Portal de Periódicos da CAPES

Structure-specific tRNA Determinants for Editing a Mischarged Amino Acid

2003; Elsevier BV; Volume: 278; Issue: 46 Linguagem: Inglês

10.1074/jbc.m307080200

1083-351X

Kirk Beebe, Eve Merriman, Paul Schimmel,

RNA Research and Splicing

Alanyl-tRNA synthetase efficiently aminoacylates tRNAAla and an RNA minihelix that comprises just one domain of the two-domain L-shaped tRNA structure. It also clears mischarged tRNAAla using a specialized domain in its C-terminal half. In contrast to full-length tRNAAla, minihelixAla was robustly mischarged and could not be edited. Addition in trans of the missing anticodon-containing domain did not activate editing of mischarged minihelixAla. To understand these differences between minihelixAla and tRNAAla, several chimeric full tRNAs were constructed. These had the acceptor stem of a non-cognate tRNA replaced with the stem of tRNAAla. The chimeric tRNAs collectively introduced multiple sequence changes in all parts but the acceptor stem. However, although the acceptor stem in isolation (as the minihelix) lacked determinants for editing, alanyl-tRNA synthetase effectively cleared a mischarged amino acid from each chimeric tRNA. Thus, a covalently continuous two-domain structure per se, not sequence, is a major determinant for clearance of errors of aminoacylation by alanyl-tRNA synthetase. Because errors of aminoacylation are known to be deleterious to cell growth, structure-specific determinants constitute a powerful selective pressure to retain the format of the two-domain L-shaped tRNA. Alanyl-tRNA synthetase efficiently aminoacylates tRNAAla and an RNA minihelix that comprises just one domain of the two-domain L-shaped tRNA structure. It also clears mischarged tRNAAla using a specialized domain in its C-terminal half. In contrast to full-length tRNAAla, minihelixAla was robustly mischarged and could not be edited. Addition in trans of the missing anticodon-containing domain did not activate editing of mischarged minihelixAla. To understand these differences between minihelixAla and tRNAAla, several chimeric full tRNAs were constructed. These had the acceptor stem of a non-cognate tRNA replaced with the stem of tRNAAla. The chimeric tRNAs collectively introduced multiple sequence changes in all parts but the acceptor stem. However, although the acceptor stem in isolation (as the minihelix) lacked determinants for editing, alanyl-tRNA synthetase effectively cleared a mischarged amino acid from each chimeric tRNA. Thus, a covalently continuous two-domain structure per se, not sequence, is a major determinant for clearance of errors of aminoacylation by alanyl-tRNA synthetase. Because errors of aminoacylation are known to be deleterious to cell growth, structure-specific determinants constitute a powerful selective pressure to retain the format of the two-domain L-shaped tRNA. Aminoacyl-tRNA synthetases are ancient proteins that establish the rules of the genetic code through aminoacylation reactions, in which a specific amino acid is joined to its cognate tRNA (1Lapointe J. Giegé R. Trachsel H. Translation in Eukaryotes. CRC Press, Inc., Boca Raton, FL1991: 35-69Google Scholar, 2Moras D. Trends Biochem. Sci. 1992; 17: 159-164Abstract Full Text PDF PubMed Scopus (200) Google Scholar, 3Carter Jr., C.W. Annu. Rev. Biochem. 1993; 62: 715-748Crossref PubMed Scopus (329) Google Scholar, 4Martinis S.A. Schimmel P. Söll D. RajBhandary U.L. tRNA: Structure, Biosynthesis and Function. American Society for Microbiology, Washington D. C.1995: 349-370Google Scholar, 5Ibba M. Söll D. Annu. Rev. Biochem. 2000; 69: 617-650Crossref PubMed Scopus (1115) Google Scholar). The tRNA, in turn, harbors the anticodon triplet of the code that corresponds to the attached amino acid. Each amino acid has its own specific tRNA synthetase, so that a total of 20 enzymes are responsible for all of the aminoacylation reactions. In bacteria, a single gene typically encodes these 20 enzymes, whereas in eukaryotes, separate genes usually code for cytoplasmic and mitochondrial forms. Because living systems are based on the genetic code, and because tRNA synthetases are key components of the code and only have single genes, a mutation that disrupts aminoacylation activity of any synthetase results in cell death (6Anderson J.J. Neidhardt F.C. J. Bacteriol. 1972; 109: 315-325Crossref PubMed Google Scholar, 7Schimmel P.R. Söll D. Annu. Rev. Biochem. 