Functional implication of archaeal homologues of human RNase P protein pair Pop5 and Rpp30

Abstract

PhoPop5 and PhoRpp30 in the hyperthermophilic archaeon Pyrococcus horikoshii, homologues of human ribonuclease P (RNase P) proteins hPop5 and Rpp30, respectively, fold into a heterotetramer [PhoRpp30–(PhoPop5)₂–PhoRpp30], which plays a crucial role in the activation of RNase P RNA (PhopRNA). Here, we examined the functional implication of PhoPop5 and PhoRpp30 in the tetramer. Surface plasmon resonance (SPR) analysis revealed that the tetramer strongly interacts with an oligonucleotide including the nucleotide sequence of a stem-loop SL3 in PhopRNA. In contrast, PhoPop5 had markedly reduced affinity to SL3, whereas PhoRpp30 had little affinity to SL3. SPR studies of PhoPop5 mutants further revealed that the C-terminal helix (α4) in PhoPop5 functions as a molecular recognition element for SL3. Moreover, gel filtration indicated that PhoRpp30 exists as a monomer, whereas PhoPop5 is an oligomer in solution, suggesting that PhoRpp30 assists PhoPop5 in attaining a functionally active conformation by shielding hydrophobic surfaces of PhoPop5. These results, together with available data, allow us to generate a structural and mechanistic model for the PhopRNA activation by PhoPop5 and PhoRpp30, in which the two C-terminal helices (α4) of PhoPop5 in the tetramer whose formation is assisted by PhoRpp30 act as binding elements and bridge SL3 and SL16 in PhopRNA.

Ribonuclease P (RNase P) is a ubiquitous trans-acting ribozyme that catalyses the processing of 5′ leader sequences from tRNA precursors (pre-tRNA) and other noncoding RNAs in all living cells (1, 2). In contrast to eubacterial RNase P RNAs, the RNA components in archaea and eukaryotes alone have little catalytic activity in vitro and function in cooperation with protein subunits in substrate recognition and catalysis (3). Hence, archaeal and eukaryotic RNase Ps may serve as a model ribonucleoprotein (RNP) for studying how a functional RNA can be activated by protein cofactors and how the RNP enzymes catalyse biological processes.

We earlier found that RNase P RNA (PhopRNA) and five proteins in the hyperthermophilic archaeon Pyrococcus horikoshii OT3 reconstituted RNase P activity that exhibits enzymatic properties like those of the authentic enzyme (4, 5). The P. horikoshii RNase P proteins were designated PhoPop5, PhoRpp21, PhoRpp29, PhoRpp30 and PhoRpp38, according to their sequence homology with the human RNase P proteins hPop5, Rpp21, Rpp29, Rpp30 and Rpp38, respectively (6). Biochemical and structural studies revealed that PhoPop5 and PhoRpp21 form a complex with PhoRpp30 and PhoRpp29, and the resulting complexes, PhoPop5–PhoRpp30 and PhoRpp21–PhoRpp29, are involved in activation of the C- and S-domains, respectively (7–9). Functional reconstitutions of the RNA component and individual proteins, and the cooperative of archaeal homologues of human RNase P protein pairs of Pop5 with Rpp30 and Rpp21 with Rpp29 were reported on other archaeal RNase Ps (10–13).

The X-ray structure shows that PhoPop5 and PhoRpp30 fold into a heterotetramer [PhoRpp30–(PhoPop5)₂–PhoRpp30], in which a homodimer of PhoPop5 sits between two PhoRpp30 monomers (Fig. 1). PhoPop5 dimerizes through hydrogen bonding interaction from the loop between α1 and α2 helices. The reconstituted particle containing the PhoPop5 mutant termed ΔL43-48, in which the α1–α2 loop in PhoPop5 was deleted, had significantly reduced pre-tRNA cleavage activity (7, 14). Furthermore, reconstitution experiments indicated that deletion of the C-terminal helices α4 and α5 (3₁₀-helix) significantly influenced the pre-tRNA cleavage activity, whereas that of α5 had little effect on this activity (14). These results indicate that the heterotetrameric structure is essential for activation of PhopRNA, and that C-terminal helix α4 in PhoPop5 plays a crucial role in the activation of PhopRNA. Recently, we found that PhopRNA mutants ΔP3 and ΔP16, in which stem-loops SL3 and SL16 containing P3 and P16 in PhopRNA were deleted, respectively, had little ability to bind PhoPop5 and PhoRpp30, suggesting that the PhoPop5–PhoRpp30 complex specifically recognizes SL3 and SL16 (15). Abundant evidence demonstrates that the tetrameric formation of the archaeal RNase P protein pair Pop5 and Rpp30 is indispensable for the catalytic activity of archaeal RNase Ps, but the mechanism behind this remains unclear.

Fig. 1.

