Tryptophanyl transfer ribonucleic acid synthetase of Escherichia coli. I. Purification of the enzyme and of tryptrophan transfer ribonucleic acid.

Abstract Tryptophanyl-tRNA synthetase of Escherichia coli B was purified 1000-fold by MnCl2 precipitation and column chromatography on DEAE-cellulose, hydroxylapatite, and Amberlite CG-50. The purified enzyme gave one peak of constant specific activity on Sephadex G-200, one protein band on disc gel electrophoresis, and contained no other aminoacyl-tRNA synthetases. The rate of tryptophanyl-tRNA formation was optimal at pH 8.8. The kinetic constants for tryptophan, ATP, and tRNAtrp were determined at pH 7.0 and 8.8, and the turnover number was 1200 at pH 8.8. The amino acid composition is similar to the composition of other aminoacyl-tRNA synthetases, except for a low tryptophan and half-cystine content. tRNAtrp was purified on benzoylated DEAE-cellulose and hydroxylapatite columns. The final product had a specific activity of 1.8 nmoles per A260 unit and an estimated purity of 100%. The calculated ratio of tRNAtrp to tryptophanyl-tRNA synthetase in vivo is 1:1.

The purified enzyme gave one peak of constant specific activity on Sephadex G-200, one protein band on disc gel electrophoresis, and contained no other aminoacyl-tRNA synthetases.
The rate of tryptophanyl-tRNA formation was optimal at pH 8.8. The kinetic constants for tryptophan, ATP, and tRNATrp were determined at pH 7.0 and 8.8, and the turnover number was 1200 at pH 8.8. The amino acid composition is similar to the composition of other aminoacyl-tRNA synthetases, except for a low tryptophan and halfcystine content. tRNATPp was purified on benzoylated DEAE-cellulose and hydroxylapatite columns.
The final product had a specific activity of 1.8 nmoles per As60 unit and an estimated purity of 100%. The calculated ratio of tRNATrP to tryptophanyl-tRNA synthetase in vivo is 1: 1.
Although many aminoacyl-tRNil synthetases and their cognate tRNhs ,,'.vve been purified to homogeneity (l-3), the mechanism involved in the recognition of the specific tRNAs by the aminoacyl-tRNA synthetases is not underst.ood. We have approached that problem with a study of the tryptophanyl-tRNA synthetase (L-tryptophan + ATP + tRNA = L-tryptophanyl-tRNB + AMP + PPi,' EC 6.1. 1.2) and tRNATrp from Escherichia coli because of some unusual aspects of this system.

Although tryptophan
comprises only a small portion of the amino acids in fl. coli protein (4), and tRNATrp is one of the least prevalent of all tRNAs in E. co& the tryptophanyl-tRNA thetase is one of the most active aminoacyl-tRNA synthetases in E. coli extracts (5). Secondly, as the tryptophanyl-tRNA synthetase from beef pancreas (6), the enzyme from E. coli catalyzes the formation of a tryptophanyl-ATP ester (7). This reaction is analogous to the esterification of tryptophan to the terminal adenylic acid residue of tRNA.
Finally, tRNATrp from E. coli exists in active and inactive conformations (8-lo), and only the active conformation is a substrate for tryptophanyl-tRNh synthetase.
That enzyme is of historical interest because it was the first aminoacyl-tRNA synthetase purified to homogeneity (14). We here report the purification and partial characterization of tryptophanyl-tRNA synthetase and the purification of tRNATrP from E. coli. The following communication (15)  Other mat.erials were from sources marned previously (7,10).
