www.biochemj.org

Biochem. J. (2010) 426, 337–344 (Printed in Great Britain) doi:10.1042/BJ20091553 337

Characterization of RNase HII substrate recognition using RNase HII–argonaute chimaeric enzymes from Pyrococcus furiosus Sayaka KITAMURA*†, Kosuke FUJISHIMA*, Asako SATO*, Daisuke TSUCHIYA*, Masaru TOMITA*† and Akio KANAI*†1 *Institute for Advanced Biosciences, Keio University, Tsuruoka 997-0017, Japan, and †Systems Biology Program, Graduate School of Media and Governance, Keio University, Fujisawa 252-8520, Japan

RNase H (ribonuclease H) is an endonuclease that cleaves one of the conserved secondary structural regions (the fourth the RNA strand of RNA–DNA duplexes. It has been reported β-sheet and the fifth α-helix of Pf-RNase HII) contains family- that the three-dimensional structure of RNase H is similar to that specific amino acid residues. Using a series of Pf-RNase HII–Pf - of the PIWI domain of the Pyrococcus furiosus Ago (argonaute) Ago chimaeric mutants of the region, we discovered that residues protein, although the two enzymes share almost no similarity Asp110,Arg113 and Phe114 are responsible for the dsRNA (double- in their amino acid sequences. Eukaryotic Ago proteins are stranded RNA) digestion activity of Pf-RNase HII. On the basis of key components of the RNA-induced silencing complex and the reported three-dimensional structure of Ph-RNase HII from are involved in microRNA or siRNA (small interfering RNA) Pyrococcus horikoshii, we built a three-dimensional structural recognition. In contrast, prokaryotic Ago proteins show greater model of RNase HII complexed with its substrate, which suggests affinity for RNA–DNA hybrids than for RNA–RNA hybrids. that these amino acids are located in the region that discriminates Interestingly, we found that wild-type Pf-RNase HII (P. furiosus, DNA from RNA in the non-substrate strand of the duplexes. RNase HII) digests RNA–RNA duplexes in the presence of Mn2+ ions. To characterize the substrate specificity of Pf-RNase HII, we aligned the amino acid sequences of Pf-RNase HII and Pf - Key words: archaeon, argonaute, double-stranded RNA (dsRNA), Ago, based on their protein secondary structures. We found that ribonuclease H, RNA–DNA duplex, site-directed mutagenesis.

INTRODUCTION abyssi and Pyrococcus horikoshii. Interestingly, several bacterial Ago proteins, such as Aa-Ago (Aquifex aeolicus Ago) and RNase H (ribonuclease H) is a ubiquitous enzyme found in all Tt-Ago (Thermus thermophilus Ago), are reported to be DNA- three kingdoms of the tree of life: archaea, bacteria and eukarya. strand-mediated site-specific RNA endonucleases [8–10]. In other

Bacteria and eukaryotes contain two or more RNase H-encoding words, these Ago proteins can digest the RNA strand of RNA– Biochemical Journal genes, whereas the hyperthermophilic archaeon Pyrococcus DNA hybrids, similarly to the RNase H proteins. It has also furiosus has only one gene, designated rnhB [1]. We have pre- been reported that the archaeal PIWI protein from Archaeoglobus viously constructed a reconstitution system for Okazaki fragment fulgidus (Af -PIWI) binds to DNA more tightly than it does to processing using two P.furiosus recombinant enzymes, Pf-RNase RNA [11], suggesting an evolutionary relationship between the HII and Pf -FEN-1 (Flap endonuclease 1) [2]. We showed that both RNase H and Ago proteins in prokaryotes. enzymes are required for the effective degradation of the RNA To characterize the substrate specificity of Pf-RNase HII, we moiety of an RNA–DNA–DNA substrate (the Okazaki substrate). first aligned the amino acid sequences of Pf-RNase HII and Pf - Song et al. [3] have reported that the three-dimensional structure Ago based on the protein secondary structures. We found that of the PIWI domain of the P. furiosus Ago (argonaute) protein one of the helical regions corresponding to the same secondary is similar to that of RNase H, although the two enzymes show structure in both proteins contains family-specific conserved almost no similarity in their amino acid sequences. It has also been amino acid residues. We then used Pf -RNase HII as the basic reported that in the eukarya, the Ago proteins are key components enzyme to produce a series of chimaeric Ago–RNase HII enzymes of the RNA-induced silencing complex and are involved in in Escherichia coli that were mutated in the region of interest, microRNA or siRNA (small interfering RNA) recognition [4–7]. and characterized their specificities. As a result, we unexpectedly Biochemical studies of Ago proteins from the eukarya have shown found that wild-type Pf -RNase HII digests the RNA strand of that some have endonuclease (slicer) activity and can digest RNA–RNA duplexes. We also discovered that amino acid residues one RNA moiety of RNA–RNA duplexes. Unlike eukaryotic Asp110,Arg113 and Phe114 are responsible for the dsRNA (double- genomes, which contain one or more Ago genes (and/or Ago- stranded RNA) digestion activity of Pf-RNase HII and are located related piwi genes), few genomes of either bacterial or archaeal in the region that discriminates the non-substrate strand (the so- origin encode Ago or Ago-related PIWI proteins. For example, the called ‘guide strand’) in RNA–DNA and RNA–RNA duplexes, as P.furiosus genome has one Ago gene (Pf-Ago), whereas there is no revealed by a three-dimensional structural model of Pyrococcus such gene in the genomes of the closely related species Pyrococcus enzymes.