The crystal structure of the PhoPop5–PhoRpp30 complex. The crystal structure of the PhoPop5–PhoRpp30 complex (Protein Data Bank accession code 2CZV) was drawn using Pymol (http://www.pymol.org). N and C indicate the N- and C-termini of PhoPop5, respectively. The C-terminal helices α4 and α5 are highlighted in red, and the α1–α2 loop in PhoPop5 is green. (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

To address this issue, surface plasmon resonance (SPR) analysis using Biacore X-100 was employed to analyse interactions of PhoPop5, its mutants, and PhoRpp30 with SL3 in PhopRNA. Furthermore, the functional implication of PhoRpp30 in the tetramer was investigated by gel filtration and kinetic analysis of the reconstituted particles containing its mutants. The results indicated that the C-terminal helix (α4) in PhoPop5 functions as a molecular recognition element for SL3, whereas PhoRpp30 primarily serves as a molecular chaperone for PhoPop5 to form a functionally active structure in the tetramer. On the basis of these results, we present a three-dimensional (3-D) model, in which the two C-terminal helices (α4) of PhoPop5 in the tetramer bridge SL3 and SL16, which suggests a molecular basis for the PhopRNA activation by PhoPop5 and PhoRpp30.

Materials and Methods

Materials

Ex Taq DNA polymerase and the DNA ligation kit were purchased from Takara Bio (Shiga, Japan). SA sensor chips and Biacore consumables were obtained from GE Healthcare. The 5′-biotinylated oligonucleotides and template oligonucleotides for RNA synthesis were purchased from Eurofins MWG Operon (Ebersberg, Germany). All other chemicals were of analytical grade for biochemical use.

Preparation of proteins and RNAs

Five RNase P proteins (PhoPop5, PhoRpp21, PhoRpp29, PhoRpp30 and PhoRpp38), PhopRNA and pre-tRNA^Tyr in P. horikoshii were prepared as described previously (4, 5). The deletion mutant ΔL43-48, in which the residues 43–48 including Glu44 and Glu48 in PhoPop5 were deleted, and the C-terminal truncated mutant (Δ14 C) of PhoPop5 were prepared as described previously (7, 14). Mutants of PhoRpp30, in which Arg90, Arg107, Lys123, Arg176, Asp180 and Lys196 in PhoRpp30 were individually replaced by Ala, were prepared as described previously (16).

Biosensor experiments

All SPR experiments were performed with a Biacore X-100® instrument (GE-Healthcare Biacore, Uppsala, Sweden). The running buffer composition was 20 mM phosphate buffer (pH 7.2), 500 mM NH₄Cl and 0.05% Tween20. 5′-labeled biotin-oligonucleotides (Fig. 2A) were immobilized on SA sensor chips. These SA sensor chips had been preconditioned by triplicate injections of 50 mM NaOH and 2 M NaCl, each with a contact time of 60 s. Before immobilization, the single-strand oligonucleotides, diluted at a low concentration of running buffer, were heated at 90°C for 1 min and put on ice before injection onto the SA sensor chip at 10 µl/min at 25°C to generate 100–150 RU on the surface. RNA chain Ta-SL3 with stem-loop SL3 was prepared by in vitro transcription and preformed at 500 nM in running buffer, with heating at 95°C for 15 min and then slowly allowing cooling to 4°C. The product was used for injection onto the SA sensor chip to reach a density of 200 RU. Proteins were serially diluted in running buffer to the final concentrations indicated in Figs 2B–D and injected at a flow rate of 25 µl/min for 35–60 s. The responses were also monitored during a 3- to 6-min dissociation phase. After this dissociation step, samples were kept in 50 mM NaOH in water, each for 60 s, to remove any residual protein and Ta-SL3. Each concentration series was run in duplicate (and sometimes in triplicate or quadruplicate). All binding data were collected at 25°C.

Fig. 2.

Sensorgrams of real-time analyses by Biacore. (A) A target oligonucleotide termed Ta-SL3 containing the SL3 sequence (red) was immobilized via biotinylated oligonucleotide complementary to the nucleotide sequence at the 3′-end of Ta-SL3 on the SA sensor chip. (B) The interaction of the PhoPop5–PhoRpp30 complex with Ta-SL3 was analysed, as described in Materials and Methods. The PhoPop5–PhoRpp30 complex was injected sequentially at increasing concentrations 4.8, 24, 120, 600 nM and 3 µM. k_on and k_off for the interaction of the PhoPop5–PhoRpp30 complex with Ta-SL3 were obtained by fitting sensorgrams to a Langmuir binding model shown by broken lines in red. (C, D) Sensorgrams of the interaction of PhoPop5 and PhoRpp30 with Ta-SL3, respectively. The proteins were injected in the same manner as that described above. K_D values for the binding of PhoPop5 and PhoRpp30 to Ta-SL3 were calculated by non-linear regression analysis of the steady-state values using a single site model, respectively, as presented in Supplementary Figs S1A and B. The resulting parameter values are given in Table I. (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

