Mechanism for Template-Independent Terminal Adenylation Activity of Qβ Replicase

Structure

Article Mechanism for Template-Independent Terminal Adenylation Activity of Qb Replicase Daijiro Takeshita,1 Seisuke Yamashita,1 and Kozo Tomita1,2,* 1Biomedical Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1, Higashi, Tsukuba, Ibaraki 305-8566, Japan 2PRESTO, Japan Science and Technology Agency (JST), 4-1-8, Honcho, Kawaguchi, Saitama, 332-0012, Japan *Correspondence: [email protected] http://dx.doi.org/10.1016/j.str.2012.07.004

SUMMARY

The genomic RNA of Qb virus is replicated by Qb replicase, a template-dependent RNA polymerase complex. Qb replicase has an intrinsic templateindependent RNA 30 -adenylation activity, which is required for efficient viral RNA amplification in the host cells. However, the mechanism of the template-independent 30 -adenylation of RNAs by Qb replicase has remained elusive. We determined the structure of a complex that includes Qb replicase, a template RNA, a growing RNA complementary to the template RNA, and ATP. The structure represents the terminal stage of RNA polymerization and reveals that the shape and size of the nucleotidebinding pocket becomes available for ATP accommodation after the 30 -penultimate template-dependent C-addition. The stacking interaction between the ATP and the neighboring Watson-Crick base pair, between the 50 -G in the template and the 30 -C in the growing RNA, contributes to the nucleotide specificity. Thus, the template for the template-independent 30 -adenylation by Qb replicase is the RNA and protein ribonucleoprotein complex. INTRODUCTION Template-independent terminal nucleotide addition onto the 30 -end of RNAs is involved in various cellular functions, and is catalyzed by RNA-specific terminal nucleotidyltransferases. Template-independent CCA-addition onto the 30 -end of tRNA, by CCA-adding enzymes, is required for amino acid attachment to the 30 -terminus of tRNA by aminoacyl tRNA synthetases (Sprinzl and Cramer, 1979), as well as for peptide bond formation on ribosomes (Green and Noller, 1997; Kim and Green, 1999; Nissen et al., 2000). Template-independent polyA addition by polyA polymerase onto mRNAs is involved in mRNA stability, maturation, and translation (Beelman and Parker, 1995; Sachs et al., 1997; Wahle and Ru¨egsegger, 1999). In some cases, polyA addition onto various RNAs acts as an RNA-destabilizing signal, and is involved in RNA quality control (Carpousis et al., 1999; Dreyfus and Re´gnier, 2002; Re´gnier and Hajnsdorf, 2009). Terminal nucleotidyltransferase activity is reportedly associated with several RNA-dependent RNA polymerases (RdRps)

from single-stranded RNA viruses (see references in Poranen et al., 2008). The intrinsic terminal nucleotidyltransferase activity of some viral RdRps affects the initiation of RNA synthesis (Ball, 1995; Miller et al., 1985; Kao et al., 2000; Oh et al., 2000). In some viruses, RNA polymerization is initiated by using template RNAs containing additional nucleosides, and in others, the initiation starts from a nucleoside that is not present at the very 30 -terminus of the template RNA. The template-independent nucleotide addition by viral RdRps onto the 30 -ends of the viral RNAs is also advantageous for the virus, to protect the RNAs from nucleases in the host cells, and to preserve the viral genome information (Ranjith-Kumar et al., 2001; Burgya´n and Garcı´a-Arenal, 1998; Guan and Simon, 2000; Teramoto et al., 2008). Qb virus has a single, positive strand genomic RNA, and infects Escherichia coli. Qb viral genomic RNA is replicated and amplified in the infected cells by Qb replicase (RanjithKumar et al., 2001), a tetrameric protein complex comprising the Qb virus-encoded catalytic RdRp (b-subunit) and the host E. coli translational elongation factors (EFs) -Tu, -Ts, and ribosomal protein S1 (Blumenthal and Carmichael, 1979; Blumenthal et al., 1972; Wahba et al., 1974). Recent crystallographic analyses of the structures of Qb replicase, representing several elongation stages of RNA polymerization, revealed that EF-Tu in Qb replicase can modulate processive RNA polymerization at the elongation stages, beyond its established function in protein synthesis (Takeshita and Tomita, 2010; Kidmose et al., 2010; Takeshita and Tomita, 2012). EF-Tu in the complex facilitates the separation of the double helix between the template and growing RNAs at the elongation stage, and composes the exit channel for the single-stranded template RNA (Takeshita and Tomita, 2012). Qb genomic RNA has a tRNA-like structure at the 30 -region, and the 30 -end sequence is CCA-30 , as in every tRNA (Dahlberg, 1968; Weissmann et al., 1973). The 30 -A of Qb virus genomic RNA does not serve as a template. Instead, the 30 -penultimate cytidine functions as the first template nucleoside, and RNA polymerization by Qb replicase starts with GTP, without an RNA primer (de novo initiation) (Blumenthal and Carmichael, 1979; Silverman, 1973; Blumenthal, 1980). A recent crystallographic analysis of the structure of Qb replicase, representing the initiation of RNA polymerization, revealed that the 30 -A of Qb genomic RNA could provide a stable platform for the construction of a priming complex, composed of the template RNA, the priming GTP and the incoming GTP (Takeshita and Tomita, 2012). The 30 -A of the template RNA stacks onto the first

Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved 1661

Structure 0

3 -Adenylation of RNA by Qb Replicase

Watson-Crick base pair between the priming GTP and the 30 -penultimate C in the template, and the guanine base of the priming GTP stacks onto the guanine base of the incoming GTP, thus stabilizing the initiation complex. The sequence of the 50 -end of the positive strand Qb viral genomic RNA is 50 -GG (Blumenthal and Carmichael, 1979). After the synthesis of the negative strand, using the positive strand in a template-dependent manner, a single adenosine residue is incorporated, in a template-independent manner, onto the 30 -end of the negative strand RNA by the intrinsic terminal nucleotidyltransferase activity of Qb replicase. This results in the synthesis of the negative strand RNA with the CCA-30 sequence, which is, in turn, required for the efficient initiation of positive strand RNA synthesis from the negative strand RNA. Thus, the template-independent 30 -terminal adenylation of the positive and negative strand RNAs of Qb virus is essential for the efficient replication of the genomic RNA in the infected cell. The template-independent terminal adenylation by Qb replicase reportedly occurs prior to the release of the product RNAs from the enzyme, and a Qb viral RNA lacking its 30 -A could not be adenylated (Blumenthal and Carmichael, 1979; Weber and Weissmann, 1970). Furthermore, a particular sequence at the 50 end of the template RNA is reportedly required for the template-independent terminal adenylation (Bausch et al., 1983). However, the mechanism of the template-independent terminal adenylation by Qb replicase has remained elusive. Here, we present crystal structures representing the terminal stage of RNA synthesis by Qb replicase, including the template-dependent 30 -penultimate C-addition and the template-independent 30 -terminal A-addition. These structures, along with biochemical studies using the mutant RNAs, provide the molecular basis for the terminal adenylation activity of Qb replicase. The specificity of the template-independent adenylation at the 30 -end of the RNA is determined by the enzyme and the 30 -blunt-ended double helix between the template and growing RNAs in a collaborative manner, and the sequential reactions at the termination stage of RNA polymerization determine the nucleotide specificity, by reshaping the nucleotide pocket formed by the ribonucleotide-protein (RNP) complex of RNA and proteins. RESULTS Structure of Template-Independent Terminal Adenylation To clarify the molecular basis for the template-independent 30 -terminal adenylation activity of Qb replicase, the core Qb replicase, comprising the b-subunit, EF-Tu and EF-Ts, was crystallized with a template RNA possessing a 50 -guanosine (50 -G1) and a growing RNA complementary to the 30 -part of the template RNA, except for the 50 -G1, in the presence of a nucleotide (ATP) and a nucleotide analog (30 -deoxy CTP, 30 -dCTP). Prior to the crystallization, the crystallization mixture was incubated in the presence of magnesium ions, to allow Qb replicase to transfer one 30 -dCTP to the 30 end of the growing RNA in a templatedependent manner (Figure 1A). In the determined structure (Table 1), the template and growing RNAs form a double helix, which is directed toward EF-Tu in the core Qb replicase (Figure 1B). In the structure,

one 30 -dCTP is transferred onto the 30 -end of the growing RNA in a template-dependent manner, and the ATP is positioned as the incoming nucleotide in the catalytic pocket. The geometries of the two metals, the ATP, and the 30 -dC incorporated into the growing RNA, relative to those of the three catalytic carboxylates (Asp273, Asp358, and Asp359), suggested that the structure represents the stage of template-independent adenylation, where the 30 -OH group of the growing RNA is ready to attack the triphosphate of the incoming ATP (hereafter, this structure is referred to as the ‘‘30 -A stage,’’ Figure 1C; see Figures S1A and S1B available online). The positions of the triphosphate moiety of ATP, the two metals, and the 30 -dC of the growing RNA, relative to those of the three catalytic carboxylates, are almost identical to those observed in the structures representing the template-dependent elongation stages of RNA polymerization (9nt-, 10nt-, and 14ntstages in Takeshita and Tomita, 2012). However, the base and ribose moieties of the ATP in the 30 -A stage are shifted slightly toward the center of the double helix formed by the template and growing RNAs (Figure 1D), as compared to the position of the incoming nucleotide in the structure representing the template-dependent elongation stage (9nt-stage, Figure 1E). The shifting of the base and ribose moieties of the ATP in the 30 -A stage may be due to the absence of the template nucleoside, as described below. In the structures representing the template-dependent elongation stages, the 30 -OH group of the ATP ribose hydrogen bonds with the Od atom of Ser278 and the Oε1 atom of Glu331, and the 20 -OH group of the ATP ribose hydrogen bonds with the Oε2 atom of Glu331. However, due to the slight shift of the ATP toward the center of the RNA double helix, the ribose of the ATP forms a van der Waals interaction with the cytosine base at the 30 -terminal cytidine of the growing RNA (Figures 1C and 1D). As a result, the 30 -cytidine in the growing RNA, which forms Watson-Crick hydrogen bonds with the 50 -G1 of the template RNA, is twisted by 14 degrees (Figure 1C). In the 30 A stage structure, the 50 -part of the template RNA is only recognized by Arg152. Thus, the few interactions between the 50 -part of the template RNA and Qb replicase result in the twisting of the base of the 30 -cytidine in the growing RNA, upon ATP binding in the pocket. In the 30 -A-stage structure, the six-membered adenine ring of the ATP is sandwiched between the six-membered guanine ring of the 50 -G1 and the side chain of Met322. The Sd atom of Met322 hydrogen bonds with the Nh2 atom of Arg219, and the Nh1 atom of Arg219 hydrogen bonds with the N7 atom of ATP. Therefore, this hydrogen-bond network would further stabilize the stacking interaction between the side chain of Met322 and the adenine base of ATP. The adenine base of ATP also forms van der Waals interactions with the main chain amino group of Gly323 and the side chain of Ile221, which compose the adenine-binding pocket, together with the guanine base of 50 -G1 and the side chain of Met322. Selection of ATP in a Template-Independent Manner The structure of the 30 -A stage (Figures 1C and 1D) suggested that the size and shape of the nucleobase-interacting pocket in the catalytic site, formed at the 30 -A stage, are suitable for the accommodation of ATP. In addition, the stacking interaction