Abbreviations used: Aa, Aquifex aeolicus; Af, Archaeoglobus fulgidus; Ago, argonaute; dsRNA, double-stranded RNA; FAM, carboxyfluorescein; FEN, Flap endonuclease; Mj, Methanococcus jannaschii; Mk, Methanopyrus kandleri; Ni-IMAC, nickel-immobilized metal-ion-affinity chromatography; Pf, Pyrococcus furiosus; Ph, Pyrococcus horikoshii; RNase H, ribonuclease H; rnhB, RNase H-encoding gene; RT, reverse transcriptase; ssDNA, single- stranded DNA; Tt, Thermus thermophilus; WT, wild-type. 1 To whom correspondence should be addressed (email [email protected]).

c The Authors Journal compilation c 2010 Biochemical Society 338 S. Kitamura and others

EXPERIMENTAL Table S1 for details). Standard processing assays were performed in a final volume of 20 μl which contained 20 mM Tris/HCl Dataset (pH 8.0), 1 mM dithiothreitol, 50 mM KCl, 4 mM MnCl2, The UniProt accession codes for each protein are as follows: 40 pmol of substrate and 5–100 ng of the purified recombinant Pf -RNase HII, Q8U036; Ph-RNase HII, O59351; Af -RNase Pf -RNase HII protein. The reaction mixture was incubated at HII, O29634; Mj (Methanococcus jannaschii)-RNase HII, 50 ◦C for 15 min and quenched with an equal volume of STOP Q57599; Mk (Methanopyrus kandleri)-RNase HII, Q8TYV5; solution [1 M Tris/HCl (pH 7.2), 8 M urea and a small amount

Pf -Ago, Q8U3D2; Af -PIWI protein, O28951; Mj-PIWI protein of Blue Dextran (Sigma–Aldrich)]. In some experiments, MgCl2 (uncharacterized protein MJ1321), Q58717; and Mk-Ago (the was used instead of MnCl2. The samples were denatured at homologue of the eukaryotic protein), Q8TVS7. The protein 70 ◦C for 5 min and the cleavage products were separated on secondary structures of the Pf -RNase HII and the Pf -Ago PIWI a20% polyacrylamide gel containing 8 M urea. Gel images domain proteins were generated from the structures of Ph- were visualized and analysed using a Molecular Imager FX Pro RNase HII (PDB entry 1UAX) and Pf -Ago (PDB entry 1U04) apparatus (Bio-Rad Laboratories). respectively. Structural model of Pf-RNase HII complexed with a dsRNA ligand Construction of the expression vectors A structural model was constructed using the molecular graphics The Pf -RNase HII M1–M4 expression plasmids were constructed software Visual Molecular Dynamics [12] with the atomic co- by replacing the NdeI–HindIII fragment of the WT (wild- ordinates of Ph-RNase HII (PDB 1UAX) and an A-form dsRNA type) Pf -RNase HII plasmid pET-rnh2 [2] with each of the (PDB 2Q1R), which were downloaded from the Protein Data NdeI–HindIII fragments amplified using the mutant-specific Bank. The dsRNA structure was manually positioned on to the primers described in Supplementary Table S1 (available at RNase HII model, with reference to the crystal structure of http://www.BiochemJ.org/bj/426/bj4260337add.htm). The Pf - the complex between Bacillus halodurans RNase H and an RNA– RNase HII M5–M9 plasmids were constructed using the DNA hybrid (PDB 1ZBI). QuikChange® site-directed mutagenesis kit (Stratagene) and specific primers (Supplementary Table S1) with pET-rnh2 as the template. The resulting vectors, pET-rnhM1–M9, encoded each of RESULTS the Pf -RNase HII mutant proteins with a His6 tag at the C-terminal ends. The nucleotide sequences of the expression plasmids were Alignment of the amino acid residues of Pf-RNase HII and the PIWI verified by sequencing. domain of Pf-Ago, based on their secondary structures As several bacterial Ago proteins (e.g. Aa-Ago and Tt-Ago) Expression and purification of His6-tagged recombinant proteins are reportedly DNA-strand-mediated site-specific RNA endo- E. coli strain BL21(DE3) was transformed with the expression nucleases [8–10], and because the archaeal Af -PIWI protein plasmids. The transformants were grown exponentially in Luria– binds to DNA more tightly than to RNA [11], we first focused Bertani medium containing 50 μg/ml ampicillin and 0.4 mM on the biochemical properties of recombinant Pf-Ago protein. ◦ IPTG (isopropyl β-D-thiogalactoside) at 37 C. After 14–16 h of We used a gel-shift assay to confirm that the purified Pf-Ago further growth at 30 ◦C, the cells were harvested by centrifugation protein preferentially binds to ssDNA (single-stranded DNA) (5000 g for 5 min at 4 ◦C) and the recombinant proteins were over dsDNA (double-stranded DNA) or RNA–DNA duplexes, extracted by sonication with a Handy Sonic model UR-20P and obtained no clear band-shift for dsRNA (results not shown). (TOMY SEIKO, Tokyo, Japan) at maximal power level for Therefore we concluded that the nucleic-acid-binding activity 0.5 min in buffer A [20 mM Tris/HCl (pH 8.0), 5 mM imidazole, of the Pf-Ago protein is similar to that of the Aa-Ago and 500 mM NaCl and 0.1% NP40 (Nonidet P40)]. These extracts Af-PIWI proteins. Furthermore, we did not detect any Pf -Ago were heat-treated at 85 ◦C for 15 min to denature the endogenous endonuclease activity for either the RNA–RNA or RNA–DNA ribonuclease-related proteins from E. coli and centrifuged at duplexes (results not shown). 12000 g for 10 min at 4 ◦C to remove debris. The recombinant Pf - Next, we compared the amino acid sequences of the Pf -RNase RNase HII and mutant proteins were purified using the Proteus Ni- HII and Pf -Ago proteins, based on their secondary structures. As IMAC (nickel immobilized metal-ion-affinity chromatography) described previously [2], the Pf-rnhB gene encodes a protein of protein purification kit (Prochem). The eluted protein peaks 224 amino acids. It has also been reported that the Pf-Ago gene were dialysed against buffer B [50 mM Tris/HCl (pH 7.5), 1 mM encodes a protein of 770 amino acids, with four distinct domains: EDTA, 0.2% Tween 20, 7 mM 2-mercaptoethanol and 10% (v/v) an N-terminal domain (residues 1–151), a PAZ (piwi argonaute glycerol]. zwille) domain (residues 152–275), a middle domain (residues For the detection of dsRNA-digestion activity, 1 ml of the 362–544) and a PIWI domain (residues 545–770) [3]. Thus the dialysed lysate (approx. 50 μg/ml) was loaded on to a 1 ml PIWI domain of the Pf -Ago protein is 226 amino acids long, RESOURCETM S column (GE Healthcare) equilibrated with similar to the length of Pf -RNase HII. buffer B and then eluted with a linear gradient of NaCl [from The three-dimensional structures of Pf -RNase HII and the 0 M (fraction 7) to 0.5 M (fraction 16)] in buffer B using an Pf -Ago PIWI domain revealed the presence of topologically β α AKTA¨ Purifier FPLC system (GE Healthcare). common secondary structures: five -sheets and three - helixes (Figure 1). We identified the amino acid residues that are evolutionarily conserved in each of the respective protein RNase H assay families (see Supplementary Figures S1 and S2 available at Oligonucleotides, either unlabelled or 3-end-labelled with the http://www.BiochemJ.org/bj/426/bj4260337add.htm). We found fluorescent marker FAM (carboxyfluorescein), were chemically that only four archaeal species (M. jannaschii, P. furiosus, synthesized by Hokkaido System Science. These oligonucleotides A. fulgidus and M. kandleri) contain Ago or PIWI genes in their were annealed to the complementary DNA or RNA molecules and genomes, so we used these species to analyse the conservation the resulting duplexes were used as substrates (see Supplementary of the amino acid residues of RNase HII. We mapped the