Data processing

Data processing and analysis were performed using Biacore X-100® software package in BiaEval 4.1. The responses were double-referenced by subtracting from the signal on the 5′-labeled biotin-oligonucleotide surface, both the signal recorded on the reference surface and the signal of a blank injection with the running buffer (17). Binding experiments were performed to determine the association and dissociation rate constants, k_on and k_off, respectively, for the complex formation. The dissociation equilibrium constant, K_D, was calculated as k_off/k_on. The sensorgrams were fitted assuming simple reversible bimolecular reactions as described by Palau and Primo (18). When sensorgrams could not be fitted, the steady-state values were plotted against the peptide concentration and the K_D values were determined by non-linear regression analysis of the steady-state values using a single site model: [Comp] = R_max-exp × [Pro]/(K_D + [Pro]), where R_max-exp is the maximal binding capacity, [Comp] the concentration of complex at a steady state, and [Pro] the concentration of the protein.

Assay for pre-tRNA cleavage activity

The RNase P activity for the reconstituted particles was analysed in the reconstitution buffer, 50 mM Tris–HCl (pH 7.6) containing 50 mM MgCl₂, 600 mM NH₄OAc and 60 mM NH₄Cl, principally as described previously (15). To examine the kinetic constants of the RNase P activity of the reconstituted particles, the pre-tRNA cleavage activity was analysed in the presence of PhoRpp30 or its mutants (5 pmol), the other four proteins (each 5 pmol), PhopRNA (5 pmol) and pre-tRNA^Tyr (1.88–19.4 µg) for 15 min. As for the analysis of the stoichiometry, various amounts of PhoPop5 and PhoRpp30 or PhoRpp21 and PhoRpp29 were incubated with the reconstituted particles containing PhopRNA (10 pmol) and the other three proteins (each 10 pmol) for 15 min at 75°C. The reactions were stopped by adding phenol, and the reaction products were separated on 10% polyacrylamide denaturing gels in TBE buffer (900 mM Tris-borate containing 10 mM EDTA) at 150 V for 1 h. After electrophoresis, the reaction products were visualized by staining in a 0.1% toluidine blue solution. The resulting image was used to obtain values for the pre-tRNA^Tyr processing activity with various incubation times. The cleavage efficiency was calculated as follows: the quantity of (matured tRNA^Tyr + leader fragment)/the quantity of (pre-tRNA^Tyr + matured tRNA^Tyr + leader fragment), and the percentage was plotted against the incubation times.

Modelling of PhoPop5–PhoRpp30 bound PhopRNA

Comparative analysis of the RNase P RNA sequences and existing crystallographic structural information of the bacterial RNase P RNAs were combined to generate a phylogenetically supported 3-D model of PhopRNA (19). The crystal structure of PhoPop5 in complex with PhoRpp30 was solved previously (Protein Data Bank accession code 2CZV) (7). The two PhoPop5 α4 helices in the tetramer were docked onto the PhopRNA model so as to interact with double-stranded structures in SL3 and SL16 using the program Coot (20). Because the model structure has not been subjected to energy minimization, the coordinates of the structure have not been deposited in Protein Data Bank. The coordinates will, however, be provided on request from the corresponding author.

Results

Binding affinities of PhoPop5 and PhoRpp30 to SL3 measured by SPR analysis

To examine the interaction of PhoPop5 and PhoRpp30 with PhopRNA quantitatively, SPR analysis with Biacore system was chosen because one of the partners is fixed so that events subsequent to the binding step may be prevented in this analysis, as described by Baltzinger et al. (21). We found recently that PhoPop5 and PhoRpp30 specifically bind to SL3 and SL16 in the C-domain of PhopRNA (15). It is known that residues C252-G258 connecting P16 interact with a single-strand region (residues C55-G61) to form P6 (19). It is thus likely that PhoPop5–PhoRpp30 recognizes a specific structure of SL16 formed in P6. Hence, SL3 was chosen as a target RNA for SPR analysis. To this end, a target oligonucleotide termed Ta-SL3 containing the SL3 sequence prepared by in vitro transcription was immobilized via biotinylated oligonucleotide complementary to the nucleotide sequence at the 3′-end of Ta-SL3 on the SA sensor chip, as shown in Fig. 2A. In SPR analysis, we employed the single-cycle kinetics introduced by Karlsson et al. (22), in which sequential injections of proteins (analyte) at increasing concentrations without regeneration steps between each sample injection are possible in a single binding cycle. Figure 2B–D shows typical sensorgrams of the association and dissociation of the proteins with Ta-SL3, where the proteins were injected sequentially at increasing concentrations (4.8, 24, 120, 600 nM and 3 µM). The sensorgram clearly shows that the PhoPop5–PhoRpp30 complex associates with Ta-SL3 quickly and dissociates slowly, indicating a specific interaction with Ta-SL3 (Fig. 2B). By fitting sensorgrams to a Langmuir binding model, k_on and k_off were unambiguously obtained, and K_D was calculated to be 2.27 nM, as presented in Table I. In contrast, PhoPop5 or PhoRpp30 was found to dissociate from Ta-SL3 more rapidly than the complex as shown in Figs 2C and D, respectively, and hence, k_on and k_off could not be determined accurately. In particular, no or little binding was detected when PhoRpp30 was injected over Ta-SL3 immobilized on the SA chip (Fig. 2D). K_D values for the binding of PhoPop5 and PhoRpp30 to Ta-SL3 were therefore calculated to be 2.41 and 55.79 µM, respectively, by non-linear regression analysis of the steady-state values using a single-site model (Supplementary Figs S1A and B), as described in Materials and Methods. This result indicated that PhoPop5 is primarily responsible for binding to SL3, and that the complex formation with PhoRpp30 remarkably enhanced binding affinity by 830- and 19,200-fold compared with those for PhoPop5 and PhoRpp30, respectively.