1662 Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved

Structure 30 -Adenylation of RNA by Qb Replicase

between the six-membered adenine ring of the ATP and the six-membered guanine ring of the 50 -G1 in the template RNA further stabilizes the ATP binding in the pocket. These structural features can explain the previous observation that the 30 -terminal nucleoside of the RNA replicated from a positive strand genomic RNA is predominantly adenosine (Rensing and August, 1969). These features are also consistent with the previous results showing that the RNAs replicated by Qb replicase have the A-30 terminus (Schaffner et al., 1977). Our biochemical studies revealed that the affinity of ATP for the template-independent nucleotide addition is, indeed, higher than those of the other three nucleotides (Figures 2A, 2B, and S2). The Km value of ATP for template-independent nucleotide addition is 83 mM. On the other hand, the Km values of CTP and UTP for template-independent nucleotide addition are larger than 1 mM, and the Km value of GTP is 397 mM. ATP is snugly accommodated within the pocket, in a template-independent manner (Figure 2C). Models of nucleotide-binding in the pocket revealed that, although the smaller pyrimidines (CTP and UTP) can also bind in the pocket, they do not snugly fit, and the stacking interactions between the pyrimidine nucleobase, the guanine base of 50 -G1 in the template RNA, and the side chain of Met322 are weaker than those observed in the template-independent 30 -A stage (Figures 2D and 2E). In the GTP binding model (Figure 2F), the 2-NH2 group of the guanine base of GTP would clash with the main chain of Gly323. The Km value of GTP (397 mM) is larger than that of ATP (83 mM), but it is significantly smaller than those of CTP and UTP. This suggests that GTP could bind within the pocket. Consistent with this model, the electron density of GTP was visible in the pocket of the structure representing template-independent G-addition (30 -G stage, Figures 2G and 2H), although the 2-NH2 group of guanine of the GTP is close to the main chain amino of Gly323 (distance is 2.5 A˚). Although the GTP binding in the nucleotide pocket is cramped, the strong stacking interactions between the GTP and the neighboring Watson-Crick base pairs contribute to the affinity of GTP toward the pocket, as described below. Our biochemical analysis showed that the b-subunit mutants, such as Met322Ala, Ile221Ala, and Arg219Ala, all completely lost the template-dependent RNA polymerization activity (Figure S3). The loss of the activity by the mutation of noncatalytic residues is described below. Since the template-dependent nucleotide incorporation and the template-independent terminal A-addition are sequential and coupled (Bausch et al., 1983), the assessment of these residues in the catalytic pocket of the b-subunit (Met322, Ile221, and Arg219) on the terminal adenylation of RNA was not pursued.

Figure 1. Structure Representing Template-Independent Terminal A-Addition (30 -A Stage) (A) Procedure for preparing the crystal representing the 30 -A stage. (B) Overall structure of the 30 -A stage. The b-subunit, EF-Tu, and EF-Ts are colored yellow, red, and blue, respectively. The template and growing RNAs are depicted by blue and red stick models, respectively. The calcium ion in the structure of the b-subunit is colored gray. (C) A stereo view of the structure of template-independent ATP binding at the 30 -A stage.

The Role of the 50 -Part of the Template RNA in TemplateIndependent Adenylation The sequence of the 30 -end of most RNAs replicated by Qb replicase is CCA-30 , and that of the 50 end is 50 -GG (Schaffner et al., 1977; Biebricher and Luce, 1993). The 30 -template-independent adenylation is considered to proceed in a template-determined (D and E) Comparison of the structures representing the 30 -A stage and the 9nt-stage. RNAs and nucleotides in the 30 -A stage (D) and the 9nt-stage (E). Proteins are omitted from the figures, for clarity. See also Figure S1.

Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved 1663

Structure 0

3 -Adenylation of RNA by Qb Replicase

Table 1. Data Collection and Refinement Statistics 30 -A Stage

30 -Penultimate C Stage

30 -G Stage

C2221

C2221

C2221

139.00, 255.92, 101.39

139.73, 256.04, 101.72

137.51, 255.05, 100.90

Resolution (A˚)a

50–3.2 (3.26– 3.20)

50–2.6 (2.64– 2.60)

50–3.2 (3.26– 3.20)

Rsym

0.112 (0.361)

0.098 (0.428)

0.086 (0.278)

I / sI

10.2 (2.5)

28.6 (2.9)

12.4 (3.0)

Completeness (%)

95.9 (87.0)

99.4 (98.9)

90.1 (79.4)

Redundancy

6.0 (2.4)

9.2 (5.6)

5.5 (2.3)

Resolution (A˚)

20–3.2

20–2.6

20–3.2

No. reflections

27520

55640

25474

Rwork / Rfree

27.4/31.6

22.0/27.9

24.5/31.4

9287

9287

9287

Data collection Space group Cell dimensions a, b, c (A˚)

Refinement

No. atoms Protein Ligand/ion

283/2

344/2

285/2

Water

12

47

8

Bond lengths (A˚)

0.010

0.010

0.010

Bond angles ( )

1.3

1.3

1.4 78.6

Rmsds

Ramachandran plot Most favored(%)

81.3

86.2

Allowed (%)

18.3

12.9

20.7

Generously allowed (%)

0.5

0.9

0.8

Disallowed (%)

0.0

0.0

0.0

a

Values in parentheses are for highest resolution shell.

