Mechanistic Studies Reveal Similar Catalytic Strategies for Phosphodiester Bond Hydrolysis by Protein-only and RNA-dependent Ribonuclease P*

Background: A protein-only ribonuclease P (PRORP) has been recently discovered. Results: PRORP activity has a single ionization (pKa ∼ 8.7) important for catalysis and a cooperative dependence on Mg2+ (nH = 2). Conclusion: PRORP uses catalytic strategies similar to RNA-dependent RNase P. Significance: These results provide evidence for the mechanistic convergence of two different enzymatic macromolecules (RNA and protein) that perform the same biological function. Ribonuclease P (RNase P) is an endonuclease that catalyzes the essential removal of the 5′ end of tRNA precursors. Until recently, all identified RNase P enzymes were a ribonucleoprotein with a conserved catalytic RNA component. However, the discovery of protein-only RNase P (PRORP) shifted this paradigm, affording a unique opportunity to compare mechanistic strategies used by naturally evolved protein and RNA-based enzymes that catalyze the same reaction. Here we investigate the enzymatic mechanism of pre-tRNA hydrolysis catalyzed by the NYN (Nedd4-BP1, YacP nuclease) metallonuclease of Arabidopsis thaliana, PRORP1. Multiple and single turnover kinetic data support a mechanism where a step at or before chemistry is rate-limiting and provide a kinetic framework to interpret the results of metal alteration, mutations, and pH dependence. Catalytic activity has a cooperative dependence on the magnesium concentration (nH = 2) under kcat/Km conditions, suggesting that PRORP1 catalysis is optimal with at least two active site metal ions, consistent with the crystal structure. Metal rescue of Asp-to-Ala mutations identified two aspartates important for enhancing metal ion affinity. The single turnover pH dependence of pre-tRNA cleavage revealed a single ionization (pKa ∼ 8.7) important for catalysis, consistent with deprotonation of a metal-bound water nucleophile. The pH and metal dependence mirrors that observed for the RNA-based RNase P, suggesting similar catalytic mechanisms. Thus, despite different macromolecular composition, the RNA and protein-based RNase P act as dynamic scaffolds for the binding and positioning of magnesium ions to catalyze phosphodiester bond hydrolysis.

and 3Ј (trailer) ends. Removal of these extra sequences is essential for tRNA function, making the enzymes that catalyze pre-tRNA end maturation essential. Ribonuclease P (RNase P) catalyzes the formation of the mature 5Ј end of pre-tRNA and has remarkable diversity in subunit and macromolecular composition throughout domains and even within the same species (1). RNA-based RNase P (ribozyme) contains varying numbers of protein subunits ranging from one in bacteria to at least 10 in eukaryotes (2).
Some eukaryotic species, such as Arabidopsis thaliana, are seemingly devoid of an RNA-based RNase P and instead encode a protein-only RNase P (PRORP) (3). Mammals, including humans, use both RNA-and protein-based RNase P to catalyze pre-tRNA maturation (4). This redundancy in humans can be partly explained by differential localization; PRORP (or MRPP3) functions in the mitochondria, whereas the RNAbased RNase P functions in the nucleus (4,5). Knockdown of human PRORP results in accumulation of mitochondrial pre-tRNA in HeLa cells, and knock-out of the homologous PRORP1 in A. thaliana is lethal (3,4,6). Thus, PRORP enzymes play an essential role in mitochondria.
In A. thaliana, three PRORPs are encoded within the nuclear genome (PRORP1, -2, and -3). PRORP2 and PRORP3 co-localize to the nucleus, whereas PRORP1 functions within both the mitochondria and chloroplasts (6). PRORP is homologous to the nuclease component of the human mitochondrial RNase P (MRPP3/human PRORP). The structure of A. thaliana PRORP1 revealed three domains: a Nedd4-BP1, YacP nuclease (NYN) metallonuclease domain, a central structural zinc-binding domain, and a pentatricopeptide repeat domain involved in pre-tRNA binding (Fig. 1A) (7). The NYN domain is a novel metallonuclease domain sharing structural homology to the PIN (PilT N terminus) and flap nuclease families (8). NYN domains have a relatively exposed active site and contain four conserved aspartates as compared with the flap nuclease family that has six conserved aspartates (7,8). Despite having only four conserved potential metal ligands within its active site, the crystal structure visualized the binding of two manganese ions to the NYN domain of PRORP1 in positions similar to those observed in the flap nuclease family (7,9). These similarities led to the proposal that NYN metallonuclease domains catalyze phosphodiester bond hydrolysis using a classical two-metal ion mechanism (6,7,10). In this mechanism, metal I is proposed to activate a coordinated water molecule for nucleophilic attack, both metals are proposed to stabilize the negative charge buildup in the transition state, and metal II is proposed to stabilize the developing charge on the 3Ј-oxyanion leaving group (10).
