Research Article

Structural insight into precursor tRNA processing by yeast ribonuclease P

See allHide authors and affiliations

Science  09 Nov 2018:
Vol. 362, Issue 6415, eaat6678
DOI: 10.1126/science.aat6678

Structures of eukaryotic ribonuclease P

Ribonuclease P (RNase P) is a ribozyme that processes transfer RNA (tRNA) precursors and is found in all three kingdoms of life. Now, Lan et al. report the structures of yeast RNase P (see the Perspective by Scott and Nagai). The aporibozyme structure reveals how the protein components stabilize the RNA and explains how the structural roles of bacterial RNA elements have been delegated to the protein components in RNase P of higher organisms during evolution. The structure of yeast RNase P in complex with its natural substrate, a tRNA precursor, demonstrates the structural basis for substrate recognition and provides insights into its catalytic mechanism.

Science, this issue p. eaat6678; see also p. 644

Structured Abstract

INTRODUCTION

Ribonuclease P (RNase P), a universal ribozyme that has been found in organisms from all three domains of life, processes the 5′ end of transfer RNA (tRNA). RNase P is a ribonucleoprotein complex, composed of a single catalytic RNA component and a variable number of proteins. Unlike bacterial RNase P, which contains only one small protein cofactor, archaeal and eukaryotic nuclear RNase Ps have evolved considerably more complex protein subunits: five in archaea and 9 to 10 in eukarya. The pre-tRNA processing reaction can be described by a kinetic mechanism that includes four distinct events: (i) rapid and irreversible binding of RNase P (E) to pre-tRNA (S) to form the initial RNase P-pre-tRNA complex (ES); (ii) a conformational change isomerizing the ES complex to a catalytically competent conformer (ES*) in a magnesium ion (Mg2+)–dependent manner; (iii) the cleavage of the phosphodiester bond; and (iv) rapid dissociation of the 5′ leader and slow, rate-limiting release of the mature tRNA (see the figure, right).

RATIONALE

Despite extensive biochemical and genetic studies, however, the role of protein components and the reason for the increased complexity of the protein moieties in eukaryotic nuclear RNase P are still poorly understood. It is still enigmatic how the pre-tRNA substrate, especially the 5′-leader, is recognized by eukaryotic RNase P; how the catalytically important metal ions are coordinated in the active site; and what the chemical mechanism is of pre-tRNA 5′ cleavage. High-resolution structures of eukaryotic RNase Ps are required to answer these key questions.

RESULTS

Here, we report the 3.5-Å cryo–electron microscopy structures of Saccharomyces cerevisiae RNase P holoenzyme alone and in complex with pre-tRNAPhe. The yeast RNase P holoenzyme consists of one catalytic RNA Rpr1 and nine protein components. The Rpr1 RNA adopts an extended single-layered conformation that maintains a central helical core but lacks most of the long-range RNA-RNA interactions that are essential for structural stability in bacterial RNase P. The protein components form an interconnected hook-shaped architecture that tightly wraps around the RNA and stabilizes yeast RNase P into a “measuring device,” with two fixed anchors that recognize the L-shaped structure rather than specific sequences of pre-tRNA substrates (see the figure, left). This “measuring device” mediates the initial engagement with pre-tRNA to form the low-affinity ES complex. The recognition of the 5′-leader of pre-tRNA involves both the Rpr1 RNA and the protein subunit Pop5. Two catalytically important Mg2+ ions are coordinated in the catalytic center by highly conserved uridine U93 and the phosphate backbone of Rpr1, together with the scissile phosphate and the O3′ leaving group of pre-tRNA (see the figure, right). The configuration of this RNA-based catalytic center is universally conserved in all RNase Ps, from bacteria to eukarya. Pre-tRNA binding induces a dramatic conformational change in the catalytic center, corresponding to the isomerization step to the ES* state. Moreover, our simulation analysis visualized the mechanistic details of phosphodiester bond hydrolysis of pre-tRNA, which is a two-Mg2+-ion–mediated SN2 reaction (see the figure, right).

CONCLUSION

The structures presented here represent a major step forward for mechanistic understanding of the function of eukaryotic RNase P. Our data support that all RNase P ribozymes share an RNA-based, substrate-induced catalytic mechanism of pre-RNA processing. Whereas bacterial RNase P RNA is catalytically active by itself, eukaryotic RNase P is a protein-controlled ribozyme; its protein components not only directly participate in substrate recognition but also stabilize the catalytic RNA in a conformation optimal for pre-tRNA binding and cleavage reaction.

Catalytic mechanism of pre-tRNA processing catalyzed by yeast RNase P.

(Left) The overall structure of yeast RNase P holoenzyme in complex with pre-tRNAPhe. The protein hook and the RNAs [the Rpr1 RNA (gray) and the pre-tRNA (cyan)] are in surface and ribbon representations, respectively. (Right) Pre-tRNA is cleaved by yeast RNase P by means of a kinetic mechanism that includes four distinct events. First, pre-tRNA is recognized by RNase P through a double-anchor mechanism to form the initial ES complex, which induces a local conformational change in the catalytic center of RNase P. In particular, nucleotide U93 of the Rpr1 RNA undergoes a dramatic conformational change to mediate an inner-sphere coordination of the catalytically important Mg2+ ion, so that the ES complex is isomerized to the active ES* state. Next, the activated ES* complex catalyzes the phosphodiester bond cleavage of pre-tRNA through a two-metal-ion SN2 mechanism to release the 5′-leader of pre-tRNA. Last, the mature tRNA dissociates from the holoenzyme in a slow, rate-limiting step, and RNase P is ready for the next round of catalysis.

Abstract

Ribonuclease P (RNase P) is a universal ribozyme responsible for processing the 5′-leader of pre–transfer RNA (pre-tRNA). Here, we report the 3.5-angstrom cryo–electron microscopy structures of Saccharomyces cerevisiae RNase P alone and in complex with pre-tRNAPhe. The protein components form a hook-shaped architecture that wraps around the RNA and stabilizes RNase P into a “measuring device” with two fixed anchors that recognize the L-shaped pre-tRNA. A universally conserved uridine nucleobase and phosphate backbone in the catalytic center together with the scissile phosphate and the O3′ leaving group of pre-tRNA jointly coordinate two catalytic magnesium ions. Binding of pre-tRNA induces a conformational change in the catalytic center that is required for catalysis. Moreover, simulation analysis suggests a two-metal-ion SN2 reaction pathway of pre-tRNA cleavage. These results not only reveal the architecture of yeast RNase P but also provide a molecular basis of how the 5′-leader of pre-tRNA is processed by eukaryotic RNase P.