1979; 48: 601-648Crossref PubMed Scopus (472) Google Scholar). The existence of single-copy genes is believed to reflect, in part, the result of selective pressure against any gene proliferations that could result in corruption of the code (8Lovato M.A. Chihade J.W. Schimmel P. EMBO J. 2001; 20: 4846-4853Crossref PubMed Scopus (37) Google Scholar). For example, were the gene for a specific synthetase to be duplicated, one of the two copies could absorb mutations that might make possible new amino acid or tRNA specificities, whereas the other copy continued to maintain cell growth through the normal aminoacylation reaction. This circumstance could create genetic code ambiguities that, in turn, would place the organism at a selective disadvantage. Genetic code ambiguity can also be caused by misacylation reactions caused by the inherent inability of the active site domain to discriminate between closely similar amino acids (9Jakubowski H. Goldman E. Microbiol. Rev. 1992; 56: 412-429Crossref PubMed Google Scholar, 10Döring V. Mootz H.D. Nangle L.A. Hendrickson T.L. de Crécy Lagard V. Schimmel P. Marliere P. Science. 2001; 292: 501-504Crossref PubMed Scopus (162) Google Scholar, 11Nangle L.A. de Crécy-Lagard V. Döring V. Schimmel P. J. Biol. Chem. 2002; 277: 45729-45733Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). This problem is particularly relevant for hydrophobic amino acids, such as isoleucine, leucine, valine, and alanine (12Baldwin A.N. Berg P. J. Biol. Chem. 1966; 241: 839-845Abstract Full Text PDF PubMed Google Scholar, 13Tsui W.C. Fersht A.R. Nucleic Acids Res. 1981; 9: 4627-4637Crossref PubMed Scopus (79) Google Scholar, 14Englisch S. Englisch U. von der Haar F. Cramer F. Nucleic Acids Res. 1986; 14: 7529-7539Crossref PubMed Scopus (78) Google Scholar, 15Schmidt E. Schimmel P. Science. 1994; 264: 265-267Crossref PubMed Scopus (138) Google Scholar, 16Martinis S.A. Fox G.E. Nucleic Acids Symp. Ser. 1997; 36: 125-128PubMed Google Scholar, 17Chen J.F. Guo N.N. Li T. Wang E.D. Wang Y.L. Biochemistry. 2000; 39: 6726-6731Crossref PubMed Scopus (116) Google Scholar). In these circumstances, a second active site acts as a filter of amino acid fine structure. This discrimination is typically tRNA-dependent; that is, it occurs only in the context of the cognate tRNA (12Baldwin A.N. Berg P. J. Biol. Chem. 1966; 241: 839-845Abstract Full Text PDF PubMed Google Scholar, 18Fersht A. Enzyme Structure and Mechanism. 2nd ed. W. H. Freeman and Company, New York1985Google Scholar, 19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). An example is provided by alanyl-tRNA synthetase (AlaRS), 1The abbreviation used is: AlaRS, alanyl-tRNA synthetase. which misactivates glycine to form the aminoacyl adenylate (Gly-AMP). In the presence of tRNAAla, the misactivated Gly-AMP is cleared by hydrolysis at the editing site of AlaRS (13Tsui W.C. Fersht A.R. Nucleic Acids Res. 1981; 9: 4627-4637Crossref PubMed Scopus (79) Google Scholar, 19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). AlaRS+Gly+ATP→AlaRS(Gly-AMP)+AMP+PPi(Eq. 1) AlaRS(Gly-AMP)+tRNAAla→AlaRS+Gly+AMP+tRNAAla(Eq. 2) Thus, tRNAAla acts as a co-factor for the hydrolytic reaction so that, in the presence of AlaRS, Gly, ATP, and tRNAAla, an abortive cycle of ATP hydrolysis is established. Although the editing site in AlaRS has been localized to a discrete domain that is distinct from the active site (Fig. 1A) (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar, 20Dock-Bregeon A. Sankaranarayanan R. Romby P. Caillet J. Springer M. Rees B. Francklyn C.S. Ehresmann C. Moras D. Cell. 2000; 103: 877-884Abstract Full Text Full Text PDF PubMed Scopus (153) Google Scholar), and although amino acids important for editing have been identified, the role of the tRNAAla cofactor is not understood in detail. However, mischarged Gly-tRNAAla is specifically deacylated by AlaRS, whereas Ala-tRNAAla is not cleared. Gly-tRNAAla+AlaRS→Gly+tRNAAla+AlaRS(Eq. 3) Ala-tRNAAla+AlaRS→littleornoreaction(Eq. 4) Thus, although AlaRS has some discrimination, though limited, of Ala from Gly at the active site for aminoacylation, when in the context of tRNAAla, the two amino acids are further distinguished so that little or no stable Gly-tRNAAla can be generated. The stable production of Gly-tRNAAla would result in genetic code ambiguity from incorporation of glycine at codons for alanine that, ultimately, would lead to cell toxicity (11Nangle L.A. de Crécy-Lagard V. Döring V. Schimmel P. J. Biol. Chem. 2002; 277: 45729-45733Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar, 19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). Among the questions raised by these observations is the basis for the tRNAAla-dependent discrimination of Ala from Gly. Of particular interest is whether the RNA structure needed for aminoacylation is also critical for editing or whether the two activities require, at least in part, different components of the tRNA structure. For this purpose, we took advantage of a model RNA substrate for aminoacylation that lacked roughly half of the tRNA structure. Investigation of the model substrate, and of chimeric tRNAs in which the part of the tRNA that was missing from the model substrate was replaced with pieces from non-alanine tRNAs, provided a strategy to understand better the role of tRNA structural parameters in the discrimination of Gly from Ala. Construction of Plasmids—Construction of chimeric tRNAs was accomplished by annealing four overlapping oligonucleotides encoding the nucleotide sequence of the tRNA using a procedure detailed previously (21Farrow M.A. Nordin B.E. Schimmel P. Biochemistry. 