Functional Interaction of Mammalian Valyl-tRNA Synthetase with Elongation Factor EF-1α in the Complex with EF-1H*

In mammalian cells valyl-tRNA synthetase (ValRS) forms a high M r complex with the four subunits of elongation factor EF-1H. The β, γ, and δ subunits, that contribute the guanine nucleotide exchange activity of EF-1H, are tightly associated with the NH2-terminal polypeptide extension of valyl-tRNA synthetase. In this study, we have examined the possibility that the functioning of the companion enzyme EF-1α could regulate valyl-tRNA synthetase activity. We show here that the addition of EF-1α and GTP in excess in the aminoacylation mixture is accompanied by a 2-fold stimulation of valyl-tRNAValsynthesis catalyzed by the valyl-tRNA synthetase component of the ValRS·EF-1H complex. This effect is not observed in the presence of EF-1α and GDP or EF-Tu·GTP and requires association of valyl-tRNA synthetase within the ValRS·EF-1H complex. Since valyl-tRNA synthetase and elongation factor EF-1α catalyze two consecutive steps of the in vivo tRNA cycle, aminoacylation and formation of the ternary complex EF-1α·GTP·Val-tRNAVal that serves as a vector of tRNA from the synthetase to the ribosome, the data suggest a coordinate regulation of these two successive reactions. The EF-1α·GTP-dependent stimulation of valyl-tRNA synthetase activity provides further evidence for tRNA channeling during protein synthesis in mammalian cells.

Aminoacyl-tRNA is the donor of amino acid in ribosomal protein synthesis. The tRNA molecule is aminoacylated with the corresponding amino acid by an aminoacyl-tRNA synthetase, the aminoacyl-tRNA is converted to a ternary complex with elongation factor 1␣, to give the immediate precursor of amino acid for protein synthesis: EF-1␣⅐GTP⅐aminoacyl-tRNA. Several lines of evidence have suggested that in mammalian cells the translational apparatus is highly organized. In particular, association of the protein vectors of tRNA, aminoacyl-tRNA synthetases and elongation factors, with the cytoskeletal framework has been reported (1-3) and colocalization of these components described (4). The isolation and characterization of supramolecular assemblies of aminoacyl-tRNA synthetases and elongation factors (5-7) have provided structural evidence for the subcellular organization of the protein synthesis machinery. The existence of a channeled tRNA cycle during mammalian protein synthesis provided functional evidence for cellular compartmentalization of translation (8 -10). According to the proposed channeling scheme, aminoacyl-tRNAs are vectorially transferred from the aminoacyl-tRNA synthetases to the ribosomes as ternary complexes EF-1␣⅐GTP⅐aminoacyl-tRNA (8,10). Moreover, the GDP form of EF-1␣ could be involved in the capture of deacylated tRNA at the exit site of the ribosome and its delivery to the synthetase (11).
Channeling, or direct transfer of metabolites from one enzyme to another in a metabolic pathway, is believed to increase significantly the efficiency of the overall reaction (12). For sequential metabolic enzymes, the stimulation of activity of the first enzyme induced by a protein-protein interaction with the second enzyme provides a structural basis for a channeling mechanism. As far as the protein biosynthesis machinery is concerned, the possible regulation of aspartyl- (13) and phenylalanyl- (14) tRNA synthetase activities by elongation factor EF-1␣ has been reported. In both cases, no stable proteinprotein interaction between the synthetase and the elongation factor was detected. In connection with tRNA channeling from aminoacyl-tRNA synthetase to EF-1␣, two enzymes ensuring consecutive steps of the tRNA cycle, the multienzyme complex containing valyl-tRNA synthetase and EF-1H deserves special mention.