Ribonuclease P (RNase P), one of only two universal ribozymes that have been found in organisms from all three domains of life, is responsible for the maturation of the 5′ end of transfer RNA (tRNA) (1). RNase P is ribonucleoprotein complex, composed of a single catalytic RNA component and a variable number of proteins (24). The RNA component of RNase P from all organisms shows marked similarities at the primary and secondary structure level, which is suggestive of the presence of a universally conserved catalytic RNA core (2, 4). All RNase P RNAs can be divided into two independent folded domains, the catalytic domain (C domain) and the specificity domain (S domain), which play key roles in substrate cleavage and substrate binding, respectively (5, 6). Bacterial RNase P contains a single small protein (Rpp) that is essential for substrate recognition and cleavage under physiological conditions (79). By contrast, archaeal and eukaryotic nuclear RNase Ps have evolved considerably more complex subunit compositions with an increased number of protein components, five in archaea and 9 to 10 in eukarya (2). Unlike bacterial RNAs, the RNAs themselves in archaeal and eukaryotic RNase Ps are not generally catalytically active, and protein subunits are required to enhance the pre-tRNA substrate binding affinity and cleavage efficiency (1013).

Very little is known about the structure of eukaryotic RNase Ps. Only crystal structures of human Pop6-Pop7 subcomplex and Saccharomyces cerevisiae Pop6-Pop7 complexed with the P3 element of RNase MRP (mitochondrial RNA processing) RNA are available (14, 15). In addition, a low-resolution structure of S. cerevisiae nuclear RNase P has been solved by means of cryo–negative staining electron microscopy (EM) (16). A high-resolution structure of eukaryotic nuclear RNase P holoenzyme still has yet to be determined, hindering our understanding of the structural organization and mechanism of action of RNase P from higher organisms.

RNase P is a multiple turnover ribozyme that recognizes its substrates in trans (17). Previous studies suggest that the cleavage of pre-tRNA by RNase P can be described by a kinetic mechanism that includes at least four distinct events: (i) rapid and irreversible binding of RNase P (E) to pre-tRNA (S) to form the initial RNase P–pre-tRNA complex (ES); (ii) the ES complex then undergoes a conformational change and is isomerized to a catalytically competent conformer (ES*) in a magnesium ion(Mg2+)–dependent manner; (iii) the cleavage of the phosphodiester bond; and (iv) rapid dissociation of the 5′-leader and slow, rate-limiting release of the mature tRNA (1821). Although this kinetic model was proposed about a decade ago, how RNase P facilitates such a reaction remains largely enigmatic.

The first step of this kinetic scheme represents the formation of the initial low-affinity ES complex. RNase P recognizes the three-dimensional structural feature rather than specific sequences of pre-tRNAs (1, 22, 23). It has been proposed that RNase P contains a “measuring device” that recognizes a common structural feature of all pre-tRNAs; the coaxially stacked acceptor and T stems have a fixed length of 12 base pairs of nucleotides (1). The crystal structure of the bacteria Thermotoga maritima RNase P–tRNA complex with a soaked 5′-leader provides the structural basis of a simple, RNA-based “measuring device” and some insight into the binding of pre-tRNA 5′-leader (23). However, pre-tRNA recognition by the much more complex and indispensable protein components of eukaryotic RNase Ps is still poorly understood. Kinetic evidences suggested that a conformational change occurs after the formation of the initial ES complex and before 5′-leader cleavage, transforming the ES complex into an active ES* state (7, 9, 18, 24, 25). The ES* complex is stabilized by at least two Mg2+ ions during the transition from ES to ES*, and the functional groups and metal ions in the active site are believed to be repositioned to coordinate the substrate for catalysis (19, 26). However, there is still no structural evidence of the existence of this conformational change. The phosphodiester bond cleavage is an SN2-type transesterification reaction, which uses a two-Mg2+-ion mechanism that is also used by the group I and group II introns as well as the spliceosome (2730). Despite extensive studies, the positions of the catalytic Mg2+ ions and the composition of the catalytic center in RNase P are still enigmatic because of the lack of structural information of a pre-tRNA substrate–bound RNase P (24, 3139). Consequently, the chemical mechanism of this two-Mg2+-ion–mediated reaction still awaits to be revealed.

Here, we present the 3.5-Å cryo-EM structures of S. cerevisiae nuclear RNase P holoenzyme alone and in complex with a pre-tRNA substrate. The structures unveil the arrangement and function of all the subunits within yeast RNase P and provide an integrated model that depicts how the pre-tRNA substrate is recognized and how the hydrolysis of the 5′-leader of pre-tRNA is catalyzed by eukaryotic RNase P.

Overall structure of yeast RNase P

We used a two-step affinity purification scheme to obtain the endogenous S. cerevisiae RNase P complex. The highly purified RNase P was analyzed by means of SDS–polyacrylamide gel electrophoresis followed by mass spectrometry confirmation (fig. S1, A and B, and supplementary materials, materials and methods). Yeast RNase P exhibited a robust pre-tRNA cleavage activity in the presence of magnesium ions, indicating the recovery of a fully functional ribozyme (fig. S1C). Single-particle EM analysis of yeast RNase P yielded a well-defined EM density map at an overall resolution of 3.5 Å (figs. S1, D to J, and S2). We combined de novo model building and homologous modeling to generate an atomic structure for the RNase P complex (Fig. 1, A and B; fig. S3; and table S1). The final refined model of yeast RNase P contains all the previously identified components, one catalytic RNA—Rpr1—and nine proteins (Fig. 1B).

Fig. 1 EM Structure of the S. cerevisiae RNase P holoenzyme.

(A) The EM density map of S. cerevisiae RNase P at an average resolution of 3.5 Å. The individual protein and RNA subunits of RNase P are colored according to the scheme shown at the bottom right of the figure. (B) Overall structure of the S. cerevisiae RNase P complex. (C) Two orthogonal views of the overall structure of the protein hook.

The protein components of yeast RNase P form an intimately interconnected hook-shaped architecture, with Pop1, Pop6, and Pop7 being the head and Pop3, Pop4, Pop5, Pop8, Rpp1, and Rpr2 being the arm (Fig. 1C). Structurally, the protein hook can be considered as the assembly of Pop1, Pop6-Pop7 heterodimer, Pop5-Pop8-(Rpp1)2 heterotetramer, and Pop4-Rpr2-Pop3 heterotrimer (Fig. 1C). The Rpr1 RNA adopts an extended and slightly curved single-layered configuration, with the C and S domains packing against each other (Fig. 2, A and B). Three coaxially stacked helical stems, P2-P19 and P4-P1 of the C domain and P8-P9 of the S domain, form the core of the RNA, which is covered by Pop1 on one side and by the Pop5-Pop8-(Rpp1)2 heterotetramer on the other (Fig. 1B). In the head module of the hook, the Pop6-Pop7 heterodimer together with Pop1 encircles the P3 branch of Rpr1 (Fig. 1B). The arm of the hook packs along one side of Rpr1 (Fig. 1B). At the end of the hook, the Pop4-Rpr2-Pop3 heterotrimer functions as a bridge between the C and S domains of Rpr1 (Fig. 1B). Together, the protein hook tightly wraps around the Rpr1 RNA, burying a total of ~14,715 Å2 surface area between the RNA and the proteins (Fig. 1B).

Fig. 2 Structure of the catalytic RNA subunit Rpr1.