Here we investigate the catalytic mechanism of PRORP1 by examining the metal and pH dependence of pre-tRNA cleavage. These data provide evidence for a two-metal ion mechanism and identify a single ionization that is important for catalysis. The pH dependence is consistent with deprotonation of a catalytic metal-bound water and provides no evidence for an amino acid side chain acting as a general acid to protonate the leaving group. Alanine mutations of the conserved aspartate residues that coordinate metal ions significantly reduce single turnover activity as observed previously (6,7). However, addition of high concentrations of Mg 2ϩ can partially rescue the activity of the D474A and D475A mutant enzymes, indicating that these residues are important for enhancing metal ion affinity. These findings provide insight into the enzymatic mechanism and metal binding properties of PRORP1 that can be extended to the NYN metallonuclease domains found throughout Eukaryota. Furthermore, the mechanism of PRORP1 has important similarities to the RNA-dependent RNase P.

Experimental Procedures
Enzyme Preparation-A truncation of PRORP1 lacking the N-terminal 76 amino acids, encompassing the mitochondrial localization sequence, was used in this study. PRORP1 WT and mutants were purified as described previously (7). Enzyme concentration was determined by absorbance using an extinction coefficient at 280 nm of 84,630 M Ϫ1 cm Ϫ1 . Purified enzymes were aliquoted, flash frozen, and stored at Ϫ80°C in 20 mM MOPS, pH 7.8, 100 mM NaCl, and 1 mM TCEP.
Pre-tRNA Preparation-The A. thaliana pre-tRNA was produced by runoff transcription from a linearized plasmid containing a T7 promoter, a 5:1 (leader nt length:discriminator base) mitochondrial pre-tRNA Cys gene, and a 3Ј-BstNI restriction site. The Bacillus subtilis pre-tRNA Asp was transcribed from a PCR product containing a T7 promoter with a 5-nt leader and a discriminator base (1-nt trailer) using a plasmid encoding B. subtilis pre-tRNA Asp with a 5-nucleotide leader and GCCA trailer as the template (11). Transcription reactions were performed in the presence of GMPS and the 5Ј-phosphorothioate was incubated with 5-(iodoacetamido)fluorescein to fluorescently label the 5Ј end (12). Substrates were purified on 12% urea-polyacrylamide gels, crush-soaked to elute, then washed, and concentrated. The concentration of pre-tRNA substrates was calculated by absorbance using extinction coefficients of 674,390 and 685,000 M Ϫ1 cm Ϫ1 at 260 nm for A. thaliana and B. subtilis pre-tRNA, respectively. Labeling efficiency was assessed by fluorescein absorbance at 492 nm using an extinction coefficient of 78,000 M Ϫ1 cm Ϫ1 . Samples were stored at Ϫ20 or Ϫ80°C in 10 mM Tris, pH 7.4. Proposed tRNA secondary structures were generated with tRNAscan-SE 1.21 (13).
Enzyme Assays-Standard reaction conditions were 30 mM MOPS, pH 7.8, 150 mM NaCl, 1 mM TCEP, and 1 mM MgCl 2 at 25°C. Pre-tRNA substrates were denatured by heating at 95°C for 1 min in Milli-Q H 2 O and then refolded by incubating at 25°C for 15 min, adding reaction buffer, and incubating at 25°C Completely conserved residues are represented as blue sticks (Asp-399, Asp-493, Asp-474, Asp-475, and His-498). Partially conserved residues are in cyan (His-438 and Asp-497). Potential hydrogen bonds are depicted by magenta dashes (with other amino acids) and yellow dashes (with metal-bound waters). Black dashes represent proposed inner sphere coordination of the metal ions. In the crystal structure, metal II has a lower occupancy (ϳ80%), and the metal-bound waters are not visible (7).
for at least 15 min before use. For single turnover reactions, the enzyme and pre-tRNA concentrations were 5 M and 20 nM, respectively, unless otherwise noted. Single turnover reactions were initiated by addition of enzyme (5 M) and quenched at specified times with an equal volume of 100 mM EDTA, 6 M urea, 0.1% bromphenol blue, 0.1% xylene cyanol, and 2 g/l yeast tRNA. The labeled 5Ј leader product was separated from pre-tRNA by 20 or 22.5% urea-PAGE. Gels were visualized using a Typhoon 9410 scanner, and the fraction product was quantified using ImageQuant 5.2 software. The observed single turnover rate constant was calculated from a fit of a first order exponential equation to the data using KaleidaGraph fitting software. The k obs was not significantly dependent on the PRORP1 concentration (2 and 5 M), confirming that the enzyme was saturating (k max conditions).