c The Authors Journal compilation c 2010 Biochemical Society Substrate recognition of archaeal RNase HII 339

Figure 1 Alignment and comparison of the amino acid residues of Pf-RNase HII and PIWI domain from Pf-Ago based on their secondary structures

The amino acid sequences were aligned using ClustalX [33]. Gaps (–) were inserted to maximize the number of amino acid matches. Secondary structural elements are denoted by arrows [β-sheets (E1–E9)], waves [α-helices (H1–H9)] and humps (turns). The corresponding secondary structures, as determined by a comparison of the three-dimensional structures of Pf-RNase HII and PIWI domain from Pf-Ago, are enclosed in boxes. Amino acid residues that are evolutionarily conserved among the four related species are indicated by black (completely conserved) and grey (partially conserved) arrows (also see Supplementary Figures S1 and S2). *, identity; :, conservative replacement; ., non-conservative replacement. evolutionarily conserved amino acid residues in the secondary- structure-based amino acid alignment of the Pf -RNase HII and Pf -Ago proteins. We found that the RNase H family of proteins has a greater number of conserved amino acid residues than does the Ago or PIWI protein families. The N-terminal (residues 5– 80) and C-terminal (residues 140–238) regions of Pf-RNase HII are particularly well conserved. Conversely, there were relatively few conserved amino acid residues in the Ago or PIWI protein families. We paid particular attention to one of the conserved regions (the fourth β-sheet and fifth α-helix of Pf-RNase HII, corresponding to the fourth β-sheet and second α-helix of Pf-Ago, i.e. the PIWI domain), which had the same protein secondary structure in the two proteins and family-specifically conserved amino acid residues in each of the proteins. The conserved region of Pf-RNase HII contains the amino acid residue that is located 105 in the centre of its active site (i.e. residue Asp ). Figure 2 Summary of Pf-RNase HII–Pf-Ago chimaeric proteins

The positions and sequences of the amino acid residues in the Pf-RNase HII–Pf-Ago chimaeric Nuclease activity of the Pf-RNase HII–Pf-Ago chimaeric proteins regions are shown. The amino acid sequences that were replaced between Pf-RNase HII and Pf-Ago are indicated in bold. The two corresponding secondary protein structures (the fourth As we had established an in vitro system for measuring Pf -RNase β-sheet and fifth α-helix of Pf-RNase HII and the fourth β-sheet and second α-helix of Pf-Ago) HII activity previously [2], and had experienced difficulty in are enclosed in boxes. producing a detectable amount of nuclease activity for the Pf- Ago protein, we used Pf -RNase HII as the base enzyme in this system and designed a series of amino-acid-substituted mutants by replacing amino acid residues in the fourth β-sheet and fifth 85 ◦C for 15 min, followed by purification using the Proteus Ni- α-helix of Pf-RNase HII with those located in the fourth β- IMAC protein purification kit. The purified proteins had molecular sheet and second α-helix of the Pf-Ago protein (Figure 2). To masses of approx. 26 kDa, as determined by SDS/PAGE (see generate large amounts of the WT and mutant chimaeric enzymes Supplementary Figure S3 available at http://www.BiochemJ.org/ for biochemical characterization, recombinant C-terminal His6- bj/426/bj4260337add.htm). This finding is consistent with the tagged proteins were overexpressed in E. coli strain BL21(DE3) size predicted from the amino acid sequences deduced from and were purified to near-homogeneity using heat treatment at the corresponding genes.

c The Authors Journal compilation c 2010 Biochemical Society 340 S. Kitamura and others

Figure 4 Cleavage specificity of the purified recombinant Pf-RNase HII and site-directed mutants of Pf-RNase HII (M5–M9) on the RNA–DNA hybrid substrate

(A) Digestion products were analysed by denaturing PAGE as described in the Experimental section. A 3-FAM-labelled AGR1–AGD2 hybrid substrate was used in the processing assay. The reaction was performed with 0–20 ng of purified protein in the presence of 4 mM MnCl2. Similar results were obtained in at least three independent experiments. (B) Graphical comparison of the relative amounts of the 21-mer substrate when digested by WT, M5, M6 and M7 enzymes. The amount of 21-mer substrate in the absence of enzyme was set to 100%. The results represent + the means − S.D. for three experiments.