1999; 38: 16898-16903Crossref PubMed Scopus (30) Google Scholar). The annealed oligonucleotide cassette was ligated into PstNI/BamHI-digested pUC18 (22Yanisch-Perron C. Viera J. Messing J. Gene. 1985; 33: 103-119Crossref PubMed Scopus (11472) Google Scholar). The coding sequences within the plasmids were confirmed by DNA sequencing. The plasmid for expression of Escherichia coli AlaRS was constructed through polymerase chain reaction amplification of the gene for AlaRS with oligonucleotides containing either an NdeI or XhoI site. Amplification was carried out on plasmid pQE-alaS-6H (23Ribas de Pouplana L. Schimmel P. Biochemistry. 1997; 36: 15041-15048Crossref PubMed Scopus (21) Google Scholar). The polymerase chain reaction product was digested with NdeI and XhoI and ligated into pET21b (Novagen, Madison, WI) to create a coding sequence for a C-terminal hexahistidine tag for AlaRS in plasmid pET21b-AlaRS-H6. DNA sequencing was again used to verify the sequences of all constructs. RNA Preparation—Karin Musier-Forsyth (University of Minnesota, Minneapolis, MN) kindly supplied a plasmid for the preparation of the in vitro transcript of E. coli tRNAAla. Plasmids were linearized with BstNI, and transcription with T7 RNA polymerase was carried out as described previously (24Steer B.A. Schimmel P. Biochemistry. 1999; 38: 4965-4971Crossref PubMed Scopus (13) Google Scholar). The transcript was resuspended in 50 mm HEPES, pH 7.5, and folded as previously detailed by heat denaturation and slow cooling to room temperature (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). MinihelixAla and the sequence incorporating the second anticodon-containing domain of tRNAAla (Fig. 1C) were synthetically produced and purified as described previously (8Lovato M.A. Chihade J.W. Schimmel P. EMBO J. 2001; 20: 4846-4853Crossref PubMed Scopus (37) Google Scholar). Mischarged tRNAAla was produced in vitro using C666A/Q584H AlaRS, as described previously (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar) in aminoacylation buffer (50 mm HEPES, pH 7.5, 20 mm KCl, 0.1 mg/ml bovine serum albumin, 20 mm β-mercaptoethanol, and 10 mm MgCl2). Labeling of RNA at the 3′-terminal adenosine was accomplished using [γ-32P]ATP and E. coli tRNA-terminal nucleotidyl transferase, as described by Wolfson et al. (25Wolfson A.D. Uhlenbeck O.C. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5965-5970Crossref PubMed Scopus (108) Google Scholar). Protein Expression and Purification—AlaRS was prepared from expression of its gene encoded by plasmid pET21b-AlaRS-H6 in BL21-CodonPlus(DE3)-RIL cells (Stratagene, La Jolla, CA). Expression and purification were carried out as described previously (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar) except that cells were induced with 500 μm isopropyl-1-thio-β-d-galactopyranoside for 3–5 h. T7 RNA polymerase was purified as described previously by nickel-nitrilotriacetic acid chromatography (21Farrow M.A. Nordin B.E. Schimmel P. Biochemistry. 1999; 38: 16898-16903Crossref PubMed Scopus (30) Google Scholar). The plasmid encoding tRNA-terminal nucleotidyltransferase was a generous gift from Nancy Maizels and Alan Weiner and was purified as detailed elsewhere (27Shi P.Y. Weiner A.M. Maizels N. RNA. 1998; 4: 276-284PubMed Google Scholar). Enzyme concentrations were determined by the Bradford assay (28Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (217544) Google Scholar). Aminoacylation and Deacylation Assays—Aminoacylation assays were performed at 37 °C as described previously (29Buechter D.D. Schimmel P. Biochemistry. 1993; 32: 5267-5272Crossref PubMed Scopus (42) Google Scholar) in charging buffer (50 mm HEPES, pH 7.5, 20 mm KCl, 0.1 mg/ml bovine serum albumin, 20 mm β-mercaptoethanol, and10 mm MgCl2). Mischarging with wild-type AlaRS (5 μm) was done in the presence of tRNAAla transcript (5 μm) or minihelixAla (15 μm) and [3H]Gly (10.7 μm) in charging buffer at 37 °C. Aminoacylation of 32P-labeled minihelixAla was followed by the assay described by Wolfson et al. (25Wolfson A.D. Uhlenbeck O.C. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5965-5970Crossref PubMed Scopus (108) Google Scholar). The aminoacylation reaction was carried out with 3 mm alanine (or 175 mm glycine), 500 nm AlaRS, and 5.1 μm labeled minihelixAla. Briefly, at various time points, aliquots were quenched in 200 mm sodium acetate, pH 5.0, containing nuclease P1. After digestion to completion with nuclease P1, aliquots containing a mixture of [32P]AMP and 32P-Ala (or -Gly)-AMP were spotted onto thin layer chromatography plates. Separation was achieved on polyethylenimine cellulose and an SI PhosphorImager (Amersham Biosciences) was used to quantitate the amount of product generated. For deacylation assays, [3H]Gly-tRNAAla (1 μm) or [3H]Gly-minihelixAla (5 μm) was added to AlaRS (20–200 nm) in charging buffer at 37 °C, and the reaction was stopped at several time intervals by addition to filter paper saturated in 5% trichloroacetic acid (21Farrow M.A. Nordin B.E. Schimmel P. Biochemistry. 1999; 38: 16898-16903Crossref PubMed Scopus (30) Google Scholar). Strategy for Identifying Role of tRNA Structure in Editing Response—The tRNA sequence is typically composed of 76 nucleotides arranged in a cloverleaf secondary structure with four stems and three loops. The loops with their associated hydrogen-bonded stems are designated as the dihydrouridine (D), anticodon, and TψC-stem loops (Fig. 1B). Aminoacylation occurs at the 3′-end of the acceptor stem that terminates in the universal sequence XCCA76, where X is the “discriminator base” that can be any of the four nucleotides. (For tRNAAla, X is an A.) The tRNA cloverleaf is arranged in three dimensions as a two-domain L-shaped structure, with the TψC-stem stacking on the acceptor stem to give a 12-bp minihelix domain (capped by the TψC-loop) and the anticodon stem stacking on the D-stem to give a 10-bp helix capped at opposite ends by the D-loop and anticodon loop, respectively, to give the anticodon-containing domain. Tertiary interactions that include contacts between the D- and TψC-loops bring together the minihelix and anticodon-containing domains at a corner (30Kim S.H. Suddath F.L. Quigley G.J. McPherson A. Sussman J.L. Wang A.H.J. Seeman N.C. Rich A. Science. 1974; 185: 435-440Crossref PubMed Scopus (758) Google Scholar, 31Robertus J.D. Ladner J.E. Finch J.T. Rhodes D. Brown R.S. Clark B.F.C. Klug A. Nature. 1974; 250: 546-551Crossref PubMed Scopus (808) Google Scholar). For tRNAAla, the major determinant for aminoacylation is a G3:U70 base pair in the acceptor stem (32Hou Y.M. Schimmel P. Nature. 1988; 333: 140-145Crossref PubMed Scopus (516) Google Scholar, 33McClain W.H. Foss K. Science. 1988; 240: 793-796Crossref PubMed Scopus (292) Google Scholar). Transfer of this base pair into other tRNA confers aminoacylation with alanine (32Hou Y.M. Schimmel P. Nature. 1988; 333: 140-145Crossref PubMed Scopus (516) Google Scholar, 33McClain W.H. Foss K. Science. 1988; 240: 793-796Crossref PubMed Scopus (292) Google Scholar, 34Francklyn C. Schimmel P. Nature. 1989; 337: 478-481Crossref PubMed Scopus (280) Google Scholar). In addition, the 12-bp minihelix domain of tRNAAla (Fig. 1C) is an efficient substrate for aminoacylation with an elevation in the Km but no change in V max (34Francklyn C. Schimmel P. Nature. 1989; 337: 478-481Crossref PubMed Scopus (280) Google Scholar). Minihelices and smaller oligonucleotide substrates with as few as four base pairs are also substrates for charging by AlaRS provided they contain the G3:U70 base pair (34Francklyn C. Schimmel P. Nature. 1989; 337: 478-481Crossref PubMed Scopus (280) Google Scholar, 35Shi J.-P. Martinis S.A. Schimmel P. Biochemistry. 1992; 31: 4931-4936Crossref PubMed Scopus (45) Google Scholar, 36Musier-Forsyth K. Schimmel P. Acct. Chem. Res. 1999; 32: 368-375Crossref Scopus (41) Google Scholar). Thus, in this instance, the second domain of the tRNA structure is not essential for aminoacylation. Indeed, AlaRS makes no contact with the anticodon triplet, although some contacts are made with other parts of the anticodon-containing domain (37Park S.J. Schimmel P. J. Biol. Chem. 1988; 263: 16527-16530Abstract Full Text PDF PubMed Google Scholar). These contacts mainly lower the Km for aminoacylation beyond that achieved with minihelixAla alone (34Francklyn C. Schimmel P. Nature. 