Multiple turnover reactions were performed in black Costar half-area 96-well plates using a fluorescence polarization (FP) assay (as described in Ref. 14). Standard reaction conditions were used with an enzyme concentration of 20 nM. The concentration of fluorescently labeled pre-tRNA was held constant at 40 nM in the reactions, whereas the concentration of unlabeled pre-tRNA substrate was varied. The ratio of labeled to unlabeled substrate did not alter the measured initial rates. Initial rates were calculated from the linear decrease in fluorescence anisotropy (excitation, 488 nm; emission, 535 nm). The steady state kinetic parameters were calculated from a fit of the Michaelis-Menten (15) equation to the concentration dependence of the initial rates using KaleidaGraph.
Metal Dependence-The metal dependence of cleavage of B. subtilis pre-tRNA was measured under multiple turnover conditions using the FP assay. The pre-tRNA concentration dependence of the initial velocity was measured at seven MgCl 2 concentrations with the ionic strength held constant by adjusting the NaCl concentration ( ϭ 200 mM taking only the NaCl and MgCl 2 concentrations into account). The metal dependence of PRORP-catalyzedcleavageofA.thalianapre-tRNA Cys wasmeasured at two substrate concentrations, 250 nM and 5 M, mimicking k cat /K m and k cat conditions, respectively. Metal dependence reactions were similarly performed in the presence of 1 mM CaCl 2 . Equation 1 was fit to the Mg 2ϩ dependence of the apparent steady state kinetic parameters where k represents the apparent k cat or k cat /K m , k Mg represents the maximal rate constant (k cat Mg or (k cat /K m ) Mg ) at saturating Mg 2ϩ , and K m Mg is the concentration of Mg 2ϩ at which k Mg is half-maximal. For WT, the NaCl concentration was adjusted to correct for ionic strength, whereas 100 mM NaCl was used to assay the mutant enzymes. These single turnover reactions were quenched at 2 h with an equal volume of 100 mM EDTA, 6 M urea, 0.1% bromphenol blue, 0.1% xylene cyanol, and 2 g/l yeast tRNA. The fraction product was determined using 22.5% urea-PAGE.
pH Dependence-Reactions were performed under single turnover conditions (5 M PRORP and 20 nM pre-tRNA) in a three-component buffer system (MTA buffer): 50 mM MES (pK a 6.15), Tris (pK a 8), and 2-amino-2-methyl-1-propanol (pK a 9.7) (16,17). Reactions also included 1 mM TCEP, 1 mM MgCl 2 , and NaCl to maintain ionic strength ( ϭ 240 Ϯ 10 mM). The pH range assayed was 6.0 -9.75. Doubling the concentration of PRORP to 10 M at pH 6.5 and 8.0 increased the observed rate constant Ͻ20%, indicating that the enzyme concentration was saturating. The rate constants measured in MTA buffer were within 20% of assays performed in 50 mM PIPES, pH 6.95; Tris, pH 8.0; and CHES, pH 9.5. Reactions performed with A. thaliana pre-tRNA at and above pH 8.5 displayed biphasic kinetics possibly due to misfolded substrate (18). The observed rate constant for the first phase (largest amplitude) was used in the pH profile. Reactions at pH 9.5 and 9.75 for cleavage of B. subtilis pre-tRNA were performed using a quench-flow apparatus. A single ionization model (Equation 2) was fit to the pH dependence of the single turnover reaction where k max is the observed rate constant at a given pH, k max pH is the pH-independent rate constant, and the pK a is the pH when the activity is half-maximal.
Single turnover (STO) reactions performed in the presence of various divalent ions contained 8 M PRORP1, 50 mM MTA buffer at pH 6.5, 10 mM divalent metal, 1 mM TCEP, 250 M hexamminecobalt(III) chloride, 165 mM NaCl, and 30 nM B. subtilis pre-tRNA. k max was measured under various metal concentrations (1, 5, 10, and 20 mM) to ensure metal saturation and lack of inhibition. Anisotropy Binding Assays-Binding assays were performed as described previously (7). Briefly, the concentration of fluorescein labeled pre-tRNA was maintained at 20 nM, whereas the concentration of PRORP was varied (0.005-5 M). Binding experiments were performed in 20 mM MOPS, pH 7.8, 300 mM NaCl, 1 mM TCEP, and 1 mM CaCl 2 in a 96-well plate format. Fluorescence anisotropy was measured at an excitation wavelength of 488 nm and an emission wavelength of 535 nm (14). PRORP1 is not active in CaCl 2 (7). Equation 3 was fit to the dependence of the anisotropy on the protein concentration where A is the observed anisotropy, A 0 is the initial anisotropy, ⌬A is the total change in anisotropy, P is the concentration of PRORP, and K d is the apparent dissociation constant.