2+ Figure 3 Cleavage specificity of purified recombinant Pf-RNase HII and presence of Mn , M1–M3 showed a gradual reduction in nuclease Pf-RNase HII–Pf-Ago chimaeric proteins (M1–M4) on the RNA–DNA hybrid activity, according to the number of amino acid replacements. substrate M4 showed no detectable nuclease activity, because the replaced 105 (A) Digestion products were analysed by denaturing PAGE as described in the Experimental region contained amino acid residue Asp , which is located in section. A 3-FAM-labelled AGR1–AGD2 hybrid substrate was used in the processing assay. The the active site and is necessary for the enzymatic activity. There reaction was performed with 0–100 ng of purified protein in the presence of 4 mM MnCl2. The was virtually no difference in the digestion patterns of the WT cleavage sites (for the RNA strand only) are shown as follows: large and small arrows represent and mutant enzymes. However, the sizes of the major digested major and intermediate cleavage sites respectively and arrowheads represent minor cleavage bands became larger, resulting from an apparent shift in the major sites. (B) The analysis depicted in (A) was performed with 4 mM MgCl2 instead of 4 mM MnCl2. digestion sites on the substrate. We confirmed that divalent ions Similar results were obtained in at least two independent experiments. were necessary for enzyme activity, but the concentrations of these ions did not affect the fundamental patterns of substrate digestion To examine the nuclease activity of the purified recombinant (results not shown). These results suggest that the mutations enzymes, we used either a 21-mer RNA–DNA duplex reduced the basic activity of the enzyme, which resulted in the (oligonucleotides AGR1 and AGD2; Supplementary Table S1) or accumulation of longer digested bands over a given period. a 21-mer RNA–RNA duplex (oligonucleotides AGR1 and AGR2; The effect of amino acid replacements on the nuclease activity Supplementary Table S1) with two overhanging nucleotides at was more pronounced in the presence of Mg2+. The observations each of the 3 ends, to mimic dicer products [13]. However, we did that the replacement of five amino acids (M1) resulted in a marked not observe any significant differences between the overhanging reduction in enzymatic activity and that additional replacements and blunt-ended substrates (results not shown). Furthermore, 5- abolished the activity altogether suggest that the Asn111–Glu115 end kination of the duplex, which is required for Ago nuclease region also corresponds to the divalent-ion recognition site. To activity [14,15], did not affect the nuclease activity of the mutants identify the amino acids responsible for the nuclease activity, (results not shown). First, we checked the nuclease activity of we characterized further point mutations in the region (M5– the WT and mutants M1–M4 using a 21-mer RNA–DNA duplex M9). The results revealed that the F114E (M6) and R113E (M7) with two overhanging nucleotides at the 3 ends in the presence of mutations caused reduced and partially reduced nuclease activity either Mn2+ (Figure 3A) or Mg2+ (Figure 3B). The mutants M1– respectively for the RNA–DNA duplex substrate. In contrast, M4 had 5, 8, 11 and 16 amino acids replaced respectively. In the the A115E (M5) mutation caused partially enhanced nuclease

c The Authors Journal compilation c 2010 Biochemical Society Substrate recognition of archaeal RNase HII 341

Table 1 Comparison of the Pf-RNase HII kinetic parameters 3-FAM-labelled AGR1–AGD2 or AGR1–AGR2 hybrid substrates were used for the processing ◦ assay. The reaction was carried out at 50 C for 15 min in the presence of 4 mM MnCl2 or MgCl2. The amount of enzyme was controlled so that the reaction of the hydrolysed substrate did not exceed 30% of the total, as described previously [32]. Hydrolysis of the substrates with the enzyme followed Michaelis–Menten kinetics and the kinetic parameters were calculated via a Lineweaver–Burk plot. The results shown are the means of at least three experiments; the S. D. was less than 10% in each case. –, no detectable activity.

−1 Enzyme Substrate Divalent ion K m (μM) k cat (min )

WT RNA–DNA Mn2+ 0.12 11.3 WT RNA–RNA Mn2+ 1.30 0.093 WT RNA–DNA Mg2+ 0.12 17.0 WT RNA–RNA Mg2+ –– M5 RNA–DNA Mn2+ 0.14 15.3 M6 RNA–DNA Mn2+ 0.12 0.41 M7 RNA–DNA Mn2+ 0.20 6.6