The only stable macromolecular assemblage that involves an aminoacyl-tRNA synthetase and elongation factor 1␣ is the ValRS 1 ⅐EF-1H complex (15,16). EF-1H, the "heavy" form of the translation elongation factor 1, is a pentameric complex of the four subunits ␣, ␤, ␥, and ␦ in molar ratio 2:1:1:1 (17). EF-1␣ forms a ternary complex with aminoacyl-tRNA and GTP to give the active species of amino acid for protein synthesis and delivers aminoacyl-tRNA to the A site of the ribosome. The EF-1␤␥␦ subunits contribute the guanine nucleotide exchange activity to regenerate EF-1␣⅐GTP from EF-1␣⅐GDP. The finding that in mammalian cells valyl-tRNA synthetase is exclusively found as a complex with EF-1H has suggested that this association might contribute an essential function in vivo. Here we show that EF-1␣ controls the aminoacylation reaction catalyzed by ValRS. Our results demonstrate that association of ValRS with EF-1H is absolutely required for the stimulation in trans by EF-1␣. The ValRS⅐EF-1H complex provides a structural basis for the functional interaction between ValRS and * This work was supported in part by grants from the CNRS and the Association pour la Recherche sur le Cancer (France), by Grant  EF-1␣. The coupling of these two consecutive reactions during protein biosynthesis could have a regulatory function.

EXPERIMENTAL PROCEDURES
Protein Purification-The ValRS⅐EF-1H complex was isolated from rabbit liver according to a procedure adapted from that used to isolate the multisynthetase complex (18). Briefly, livers (1 kg) from 13 rabbits were homogenized in three portions (w/v; total volume 2 liters) of extraction buffer (50 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 0.1 mM EDTA, 10% glycerol, 1 mM DTE (1,4-dithioerythritol)) containing 1 mM diisopropyl fluorophosphate. Postmitochondrial supernatant was adjusted to 2% polyethylene glycol-6000 by addition of a 50% stock solution in extraction buffer and stirred for 30 min at 4°C. After centrifugation at 10,000 ϫ g for 30 min, the supernatant was recovered, adjusted to 5% polyethylene glycol-6000, and stirred for 30 min at 4°C. After centrifugation, the precipitate was dissolved in 200 ml of 25 mM potassium phosphate (pH 7.6), 5% glycerol, and 10 mM 2-mercaptoethanol. The homogeneous solution was applied to a Bio-Gel A-5m column (Bio-Rad; 90 ϫ 910 mm) equilibrated with the same buffer containing 10% glycerol. Fractions containing valyl-tRNA synthetase activity were combined and loaded on a tRNA-Sepharose column (50 ϫ 315 mm) equilibrated with the same buffer. The enzyme was eluted by a linear gradient of 25-400 mM potassium phosphate buffer (pH 7.6). Fractions with valyl-tRNA synthetase activity were pooled and applied to a Source 15Q column (Amersham Pharmacia Biotech; 20 ϫ 95 mm) equilibrated in 25 mM Tris-HCl (pH 7.5), 50 mM KCl, 10% glycerol, and 10 mM 2-mercaptoethanol and developed with a 50 -1000 mM KCl gradient in the same buffer. Fractions with ValRS activity were pooled, concentrated by vacuum dialysis, and stored at Ϫ20°C after dialysis against 25 mM potassium phosphate (pH 7.6), 50% glycerol, and 2 mM DTE.
Dissociation of the ValRS⅐EF-1H complex and isolation of its ValRS, EF-1␦, and EF-1␤␥ subunits in their native state was conducted in the presence of 0.5 M NaSCN as reported previously (19). The free truncated ValRS species deprived of its NH 2 -terminal extension was obtained by controlled elastase treatment, as described (20).
The free form of elongation factor 1␣ was purified to homogeneity from rabbit liver as described in (14) and stored at Ϫ80°C in 25 mM Tris-HCl (pH 7.5), 25 mM NH 4 Cl, 2 mM MgCl 2 , 2 mM DTE, and 25% glycerol.
