and the recA allele serving to suppress plasmid recombination. Cell growth was at 37 °C in LB (Luria–Bertani) medium (400 ml) containing chloramphenicol (40 μg/ml) and ampicillin (100 μg/ml). IPTG (1 mM) was added when cells reached an attenuance of 0.4 (D395), and the incubation continued with aeration for 4 h. Aliquots were removed and analysed for protein production by SDS/PAGE (12% gels) and staining with Coomassie Brilliant Blue R. Cells were collected by centrifugation at 3500 g for 20 min at 5 °C, and stored at −20 °C until further use. Cells were resuspended in 30 ml of Buffer A (500 mM NaCl, 5 mM imidazole and 20 mM Tris/HCl, pH 8) and subjected to repeated sonication on ice. The following steps were carried out at ∼5 °C. The sonicated material was clarified by centrifugation at 3500 g for 20 min at 5 °C. The first chromatographic step employed an Ni-NTA (Ni2+-nitrilotriacetate) column (HisBind resin; Novagen; 1.5 ml), initially washed with 10 column volumes of Buffer A. The column then was charged using an NiSO4 solution (50 mM). The clarified cell sonicate (∼30 ml) was slowly loaded on to the column, which was then washed with 20 column volumes of Buffer A and 10 column volumes of Buffer B (500 mM NaCl, 60 mM imidazole and 20 mM Tris/HCl, pH 8.0). The protein was eluted with 5 column volumes of Buffer C (1 M NaCl, 400 mM imidazole and 20 mM Tris/HCl, pH 8.0). The protein was contained mainly in the first three elute volumes. The eluates were combined and diluted 2-fold in a buffer consisting of 65 mM Tris/HCl (pH 8.0) and 3 mM CaCl2 and loaded on to a calmodulin column (Stratagene; 1 ml). Before loading, the column was prepared by washing with 10 column volumes of Buffer D (333 mM NaCl, 2 mM CaCl2 and 50 mM Tris/HCl, pH 8.0). After sample loading, the resin was washed with 20 column volumes of Buffer D, and the protein was eluted with three 1 ml aliquots of Buffer E (1 M NaCl, 2 mM EDTA and 50 mM Tris/HCl, pH 8.0). The eluted protein was dialysed overnight against a buffer consisting of 1 M NaCl, 60 mM Tris/HCl (pH 8.0), 1 mM EDTA and 1 mM DTT (dithiothreitol). The purified RNase III heterodimer preparations were stored at −20 °C in 50% (v/v) glycerol, 0.5 M NaCl, 30 mM Tris/HCl (pH 8.0), 0.5 mM EDTA and 0.5 mM DTT. The purity of the preparation was estimated by SDS/PAGE (12% gels) to be ∼90% and the preparation was free of contaminating nucleases.
The RNase III[NucD/wt] (where wt is wild-type) and RNase III]NucD/E117A] mutant heterodimers (see the Results section) were purified as follows. The inclusion body, obtained from low-speed centrifugation of the sonicated material (see above), was resuspended in 4 ml of Buffer A containing 6 M urea. The solution was clarified by centrifugation at 3500 g for 20 min at 5 °C, and the supernatant was purified on an Ni-NTA column, followed by calmodulin column chromatography as described above. The proteins purified from the soluble fraction and the inclusion body exhibited essentially the same catalytic activities. However, in order to provide accurate comparison of activity, cleavage assays involving the two mutant heterodimers also involved wt enzyme purified in the same manner. E. coli RNase III carrying a CBP tag on both subunits was active, but the protein was significantly less soluble than the His6-tagged homodimer (W. Meng and A. W. Nicholson, unpublished work). Finally, the E. coli BLR(DE3) chromosome carries the rnc gene, which is expressed. However, the endogenous RNase III polypeptide is not expected to co-purify either as a homodimer or as a heterodimer, since it lacks an affinity tag.
