tRNA recognition by glutamyl-tRNA reductase.

During the first step of porphyrin biosynthesis in Archaea, most bacteria, and in chloroplasts glutamyl-tRNA reductase (GluTR) catalyzes the NADPH-dependent reduction of glutamyl-tRNA to glutamate-1-semialdehyde. Elements in tRNA(Glu) important for utilization by Escherichia coli GluTR were determined by kinetic analysis of 51 variant transcripts of E. coli Glu-tRNA(Glu). Base U8, the U13*G22**A46 base triple, the tertiary Watson-Crick base pair 19*56, and the lack of residue 47 are required for GluTR recognition. All of these bases contribute to the formation of the unique tertiary core of E. coli tRNA-(Glu). Two tRNA(Glu) molecules lacking the entire anticodon stem/loop but retaining the tertiary core structure remained substrates for GluTR, while further decreasing tRNA size toward a minihelix abolished GluTR activity. RNA footprinting experiments revealed the physical interaction of GluTR with the tertiary core of Glu-tRNA(Glu). E. coli GluTR showed clear selectivity against mischarged Glu-tRNA(Gln). We concluded that the unique tertiary core structure of E. coli tRNA(Glu) was sufficient for E. coli GluTR to distinguish specifically its glutamyl-tRNA substrate.

NADPH-dependent reduction of glutamate to glutamate-1semialdehyde (GSA), which is subsequently converted to 5-aminolevulinic acid by glutamate-1-semialdehyde-2,1-aminomutase (GSA-AM) (4,5). To maintain the fidelity of protein biosynthesis and, at the same time, an adequate metabolite flux for 5-aminolevulinic acid formation, accurate recognition of various cellular tRNAs by their cognate aminoacyl-tRNA synthetase and of Glu-tRNA by GluTR must be secured. The characteristic nucleotides in tRNAs (identity elements) needed for recognition by their cognate aminoacyl-tRNA synthetases have been characterized for many systems (6). For the Escherichia coli GluRS⅐tRNA Glu system, 5-methylaminomethyl-2thiouridine in position 34, U35, C36, A37 in the anticodon loop, G1*C72, U2*A71 in the acceptor stem, the U11*A24 base pair, the U13*G22**A46 base triple, and the lack of residue 47 (⌬47) determine the charging identity (7,8,33). The crystal structure of the GluRS⅐tRNA Glu complex (9,10) provided further insight to tRNA Glu recognition by GluRS. In contrast, there is little information on the recognition of Glu-tRNA Glu by GluTR. Specific sequence information is required, as 5-aminolevulinic acid synthesis in barley chloroplasts demands chloroplast tRNA Glu and not cytoplasmic tRNA Glu (11). Specificity of GluTR for tRNA Glu was also shown for the enzyme from Chlorella vulgaris (31), Chlorobium vibrioforme (32), Chlamydomonas reinhardtii (13), and Synechocystis 6803 (12). Based on a sequence comparison of tRNA Glu species that are substrates for various GluTRs with tRNA species which are not utilized, putative identity elements were postulated (14). A point mutation (C56 -U56) in Euglena gracilis chloroplast tRNA Glu was reported to uncouple protein and chlorophyll biosynthesis. This mutant tRNA was still aminoacylated by chloroplast GluRS and utilized in protein biosynthesis but was not a substrate for GluTR and therefore did not support tetrapyrrole biosynthesis (15). The solution of the long sought-after crystal structure of GluTR (30) and the biochemical characterization of E. coli GluTR (16,17) generated much interest in the details of tRNA recognition by this protein. The role of the glutamate part of the substrate was elucidated by site-directed mutagenesis of the E. coli GluTR substrate binding pocket and the co-crystallization of glutamycin, a glutamate analogue, with the Methanopyrus kandleri enzyme (29). Glutamycin was found tightly coordinated in the highly conserved catalytic pocket, which also includes the cysteine residue (Cys-50) known to form an intermediate thioester bond between the enzyme and glutamate (16). In contrast, much less is known about the contribution of the tRNA portion of the substrate to recognition and catalysis. Here we present an analysis of tRNA identity in GluTR recognition.

EXPERIMENTAL PROCEDURES
Production of Recombinant E. coli GluTR, GluRS, GSA-AM, and T7 RNAP-E. coli GluTR was produced as a six-histidine N-terminal fusion protein, renatured from inclusion bodies and purified as described (17). Recombinant E. coli GluRS and GSA-AM were purified to apparent homogeneity according to published procedures (18,25). Bacteriophage T7 RNAP was purified from an overproducing E. coli strain harboring the plasmid pAR1219 as outlined previously (19).
