The HD domain of the Escherichia coli tRNA nucleotidyltransferase has 2',3'-cyclic phosphodiesterase, 2'-nucleotidase, and phosphatase activities.

In all mature tRNAs, the 3'-terminal CCA sequence is synthesized or repaired by a template-independent nucleotidyltransferase (ATP(CTP):tRNA nucleotidyltransferase; EC 2.7.7.25). The Escherichia coli enzyme comprises two domains: an N-terminal domain containing the nucleotidyltransferase activity and an uncharacterized C-terminal HD domain. The HD motif defines a superfamily of metal-dependent phosphohydrolases that includes a variety of uncharacterized proteins and domains associated with nucleotidyltransferases and helicases from bacteria, archaea, and eukaryotes. The C-terminal HD domain in E. coli tRNA nucleotidyltransferase demonstrated Ni(2+)-dependent phosphatase activity toward pyrophosphate, canonical 5'-nucleoside tri- and diphosphates, NADP, and 2'-AMP. Assays with phosphodiesterase substrates revealed surprising metal-independent phosphodiesterase activity toward 2',3'-cAMP, -cGMP, and -cCMP. Without metal or in the presence of Mg(2+), the tRNA nucleotidyltransferase hydrolyzed 2',3'-cyclic substrates with the formation of 2'-nucleotides, whereas in the presence of Ni(2+), the protein also produced some 3'-nucleotides. Mutations at the conserved His-255 and Asp-256 residues comprising the C-terminal HD domain of this protein inactivated both phosphodiesterase and phosphatase activities, indicating that these activities are associated with the HD domain. Low concentrations of the E. coli tRNA (10 nm) had a strong inhibiting effect on both phosphatase and phosphodiesterase activities. The competitive character of inhibition by tRNA suggests that it might be a natural substrate for these activities. This inhibition was completely abolished by the addition of Mg(2+), Mn(2+), or Ca(2+), but not Ni(2+). The data suggest that the phosphohydrolase activities of the HD domain of the E. coli tRNA nucleotidyltransferase are involved in the repair of the 3'-CCA end of tRNA.

In all organisms, the 3Ј-terminal CCA sequence in all mature tRNAs is required for tRNA aminoacylation reaction and for translation on the ribosome (1,2). The CCA-adding enzyme (ATP(CTP):tRNA nucleotidyltransferase; EC 2.7.7.25) adds these three nucleotides in the order of C, C, and A to the tRNA nucleotide-73, using CTP and ATP as substrates and producing inorganic pyrophosphate (3). tRNA nucleotidyltransferases (tRNA-NTs) 1 are highly conserved throughout evolution, and they have been identified in all three kingdoms (4). The CCAadding enzyme is essential in eukaryotes, archaea, and certain eubacteria where some or all tRNA genes do not encode the CCA sequence (5). In these organisms, 3Ј trailer sequences are removed from tRNA precursors by nucleases that stop at the discriminator base (position 73), leaving the acceptor stem intact as a substrate for the CCA-adding enzyme (6). In organisms in which tRNA genes encode the CCA sequence (Escherichia coli and many eubacteria), tRNA-NT repairs the tRNA 3Ј CCA sequence depleted by exonucleases, and inactivation of the tRNA-NT gene impairs cell growth (7).
The CCA-adding enzyme is unique among nucleotidyltransferases in that it adds an ordered nucleotide sequence to a specific primer without using a nucleic acid template (6,8). Moreover, the enzyme can monitor and faithfully rebuild tRNAs with incomplete 3Ј ends (tRNA-NCC, tRNA-NC, and tRNA-N, where N is the discriminator base). It completes the terminal CCA sequence by adding one nucleotide at a time. The protein consists of a single polypeptide with dual specificity for AMP and CMP incorporation, and a single active site is responsible for addition of both nucleotides (4). Several models have been proposed to explain these properties (9 -14), but the molecular details remain unclear.