(A) Two orthogonal views of the overall structure of the Rpr1 RNA. The three open holes are highlighted by the red triangle, circle, and rectangle, respectively. (B) Secondary structure diagram of the yeast Rpr1 RNA. RNA elements are colored as in (A). The C and S domains are denoted. The canonical Watson-Crick and noncanonical base-pairing interactions are shown as solid lines and dots, respectively. Nucleotides in CRs that are universally conserved in bacterial, archaeal, and eukaryotic RNase P RNAs are highlighted with solid circles, whereas those that are invariant only within eukaryotic RNAs are highlighted with open circles. (C) Structure of the universally conserved pseudo-knot formed by CR-I, R-IV, and CR-V. (D) The tetraloop-tetraloop receptor interactions between stems P4 and P8 and between P1 and P9. The interacting regions are highlighted in red boxes. (E) A close-up cartoon representation of the CR-II/CR-III region in the S domain.

Structure of the Rpr1 RNA

In the coaxial helical core of Rpr1, three highly conserved regions—CR-I, CR-IV, and CR-V—fold together into a distinct pseudo-knot structure (Fig. 2, A and C). This pseudo-knot structure brings the P2-P19 stem into close vicinity to stem P4-P1 so that the two long stems pack parallel to each other (Fig. 2A). By contrast, stem P8-P9 in the S domain is not close to stem P4-P1 in primary sequence (Fig. 2B). Both the terminal loops of stems P8 and P9 adopt a tetraloop conformation and respectively interact with the minor grooves of stems P4 and P1 through tetraloop-tetraloop receptor interactions (Fig. 2, A and D). Beside CR-I, CR-IV, and CR-V, there are two additional universally conserved regions CR-II and CR-III, corresponding to two single-stranded junctions J11/12 and J12/11, between stems P10/11 and P12 in the S domain (Fig. 2B). J11/12 and J12/11 fold into two interleaving T-loop motifs that are stabilized by an intricate network of non–Watson-Crick interactions among the conserved nucleotides (Fig. 2E).

The most striking feature of yeast RNA is that six stems (P3, P7, P8′, P10/11, P12, and P15) are loosely connected to the helical core, expanding outward from the RNA center (Fig. 2A). This expanded conformation of Rpr1 yields three big open holes in the RNA structure (Fig. 2A). In sharp contrast, RNAs of bacterial RNase Ps adopt much more compact, two-layered configurations, with auxiliary elements that mediate long-range interactions among different RNA regions to ensure the correct fold and the stability of the RNA (fig. S5, A and B) (23, 40, 41).

Another difference between bacterial and yeast RNase P RNAs is from the P3 branch. In bacteria RNase P RNA, the P3 branch is a continuous base-paired stem (fig. S4, A and B). By contrast, the P3 branch of yeast Rpr1 contains a large unpaired region that separates the P3 branch into two helical stems (Fig. 2, A and B). Hereafter, we will refer to the proximal and distal helical stems as P3 and P3′, respectively, and the two unpaired single-stranded regions between stems P3 and P3′ as junctions J3/3′ and J3′/3, respectively (Fig. 2, A and B).

Head module of the protein hook and its interaction with Rpr1

The largest protein subunit Pop1 is specific to eukaryotic RNase Ps (Fig. 1, A to C) (42). From the N to C termini, Pop1 contains three motifs: an N-terminal motif (NTM), an internal motif (INM), and a large C-terminal globular domain (CTD) (Fig. 3A). A long helix α1 of Pop1NTM sticks out and mediates the interaction with Pop5 (Fig. 3A). The rest of Pop1NTM folds into a small helical and highly basic structure, plugging into the open junction between CR-IV and stems P4, P7, and P15 of Rpr1 (Fig. 3B). The side chains of multiple arginine residues in Pop1NTM make extensive stacking and electrostatic interactions with nucleotides in stems P4, P7 and P15, and CR-IV from four different directions (Fig. 3B and fig. S5A). These arginine residues are highly conserved from yeast to humans (fig. S5B), suggesting that Pop1NTM plays an important role in stabilizing the RNA structure in all eukaryotic RNase Ps. Consistent with this notion, mutation of conserved Arg97Arg98Arg99 in yeast Pop1 resulted in severe defect in pre-tRNA processing (43). The INM of Pop1 contains three helices and a long loop, fitting into the big open hole among stems P7, P8′, P8, P9, and P11 and stabilizing the conformation of the S domain of Rpr1 (Fig. 3C and fig. S5C). Pop1CTD is located at the concave side of the Rpr1 RNA opposite to Pop1NTM (Fig. 3D). Three conserved basic patches in Pop1CTD mediate interactions with different regions of Rpr1, covering a large surface area of the helical core of Rpr1 and stabilizing the single-layered architecture of the RNA (Fig. 3D and fig. S5D). Consistent with this notion, mutations of key residues on all three patches cause defects in pre-tRNA processing (fig. S5B) (43).

Fig. 3 The protein hook and its interaction with the Rpr1 RNA.

(A) Overall structure of Pop1. (B) An overall view of the interaction between Pop1NTM and the Rpr1 RNA. The RNA is shown as cartoon with semitransparent surface. CR-IV and stems P4, P7, and P15 that forms the big open hole are denoted. (C) An overall view of Pop1INM (colored in yellow) that fits into the big open hole among stems P7, P8′, P8, P9, and P11. (D) Interactions between Pop1CTD and Rpr1. (E) The head module of the protein hook wraps around the C domain and part of the S domain of Rpr1, stabilizing stems P3′ and P15. (F) Overall structure of the Pop5-Pop8-(Rpp1)2 heterotetramer. (G) The Pop5-Pop8-(Rpp1)2 heterotetramer sits on one side of the C domain of Rpr1, mediating extensive interactions with the pseudo-knot. (H) Analysis of the electrostatic surface potential of Pop5 reveals a highly basic deep cleft that holds CR-IV tightly (red, negative; blue, positive). (I) Overall structure of the Pop4-Rpr2-Pop3 heterotrimer. (J) Overall view of the interaction between the Pop4-Rpr2-Pop3 heterotrimer and the Rpr1 RNA. (K) The Pop4-Rpr2-Pop3 heterotrimer functions as a bridge between the C and S domains of Rpr1.

The P3 branch of Rpr1 binds to the concaved surface of the saddle-shaped Pop6-Pop7 heterodimer, with junctions J3/3′ and J3′/3 sitting on the basic center of the saddle (fig. S6, A and B). In contrast to Pop6, which only interacts with the P3 branch, Pop7 also contacts stem P15 and is sandwiched in the wedge between stems P3 and P15 of Rpr1 (fig. S6C). Binding with Pop6-Pop7 induces an ~120° bend between the P3 branch and the single-layered core of Rpr1, so that Pop7 and Pop1 form an extensive interface and the P3 branch of Rpr1 is encircled by Pop1, Pop6, and Pop7 (fig. S6D). Taken together, the head module of the protein hook wraps around the C domain and part of the S domain of Rpr1, stabilizing the helical RNA core as well as loosely connected stems P3′ and P15 (Fig. 3E).