Construction of Expression Vectors for tRNA Glu Variants-The majority of expression vectors for the production of E. coli tRNA Glu variants were described before (7). The plasmid pKR320 (20) contains the E. coli tRNA Glu gene cloned downstream of a T7 RNA polymerase promoter. The tRNA Glu variants encoding tRNAs harboring the point mutations C56U, G19A, C56U/G19A, U8C, C56G, minimalist tDNA constructs ⌬AC1 and ⌬AC2, and the minihelix (see Fig. 1) were generated by site-directed mutagenesis of the plasmid pKR320 using the QuikChange TM site-directed mutagenesis kit (Stratagene, Heidelberg, Germany). To obtain the minimalist tDNA constructs an expression vector encoding a tRNA Glu transcript that lacks the entire anticodon stem/loop (pKR320⌬AC) was generated by deletion of 22 nucleotides using the oligonucleotide 5Ј-CTAGAGGCCCAGGACAGGGGTTCGAA-TCC-3Ј. The following oligonucleotides were employed to insert the linker region in plasmid pKR320⌬AC underlined, 5Ј-CTAGAGGCCCA-GGACAAATATAACAGGGGTTCGAATCC-3Ј (⌬AC1) and 5Ј-CTAGAG-GCCCAGGACAATTAACAGGGGTTCGAATCC-3Ј (⌬AC2). The minihelix was constructed by the deletion of 19 nucleotides of plasmid pKR-320⌬AC using the oligonucleotide 5Ј-CACTATAGTCCCCTAGGGGTT-CGAATC-3Ј. All introduced mutations were verified by complete DNA sequence determination.
Preparation and Purification of tRNA Gene Transcripts-E. coli tRNA Glu (UUC) variants and minimalist tRNA Glu molecules were synthesized by in vitro T7 RNAP run-off transcription (21). The DNA fragments carrying the E. coli tRNA Glu gene variants including the T7 RNAP promoter were amplified using PCR. To generate the appropriate 3Ј-CCA-end of the tRNA Glu transcripts, the various PCR-amplified template DNAs carrying the tRNA Glu genes were digested with NspI and with FokI in the case of the minimalist tRNA molecules. The in vitro transcription reaction was performed at 37°C for 3 h in a buffer containing 40 mM Na-HEPES, pH 8.1, 10 mM MgCl 2 ,40 mM KCl, 10 mM dithiothreitol, 2 mM spermidine, 200 g/ml bovine serum albumin, 2.0 -5.0 mM each nucleoside triphosphate (depending on the template), the PCR-amplified template DNA (ϳ0.5 M), and 100 g/ml (1 mM) T7 RNAP. The E. coli tRNA Gln (CUG) transcript was prepared according to published procedures (22). The RNA transcripts were subsequently purified by MonoQ anion exchange chromatography as described (23).
Aminoacylation of tRNA Glu -The GluTR substrate [ 14 C]Glu-tRNA Glu was prepared in 100 or 500 l aminoacylation reaction mixtures containing 30 mM Na-HEPES, pH 7.5, 15 mM MgCl 2 , 25 mM KCl, 3 mM dithiothreitol, 4 mM ATP, and 40 M [ 14 C]glutamate (10 Ci) with a specific activity of 284 mCi/mmol (10.5 GBq/mmol). The tRNA transcript concentrations ranged from 1 to 20 M and GluRS concentrations varied from 400 nM to 8 M depending on the specific tRNA variant. Minimalist tRNA substrates were denatured at 90°C for 3 min and quickly cooled on ice to facilitate RNA folding.
GluTR Assay-As GSA, the product of GluTR catalysis, is highly unstable under physiological conditions and difficult to isolate in reproducible quantities, GluTR activity was measured by the substrate depletion assay, which is currently the most reliable method of analysis (24). This assay determines the decrease of The tRNA was precipitated, washed, and quantified as described previously (27). Reaction assays without GluTR for each employed condition served as background controls for spontaneous substrate hydrolysis (ϳ5% after a 3-min reaction time). As some of the mutations in tRNA Glu also affected glutamylation, the amounts of the GluTR substrate, Glu-tRNA Glu , were limited and did not permit use of the appropriate excess of substrate required for classical Michaelis-Menten ki-netics. Therefore, GluTR specific activity was deduced from first-order equation by a method described previously (44).