Results
Single and Multiple Turnover Kinetics of PRORP1-catalyzed Pre-tRNA Cleavage-We measured the STO and multiple turnover (MTO) kinetics of pre-tRNA cleavage catalyzed by PRORP1 to establish a kinetic framework to evaluate the effects Enzymatic Mechanism of Protein-only RNase P of pH, metal ion concentration, and mutations. Cleavage reactions were carried out using two substrates, A. thaliana mitochondrial pre-tRNA Cys (A. thaliana pre-tRNA) and B. subtilis pre-tRNA Asp (B. subtilis pre-tRNA), a bacterial substrate extensively used to analyze the reactivity of the RNA-based RNase P (Fig. 2). These substrates were 5Ј end-labeled with fluorescein and contain a 5-nt leader and a 3Ј discriminator base. For MTO reactions, we used a recently developed real time FP assay to monitor the dependence of PRORP1-catalyzed pre-tRNA cleavage on the substrate concentration (14). Representative MTO data from the FP assay are shown in Fig. 3A; the value of k cat /K m is ϳ 1 ϫ 10 5 M Ϫ1 s Ϫ1 , which is ϳ35-fold slower than the reaction catalyzed by B. subtilis RNA-dependent RNase P (19), and the value of k cat is 0.03-0.06 s Ϫ1 ( Table 1). Cleavage of pre-tRNA catalyzed by PRORP1 under STO conditions was too fast to measure using the FP assay. Therefore, these reactions were performed in a stopped assay format at saturating enzyme concentrations, and product formation was analyzed by urea-PAGE (Fig. 3C) as described under "Experimental Procedures." A single exponential equation was fit to the data (Fig. 3D). The values of the STO rate constant at saturating enzyme (k max ), 0.03-0.04 s Ϫ1 , are comparable with the multiple turnover parameter at saturating substrate (k cat ), suggesting that a step at or before cleavage is rate-limiting for both substrates under these conditions. PRORP1 Magnesium Dependence-The crystal structure of PRORP1 in the presence of Mn 2ϩ revealed two metal ion binding sites within the active site ( Fig. 1B) (7). Based on this observation and the similarity of the NYN domain of PRORP1 to other nucleases, it has been proposed that PRORP enzymes use a two-metal ion mechanism for catalyzing phosphodiester  subtilis pre-tRNA under standard reaction conditions. D, the gel was quantified, and the fraction product was calculated at each time point. A nonlinear regression curve of a single exponential fit to the data is shown. Kinetic parameters are summarized in Table 1. Fl, fluorescein.  Fig. 3. Values reported reflect the mean of two independent experiments with the error reported as the standard deviation.   4C). To determine whether these effects are substrate-specific, we also examined the Mg 2ϩ dependence of the A. thaliana pre-tRNA at a subsaturating substrate concentration (250 nM) reflecting k cat /K m conditions and a saturating substrate con-centration (5 M) reflecting k cat conditions. The magnesium dependence of the PRORP1-catalyzed cleavage of the A. thaliana pre-tRNA substrate mirrors that observed for the B. subtilis pre-tRNA substrate; a cooperative dependence on Mg 2ϩ concentration was observed under k cat /K m conditions (Fig. 4E) but disappeared under saturating substrate (k cat ) conditions (Fig. 4D). Finally, to test whether the magnesium cooperativity observed under k cat /K m conditions is due to stabilization of the structure of pre-tRNA, we repeated these activity measurements in the presence of 1 mM CaCl 2 . Calcium stabilizes tRNA structure (20) but does not activate PRORP1 catalysis (7). However, we still observed an n H of 2.1 Ϯ 0.2 for the Mg 2ϩ dependence of k cat /K m in the presence of 1 mM CaCl 2 (Fig. 4F). Potential models for the Mg 2ϩ dependence under k cat /K m conditions and k cat conditions are represented in Schemes 1 and 2, respectively.