Enzymatic Assays-Initial rates of tRNA aminoacylation were measured at 25°C in 0.1 ml of 20 mM imidazole-HCl buffer (pH 7.5), 3 mM KCl, 15% glycerol, 1 mg/ml bovine serum albumin, 0.5 mM DTE, 5 mM MgCl 2 , 3 mM ATP, 60 M [ 14 C]valine (NEN Life Science Products, 50 Ci/mol), and saturating amounts (5-10 M) of partially purified beef liver tRNA (valine acceptance of 325 pmol/A 260 ). Where indicated, the incubation mixture also contained GDP, GTP, or GMP-PNP, with or without EF-1␣. Catalytic amounts of ValRS were added (1-5 nM) after appropriate dilution in 20 mM Tris-HCl (pH 7.5), 10% glycerol, 0.2 mM DTE, and 4 mg/ml bovine serum albumin. All dilutions and incubations were conducted in plastic tubes. The reaction mixture was preincubated at 25°C for 2 min and the aminoacylation reaction was started by addition of tRNA. Incubation was conducted at 25°C. Aliquots were withdrawn at times indicated and immediately quenched in 5% trichloroacetic acid. Precipitated tRNA was collected on GF/C filters (Whatman). Filters were dried and counted in Lipoluma scintillation fluid (Lumac LSC). One unit of activity is the amount of enzyme producing one nmol of aminoacyl-tRNA/min, at 25°C.
Saturation kinetics of Val-tRNA Val synthesis by ValRS (2 nM) from the ValRS⅐EF-1H complex in the presence or in the absence of an excess of free EF-1␣ (500 nM) were obtained in the presence of 100 M GTP in the incubation mixture. Michaelian parameters K m and k cat were determined by nonlinear regression of the theoretical Michaelis equation to the experimental curves using the KaleidaGraph 3.0.4 software (Abelbeck Software).
For measuring the activity of EF-1␣, the guanine nucleotide exchange assay was carried out as described previously (20). The EF-1␣⅐[ 3 H]GDP complex was prepared following incubation of 6 M EF-1␣ with 4 M [ 3 H]GDP (Amersham Pharmacia Biotech; 1500 Ci/mol) in 80 l of 35 mM Tris-HCl (pH 7.5) containing 0.5 mM DTE, 8.6 mM magnesium acetate, 100 mM NH 4 Cl, 1 mg/ml bovine serum albumin, and 18% glycerol for 5 min at 37°C. The reaction mixture was put on ice and diluted by addition of 800 l of ice-cold exchange buffer (20 mM Tris-HCl (pH 7.5), 10 mM magnesium acetate, 50 mM NH 4 Cl, 0.1 mg/ml bovine serum albumin). The exchange reaction was conducted at 0°C after addition of exchange buffer containing nucleotide and specified exchange factors. Aliquots were taken at times indicated and immediately filtered through nitrocellulose filters (Sartorius; pore size 0.45 m). Filters were washed three times with 1 ml of ice-cold washing buffer (20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 100 mM NH 4 Cl, 0.1 mg/ml bovine serum albumin), dried, and counted in Lipoluma scintillation fluid.
To obtain the GTP form of bacterial EF-Tu, purified EF-Tu was incubated with 100 M GTP in an incubation mixture containing 25 mM Tris-HCl (pH 7.5), 50 mM NH 4 Cl, 10 mM MgCl 2 , 1 mM DTE, 0.5 mM EDTA, in the presence of pyruvate kinase (30 g/ml) and phosphoenolpyruvate (1 mM) to remove traces of GDP. Incubation was conducted at 30°C for 15 min, and the EF-Tu⅐GTP preparation was used immediately.

EF-1␣⅐GTP Stimulates ValRS Activity in the ValRS⅐EF-1H
Complex-Earlier studies suggested that tRNA aminoacylation catalyzed by ValRS is independent of its association with EF-1␣. Indeed, the intrinsic specific activity of the enzyme, determined in the aminoacylation reaction, was not affected following its dissociation from the EF-1H complex (15,19). However, whereas the EF-1␤, -␥, and -␦ subunits are tightly bound to ValRS, the EF-1␣ subunit can be easily depleted from the ValRS⅐EF-1H complex (21). This finding suggested that the ternary complex EF-1␣⅐GTP⅐Val-tRNA Val is readily dissociated from the ValRS⅐EF-1H complex following completion of a single aminoacylation cycle. We reasoned that if valylation of tRNA Val by ValRS is controlled by EF-1␣, this effect should be detected only in the presence of an excess of free EF-1␣. If ValRS and EF-1␣ activities are coupled, as expected if tRNA is channeled from the synthetase to the elongation factor, addition in the reaction mixture of free EF-1␣ in excess, that could be effectively transformed into its EF-1␣⅐GTP species by the ␤, ␥, and ␦ subunits of EF-1H in the presence of GTP, should result in an enhanced rate of Val-tRNA Val formation. When all EF-1␣ is transformed to a ternary complex with the aminoacylated tRNA, this stimulation should ceased.