RNAs were enzymatically synthesized in vitro by using T7 RNA polymerase and oligodeoxynucleotide templates according to established protocols [38,39] with modifications as described previously . The sequences of the oligodeoxynucleotide transcription templates are available on request. For 5′-32Plabelling, enzymatically synthesized, non-radioactive RNA (100 pmol) was treated with calf intestine alkaline phosphatase (4 units) at 37 °C for 1 h using the supplied buffer. The RNA was purified by phenol/chloroform extraction and ethanol precipitation. Dephosphorylated RNA (∼5 pmol) was incubated at 37 °C for ∼30 min with 10 μCi of [γ-32P]ATP (3000 Ci/mmol) and T4 polynucleotide kinase (10 units), using the supplied buffer. The reaction was electrophoresed in a 15% (w/v) polyacrylamide gel containing TBE (Tris/borate/EDTA; 1×TBE=45 mM Tris/borate and 1 mM EDTA) buffer and 7 M urea, and the RNA was isolated as described previously . RNA was stored at −20 °C in Tris/EDTA buffer (10 mM Tris/HCl and 1 mM EDTA, pH 7.0).
Substrate cleavage assay
RNase III cleavage assays were performed using internally 32P-labelled RNA (U-labelled) according to an established protocol . Specific features of the assays are detailed in the appropriate Figure or Table legend. In general, assays were conducted by first incubating the RNA (specific amounts indicated in the relevant Figure legends) at 37 °C in a buffer (160 mM NaCl and 30 mM Tris/HCl, pH 8) for 5 min. RNase III was then added at the specified concentration, and the sample was incubated for 1 min at 37 °C. Mg2+ was then added at the specified concentration to initiate the reaction. Reactions were stopped by adding one-half volume of loading dye that contained 20 mM EDTA . Aliquots were analysed by electrophoresis in 15% (w/v) polyacrylamide gels containing 7 M urea and TBE buffer. The results were visualized by phosphoimaging (Typhoon 9400 system) and quantified by using ImageQuant software. Curve fitting for determination of the kinetic parameters was carried out using Kaleidagraph software (v.3.5) [29,36,37].
Substrate binding assay
Gel mobility-shift assays were performed as described previously [36,37], using 5′-32P-labelled R1.1[WC] RNA (see Figure 4A). Ca2+ (10 mM) was included in the gel shift and electrophoresis buffers. Electrophoresis was carried out at ∼5 °C in an 8% (w/v) polyacrylamide gel containing 0.5×TBE buffer and 10 mM CaCl2. Reactions were visualized by phosphoimaging and quantified using ImageQuant software. K′d values (apparent dissociation constants) were determined by curve fitting using Kaleidagraph software (v3.5) and established protocols [28,40,41].
RNase III heterodimer production involved the expression in vivo of two RNase III (rnc) genes, carried on separate, compatible plasmids. Each rnc gene was under control of an IPTG-inducible T7 promoter, with one of the encoded polypeptides fused to an N-terminal His6 tag and the other polypeptide fused to an N-terminal, 26-amino-acid CBP. Both polypeptides were produced in an IPTG-dependent manner, as revealed by SDS/PAGE analysis of total cellular protein (results not shown). The polypeptides accumulated to different levels, which in part may reflect the different plasmid copy numbers. Thus the amount of the RNase III polypeptide produced from the recombinant pET-15b plasmid (copy number ∼30) was greater than that of the polypeptide produced from the recombinant pACYC184 plasmid (copy number ∼12). Serial affinity chromatography on Ni-NTA and calmodulin columns yielded a purified preparation consisting of two polypeptides of comparable dye staining intensity, consistent with a heterodimeric form of the holoenzyme (Figure 2, lanes 3–6). While it was shown that the N-terminal His6 tag does not significantly affect RNase III catalytic activity , the effect of an N-terminal CBP tag is not known. We therefore compared the steady-state kinetic parameters of RNase III with a CBP tag on one subunit and a His6 tag on the other subunit, with RNase III containing His6 tags on both subunits. The substrate is R1.1 RNA (Figure 3A) , a 60 nt hairpin based on the bacteriophage T7 R1.1 RNase III processing signal, which contains a single primary cleavage site that is recognized in vivo and in vitro, and a secondary site that is cleaved only under specific conditions in vitro [1,27,43]. Table 1 shows that the two tagged forms of wt RNase III exhibit similar Km and kcat values with respect to cleavage of R1.1 RNA. We conclude that the N-terminal CBP tag also does not greatly alter catalytic activity. Thus the affinity tags were retained on the heterodimers in the analyses presented below, and the mutant heterodimers and wt enzymes all contained the His6/CBP affinity tag pair. For convenience, CBP/His6-tagged wt RNase III is referred to as RNase III.