Multiple turnover
Taken together, these magnesium-dependent data indicate that PRORP1 requires at least two Mg 2ϩ ions for optimal catalysis and that the Mg 2ϩ ions bind cooperatively to PRORP1. Additionally, because Mg 2ϩ cooperativity was only observed

Enzymatic Mechanism of Protein-only RNase P
under k cat /K m conditions, these data suggest that Mg 2ϩ binding to PRORP1 is coupled to pre-tRNA binding.
Metal Dependence of Aspartate to Alanine Mutants-The PRORP1 crystal structure also revealed amino acid residues that interact with the active site metal ions (7). Two manganese ions are observed bound to the active site interacting with four conserved aspartate residues (D399A, D474A, D475A, and D493A) through both inner and outer sphere interactions (Fig.  1B). Mutation of any one of these conserved aspartate residues to alanine significantly reduces activity (7); no product formation is observed after a 30-min incubation under standard STO assay conditions (1 mM MgCl 2 ). To evaluate whether the activity of these mutants could be enhanced by higher Mg 2ϩ concentrations, we compared the STO cleavage activity at 1 and 20 mM Mg 2ϩ (Fig. 5). The activity of the D474A and D475A increased significantly at the higher magnesium concentration (Fig. 5). However, no increase in activity was observed with the D399A and D493A mutants (even after incubation at 50 mM Mg 2ϩ for 2 h; data not shown).
We then measured the magnesium dependence of the activity of WT and the D474A and D475A mutants under STO conditions with saturating enzyme (Fig. 6). For WT, the observed rate constant shows a hyperbolic dependence on Mg 2ϩ concentration, yielding k max Mg ϭ 0.12 Ϯ 0.01 s Ϫ1 , K 1/2 Mg ϭ 10 Ϯ 3.3 mM, and n H ϭ 1.3 Ϯ 0.3. High concentrations of Mg 2ϩ inhibited both wild-type PRORP1 and the aspartate mutants (Fig. 6, open symbols), and these points were not included in the analysis of the data. For the mutants, the K 1/2 Mg value was estimated as Ͼ90 mM, and k max Mg was Ն0.007 s Ϫ1 (D474A) and Ն0.002 s Ϫ1 (D475A), indicating that the value of K 1/2 Mg was increased at least 9-fold and the activity was decreased at least 17-fold as compared with WT ( Table 2). The metal-dependent parameters measured for WT PRORP1 under STO conditions (K 1/2 Mg ϭ 10 Ϯ 3.3 mM and k max Mg ϭ 0.12 Ϯ 0.01 s Ϫ1 ) were larger than the comparable values measured under MTO conditions (K m Mg (k cat ) ϭ 1.4 Ϯ 0.3 mM and k cat Mg ϭ 0.06 Ϯ 0.01 s Ϫ1 ); this could potentially be a result of different assay conditions or a change in the rate-limiting step at high Mg 2ϩ concentrations for MTO reactions. Nonetheless, these data indicate that one function of Asp-474 and Asp-475 is to enhance metal ion affinity; however, the decrease in activity suggests that these side chains also increase the reactivity of the metal ions, possi-bly by correct positioning. This proposal is supported by the observation that the D474A and D475A PRORP1 variants catalyze low amounts of miscleaved product (Fig. 5).
pH Dependence-To provide further insight into the mechanism of PRORP-catalyzed phosphodiester bond hydrolysis, we SCHEME 1. Mg 2؉ dependence under k cat /K m conditions. SCHEME 2. Mg 2؉ dependence under k cat conditions.   Table 2.

Enzymatic Mechanism of Protein-only RNase P
measured the pH dependence of STO cleavage with both A. thaliana and B. subtilis pre-tRNA substrates. The k obs values for both substrates show a similar dependence on pH: a loglinear increase with increasing pH that becomes pH-independent under alkaline conditions (Fig. 7A). These data are well described by a single ionization that increases activity with pK a values of 7.9 Ϯ 0.1 (A. thaliana pre-tRNA) and 8.7 Ϯ 0.4 (B. subtilis pre-tRNA), suggesting that deprotonation of a single group enhances activity. The lack of an alkaline limb suggests that no protein side chain acts as a general acid.
