The Monomeric Glutamyl-tRNA Synthetase of Escherichia coli PURIFICATION AND RELATION BETWEEN ITS STRUCTURAL AND CATALYTIC PROPERTIES*

The glutamyl-tRNA synthetase has been purified to homogeneity from Escherichia coli with a yield of about 50%. It is a monomer with a molecular weight of 56,000 and has the same kinetic properties as those of the alpha chain of the dimeric alphabeta-glutamyl-tRNA synthetase described previously (Lapointe, J., and Söll, D. (1972) J. Biol. Chem. 247, 4966-4974). It is the smallest amino-acyl-tRNA synthetase purified from E. coli and contains no important sequence repetition. It is also the only monomeric aminoacyl-tRNA synthetase reported so far to contain no major sequence duplication. Considering its structural and mechanistic similarities with the glutaminyl- and the arginyl-tRNA synthetases of E. coli, we propose the existence of a relation between the true monomeric character of the glutamyl-tRNA synthetase (as opposed to monomers with sequence duplications) and its requirement for tRNA in the activation of glutamate. A single sulfhydryl group of the native enzyme reacts with 5,5'-dithiobis(2-nitrobenzoic acid) causing no loss of enzymatic activity, whereas four such groups per enzyme react in the presence of 4 M guanidine HCl.

The glutamyl-tRNA synthetase has been purified to homogeneity from Escherichia coli with a yield of about 50%. It is a monomer with a molecular weight of 56,000 and has the same kinetic properties as those of the (Y chain of the dimeric cup-glutamyl-tRNA synthetase described previously ( It is the smallest aminoacyl-tRNA synthetase purified from E. coli and contains no important sequence repetition.
It is also the only monomeric aminoacyl-tRNA synthetase reported so far to contain no major sequence duplication.
Considering its structural and mechanistic similarities with the glutaminyl-and the arginyl-tRNA synthetases of E. coli, we propose the existence of a relation between the true monomeric character of the glutamyl-tRNA synthetase (as opposed to monomers with sequence duplications) and its requirement for tRNA in the activation of glutamate.
A single sulfhydryl group of the native enzyme reacts with 5,5'-dithiobis (2-nitrobenzoic acid) causing no loss of enzymatic activity, whereas four such groups per enzyme react in the presence of 4 M guanidine HCl.
According to the theory of the co-evolution of the genetic code and of the amino acids biosynthetic pathways (1,2), glutamate is one of the seven amino acids present in the "paleokaryotes." In this context, the glutamyl-tRNA synthetase is probably one of the "oldest" aminoacyl-tRNA synthetases. This model is supported by the unusual properties of this enzyme (3) and the absence of a glutaminyl-tRNA synthetase in Bacillus subtilis (3,4). A better understanding of its structure and properties might help to unify our view of the structure of the aminoacyl-tRNA synthetases (5)(6)(7) and of the evolution of their structural genes.
The purification of a dimeric form (a/?) of this enzyme from Escherichia coli has been reported previously (8). Following a separation of these two "subunits" by a mild procedure (isoelectric focusing), only the (Y polypeptide (M, = 56,000) can catalyze the formation of Glu-tRNA, while the ,B polypeptide (M, = 46,000) increases the affinity of (Y for glutamate and ATP and its stability (9). In view of the weakness of the interaction between (Y and p (9), we now consider the glutamyl-tRNA synthetase as a monomeric enzyme (a) which can interact and is sometimes co-purified with the p polypeptide. This monomeric enzyme constitutes, with the glutaminyl-tRNA synthetase and the arginyl-tRNA synthetase, a subgroup of aminoacyl-tRNA synthetases sharing the following structural and catalytic properties. They require their cognate tRNA to catalyze the incorporation of ["'PIPPi into ATP, and they are monomeric enzymes of similar molecular weights, respectively, 56,000, 68,000, and 64,000 (8,10,11). The glutamyl-tRNA synthetase appears to a have a strategic position in this family since glutamate is a metabolic precursor to both glutamine and arginine. It is also the smallest monomeric aminoacyl-tRNA synthetase of this subgroup and of all the aminoacyl-tRNA synthetases studied up to now in E. coli. We describe here a new technique for the purification of this monomeric enzyme with a yield of about 50%. We have studied some of its properties and compared them to those of the aP enzyme. We suggest a model for the evolution of the structural genes of the aminoacyl-tRNA synthetases specific for glutamate, glutamine, and arginine and present a correlation between the fact that these three synthetases are monomers, and their requirement for tRNA to catalyze the ATP-PPi exchange. The catalytic mechanism of this monomeric enzyme will be described elsewhere.'