manner, and to occur during RNA synthesis (Bausch et al., 1983). The adenylation of the 30 -end of the RNA does not occur with a template RNA lacking the 50 -GG, suggesting that a particular sequence and/or structure at the 50 -part of the template RNA is required for adenylation to occur (Bausch et al., 1983). We examined the involvement of the base-pairing between the 50 -part of the template RNA and the 30 -part of the growing RNAs, in the template-independent terminal adenylation of the RNA. We also investigated the influence of the stacking interactions between the nucleobases in the template RNA within the double helix of the template and growing RNAs, in the template-independent terminal adenylation of RNA (Figures 3A and 3B). A template RNA with either 50 -C or 50 -U is less competent (40–50% of that of RNA with 50 -G) for template-independent terminal A-addition (Figure 3B). A template RNA with 50 -A is also less competent, but still retained about 70% competency relative to the RNA with 50 -G. RNAs with mutations of the 50 -second G (G2) to either C (C2) or U (U2) exhibited reduced template competency for template-independent terminal A-addition (40–50% of that of RNA with 50 -G2), but an RNA with

a 50 -second A (A2) was as competent as the template RNA with G2 for template-independent terminal A-addition (Figure 3B). An RNA with either 50 -U1U2, 50 -C1C2 or 50 -A1A2 is less competent for template-independent terminal A-addition (Figure 3B). Taken together, these results indicated that the successive 50 -purine–purine coaxial nucleobase stacking in the template RNA in the double helix and the strong G:C base-pairing between the 50 -nucleobase of the template RNA and the 30 -base of the growing RNA are required for efficient templateindependent 30 -terminal adenylation. The Watson-Crick hydrogen bonds between the 50 -G1 in the template RNA and the 30 -C of the growing RNA would stack with the adenine of an incoming ATP, thus stabilizing ATP accommodation in the pocket. These structural and biochemical results are consistent with the previous observations (Bausch et al., 1983, Schaffner et al., 1977, Biebricher and Luce, 1993). RNAs with 50 -A1G2 or 50 -G1A2 are competent for the terminal template-independent 30 -A addition (about 70% and 100% competency relative to the RNA with 50 -G1G2, respectively) (Figure 3B). The RNAs ending with CUA-30 and UCA-30 , which are synthesized from the template RNAs with 50 -A1G2 and 50 -G1A2, respectively, are less competent as templates at the initiation stage of RNA polymerization (Takeshita and Tomita, 2012). Thus, only the RNAs with 50 -G1G2 and CCA-30 are predominantly amplified during the exponential replication cycles, and they are isolated as the replicated RNAs by Qb replicase (Bausch et al., 1983). Considering the size and shape of the nucleotide-binding pocket, the stacking interaction between ATP and the neighboring Watson-Crick base pairs, and the successive G-C Watson-Crick base-pairings in the double helix between the template and growing RNAs, ATP would predominantly be selected, as a consequence of the competition between NTPs for the nucleotide-binding pocket. Structure of Template-Dependent 30 -Penultimate C-Addition To clarify the dynamic transition from template-dependent to -independent RNA polymerization by Qb replicase at the terminal stage of RNA polymerization, the core Qb replicase was also crystallized with a template RNA with 50 -G1, and a growing RNA complementary to the 30 -part of the template RNA, except for the 50 -terminal G1, in the presence of only 30 -dCTP. In this crystallization, magnesium ions were omitted from the solution (Figure 4A). In the determined structure, the template and growing RNAs form a double helix, and one 30 -dCTP is positioned as an incoming nucleotide in the catalytic pocket, and forms a Watson-Crick base pair with the guanine of 50 -G1 in the template RNA (hereafter, this structure is referred to as the ‘‘30 -penultimate C stage’’). The geometries of the two metals, the 30 -dCTP, and the 30 -terminal C of the growing RNA, relative to those of the three catalytic carboxylates (Asp273, Asp358, and Asp359), are almost identical to those observed in the other elongation stages of RNA polymerization by Qb replicase (Figures 4B, S1C, and S1D). In this stage, the phosphate backbone of the 50 -penultimate G (G2) and the 50 -third G (G3) in the template RNA hydrogen-bond with the main chains of Gly156 and Arg152, respectively. The side chain of Ile221 in the b strand (b2) stacks onto the 50 -G1 in the template RNA. The side chain of Arg219 in the b2 strand and

1664 Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved

Structure 30 -Adenylation of RNA by Qb Replicase

Figure 2. Selection of ATP for TemplateIndependent Terminal Adenylation by Qb Replicase (A) Procedure for assaying template-independent terminal nucleotide addition. (B) Km values of nucleotides for template-independent terminal nucleotide addition. The concentration of the respective nucleotide was titrated (0–1,000 mM) under conditions where the concentrations of the other nucleotides were fixed at 50 mM. The reaction solutions were incubated for 15 min. Under these conditions, the reactions proceeded in linear ranges (Figure S2), and the nucleotides other than ATP were not incorporated in template-independent manners. The efficiencies of template-independent nucleotide addition onto the RNAs were calculated as the fraction of the template-independent, terminally nucleotideadded RNA (arrows, upper bands in the gels) out of the total template-dependent RNA polymerization (lower bands in the gels). The asterisks in the gel panel indicate the bands where two nucleotides were incorporated in a template-independent manner. The bars in the graphs are the standard deviations of two independent experiments. The Km values for ATP and GTP were 83 and 397 mM, respectively. The Km values for CTP and UTP are larger than 1 mM. (C) Structure of ATP binding in the catalytic pocket. (D–F) Nucleotide-binding models in the catalytic pocket. CTP (D), UTP (E), and GTP (F). Nucleotides were manually modeled and are depicted by stick models. The RNA double helixes of the template and growing RNAs are shown as line models, for clarity. (G) Structure of 30 -G addition (30 -G stage) in a template-independent manner. The 2Fo-Fc map of GTP in the structure of the 30 -G stage, contoured at 1.0 s, is colored gray. (H) Superposition of the structures of the 30 -A (colored cyan) stage and the 30 -G stage (colored yellow). See also Figure S2.