According to their primary sequence, the CCA-adding enzymes belong to an ancient superfamily of nucleotidyltransferases, which is divided into two classes (6,15). Class I contains archaeal CCA-adding enzymes, whereas class II comprises eubacterial and eukaryotic proteins. Class II enzymes have similar sizes (400 -470 amino acids) and share a highly conserved 25-kDa N-terminal region with characteristic active-site signature motifs (DxD and RRD), but differ from one another in their C-terminal regions. The E. coli tRNA-NT belongs to the class II enzymes, and its nucleotidyltransferase activity has been characterized (16,17). This protein has two domains, the N-terminal nucleotidyltransferase domain and * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.This work was supported financially by Genome Canada, the Ontario Research and Development Challenge Fund, and the Protein Structure Initiative of the National Institutes of Health (GM62414 and GM62413). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed. the C-terminal HD domain named after the conserved HD motif (His-255 and Asp-256 in E. coli). The HD motif defines a superfamily of metal-dependent phosphohydrolases that includes a variety of uncharacterized proteins and domains associated with nucleotidyltransferases and helicases from bacteria, archaea, and eukaryotes (18). The HD superfamily is related to a well characterized eukaryotic 3Ј,5Ј-cyclic phosphodiesterase (19) that is, however, distinct from the HD superfamily and contains additional conserved regions. The HD domain was predicted to have a metal-dependent phosphohydrolase activity (18).
Here, we report that the E. coli tRNA-NT has 2Ј,3Ј-cyclic phosphodiesterase, 2Ј-nucleotidase, and phosphatase activities. Site-directed mutagenesis demonstrated that these three activities are associated with the conserved C-terminal HD motif and probably have partially overlapping active sites. Low concentrations of total E. coli tRNA produced strong competitive inhibition on all three activities, suggesting that tRNA might be a natural substrate for the HD domain-associated activities. Based on our results, we proposed a model for the role of the HD domain-associated activities in the repair of the tRNA 3Ј end.

MATERIALS AND METHODS
Gene Cloning and Protein Purification-Recombinant proteins were expressed with a His 6 fusion tag at the N terminus using a modified pET15b (Novagen) expression vector in E. coli BL21 (DE3) as described previously (20). Purification of the E. coli tRNA nucleotidyltransferase (from 4 L E. coli cultures) was performed using a combination of affinity chromatography on nickel-nitrilotriacetic acid resin (Qiagen) and anion-exchange chromatography on Q-Sepharose (Amersham Biosciences; elution by NaCl gradient 50 -200 mM).
Enzymatic Assays-The E. coli tRNA-NT was screened for a variety of catalytic activities as follows. Reactions were performed in 96-well microplates using 0.2-ml reaction mixtures and the following spectrophotometric procedures. After addition of protein samples (0.1-1.0 g), the plates were incubated for 1-3 h at 37°C before taking the reading at indicated wavelengths. Positive controls with small amount of commercial enzymes were included in all tests. Phosphatase activity was measured in a reaction mixture containing 50 mM HEPES-K buffer, pH 7.5, 5 mM MgCl 2 , 0.1 mM MnCl 2 , and 40 mM p-nitrophenyl phosphate (pNPP) by following the absorption increase at 410 nm. Phosphodiesterase and nuclease activities were measured using 50 mM Tricine buffer, pH 8.5, 5 mM MgCl 2 , 0.1 mM MnCl 2 , and 0.83 mM bis-p-nitrophenyl phosphate (bis-pNPP) with detection at 410 nm. Esterase activity was tested using 1 mM p-nitrophenyl palmitate in 50 mM Tris-HCl, pH 8.0, 0.4% Triton X-100, and 0.1% of Gum Arabic (detection at 410 nm). Protease activity was determined with a mixture of two chromogenic substrates (benzoyl-Arg-p-nitroanilide and Leu-p-nitroanilide; 0.2 mM each) was measured in 50 mM HEPES-K buffer, pH 7.5, 0.5 mM Ca 2ϩ , 0.5 mM Zn 2ϩ , and 1 mM dithiothreitol (detection at 405 nm). Dehydrogenase and oxidase activities were measured using three pools of different substrates: 20 amino acids (0.2 mM concentrations of each in reaction mixture), eight different alcohols (methanol, ethanol, adonitol, xylitol, dulcitol, mannitol, sorbitol, arabitol; 0.5 mM concentrations of each in reaction mixture), or nine different organic acids (acetate, fumarate, malate, lactate, pyruvate, isocitrate, succinate, oxaloacetate, ␣-ketoglutarate; 0.5 mM concentrations of each in reaction mixture). Dehydrogenase activity was measured in a reaction mixture containing 50 mM Tricine buffer, pH 8.5, 0.5 mM NAD, 0.5 mM NADP, the substrate solution (described above), 1 mM Mg 2ϩ , and 0.1 mM Mn 2ϩ (detection at 340 nm). Oxidase activity was monitored using the chromogenic reaction of peroxidase and o-dianisidine (21) in a reaction mixture containing 50 mM HEPES-K, pH 8.0, the substrate solution (described above), 0.1 mM o-dianisidine, and 2 g of peroxidase (detection at 460 nm).