Table I.

Kinetic constants for the interaction of the proteins with Ta-SL3

Proteins	kon (1/Ms)	koff (1/s)	KD (nM)	Rmax (RU)
PhoPop5-PhoRpp30	2.59 × 107	5.88 × 10−2	2.27	793.16
PhoPop5	ND	ND	2.41 × 103	2,145.74
PhoRpp30	ND	ND	55.79 × 103	55.74
ΔL43-48-PhoRpp30	3.68 × 106	2.40 × 10−2	6.52	734.52
Δ14 C-PhoRpp30	ND	ND	1.77 × 103	1,112

ND, not determined

The C-terminal helix in PhoPop5 is a crucial element for binding to SL3

We found previously that deletion of the loop (residues 43–48) between α1 and α2 in PhoPop5 abolished its ability to homodimerize itself and resulted in heterodimerization with PhoRpp30, and that the reconstituted particle containing the mutant, referred to as ΔL43-48, had little pre-tRNA cleavage activity (7, 14). In addition, the reconstituted particle containing the C-terminal-truncated PhoPop5 mutant, Δ14 C, exhibited little pre-tRNA cleavage activity (14). Next, to gain more insight into the functional implication of PhoPop5, the binding affinity of ΔL43-48 and Δ14 C to Ta-SL3 was evaluated in the presence of PhoRpp30 (Supplementary Fig. S2A and B). Their binding parameters measured are shown in Table I. ΔL43-48 and PhoRpp30 were able to interact with Ta-SL3 at a level comparable to the wild-type PhoPop5 and PhoRpp30 (Supplementary Fig. S2A); the K_D value was 6.52 nM. This result indicated that the heterodimer composed of ΔL43-48 and PhoRpp30 is sufficient for specific binding to Ta-SL3. In contrast, Δ14 C and PhoRpp30 were found to dissociate from Ta-SL3 more rapidly than the PhoPop5–PhoRpp30 complex as shown in Supplementary Fig. S2B. Hence, the K_D value for binding of Δ14 C and PhoRpp30 to Ta-SL3 was calculated by non-linear regression analysis of the steady-state values using a single-site model (Supplementary Fig. S2C), as described above. The K_D value (1.77 µM) for the binding of Δ14 C and PhoRpp30 to Ta-SL3 was comparable to that (2.41 µM) for PhoPop5 to Ta-SL3 (Table I), suggesting that Δ14 C retains the ability to bind Ta-SL3 comparable to PhoPop5, but has no ability to complex with PhoRpp30. Although we found previously that the 14 C-terminal residues in PhoPop5 are not implicated in the formation of a complex with PhoRpp30 (Hazeyama et al., unpublished results), we measured the binding potency of Δ14 C alone to Ta-SL3 to exclude this assumption. The sensorgram clearly shows that Δ14 C had no ability to bind to Ta-SL3 (Supplementary Fig. S2D). That is, Δ14 C and PhoRpp30 were found to decrease the binding affinity to Ta-SL3 by 780-fold compared with the wild-type complex, indicating that the C-terminal helix (α4) in PhoPop5 acts as an essential element for PhopRNA binding.

PhoRpp30 serves as a molecular chaperone for PhoPop5

We found previously that PhopRNA mutant ΔP3, in which stem-loop SL3 containing P3 was deleted, had little ability to bind PhoPop5 and PhoRpp30 (15). In this study, SPR analysis indicated that PhoRpp30 has little ability to bind Ta-SL3 and its tetrameric formation with PhoPop5 remarkably enhanced the binding affinity to Ta-SL3. These results suggested that PhoRpp30 serves as a molecular chaperone in the dimeric formation of PhoPop5 in the tetramer. To support this assumption, the quaternary structure in solution was estimated by gel filtration on Superdex S-200. As shown in Fig. 3, PhoRpp30 and PhoPop5–PhoRpp30 were eluted at positions corresponding to their molecular masses, around 25,000 and 70,000, respectively, whereas PhoPop5 was eluted at an earlier position corresponding to around 50,000, even though it has a molecular mass of 14,043. This result indicated that PhoRpp30 exists as a monomer, whereas PhoPop5 is an oligomer in solution. The crystal structure shows the extensive hydrophobic and ionic interactions between PhoRpp30 and PhoPop5 (7). It is thus likely that PhoRpp30 assists PhoPop5 in attaining a functionally active conformation probably by shielding the hydrophobic surfaces of PhoPop5.