activity for the RNA–DNA duplex substrate, whereas the other mutants (M8 and M9) displayed activities similar to that of the WT (Figure 4). In these experiments, we used an excess of enzyme to enhance the enzymatic effects of the mutations. We estimated the kinetic parameters of the WT, M5, M6, and M7 enzymes in a separate experiment (Table 1). dsRNA cleavage activity of purified recombinant Pf-RNase HII We checked the nuclease activity of the WT and mutant Pf -RNase HII enzymes using a RNA duplex substrate. For this purpose, we Figure 5 Identification of the dsRNA cleavage activity of the purified further purified the elution peaks from the Ni-IMAC column using recombinant Pf-RNase HII TM a RESOURCE S column, which is an ion-exchange column- TM ¨ (A) SDS/PAGE of WT and M7 Pf-RNase HII proteins purified with RESOURCE S column chromatographic method that uses an AKTA Purifier FPLC chromatography. Aliquots of the fractions from the column were subjected to SDS/PAGE system, to exclude the possibility of any contamination from (10–20% gels) and the gels were stained with Coomassie Brilliant Blue. Lane M, molecular mass E. coli nucleases. We also compared the purification profiles and markers, the size in kDa is indicated on the left-hand side; arrowheads indicate the positions the nuclease activity profiles of the WT and mutant enzyme M7, of the purified proteins. (B) dsRNA cleavage activity and (C) RNA–DNA digestion activity using aliquots of the fractions from the RESOURCETM S column. A 3-FAM-labelled AGR1–AGR2 which is unable to digest the dsRNA substrate (described below). Both WT Pf -RNase HII and the mutant M7 enzyme were eluted hybrid substrate or a 3 -FAM-labelled AGR1–AGD2 hybrid substrate was used in the processing assay respectively. The digestion products were analysed by denaturing PAGE as described in using approx. 0.2 M NaCl (fractions 9 and 10, Figure 5A). WT Pf- the Experimental section. Lo, load fraction (the eluted protein peak from the Ni-IMAC column); 2+ RNase HII digested the dsRNA substrate in the presence of Mn , (-), no fraction added. whereas the mutant M7 enzyme showed dramatically reduced activity (Figure 5B). The digested band intensities were similar to those obtained with SDS/PAGE, which clearly shows that WT Pf-RNase HII has dsRNA digestion activity. There was no dsRNA (M6) and R113E (M7) mutation resulted in reduced nuclease digestion activity in the presence of Mg2+ (results not shown). In activity for the RNA–RNA duplex, similar to that observed contrast, both enzymes exhibited the activity necessary to digest a for the RNA–DNA duplex, one other mutation, D110T (M9), RNA–DNA hybrid substrate (Figure 5C). These results, combined caused the loss of dsRNA digestion activity (Figure 6, M5–M9). with the observation that no other fractions showed similar Among these mutants, M7 had the most severe effect on this nuclease activity, demonstrate that under our assay conditions, activity. we only detected nuclease activity from P. furiosus enzymes in Finally, to determine the effects of the mutations in the M6 the heat-treated and Ni-IMAC-purified protein fractions. (F114E), M7 (R113E) and M9 (D110T) proteins, we constructed Next, we compared the dsRNA digestion activity of the WT a structural model of Pyrococcus RNase HII complexed with and mutant enzymes. As shown in Figure 6, major RNA cleavage dsRNA (Figure 7). As the three-dimensional structure of Pf - produced by Pf-RNase HII was observed at a single ribonucleotide RNase HII has yet to be determined, we used the currently position in the RNA strand of the RNA duplex. We determined the available structural data for Ph-RNase HII. The amino acid kcat value for the RNA–RNA duplex substrate, which was approx. sequence of Pf -RNase HII has 84% similarity to that of Ph- 120-fold lower than that determined for the RNA–DNA duplex RNase HII; the sequence positions Phe114,Arg113 and Asp110 are substrate (Table 1). In contrast, the Km value for the RNA–RNA completely conserved between the two enzymes. The structure duplex substrate was approx. 38-fold higher than that determined of the substrate was in the A-form, because RNA inevitably for the RNA–DNA duplex substrate (Table 1). The introduction assumes this form [16]. On the basis of previous reports [17– of several Pf-Ago protein sequences into Pf-RNase HII abolished 20], we mapped the active-site amino acid residues required for its enzymatic activity completely (Figure 6, M1–M4), which cleavage activity in our model (Asp7,Glu8,Asp105 and Asp135; suggests that the dsRNA digestion activity in the presence of Figure 7, shown in red). A possible cleavage site (Figure 7, Mn2+ did not stem from contamination. Instead, the WT enzyme blue arrowhead) is located in the middle of the active site and appeared to possess this activity. Interestingly, whereas the F114E within the region that interacts directly with the digestion substrate

c The Authors Journal compilation c 2010 Biochemical Society 342 S. Kitamura and others

degradation of one RNA strand of an RNA–RNA duplex in the presence of Mn2+ [22–24]. In all these cases, the cleavage activity for a RNA–RNA duplex is much lower than that observed for an

RNA–DNA hybrid. Indeed, the kcat value of Pf -RNase HII for the dsRNA substrate is just 0.093 min−1 (Table 1). As many RNase H enzymes, including virus RTs, have this activity, this weak dsRNA digestion activity may constitute a fundamental function of the enzyme. Notably, in some cases, RNase H activity for RNA–DNA substrates is regulated by divalent ions. For example, a Mn2+- dependent RNase HII enzyme and an Mg2+-dependent RNase III enzyme have been identified in Bacillus subtilis [25].Wehave also identified Mn2+-dependent RNase HI and Mg2+-dependent RNase HI enzymes in Caenorhabditis elegans previously [26]. It is reported that FEN has two conserved metal-binding sites (high- and low-affinity cofactor-binding sites) and mutation analysis of these sites has suggested that the types of metal ions and the local concentrations of these free cofactors regulate the FEN- catalysed exonucleolytic reaction [27]. It has also been reported that the enzymatic activity of EcoRV is modulated by Mg2+/Mn2+ ions [28]. Therefore the present research suggests that these divalent ions were selected to regulate or modulate ribonuclease activity, depending on the type of substrate, although their in vivo biological roles remain unknown. We also found that amino acid residues Asp110,Arg113 and Phe114 of Pf-RNase HII are strongly involved in the dsRNA-digestion activity of the enzyme. Many studies have identified amino acid residues that are important for the catalytic activity of RNase H. Structural biological approaches, combined with site-directed mutagenesis, provide powerful tools in this field. On the basis of on these techniques, Kanaya [20] reported that five acidic amino acid residues (Asp7,Glu8,Asp40,Asp105 and Asp135)are important for RNase H activity in RNase HII from Thermococcus kodakaraensis. These amino acids all map to a catalytic centre that is homologous with that of Pf -RNase HII (Figure 7, shown in red) and some of these residues are reported to be metal-binding sites [19,20,29,30]. Figure 6 dsRNA cleavage activity of the purified recombinant Pf-RNase HII In the present study, we identified novel amino acid residues and the mutants of Pf-RNase HII (M1–M9) that are not located in the catalytic centre but are important for the dsRNA digestion activity of Pf -RNase HII. Because A3-FAM-labelled AGR1–AGR2 hybrid substrate was used in the processing assay. The reaction Phe114 and Arg113 were replaced with a glutamate residue in the was performed with 0–100 ng of purified protein in the presence of 4 mM MnCl2. The bottom panel shows cleavage sites for the labelled strand only. The arrow and arrowheads represent M6 and M7 mutants, the negative charge on the mutants may have major and minor cleavage sites respectively. Similar results were obtained for at least two electrostatically repulsed the main-chain phosphate(s) to inhibit independent experiments. substrate binding. It should be noted that the R113E mutation significantly reduced the RNase activity of the enzyme for RNA– RNA duplexes (Figure 6), but only partially reduced the activity (Figure 7). Conversely, the important mutated residues (F114E, for RNA–DNA hybrids (Figure 4). F114E reduced the enzyme R113E and D110T) identified in the present study (Figure 7, activity for both substrates, whereas D110T was also exclusively shown in orange, blue and yellow respectively) are not located in important for dsRNA digestion. The structure of the RNA–DNA the active site, but are accessible to the RNA–DNA or RNA–RNA hybrid is dissimilar to the ideal A-form in the crystal structure of its complex with B. halodurans RNase H [31]. The RNA strand duplex substrates. It is noteworthy that these mutated amino acid residues are strategically located so as to only interact with the is in the A-form with 3 -endo sugars, whereas the 2 -endo or 1 - non-digested strand (the guide strand). exo sugars of the DNA strand produce the B-form conformation. This may explain the differences in substrate recognition detected among the mutant enzymes. Furthermore, the A115E (M5) mutant k DISCUSSION showed an increase in its cat of approx. 140% for the RNA–DNA substrate. This again suggests that the region encompassing amino In the present study we found that Pf-RNase HII digested an RNA– acid residues Asp110–Ala115 is responsible for the modulation of RNA duplex in the presence of Mn2+. This is the first report to RNase H-derived enzymatic activities. As the position of amino demonstrate that RNase HII, one of the RNase H-type protein acid residue Ala115 is hidden behind residues Ala113 and Asp110, family members, has the ability to cleave dsRNA. However, it we could not show its position in Figure 7. Considering that the has been reported that RNase HI from the thermoacidophilic mutation sites only interact with the guide strand in the duplex archaeon Sulfolobus tokodaii cleaves dsRNA in the presence (Figure 7), it is possible that the differences in enzymatic activity of Mn2+ and Co2+ ions [21]. The RNase H domain of the RT observed in the presence of different substrates are regulated by (reverse transcriptase) enzymes of some retroviruses, such as HIV the enzyme’s interactions with the guide strand. Further analysis type-1 and Moloney-murine-leukaemia virus, is involved in the is required to test this hypothesis.