Therefore, we devised an aminoacylation assay designated to test a putative effect of EF-1␣⅐GTP on the valylation efficiency catalyzed by the ValRS component of the ValRS⅐EF-1H complex. In the assay procedure used in our study, the reaction mixture containing the ValRS⅐EF-1H complex and free EF-1␣ in excess, but deprived of tRNA, was incubated 2 min at 25°C before the aminoacylation reaction was started by addition of the tRNA substrate. This preincubation was necessary to avoid a lag in the time course experiments described below. Presumably, EF-1H is preloaded with EF-1␣⅐GTP during this initial incubation. When catalytic amounts of the ValRS⅐EF-1H complex (1.5 nM) is incubated in the presence of an excess of free EF-1␣ (500 nM) and saturating amounts of the substrates for the aminoacylation reaction, in the presence of GTP to produce the active species EF-1␣⅐GTP, the rate of valyl-tRNA Val synthesis was 2-fold increased, raising from 1.2 pmol/min to 2.1 pmol/min (Fig. 1A). GTP alone had no effect on the rate of Val-tRNA Val formation. This increased aminoacylation rate was especially observed during the first minutes of the incubation, when the free form of EF-1␣ is still in excess. As exemplified in Fig. 1B with a higher amount of the ValRS⅐EF-1H complex (3 nM), when consumption of free EF-1␣ is faster, the time course of valyl-tRNA Val formation is clearly biphasic. The initial rate in the presence of EF-1␣⅐GTP (4.7 pmol of valyl-tRNA Val formed/min) is 2-fold that observed in the absence of EF-1␣ (2.4 pmol/min). After the reaction proceeded for about 9 min, producing 45 pmol of valyl-tRNA Val , corresponding approximately to the amount of EF-1␣ added in the incubation mixture (50 pmol), the time course of aminoacylation returned to its control value determined in the absence of EF-1␣. As shown in the inset of Fig. 1A, in the presence of GTP the duration of the stimulation and the apparent initial rate of valyl-tRNA Val synthesis are functions of the amount of free EF-1␣ added in the incubation mixture.
EF-1␣ Stimulation of ValRS Activity Is GTP-dependent-The GTP form of elongation factor EF-1␣, and therefore its ability to contribute a ternary complex EF-1␣⅐GTP⅐Val-tRNA-Val , is absolutely required to stimulate the aminoacylation activity of ValRS. Indeed, in the presence of GDP instead of GTP, the rate of valylation of tRNA Val by the ValRS⅐EF-1H complex is not affected by the presence of a large excess (200-fold) of free EF-1␣ (Fig. 2). The nonhydrolyzable GTP analogue GMP-PNP did produce a stimulation of ValRS activity in the presence of EF-1␣, albeit to a lesser extent, as compared with GTP (Fig. 2). This lower efficiency could be due to a slower rate of GDP/GMP-PNP exchange, as compared with the GDP/GTP exchange, leading to a decreased rate of dissociation of EF-1␣⅐GMP-PNP⅐Val-tRNA Val from the ValRS⅐EF-1H complex, as compared with the regular ternary complex EF-1␣⅐GTP⅐Val-tRNA-Val . Accordingly, the rate of dissociation of GDP from the EF-1␣⅐[ 3 H]GDP complex in the presence of an excess of free GMP-PNP is not as fast as that observed in the presence of GTP (Fig.  3). Similarly, EF-1␣⅐GMP-PCP has a stronger affinity for the EF-1␤␥ subunits than EF-1␣⅐GTP, thus slowing down the exchange of GDP from EF-1␣⅐GDP (22). Alternatively, the inability of GMP-PNP to mimic GTP for the stimulation of ValRS by EF-1␣ also suggests that GMP-PNP, from which P i cannot be released, has not the potential of GTP to produce a conformational change in the ValRS⅐EF-1H complex that would favor the dissociation of the ternary complex EF-1␣⅐GTP⅐Val-tRNA Val .