the side chain of Met322 stack onto the base of the incoming CTP, and Arg219 also hydrogen bonds with the triphosphate of the incoming CTP, as observed in the other template-dependent elongation stages. Ile221 in b2 anchors the nucleobase (50 -G1) of the template RNA. As a result, the incoming nucleotide (CTP) complementary to the template nucleobase is accommodated in the catalytic pocket, and Arg219 and Met322 participate in the recognition and stabilization of the incoming nucleotide. Consistent with this, as described, the Met322Ala, Ile221Ala, and Arg219Ala mutants of the b-subunit lost the template-dependent RNA polymerization activity (Figure S3). These amino acid residues in the catalytic pocket, besides the catalytic residues (Asp273, Asp358, and Asp359), are required for the templatedependent RNA polymerization and also participate in the template-independent adenylation (Figure 2). Transition from Template-Dependent to -Independent Nucleotidyltransfer A comparison of the 30 -A and 30 -penultimate C stage structures revealed the conformational differences in the catalytic pocket of

the b-subunit at the terminal stage of RNA polymerization (Figures 4C, S4A, and S4B), and thus provided a mechanism for the detailed sequential transition from template-dependent to -independent RNA polymerization by Qb replicase (Figure 5). In the 30 -penultimate C stage, the side chain of Ile221 in b2 stacks onto the 50 -G1 in the template RNA (Figures 4C and 5A). After template-dependent 30 -C addition, the template and growing RNA double helix translocates toward EF-Tu in the core Qb replicase (Figure 5B). As a result, the side chain of Ile221 moves toward the double helix of the template and growing RNAs by 2.5 A˚, due to the absence of the template nucleoside, and participates in the formation of an ATP binding pocket at the 30 -A stage (Figures 1C, 4C, and 5C). In the transition, the side chain of Arg219, which stacks onto the base of the incoming CTP at the 30 -penultimate stage, rotates, and the Nh2 atom of Arg219 hydrogen bonds with the Sd atom of Met322, to stabilize the stacking interaction between the side chain of Met322 and the adenine base of ATP at the 30 -A stage (Figure 1C). After 30 -terminal adenylation in a template-independent manner,

Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved 1665

Structure 0

3 -Adenylation of RNA by Qb Replicase

Figure 3. In Vitro Terminal Adenylation Using Mutant Template RNAs (A) Procedures for assays of template-independent terminal adenylation. The 50 -G1G2 sequence (boxed) of DN26 RNA was mutagenized. (B) The efficiencies of template-independent 30 -A addition onto RNA variants after a 15 min incubation. The concentration of nucleotides was 50 mM. The efficiencies of template-independent adenylation onto the RNAs were calculated as the fraction of the template-independent, terminally adenylated RNA (in the presence of ATP: upper bands in the gels) out of the total templatedependent RNA polymerization (in the absence of ATP in the reactions: lower bands in the gels). The efficiencies of the template-independent adenylation of the template RNA were normalized, relative to the efficiency of the RNA with 50 -G1G2 as 1.0. The bars in the graphs are standard deviations of three independent experiments. See also Figure S3.

pyrophosphate release would drive the double helix of the template and growing RNAs toward EF-Tu in the complex. The absence of the preceding neighboring Watson-Crick base-pairing between the template and growing RNAs would result in the termination of the RNA polymerization. Thus, at the transition, the catalytic pocket structure is sequentially reshaped by the preceding base pair between the template and growing RNAs and the protein, in a collaborative manner. DISCUSSION Template-independent terminal adenylation of Qb viral genomic RNA by Qb virus is required for efficient replication of the genomic RNA (Takeshita and Tomita, 2012). The present study provided a detailed molecular basis for the template-independent adenylation of RNA by Qb replicase. The nucleotide specificity is dictated by the size and shape of the nucleotide pocket, which is composed of the amino acid residues and the neighboring Watson-Crick base pair between the 50 -nucleoside (G) in the template RNA and the 30 -nucleoside (C) of the growing RNA, after 30 -penultimate template-dependent C-addition. In addition, the stacking interaction between an incoming nucleotide and the neighboring Watson-Crick base-pairing contributes toward determining the specificity of the terminal adenylation. More importantly, in Qb replicase, the sequential reaction at the termination stage of RNA polymerization determines the nucleotide specificity, by reshaping the nucleotide pocket formed by the protein-RNA complex.

The mechanisms of template-independent 30 -terminal adenylation by RNAspecific terminal nucleotidyltransferases, such as CCA-adding enzymes and polyA polymerases, have been extensively studied and characterized (Shi et al., 1998; Xiong and Steitz, 2004; Tomita et al., 2004, 2006; Toh et al., 2009, 2011; Pan et al., 2010; Martin et al., 2000; Bard et al., 2000). The specificity of the template-independent A-addition at position of 76 of tRNA by the CCA-adding enzymes is dictated by the interactions between the 30 -end of the tRNA and the enzymes. In the eubacterial CCA-adding enzymes, the conformations of two key amino acid residues (Asp and Arg) interacting with the adenine base become suitable for accommodating only ATP in the pocket, by the interaction between the flexible loop and the b-turn in the catalytic domain and the growing CC-30 of the RNA primer (Tomita et al., 2004, Toh et al., 2009). In the archaeal CCA-adding enzymes, the size and shape of the nucleobase-interacting pocket become suitable only for ATP, by the interaction between the 30 -CC of the RNA and the proteins, and ATP is recognized by a single Arg residue and the phosphate group of the 30 -end of the RNA primer (Xiong and Steitz, 2004, Tomita et al., 2006, Pan et al., 2010). On the other hand, the eubacterial and eukaryotic polyA polymerases may only use the protein to select ATP (Toh et al., 2011, Martin et al., 2000, Bard et al., 2000), and the specific pocket for ATP exists in the catalytic site of the enzymes, even in the absence of RNA primers. Although the detailed mechanisms for the template-independent 30 -terminal A-addition by Qb replicase and the CCA-adding enzymes are different, as described above, both enzymes utilize the common strategy of forming a specific nucleotide pocket by RNA and proteins in a collaborative manner, and during the sequential nucleotide addition, the RNA and protein together remodel the pocket structure. Thus, the mechanism for the template-independent adenylation by Qb replicase is similar to those of the CCA-adding enzymes, rather than those of the polyA