Polypeptide profiles of RNase III heterodimers
Catalytic activity of RNase III[E117A/wt]
RNA processing behaviour of an RNase III heterodimer containing a single functional catalytic site
The existence of two catalytic sites in bacterial RNase III was demonstrated by the retained activity of an E. coli RNase III heterodimer in which one subunit contained the E117K mutation . E117 binds a catalytically essential Mg2+ ion , and the E117K mutation fully suppresses processing activity when present in both subunits of E. coli RNase III [44,45]. The two catalytic sites are symmetrically positioned at the subunit interface, with each site capable of cleaving one of the two bonds in a target site phosphodiester pair. Determining the kinetic behaviour and bivalent metal ion dependence of a single-catalytic-site RNase III heterodimer could reveal a functional interdependence of the catalytic sites. We prepared an E. coli RNase III heterodimer in which one subunit contained the E117A mutation. This mutation is similar to the E117K mutation, in that when the E117A mutation is present in both subunits, it suppresses catalytic activity without affecting substrate binding . The polypeptide profile of purified RNase III[E117A/wt] is shown in Figure 2 (lane 4), which reveals two closely spaced bands of comparable dye staining intensity. The difference in polypeptide size is attributable to the different affinity tags, with the CBP tag slightly larger than the His6 tag. An issue is whether the holoenzyme subunits are stable to exchange, as such a process would compromise the biochemical analyses. However, a previous study of an E. coli RNase III heterodimer provided experimental evidence against such an exchange process . This finding is also consistent with the extensive subunit interface observed in protein crystal structures [22–24], which is stabilized by multiple hydrophobic interactions and hydrogen bonds, and therefore is expected to provide a strong kinetic as well as thermodynamic barrier to dimer dissociation.
|wt/wt (His6/His6)||wt/wt (His6/CBP)||E117A/wt||NucD/wt|
|wt/wt (His6/His6)||wt/wt (His6/CBP)||E117A/wt||NucD/wt|
A representative cleavage assay using RNase III[E117A/wt] is shown in Figure 3(B), which reveals that the heterodimer is catalytically active and cleaves R1.1 RNA with the same specificity as RNase III (Figure 3B, compare lanes 2–6 with lanes 7–11). Given the asymmetry of R1.1 RNA and RNase III[E117A/wt], the binding of this RNA to the mutant heterodimer can occur in either of two ways: one binding mode would place the single scissile bond in the functional catalytic site, leading to cleavage, whereas the other binding mode would place the scissile bond in the inactive site, disallowing cleavage. Assuming comparable affinities for both binding modes, there is expected to be an equal probability of forming productive and non-productive enzyme–substrate complexes. As such, at saturating substrate concentrations, the Vmax (and thus also the kcat) for the reaction involving RNase III[E117A/wt] would be expected to be one-half that of the reaction involving RNase III, whereas the Km values would be predicted to be essentially the same. In fact, Table 1 shows that the kcat value for RNase III[E117A/wt] cleavage of R1.1 RNA is 1.2±0.3 min−1, whereas that of RNase III is 2.9±0.3 min−1. In addition, the Km values are similar (41±5 and 48±2 nM respectively).
Gel shift assays were performed to determine the affinity of RNase III[E117A/wt] for R1.1 RNA. In this assay, Mg2+ was replaced by Ca2+, which promotes substrate binding but does not support cleavage . The K′d values are provided in Table 1, where it is seen that the RNase III[E117A/wt]·R1.1[WC] RNA complex has a K′d value of 12.0±1.3 nM, which is only modestly higher than that of the RNase III·R1.1[WC] RNA complex (K′d=8.6±0.6 nM). The comparable K′d values are consistent with the comparable Km values. The kinetic and substrate binding data support a model of interaction of a single-catalytic-site heterodimer with a single-scissile-bond substrate that involves alternative productive/non-productive binding modes. In addition, the inactivation of a catalytic site does not affect the steady-state catalytic parameters above and beyond that which would be expected for the productive/non-productive binding model.