EXPERIMENTAL PROCEDURES
Materials-E. coli MRE-600 was grown in minimal medium (12)  All the operations were performed between 0' and 4°C. All the buffers contained 10% (v/v) glycerol, 20 mM 2-mercaptoethanol, and 0.1 mM PMSF as protective agents against proteases. In the buffer used for cell lysis, 10 mM PMSF was present. The centrifugations were made in a GSA rotor in a Sorvall RC2-B.
Step 1: Cell Lysis-Wet cells (1 kg) were suspended in 2 liters of 10 mM potassium phosphate, pH 8.0, and broken by sonication during 10 min in a Raytheon sonic oscillator (model DFlOl), by fractions of 75 ml. The lysate was centrifuged at 8000 rpm during 30 min to remove cell debris and intact cells, yielding 2350 ml of supernatant.
Step 2: Partition in a Polyethylene Glycol-Dextran Two-Phase System-Potassium phosphate, pH 8.0 (125 ml, 1 M), was added to the supernatant for reasons described previously (8). Then, concentrated solutions of PEG-6000 and dextran T-500 were added to reach the final concentrations of 7 and 1.5%, respectively, in the supernatant. This suspension was mixed during 2 h, and the two phases were separated by centrifugation at 5000 rpm during 20 min. The PEG-rich top phase contains most of the glutamyl-tRNA synthetase activity. Step 4: Chromatography on Hydroxylapatite-Fraction DEAE obtained from the chromatography of the 5 liters of diluted top phase was adsorbed on a hydroxylapatite column (6 X 12 cm) equilibrated against 10 mM potassium phosphate, pH 6.8. The column was washed at 300 ml/h with 200 ml of the same buffer, then with 2 liters of a linear gradient from 20 to 200 mM potassium phosphate, pH 6.8. Two peaks of glutamyl-tRNA synthetase activity were eluted (Fig. 1B); the first, representing only a small percentage of the total activity was eluted at a conductivity of 3.6 mmho (at 4"C), whereas the second and major peak was eluted at 5.7 mmho. Only the most active fractions of this major peak were pooled (Fraction HA) and used in the following purification steps.
Step 5a: Preparative Polyacrylamide Gel Electrophoresis: Last Purification Step-Fraction HA was concentrated 2-to 3-fold by dialysis against 30% polyethylene glycol and then against 0.01 M Tris, 0.077 M glycine (pH 8.3), 20 mM Z-mercaptoethanol, 50% glycerol. About 100 mg of proteins present in 8 ml of this concentrated Fraction HA were mixed with 0.1 ml of a saturated solution of bromphenol blue and layered on the top of a column (I2 cm x 3.4 cm') of polyacrylamide gel whose preparation is described under "Experimental Procedures." A constant current of 40 mA was passed through the gel, whose electric resistance gradually reached a constant value of about 10,000 ohms. The bottom surface of the gel was continuously washed with a 80-ml/h stream of 0.01 M Tris, 0.077 M glycine (pH 8.3), 10% glycerol, and 20 mM P-mercaptoethanol, which was collected in &ml fractions. Following the elution of bromphenol blue (4 h of electrophoresis), the glutamyl-tRNA synthetase was the first protein eluted (Fig. lC), as was observed for the ap enzyme (8). The active fractions were pooled and concentrated (Fraction PAGE).