HPLC Analysis of GluTR Product Formation-Reaction products were analyzed on a Waters Bondapack™ C18 reversed phase column (3.9 ϫ 150 mm, 125-Å pore size, 10-m particle diameter) as described previously (24). The reaction product GSA was identified by its specific conversion into 5-aminolevulinic acid using purified E. coli GSA-AM (25).
tRNA Footprinting Using RNase Digestion-The E. coli tRNA Glu transcript was 3Ј-end-labeled using published procedures (26) and charged with glutamate as described above. RNase footprintings were performed according to the manufacturer's instruction using RNase V1, RNase T1, and Nuclease S1 purchased from Ambion (Austin, TX). Prior to the reaction 2-3 g of end-labeled Glu-tRNA Glu , 0.1-10 M GluTR, and 10 mM of the GluTR inhibitor ␤-nicotinamide-mononucleotide (reduced form) (Sigma) were incubated at 4°C for 3 min to allow the formation of the GluTR⅐tRNA complex. Control experiments omitted either RNase or GluTR. The cleavage pattern of the footprinting reactions was analyzed by electrophoresis on denaturing 15% polyacrylamide gels. The gels were exposed to a Fuji Bas-III imaging screen for 16 h at Ϫ80°C.
Circular Dichroism (CD) and UV Absorption Measurements-CD and UV molecular absorption spectra (195-320 nm) were recorded on a Jasco J-810 spectropolarimeter equipped with Neslab RET-110 temperature control unit. Each tRNA sample (ϳ0.3 mg/ml) was initially stabilized at 27°C for 15 min. Melting experiments were performed at 1°C increments using a temperature rate of 30°C/h in the presence of either 10 mM EDTA or 10 mM MgCl 2 .

RESULTS AND DISCUSSION
Because both GluTR and GluRS recognize the same tRNA Glu , we wanted to compare which nucleotides are crucial for recognition. The earlier studies on GluRS identity not only presented a basis for comparison but also provided the large set of in vitro synthesized Glu-tRNA Glu variants required to identify residues in E. coli tRNA Glu that are important for E. coli GluTR activity.
Kinetic Analysis of tRNA Glu Variants-The recognition of Glu-tRNA Glu by GluTR was examined by the kinetic analysis of unmodified wild type tRNA Glu (Fig. 1) and 51 tRNA Glu variants that were prepared by in vitro run-off transcription. All transcripts were aminoacylated with [ 14 C]glutamate using E. coli GluRS and tested as GluTR substrates in the depletion assay by monitoring the loss of charged tRNA Glu (24). The formation of the reaction product GSA was verified by HPLC analysis. It was shown previously that E. coli GluTR utilizes fully modified E. coli Glu-tRNA Glu and unmodified Glu-tRNA Glu transcripts at comparable catalytic rates (16). The specific activity for this process was measured as 0.47 mol min Ϫ1 mg Ϫ1 (k cat ϭ 0.38 s Ϫ1 ) and was set to 100%. All other values obtained in this study were related to this value. The majority of charged tRNA Glu transcripts harboring single, double, or triple mutations were utilized by E. coli GluTR without any significant reduction of the catalytic activity (Table I). Interestingly, this set also includes tRNA Glu transcripts carrying the mutations of bases that are known to be identity elements in tRNA Glu for recognition by E. coli GluRS (7,8).
GluTR Does Not Recognize the Acceptor Arm of tRNA and the Anticodon Stem/Loop-Each single base pair of the anticodon stem and all bases of the anticodon loop of tRNA Glu were exchanged without affecting GluTR activity. This contrasts with observations made with GluRS, which is known to recognize the residues U34, U35, C36, and A37 of the anticodon loop. In agreement with this, the loss of the 5-methylaminomethyl-2-thiol (mnm 5 s 2 ) modification of the base U34 in the tRNA Glu transcript, which significantly contributes to tRNA Glu identity for GluRS (33), does not affect GluTR activity (16). Surprisingly, no tRNA Glu identity elements for GluTR were found in the acceptor stem. The discriminator base G73 and all base pairs of the acceptor stem (G1*C72, U2*A71, C3*G70, C4*G69, C5*G68, C6*G67, U7*A66) were mutated without affecting tRNA Recognition by Glutamyl-tRNA Reductase GluTR activity. In contrast, GluRS was shown to recognize G1*C72 and U2*A71 (7).