In addition to the conserved aspartate residues, the active site of PRORP1 contains two histidine side chains that are located ϳ6 Å from the metal binding sites (Fig. 1B). His-498 is invariant, whereas His-438 is conserved in plants. To examine whether either of these histidine side chains plays a role in PRORP1 catalysis or pH dependence, we mutated His-498 and His-438 to alanine. Under the standard STO assay conditions, the H498A and H438A mutations reduced the observed rate constant to 5.0 Ϯ 0.2 ϫ 10 Ϫ4 and 0.02 Ϯ 0.001 s Ϫ1 , representing an 80-and 2-fold decrease, respectively, as compared with WT (0.04 s Ϫ1 ; Table 1). However, the more conservative mutation of His-498 to glutamine (H498Q) decreased the STO rate constant only 4-fold compared with WT. Furthermore, a single ionization was observed in the pH dependence of the STO cleavage rate constants for both the H498Q and H438A mutants with pK a values of 8.2 Ϯ 0.1 and 7.8 Ϯ 0.1, respectively. These mutations decreased the value of k max pH to 0.03 Ϯ 0.002 and 0.04 Ϯ 0.003 s Ϫ1 , respectively, representing an ϳ3-fold decrease for both as compared with WT (Fig. 7B). These modest alterations in the activity and pK a values for the H438A and H498Q mutants are not consistent with a function as a general acid/base catalyst; mutagenesis of a general acid/base group typically decreases activity by ϳ10 3 -10 4 in hydrolytic reactions (21,22).
To determine whether the side chains of His-498 and His-438 enhance pre-tRNA binding, we measured the affinity of the variants for A. thaliana pre-tRNA using an anisotropy assay (Fig. 8). No significant change in binding affinity was observed for these histidine mutants in comparison with WT with dissociation constants of ϳ700 nM. However, the end point anisotropy, reflecting the anisotropy of the ES complex, for the His-498 mutants varied significantly from the values measured for the WT and H438A enzymes. This result may indicate that the fluorescein and potentially the 5Ј leader adopt a different position and/or conformation when bound to the His-498 mutants as compared with WT.
To examine whether the observed ionization reflects deprotonation of a metal-bound water, we measured the dependence of the activity on the identity of the metal ion. PRORP1 is activated by Mg 2ϩ , Mn 2ϩ , Co 2ϩ , and Ni 2ϩ (but not Ca 2ϩ , Zn 2ϩ , or Cd 2ϩ (7,23)). We measured the rate constant for single turnover cleavage (k max ) at pH 6.5 with saturating concentrations of PRORP1 and metal (Co 2ϩ , Ni 2ϩ , Mn 2ϩ , or Mg 2ϩ ). At pH 6.5, PRORP1 catalysis is within a log-linear region of the pH profile (Fig. 7). Furthermore, this pH is below the pK a values of the hydrated metals used (Co 2ϩ ϭ 9.65, Ni 2ϩ ϭ 10.0, Mn 2ϩ ϭ 10.46, and Mg ϭ 11.4) (24, 25). Therefore, if the observed ionization reflects metal-water deprotonation, the value of k max  Inner sphere with metal II No a Interactions of Asp side chains with bound metal ions are predicted from the crystal structure (see Fig. 1 and Ref. 7). b The K 1/2 Mg values were determined as described in the legend of Fig. 6. should correlate with the concentration of the metal-hydroxide, as indicated by the pK a value. Thus, the k max values are predicted to follow the order Co 2ϩ Ͼ Ni 2ϩ Ͼ Mn 2ϩ Ͼ Mg 2ϩ . The measured k max values for PRORP1 are 0.417, 0.042, 0.004, and 0.021 s Ϫ1 for Co 2ϩ , Ni 2ϩ , Mn 2ϩ , and Mg 2ϩ , respectively, representing a trend of Co 2ϩ Ͼ Ni 2ϩ Ͼ Mg 2ϩ Ͼ Mn 2ϩ . The faster reactivity observed in reactions with Mg 2ϩ versus Mn 2ϩ could be a result of PRORP1 evolving to specifically use Mg 2ϩ ions; thus others factors such as differences in coordination geometries and preferred ligands of the metals used may explain this deviation (26). Excluding the Mg 2ϩ data, log k max decreased linearly with the value of the pK a of the hydrated metal (r ϭ 0.998), consistent with the observed ionization in catalysis arising from a metal-bound water.

Discussion
A Minimal Kinetic Mechanism Reveals That Product Release Is Not Rate-limiting at Neutral pH-We have determined kinetic parameters for PRORP1-catalyzed pre-tRNA cleavage for a cognate A. thaliana substrate and a bacterial pre-tRNA substrate. Despite being from different organisms and domains of life, these substrates show similar kinetic parameters (Table  1). This is not surprising given that both of these tRNAs have conserved secondary and tertiary features (27). Furthermore, the similarity between the STO (k max ) and MTO (k cat ) values suggests that a step observable under STO conditions is ratelimiting (e.g. a step at or before cleavage) and rules out product release as the rate-limiting step under MTO conditions. Consistent with this proposal, we did not observe burst kinetics in the steady state turnover with the substrates used in this study (Fig. 3A). This is in contrast to other RNase P enzymes that have been kinetically characterized. For instance, bacterial and yeast nuclear RNase P (RNA-based enzymes) are both limited by product dissociation and display burst kinetics under MTO conditions (28,29). The standard assay condition (pH 7.8) is within the log-linear region of the pH dependence for PRORP1-catalyzed hydrolysis of B. subtilis pre-tRNA, suggesting that the cleavage step is rate-limiting.