Alternative Last Purification
Step-Because no more than 100 mg of proteins from Fraction HA could be purified to homogeneity by electrophoresis on the polyacrylamide gel column described above, we replaced this step in certain cases by a chromatography of Fraction HA dialyzed against 10 mM potassium phosphate, pH 7.0, on a column (3.5 X 35 cm) of phosphocellulose (Whatman P-11) equilibrated against the same buffer. After adsorption of the protein sample, the column was washed at a rate of 100 ml/h with 750 ml of this buffer, then with the same buffer containing 0.1 M KCl. Under these conditions, the glutamyl-tRNA synthetase was retarded compared to all the other proteins present in the Fraction HA (Fig. 1D). The fractions containing this enzymatic activity were pooled and concentrated (Fraction Phosphocellulose). Purity of the Glutamyl-tRNA Synthetase Obtained after Steps 5a and 5b-The analysis of the Fraction PAGE by analytical polyacrylamide gel electrophoresis in the absence or presence of a denaturing agent revealed, respectively, the presence of one protein band ( Fig. 2A) and of one polypeptide chain (Fig. 2B), indicating that this fraction contains only the pure enzyme. A similar analysis of Fraction Phosphocellulose shows that the glutamyl-tRNA synthetase represents more than 90% of its protein content (Fig. 2C).
Yield of these Purifications-The results of two purification procedures, using as the last step an electrophoresis on polyacrylamide gel and a chromatography on phosphocellulose, respectively, are summarized in Table I, A and B. For the first purification (Table IA), the cell extract was obtained by sonication of cells grown as described under "Experimental Procedures." The final step (electrophoresis) had to be performed three times because of the low capacity of our column. For the second purification procedure (Table IB), the starting material was a 100,000 x g supernatant (cf "Experimental Procedures").
The specific activity of the glutamyl-tRNA synthetase obtained with the second procedure is slightly superior to that of the Fraction PAGE obtained with the fist About    The values obtained are, respectively, 53,100, 71,100, and 60,800. By electrophoresis of the native enzyme in polyacrylamide gels of various concentrations in the presence of M, markers, we obtained a value of about 58,000 (Fig. 3A).
The reduced and denatured enzyme migrates, during electrophoresis in the presence of SDS (cfi "Experimental Procedures"), as a single polypeptide chain (Fig. 2B) of M, = 56,000 (Fig. 3B). These results indicate that the native enzyme is a monomer.
Amino Acid Composition and Tryptic Map-The amino acid composition of the glutamyl-tRNA synthetase is presented in Table II. A two-dimensional analysis of a tryptic digest of the pure enzyme previously labeled with ['4C]iodoacetate reveals the presence of about 55 peptides (Fig. 4). About 30 peptides contain arginine, 5 contain tryptophan, and 5 react with ["'Cliodoacetate.
The amino acid analysis of these 5 '?-labeled peptides shows the presence of one carboxymethylcysteine per peptide, indicating that only cysteines have reacted with [Wliodoacetate.
These results are in agreement with the amino acid composition: 31 arginines, 26 lysines, 5 tryptophans, and 5 cysteines or half-cystines per enzyme molecule (Table II).
Titration of the Sulfhydryl Groups of the Glutamyl-tRNA Synthetase with DTNB and p-CMB and their Influence on the Enzymatic Activities-When a solution of enzyme was dialyzed twice for 12 h successively against 1000 times its volume of a neutral buffer containing no sulfhydryl group protector, about 50% of its aminoacylation and of its ATP-[32P]PP, exchange activities was lost. The initial activity could be completely recovered by addition of 20 mM 2-mercaptoethanol. On the other hand, dialysis against the same buffer containing no 02 (removed by a stream of N2) did not inactivate it. The titration of sulfhydryl groups was conducted in the absence and in the presence of a denaturing agent. In the absence of a denaturing agent, the reaction of DTNB with the native enzyme (followed by measuring the change of absorbance at 412 nm) is completed after 30 min, when 0.96 sulfhydryl group/enzyme molecule has reacted with DTNB (Fig. 5). The kinetics of this reaction is biphasic. Half of the inhibits both of these catalytic activities in parallel (Fig. 6). In the presence of a stoichiometric concentration of this reagent, 50% of the activity is lost after 2 min, and 80% after 30 min. A 3-fold excess ofp-HMB causes the loss of 85% of the glutamyl-tRNA synthetase activity within 2 min.
In the presence of 5 M urea, the reaction of the glutamyl-  tRNA synthetase with DTNB is over after 30 min, when 2.08 sulfhydryl groups/enzyme have reacted. Here again, the reaction has a faster initial phase, 0.9 sulfhydryl group/enzyme reacting during the 1st min (Fig. 5). In the presence of 4 M guanidine HCl, 4.1 sulfhydryl groups/enzyme react with DTNB within 2 min (Fig. 5).