1989; 337: 478-481Crossref PubMed Scopus (280) Google Scholar). In our experiments, we first sought to determine whether determinants for aminoacylation could be separated from those for editing. Therefore, we investigated minihelix substrates because they are missing about half of the tRNA structure. The idea was to test whether the second domain of the tRNA structure, although not required for aminoacylation, was needed for editing. If it turned out that the anticodon-containing domain was important for editing, the question then was whether specific determinants for editing could be identified in that domain. MinihelixAla but Not tRNAAla Is Mischarged with Glycine—As stated, tRNAAla and minihelixAla are efficiently aminoacylated with alanine by AlaRS. In contrast, tRNAAla is not charged with glycine because of the enzyme's robust editing activity (13Tsui W.C. Fersht A.R. Nucleic Acids Res. 1981; 9: 4627-4637Crossref PubMed Scopus (79) Google Scholar, 19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). In our first experiments, charging of minihelixAla with glycine was investigated. Remarkably, minihelixAla was efficiently mischarged with glycine under conditions in which no mischarging of tRNAAla was observed (Fig 2A). The experiments in Fig. 2A used concentrations of glycine (10.8 μm) well below the Km determined by the inorganic pyrophosphate exchange assay (58 mm), which is 50–100-fold higher than the Km for alanine (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). The low concentration of glycine was necessary because the assay used [3H]Gly to detect covalent attachment of glycine to minihelixAla, and raising the concentration of glycine by addition of unlabeled glycine would dilute the specific activity. To investigate further the mischarging of minihelixAla with glycine, efforts were focused on a direct comparison of alanine versus glycine, using concentrations of amino acid that were the same relative to the respective Km values. For this purpose, non-radioactive amino acid was used to aminoacylate 32P-end-labeled tRNAAla using an assay described by Wolfson et al. (25Wolfson A.D. Uhlenbeck O.C. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5965-5970Crossref PubMed Scopus (108) Google Scholar). This assay allowed for use of amino acid concentrations well above the Km. With a ratio of concentration/Km of 3 for each amino acid, a direct comparison of charging of minihelixAla with either amino acid could be done. This experiment showed that, under these conditions, Ala and Gly were distinguished by less than 2-fold in the rate at which minihelixAla was charged (Fig. 2B). Thus, apart from RNA-dependent editing, no other activities or functions of AlaRS allow discrimination of Ala from Gly. To test whether the anticodon-containing domain that is missing from minihelixAla could be added in trans to restore editing, an aminoacylation mixture containing minihelixAla and the anticodon-containing domain of tRNAAla was set up. This domain is strongly predicted by MFOLD (38Walter A.E. Turner D.H. Kim J. Lyttle M.H. Muller P. Mathews D.H. Zuker M. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9218-9222Crossref PubMed Scopus (428) Google Scholar) to approximate the stem-biloop structure of the second domain of tRNA (Fig. 1C). With this mixture, no effect of the addition of the anticodon-containing domain could be seen; minihelixAla was still efficiently mischarged with glycine (Fig. 2C). Thus, the two domains of tRNAAla must be covalently linked for efficient clearance of any glycine that has been activated by AlaRS. AlaRS Clears Gly-tRNAAla but Not Gly-MinihelixAla—The results in Fig. 2 collectively show that minihelixAla is missing determinants for RNA-dependent editing. Because Gly-tRNAAla is an efficient substrate for editing by AlaRS, the above results implied that Gly-minihelixAla was not a substrate for editing. To investigate directly this possibility, Gly-tRNAAla and Gly-minihelixAla were prepared by using editing-defective C666A/Q584H AlaRS. (Because of two mutations in the center for editing, this mutant enzyme mischarges tRNAAla with glycine, allowing for the straightforward preparation of tRNAAla or other RNA substrates that are mis-acylated with glycine (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar).) AlaRS efficiently cleared Gly-tRNAAla. However, even at enzyme concentrations 10-fold higher than required to rapidly deacylate Gly-tRNAAla (data not shown) and substrate concentrations 25-fold higher than that of Gly-tRNAAla, no deacylation of Gly-minihelixAla was observed (Fig. 3). In fact, GlyminihelixAla was as stable in the presence as in the absence of AlaRS. Design and Analysis of Chimeric tRNAs with Acceptor Stem Replacements—The failure of AlaRS to clear mischarged minihelixAla (Fig. 3), and the lack of an effect from addition of the anticodon-containing domain to the aminoacylation mixture containing minihelixAla (Fig. 2c), prompted investigation of structural determinants for editing that were missing from minihelixAla. For this purpose, chimeric tRNAs were constructed in which the second, anticodon-containing domain from non-alanine tRNAs was grafted onto minihelixAla. Three different tRNAs were selected for this purpose (Fig. 4). E. coli tRNAPhe is identical in length to tRNAAla and, like tRNAAla, is classified as type I tRNA (39Rich A. RajBhandary U.L. Annu. Rev. Biochem. 