Screening with natural phosphatase substrates was performed in 96-well plates using 160-l reaction mixtures containing 50 mM HEPES-K, pH 7.5, 0.1 mM substrate (various nucleotides and phosphorylated sugars, amino acids, and organic acids from Sigma), 6.25 mM Mg 2ϩ , 0.6 mM Mn 2ϩ , 0.6 mM Ni 2ϩ , and 2-5 g of protein. After a 30 -60 min incubation at 37°C, the reaction was terminated by the addition of 40 l of Malachite Green reagent (22), and the production of P i was measured at 630 nm.
Phosphodiesterase activity of the E. coli tRNA-NT against bis-pNPP was measured in a reaction mixture containing 50 mM HEPES-K, pH 7.0, 3.3 mM bis-pNPP, with or without 0.5 mM Ni 2ϩ . After a 30-min incubation at 37°C, the reaction was stopped by the addition of 0.2 ml of 6 N NaOH and produced p-nitrophenol was determined from the absorbance at 410 nm. Phosphodiesterase activity with natural substrates (various 2Ј,3Ј-and 3Ј,5Ј-cyclic nucleoside monophosphates from Sigma) was measured using a quantitative assay based on a measurement of the alkaline phosphatase-sensitive nucleotide product. All incubations (0.4 ml) contained 50 mM HEPES-K, pH 7.0, 1-2 mM substrate, and metal (if indicated). After a 20-min incubation at 37°C with enzyme (0.2-0.3 g), the reaction was stopped by the addition of 0.4 ml of 2ϫ alkaline phosphatase buffer (0.2 M CHES buffer, pH 9.0, and 10 mM MgCl 2 ) and incubated with 1 unit of alkaline phosphatase for 10 min at 37°C. Liberated P i was assayed with Malachite Green reagent (22).
For determination of the K m and V max , the phosphatase and phosphodiesterase assays contained substrates at concentrations of 0.1-10 mM. Kinetic parameters were determined by non-linear curve fitting from the Lineweaver-Burk plot using GraphPad Prism software (version 4.00 for Windows, GraphPad Software, San Diego, CA).
TLC Analysis-Products of enzymatic digestions, performed as described above, were analyzed by TLC on cellulose plates in solvent A (saturated ammonium sulfate/3 M sodium acetate/isopropyl alcohol; 80:6:2) (23). The reaction products and nucleotide standards were visu-  Site-directed mutagenesis was carried out using QuikChange (Stratagene) and mutagenic primers designed to replace the selected amino acid residues by Ala. Wild-type E. coli tRNA-NT gene cloned into pET15b was used as a template for mutagenesis. Plasmids were purified using Qiaprep Spin Mini Prep Kit (Qiagen), and all mutations were verified by DNA sequencing.