Comparison
of the Kinetic Properties of the Monomeric and of the &Glutamyl-tRNA Synthetases The K, values of the monomeric enzyme for its substrates were measured at the same pH (7.2) used for the determination of those of the @ enzyme and of the LY polypeptide derived from it (9). As shown in Table   III, the Km of the monomeric enzyme whose purification is described here is very similar to those of (11 obtained by isoelectric focusing of & (8). These two monomeric glutamyl-tRNA synthetases also have a lower affinity for glutamate and ATP than that of the a/~! form. DISCUSSION The Dimeric Glutamyl-tRNA Synthetase-A dimeric structure of the type L$ has been proposed (8) for the glutamyl-tRNA synthetase of E. coli on the basis of the following observations. Following five purification steps, the enzymatic activity co-migrates with one protein band during gel electrophoresis in the absence of a denaturing agent. An analysis of this protein by gel electrophoresis in the presence of SDS shows the presence of approximately equimolar amounts of two polypeptide chains of M, = 56,000 ((Y) and 46,000 (p). Finally, the polypeptide chains (Y and ,L? were separated by isoelectric focusing; (Y, but not ,& had glutamyl-tRNA synthetase activity. The ,f3 chain increased the thermal stability of (Y and its affinity for glutamate and ATP in the aminoacylation reaction (9).
The interactions observed between (Y and p were relatively weak. Indeed, they could be separated by isoelectric focusing, and the @ enzyme sediments as a globular protein of &l, = 60,000 on sucrose gradient (8). This interaction is thus much weaker than that observed between the subunits of other aminoacyl-tRNA synthetases, and is compatible with the regulatory function suggested for ,L? by its in vitro properties (9). Consequently, the properties and the interaction of n and ,L? are such that we now consider (Y as a monomeric glutamyl-tRNA synthetase and ,B as a polypeptide chain which can interact with it. The physiological role of the ,L? protein is still unknown, but its co-purification with (Y is not purely accidental since it does influence its activity (9) and was found in several independent purifications. Moreover, the glutamyl-tRNA synthetase of B. subtilis is a monomer of ikf, = 65,000, and is also weakly associated with a protein of M, =: 46,000 (3). The possibility that the /3 protein is a product of the partial proteolysis of (Y is extremely weak. Indeed, its calculated amino acid composition is very different from that of (Y (Ref. 8, and Table II). Moreover, partially proteolyzed forms of several aminoacyl-tRNA synthetases generally retain some catalytic activity and do not co-purify with the intact form. For instance, the partially proteolyzed form of the valyl-tRNA synthetase from yeast is active and is eluted from hydroxylapatite before the native enzyme (25). In this context, the very minor peak of glutamyl-tRNA synthetase activity eluted from the hydroxylapatite column before the major one ( Fig.  1C) is likely to be due to a partially proteolyzed enzyme. This possibility is strengthened by the fact that this first peak of activity eluted from hydroxylapatite is much smaller when the serine proteases inhibitor PMSF is present during the purification, than in the absence of protease inhibitor (results not shown).
It is conceivable that the association between o and 8 may be due to relatively labile disulfide bond(s), since both polypeptides migrate together during electrophoresis on polyacrylamide gel in the absence of a reducing agent, whereas their interaction was not detected during a sedimentation of the ap enzyme on a H1O:DZO density gradient in the presence of 0.01 M 2-mercaptoethanol (8)." New Purification Procedure of the Monomeric Glutamyl-tRNA Synthetase-The approach described here for the purification of this enzyme has two major differences from the one leading to the afi form (8). First, we used a different linear salt and pH gradient to elute the enzyme from the DEAEcellulose column. Secondly, none of the buffers used for this purification contained MgC12. The enzyme obtained after four purification steps is a single polypeptide chain of M, = 56,000. In the absence of a denaturing agent, it migrates on a gel (Fig.  3) or sediments on a sucrose gradient as a globular protein of M, = 55,000 to 60,000, indicating a structure of the type a. This monomeric enzyme corresponds, by its structural (M, = 56,000), catalytic, and kinetic properties (Ref. 9, and Table  III) to the LY subunit of the c& complex described previously (8,9). While we were working on this project, Willick and Kay (26) reported the purification of the "catalytically active subunit" of the glutamyl-tRNA synthetase using a modification of our initial procedure (8).