1976; 45: 805-860Crossref PubMed Scopus (462) Google Scholar). These tRNAs are in the most common class of tRNAs, containing 76 nucleotides, a variable loop of 5 nucleotides, and a D-stem composed of 4 base pairs. They also have eight nucleotides in the D-loop. As a less conservative example, tRNAGlu was selected, substantially differing in sequence from tRNAAla in the D-stem, and having a D-loop containing nine nucleotides. To complete the range of chimeras with a more dramatic contrast, bovine mitochondrial tRNASer was chosen. This unusual tRNA has only three nucleotides in the variable loop and a D-loop of five instead of eight bases. (Although mitochondrial tRNASer is substantially different in many ways from type I tRNAs, the overall l-shape can still be adopted (40de Bruijn M.H.L. Klug A. EMBO J. 1983; 2: 1309-1321Crossref PubMed Scopus (115) Google Scholar, 41Steinberg S. Leclerc F. Cedergren R. J. Mol. Biol. 1997; 266: 269-282Crossref PubMed Scopus (62) Google Scholar). Thus, the use of this particular tRNA for making a chimera allowed formation of the basic L-like three-dimensional structure.) Collectively, these chimeric tRNAs introduced sequence differences through all parts of the anticodon-containing domain and, with the tRNAAla/Ser chimera, also offered a variation on the canonical structural format of type I tRNAs. Only U33 and A37 in the anticodon loop remained fixed throughout all the tRNAs tested. AlaRS efficiently charged each of the chimeric tRNAs with alanine. Indeed, the apparent k cat /Km for aminoacylation varied no more than 5-fold (arising from either slight alterations in the k cat and/or Km) from that seen with tRNAAla (data not shown). In contrast, like tRNAAla, none of the chimeric molecules could be mischarged with glycine, even using stoichiometric amounts of AlaRS (Fig. 5a). These results imply that all of the chimeric tRNAs are substrates for editing by AlaRS. To test directly the editing response with the chimeric structures, the editing-defective C666A/Q584H AlaRS was used to prepare the misacylated species. Strikingly, each tRNA was deacylated by AlaRS (Fig. 5B). Using catalytic amounts of enzyme, the rates of deacylation were not significantly different (within ∼15%) than that achieved with mischarged tRNAAla as substrate. Because of the collective differences between the three chimeric tRNAs and tRNAAla, and considering especially the differences introduced with the tRNAAla/Ser chimera, the results support the idea that tRNA structure per se and not sequence is essential for RNA-dependent editing by AlaRS. Aminoacyl-tRNA synthetases are divided into two classes of ten enzymes each. The two classes are defined by the active site structures shared by members of the same class (42Webster T.A. Tsai H. Kula M. Mackie G.A. Schimmel P. Science. 1984; 226: 1315-1317Crossref PubMed Scopus (210) Google Scholar, 43Ludmerer S.W. Schimmel P. J. Biol. Chem. 1987; 262: 10801-10806Abstract Full Text PDF PubMed Google Scholar, 44Cusack S. Berthet-Colominas C. Hartlein M. Nassar N. Leberman R. Nature. 1990; 347: 249-255Crossref PubMed Scopus (569) Google Scholar, 45Eriani G. Delarue M. Poch O. Gangloff J. Moras D. Nature. 1990; 347: 203-206Crossref PubMed Scopus (1192) Google Scholar). These classes seem to have independent origins and are thought to reflect, at least in part, an early system of pairing synthetases of opposite classes on the same tRNA acceptor stem (46Ribas de Pouplana L. Schimmel P. Cell. 2001; 104: 191-193Abstract Full Text Full Text PDF PubMed Scopus (147) Google Scholar). AlaRS is a class II enzyme, defined by a catalytic domain for aminoacylation having a seven-stranded β-structure with flanking α-helices (44Cusack S. Berthet-Colominas C. Hartlein M. Nassar N. Leberman R. Nature. 1990; 347: 249-255Crossref PubMed Scopus (569) Google Scholar, 45Eriani G. Delarue M. Poch O. Gangloff J. Moras D. Nature. 1990; 347: 203-206Crossref PubMed Scopus (1192) Google Scholar, 47Ribas de Pouplana L. Buechter D.D. Davis M.W. Schimmel P. Prot. Sci. 1993; 2: 2259-2262Crossref PubMed Scopus (27) Google Scholar). Joined to the C-terminal side of this structure is a domain that contains the active site for editing (Fig. 1A). Class I enzymes are characterized by a Rossmann nucleotide binding fold (48Rao S.T. Rossmann M.G. J. Mol. Biol. 1973; 76: 241-256Crossref PubMed Scopus (717) Google Scholar) that forms the catalytic site for amino acid activation (3Carter Jr., C.W. Annu. Rev. Biochem. 1993; 62: 715-748Crossref PubMed Scopus (329) Google Scholar, 42Webster T.A. Tsai H. Kula M. Mackie G.A. Schimmel P. Science. 1984; 226: 1315-1317Crossref PubMed Scopus (210) Google Scholar, 45Eriani G. Delarue M. Poch O. Gangloff J. Moras D. Nature. 