Bacterial EF-Tu Cannot Replace Eukaryotic EF-1␣-The simplest mechanism that might account for the observed EF-1␣⅐GTP-dependent stimulation of ValRS activity would merely imply that the Val-tRNA Val formed is sequestered by EF-1␣⅐GTP, thereby preventing product inhibition of ValRS activity. We tested this possibility by adding to the aminoacylation mixture the prokaryotic analogue of EF-1␣, EF-Tu, that is known to interact, in its GTP-bound form, with eukaryotic aminoacyl-tRNAs (23).
Since EF-Tu has a much lower affinity for GTP than for GDP (24) and cannot be charged with GTP by the nucleotide exchange subunits of EF-1H (unpublished observation), nucleotide exchange was performed by preincubation of the factor with an excess of GTP in the presence of phosphoenolpyruvate and pyruvate kinase. When EF-1␣⅐GTP was substituted by EF-Tu⅐GTP in the aminoacylation reaction, no stimulatory effect of EF-Tu⅐GTP on the ValRS aminoacylation activity was observed at different concentrations of the factor (0.5-2 M) (Fig. 4). This result suggested that the EF-1␣-induced stimulation of Val-tRNA synthesis catalyzed by the ValRS component of the ValRS⅐EF-1H complex is contributed by a proteinprotein interaction requiring cognate factors from higher eukaryotic origin.

EF-1␣⅐GTP Stimulation Requires a Native ValRS⅐EF-1H
Complex-To test for the requirement of the native ValRS⅐ EF-1H assembly in the EF-1␣⅐GTP stimulation of ValRS activity, two ValRS derivatives that behave as free species were isolated from the complex. It was shown previously that the ability of ValRS to associate with EF-1H is lost upon conversion by elastase of the native enzyme of 140 kDa to a fully active truncated form of 125 kDa (20). The 200-amino acid extension of ValRS lost upon elastase treatment is involved in the assembly of the ValRS⅐EF-1H complex through protein-protein interaction with the ␦ subunit of EF-1H. Furthermore, native ValRS can be dissociated from the EF-1H components in the presence of 0.5 M NaSCN, a chaotropic salt (19). The resulting monomeric ValRS species is isolated without loss of its aminoacylation activity. These two monomeric ValRS species were assayed for their potential to be stimulated by EF-1␣⅐GTP in conditions where the native ValRS⅐EF-1H species does. Neither the truncated free ValRS, nor the native free ValRS, could be activated by the addition of preformed EF-1␣⅐GTP in the incubation mixture (Fig. 5). Further addition of the EF-1␤␥ subunits in the assay had no effect on ValRS activity (Fig. 5). As shown previously, the EF-1␤ and -␥ subunits form a stable complex that efficiently exchanges the nucleotide from EF-1␣⅐GDP but does not associate to the isolated 140-or 125-kDa ValRS species (20). In the absence of the EF-1␦ subunit the ValRS⅐EF-1H complex cannot be reconstituted. On the other hand, the EF-1␦ subunit associates with the native 140-kDa ValRS form (20), but in the absence of the ␤ and ␥ subunits, this nucleotide exchange factor alone did not confer on EF-1␣ the ability to stimulate ValRS activity (not shown). Therefore, the association of ValRS with the ␤, ␥, and ␦ subunits of EF-1H seems to be absolutely required to produce a functional interaction between the synthetase and elongation factor EF-1␣. This result suggests that a proper positioning of EF-1␣⅐GTP and ValRS⅐Val-tRNA Val contributes an essential step for the activation of the synthetase activity. The presence in solution of a competitor protein (EF-1␣⅐GTP or EF-Tu⅐GTP) with a high affinity for the newly synthesized aminoacyl-tRNA cannot per se explain the observed stimulation effect.