RESULTS
tRNA Nucleotidyltransferase Has Phosphatase Activity-General enzymatic screens with pNPP as a substrate identi-fied a new activity in the E. coli tRNA nucleotidyltransferase (tRNA-NT, CCA adding enzyme) (data not shown). The protein was purified in a large scale by a combination of nickelchelate affinity and anion-exchange chromatography to Ͼ95% homogeneity. In solution, most of purified tRNA-NT existed as a monomer (44.8 Ϯ 0.5 kDa) with low levels (up to 20%) of dimers (103.0 Ϯ 10.0 kDa) and oligomers (784.2 Ϯ 33.5 kDa). All three fractions (monomers, dimers, and oligomers) had approximately the same level of phosphatase activity with pNPP. In contrast to the characterized nucleotidyltransferase activity of this protein (16,17), the phosphatase activity had a neutral pH optimum, pH 7.0, and was greatly stimulated by Ni 2ϩ (Fig. 1). With pNPP as a substrate, tRNA-NT showed high affinity to Ni 2ϩ (K D ϭ 6.0 Ϯ 0.7 M in the presence of 8 mM pNPP) and low affinity to pNPP (K m ϭ 6.2 Ϯ 0.46 mM in the presence of 50 M Ni 2ϩ ) (Fig. 1). However, the specific activity toward pNPP (V max ϭ 12.4 Ϯ 0.34 mol/min/mg of protein) was sufficiently high to produce positive results in general screens without Ni 2ϩ .
Phosphatase Activity with Natural Substrates-Secondary catalytic screens with natural phosphatase substrates (57 compounds) in the presence of Ni 2ϩ identified high phosphatase activity of the E. coli tRNA-NT against a broad range of canonical nucleoside di-and triphosphates, NADP, and pyrophosphate (Fig. 2). Highest phosphatase activity was found against NADP, ATP, CTP, TTP, and dCTP. Nucleoside diphosphates (ADP, CDP, GDP, and UDP) were also good substrates for the E. coli tRNA-NT (data not shown). With all substrates, the protein showed saturation kinetics with highest affinity to pyrophosphate and NADP ( Fig. 3 and Table I). Cellulose TLC analysis of the hydrolysis of ATP identified ADP and AMP as reaction products (Fig. 4A), indicating that the E. coli tRNA-NT sequentially removes monophosphates from ATP. Enzymatic analysis of the reaction products produced during hydrolysis of ATP or ADP yielded P i , AMP, and ADP as the reaction products (data not shown). With nucleoside monophosphates as substrates, high phosphatase activity was observed against 2Ј-AMP (Figs. 2 and 3), low activity against 3Ј-AMP (12.9 nmol/ min/mg of protein), and no activity toward 5Ј-mononucleotides.
As with pNPP, hydrolysis of natural substrates had a neutral pH optimum, pH 7.0, and was greatly stimulated by Ni 2ϩ (Fig.  5). With natural substrates, the protein had even higher affinity for Ni 2ϩ (K D ϭ 1.06 Ϯ 0.1 M with pyrophosphate as a substrate). Cu 2ϩ stimulated the hydrolysis of pyrophosphate and ATP and completely inhibited the hydrolysis of 2Ј-AMP, suggesting some difference in the active sites involved in these reactions. Because the nucleotidyltransferase activity of the E. coli tRNA-NT (which also involves the hydrolysis of ATP and CTP) had an alkaline pH optimum (pH 9.0 -9.5) and required Mg 2ϩ (16,17), it is not related to the Ni 2ϩ -dependent neutral phosphatase and 2Ј-nucleotidase activities observed in our experiments.
We also determined the kinetic parameters for the hydrolysis of several natural substrates (Fig. 3, Table I). With all substrates, the protein showed saturation kinetics with highest affinity for pyrophosphate and NADP ( Fig. 3; Table I). For the nucleotidyltransferase reaction, 1 mM PP i caused only 20% inhibition (16). This insensitivity to PP i can be explained by the presence of pyrophosphatase activity in the E. coli tRNA-NT. Whereas the K m for ATP (0.18 mM) was near the K m for AMP incorporation (0.33 mM) (17), the K m for CTP (0.13 mM) was significantly higher than the K m for CMP incorporation (0.03 mM) (17). The protein showed much higher activity and affinity to NADP (which also has a 2Ј-phosphate) than to 2Ј-AMP (Table I), suggesting that it might prefer large substrates like tRNAs.  tRNA-NT Has Phosphodiesterase Activity-Because a number of phosphodiesterases have been shown to possess both nucleotidase and phosphodiesterase activities (24,25), the E. coli tRNA-NT was tested for the presence of phosphodiesterase activity toward 2Ј,3Ј-and 3Ј,5Ј-cyclic nucleotides. The protein exhibited high phosphodiesterase activity against 2Ј,3Ј-cAMP and 2Ј,3Ј-cGMP, low activity against 2Ј,3Ј-cCMP, and no activity with 3Ј,5Ј-cyclic nucleotides (Fig. 3). Like most known 2Ј,3Јcyclic phosphodiesterases, the phosphodiesterase activity of the E. coli tRNA-NT had a neutral pH optimum, pH 7.0, and no metal dependence (Fig. 5D). Similar levels of phosphodiesterase activity were observed without metal addition or in the presence of Mg 2ϩ , Mn 2ϩ , Ca 2ϩ , or Ni 2ϩ , whereas Co 2ϩ , Cu 2ϩ , and Zn 2ϩ were inhibiting (Fig. 5D). The protein showed highest activity and affinity to 2Ј,3Ј-cAMP (Table I).