1990; 347: 203-206Crossref PubMed Scopus (1192) Google Scholar, 49Steitz T.A. Curr. Opin. Struct. Biol. 1991; 1: 139-143Crossref Scopus (12) Google Scholar, 50Cavarelli J. Moras D. FASEB J. 1993; 7: 79-86Crossref PubMed Scopus (76) Google Scholar, 51Cusack S. Nat. Struct. Biol. 1995; 2: 824-831Crossref PubMed Scopus (201) Google Scholar, 52Nureki O. Vassylyev D.G. Tateno M. Shimada A. Nakama T. Fukai S. Konno M. Hendrickson T.L. Schimmel P. Yokoyama S. Science. 1998; 280: 578-582Crossref PubMed Scopus (316) Google Scholar). For the closely related class I isoleucyl-, leucyl-, and valyl-tRNA synthetases, an insertion known as CP1 splits the active site (42Webster T.A. Tsai H. Kula M. Mackie G.A. Schimmel P. Science. 1984; 226: 1315-1317Crossref PubMed Scopus (210) Google Scholar, 52Nureki O. Vassylyev D.G. Tateno M. Shimada A. Nakama T. Fukai S. Konno M. Hendrickson T.L. Schimmel P. Yokoyama S. Science. 1998; 280: 578-582Crossref PubMed Scopus (316) Google Scholar, 53Starzyk R.M. Webster T.A. Schimmel P. Science. 1987; 237: 1614-1618Crossref PubMed Scopus (141) Google Scholar, 54Schmidt E. Schimmel P. Biochemistry. 1995; 34: 11204-11210Crossref PubMed Scopus (78) Google Scholar, 55Hale S.P. Schimmel P. Tetrahedron. 1997; 53: 11985-11994Crossref Scopus (13) Google Scholar). This insertion harbors the center for editing that, in the three-dimensional structures, is about 30 Å from the site for amino acid activation (15Schmidt E. Schimmel P. Science. 1994; 264: 265-267Crossref PubMed Scopus (138) Google Scholar, 52Nureki O. Vassylyev D.G. Tateno M. Shimada A. Nakama T. Fukai S. Konno M. Hendrickson T.L. Schimmel P. Yokoyama S. Science. 1998; 280: 578-582Crossref PubMed Scopus (316) Google Scholar, 54Schmidt E. Schimmel P. Biochemistry. 1995; 34: 11204-11210Crossref PubMed Scopus (78) Google Scholar, 56Lin L. Hale S.P. Schimmel P. Nature. 1996; 384: 33-34Crossref PubMed Scopus (92) Google Scholar, 57Silvian L.F. Wang J. Steitz T.A. Science. 1999; 285: 1074-1077Crossref PubMed Scopus (360) Google Scholar, 58Hendrickson T.L. Nomanbhoy T.K. de Crécy Lagard V. Fukai S. Nureki O. Yokoyama S. Schimmel P. Mol. Cell. 2002; 9: 353-362Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar). In the cases of at least isoleucyl- and valyl-tRNA synthetases, the role of the cognate tRNA is to trigger translocation in cis of the misactivated amino acid from the catalytic site for activation to that for editing (56Lin L. Hale S.P. Schimmel P. Nature. 1996; 384: 33-34Crossref PubMed Scopus (92) Google Scholar, 59Nomanbhoy T.K. Hendrickson T.L. Schimmel P. Mol. Cell. 1999; 4: 519-528Abstract Full Text Full Text PDF PubMed Scopus (72) Google Scholar). In addition, the editing site can act “in trans” on exogenously added mischarged tRNAs (18Fersht A. Enzyme Structure and Mechanism. 2nd ed. W. H. Freeman and Company, New York1985Google Scholar, 57Silvian L.F. Wang J. Steitz T.A. Science. 1999; 285: 1074-1077Crossref PubMed Scopus (360) Google Scholar, 60Eldred E.W. Schimmel P.R. J. Biol. Chem. 1972; 247: 2961-2964Abstract Full Text PDF PubMed Google Scholar, 61Schreier A.A. Schimmel P.R. Biochemistry. 1972; 11: 1582-1589Crossref PubMed Scopus (140) Google Scholar). Whether tRNAAla has the same role (a cofactor to stimulate translocation) is not known for AlaRS or for any other class II enzyme. For at least the class I isoleucyl- and valyl-tRNA synthetase, the center for editing can clear either the misactivated aminoacyl adenylate (pretransfer editing) or the mischarged aminoacyl ester (in the form of mischarged tRNA (post-transfer editing)) (11Nangle L.A. de Crécy-Lagard V. Döring V. Schimmel P. J. Biol. Chem. 2002; 277: 45729-45733Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar, 18Fersht A. Enzyme Structure and Mechanism. 2nd ed. W. H. Freeman and Company, New York1985Google Scholar, 56Lin L. Hale S.P. Schimmel P. Nature. 1996; 384: 33-34Crossref PubMed Scopus (92) Google Scholar, 59Nomanbhoy T.K. Hendrickson T.L. Schimmel P. Mol. Cell. 1999; 4: 519-528Abstract Full Text Full Text PDF PubMed Scopus (72) Google Scholar, 60Eldred E.W. Schimmel P.R. J. Biol. Chem. 1972; 247: 2961-2964Abstract Full Text PDF PubMed Google Scholar, 61Schreier A.A. Schimmel P.R. Biochemistry. 1972; 11: 1582-1589Crossref PubMed Scopus (140) Google Scholar, 62Fersht A.R. Kaethner M.M. Biochemistry. 1976; 15: 3342-3346Crossref PubMed Scopus (180) Google Scholar, 63Fersht A.R. Biochemistry. 1977; 16: 1025-1030Crossref PubMed Scopus (202) Google Scholar, 64Bishop A.C. Nomanbhoy T.K. Schimmel P. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 585-590Crossref PubMed Scopus (51) Google Scholar, 65Bishop A.C. Beebe K. Schimmel P.R. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 490-494Crossref PubMed Scopus (26) Google Scholar). The overall editing reactions can be followed by the continuous abortive hydrolysis of ATP (analogous to the summation of Equations 1 and 2 for AlaRS). Not known is whether AlaRS can directly clear Gly-AMP or all of the clearance is through Gly-tRNAAla. However, we confirmed that minihelixAla is inactive in stimulating overall ATP hydrolysis, in the presence of glycine, ATP, and AlaRS. 2K. Beebe, unpublished observations. Thus, if pretransfer editing occurs with AlaRS, then minihelixAla is also inactive for this reaction. Like the class II AlaRS, the class I isoleucyl-tRNA synthetase also needs the entire tRNA structure to activate the editing reaction, with minihelix substrates failing to support editing (66Nordin B.E. Schimmel P. J. Biol. Chem. 1999; 274: 6835-6838Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar). Many chimeric tRNAs (composed of pieces of tRNAIle and tRNAVal) were shown to be active for aminoacylation but not for editing (67Hale S.P. Auld D.S. Schmidt E. Schimmel P. Science. 1997; 276: 1250-1252Crossref PubMed Scopus (79) Google Scholar). After construction of a series of chimeric molecules and of site-specific mutations, three nucleotides in the D-loop were identified as essential for editing (21Farrow M.A. Nordin B.E. Schimmel P. Biochemistry. 1999; 38: 16898-16903Crossref PubMed Scopus (30) Google Scholar, 67Hale S.P. Auld D.S. Schmidt E. Schimmel P. Science. 1997; 276: 1250-1252Crossref PubMed Scopus (79) Google Scholar). Indeed, the editing response could be switched off and on by manipulation of these specific nucleotides located at the corner of the L-shaped structure. These particular nucleotides are important specifically for translocation of a misactivated amino acid from the catalytic site for adenylate synthesis to the center for editing (21Farrow M.A. Nordin B.E. Schimmel P. Biochemistry. 1999; 38: 16898-16903Crossref PubMed Scopus (30) Google Scholar). In contrast, for the alanine system studied here, all chimeric molecules were active in editing, even though they collectively changed nearly every nucleotide in the second domain of the tRNA structure, including those in the D-loop. Particularly striking was the success of the chimera with bovine mitochondrial tRNASer that lacked several nucleotides compared with those seen in most canonical tRNAs. Thus, at least for this class II system, the L-shaped structure per se seems to be the critical factor sensed by the enzyme. Previous work showed that, under specific conditions, disruption of the editing function of AlaRS is cytotoxic (19Beebe K. Ribas de Pouplana L. Schimmel P. EMBO J. 2003; 22: 668-675Crossref PubMed Scopus (141) Google Scholar). Thus, strong selective pressure assures retention of the editing activity. Several lines of evidence suggest that the two domains of tRNA had separate origins and were brought together into a two-domain tRNA to make possible the modern genetic code (36Musier-Forsyth K. Schimmel P. Acct. Chem. Res. 1999; 32: 368-375Crossref Scopus (41) Google Scholar, 68Weiner A.M. Maizels N. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 7383-7387Crossref PubMed Scopus (259) Google Scholar, 69Di Giulio M. J. Mol. Evol. 1992; 159: 199-214Google Scholar, 70Noller H.F. Gestland R.F. Atkins J.F. The RNA World. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1993: 137-156Google Scholar, 71Schimmel P. Giegé R. Moras D. Yokoyama S. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8763-8768Crossref PubMed Scopus (352) Google Scholar, 72Maizels N. Weiner A. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6729-6734Crossref PubMed Scopus (170) Google Scholar, 73Schimmel P. Ribas de Pouplana L. Cell. 1995; 81: 983-986Abstract Full Text PDF PubMed Scopus (158) Google Scholar). Results with the class I isoleucyl-tRNA synthetase and class II alanyl-tRNA synthetase showed that, for both systems, a full tRNA structure is required to trigger the editing response. This requirement is one of several pressures for retaining the two domains of tRNA and the L-shape. Significantly, unlike isoleucyl-tRNA synthetase, in which the anticodon is a major determinant for aminoacylation (26Nureki O. Niimi T. Muramatsu T. Kanno H. Kohno T. Florentz C. Giegé R. Yokoyama S. J. Mol. Biol. 1994; 236: 710-724Crossref PubMed Scopus (101) Google Scholar, 74Muramatsu T. Nishikawa K. Nemoto F. Kuchino Y. Nishimura S. Miyazawa T. Yokoyama S. Nature. 1988; 336: 179-181Crossref PubMed Scopus (362) Google Scholar), alanyl-tRNA synthetase makes no contact with the anticodon and has its major determinants for aminoacylation in the minihelix domain. Thus, even when the anticodon-containing domain has a minor role in determining aminoacylation efficiency, the editing function exerts strong pressure to bring together the anticodon-containing domain with the minihelix to create the L-shaped structure.