lllethods Assay of Tryptophanyl-tRNA Fornzation-The enzymatic act,ivity was measured by the rate of tryptophanyl-tRNA Eormat,ion at 37". The final 0.5.ml reaction mixture contained 100 mnf potassium Bicine buffer, pH 8.8, 5.0 mM MgCl,, 4.0 mM reduced glutsthione, 2.5 mM chloroquine, 1.0 rnhx ATP, l-2 pM tRNATrp (45 to 50 AzeO units of tRNA), and 0.10 mM L- [3-14C]tryptophan, 4 to 8 cpm per pmole. After these components in 0.45 to 0.49 ml were incubated at 37" for 5 min, 0.05 to 0.2 unit of enzyme was added in 0.010 to 0.050 ml of 10 mM potassium phosphate buffer, pH 6.9, containing 20 mM 2-mercaptoethanol and 10% glycerol, and the complete mixture was incubated for 2 to 10 min at 37". The reaction was stopped by the addition of 1.3 ml of 12: 1 (v/v) cold 95'$$ ethanol-2 M sodium acetate buffer, pH 5, and the tryptophanyl-tRNA formed was separated from free tryptophan by centrifugation, solution in buffer, precipitation with HCl, and collection on glass fiber filters as previously described (5). Alternatively, the reaction was stopped by the addition of 2 ml of 12 : 1 (v/v) cold 95 y0 ethanol-2 M sodium acetate buffer, pH 5. The tubes were placed at 0". The precipitates were collected on dry GF/C glass fiber filters, washed with 2 ml of 2: 1 (v/v) 95% ethanol-O.2 M sodium acetate buffer, pH 5, and then washed five times with 3 ml of 2 s IICl.
In both methods the filters were dried and counted in a liquid scintillation counter. The two procedures for measuring tryptophanyl-tRNA give identical results.
Under these assay conditions the amount of tryptophanyl-tRNA formed is proportional to time and t.o the amount of enzyme added.
One unit of enzyme forms 1 nmole of tryptophanyl-tRNA in 10 min at 37". The concentration of tRNATrp was measured by extent of tryptophanyl-tRNA formation under the same conditions but with limiting amounts of tRNATrQ and excess enzyme.
Therefore, during the routine 5-min assay at pH 8.8, only a negligible amount of the total tryptophanyl-tRNA formed was hydrolyzed.
At 0" and pH 4.8 in 100 mM potassium acetate buffer, the tryptophanyl-tRNA was stable for at least 2 hours. Other Assays-Alkaline phosphatasc was assayed according to Garen and Levinthal (21).
Protein was measured by the method of Lowry et al. (22), with bovine serum albumin as a standard, or by nitrogen determination.
Nitrogen wa.s determined with a calorimetric method similar to that described by Fels and Veatch (23), as modified by Strid (24). Ammonium sulfate (4 to 10 pg nitrogen) or 20 to 100 pg of protein were placed in a conical centrifuge tube, and 0.100 ml of 50% HzS04 and 0.025 ml of 30y0 II202 were added to each tube in that order. The tubes were placed in hot sand, and the &SO4 was allowed to reflux for 3 hours after the I-120 evaporated.
After the tubes were cooled to room temperature, 6 ml of 1.33 M sodium acetate buffer, pH 5.5, were added to each tube. Three I-ml aliquots were placed in 20.ml screw-cap culture tubes, and aft,er 1 ml of ninhydrin solution (1 g of ninhydrin and 0.15 g of hydrindantin in 37.5 ml of peroxide-free methylcellosolve) was added to each tube, the mixtures were heated at 100" for 15 min. A stream of O2 was bubblrd through each cooled reaction mixture for a few seconds to oxidize residual hydrindantin (25), and 5 ml of 50% ethanol was added. The absorbance at 570 nm was measured, and the nitrogen content of t,he protein was detrrmincd l)y comparison with the ammonium sulfate standards.
I~orinc serum albumin was found to contain 16.1 y0 nitrogen, corresponding exactly with the known value.
The concentration of tryptophanyl-tRNA synthetase was determined on the assumption that the enzyme is 16% nitrogen. For pure tryptophanyl-tRNA synthetase the nitrogen method gave a value 91% of t#hat determined by the method of Lowry et al. (22). Purification-The purification of the enzyme extends and modifies that previously published (7) and is summarized in Table I. All steps were at O--2". All phosphate buffers were pH 6.9, made by mixing equimolnr amounts of KH2P04 and KzHPO4.
Frozen E. coli cells (950 g) were suspended in a final volume of 3 liters by addition of 1 IYLM phosphate buffer containing 10 rnM MgC12, 20 mM 2-mercaptoethanol, and 10% glycerol. The cells were thawed and passed through a French press at 10,000 p.s.i. The suspension was centrifuged at 16,000 x g for 2 hours, and the turbid, supernatant fluid (2185 ml) was collect,ed by decantation.
The extract was placid in a 4-liter flask, and 109 ml of 1 M MnCl% (26) were added dropwise with stirring.