When tested with the general phosphodiesterase substrate bis-pNPP, tRNA-NT had maximal activity at pH 7.0 and reduced activity (25%) at pH 8.5 (used in screens), and it showed a strong requirement for Ni 2ϩ (little activity with Mn 2ϩ and no activity without metal or in the presence of Mg 2ϩ ). This strong dependence of the bis-pNPP hydrolysis by the E. coli tRNA-NT  on Ni 2ϩ explains why the general phosphodiesterase screens failed to detect this activity. In addition, our results demonstrated a different metal requirement for the hydrolysis of natural (2Ј,3Ј-cAMP) and artificial (bis-pNPP) substrates by the E. coli tRNA-NT. Without metal or in the presence of Mg 2ϩ , the E. coli tRNA-NT hydrolyzed 2Ј,3Ј-cAMP to 2Ј-AMP as analyzed by cellulose TLC (Fig. 4B). However, in the presence of Ni 2ϩ , the protein produced both 2Ј-AMP and 3Ј-AMP. Because Ni 2ϩ stimulates 2Ј-nucleotidase activity of the E. coli tRNA-NT (Fig. 5C), in the presence of this metal, a significant proportion of 2Ј-AMP product was converted to adenosine (Fig. 4C). As presented on Fig. 4B, in the presence of Ni 2ϩ , 2Ј-AMP was still the main product (ϳ70%) of 2Ј,3Ј-cAMP hydrolysis. Similar results were obtained with 2Ј,3Ј-cGMP (Fig. 4C) and 2Ј,3Ј-cCMP (data not shown) as substrates. Other metals (Mg 2ϩ , Mn 2ϩ , Ca 2ϩ ) and the E. coli tRNA did not affect the hydrolysis of 2Ј,3Ј-cAMP (data not shown). Thus, the E. coli tRNA-NT has 2Ј,3Ј-cyclic nucleotide 3Ј-phosphodiesterase activity in the absence of Ni 2ϩ and 2Ј,3Ј-cyclic nucleotide 2Ј,3Ј-phosphodiesterase activity in the presence of Ni 2ϩ . At the moment, two groups of 2Ј,3Ј-cyclic phosphodiesterases are known (BRENDA data base). Members of the first group (EC 3.1.4.16; mostly prokaryotic proteins) convert a 2Ј,3Ј-cyclic nucleotide exclusively to a 3Ј-nucleotide, which is further hydrolyzed to a ribonucleotide and P i . The second group (EC 3.1.4.37; mostly eukaryotic proteins) hydrolyzes the same substrate exclusively to a 2Ј-nucleotide and then degrades it to a ribonucleotide and P i . Thus, the E. coli tRNA-NT is the first 2Ј,3Ј-cyclic phosphodiesterase that can hydrolyze a 2Ј,3Ј-cyclic phosphodiester at both 2Ј-and 3Ј-bonds and produce the corresponding 2Ј-and 3Ј-nucleotides.