The RNase III[E117A/wt] heterodimer is also predicted to exhibit an altered manner of cleavage of substrates with two scissile bonds at the target site. If each catalytic site cleaves only one phosphodiester during a single binding event, then the release and rebinding of the singly cleaved product would be required for cleavage of the remaining bond. The release–rebinding event may be reflected in the accumulation of singly cleaved species, especially if the species exhibit a weaker affinity for enzyme than uncleaved substrate. R1.1[WC] RNA is a fully base-paired variant of R1.1 RNA and contains two scissile bonds at the target site (Figure 4A) . A time course assay using RNase III[E117A/wt] reveals the production of singly cleaved intermediates, which are slowly converted into fully cleaved products (Figure 4B, lanes 8–13). In contrast, efficient generation of the fully cleaved products occurs with RNase III, with negligible amounts of single cleaved species (Figure 4B, lanes 2–7). If the singly cleaved species are released from RNase III[wt/E117A] prior to cleavage of the second bond, then the addition of excess unlabelled substrate would be expected to block the second cleavage step. Figure 5(B) shows an experiment in which a 10-fold molar excess of unlabelled R1.1[WC] RNA was added shortly after initiation of a cleavage reaction involving RNase III[E117A/wt]. The presence of excess substrate blocks further cleavage of the intermediates, as well as preventing cleavage of unchanged 32P-labelled substrate (Figure 5B, compare lanes 7–12 with lanes 1–6). The specificity of inhibition is demonstrated by the lack of effect of a 10-fold molar excess of tRNA (Figure 5B, lanes 13–18). For RNase III, the addition of an excess of R1.1[WC] RNA also blocked further cleavage of unchanged, radiolabelled substrate, but did not cause an accumulation of singly cleaved intermediates (Figure 5A, compare lanes 8–13 with lanes 1–7). Thus, in contrast with RNase III[wt/E117A], wt RNase III efficiently cleaves both phosphodiesters of R1.1[WC] RNA during a single binding event.
RNase III[E117A/wt] processing of R1.1[WC] RNA and the generation of singly cleaved intermediates
Effect of excess unlabelled substrate on RNase III heterodimer cleavage of R1.1[WC] RNA
Bivalent metal ion dependence of RNase III[wt/E117A] catalytic activity
Kinetic, inhibitor and structural studies on RNase III and Dicer [24,29,30] indicate a two-metal-ion catalytic mechanism [47,48]. If the two-metal-ion mechanism is strictly a function of a single catalytic site, then RNase III[wt/E117A] would exhibit the same metal dependence as the wt enzyme. We measured the Mg2+ concentration dependence of RNase III[E117A/wt] cleavage of R1.1 RNA, using single-turnover conditions in which phosphodiester cleavage is the rate-limiting step [26,29]. The measured Hill coefficient (h) is 2.0±0.1 (Table 1), which is the same value as that obtained for the wt enzyme, and indicates that the two-Mg2+-ion mechanism in fact reflects the function of a single catalytic site.
E. coli RNase III catalytic activity is supported by Mn2+ ion in the 0.1–1 mM concentration range, with concentrations above ∼1 mM causing inhibition of cleavage [27,49]. The inhibition is proposed to be due to Mn2+ binding to a site near the catalytic site . To determine whether Mn2+ inhibition is a function of a single catalytic site, the cleavage of R1.1 RNA by RNase III[E117A/wt] was measured as a function of the Mn2+ concentration. Figure 6 shows that Mn2+ supports RNase III[E117A/wt] cleavage of R1.1 RNA at low concentrations, but inhibition is observed at higher concentrations, similar to what is observed with RNase III. The heterodimer has maximal activity at the same Mn2+ concentration as that of the wt enzyme, but is inhibited to a greater degree. The retained sensitivity of RNase III[E117A/wt] to Mn2+ ion indicates that the inhibition is not a strict function of both catalytic sites.