c The Authors Journal compilation c 2010 Biochemical Society Substrate recognition of archaeal RNase HII 343

Figure 7 Possible structural model of Pf-RNase HII complexed with a dsRNA substrate

The structure of Ph-RNase HII is shown as the molecular surface. The amino acid residues required for its cleavage activity are shown in red (Asp7,Glu8,Asp105 and Asp135). The mutated residues are colour-coded as follows: orange, M6 (F114E); blue, M7 (R113E); and yellow, M9 (D110T). The two RNA strands are illustrated as cyan and green tubes. The blue arrowhead indicates the possible cleavage site.

The recent exponential growth in PDB entries has allowed 4 Rivas, F. V., Tolia, N. H., Song, J. J., Aragon, J. P., Liu, J., Hannon, G. J. and Joshua-Tor, largely accurate functional predictions to be made on the basis of L. (2005) Purified Argonaute2 and an siRNA form recombinant human RISC. Nat. Struct. structural similarities in the three-dimensional profiles of proteins. Mol. Biol. 12, 340–349 We propose that analysis of these structurally similar proteins 5 Okamura, K., Ishizuka, A., Siomi, H. and Siomi, M. C. (2004) Distinct roles for Argonaute from an evolutionary perspective could result in the identification proteins in small RNA-directed RNA cleavage pathways. Genes Dev. 18, 1655–1666 of novel functional domains, as is the case in the present study. 6 Peters, L. and Meister, G. (2007) Argonaute proteins: mediators of RNA silencing. Mol. Cell 26, 611–623 7 Hock, J. and Meister, G. (2008) The Argonaute protein family. Genome Biol. 9, 210 AUTHOR CONTRIBUTION 8 Yuan, Y. R., Pei, Y., Ma, J. B., Kuryavyi, V., Zhadina, M., Meister, G., Chen, H. Y., Dauter, Z., Tuschl, T. and Patel, D. J. (2005) Crystal structure of A. aeolicus argonaute, a Sayaka Kitamura and Akio Kanai designed research; Sayaka Kitamura, Kosuke Fujishima, site-specific DNA-guided endoribonuclease, provides insights into RISC-mediated mRNA Asako Sato, Daisuke Tsuchiya and Akio Kanai performed research; Daisuke Tsuchiya and cleavage. Mol. Cell 19, 405–419 Kosuke Fujishima contributed to the three-dimensional model analysis; Sayaka Kitamura, 9 Wang, Y., Juranek, S., Li, H., Sheng, G., Tuschl, T. and Patel, D. J. (2008) Structure of an Kosuke Fujishima, Asako Sato, Daisuke Tsuchiya, Masaru Tomita and Akio Kanai analysed argonaute silencing complex with a seed-containing guide DNA and target RNA duplex. data; and Akio Kanai wrote the paper. Nature 456, 921–926 10 Wang, Y., Sheng, G., Juranek, S., Tuschl, T. and Patel, D. J. (2008) Structure of the ACKNOWLEDGEMENTS guide-strand-containing argonaute silencing complex. Nature 456, 209–213 11 Ma, J. B., Yuan, Y. R., Meister, G., Pei, Y., Tuschl, T. and Patel, D. J. (2005) Structural We thank Dr Naoto Ohtani, Dr Hiromi Kochiwa, Dr Mitsuhiro Itaya and Dr Arun Krishnan basis for 5-end-specific recognition of guide RNA by the A. fulgidus Piwi protein. Nature (Keio University, Japan) for their helpful discussions. We also thank Dr Robert J. Crouch 434, 666–670 (Program in Genomics of Differentiation, Eunice Kennedy Shriver National Institute of 12 Humphrey, W., Dalke, A. and Schulten, K. (1996) VMD: visual molecular dynamics. J. Child Health and Human Development, Bethesda, MD, U.S.A.) for his encouragement. Mol. Graph. 14, 27–38 13 Rose, S. D., Kim, D. H., Amarzguioui, M., Heidel, J. D., Collingwood, M. A., Davis, M. E., Rossi, J. J. and Behlke, M. A. (2005) Functional polarity is introduced by Dicer FUNDING processing of short substrate RNAs. Nucleic Acids Res. 33, 4140–4156 This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas; 14 Pellino, J. L., Jaskiewicz, L., Filipowicz, W. and Sontheimer, E. J. (2005) ATP modulates a Grant-in-Aid from the Keio University 21st Century Centre of Excellence Programme, siRNA interactions with an endogenous human Dicer complex. RNA 11, 1719–1724 (Understanding and Control of Life’s Function via Systems Biology); and by research 15 Weitzer, S. and Martinez, J. (2007) The human RNA kinase hClp1 is active on 3 transfer funds from the Yamagata Prefectural Government and Tsuruoka City, Japan. RNA exons and short interfering RNAs. Nature 447, 222–226 16 Gyi, J. I., Lane, A. N., Conn, G. L. and Brown, T. (1998) Solution structures of DNA–RNA hybrids with purine-rich and pyrimidine-rich strands: comparison with the homologous REFERENCES DNA and RNA duplexes. Biochemistry 37, 73–80 1 Kochiwa, H., Tomita, M. and Kanai, A. (2007) Evolution of ribonuclease H genes in 17 Kanaya, S., Kohara, A., Miura, Y., Sekiguchi, A., Iwai, S., Inoue, H., Ohtsuka, E. and prokaryotes to avoid inheritance of redundant genes. BMC Evol. Biol. 7, 128 Ikehara, M. (1990) Identification of the amino acid residues involved in an active site of 2 Sato, A., Kanai, A., Itaya, M. and Tomita, M. (2003) Cooperative regulation for Okazaki Escherichia coli ribonuclease H by site-directed mutagenesis. J. Biol. Chem. 265, fragment processing by RNase HII and FEN-1 purified from a hyperthermophilic 4615–4621 archaeon, Pyrococcus furiosus. Biochem. Biophys. Res. Commun. 309, 247–252 18 Nakamura, H., Oda, Y., Iwai, S., Inoue, H., Ohtsuka, E., Kanaya, S., Kimura, S., Katsuda, 3 Song, J. J., Smith, S. K., Hannon, G. J. and Joshua-Tor, L. (2004) Crystal structure of C., Katayanagi, K., Morikawa, K. et al. (1991) How does RNase H recognize a DNA–RNA Argonaute and its implications for RISC slicer activity. Science 305, 1434–1437 hybrid? Proc. Natl. Acad. Sci. U.S.A. 88, 11535–11539