Effect of tRNA and Metals-In the absence of metal or in the presence of Ni 2ϩ , low concentrations of total tRNA from E. coli (10 nM) had a strong inhibiting effect on both phosphatase and phosphodiesterase activities of the E. coli tRNA-NT (Fig. 6A). The activities toward natural substrates were especially sensitive to tRNA inhibition (K i ϭ 0.29 nM for ATP hydrolysis, 1.08 nM for 2Ј-AMP hydrolysis, and 1.68 nM for 2Ј,3Ј-cAMP hydrolysis); the inhibition of pNPP hydrolysis by 30 nM tRNA was minimal (30%) (Fig. 6A). These inhibition constants are at least 10,000 times lower than the K m for tRNA in nucleotidyltransferase reaction (15-20 M) (17). As shown on Fig. 6 (B and C), the E. coli tRNA acted as a competitive inhibitor of the hydrolysis of both 2Ј-AMP and 2Ј,3Ј-cAMP and induced a significant increase in K m for these substrates (Table II). The inhibition by tRNA was released at high substrate concentrations (Fig. 6, B and C), suggesting that tRNA might be a natural substrate for new phosphohydrolase activities. Total tRNA from yeast had similar inhibiting effects (data not shown). Moreover, we found that the phosphohydrolase activities of the E. coli tRNA-NT were also inhibited in a dose-dependent manner by nanogram quantities of the E. coli total rRNA (with 50% reduction caused by 8.86 -22.2 ng of rRNA), genomic DNA (8.95-17.9 ng), or synthetic oligonucleotides poly(A) (36.3-97.6 ng) or poly(C) (14.4 -19.1 ng), whereas single nucleotides (like 5Ј-AMP) had weak inhibiting effect (K i ϭ 0.59 mM). This sensitivity to various nucleic acids can be explained by broad substrate specificity of tRNA-NTs that can efficiently recognize all tRNAs regardless of amino acid acceptor specificity, viral RNAs, U2 small nuclear RNAs, synthetic tDNA, and DNA oligonucleotides (10, 26 -29).
The inhibitory effect of E. coli tRNA on the tRNA-NT phosphohydrolase activities was suppressed in a concentration-dependent manner by the addition of Mg 2ϩ , Mn 2ϩ , or Ca 2ϩ ( Fig.  7; data shown for Mg 2ϩ ). The stimulatory effect of Mg 2ϩ on ATP hydrolysis in the presence of tRNA was observed at a very narrow range of metal concentrations (0.1-1.0 mM), which changed to strong inhibition at Mg 2ϩ concentrations Ͼ1 mM (Fig. 7A). The hydrolysis of 2Ј-AMP (in the presence of tRNA) was stimulated by Mg 2ϩ up to 2.5 mM with significant inhibition (40%) observed at 10 mM Mg 2ϩ , whereas no inhibition by Mg 2ϩ (up to 10 mM) was found for the hydrolysis of 2Ј,3Ј-cAMP (Fig. 7, B and C). In these experiments, the affinity to Mg 2ϩ was highest in the hydrolysis of ATP (apparent K D , ϳ0.1 mM), followed by 2Ј-AMP (K D ϭ 0.24 Ϯ 0.1 mM) and 2Ј,3Ј-cAMP (K D ϭ 0.49 Ϯ 0.1 mM). With 2Ј-AMP as a substrate, it was found that simultaneous presence of Mg 2ϩ and tRNA resulted in a 2-fold time increase in V max and K m (data not shown). In control experiments without tRNA, all three metals (Mg 2ϩ , Mn 2ϩ , Ca 2ϩ ; 0.01-1 mM) produced small stimulating effects (10 -40%) on both phosphatase and phosphodiesterase activi-  ties of tRNA-NT (data not shown). In conclusion, the three new phosphohydrolase activities of the E. coli tRNA-NT differ in their sensitivities to tRNA and Mg 2ϩ , suggesting that the corresponding active sites are different.