Fig. 3.

Gel filtration chromatography of proteins. The proteins PhoPop5 (▪ green), PhoRpp30 (• blue), and the PhoPop5–PhoRpp30 complex (○ red) were loaded onto a HiLoad 16/60 Superdex 200 column (1.6 cm × 60 cm). The arrows indicate the estimated elution positions for proteins with the molecular masses of 45,000 and 25,000. The molecular mass marker proteins used were ovalbumin (45,000) and α-chymotrypsinogen (24,800). (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

The present result suggested that PhoRpp30 primarily acts as a molecular chaperone for formation of the active dimeric structure of PhoPop5. Previous Ala-scanning site-directed mutagenesis of amino acids conserved in the Rpp30 family indicated that mutations of Arg90, Arg107, Lys123, Arg176, Asp180 and Lys196 in PhoRpp30 moderately reduced pre-tRNA cleavage activity (32–50%) by the reconstituted particles containing individual mutants (16). Mapping these amino acids on the heterotetrameric structure, Arg90, Lys123 and Lys196 are located far from the interface of PhoPop5, although Arg107, Arg176 and Asp180 appear to be involved in the interaction with PhoPop5 (Supplementary Fig. S3A). This finding suggested that PhoRpp30 itself appeared to be involved in the activation of PhopRNA. To examine the functional implication of PhoRpp30 further, we prepared these mutants and reinvestigated them with respect to the involvement in pre-tRNA cleavage activity by analysing kinetic parameters (Supplementary Fig. S3B and C).

Kinetic constants for pre-tRNA cleavage activity of the reconstituted particles containing individual mutants are summarized in Table II. The k_cat values (0.79–1.26 min⁻¹) of the mutant particles except for that containing R107A were approximately the same as that (1.11 min⁻¹) of the wild type. Mutation of Arg107 to Ala resulted in a 2.4-fold decrease in k_cat. In contrast, the K_M values (8.16–14.81 µM) of the mutant particles except for that of R107A were slightly increased by about 2- to 3-fold compared with that (5.03 µM) of the wild type. The resulting specificity constants (k_cat/K_M) of the reconstituted particles containing mutants were reduced 1.57- to 2.75-fold in comparison to that of the particle containing the wild-type PhoRpp30. This result indicated that PhoRpp30 itself has little contribution to the activation of PhopRNA, although one cannot exclude the possibility that PhoRpp30 is moderately involved in substrate binding.

Table II.

Kinetic constants of the reconstituted particles containing PhoRpp30 or its mutants

Proteins	Vmax (µM/min)	KM (µM)	kcat (min−1)	kcat/KM (min−1/µM)
PhoRpp30	0.28 ± 0.05	5.03 ± 2.13	1.11	0.22
R90A	0.20 ± 0.03	10.45 ± 2.29	0.79	0.08
R107A	0.12 ± 0.03	4.99 ± 2.69	0.46	0.09
K123A	0.29 ± 0.05	8.16 ± 2.66	1.14	0.14
R176A	0.26 ± 0.07	8.37 ± 4.31	1.03	0.12
D180A	0.31 ± 0.08	14.81 ± 5.63	1.26	0.08
K196A	0.26 ± 0.06	11.01 ± 4.61	1.03	0