Arm module of the protein hook and its interaction with Rpr1

The arm module of the protein hook contains two subcomplexes, the Pop5-Pop8-(Rpp1)2 heterotetramer and the Pop4-Rpr2-Pop3 heterotrimer (Fig. 1C). Pop5, Pop8, and two copies of Rpp1 form a heterotetramer with a pseudo-twofold symmetry (Fig. 3F). The heterotetramer is connected with the head module through two discrete, reciprocal interactions with Pop1 involving Pop5 and one Rpp1 molecule (Fig. 1C). Hereafter, we will refer to the Rpp1 molecule that interacts with Pop1 as Rpp1A and the other as Rpp1B (Fig. 1C). Although there is very limited sequence similarity between Pop5 and Pop8, their structures highly resemble each other, with a root mean square deviation of 2.5 Å (fig. S7A). The Pop5-Pop8-(Rpp1)2 heterotetramer sits on one side of the C domain of Rpr1 opposite to Pop1CTD (Figs. 1B and 3G and fig. S7, B and C). The surface of one side of Pop5 is highly basic, forming a deep cleft that tightly holds the zig-zagged CR-IV of Rpr1 (Fig. 3H). The highly conserved basic N terminus of Pop5 embraces CR-IV, with the side chains of Arg3 and Lys5 pointing into the narrow space between CR-I, CR-IV, and CR-V (fig. S7D). In addition to Pop5, Rpp1B also coordinates electrostatic interactions with both CR-I and CR-V (fig. S7B), so that Pop5 and Rpp1B together stabilize the pseudo-knot structure of Rpr1 by means of a large extended complementary interface (Fig. 3G).

This Pop4-Rpr2-Pop3 heterotrimer is connected to the Pop5-Pop8-(Rpp1)2 heterotetramer through extensive interactions between the β-barrel of Pop4 and the C-terminal long coiled coil of Rpp1B, so that the two subcomplexes together form the rod-shaped arm module of the protein hook (Figs. 1C and 3I and fig. S8A). The β-barrel of Pop4 interacts with the minor groove of stem P1 of Rpr1 through electrostatic interactions (Fig. 3J and fig. S8B). The two ends of long helix α3 of Pop4 respectively pack on the termini of stems P1 and P9, stabilizing the tetraloop-tetraloop receptor interaction between stems P1 and P9 (Fig. 3J and fig. S8C). Rpr2 mainly interacts with the S domain of Rpr1; helices α2 and α3, one edge of the β-sheet, and peripheral loops of Rpr2 constitute a positively charged depression that holds the U-shaped T-loop of J12/11 of Rpr1 (Fig. 3J and fig. S8D). In contrast to Pop4 and Rpr2, Pop3 makes limited direct interactions with the RNA; only the extended β-sheet of Pop3 contacts J11/12 of Rpr1 (Fig. 3J and fig. S8E). Taken together, the Pop4-Rpr2-Pop3 heterotrimer functions as a bridge between the C and S domains of Rpr1, stabilizing the position of J11/12 and J12/11 relative to the helical core of the RNA (Fig. 3K).

The protein hook stabilizes the Rpr1 RNA

To corroborate our structural study, we carried out molecular dynamics (MD) simulations to investigate how the protein components affect the structural stability of the Rpr1 RNA. MD simulation trajectories indicated that Rpr1 alone exhibited a very large fluctuation in the backbone phosphate atoms (Fig. 4A). Association of the protein hook with Rpr1 greatly suppressed this fluctuation except for stems P8′ and P12, which make no direct contact with the proteins (Fig. 4A). Binding with Pop1 alone was able to substantially reduce the fluctuation of Rpr1, especially the P3 branch and stem P15 (Fig. 4A). Other than the structural variation of individual nucleotides, MD simulation also showed that except for stems P8′ and P12, local cross-correlation in Rpr1 was suppressed upon protein hook association (fig. S9), suggesting that binding with proteins reduced the relative movements of individual structural elements within Rpr1 so that the RNA behaves more as a whole in the RNase P complex.

Fig. 4 The protein hook stabilizes the Rpr1 RNA.

(A) The root mean square fluctuation (RMSF) of backbone phosphorous atoms of the Rpr1 RNA alone (black), in the presence of all protein components (red), or only Pop1 (cyan), the Pop6-Pop7 heterodimer (magenta), the Pop5-Pop8-(Rpp1)2 heterotetramer (blue), or the Pop4-Rpr2-Pop3 heterotrimer (green). (B) Porcupine plots of the first two eigenvectors generated by means of PCA. The vectors represented as red arrows illustrate the tendency of the movement of the Rpr1 RNA in the absence (left) or in the presence (right) of the protein components. (C) Free-energy landscape calculated by projecting the conformational space onto the two principal components (PC1 and PC2) in the absence (top) or in the presence (bottom) of the protein components.

To obtain an overall picture of the atomic motions in Rpr1, we performed principle component analysis (PCA) using Cartesian coordinates of the phosphate atoms of Rpr1. The projection of trajectories onto the first two principal components, PC1 and PC2, accounted for a substantial amount of overall motion of Rpr1 in phase space. Consistent with the cross-correlation analysis, except for stems P8′ and P12, the overall motion of Rpr1 along PC1 and PC2 was greatly constrained by proteins in the RNase P complex (Fig. 4B and Movie 1). Furthermore, Gibbs free-energy landscape (FEL) analysis showed that for Rpr1 in the RNase P complex, a single global energy minimum was observed, indicating that the conformational state of Rpr1 is well restricted by the protein components (Fig. 4C). In sharp contrast, many local minimum-energy basins were observed for Rpr1 alone, which is suggestive of thermodynamically less stable conformations (Fig. 4C). Collectively, MD simulation revealed an obvious stabilizing effect on the Rpr1 RNA by the protein hook.

Movie 1. The first two essential modes of Rpr1 obtained from PCA using MD simulation trajectories.

(Left) Absent of the protein components. (Right) Present of the protein components.

pre-tRNA recognition

To gain insights into the pre-tRNA recognition and 5′-leader processing mechanism, we determined the cryo-EM structure of yeast RNase P in complex with a yeast pre-tRNAPhe at a resolution of 3.5 Å (figs. S10, A to H, and S11, A to E, and table S1). In this pre-tRNA substrate bound structure, the pre-tRNA substrate sits in a large open pocket on one side of the holoenzyme (Fig. 5, A to C). The coaxially stacked acceptor and T stems of the pre-tRNA substrate make extensive intermolecular interactions with both the Rpr1 RNA and protein components, positioning the scissile phosphate of pre-tRNA right at the active site of the ribozyme (Fig. 6A and fig. S11, A to C). In addition, one side of the anticodon arm of pre-tRNA also mediates extensive contacts with Rpp1B through complementary interface, further stabilizing the pre-tRNA in the binding pocket (Fig. 5, A to C).

Fig. 5 Overall structure of the S. cerevisiae RNase P holoenzyme complexed with pre-tRNAPhe.