In the case of yeast ValRS, for which dissociation of the aminoacyl-tRNA was shown to be the rate-limiting step in the aminoacylation reaction (25), addition of EF-1␣, EF-1␤␥, and GTP proved to have no stimulatory effect (result not shown). Similarly, incubation of the multisynthetase complex from rab- bit containing nine aminoacyl-tRNA synthetases (18) in the presence of EF-1␣⅐GTP did not confer on methionyl-, lysyl-, or aspartyl-tRNA synthetases increased aminoacylation rates (results not shown).

EF-1␣⅐GTP Increases the Apparent k cat of the Aminoacylation Reaction Catalyzed by ValRS-The
Michaelian parameters for tRNA in the aminoacylation reaction catalyzed by ValRS from the ValRS⅐EF-1H complex were determined in the presence or in the absence of EF-1␣⅐GTP in the incubation mixture. Kinetic parameters were determined using 0.05-3 M beef liver tRNA enriched in tRNA Val (325 pmol/A 260 ). The apparent dissociation constant (K m ) for tRNA of 0.33 Ϯ 0.08 M was not affected by the addition of EF-1␣⅐GTP in the aminoacylation reaction. By contrast, the catalytic constant k cat was found to be increased 2-fold, raising from 0.17 Ϯ 0.03 s Ϫ1 to 0.32 Ϯ 0.04 s Ϫ1 in the presence of EF-1␣⅐GTP. Therefore, we suggest that the major effect of the interaction of ValRS⅐EF-1H with EF-1␣⅐GTP is in increasing the turnover number of ValRS, through a conformational change that would promote the release of Val-tRNA Val .

DISCUSSION
The EF-1␣-induced stimulation of ValRS can be detected in the presence of GTP but not of GDP. This finding clearly demonstrates the importance of the ternary complex EF-1␣⅐GTP⅐Val-tRNA Val for the stimulation mechanism. As expected, activation is observed as long as some free EF-1␣⅐GTP is available in the incubation mixture. When all EF-1␣ is converted to EF-1␣⅐GTP⅐Val-tRNA Val , ValRS is no longer stimulated, suggesting that EF-1␣ does not rapidly dissociate from GTP and Val-tRNA Val . Although the intrinsic GTPase activity of EF-1␣ is stimulated by the aminoacyl-tRNA (26), EF-1␣ does not appear to be efficiently recycled in this assay. Accordingly, it is possible to substitute GTP by GMP-PNP. This nonhydrolyzable GTP analogue can be regarded as mimicking the nucleotide substrate in its ground state complex. Therefore, the endogenous GTPase activity of EF-1␣ is not involved in the stimulation of ValRS activity.
Mammalian ValRS is a 1265-amino acid protein (27). As compared with its bacterial counterpart, the human enzyme displays a large NH 2 -terminal polypeptide extension dispensa-ble for its activity (20). In vitro reconstitution experiments have shown that this domain interacts with the ␦ subunit of EF-1H (20). The finding that neither the native enzyme of 140 kDa nor the truncated 125-kDa ValRS species are susceptible to EF-1␣ mediated stimulation also demonstrates that the NH 2 -terminal extension of ValRS is not per se engaged in a functional interaction with EF-1␣. Noteworthy, the NH 2 -terminal polypeptide extension of the multifunctional glutamyl-prolyl-tRNA synthetase, and the p18 auxiliary component of the multisynthetase complex display sequence similarities with this domain (28). This observation led to the suggestion that these polypeptide domains could be involved in the transient anchoring of EF-1␣ to the complex. Though we were unable to detect any effect of EF-1␣⅐GTP on the activity of three components of the multisynthetase complex, namely lysyl-, aspartyl-, and methionyl-tRNA synthetases, we cannot rule out the possibility that a specialized adaptor molecule could be involved to provide a functional interaction between EF-1␣ and those synthetases, a role played by the EF-1␤␥␦ subunits in the case of the ValRS⅐EF-1H complex.