After addition of the MnClz the suspension was allowed to stand for 5 hours and then centrifuged as above. The cloudy, supernatant solution was dialyzed 24 hours against 12 liters of 5 mM phosphate buffer containing 20 mM 2-mercaptoethanol, 10% glycerol, and 15aj, polyethylene glycol 6000. The content of the dialysis sac was centrifuged as before; 500 ml of turbid supernatant fluid were collected and diluted to 1600 ml with Solution A (20 mM 2-mercaptoethanol, 10% glycerol), in order to decrease the conductiiity to that of 10 mM phosphate buffer or less. In a stepwise gradient from 0 to 707, glycerol a pH gradient was established with 1% ampholyte, pH range 5 to 8, at 0" by operation of the LKB model 8101 apparatus at 360 to 1270 volts for 28 hours. The higher glycerol concentrat,ion was at the higher pH. Then 330 units of hydroxylapatite fraction in 0.050 ml of 3 mM potassium phosphate buffer, pH 6.9, 10 rnM 2-mercaptoethanol, 50% glycerol were placed at the 50% glycerol level (pH N 8) and focused for 17 hours at 0", 900 volts. The pH (0) was measured directly on 2-ml fractions at 0". The activity (a) of tryptophanyl-tRNA synthetase was measured as under "Methods." Recovery was 537, of input activity.
The 1600 ml were then placed over a DE52 column, 4 x 24 cm, previously equilibrated with 10 mM phosphate buffer in Solution A. The column was washed with 1300 ml of the buffer in Solution A, then developed at 60 ml per hour by a linear gradient from 10 to 100 mM phosphate buffer in 4000 ml of Solution A. The enzyme emerged as a sharp, symmetrical peak centered at a buffer concentration of 70 mM. Fractions comprising the peak were pooled and applied directly to a hydroxylapatite column, 2.5 x 30 cm, equilibrated with 100 mM phosphate buffer in Solution A. The column was washed with 450 ml of the buffer in Solution A, then developed at 70 ml per hour by a linear gradient from 100 to 200 mM phosphate buffer in 2000 ml of Solution A. The fractions containing the enzyme, centered at a buffer concentration of 150 mM, were pooled and dialyzed against 4000 ml of 10 mM phosphate buffer and 30 To polyethylene glycol6000 in Solution A. The content of the dialysis sacs was adsorbed on a Amberlite CG-50 column, 1.5 X 15 cm, equilibrated with 10 mM phosphate buffer in Solution A. After a wash with 50 ml of the buffer in Solution A the column was developed at 50 ml per hour by a linear gradient from 10 to 300 mM phosphate buffer in 500 ml of Solution A. The enzyme appeared in a single, symmetrical peak centered at a buffer concentration of 130 mM.
The purified enzyme was concentrated to 2 mg per ml by dialysis as before and stored at 2" in 20 mM phosphate buffer in Solution A.
Notes on Pur$cation--Streptomycin sulfate treatment of the extract (1.3 g/62 ml) allowed recovery of 95 y. of the activity but removed less nucleic acid than did the MnCla precipitation. Autolysis for 5 hours at 37" destroyed 67% of the extract activity. Fractional precipitation of the MnClz fraction with polyethylene glycol 6000 or with ammonium sulfate provided little or no purification.
Another purification from 40 g of cells produced a homogeneous enzyme after only a 350-fold purification.
The difference is ex- Polyacrylamide disc gel electrophoresis of the Amberlite fraction.
The enzyme, 10 rg (right), 40 pg (renter), and 100 pg (left), was subjected to disc gel electrophoresis (28) in the system described by the Canalco brochure (April, 1965) with the st-andard 7.Ooj, polyacrylamide gel, stacking at pH 8.9 and running at pH 9.5. Tris-glycine buffer, pH 8.5, containing bromphenol blue was at the anode and cathode. At room temperature the gels, 6.3 X 0.5 cm, were subjected to 3 ma per gel in a Canalco model 12 apparatus until the dye front reached the bottom of the gels. The gels were stained with 0.1% Amido Schwarz and washed with iOyo acetic acid. plainable, since a different batch of cells was used in each preparation, and the specific activities of the extracts differed by a factor of 4.
Isoelectric Focusing-As shown in Fig. 1 the enzyme has a pI of 6.2. The tryptophanyl-tRNA synthetase purified from human lymphocytes has a p1 of 5.2 (27). The chromatographic properties of the two enzymes on DEAE-cellulose and on Amberlite CG-50 are in direct agreement with the p1 values.