(A) Overall view of the EM density map of S. cerevisiae RNase P complexed with pre-tRNAPhe. The individual protein and RNA subunits of RNase P and the pre-tRNA substrate are colored according to the scheme shown at the right of the figure. A close-up view of the active center bound with the pre-tRNA substrate is shown on the right. (B to C) Overall structure of the S. cerevisiae RNase P–pre-tRNA complex. The protein components are in cartoon (B) or surface (C) representation. The RNA component and pre-tRNA are shown in cartoon.

Fig. 6 The pre-tRNAPhe substrate recognition.

(A) Overview of the interactions between the yeast RNase P holoenzyme and the pre-tRNA substrate. Interactions at the two anchor sites are highlighted in red and blue boxes, respectively. (B) A close-up view of the stacking interactions between the TψC and D-loops of pre-tRNA and the T-loops in CR-II and CR-III of the Rpr1 RNA. (C) Two related close-up views of the interactions between Pop1NTM and pre-tRNA. The pre-tRNA, Pop1NTM, and the pseudoknot of Rpr1 were colored in cyan, green, and orange, respectively. (D) A close-up view of the interaction between the 3′-tailor of pre-tRNA and Pop1NTM. (E) A close-up view of the interactions between Pop1NTM and the first three base pairs in the acceptor stem of pre-tRNA. (F) Distribution of the distance between the backbone phosphorus atom of A344 and the C1′ atom of G245 of Rpr1, sampled from a total of ~1.8 μs integrated accelerated molecular dynamics simulation trajectories of each system. (G) Overall view of the distance between the conserved A344 and G245 located in C and S domains of Rpr1, respectively. A pre-tRNA in ribbon representation is also shown. The distance between the cleavage site and the TΨC-loop is denoted. (H) Recognition of the 5′-leader of pre-tRNA by yeast RNase P. The protein subunits Pop5 and Rpp1B are shown in surface (left) or electrostatic surface (right) representation, respectively. Four nucleotides in the 5′-leader were colored in dark pink and shown in cartoon. The rest of pre-tRNA and the conserved pseudoknot of Rpr1 were shown in cartoon and colored in cyan and orange, respectively. (I) A close-up view of the interactions between the 5′-leader and RNase P. The individual components of RNase P are colored as in (H).

The cryo-EM structure of yeast RNase P–pre-tRNA complex provided an atomic model of the “measuring device” of eukaryotic nuclear RNase Ps. The “measuring device” for pre-tRNA recognition is mainly composed of two anchors in RNase P (Fig. 6A). At the first anchor site, two unpaired nucleotides—A204 and G245, from the conserved T-loops in the S domain of Rpr1—form π-π stacking interactions with two unstacked bases, C56 from the TψC loop and G19 from the D loop of the pre-tRNA, respectively (Fig. 6B and fig. S11D).

At the other end of the acceptor stem, where the scissile phosphate resides, the highly conserved N-terminal motif of Pop1 (Pop1NTM) functions as the second anchor to stabilize the pre-tRNA at the catalytic center in the C domain of RNase P (Fig. 6C). The N-terminal long helix α1 in Pop1NTM that packs with Pop5 sits right on the acceptor stem of pre-tRNA so that the loop between helices α1 and η1 in Pop1NTM (L1-1) packs on the G1-C72 base pair of pre-tRNA, forcing the 3′-tailor of pre-tRNA to adopt a sharp turn at A73 and fold back to the 5′-leader (Fig. 6C). The base of pre-tRNA A73 flips out away from the acceptor stem and is sandwiched between the base of Rpr1 G300 and the aliphatic side chain of Arg115 in Pop1NTM (Fig. 6D). The helical core of Pop1NTM also mediates extensive hydrogen-bonding and electrostatic interactions with the first three base pairs of the pre-tRNA (Fig. 6E).

MD simulation analysis revealed a broad distribution of the distance between the two anchor points of Rpr1, which is suggestive of a very dynamic behavior of the RNA (Fig. 6F). In the presence of the protein hook, this distance is tightly constrained to ~51 Å and is optimal to accommodate the coaxially stacked acceptor stem and T-stem of pre-tRNAs, which have a fixed length of ~45 Å from the cleavage site to the TψC loop (Fig. 6G). The Pop4-Rpr2-Pop3 heterotrimer by itself already had a strong stabilizing effect (Fig. 6F), which is consistent with this trimer functioning as a bridge between the C and S domains, stabilizing the position of the T-loops relative to the catalytic center (Fig. 5, B and C). Taken together, our structural and MD simulation results suggested that the protein hook stabilizes yeast RNase P into a “measuring device,” with two fixed anchors that recognize the L-shaped structure rather than specific sequences of pre-tRNA substrates (1).

5′-leader recognition

In the complex structure, at least four nucleotides in the 5′-leader of pre-tRNA were visible in the cryo-EM density map (fig. S12A). The recognition of the 5′-leader involves both the Rpr1 RNA and protein subunits. Pop5 and Rpp1B form a continuous basic surface that holds the 5′-leader (Fig. 6H). The 5′ terminal nucleotide A(–4) of pre-tRNA sticks into a hydrophobic pocket between Pop5 and Rpp1B, with its base stacking with the side chain of Phe140 of Rpp1B (Fig. 6I). Nucleotides at the –1 and –2 positions of pre-tRNA—A(–2) and A(–1), respectively—sequentially stack on G1 following the double-stranded trajectory of the acceptor stem (Fig. 6I). This continuous base-stacking is extended into nucleotide A314 in the zig-zagged turn of CR-IV in Rpr1 (Fig. 6I). Nucleotide G(–3) of the 5′-leader diverges from this base stack and packs on C313 of Rpr1 (Fig. 6I). In addition to these stacking interactions, G(–3) and A(–2) also make multiple hydrogen-bonding interactions with C313 and A314 of Rpr1 (fig. S12B), further strengthening the connection between the 5′-leader of pre-tRNA and CR-IV of Rpr1. The zig-zagged turn of CR-IV is tightly grabbed by a highly basic cleft of Pop5, emphasizing that one of the major functions of Pop5 is to stabilize CR-IV of Rpr1 for 5′-leader recognition. In addition to this indirect recognition of the 5′-leader via CR-IV of Rpr1, Pop5 also directly contacts the backbone of nucleotides A(–1) and A(–2) of pre-tRNA through hydrogen-bonding interactions (Fig. 6I).

Active site

The location of the catalytic center is inferred from the scissile phosphate of the pre-tRNA substrate, which resides right on the P4 stem of the Rpr1 RNA (Fig. 7A). The distance between the scissile phosphate and the closest protein subunit Pop5 is ~7 Å, arguing against protein components playing any direct role in the catalysis (fig. S13A). By contrast, the scissile phosphate of pre-tRNA is closely surrounded by a panel of Rpr1 nucleotides: A91, U92, and U93 from CR-I and G343 and A344 from CR-V (Fig. 7A). Equivalent nucleotides of A91 and U92 in Bacillus subtilis and Escherichia coli RNase P RNAs have been implicated in metal ion coordination and catalysis (24, 35, 36, 39, 44, 45).