The EF-1␣-induced activation of tRNA Val aminoacylation by ValRS requires the whole ValRS⅐EF-1H complex to occur. Neither bacterial EF-Tu can substitute for mammalian EF-1␣, nor free dissociated ValRS can mimic the ValRS component of the ValRS⅐EF-1H complex. Furthermore, the isolated guanine- nucleotide exchange factors, EF-1␦ or EF-1␤␥, cannot individually sustain the activation of ValRS by EF-1␣. All these results illustrate the direct connection between regulation of ValRS activity and the adequate association of ValRS and EF-1␣ within a macromolecular assemblage of defined composition and structure. The ValRS⅐EF-1H complex should provide a structural support for the functional interaction of ValRS with EF-1␣. A synoptic model is outlined in Fig. 6. ValRS is tightly associated to the ␤, ␥, and ␦ subunits of EF-1H. EF-1␣ is loosely bound to this complex: (i) it is easily dissociated from the other components following chromatography on a Mono Q column (21) or following incubation with aminoacyl-tRNA and GTP (17); (ii) catalytic amount of the ValRS⅐EF-1H complex can efficiently exchange GTP for GDP from a 100-fold molar excess of EF-1␣⅐[ 3 H]GDP (Fig. 3). Therefore, EF-1␣⅐GDP can be recycled into EF-1␣⅐GTP by the exchange factors bound to ValRS even in the absence of synthetase activity, at least in vitro (Fig.  6, top right). In parallel, the ValRS component of the complex is able to catalyze tRNA Val aminoacylation even in the absence of GTP or of an excess of free EF-1␣, that is when the elongation factors are not functioning (Fig. 6, top left). However, the concomitant functioning of ValRS and of the EF-1 subunits is accompanied by a 2-fold increase in the valylation rate. Since the k cat parameter is primarily affected, the EF-1␣⅐GTP facilitated ValRS activity is most likely the result of a proteinprotein interaction that induces a conformational change in ValRS, and promotes the release of Val-tRNA Val from the enzyme, suggesting that product release is the rate-limiting step. Coupling of these two parallel reactions would favor the direct transfer of the aminoacylated tRNA to the elongation factor, with formation of the ternary complex EF-1␣⅐GTP⅐Val-tRNA Val (Fig. 6, bottom). In vivo, this coupling could be responsible for the channeling of tRNA reported by Deutscher and co-workers (8,10). Using a permeabilized cell system, they observed that exogenously added tRNA or aminoacyl-tRNA are not effective precursors for protein synthesis, suggesting that there is a channeled tRNA cycle in mammalian cells. In the case of the ValRS system, our results suggest that the release of the aminoacylated tRNA from the synthetase could be controlled by the availability of free EF-1␣, therefore providing a rational explanation for channeling. In that connection, it should be stressed that a great deal of data reports that members of the EF-1H and of the ValRS⅐EF-1H complexes are the targets of phosphorylation events (21, 29 -32). Hyperphosphorylation of elongation factor 1␦ is also observed in virus infected cells (33). The functional significance of these modifications is poorly understood but could alter translational efficiency. Whether these postranslational modifications have a role on the EF-1␣induced stimulation of ValRS, and therefore on tRNA channeling in vivo remains to be shown, but could be a means to regulate the efficiency of tRNA delivery for protein synthesis.
FIG. 6. Channeling of tRNA Val in the valyl-tRNA synthetase⅐ elongation factor 1H complex. The ValRS⅐EF-1H complex is a dimer of an elementary core made of equimolar amounts of each subunit. For clarity, only one monomeric core containing one copy of ValRS and of the ␣, ␤, ␥, and ␦ subunits of EF-1 is shown. Two parallel reactions take place on this complex: specific recognition of tRNA Val by the synthetase followed by its aminoacylation by valine and association of EF-1␣⅐GDP to the ␤ and/or ␦ subunits of EF-1H to exchange its nucleotide with GTP. The EF-1␣⅐GTP-dependent stimulation of tRNA Val aminoacylation suggests that these two reactions are concerted mechanisms. Formation of the ternary complex EF-1␣⅐GTP⅐Val-tRNA Val on the ValRS⅐EF-1H complex would be responsible for the vectorial transfer of tRNA from the synthetase to the ribosome.