Polyacrylamiak Disc Gel Electrophoresis-By disc gel electrophoresis the DE52 fraction exhibited at least 13 protein bands, and the hydroxylapatite fraction contained three bands, one major and two minor.
The Amberlite fraction migrated as a single protein band (Fig. 2) corresponding with the major band of the hydroxylapatite fraction. Whereas 10 pg gave a clearly visible band, even 240 pg of protein (not shown) gave no trace of a second band.
Moreover, the Amberlite fraction migrated as a single band under denaturing conditions (15). During electrophoresis of the hydroxylapatite fraction at 2" with 10% glycerol and 20 mM 2mercaptoethanol in the buffers all of the tryptophanyl-tRNA synthetase recovered by crushing l-mm gel slices in enzyme dilution buffer (see "Methods") migrated with the major protein band.
Gel Filtration on Sephadex G-200-A sample of the Amberlite fraction was subjected to gel filtration on a Sephadex G-200 column (29). The enzyme emerged in a single, symmetrical peak with constant specific activity, equal to that of the input.  synthetase cochromatographed with E. coli alkaline phosphatase, molecular weight 74,000 (29). This apparent2 molecular weight was confirmed on calibrated Sephadex G-100 and G-150 columns (15).
Absence of Other Amino Acid-activating Enzymes--The absence of other amino acid-activating enzymes in the purified enzyme was determined by measuring the stimulation of ATP-Pl'i exchange by the other 19 coded amino acids (Table II).
In the hydroxylapatite fraction the combined activities of all 19 other amino acid-activating enzymes added up to 1% of the activity of tryptophan-activating enzyme, and in the Amberlite fraction the sum was 0.6 To.
Optimal Assay Conditions-The rate of tryptophanyl-tRNA formation as a function of pH was determined in 100 mM potassium Bicine buffer and 100 mM potassium cacodylate buffer (Fig.  3). The optimal p1-I for tryptophanyl-tRNA formation was 8.8, the same as that for tryptophan-dependent ATP-PP; exchange. Sodium cacodylate, sodium Bicine, potassium Taps, and Tris-HCl buffers (100 mM) did not inhibit tryptophanyl-tRNA formation in their buffering ranges. Potassium phosphate buffer, 100 mM, inhibited the enzyme at all pH values tested between 5.6 and 7.3.
The optimal concentration of potassium Bicine buffer, pH 8.8, for tryptophanyl-tRNA formation was 50 to 100 mM. The optimal Mg2f concentration was 5 mM in the presence of 1 mM ATP and 100 rnM potassium Bicine buffer, pH 8.0. No activity was observed in the absence of Mg2f.
The optimal ATP concentration was 1 m&l; 3 mM ATP decreased the rate by 40%, and 6 111M ATP by 80%. The enzyme synthesizes tryptophanyl-ATP ester in 6 mM ATP (7), and that 2 In this and the following paper the word "apparent" in connection with molecular weights determined by gel filtration of native proteins indicates our cognizance of the limitations of the method. Whole tRNA (50 A2c0 units, 10 to 20 pmoles of tRNATrp per A2G0 unit) and 2.5 mM chloroquine were routinely used in the enzyme assay (7). In the presence of 2.5 rnM chloroquine, tRNATlp is totally in the active form at pH 7 (10). To be certain that chloroquine has the same effect on tRNATrp at pH 8.8 we measured the acceptor capacity of tRNATrp for [14C]tryptophan with and without chloroquine in sodium cacodylate buffer, pH 7, and potassium Bicine buffer, pH 8.8. The specific acceptor activity was 20 pmoles of tryptophan per A260 unit of tRNA in the presence of chloroquine in both systems. However, in the absence of chloroquine, the specific activity was 6.0 prnoles per n2eo unit at pH 7.0 and 9.5 pmoles per AS,, unit at pH 8.8. The result suggests that at pH 8.8 tRNATTp was partially activated during the 30-min incubation at 37". Chloroquine has pK values (30) of 8.1 (quinoline ring) and 10.2 (side chain tertiary amine).