A 3-D model of PhoPop5 and PhoRpp30 in complex with PhopRNA

The present study indicated that the C-terminal helix (α4) in PhoPop5 acts as an essential element for PhopRNA binding. Furthermore, reconstitution experiments indicated that the tetrameric structure is essential for the activation of PhopRNA, and that the C-terminal helix α4 in PhoPop5 plays a crucial role in the activation of PhopRNA (7, 14). We found further that the PhoPop5–PhoRpp30 complex specifically recognizes SL3 and SL16 (15). In the PhopRNA model structure (19), SL3 is approximately 65 Å away from SL16 (Fig. 4A). On the other hand, the two PhoPop5 C-terminal helices α4 in the tetramer, which are crucial RNA-binding elements found in this study, are similarly 65 Å away from each other (Fig. 4B). This finding, together with previous data allowed us to generate a 3-D model, in which the two PhoPop5 C-terminal helices (α4) in the tetramer bind SL3 and SL16 in PhopRNA, as presented in Fig. 4C. In this 3-D model, the tetramer composed of two PhoPop5 and two PhoRpp30 binds SL3 and SL16 in PhopRNA. On the one hand, the crystal structure of the PhoRpp21–PhoRpp29 complex reveals that PhoRpp21 and PhoRpp29 fold into a dimeric structure (8), suggesting that the stoichiometry of the PhoPop5–PhoRpp30 complex with the PhoRpp21–PhoRpp29 complex is 2:1. Recently, Ma et al. (23) described that the stochiometry of the proteins is 1:1 in the hyperthermophilic archaea P. furiosus RNase P, even though PfuPop5 and PfuRpp30, homologues of PhoPop5 and PhoRpp30, respectively, fold into the heterotetrameric structure in solution. To address this issue, the stoichiometry of the P. horikoshii RNase P proteins was examined by analysing the pre-tRNA cleavage activity of the reconstituted particles containing PhopRNA and various amounts of the proteins. When an increasing amount of PhoRpp21 and PhoRpp29 was added to the reconstituted mixture containing an equimolar amount (10 pmol) of PhopRNA and the other three proteins (PhoPop5, PhoRpp30 and PhoRpp38), the reconstituted particle containing 5 pmol PhoRpp21 and PhoRpp29 had the full pre-tRNA cleavage activity, as shown in Fig. 5. In contrast, 20 pmol PhoPop5 and PhoRpp30 were required to reach full pre-tRNA cleavage activity of the reconstituted mixture containing an equimolar amount (10 pmol) of PhopRNA and the other three proteins (PhoRpp21, PhoRpp29 and PhoRpp38) (Fig. 5). This result indicated a 2:1 stoichiometry of the two protein pairs, i.e. the tetramer composed of PhoPop5 and PhoRpp30 interacts with PhopRNA in RNase P, consistent with the model structure proposed in this study.

Fig. 4.

A 3-D model of the PhoPop5–PhoRpp30 complex bound PhopRNA. (A, B) The 3-D model of PhopRNA and the crystal structure of the PhoPop5–PhoRpp30 complex, respectively. In the PhopRNA model structure, SL3 in violet is approximately 65 Å away from SL16 in green, and the two PhoPop5 C-terminal helices α4 (red) in the tetramer are similarly 65 Å away from each other. (C) A 3-D model of the PhoPop5–PhoRpp30 complex bound PhopRNA. PhoPo5 and PhoRpp30 are in cyan and brown, respectively; and the two C-terminal helices α4 in PhoPop5 are highlighted in blue. Stem-loop structures SL3 and SL16 are highlighted in red. (D) Secondary structures of SL3 and SL16 containing P3 and P16 helices, respectively. Possible structural elements recognized by the PhoPop5–PhoRpp30 complex are boxed in red. Figures were drawn using Pymol (http://www.pymol.org). (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

Fig. 5.

Analysis of the stoichiometry of the Pyrococcus horikoshii RNase P proteins. Various amounts of PhoPop5 and PhoRpp30 or PhoRpp21 and PhoRpp29 were incubated with the reconstituted particles containing PhopRNA (10 pmol) and the other three proteins (each 10 pmol) for 15 min at 75°C. The reaction products were analysed as described previously (15). Open circles (○) indicate pre-tRNA cleavage activities of the reconstituted particles containing increasing amounts of PhoPop5 and PhoRpp30, whereas closed circles (•) indicate those of the particles containing increasing amounts of PhoRpp21 and PhoRpp29. The experiments were carried out in triplicate, and the mean values are presented.

Discussion

Reconstitution experiments suggested that PhoPop5 and PhoRpp30 synergistically activate PhopRNA (6). Structural and mutational data indicated that PhoPop5 and PhoRpp30 fold into a tetramer in solution, the quaternary structure playing a crucial role in the activation and stabilization of PhopRNA (7, 14). To understand the molecular basis of the cooperative function of PhoPop5 and PhoRpp30, we first characterized them with respect to binding affinity to PhopRNA. Previously, gel retardation assay showed that the interactions of PhoPop5 or PhoRpp30 with PhopRNA produced a smeary shift, and the protein appeared to be aggregated on PhopRNA at higher protein concentrations (4). Similar observations on some RNA-binding proteins have been reported (24). These phenomena have impeded quantitative evaluation of their interaction with RNA. Recently, Baltzinger et al. have described that SPR analysis immobilizing a target RNA on a sensor chip can avoid events subsequent to the binding step (21). Hence, we employed SPR analysis to investigate the interaction of PhoPop5 and PhoRpp30 with PhopRNA using Ta-SL3 as a target RNA (Fig. 2). In the initial step of this analysis, no specific binding of the proteins to Ta-SL3 was observed using the reconstitution buffer, 50 mM Tris–HCl (pH 7.6) containing 50 mM MgCl_2, 600 mM NH₄OAc and 60 mM NH₄Cl, as a running buffer. As it is known that the interaction between an oligonucleotide and a positively charged molecule in SPR analysis is highly dependent on salt concentrations (21), running buffer without MgCl₂ and NH₄OAc was used, as described in Materials and Methods. As a result, a specific interaction of the PhoPop5–PhoRpp30 complex and PhoPop5 with Ta-SL3 was successfully measured and their K_D values were determined (Table I and Fig. 2). The SPR analysis was, however, carried out using Ta-SL3 under binding conditions distinct from those used for pre-tRNA cleavage assay. Therefore, we cannot correlate their K_D values with the binding affinity of PhopRNA by the proteins. Nevertheless, this result indicates that PhoPop5 is primarily responsible for binding to Ta-SL3 in the tetramer, and that its complex formation with PhoRpp30 remarkably enhances binding affinity to Ta-SL3 by 830-fold compared with that of PhoPop5 alone.