The Unique Tertiary Core Structure of tRNA Glu Is Recognized by GluTR-The tRNA Glu variants that led to significantly reduced GluTR activity are summarized in Table II. A set of tRNA Glu variants harboring mutations in the tertiary core (D-stem/loop, variable loop, and T-stem/loop) was synthesized by T7 RNAP transcription. The U13*G22**A46 base triple is formed via a tertiary base pairing between the phylogenetically conserved pyrimidine 13-purine 22 motif (U13*G22 in tRNA-Glu ) in the D-stem and base 46 (A46) in the variable loop (36) (see Fig. 2C). A tRNA Glu transcript carrying a mutation of U13-C13, which allows the formation of the Watson-Crick base pair C13*G22, was still a reasonable substrate for GluTR but led to significantly reduced GluRS activity (7). However, for both GluRS and GluTR mutations of the bases G22 and A46 (to A22, G46, U46) almost abolished substrate utilization. Furthermore the tertiary core of tRNA Glu is characterized by the lack of residue 47 in the four nucleotide short variable loop. The insertion of U into position 47 decreased the catalytic activity of GluTR as well as of GluRS. As the base triple 13*22**46 was shown to play an important role in maintaining the overall folding of the tRNA (34), it seems likely that GluTR recognition strongly depends on the unique core structure of tRNA Glu . In addition, the lack of residue 47 is thought to stabilize the mentioned base triple via stable positioning of base A46 (35). Furthermore the strictly conserved tertiary base pair U8*A14, which forms a reverse Hoogsteen pairing configuration, was disrupted in a tRNA Glu variant, which harbored a mutation of the base U8 -C8. The transcript, although charged by GluRS, was only a weak substrate for GluTR (Table II). The finding that the tertiary core structure of tRNA Glu is recognized by GluTR was supported by in vivo data from E. gracilis. It was shown that a point mutation of tRNA Glu led to a chlorophylldeficient phenotype, with a tRNA Glu variant still participating in chloroplast protein biosynthesis but no longer able to be utilized by GluTR (15). The E. gracilis tRNA Glu harbored the single mutation of C56 -U56. As this base forms a tertiary Watson-Crick base pair with the base G19 it was interesting to test the consequences of this mutation in the E. coli system. We verified that the E. coli tRNA Glu variant transcript carrying a mutation of the base C56 -U56 was efficiently aminoacylated by E. coli GluRS. However, it was a poor substrate for E. coli GluTR. A single mutation of base G19 -U19 also prevented utilization by GluTR. Interestingly, the double mutant restoring this tertiary base pair (A19**U56 -G19**C56) led to nearly full GluTR utilization. Thus the base exchanges C56 -U56 and G19 -U19 destroy another important feature of the tertiary core of tRNA Glu , the G19**C56 base pair (Fig. 2). The most crucial determinant for tRNA Glu recognition by GluRS (7), the U11*A24 base pair, which does not form any tertiary interactions, was found to be unimportant for GluTR activity.

TABLE I Investigated Glu-tRNA Glu variants with no influence on
GluTR activity Glu-tRNA Glu variants were prepared as described (see "Experimental Procedures"). GluTR activity was measured by the substrate depletion assay. Product formation was verified by HPLC analysis. tRNAs Base exchange tRNA Glu transcript G73 3A73 T-stem A49*U65 3G49*C65 G50*C64 3A50*U64 G51*C63 3A51*U63 G52*C62 3A52*U62 D-stem/loop A26**U44 3G26**C44 C25*G10**A45 3U25*A10**G45 C25*G10**A45 3C25*G10**G45 U11*A24 3C11*G24 C12*G23**C9 3U12*A23**A9 C12*G23**C9 3C12*G23**A9 U13*G22 3G13*U22 A16 3G16 C20 3U20 C20a 3U20a tRNA Recognition by Glutamyl-tRNA Reductase Conformational Stability of the tRNA Transcripts-Because of the central role of the tRNA Glu backbone in GluTR recognition, the mutated tRNA Glu transcripts used in this study were analyzed for their structural integrity. The stabilities of wild type tRNA Glu transcripts and the variants affecting tertiary base interactions were investigated by thermal unfolding experiments. The changing spectroscopic features of various tRNA Glu transcripts were investigated by UV absorbance and CD spectroscopy and compared. The increasing UV absorbance during the thermal tRNA unfolding process was monitored between 250 and 270 nm. This region of the tRNA spectrum showed the highest degree of absorbance and hyperchromicity (data not shown). However, the UV absorbance spectra of the various tRNA Glu molecules changed only slightly during the thermal unfolding process. In contrast, clear changes of the corresponding CD spectra between 80 and 320 nm were observed. The CD spectrum of folded wild type tRNA Glu at 27°C showed a minimum centered at 210 nm and a maximum centered at 270 nm. During the melting process of this tRNA Glu molecule the absorbance at 210 nm increased with the rising temperature, whereas the absorbance at 270 nm maximum decreased in amplitude and shifted its maximum to 277 nm (Fig. 3A). Fig. 3B shows a differential representation of the melting curves based on changes in the ellipticity at the single wavelength at 210 nm for the wild type tRNA Glu transcript. In the same figure the corresponding data for the tRNA Glu variant transcript harboring the base exchange C56 -U56 is shown. Only minimal differences (Ͻ1°C) of the melting point between Tertiary core 0 E. coli tRNA Gln transcript 0 a GluTR activity was measured by the substrate depletion assay as described in detail under "Experimental Procedures." Product formation was verified by HPLC analysis. The specific activity (0.47 mol min Ϫ1 mg Ϫ1 ) of GluTR with wild type tRNA Glu represents 100%.