Site-directed Mutagenesis of tRNA-NT-Analysis of the available sequences of the CCA-adding enzymes (at least 84 entries in EMBL-EBI data base) revealed that at least 40 proteins contain the conserved HD motif located at the C terminus (His-255 and Asp-256 in the E. coli tRNA-NT). All HD motif-containing tRNA-NTs are found in eubacteria, mostly to the ␥-subdivision of proteobacteria, and they comprise a highly homologous group (over 50% identity for proteins from 20 or- ganisms) within the class II group (Fig. 8). There are four widely separated conserved His residues (His-231, -255, -270, and -305 in the E. coli protein), the second of which is part of an invariant dipeptide His-Asp (Fig. 8). Other conserved residues of HD domain include Pro-268, Arg-287, Pro-291, Ala-298, Asp-327, Arg-330, Arg-334, Asp-345, Arg-349, Tyr-357, Pro-358, Gln-359 (numbers for the E. coli tRNA-NT). Using mutagenizing primers, the residues comprising the conserved HD motif were replaced by Ala, creating the H255A and D256A mutant proteins. The proteins were overexpressed and purified using the same purification protocol as for the wild-type protein (see "Materials and Methods"). Both mutant proteins were folded but showed no phosphatase or phosphodiesterase activities indicating that His-255 and Asp-256 are required for both these activities. There is one more HD pair in the C-terminal domain of the E. coli tRNA-NT, His-305 and Asp-306, with His-305 being a conserved residue in the HD domain-containing tRNA-NT (Fig. 8). The H305A mutant also showed no phosphatase or phosphodiesterase activity (Table II), whereas the D306A strain was active (data not shown). Mutations at the conserved nucleotidyltransferase motif DxD (D21A and D23A) had no effect on either phosphatase or phosphodiesterase activities (Table II). These results indicate that the conserved HD motif (H255 and D256), as well as the conserved His-305, are required for both phosphatase and phosphodiesterase activities of the E. coli tRNA-NT and that they are critical for catalysis but not for substrate binding. DISCUSSION With use of general enzymatic screens and detailed biochemical analysis, we revealed the presence of at least three new activities in the E. coli tRNA-NT: a Ni 2ϩ -dependent phosphatase, a 2Ј-nucleotidase, and a metal-independent 2Ј,3Ј-cyclic phosphodiesterase. These activities have not been reported previously in tRNA nucleotidyltransferases. Site-directed mutagenesis showed that all these activities are associated with the conserved HD domain located at the C terminus of the E. coli tRNA-NT. Moreover, we found neither phosphodiesterase nor phosphatase activities in the purified CCA-adding enzyme from Methanobacterium thermoautotrophicum, which has no HD domain. 2 Thus, the new activities observed in the tRNA-NT from E. coli are associated with the conserved HD domain.
To date, only two HD domain-containing proteins have been characterized biochemically, the E. coli dGTPase and RelA/ SpoT homologues from E. coli and Streptococcus equisimilis. The E. coli dGTPase (EC 3.1.5.1) catalyzes the Mg 2ϩ -dependent hydrolysis of dGTP to deoxyguanosine and tripolyphosphate (30,31). The major role of the E. coli SpoT (EC 3.1.7.2) is the breakdown of (p)ppGpp by a Mn 2ϩ -dependent (p)ppGpp pyrophosphohydrolase activity (32,33). The E. coli RelA protein, which is a homolog of SpoT, contains substitutions in the HD domain and has no phosphohydrolase activity (34). The S. equisimilis Rel/Spo enzyme, a bifunctional (p)ppGpp synthetase/hydrolase, catalyzes both synthesis and Mn 2ϩ -dependent hydrolysis of (p)ppGpp (35). In our work, we identified the presence of 2Ј,3Ј-cyclic phosphodiesterase, 2Ј-nucleotidase, and phosphatase activities associated with the HD domain in the E. coli tRNA-NT. Thus, the HD domain is a broad substrate range phosphohydrolase that can catalyze both metal-dependent and -independent phosphomonoesterase and phosphodiesterase reactions.
The HD domain-associated activities of the E. coli tRNA-NT showed different metal dependence and sensitivity to Mg 2ϩ / tRNA, suggesting that although overlapping on the HD motif, these activities also use various residues for catalysis. To date, there is only one three-dimensional structure of the HD-domain containing protein, the N-terminal fragment of the bifunctional Rel/Spo homologue from Streptococcus dysgalactiae (36). All residues of this protein involved in the coordination of Mn 2ϩ (His-53, His-77, Asp-78, and Asp-144) are conserved in the HD domain of the E. coli tRNA-NT (His-231, His-255, Asp-256, and Asp-327) suggesting that in the latter, they also can be involved in metal coordination. The two available threedimensional structures of 2Ј,3Ј-cyclic phosphodiesterases (from Arabidopsis thaliana and rat) identified two histidine residues acting without metal to directly activate a water molecule for the nucleophilic attack of a phosphodiester bond (37,38). These conserved histidines together with two conserved hydroxyamino acids (Ser or Thr) make a signature motif for the 2H subfamily of 2Ј,3Ј-cyclic phosphodiesterases (39). Because this motif is not present in the E. coli tRNA-NT, this protein evidently uses a different mechanism for the hydrolysis of the 2Ј,3Ј-cyclic phosphodiester bond.