The pK of the ring of chloroquine bound to DNA is higher than 8.1 (18). Therefore, the chloroquine in the chloroquine-tRNATrp complex responsible for tRNATrp activation is probably diralent. Experiments with chloroquine analogs (31), for example 7-chloro-4-aminoquinoline, indicate the tertiary ammonium group on the side chain of chloroquine is not needed for activation of tRNATTp.
Icinetics-The apparent K, values for tryptophan, -kTP, and tRNATrp in the charging reaction were determined as described in Fig. 4 and Table III. The 4-fold difference in V,nv,,, at the two pH values agrees with the results seen on the pH curve (Fig. 3), which shows a 4-fold difference in the rate at pH 7.0 and 8.8. The apparent K, values for tryptophan, -\TP, and tRNATrp are typical of those reported for other aminoncyLtRNA synthetases.
Turnover Number-In addition to the experiments presented in Table III and Fig. 4  lose, being the first of all aminoacyl-tRNA synthetases of IZ'. coli to emerge from columns developed by potassium phosphate buffer gradients at pH 6.9. Conversely, the enzyme binds strongly to Bmberlite CG-50, an unusual property aiding purification.
The half-cyst'ine content was not determined accurately, although the value was less than 7 residues per mole. Most other aminoacyl-tRNA synthetases have 8 to 15 half-cystine residues per mole. Stability of &zyme-During assay at 37" as under "Methods" but with 100 m&I potassium Bicine buffer, pH 8.2, the enzyme is stable for at least 20 min, according to the method of Selwyn (46). Thus, the same amount of tryptophanyl-tRN-4 was formed by 0.04 unit of enzyme in 20 min as by 0.5 unit of enzyme in 1.6 min.
Tryptophanyl-tRNA synthetase from E. coli is stabilized by glycerol (7). The hydroxylapatite and Amberlite fractions stored at 2" in 10 y0 glycerol or at -20" in 50 y0 glycerol, both in 20 mM 2-mercaptoethanol, 20 mM potassium phosphate buffer, pH 6.9, had no detectable loss of activity in 1 year. Dilute solutions of enzyme (12 units   Partial purification of tRNAT'p was achieved by fractionating tRNA on BD-cellulose at 23" (Fig. 5). The 21-fold purified tRNATrp was then charged with L-[**C]tryptophan and rechromatographed on the same column at 0" (Fig. 6). Whereas the tRNirTrp peak had appeared in 5% ethanol (Fig. 5), the tryptophanyl-tRNA appeared in 12% ethanol (Fig. 6). The shift in elution position resulted in a further 3.5.fold purification. About 2 y0 of this material was further purified by hydroxylapatite column chromatography (Fig. 7). The results are summarized in   and reversed phase chromatography and discussed evidence for only one tRNATrp gene. The five isoacceptors resolved by gradient partition chromatography may differ only in base modifications not affecting acceptor activity or chromatographic characteristics in other systems, In E. coli tRNATrp and other tRNAs recognizing codons beginning with U contain cytokinins (54). The presence of a lipophilic cytokinin, however, is probably not responsible for the high affinity of tRNAT'p for BD-cellulose (55). State of tRNATrp and Tryptophanyl-tRNA Synthetase in Vivo-A single E. coli cell (approximately 0.3 x lop6 pg of protein) contains about 6000 molecules of phenylalanyl-tRNA synthetsse (36). C'alendar and Berg (33) calculated that there are approximately 1400 molecules of tyrosyl-tRNA synthetase and 1000 molecules of tRNATy' per cell of E. coli. Thus the concentration of tyrosyl-tRNA synthetase in E. coli is roughly 2 x 10d6 M. DeLorenzo and Ames (56) found that in Xalmonella typhimurium tRNAHi" is 2.3 X 10V6 M and histidyl-tRNA synthetase is 2.2 X 10m6 M. Because the K, for tRNAHis is 1.1 X lo-' M, histidyl-tRNA synthetase and tRNAHie may exist as a complex in vivo. If the tryptophanyl-tRNd synthetase isolated in the cell extract represents the total content of this enzyme in the cells, we can calculate with the assumptions used by Calendar and Berg (33) that there are 500 to 1100 molecules of enzyme per cell. Similarly, the tRNA Trp content is 800 molecules per cell. These concentrations, both near lOA M, and the K, of 3 X 10V7 M suggest that the enzyme and tRNATrp exist as a complex in viva, as already indicated for the histidine system (56).