1666 Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved

Structure 30 -Adenylation of RNA by Qb Replicase

Figure 4. Transition from Template-Dependent to -Independent Nucleotide Addition (A) Procedure for preparing the crystal representing the 30 -penultimate C stage. (B) A stereo view of the structure of template-dependent CTP binding at the 30 -penultimate C stage. (C) A stereo view comparing the RNA structures at the 30 -penultimate C stage (colored magenta) and the 30 -A stage (colored cyan). The b2 regions in the b-subunit and the RNAs are depicted by stick models. See also Figures S1 and S4.

polymerases. The intrinsic terminal nucleotidyltransferase activity of Qb replicase, and perhaps those of other viral RNA-dependent RNA polymerases, might be evolutionarily more closely related to those of the CCA-adding enzymes. It was previously proposed that a tRNA-like structure with the CCA-30 of RNA might function as a tag for RNA replication, and the CCA-30 might serve as a telomere to ensure the maintenance of the RNA information, and thus the CCA-adding activity would have represented an ancient telomerase activity (Weiner and Maizels, 1987). The intrinsic terminal nucleotidyltransferase activity of Qb replicase is required for efficient genomic RNA amplification. This activity also contributes to the maintenance of the genomic information during RNA replication, and the template for template-independent 30 -adenylation by Qb replicase is the ribonucleoprotein complex of RNA and protein, as in the CCA-adding enzymes. Thus, the intrinsic terminal nucleotidyltransferase activity of Qb replicase, and

Figure 5. Cartoon Representing Template-Independent Terminal Adenylation by Qb Replicase (A) Template-dependent 30 -penultimate C addition. (B) Translocation of the double helix after 30 -penultimate C addition, followed by the conformational change of b2. (C) Template-independent terminal A-addition.

those of other viral RdRps, might represent relics of a primitive telomerase. EXPERIMENTAL PROCEDURES Crystallization of Complexes A single-chained core Qb replicase, comprising EF-Ts, EF-Tu, and the b-subunit (Kita et al., 2006), was expressed and purified as described

Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved 1667

Structure 0

3 -Adenylation of RNA by Qb Replicase

(Takeshita and Tomita, 2010). To obtain a crystal representing the templateindependent adenylation stage (30 -A stage, Figure 1A), equimolar amounts of the template and growing RNAs were mixed and incubated at 70 C, and then slowly cooled to room temperature. The annealed template and growing RNAs (45 mM) were mixed with 40 mM core Qb in the presence of 0.8 mM 30 -dCTP and ATP, in a solution containing 10 mM Tris-Cl, pH 8.0, 200 mM NaCl, and 5 mM MgCl2, and the mixtures were incubated on ice for 15 hr. To obtain a crystal representing the template-dependent cytidyltransfer stage (30 -penultimate C stage, Figure 3A), the annealed template and growing RNAs (45 mM) were mixed with 40 mM core Qb in the presence of 0.8 mM 30 -dCTP, in the solution described above, except that MgCl2 was omitted. One microliter of the complex was mixed with a reservoir solution, containing 100 mM MES, pH 5.6–6.0, 200 mM Ca(OAc)2 and 28–30% (v/v) polyethylene glycol (PEG)400, and crystallized at 20 C by the sitting drop vapor-diffusion method, using the reservoir solution. A crystal representing the template-independent guanylation stage (30 -G stage, Figure 2G) was also obtained under the same conditions, as the crystal representing the 30 -A stage, where ATP was replaced with GTP. Structure Determination of Complexes The crystals were cryo-protected with 25% (v/v) PEG-400 and were flashcooled in a 100 K nitrogen stream. All data were processed using the HKL2000 program (Otwinowski and Minor, 1997). Crystal structures were solved at 2.6–3.2 A˚ resolutions by molecular replacement with the program MolRep (Vagin and Teplyakov, 2000), using the core Qb replicase structure (Takeshita and Tomita, 2010) as a search model. All complex crystals belonged to the space group C2221, and contained one complex with RNAs in the asymmetric unit cell. The initial models for RNAs and/or nucleotide(s) in the complexes were manually built using the program Xtalview/Xfit (McRee, 1999), with iterative cycles of refinement with CNS (Bru¨nger et al., 1998) and PHENIX (Adams et al., 2002). The refinement statistics are provided in Table 1. In Vitro Terminal Adenylation The activity of template-independent terminal adenylation of RNA by the core Qb replicase was assayed, using DN26 as an RNA template. DN26 is derived from the complementary sequence of DN3 (Zamora et al., 1995), and the DN26 RNA sequence is 50 -GGGCCCAAGAGGGGCCAGAAAACCCA-30 . The DN26 variants (starting with 50 -C1G2, 50 -A1G2, 50 -U1G2, 50 -G1C2, 50 -G1A2, 50 -G1U2, 50 -C1C2, 50 -A1A2, 50 -U1U2) were employed for the assays using the core Qb replicase. The RNAs used for the assays were purchased from Sigma Aldrich, Japan. The reaction solutions, containing 80 mM Tris-Cl, pH 8.0, 10 mM MgCl2, 1 mM DTT, 190 nM DN26 RNA (or its variant), 50 mM NTP, 100 nM a-32P UTP (GE Healthcare, Japan, 3,000 ci/mmol), and 190 nM core Qb replicase, were incubated for 15 min at 37 C. At this point, RNA polymerization proceeds in a linear range (Figure S2). The solutions without ATP were incubated to monitor only the template-dependent RNA polymerization. The reactions were stopped by adding EDTA (pH 8.0) to a final concentration of 25 mM, and were treated with phenol/chloroform. The RNAs were ethanol-precipitated and dissolved in a buffer containing 90% (v/v) formamide and 10 mM EDTA (pH 8.0). The RNAs were heat-denatured at 95 C for 10 min, chilled on ice, and separated by 15% (w/v) polyacrylamide gel electrophoresis under denaturing conditions. The intensities of the 32P-labeled products were analyzed and quantified by a BAS2500 Bio-Image Analyzer (Fuji Film, Japan). ACCESSION NUMBERS The Protein Data Bank accession numbers for the atomic coordinates and structural factors reported in this paper are 3VNU, 3VNV, and 4FWT. SUPPLEMENTAL INFORMATION Supplemental Information includes four figures and can be found with this article online at http://dx.doi.org/10.1016/j.str.2012.07.004. ACKNOWLEDGMENTS We thank the beam-line staff of BL-17A (KEK, Tsukuba) for technical assistance during data collection. This work was supported by grants to K.T.