b The truncated tRNA mutants are described under "Experimental Procedures."

FIG. 2. Nucleotides of E. coli tRNA Glu involved in E. coli GluTR recognition.
A, nucleotides of tRNA Glu transcript affecting GluTR activity (see Table II) are indicated in the L-shaped representation. B, three-dimensional presentation of tRNA Glu highlighting the bases required for GluTR recognition. The bases are positioned according to the crystal structure of Thermus thermophilus tRNA Glu (9). C, the tertiary triple contact U13*G22**A46 present in E. coli tRNA Glu .
tRNA Recognition by Glutamyl-tRNA Reductase wild type tRNA Glu transcript and the U56 variants were noticed, indicating an almost identical stability of these transcripts. Similar to identical results were obtained during the analysis of the tertiary core mutants G56, G56, A19, and A19**U56. In agreement with these observations, the melting temperature for wild type tRNA Glu transcript and all investi- tRNA Recognition by Glutamyl-tRNA Reductase gated variants was shifted from 33°C in the presence of 10 mM EDTA to 56°C in the presence of 10 mM MgCl 2 . In conclusion, the performed melting experiments indicate no altered conformational stability for any investigated variants. We conclude that mutations in the tertiary base pair C56*G19 cause only minor alterations in structure and stability. Nevertheless, this part of the tRNA Glu molecule represents the central recognition element for GluTR.
Minimization of the RNA Substrate-To further narrow down the influence of the tRNA tertiary core on overall tRNA Glu recognition by GluTR, minimalist tRNA Glu substrates were designed (Fig. 1). For the truncated construct ⌬AC1 the anti-codon stem/loop region was deleted and replaced by a linker region of the two nucleotides AU. For the molecule ⌬AC2 the linker consists of four nucleotides (AAUA). These linker regions were designed to allow the necessary base interactions of the tertiary core. Aminoacylation of these truncated transcripts by GluRS to wild type level was achieved by increasing the RNA and GluRS concentration up to 3 and 10 M, respectively. Both constructs ⌬AC1 and ⌬AC2 served as efficient substrates in the standard GluTR depletion assay (Table II). We concluded from these results that the whole anticodon stem/loop region is not needed for the recognition of Glu-tRNA Glu by GluTR. To examine the effect of further reduction of the tRNA substrate, a minihelix comprising the acceptor stem, the T-stem, and T-loop was constructed. This minihelix is missing the tertiary core of tRNA Glu . The minihelix transcript was charged with glutamate in an aminoacylation reaction containing 16 M GluRS for 1.5 h. This glutamyl-minihelix was not a substrate for GluTR in the standard GluTR depletion assay (Table II). In conclusion, these experiments provided further evidence for the role of the tertiary core of tRNA Glu in GluTR recognition.