Inhibition of RNase III[E117A/wt] by Mn2+ ion
Processing activity of an RNase III heterodimer containing a single dsRBD
An E. coli RNase III mutant that lacks both dsRBDs is inactive under standard reaction conditions in vitro . It is not known whether a single dsRBD is sufficient to support activity, and if so, how removal of a dsRBD may alter the kinetic parameters. For example, if strong functional co-operativity between the dsRBDs is necessary for activity, then a single-dsRBD RNase III heterodimer would be severely defective. We prepared an E. coli RNase III heterodimer containing a single dsRBD (RNase III[NucD/wt]). A representative cleavage assay using internally 32P-labelled R1.1 RNA is shown in Figure 7(A). The assay shows that RNase III[NucD/wt] is catalytically active and cleaves R1.1 RNA at the canonical site (Figure 7A, lanes 7–11). Thus a single dsRBD is sufficient for activity, and site-specificity is not altered. However, RNase III[NucD/wt] cleaves R1.1 RNA more slowly than RNase III (Figure 7A, compare lanes 7–11 with lanes 1–6). To determine the cause(s) of the reduced rate, the Km and kcat values were determined (Table 1). The Km of 157±29 nM and the kcat of 1.1±0.2 min−1 for RNase III[NucD/wt] are ∼3-fold greater and ∼2-fold lower respectively than the values for RNase III. We conclude that removal of a dsRBD weakens substrate affinity, as indicated by the larger Km, and that it also perturbs one or more steps in the enzyme–substrate complex, as indicated by the lower kcat value.
Catalytic activity of RNase III with a single dsRBD
The source of the Km and kcat effects was investigated further by examining the ability of RNase III[NucD/wt] to cleave R1.1[WC] RNA. A time course assay shows that R1.1[WC] RNA is cleaved, albeit more slowly compared with the reaction involving RNase III (Figure 7B, compare lanes 7–11 with lanes 1–6). In addition, singly cleaved intermediates are observed (Figure 7B, lanes 8–11). The addition of excess unlabelled R1.1[WC] RNA blocked further cleavage of the intermediates, as well as preventing cleavage of unchanged substrate (results not shown). The greater amount of the shorter intermediate product relative to the longer intermediate product suggests a greater affinity of the longer intermediate product for RNase III, allowing more efficient cleavage at the remaining bond (see the Discussion section). The presence of these species is indicative of a lower stability of the enzyme-cleaved intermediate and also suggests that the subunit lacking the dsRBD is catalytically defective. The latter possibility is examined further below.
A gel shift assay involving RNase III[NucD/wt] and 5′-32P-labelled R1.1[WC] RNA did not yield a specific complex. Instead, there occurred a protein-concentration-dependent disappearance of the unbound RNA, accompanied by an accumulation of the RNA at the bottom of the gel well (results not shown). This suggests protein aggregation accompanying RNA binding. In contrast, RNase III formed a discrete shifted product, with little aggregation of RNA. By measuring the fraction of unbound RNA as a function of enzyme concentration, a K′d of ∼140 nM can be estimated. In contrast, the complex involving RNase III has a K′d of 8.6 nM (Table 1). The higher K′d value for RNase III[NucD/wt] reflects a weakened affinity for the substrate and is consistent with the larger Km value (see above).
Severely defective RNA processing activity of the RNase III[NucD/E117A] double-mutant heterodimer
It is possible that the diminished activity of RNase III[NucD/wt] reflects a specifically compromised function of the subunit lacking the dsRBD, rather than a generalized effect of dsRBD removal on holoenzyme activity. In such a model, the dsRBD would function primarily in a co-ordinate manner with the catalytic domain of the same subunit. If this model were true, and if the dsRBD of one subunit cannot functionally substitute for the other dsRBD, then inactivation of the catalytic site of the full-length subunit in RNase III[NucD/wt] would be expected to severely reduce activity, since both subunits would be catalytically impaired. RNase III[NucD/E117A] was prepared to test this model. Here, one subunit lacks the dsRBD, while the full-length subunit carries the E117A mutation. Time course and enzyme titration assays reveal that, in comparison with RNase III[NucD/wt] or RNase III[E117A/wt] (see above), the RNase III[NucD/E117A] heterodimer cleaves R1.1[WC] RNA only at high enzyme concentrations, or at extended reaction times (Figure 7B, lanes 12–16). Thus point mutational inactivation of the catalytic site in one subunit further impairs activity over that caused by removal of the dsRBD from the opposing subunit. The relation of this result to the RNase III mechanism of action is discussed below.