from the PRESTO program of JST, Funding Program for Next Generation World-Leading Researchers (NEXT program) of JSPS, Takeda Science Foundation, Mitsubishi Foundation, and Cell Science Research Foundation. Received: April 19, 2012 Revised: June 29, 2012 Accepted: July 15, 2012 Published online: August 9, 2012 REFERENCES Adams, P.D., Grosse-Kunstleve, R.W., Hung, L.W., Ioerger, T.R., McCoy, A.J., Moriarty, N.W., Read, R.J., Sacchettini, J.C., Sauter, N.K., and Terwilliger, T.C. (2002). PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr. D Biol. Crystallogr. 58, 1948–1954. Ball, L.A. (1995). Requirements for the self-directed replication of flock house virus RNA 1. J. Virol. 69, 720–727. Bard, J., Zhelkovsky, A.M., Helmling, S., Earnest, T.N., Moore, C.L., and Bohm, A. (2000). Structure of yeast poly(A) polymerase alone and in complex with 30 -dATP. Science 289, 1346–1349. Bausch, J.N., Kramer, F.R., Miele, E.A., Dobkin, C., and Mills, D.R. (1983). Terminal adenylation in the synthesis of RNA by Q beta replicase. J. Biol. Chem. 258, 1978–1984. Beelman, C.A., and Parker, R. (1995). Degradation of mRNA in eukaryotes. Cell 81, 179–183. Biebricher, C.K., and Luce, R. (1993). Sequence analysis of RNA species synthesized by Q beta replicase without template. Biochemistry 32, 4848– 4854. Blumenthal, T. (1980). Q b replicase template specificity: different templates require different GTP concentrations for initiation. Proc. Natl. Acad. Sci. USA 77, 2601–2605. Blumenthal, T., and Carmichael, G.G. (1979). RNA replication: function and structure of Qbeta-replicase. Annu. Rev. Biochem. 48, 525–548. Blumenthal, T., Landers, T.A., and Weber, K. (1972). Bacteriophage Q replicase contains the protein biosynthesis elongation factors EF Tu and EF Ts. Proc. Natl. Acad. Sci. USA 69, 1313–1317. Bru¨nger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., GrosseKunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. (1998). Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. Burgya´n, J., and Garcı´a-Arenal, F. (1998). Template-independent repair of the 30 end of cucumber mosaic virus satellite RNA controlled by RNAs 1 and 2 of helper virus. J. Virol. 72, 5061–5066. Carpousis, A.J., Vanzo, N.F., and Raynal, L.C. (1999). mRNA degradation. A tale of poly(A) and multiprotein machines. Trends Genet. 15, 24–28. Dahlberg, J.E. (1968). Terminal sequences of bacteriophage RNAs. Nature 220, 548–552. Dreyfus, M., and Re´gnier, P. (2002). The poly(A) tail of mRNAs: bodyguard in eukaryotes, scavenger in bacteria. Cell 111, 611–613. Green, R., and Noller, H.F. (1997). Ribosomes and translation. Annu. Rev. Biochem. 66, 679–716. Guan, H., and Simon, A.E. (2000). Polymerization of nontemplate bases before transcription initiation at the 30 ends of templates by an RNA-dependent RNA polymerase: an activity involved in 30 end repair of viral RNAs. Proc. Natl. Acad. Sci. USA 97, 12451–12456. Kao, C.C., Yang, X., Kline, A., Wang, Q.M., Barket, D., and Heinz, B.A. (2000). Template requirements for RNA synthesis by a recombinant hepatitis C virus RNA-dependent RNA polymerase. J. Virol. 74, 11121–11128. Kidmose, R.T., Vasiliev, N.N., Chetverin, A.B., Andersen, G.R., and Knudsen, C.R. (2010). Structure of the Qbeta replicase, an RNA-dependent RNA polymerase consisting of viral and host proteins. Proc. Natl. Acad. Sci. USA 107, 10884–10889. Kim, D.F., and Green, R. (1999). Base-pairing between 23S rRNA and tRNA in the ribosomal A site. Mol. Cell 4, 859–864.