Evidence for a Two-metal Ion Mechanism-Crystal structures of metallonucleases with metals bound in the active site have fuelled mechanistic proposals for one-, two-, and even three-metal ion mechanisms (30,31). In fact, structures of PRORP1 in the presence of different divalent metal ions have revealed one (calcium) to two metal ions (manganese) bound within the active site. The metal dependence of PRORP1-catalyzed phosphodiester bond hydrolysis has a cooperativity of 2 for Mg 2ϩ under k cat /K m conditions for both substrates, indicating that at least two metal ions activate catalysis. Addition of calcium does not abrogate the cooperativity, providing evidence that this effect is not due to stabilization of the pre-tRNA structure. Given these observations, we propose a model where the binding of one Mg 2ϩ ion at the active site enhances the affinity of the second metal ion, leading to the cooperativity observed at low pre-tRNA concentrations (Scheme 1). Furthermore, the loss of cooperativity at saturating pre-tRNA concentration is most easily explained by enhancement of the affinity of one of the two metal ions by substrate. One model consistent with the data is that under k cat conditions an inactive ES complex with one tightly bound Mg 2ϩ ion forms (E Mg S) (Scheme 2). The enhanced metal binding affinity to the ES complex could be the result of either an indirect effect or the direct coordination of Mg 2ϩ to a phosphodiester oxygen from the substrate. Similarly, the bacteriophage T5 Flap endonuclease, a member of the homologous flap nuclease family, shows no cooperativity for Mg 2ϩ under k cat conditions, but a cooperativity of 2 is observed under k cat /K m conditions within the same Mg 2ϩ concentration range tested for PRORP1 (31).
Asp-474 and Asp-475 Enhance the Active Site Metal Ion Affinity-Metal dependence measurements of aspartate-to-alanine mutations identified two side chains that enhance metal ion affinity by at least 9-fold. These mutations also decreased activity a maximum of 17-and 60-fold for the D474A and D475A mutants, respectively. These residues both lie on the end of an active site helix (Fig. 1B). Based on the crystal structure, Asp-475 makes an inner sphere bidentate interaction with both active site metal ions (Fig. 1B). A bidentate Asp ligand is very common among metallonucleases where it has proposed to orient the active site metal ions for optimal catalysis (32). However, in PRORP1, other groups besides Asp-475 must also help orient the metal ions, potentially including interactions with the substrate, because this mutant has significant activity at high concentrations of Mg 2ϩ . The activity of D399A and D493A mutants could not be rescued by high Mg 2ϩ concentrations, suggesting that they serve other roles in addition to metal binding. For instance, Asp-399 makes a hydrogen bond with a metal-bound water that is positioned approximately between the two active site metals (Fig. 1B), suggesting that Asp-399 could position a metal-bound water for nucleophilic attack. In the other nonrescuable aspartate mutant, D493A, the Asp side chain forms a hydrogen bond with the backbone amide of methionine 495 (Fig. 1B); thus mutation of Asp-493 may deleteriously affect the active site architecture.
pH Dependence Reveals a Single Ionization Important for Catalysis-The pH dependence of PRORP1-catalyzed cleavage of pre-tRNA is consistent with ionization of a single group that enhances catalysis. However, for the A. thaliana pre-tRNA the values of both k max pH and pK a were decreased compared with the B. subtilis substrate. There are two models that could account for this. In the first model, the plateau in activity at high pH could result from a change in the rate-limiting step from chemistry at low pH to a pH-independent step, such as a conformational change, at high pH. In this case, the observed pK a does not reflect the thermodynamic ionization but rather a kinetic pK a due to a change in the rate-limiting step over the pH range tested. This mechanism leads to an underestimation of the pK a value as observed previously in the pH dependence of B. subtilis RNase P where the pK a reflects a change in the ratelimiting step (33). Alternatively, the difference in pK a values observed with the A. thaliana and B. subtilis pre-tRNAs could arise from differences in local active site environments in the pre-tRNA⅐PRORP complex, potentially arising from differences in nucleotide composition and/or the conformation of the bound pre-tRNA.