Fig. 7 Configuration of the active site and the conformational change of the catalytic center induced through pre-tRNA binding.

(A) (Left) An overall illustration of the catalytic center of the yeast RNase P. The Rpr1 RNA and pre-tRNA are colored in orange and cyan, respectively. Mg2+ ions (M1 and M2) are shown as green spheres. (Right) A close-up view of the catalytic center. The Rpr1 RNA and pre-tRNA are colored in white and cyan, respectively. The coordination of the two Mg2+ ions are highlighted in magenta dashed lines. (B) Comparison of the active center between the apo and pre-tRNA–bounded states of yeast RNase P. Pop5, pre-tRNA, and Rpr1 in the pre-tRNA–bound state are colored in slate, cyan, and silver, respectively, and Pop5 and Rpr1 in apo RNase P are colored in green and yellow, respectively. Conformational changes of Rpr1 U93 and the N terminus of Pop5 are denoted. (C) The RMSF of Rpr1 nucleotides along the morphed trajectory, which is generated through linearly interpolating the conformations between the apo and pre-tRNA–bound states of Rpr1 based on the cryo-EM structures. (D) A close-up view of the detailed interactions among the M1 Mg2+ ion, Arg99 of Pop1NTM, and the C94-G349 base-pair of Rpr1.

A putative Mg2+ ion (M1) was identified in the vicinity of U92 and U93 in the EM density map (fig. S13B). This site coincides with a metal site observed in the crystal structures of T. maritima and Bacillus stearothermophilus RNase Ps (23, 41). The M1 Mg2+ ion is coordinated by inner-sphere contacts with the hydroxyl of the conserved U93 in Rpr1 and three nonbridging phosphoryl oxygens, pro-Sp of Rpr1 A91, pro-Rp of Rpr1 U92, and pro-Rp of the scissile phosphate of pre-tRNA (Fig. 7A). Several lines of evidence from previous nucleotide substitution studies of bacterial RNase P support this geometry of the M1 Mg2+ ion and its environment. First, phosphorothioate-rescue experiments show that the pro-Rp but not Pro-Sp oxygen of G50 in B. subtilis RNase P RNA (equivalent to yeast Rpr1 U92) binds one metal ion (36). Second, the 4-thiouridine (4SU) modification of U51 of B. subtilis RNase P RNA (equivalent to yeast Rpr1 U93) greatly decreases phosphodiester bond cleavage of pre-tRNA (24).

A hallmark of the two–metal ion catalysis is that the two metal ions maintain a distance of ~4 Å and are located roughly in line with the phosphor-sugar backbone on the opposite sides of the scissile phosphate through interactions with the same nonbridging phosphoryl oxygen (27, 46). Although the local cryo-EM density map did not unambiguously reveal the position of the second metal ion, we could model a putative Mg2+ (M2) ion that is within the density map and is in accordance with the aforementioned constraints (fig. S13B). The M2 Mg2+ ion is coordinated by nonbridging phosphoryl oxygens of A91, G343, and A344 of Rpr1 and the O3′ leaving group of A(–1) in the pre-tRNA substrate (Fig. 7A). Consistently, previous biochemical studies have implicated major constituents of the M2 site in B. subtilis RNase P RNA, pro-Sp oxygen of A50, and pro-Rp of A390 (equivalent to yeast Rpr1 U92 and G344, respectively), in direct inner-sphere coordination of a Mg2+ ion important for RNase P activity (32, 36, 38, 39). Taken together, the structure of the catalytic center in yeast RNase P is in perfect agreement with previous biochemical data of bacterial RNase P, suggesting that the configuration of the RNA-based catalytic center is universally conserved in all RNase Ps, from bacteria to eukarya.

Substrate-induced activation of RNase P

Previous kinetic studies suggested that a conformational change may occur after the formation of the initial RNase P–pre-tRNA complex and before the cleavage step to form the product (18, 19, 25, 47, 48). To examine this conformational change, we superimposed the aporibozyme (apo) and pre-tRNA–bound yeast RNase P structures and found a dramatic conformational change in the active center of Rpr1, although the overall structure of the RNA remains largely unchanged (Fig. 7, B and C; fig. S14, A and B; and Movie 2). In the presence of pre-tRNA, the N-terminal tail of Pop5 rotates toward the catalytic center to form two hydrogen bonds with nucleotide A(–2) in pre-tRNA, helping stabilize the 5′-leader in the active site (Fig. 7B and fig. S14, A and B). Consequently, the conformation of nucleotides G343 and A344 in Rpr1 is adjusted to be optimal for coordinating a catalytic Mg2+ ion (M2) (Fig. 7B and fig. S14, A and B). The most prominent conformational change is from nucleotide U93 of Rpr1. Upon binding of pre-tRNA, U93 is forced to rotate a large angle from a position outside of stem P4 to point into the catalytic center and coordinate the other catalytic Mg2+ ion (M1) (Fig. 7, B and C, and fig. S14, A and B). The position of U93 in the active center is further stabilized by the C94-G349 base pair in stem P4 of Rpr1 through hydrogen-bonding interactions (Fig. 7D). MD simulation analysis revealed that in the apo-ribozyme, U93 bulges out of the P4 stem of Rpr1 and exhibits a dynamic conformation (fig. S15). In sharp contrast, binding of pre-tRNA tightly constrains U93 in a fix orientation that precisely matches the conformation observed in the complex structure (fig. S15). Taken together, comparative analysis of the apo and pre-tRNA–bound structures of yeast RNase P revealed a Mg2+ ion–dependent substrate induced conformational change in the catalytic center, which transforms the ribozyme from an inactive into an active state.

Movie 2. Local conformational change in the active center of RNase P upon pre-tRNA binding.

The morphed trajectory was generated through linearly interpolating the conformations of the Rpr1 RNA between the apo and pre-tRNA–bound states of yeast RNase P based on the cryo-EM structures. Although the intermediates are unphysical, the movie clearly demonstrates the major conformational differences of Rpr1 upon pre-tRNA binding.

Catalytic mechanism

On the basis of the structure of the active site, we proposed a catalytic mechanism of pre-tRNA cleavage by RNase P. In this model, the M1 Mg2+ ion coordinates the pro-Sp oxygen of the scissile phosphate and an attacking nucleophilic water molecule (Fig. 8A). The catalytic role for the M1 Mg2+ ion is to lower the pKa value of the nucleophilic water (where Ka is the acid dissociation constant and pKa = –log10Ka), thus facilitating the deprotonation of the water molecule to generate a hydroxide ion that performs an SN2 in-line attack on the phosphodiester bond with a trigonal bipyramidal transition state (Fig. 8A). The M2 Mg2+ ion coordinates the same pro-Sp oxygen of the scissile phosphate as well as the upstream ribose O3′ oxygen atom, playing a stabilizing role in maintaining the geometry of the transition state and the leaving 5′-leader (Fig. 8A).

Fig. 8 Catalytic mechanism of pre-tRNA processing catalyzed by yeast RNase P.