The present study used mutant heterodimers of E. coli RNase III to assess the consequences of single catalytic site inactivation or dsRBD deletion on substrate recognition and cleavage. RNase III heterodimer production was accomplished using an in vivo system that co-produces two RNase III polypeptides with different affinity tags. Serial affinity chromatography provided RNase III heterodimers free of the corresponding homodimers. As discussed, the E. coli RNase III heterodimers do not measurably engage in subunit exchange, thus allowing enzymological assays uncompromised by contaminating homodimers. The catalytic activity of RNase III[E117A/wt] demonstrates the functional independence of the two catalytic sites and confirms the previous report of retained processing activity of RNase III[E117K/wt] . The rate constant for phosphodiester cleavage by RNase III[E117A/wt] under single-turnover conditions has the same Mg2+ concentration dependence as the wt enzyme (h=2.0). Thus the probable two-Mg2+-ion mechanism for phosphodiester hydrolysis by RNase III [24,29] can be attributed to the operation of a single catalytic site, rather than reflecting a contribution of metal ions bound to both catalytic sites. Whereas there is no description yet of a bacterial RNase III structure with two Mg2+ ions in both catalytic sites, a crystal structure of G. intestinalis Dicer  exhibits two Europium (Eu3+) ions bound to each catalytic site, in positions consistent with a two-metal-ion catalytic mechanism. The single metal ion observed in the catalytic sites of the bacterial RNase III structures [17,24] is probably responsible for activating the water nucleophile, while the second metal ion (whose binding may accompany substrate binding) may facilitate departure of the ribose 3′-oxygen. Thus a pre-catalytic RNase III–substrate complex would contain four Mg2+ ions, with each catalytic site containing two Mg2+ ions. The inhibition of E. coli RNase III by Mn2+ at concentrations >1 mM was revealed in the initial characterization of the enzyme . The retained sensitivity of RNase III[E117A/wt] to Mn2+ indicates that the inhibitory mechanism does not require two functional catalytic sites. Determination of the mechanism of Mn2+ inhibition requires further kinetic and structural studies.
The inactivation of a single catalytic site has specific effects on the steady-state rate constants and pattern of cleavage of substrates with either one or two scissile bonds. While RNase III[E117A/wt] exhibits an unaltered Km for cleavage of a substrate with a single scissile bond, it has a kcat that is one-half that of the wt enzyme. The 2-fold difference in kcat values most likely reflects the two possible modes of substrate binding to the mutant heterodimer, which determines whether the scissile bond is placed in a functional or non-functional catalytic site. For a substrate with two scissile bonds, the pattern of cleavage indicates that a given catalytic site can cleave only one phosphodiester during a single binding event. Thus RNase III[E117A/wt] cleavage of R1.1[WC] RNA generates significant levels of singly cleaved intermediates, and these species must dissociate and rebind in order for the remaining phosphodiester to be cleaved. On dissociation of the singly cleaved species, the shorter of the two cleavage products may spontaneously disengage from the larger product, creating a modified larger product carrying either a 5′ or 3′ single-strand extension, depending on which site was cleaved. The new species may be only weakly bound by the enzyme, leading to a much slower second cleavage step, as observed. The action of single-catalytic-site E. coli RNase III heterodimers towards dsRNA is formally similar to that of Type II restriction enzyme heterodimers that carry a single functional catalytic site and that cleave only one DNA strand during a binding event [50,51]. We note that while E. coli RNase III can be used to create short dsRNAs as gene silencing reagents , a single-catalytic-site heterodimer can provide a source of nicked dsRNAs for functional analyses of dsRNA recognition and processing.