RNase Footprinting Reveals Glu-tRNA Glu ⅐GluTR Interactions at the Tertiary Core-RNase footprinting was employed to investigate GluTR⅐Glu-tRNA Glu interactions. Initial experiments were hampered by substrate destruction caused by GluTR. In the absence of NADPH GluTR possesses an esterase activity that hydrolyzes the Glu-tRNA Glu substrate into tRNA Glu and glutamate. To stabilize the GluTR-bound substrate GluTR esterase activity was inhibited by the NADPH analogue ␤-nicotinamide-mononucleotide (reduced form) as described before (24). We used RNases V1, T1, and nuclease S1 to digest Glu-tRNA Glu alone and in complex with GluTR to investigate changes in the protection pattern. During the precharging reaction nonspecific cleavage of the transcript at certain weak points was inevitable and might be induced by GluRS (28). These positions were omitted from further analysis. Surprisingly, the addition of up to 100 nM GluTR reproducibly led to increased RNase V1 digestion at the T-stem, D-stem, and parts of the anticodon stem. These findings might indicate conformational changes of the tRNA structure induced by docking to GluTR or alternatively the occurrence of GluTR RNase V1 protein-protein interactions. Thus, changes in the protection pattern of tRNA Glu were only observed by RNase T1 and nuclease S1 treatment. Protection of the base G10 in the Dstem and the bases G53, U54, and U55 in the T-loop from RNase digestion by GluTR was observed (Fig. 4A). The protection of these positions was verified by digestion of phosphorothioate containing tRNA transcripts (data not shown). Therefore, the direct interaction of GluTR with the tertiary core of tRNA Glu was demonstrated. In the current model of Glu-tRNA Glu ⅐GluTR interactions (29), based on the position of glutamycin, which mimics the 3Ј-end of Glu-tRNA Glu in the active site pocket, the tRNA Glu is bound from the inside of the L-shape of the molecule, similar to its binding to GluRS. It is in pre-cisely this region of the tRNA that the protected bases are localized (Fig. 4). The Glu-tRNA Glu ⅐GluTR model is based on the M. kandleri GluTR structure, which reveals significant catalytic and structural similarities to the E. coli enzyme (16) so that an analogue tRNA orientation is proposed.
GluTR Shows tRNA Glu Specificity-Finally, we investigated whether the unique tertiary core of tRNA Glu provides enough information to discriminate against the closely related tRNA Gln when mischarged with glutamate. The tRNA Gln transcript of E. coli was mischarged with glutamate using 31 M GlnRS and 80 M glutamate and tested for GluTR acceptance in standard depletion assays. However, [ 14 C]Glu-tRNA Gln was not a substrate for GluTR. Because many organisms carrying GluTR form glutaminyl-tRNA via naturally mischarged Glu-tRNA Gln (20) this observation is of biological significance.
General Discussion-GluTR is an enzyme that uses aminoacyl-tRNA as substrate. Specific recognition of Glu-tRNA by E. coli GluTR is brought about by the unique tertiary core of tRNA Glu created by the U13*G22**A46 base triple, ⌬47, and the tertiary base pair G19*C56. Neither the anticodon nor acceptor stem are major identity elements, as they are for most aminoacyl-tRNA synthetases. The accuracy of this process is further sustained by the tight coordination of the glutamate part of the substrate in the active site pocket of GluTR.
Do other enzymes that utilize aminoacyl-tRNA as a substrate recognize tRNA in a similar fashion? For instance, E. coli methionyl-tRNA fMet formyltransferase, the enzyme required for formylating the initiator tRNA, requires a cluster of determinants in the acceptor stem and the A11*U24 base pair in the D-stem (37,38). Furthermore, similar to GluTR, increased RNase V1 digestion was observed in the anticodon stem, although this region is located distal to the tRNA-protein contacts observed in the co-crystal structure (39). This observation was explained by conformational changes of the tRNA structure induced by Met-tRNA formyltransferase binding (40). In addition, enhanced RNase V1 cleavage in the anticodon stem was also described for the EF-Tu⅐Phe-tRNA Phe ⅐GTP and IF2⅐fMet-tRNA complexes suggesting changes in anticodon stem conformation even though these proteins interact with the distal acceptor stem (41,42). It will be interesting to explore conformational changes in Glu-tRNA Glu structure possibly induced by GluTR binding. Finally, selenocysteyl-tRNA Sec is an aminoacyl-tRNA substrate specifically recognized by selenocysteine synthase and the SELB translation factor. The unusually long eight-base-pair acceptor stem of E. coli tRNA Sec was found to be the major recognition determinant for SELB (43). The mechanisms of further enzymes utilizing an aminoacyl-tRNA substrate, e.g. Glu-tRNA Gln amidotransferase, which transamidates misacylated Glu-tRNA Gln , are still under investigation. These results suggest that enzymes utilizing aminoacyl-tRNA as substrate interact directly with the aminoacylated acceptor stem and the D-stem, whereas the anticodon domain serves as a major recognition element of aminoacyl tRNA synthetases.