1668 Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved

Structure 30 -Adenylation of RNA by Qb Replicase

Kita, H., Cho, J., Matsuura, T., Nakaishi, T., Taniguchi, I., Ichikawa, T., Shima, Y., Urabe, I., and Yomo, T. (2006). Functional Qbeta replicase genetically fusing essential subunits EF-Ts and EF-Tu with beta-subunit. J. Biosci. Bioeng. 101, 421–426. Martin, G., Keller, W., and Doublie´, S. (2000). Crystal structure of mammalian poly(A) polymerase in complex with an analog of ATP. EMBO J. 19, 4193– 4203. McRee, D.E. (1999). XtalView/Xfit—A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol. 125, 156–165. Miller, W.A., Dreher, T.W., and Hall, T.C. (1985). Synthesis of brome mosaic virus subgenomic RNA in vitro by internal initiation on (-)-sense genomic RNA. Nature 313, 68–70. Nissen, P., Hansen, J., Ban, N., Moore, P.B., and Steitz, T.A. (2000). The structural basis of ribosome activity in peptide bond synthesis. Science 289, 920–930. Oh, J.W., Sheu, G.T., and Lai, M.M.C. (2000). Template requirement and initiation site selection by hepatitis C virus polymerase on a minimal viral RNA template. J. Biol. Chem. 275, 17710–17717. Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. Pan, B., Xiong, Y., and Steitz, T.A. (2010). How the CCA-adding enzyme selects adenine over cytosine at position 76 of tRNA. Science 330, 937–940. Poranen, M.M., Koivunen, M.R., and Bamford, D.H. (2008). Nontemplated terminal nucleotidyltransferase activity of double-stranded RNA bacteriophage phi6 RNA-dependent RNA polymerase. J. Virol. 82, 9254–9264. Ranjith-Kumar, C.T., Gajewski, J., Gutshall, L., Maley, D., Sarisky, R.T., and Kao, C.C. (2001). Terminal nucleotidyl transferase activity of recombinant Flaviviridae RNA-dependent RNA polymerases: implication for viral RNA synthesis. J. Virol. 75, 8615–8623. Re´gnier, P., and Hajnsdorf, E. (2009). Poly(A)-assisted RNA decay and modulators of RNA stability. Prog. Mol. Biol. Transl. Sci. 85, 137–185. Rensing, U., and August, J.T. (1969). The 30 -terminus and the replication of phage RNA. Nature 224, 853–856. Sachs, A.B., Sarnow, P., and Hentze, M.W. (1997). Starting at the beginning, middle, and end: translation initiation in eukaryotes. Cell 89, 831–838. Schaffner, W., Ru¨egg, K.J., and Weissmann, C. (1977). Nanovariant RNAs: nucleotide sequence and interaction with bacteriophage Qbeta replicase. J. Mol. Biol. 117, 877–907. Shi, P.Y., Maizels, N., and Weiner, A.M. (1998). CCA addition by tRNA nucleotidyltransferase: polymerization without translocation? EMBO J. 17, 3197– 3206. Silverman, P.M. (1973). Replication of RNA viruses: specific binding of the Q RNA polymerase to Q RNA. Arch. Biochem. Biophys. 157, 222–233.

Sprinzl, M., and Cramer, F. (1979). The -C-C-A end of tRNA and its role in protein biosynthesis. Prog. Nucleic Acid Res. Mol. Biol. 22, 1–69. Takeshita, D., and Tomita, K. (2010). Assembly of Qb viral RNA polymerase with host translational elongation factors EF-Tu and -Ts. Proc. Natl. Acad. Sci. USA 107, 15733–15738. Takeshita, D., and Tomita, K. (2012). Molecular basis for RNA polymerization by Qb replicase. Nat. Struct. Mol. Biol. 19, 229–237. Teramoto, T., Kohno, Y., Mattoo, P., Markoff, L., Falgout, B., and Padmanabhan, R. (2008). Genome 30 -end repair in dengue virus type 2. RNA 14, 2645–2656. Toh, Y., Takeshita, D., Numata, T., Fukai, S., Nureki, O., and Tomita, K. (2009). Mechanism for the definition of elongation and termination by the class II CCAadding enzyme. EMBO J. 28, 3353–3365. Toh, Y., Takeshita, D., Nagaike, T., Numata, T., and Tomita, K. (2011). Mechanism for the alteration of the substrate specificities of templateindependent RNA polymerases. Structure 19, 232–243. Tomita, K., Fukai, S., Ishitani, R., Ueda, T., Takeuchi, N., Vassylyev, D.G., and Nureki, O. (2004). Structural basis for template-independent RNA polymerization. Nature 430, 700–704. Tomita, K., Ishitani, R., Fukai, S., and Nureki, O. (2006). Complete crystallographic analysis of the dynamics of CCA sequence addition. Nature 443, 956–960. Vagin, A., and Teplyakov, A. (2000). An approach to multi-copy search in molecular replacement. Acta Crystallogr. D Biol. Crystallogr. 56, 1622–1624. Wahba, A.J., Miller, M.J., Niveleau, A., Landers, T.A., Carmichael, G.G., Weber, K., Hawley, D.A., and Slobin, L.I. (1974). Subunit I of G beta replicase and 30 S ribosomal protein S1 of Escherichia coli. Evidence for the identity of the two proteins. J. Biol. Chem. 249, 3314–3316. Wahle, E., and Ru¨egsegger, U. (1999). 30 -End processing of pre-mRNA in eukaryotes. FEMS Microbiol. Rev. 23, 277–295. Weber, H., and Weissmann, C. (1970). The 30 -termini of bacteriophage Q-beta plus and minus strands. J. Mol. Biol. 51, 215–224. Weiner, A.M., and Maizels, N. (1987). tRNA-like structures tag the 30 ends of genomic RNA molecules for replication: implications for the origin of protein synthesis. Proc. Natl. Acad. Sci. USA 84, 7383–7387. Weissmann, C., Billeter, M.A., Goodman, H.M., Hindley, J., and Weber, H. (1973). Structure and function of phage RNA. Annu. Rev. Biochem. 42, 303–328. Xiong, Y., and Steitz, T.A. (2004). Mechanism of transfer RNA maturation by CCA-adding enzyme without using an oligonucleotide template. Nature 430, 640–645. Zamora, H., Luce, R., and Biebricher, C.K. (1995). Design of artificial shortchained RNA species that are replicated by Q beta replicase. Biochemistry 34, 1261–1266.

Structure 20, 1661–1669, October 10, 2012 ª2012 Elsevier Ltd All rights reserved 1669