It should be noted that the R_max value (2,145.74 RU) observed in the PhoPop5 binding to Ta-SL3 is significantly higher than that (793.16 RU) of the PhoPop5–PhoRpp30 complex (Table I). This observation suggests that PhoPop5 exists in an oligomeric structure in solution, whose molecular mass may be higher than that of the PhoPop5–PhoRpp30 complex. Gel filtration analysis, however, showed that PhoPop5 was eluted at a position corresponding to around 50,000, whereas PhoPop5–PhoRpp30 was at a position corresponding to its molecular mass (70,000) (Fig. 3). Although we have no appropriate explanation for the higher R_max value for PhoPop5 at the moment, it could be assumed that the interaction of PhoPop5 with RNA results in an aggregation of PhoPop5 on the SA sensor chip.

Based on available data, we propose a mechanistic model for the activation of PhopRNA by PhoPop5 and PhoRpp30 synergistically, as presented in Fig. 6. PhoRpp30 exists as a monomer, whereas PhoPop5 is an oligomer in solution, and their interaction results in attaining a functional dimeric conformation of PhoPop5 in the tetramer. The two PhoPop5 C-terminal helices α4 in the tetramer bridge SL3 and SL16, and thereby stabilize an appropriate conformation of PhopRNA. The proposed model suggests that PhoRpp30 functions to shield hydrophobic surfaces of PhoPop5 so as to prevent its aggregation and assists PhoPop5 in attaining a functionally active conformation in the tetramer. Structural analysis revealed the extensive hydrophobic and ionic interactions between PhoPop5 and PhoRpp30, including that between the N-terminal segment of PhoPop5 and residues Asp109–Asp113 in PhoRpp30 (7). Mutation of Arg107 in PhoRpp30 slightly decreased the k_cat value: mutation of Arg107 to Ala resulted in a 2.4-fold decrease in k_cat (Table II and Fig. S3C). Examination of the PhoPop5–PhoRpp30 complex’s structure reveals that the Arg107 side chain forms a salt bridge with the Asp109 side chain at the loop between β5 and α5 in PhoRpp30. The Asp109 side chain, in turn, water-mediated hydrogen bonds with the Leu5 main chain in PhoPop5 and thereby stabilizes the N-terminal segment of PhoPop5 (Supplementary Fig. S4). The N-terminal β-strand (β1) and the C-terminal strand (β4) form an antiparallel β-sheet, which acts as a platform for the C-terminal helix α4 in PhoPop5. It is thus assumed that mutation of Arg107 caused slight perturbation of an appropriate orientation of the C-terminal helix (α4) of PhoPop5 that may result in reduction of the k_cat value of the reconstituted particle containing R107A.

Fig. 6.

The proposed molecular mechanism by which PhoPop5 and PhoRpp30 activate PhopRNA. PhoRpp30 (green) exists as a monomer, whereas PhoPop5 (blue) is an oligomer in solution. Hydrophobic interactions of PhoRpp30 with PhoPop5 avoid the self-oligomerization of PhoPop5, which results in attaining a functional dimeric conformation of PhoPop5 in the heterotetramer. Then, the two PhoPop5 C-terminal helices α4 shown in red circles in the tetramer bind the stem-loop structures (red) containing P3 and P16 helices, and thereby stabilize an appropriate conformation of PhopRNA. (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

In the presented 3-D model (Fig. 4C), two PhoPop5 C-terminal helices α4 in the tetramer bridge SL3 and SL16, assuming that SL3 and SL16 have a common structural element recognized by the C-terminal helix (α4) in PhoPop5. This assumption is derived from three different observations. First, previous mutagenesis has identified nucleotides encompassing C26 to G34 as a minimum binding site for the PhoPop5–PhoRpp30 complex in SL3 (15). Second, deletion of single-stranded segments, G244-U251 and U259-A261, connecting P16 and P6 had no influence on the ability to bind the PhoPop5–PhoRpp30 complex, suggesting that the complex recognizes a double-stranded structure in SL16 (F. Kumagai and M. Kimura, unpublished results). Third, we found previously that PhoPop5 and PhoRpp30 could activate a chimeric RNase P RNA, composed of the Escherichia coli RNase P RNA (M1 RNA) C-domain and the PhopRNA S-domain (9), suggesting that a structural element in SL3 and SL16 recognized by the C-terminal helix (α4) in PhoPop5 is conserved in M1 RNA. Indeed, double-stranded structures in archaeal P3 and P16 helices are highly conserved in corresponding helices in M1 RNA. By considering these results, it is conceivable that the two PhoPop5 C-terminal helices α4 in the tetramer may recognize similar double-stranded structures formed by annealing G26-G28 and C34-U32 in P3 and G240-G243 and C265-G262 in P16 (Fig. 4D).