c The Authors Journal compilation c 2010 Biochemical Society 344 S. Kitamura and others

19 Chapados, B. R., Chai, Q., Hosfield, D. J., Qiu, J., Shen, B. and Tainer, J. A. (2001) 26 Kochiwa, H., Itaya, M., Tomita, M. and Kanai, A. (2006) Stage-specific expression of Structural biochemistry of a type 2 RNase H: RNA primer recognition and removal during Caenorhabditis elegans ribonuclease H1 enzymes with different substrate specificities DNA replication. J. Mol. Biol. 307, 541–556 and bivalent cation requirements. FEBS J. 273, 420–429 20 Kanaya, S. (2001) Prokaryotic type 2 RNases H. Methods Enzymol. 341, 27 Feng, M., Patel, D., Dervan, J. J., Ceska, T., Suck, D., Haq, I. and Sayers, J. R. (2004) 377–394 Roles of divalent metal ions in flap endonuclease-substrate interactions. Nat. Struct. Mol. 21 Ohtani, N., Yanagawa, H., Tomita, M. and Itaya, M. (2004) Cleavage of double-stranded Biol. 11, 450–456 RNA by RNase HI from a thermoacidophilic archaeon, Sulfolobus tokodaii 7. Nucleic 28 Thomas, M. P., Brady, R. L., Halford, S. E., Sessions, R. B. and Baldwin, G. S. (1999) Acids Res. 32, 5809–5819 Structural analysis of a mutational hot-spot in the EcoRV restriction endonuclease: a 22 Blain, S. W. and Goff, S. P. (1993) Nuclease activities of Moloney murine leukemia virus catalytic role for a main chain carbonyl group. Nucleic Acids Res. 27, 3438–3445 reverse transcriptase: mutants with altered substrate specificities. J. Biol. Chem. 268, 29 Chai, Q., Qiu, J., Chapados, B. R. and Shen, B. (2001) Archaeoglobus fulgidus RNase HII 23585–23592 in DNA replication: enzymological functions and activity regulation via metal cofactors. 23 Gotte, M., Fackler, S., Hermann, T., Perola, E., Cellai, L., Gross, H. J., Le Grice, S. F. and Biochem. Biophys. Res. Commun. 286, 1073–1081 Heumann, H. (1995) HIV-1 reverse transcriptase-associated RNase H cleaves RNA/RNA 30 Lai, B., Li, Y., Cao, A. and Lai, L. (2003) Metal ion binding and enzymatic mechanism of in arrested complexes: implications for the mechanism by which RNase H discriminates Methanococcus jannaschii RNase HII. Biochemistry. 42, 785–791 between RNA/RNA and RNA/DNA. EMBO J. 14, 833–841 31 Nowotny, M., Gaidamakov, S. A., Crouch, R. J. and Yang, W. (2005) Crystal structures of 24 Gotte, M., Maier, G., Gross, H. J. and Heumann, H. (1998) Localization of the active site of RNase H bound to an RNA/DNA hybrid: substrate specificity and metal-dependent HIV-1 reverse transcriptase-associated RNase H domain on a DNA template using catalysis. Cell 121, 1005–1016 site-specific generated hydroxyl radicals. J. Biol. Chem. 273, 10139–10146 32 Haruki, M., Tsunaka, Y., Morikawa, M. and Kanaya, S. (2002) Cleavage of a DNA–RNA– 25 Ohtani, N., Haruki, M., Morikawa, M., Crouch, R. J., Itaya, M. and Kanaya, S. (1999) DNA/DNA chimeric substrate containing a single ribonucleotide at the DNA–RNA junction Identification of the genes encoding Mn2+-dependent RNase HII and Mg2+-dependent with prokaryotic RNases HII. FEBS Lett. 531, 204–208 RNase HIII from Bacillus subtilis: classification of RNases H into three families. 33 Thompson, J. D., Gibson, T. J. and Higgins, D. G. (2002) Multiple sequence alignment Biochemistry 38, 605–618 using ClustalW and ClustalX. Curr. Protoc. Bioinformatics, Chapter 2, Unit 2.3