(A) Snapshot of the reactant state showing eight distances d1 to d8 that are used to describe the progress of the reaction. The Rpr1 RNA and pre-tRNA are colored in white and cyan, respectively. Mg2+ ions (M1 and M2) are shown as green spheres. The nucleophile water (W1) and the bulk water (W2) that serves as a general base to accept the nucleophile’s proton are shown in stick model and sphere, respectively. W1 was modeled in the catalytic center based on the two-metal ion mechanism, and W2 was the result of the QM/MM simulation. The rest of the catalytic center—including Mg2+ ions, Rpr1, and pre-tRNA—was based on the EM density map. (B) (Left) Bond distance changes monitored along the MFEP during the QM/MM free-energy simulations. Distances d1 to d8 are defined as in (A). d8 is the distance between the leaving proton and the nucleophilic water oxygen. (Right) Free-energy profile along the MFEP. The reactant state (R), transition state 1 (TS1), transition state 2 (TS2), and product formation state (P) are highlighted with colored areas. (C) Proposed proton transfer pathway and two-metal-ion mechanism of the phosphodiester bond cleavage reaction of pre-tRNA catalyzed by yeast RNase P. (D) Kinetic model of RNase P–mediated pre-tRNA processing. RNase P first recognizes the pre-tRNA substrate through a double-anchor mechanism to form the initial low-affinity ES complex. Binding of pre-tRNA then induces a local conformational change in the catalytic center of RNase P. In particular, nucleotide U93 within stem P4 of the Rpr1 RNA undergoes a conformational change to mediate an inner-sphere coordination of the catalytically important Mg2+ ion, so that the ES complex is isomerized to the active ES* state. The activated ES* complex then catalyzes the phosphodiester bond cleavage of pre-tRNA through a two-metal-ion SN2 mechanism to release the 5′-leader of pre-tRNA. Last, the mature tRNA dissociates from the holoenzyme in a slow, rate-limiting step, and RNase P is ready for the next round of catalysis.

We used multiscale quantum mechanical/molecular mechanical (QM/MM) free-energy simulations to unveil the underlying catalytic mechanism of yeast RNase P. First, we performed 250-ns classical MD simulations to evaluate the structural arrangement of the catalytic center, where a putative nucleophilic water molecule is proposed to coordinate the M1 Mg2+ ion (fig. S16). The key distance profiles (d1 to d7) of the active site obtained from classical MD simulations are quantitatively consistent with the structural architecture of the catalytic site observed in the cryo-EM structure (fig. S16).

Next, we carried out QM/MM simulations, combining umbrella samplings and path collective variables (CVs), to obtain the free-energy profile of the phosphodiester cleavage reaction of pre-tRNA. A path connecting the reactant and product in a seven-CV space (distance d1 to d7) was used to explicitly define the progress of the reaction. The free-energy profile along the converged minimum free-energy path (MFEP) revealed a metastable intermediate (INT) region separated by two transition states (TS1 and TS2) (Fig. 8B). After crossing the TS1 by overcoming a free-energy barrier of ~18.56 kcal/mol, the free-energy profile exhibits an intermediate region, corresponding to the inversion of stereo configuration of the pentacovalent phosphorane intermediate (Fig. 8B). Then, the reaction proceeds to cross the rate-limiting step TS2 with a free-energy barrier of ~18.75 kcal/mol (Fig. 8B). The calculated free-energy barrier is in excellent agreement with the experimentally measured rate constants of 0.35 ± 0.03 s−1, corresponding to a Gibbs free energy change (ΔG) value of ~18.19 kcal/mol (25).

On the basis of the simulations, we depicted the details of the catalytic mechanism. (i) The early stage of the MFEP starts with a decrease of the distance between the scissile phosphate and the nucleophile water (d2), indicating the beginning of the nucleophilic water to attack the scissile phosphate (Fig. 8, B and C). The distance between the scissile phosphate and the leaving O3′ atom (d1) exhibits a perceptible increase after d2 is less than 2.7 Å (path node = 3) (Fig. 8B). The distance between the M2 Mg2+ and the leaving O3′ group (d7) also decreases substantially when the nucleophilic water approaches the scissile phosphate before the TS is reached, suggesting stabilization of the leaving O3′ by the M2 Mg2+ ion. (ii) At the TS1 point, where d1 equals d2, the free-energy profile reaches the first maximum value, and the scissile phosphate evolves into a pentavalent phosphorane-like geometry (Fig. 8, B and C). Simultaneously, the nucleophilic water’s proton is released, and a hydroxyl group is formed after the TS1 (Fig. 8, B and C). We found that a bulk water molecule, stabilized by the nonbridging oxygen atoms of G1 in pre-tRNA and U93 in Rpr1, serves as a general base to accept the nucleophile’s proton (Fig. 8, A to C). A transient hydronium ion (H3O+) is identified to facilitate the proton transfer to the pro-Sp oxygen of the scissile phosphate, forming a mono-anionic phosphorane intermediate (Fig. 8, A to C). (iii) The product state is accompanied by the formation of the new phosphoryl bond between the nucleophile and the scissile phosphate (d2 ≈ 1.64 Å) and the departure of the leaving O3′ group (d1 ≈ 3.00 Å). (iv) During the late stage of the MFEP, the free-energy profile exhibits a downhill process, and the stereo configuration at the phosphorus is inverted, indicating a typical in-line SN2 nucleophilic substitution reaction mechanism (Fig. 8, B and C). We observed that the proton that is transferred onto the pro-Sp oxygen of the scissile phosphate at TS2 is spontaneously shuttled toward the leaving O3′ group after the inversion of the configuration (Fig. 8, B and C).

Discussion

The catalytic mechanism of how RNase P facilitates the cleavage of pre-tRNA is a central question in the RNase P field. Our structural and simulation data reported here provide an integrated mechanistic insight into the RNase P–catalyzed pre-tRNA processing (Fig. 8D). Our data reveal that the RNase P holoenzyme is largely preassembled to engage the pre-tRNA substrate through a two-anchor mechanism (Fig. 6A). One of the anchors, the T-loops in the S domain of Rpr1, that recognizes the TψC and D loops of pre-tRNA is universally conserved, whereas the other that recognizes the pre-tRNA cleavage site has evolved from an RNA-based apparatus in bacteria to a protein-based one in eukaryotic RNase Ps (Fig. 6C) (23). The distance between the two anchors is optimal for accommodating the coaxially stacked acceptor and T stems of pre-tRNA substrates (Fig. 6F), suggesting that these anchor sites function as the “measuring device” to recognize the pre-tRNA substrate. We propose that it is this “measuring device” of all RNase P complexes that mediates the initial engagement with pre-tRNA to form the low-affinity ES complex (Fig. 8D).