It is known that many RNA-binding proteins have modular structures and are composed of multiple repeats of RNA recognition motif (RRM) that are arranged in various ways to satisfy their diverse functional requirements (25, 26). For instance, Saccharomyces cerevisiae Prp24, an essential splicing factor, contains four RRM domains and cooperatively functions to anneal U6 and U4 RNAs during spliceosome assembly. Thus, RNA-binding proteins have evolved by combining RNA-binding domains in various structural arrangements that can recognize RNA with the affinity and selectivity that is required to find cognate RNAs in the cellular medium. Although PhoPop5 folds into RRM, it is distinct from the typical RRM in that it has an insertion of α-helix (α2) between α1 and β2, and in addition, it has additional helices (α4 and α5) at the C-terminus, which pack against one face of the β-sheet. PhoPop5 dimerizes through hydrogen bonding interaction from the loop between α1 and α2 helices with the aid of PhoRpp30, and the C-terminal helices (α4) thus adopted act as a binding element for the double-stranded structure. Hence, archaeal Pop5, unlike most RNA-binding proteins, may fulfil a functional requirement by acquiring additional structural elements that facilitate the bridging of SL3 and SL16 in PhopRNA.

Interestingly, the fourth C-terminal RRM domain in Prp24, like PhoPop5, has two additional flanking α-helices that form a large electropositive surface. It has been described that the two helices are involved in destabilization of the U6 internal stem loop, a stable helix that must be unwound during U4/U6 assembly (27). Recently, we found that PhoPop5 significantly enhanced both RNA annealing and strand displacement and thus, may assist PhopRNA in achieving a functionally active conformation by RNA chaperone activity (28). Deletion of the C-terminal helix α4 in PhoPop5 significantly reduced not only pre-tRNA cleavage activity of the reconstituted particle but also abolished its RNA annealing and strand displacement activities (14). Thus, the C-terminal helices in the tetramer, besides bridging SL3 and SL16, may be involved in correct double-stranded formation of P3 and P16 helices as an RNA chaperone.

In contrast to archaeal RNase P RNA, eubacterial RNase P RNA itself can hydrolyse pre-tRNA in vitro in the presence of a high concentration of Mg²⁺ (1). The bacterial A-type RNase P RNAs have helical stems P13, P14 and P18, which are absent in the archaeal RNase P RNAs (Fig. 7) (29, 30). It is known that tetraloop–helix interactions between P8 and P14 and P8 and P18 in the bacterial A-type RNase P RNA position the two domains (C- and S-domains) correctly to permit catalysis. A shortened RNA and an increase in the number of proteins in archaeal RNase Ps suggest that some structural roles of eubacterial RNase P RNA might be delegated to the proteins in archaeal RNase P. On the basis of the model of PhopRNA in complex with PhoPop5–PhoRpp30, it is likely that the tetraloop–helix interaction between P8 and P18 in the bacterial A-type RNase P RNA is replaced by the cross-linking of SL3 and SL16 by the PhoPop5–PhoRpp30 complex in archaeal RNase P RNAs (Fig. 7). The presented model may help guide further investigations not only into the mechanism of the PhoPop5–PhoRpp30 complex-mediated activation of PhopRNA but also into the molecular evolution of RNase Ps in the three phylogenetic domains of life.

Fig. 7.

Structural comparison of the bacterial RNase P RNA with PhopRNA in complex with PhoPop5 and PhoRpp30. The bacterial A-type RNase P RNA in Thermotoga maritima is made up of two layers; the large layer of the structure contains most of the universally conserved regions, whereas the second layer comprises helical stems P13, P14 and P18, which are absent in the archaeal RNase P RNAs (29, 30). It is assumed that the tetraloop–helix interaction between P8 (blue) and P18 (red) in the bacterial A-type RNase P RNA is replaced by the cross-linking of SL3 and SL16 by the PhoPop5–PhoRpp30 complex in PhopRNA. The C- and S-domains in the bacterial A-type RNase P RNA and PhopRNA are in grey and gold, respectively. (For interpretation of the references to colours in this figure legend, the reader is referred to the web version of this article.)

Supplementary Data

Supplementary Data are available at JB Online.

Glossary

Abbreviations

3-D

three-dimensional

; PhopRNA

ribonuclease P RNA from P. horikoshii

pre-tRNA

precursor tRNA

RNase P

ribonuclease P

RNP

ribonucleoprotein

RRM

RNA recognition motif

SL3

stem-loop containing P3 helix

SL16

stem-loop containing P16 helix

SPR

surface plasmon resonance

Funding

This work was supported in part by a grant-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (no. 22380062 to M.K.).

Conflict of Interest

None declared.