The present study has shown that a single dsRBD can support substrate cleavage by E. coli RNase III under standard conditions in vitro. However, both dsRBDs are required for full activity. The ∼3-fold higher Km for E. coli RNase III[NucD/wt] cleavage of R1.1 RNA, compared with that of RNase III, and the observation of singly cleaved intermediates in reactions involving R1.1[WC] RNA, suggest that the mutant heterodimer–substrate complex is less stable than the complex involving the wt enzyme. The higher Km probably incorporates a statistical factor, in that RNase III[NucD/wt] is only half as likely to recognize substrate as the wt enzyme. The 2-fold lower kcat for RNase III[NucD/wt] indicates that a step associated with the enzyme–substrate complex is perturbed. One possibility is that the subunit that lacks the dsRBD is specifically defective in catalysis (see below).
Do the two dsRBDs bind the substrate in an independent or co-operative manner? A thermodynamic analysis of protein–RNA recognition  may provide some insight. If binding of the two dsRBDs to the substrate is fully co-operative, then the free energies of binding of each dsRBD would be additive, and the measured apparent binding affinity (K′a) for RNase III would be the product of the K′a values of the individual dsRBD. If the binding of the two dsRBDs is fully non-co-operative, then the measured K′a for RNase III would be the sum of the K′a values of the individual dsRBDs. The K′d for the complex of R1.1[WC] RNA bound to the isolated dsRBD is ∼800 nM . Thus, if the two dsRBDs function in a fully co-operative manner, the predicted K′a (1/K′d) for RNase III would be 1.6×1012 M−1. If instead the two dsRBDs are non-co-operative, the predicted K′a for RNase III would be 2.5×106 M−1. The measured K′a for RNase III is 2.5×108 M−1, indicating that RNase III binds the substrate ∼6000 times more weakly than expected for full co-operativity, and only ∼100 times more strongly than expected for full non-co-operativity. Thus the analysis indicates a largely independent (non-co-operative) behaviour of the two dsRBDs in recognizing the substrate.
These considerations, along with the structural features of non-catalytic [22,23] or post-catalytic  RNase III·dsRNA complexes, suggest a pathway for substrate recognition by bacterial RNase III (Figure 8A). This model is based on a previous proposal by Ji and co-workers for bacterial RNase III recognition of dsRNA . The pathway involves initial recognition of the substrate by a dsRBD, followed by NucD engagement of the substrate, with an accompanying protein conformational change. The other dsRBD then engages the substrate to stabilize the catalytically competent complex (Figure 8A). As discussed previously , the two steps are dependent on the flexibility of the linker connecting the dsRBD and NucD. This model is also consistent with a stopped-flow kinetic analysis that detected an RNase III conformational change associated with substrate binding . The measured second-order rate constant (K′a) is smaller than the diffusion-controlled rate constant, suggesting that the event reflects a protein conformational change immediately following initial binding of the substrate . The presence of only a single dsRBD would provide a single pathway in step 1 (see Figure 8B), which would be reflected in a higher Km (see the Results section). The presence of only one dsRBD would also affect the catalytic step, if the dsRBD participates in catalysis. In this regard, the severely defective activity of RNase III[NucD/E117A] indicates that the dsRBD preferentially supports the action of the catalytic site of the same subunit. This functional coupling would explain the 2-fold lower kcat of RNase III[NucD/wt] cleavage of R1.1 RNA, and the generation of singly cleaved R1.1[WC] RNA intermediates, compared with the action of the wt enzyme. The behaviour of RNase III[NucD/wt] may shed light on the mechanisms of eukaryotic RNase III orthologues that contain a single dsRBD. Thus the mammalian Drosha polypeptide, which contains a single dsRBD, is only catalytically active in the presence of DGCR8 protein, which contains two dsRBDs . One possibility would be that one of the dsRBDs of DGCR8 binds the substrate, while the second dsRBD may functionally complement the single Drosha dsRBD to provide optimal catalytic activity.
Proposed pathway of dsRNA recognition and cleavage by RNase III and comparison with the RNase III[NucD/wt] pathway
DNA sequencing analyses were performed by the DNA Sequencing Facility of the University of Pennsylvania. We thank Dr R. H. Nicholson (Department of Biology, Temple University, Philadelphia, PA, U.S.A.) for helpful comments on this paper, and other members of the laboratory for their support and encouragement. This research was supported by NIH (National Institutes of Health) grant number GM56457.