The pre-tRNA–bound RNase P structure reveals that the recognition of the 5′-leader of pre-tRNA involves both Rpr1 and the protein subunit Pop5. A highly basic cleft on the surface of Pop5 tightly holds a zig-zagged turn in CR-IV of Rpr1 that mediates both stacking and hydrogen-bonding interactions with the 5′-leader (Fig. 6, H and I, and fig. S12B). Bacterial sole protein subunit Rpp also contains a similar highly basic cleft that holds CR-IV of the bacterial RNA in the same fashion (fig. S17) (23). Therefore, it is likely that the 5′-leader of pre-tRNA is recognized by bacterial RNase P through the same mechanism as in yeast RNase Ps. Unlike the 5′-leader, the 3′-tailor of pre-tRNA only makes a limited contribution to the recognition by yeast RNase P (Fig. 6D), which is consistent with the fact that the 3′-CCA sequence in eukaryotic tRNA is added posttranscriptionally (49). This is in sharp contrast to bacterial tRNA, in which the 3′-CCA is transcriptionally encoded and serves as a key element recognized by bacterial RNase P through base-pairing with nucleotides in the L15 loop of RNase P RNA (23).

A comparison of the apo and pre-tRNA–bound structures of yeast RNase P indicates that after the initial recognition of pre-tRNA by the “measuring device” of RNase P, the pre-tRNA substrate induces a local conformation change in the catalytic center, so that the phosphate oxygen–rich cleft near the P4 stem together with the scissile phosphate and the O3′ leaving group of pre-tRNA coordinate two catalytic Mg2+ ions to form the active state that is primed for catalysis (Fig. 7, A and B). In particular, universally conserved U93 undergoes a dramatic conformational change to mediate an inner-sphere coordination with the catalytically important M1 Mg2+ ion, helping to properly position the local pre-tRNA structure at the active site (Fig. 7B). Consistent with our data, previous studies suggested that nucleotide U51 in B. subtilis RNase P (equivalent to U93 in yeast Rpr1) contributes to a divalent metal ion–dependent conformational change during the RNase P–catalyzed reaction (18, 24). We proposed that this pre-tRNA–induced Mg2+ ion–dependent conformational change in the catalytic center of RNase P corresponds to the conformational change between the initial ES complex and the cleavage step. Consequently, the active site observed in the pre-tRNA–bound complex structure could represent the active ES* state or a transition configuration between ES and ES* states when the M2 ion is still not fully stabilized in the catalytic center (Fig. 8D) (1820). We further speculated that the conformation of the initial ES complex should be largely identical to that of the ES* complex except for the catalytic center, which instead adopts the conformation of the active site in the apo structure of RNase P (Fig. 8D).

Our EM structure of the pre-tRNA–bound RNase P complex revealed the configuration of the catalytic center of RNase P in the presence of a pre-tRNA substrate (Fig. 7A). On the basis of this structural information, our QM/MM simulations visualized the mechanistic details of phosphodiester bond hydrolysis of pre-tRNA, which is a two-Mg2+-ion–mediated SN2 reaction (Fig. 8C). Our simulations also reveal a two-step proton transfer pathway from the nucleophilic water to the leaving O3′ group that involves a neighboring bulk water molecule during the cleavage process (Fig. 8, B and C). A similar bulk water that facilitates the proton transfer process has also been suggested in the cleavage reaction catalyzed by group II introns (50).

The M2 site in our structure is different from the second Mg2+ (TmM2) site identified in the crystal structure of T. maritima RNase P in complex with a mature tRNA (fig. S18) (23). The TmM2 site was located in experiments in which the 5′-leader was soaked into the crystals in the presence of Sm3 (23). In contrast to the M1 and M2 sites in our structure, the geometry of the two Mg2+ ions in the T. maritima RNase P–tRNA complex is not consistent with the octahedral Mg2+ coordination geometry (27, 51). In addition, soaking a 5′-leader in a mature tRNA-bound complex does not represent a pre-tRNA substrate–bound state. Superposition analysis showed that the soaked 5′-leader did not occupy the same position as the unprocessed 5′-leader in the yeast RNase P–pre-tRNA complex (fig. S17), arguing that the TmM2 site in the T. maritima RNase P–tRNA complex is not a site for a catalytic metal ion.

The cryo-EM structures reported here represent a major step forward for mechanistic understanding of the RNase P function. Our data support that all RNase P ribozymes share an RNA-based, substrate-induced catalytic mechanism of pre-RNA processing. Whereas bacterial RNase P RNA is catalytically active by itself, eukaryotic RNase P is a protein-controlled ribozyme; its protein components not only directly participate in substrate recognition but also stabilize the catalytic RNA in a conformation optimal for pre-tRNA binding and cleavage reaction.

Methods and materials summary

Yeast strain constructions and affinity purification of the RNase P holoenzyme were performed as described previously, with minor modifications (16). Briefly, protein subunits Pop4 and Rpr2 were tagged with protein A and 3×Flag, respectively. A two-step affinity purification scheme was used to purify yeast RNase P. For substrate-bound structural studies, the in vitro transcribed yeast pre-tRNAPhe was mixed with the highly purified yeast RNase P at 4°C and immediately subjected to cryo-EM grid preparation. For cryo-EM data collection and processing, the specimens were first prepared by using a lacey carbon film with a continuous ultrathin carbon film. Images were taken on a Gatan K2 summit camera mounted on an FEI Titan Krios electron microscope operated at 300 kV. Image processing includes motion correction, CTF estimation, particles classification, and refinement. De novo atomic model building, rigid docking of known structures, and homologous structure modeling were combined to generate the atomic model for the entire S. cerevisiae RNase P holoenzyme. QM/MM free-energy simulation was used to investigate the catalytic reaction of yeast RNase P. Detailed descriptions for all the materials and methods are provided in the supplementary materials.

Supplementary Materials

www.sciencemag.org/content/362/6415/eaat6678/suppl/DC1

Materials and Methods

Figs. S1 to S18

Table S1

References (52101)

References and Notes

Acknowledgments: We thank the staff members of the Electron Microscopy System at the National Facility for Protein Science in Shanghai (NFPS), Zhangjiang Laboratory, China, and the Core Facilities for Protein Science at the Institute of Biophysics (IBP), Chinese Academy of Sciences, for providing technical support and assistance in data collection. We thank M. Cao for his help on data collection and analysis. Mass spectrometry experiments were performed at NFPS. Funding: This work was supported by grants from the National Natural Science Foundation of China (31525007 to M.L. and 21625302 to G.L.) and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB08010201 to M.L.). Author contributions: M.L., P.L., G.L., J.W., and H.C. conceived the study. M.L. supervised the whole project. P.L., S.N., and J.C. purified the yeast RNase P complex. P.L., M.T., X.W., and S.S. prepared the cryo-EM sample. M.T., P.L., X.W., S.Q., and G.C. collected the EM micrographs and processed the data. J.W. and M.L. built the atomic model. Y.Z., G.L., and X.P. performed the MD and QM/MM simulation analyses. M.L. and P.L. wrote the manuscript. Competing interests: The authors declare no competing interests. Data and materials availability: The accession numbers for the structure reported in this paper are PDB: 6AGB and 6AH3 and EMDB: EMD-9616 and EMB-9622.
View Abstract

Stay Connected to Science

Navigate This Article