The pH-dependent data for cleavage of B. subtilis pre-tRNA indicate that ionization of a functional group with a pK a of 8.7 or higher is important for catalytic activity. The crystal structure demonstrates that the active site of PRORP1 contains two His residues but no conserved lysine or arginine side chains within 8 Å of the metal binding sites. Mutagenesis data demonstrate that the ionization does not reflect one of the His side chains (His-498 and His-438). Therefore, the most likely ionizable functional group within the active site vicinity with a pK a of 8.7 or higher is a metal-bound water molecule as predicted from the [Mg(H 2 O) 6 ] 2ϩ pK a of 11.4 (24). This proposal is also consistent with the finding that the k max value is dependent on the pK a of the hydrated metal ion used in the reaction. We therefore propose that the observed ionization in PRORP1 is deprotonation of a metal-bound water to form the attacking hydroxide nucleophile. Bacterial RNase P and T5 Flap endonuclease have a similar pH dependence to PRORP1 with a single deprotonation important for catalysis of phosphodi-ester bond cleavage, which is proposed to be deprotonation of a metal-bound water (34,35).
Convergence of Enzyme Mechanism-A comparison of the proposed enzymatic mechanisms of PRORP and RNA-dependent RNase P reveals that both enzymes require catalytic metal ions, although the metal ligands vary (Fig. 9). The metal ions in PRORP1 are coordinated by conserved aspartate residues, whereas the RNA-based RNase P mainly uses non-bridging phosphodiester oxygens for metal ion coordination. Furthermore, the data suggest that one of the scissile phosphate oxygens of pre-tRNA interacts with bound metal ion(s). In PRORP, the pro-S p oxygen of pre-tRNA is hypothesized to interact with the active site metal ions (23,36), although this needs to be confirmed experimentally, whereas the pro-R p substrate oxygen coordinates a metal ion in RNA-based RNase P (37,38). Modeling of bound substrate suggests that coordination of the pro-S p or pro-R p oxygen affects the position of the metal ion relative to the scissile phosphate, leading to an alteration in the pre-tRNA binding mode. Other protein-based phosphoryl transfer enzymes also contact the pro-S p oxygen of the scissile phosphate, including RNase H, EcoRV, and DNA transposases (39 -41). The hammerhead ribozyme and spliceosome are proposed to contact the pro-R p oxygen at the cleavage site, whereas the Tetrahymena ribozyme contacts the pro-S p oxygen (42)(43)(44).
The RNA-based RNase P reaction has been proposed to catalyze cleavage using a metal-bound hydroxide nucleophile (34). The pH dependence of PRORP1 suggests a similar mechanism for the protein-catalyzed reaction. A possible alternate mechanism is activation of a metal-bound nucleophile by a general base as suggested for EcoRV and RNase H (45,46). However, two ionizations (slope ϭ 2) and pK a values of ϳ6.0 -7.0 are observed in pH profiles of these enzymes (45)(46)(47), which is significantly different from the pH dependence observed with PRORP1. Thus, it is unlikely that PRORP1 uses a general base mechanism for activation of the nucleophile. In both cases, the metal ions are proposed to stabilize the negative charge buildup in the transition state and facilitate bond cleavage by stabilizing the 3Ј-oxyanion leaving group. The absence of an alkaline limb FIGURE 9. Proposed enzymatic mechanisms for protein-only (A) and RNA-based (B) RNase P enzymes. Both protein and RNA-based RNase P enzymes use at least two metal ions for phosphodiester bond hydrolysis. The metal ions in PRORP1 are positioned through conserved aspartate residues, whereas the RNA-based RNase P coordinates metal ions mainly using non-bridging phosphate oxygens. Metal I (M1) is proposed to coordinate the nucleophilic water molecule and aid in transition state stabilization by interaction with a non-bridging phosphate oxygen. Metal II (M2) is also proposed to stabilize the transition state and potentially coordinates the 3Ј-oxyanion leaving group to facilitate bond cleavage. on the pH profile for PRORP1 and RNA-based RNase P (48) suggests that an amino acid or nucleotide with a pK a Ͻ10 does not serve as a general acid. Rather, the proton donor for the 3Ј-oxyanion leaving group is most likely a water molecule. The functional convergence of enzymes is not uncommon (49); however, the functional and mechanistic convergence of two different enzymatic macromolecules (RNA and protein) is remarkable. This observation suggests that the protein and RNA macromolecules act as dynamic scaffolds that achieve large rate enhancements mainly through the binding and positioning of hydrated magnesium ions.