Received 8 October 2009/4 January 2010; accepted 5 January 2010 Published as BJ Immediate Publication 5 January 2010, doi:10.1042/BJ20091553

c The Authors Journal compilation c 2010 Biochemical Society Biochem. J. (2010) 426, 337–344 (Printed in Great Britain) doi:10.1042/BJ20091553

SUPPLEMENTARY ONLINE DATA Characterization of RNase HII substrate recognition using RNase HII–argonaute chimaeric enzymes from Pyrococcus furiosus Sayaka KITAMURA*†, Kosuke FUJISHIMA*, Asako SATO*, Daisuke TSUCHIYA*, Masaru TOMITA*† and Akio KANAI*†1 *Institute for Advanced Biosciences, Keio University, Tsuruoka 997-0017, Japan, and †Systems Biology Program, Graduate School of Media and Governance, Keio University, Fujisawa 252-8520, Japan

Figure S1 Comparison of the amino acid residues of members of the archaeal RNase HII protein family

The amino acid sequence alignment was performed using ClustalX. Gaps (–) were inserted to maximize the number of amino-acid matches. The relative degrees of amino acid conservation are indicated by grey bars. METJA, M. jannaschii; PYRFU, P. furiosus; ARCFU, A. fulgidus; METKA, M. kandleri. *, identity; :, conservative replacement; ., non-conservative replacement.

1 To whom correspondence should be addressed (email [email protected]).

c The Authors Journal compilation c 2010 Biochemical Society S. Kitamura and others

Figure S2 Comparison of the amino acid residues of members of the archaeal argonaute and PIWI protein families

The amino acid sequence alignment was performed using ClustalX. Gaps (–) were inserted to maximize the number of amino-acid matches. The relative degrees of amino acid conservation are indicated by grey bars. METJA, M. jannaschii; PYRFU, P. furiosus; ARCFU, A. fulgidus; METKA, M. kandleri. *, identity; :, conservative replacement; ., non-conservative replacement.

c The Authors Journal compilation c 2010 Biochemical Society Substrate recognition of archaeal RNase HII

Figure S3 Purified recombinant Pf-RNase HII–Pf-Ago chimaeric proteins

Purified proteins were analysed by SDS/PAGE (14–16% gels). The gel was stained with CoomassieBrilliantBlue.Arrowheadindicatesthepositionofthepurifiedproteins.Themolecular mass in kDa is indicated on the left-hand side of the gel.

Table S1 List of the oligonucleotides used in the present study ssRNA, single-stranded RNA.

Primer name Sequence (5–3)

Cloning of chimaeric proteins: M1–M4 sense primer HMS-1L GAAGGAGATATACATATGAAAATAGGGGGAATTGAC M1 (antisense) HMA-1L TCTCTCTATCAAGCTTTCTTCTTCATTGTTATCTACATCCGCGGCATCAGCGTAAATAAG M2 (antisense) HMA-2L TCTCTCTATCAAGCTTTCTTCTTCATTGTTAGTTATGCGCGCTGCATCAGCGTAAATAAGAGCTGGTTT M3 (antisense) HMA-3L TCTCTCTATCAAGCTTTCTTCTTCATTGTTAGTTATACGCCCGTCACGATCGGCGTAAATAAGAGCTGGTTTAATCTG M4 (antisense) HMA-4L TCTCTCTATCAAGCTTTCTTCTTCATTGTTAGTTATACGCCCGTCACGAAGAAGGAGAATTTTAGCTGGTTTTATCTGCAGCGAATTTAGGGC M5 (sense) A-Es GTAGATGCCAATAGATTTGAAAGCTTGATAGAGAGAAG M5 (antisense) A-Ea CTTCTCTCTATCAAGCTTTCAAATCTATTGGCATCTAC M6 (sense) F-Es GCTGATGCAGCGGATGTAGATGCCAATAGAGAAGCAAGCTTGATAGAGAGAAGACTCAATTATAAGGCG M6 (antisense) F-Ea CGCCTTATAATTGAGTCTTCTCTCTATCAAGCTTGCTTCTCTATTGGCATCTACATCCGCTGCATCAGC M7 (sense) R-Es CAGCGGATGTAGATGCCAATGAATTTGCAAGCTTGATAGAGAG M7 (antisense) R-Ea CTCTCTATCAAGCTTGCAAATTCATTGGCATCTACATCCGCTG M8 (sense) A-Ns CTGATGCAGCGGATGTAGATAACAATAGATTTGCAAGCTTGATAG M8 (antisense) A-Na CTATCAAGCTTGCAAATCTATTGTTATCTACATCCGCTGCATCAG M9 (sense) D-Ts ACGCTGATGCAGCGGATGTAACTGCCAATAGATTTGCAAGCTTG M9 (antisense) D-Ta CAAGCTTGCAAATCTATTGGCAGTTACATCCGCTGCATCAGCGT

Probes: ssRNA (sense) AGR1 CGUACGCGGAAUACUUCGAUU–FAM ssRNA (antisense) AGR2 UCGAAGUAUUCCGCGUACGUU ssDNA (sense) AGD1 CGTACGCGGAATACTTCGATT–FITC ssDNA (antisense) AGD2 TCGAAGTATTCCGCGTACGTT

Received 8 October 2009/4 January 2010; accepted 5 January 2010 Published as BJ Immediate Publication 5 January 2010, doi:10.1042/BJ20091553

c The Authors Journal compilation c 2010 Biochemical Society