The nucleotidyltransferase domain of the E. coli tRNA-NT binds various tRNAs and needs Mg 2ϩ for catalysis (16,40). Two Asp residues located near the N terminus are involved in Mg 2ϩ coordination and are absolutely required for catalysis (4,8). Strong competitive inhibition of mononucleotide hydrolysis by tRNA (Fig. 6) suggests that tRNA might be a natural substrate for the HD domain-associated activities. Mg 2ϩ itself had small effects on both phosphodiesterase and phosphatase activities, but its addition completely abolished the inhibiting effect of tRNA. Although the mechanism of the protection by Mg 2ϩ is unclear at this stage, it indicates that the binding of Mg 2ϩ to the nucleotidyltransferase domain can be conveyed in some manner to the HD domain.
The physical association of the HD domain with the E. coli tRNA-NT suggests that the phosphohydrolase activities might be involved in the repair of the tRNA 3Ј-end. In E. coli and many other bacteria in which tRNA genes encode the CCA sequence, the role of the CCA-adding enzyme is to maintain and repair the tRNA 3Ј-CCA terminus degraded by intracellular RNases (41,42). Evidence indicates that bacterial and mammalian ribonucleases produce RNAs containing 2Ј,3Ј-cyclic phosphodiesters at the 3Ј end as true products of the reaction (43-47). The 3Ј-terminal 2Ј,3Ј-cyclic phosphate in the E. coli tRNAs can also be produced by the activity of 3Ј-terminal phosphate cyclase, which catalyzes the ATP-dependent con-version of the 3Ј-terminal phosphate of tRNA to a 2Ј,3Ј-cyclic phosphodiester (23). Thus, on the basis of available information and our biochemical data, we can propose the following model for the role of the HD domain in the E. coli tRNA-NT (Fig. 9). Our model is based on the assumption that the degradation of tRNAs by intracellular RNases produces tRNA molecules with 2Ј,3Ј-cyclic phosphate at the 3Ј-end. In the repair process, the phosphodiesterase activity of the HD domain hydrolyzes the cyclic phosphodiester to 2Ј-monophosphate, which can then be removed by the HD domain 2Ј-nucleotidase activity (Fig. 9). These activities will eventually produce the unphosphorylated 3Ј-end of tRNA suitable for the nucleotidyltransferase reaction. It is interesting that in at least seven archaeal genomes (Sulfolobus, Aeropyrum, Archaeoglobus, Pyrococcus, Methanobacterium, Methanopyrus), the tRNA-NT gene co-occurs in the same operon with a predicted 2Ј,3Ј-cyclic phosphodiesterase (39). This strong operonic association suggests that in archaea, these predicted phosphodiesterases can also be involved in the process of CCA addition to tRNA, similar to the HD domain in the E. coli tRNA-NT.
The nucleotidyltransferase reaction is accompanied by the release of inorganic pyrophosphate. The equilibrium of this reaction is shifted in the direction of polymerization if pyrophosphate is removed. Pyrophosphate hydrolysis coupled with polymerization is thought to be catalyzed by a phosphatase that has not yet been definitely identified in any system and was predicted to be performed by the HD domain in the E. coli tRNA-NT (18). Our observation of high pyrophosphatase activity of the E. coli tRNA-NT provided the first experimental proof of this hypothesis.
Thus, the E. coli tRNA-NT is a multifunctional enzyme with 2Ј,3Ј-cyclic phosphodiesterase, 2Ј-nucleotidase, phosphatase, and nucleotidyltransferase activities that we suggest act in concert to repair the 3Ј end of tRNA. Further structural and genetic studies are required to prove this model for the HDdomain associated activities of the E. coli tRNA-NT discovered by general enzymatic screens.