The role of divalent cations in the reactions of valyl transfer ribonucleic acid synthetase of Escherichia coli. Effects of spermine and ethylenediaminetetraacetate.

We have analyzed the function of spermine in the aminoacylation of tRNA-Val by the valyl-tRNA synthetase of Escherichia coli. Our results indicate that Mg2+ is required for the aminoacylation reaction as well as for the ATP-PP-i exchange catalyzed by this enzyme. The apparent stimulation by spermine is a function of the tRNA used, which appears to contain bound cations even after dialysis against 10 minus 4 M EDTA. Higher concentrations of EDTA totally abolish spermine-stimulated esterification of tRNA-Val.


SUMMARY
We have analyzed the function of spermine in the aminoacylation of tRNA Val by the valyl-tRNA synthetase of Escherichia coli. Our results indicate that Mg2+ is required for the aminoacylation reaction as well as for the ATP-PPi exchange catalyzed by this enzyme. The apparent stimulation by spermine is a function of the tRNA used, which appears to contain bound cations even after dialysis against lop4 M EDTA. Higher concentrations of EDTA totally abolish sperminestimulated esterification of tRNAVs'.
The usual representation of the aminoacylation of transfer RNA is through the intermediate formation of an aminoacyladenylate. Various workers have argued against the formation of such an intermediate. Loftfield and Eigner (1,2) have presented evidence in favor of a concerted mechanism in which an aminoacyladenylate does not participate.
Subsequently, evidence has been presented from several laboratories favoring the concerted reaction mechanism. Using spermine, spermidine, putrescine, and other polyamines, Igarashi et al. (3) observed with most of the amino acid-activating enzymes of Escherichia coli that several of these compounds can stimulate the over-all esterification reaction but fail to support ATP-PPi exchange. This observation was supported by observations by Pastuszyn and Loftfield (4) with purified valyl-tRNA synthetase and by Kayne and Cohn (5) with isoleucyl-tRNA synthetase. Therefore, it was argued that the physiological mechanism of these enzymes does not include the partial reaction necessary for ATP-PPi exchange and that the over-all reaction is concerted, not stepwise. Because independent studies in this laboratory (Midelfort et al. (6)) convinced us that a concerted mechanism does not apply to aminoacyl-tRNA synthetases, we have examined the evidence advanced by others. Crude Escherichia coli B transfer tRNA was purchased from Schwarz. Valyl-tRNA synthetase and tRNAv*l were purified according to the procedures of Yaniv and Gros (7). The enzyme was purified 500-fold from the crude extract and tRNAVnl had an acceptance of 1.3 nmol per absorbance at 260 nm. Spermine was purchased from Calbiochem. Chelex 100 was obtained from Bio-Rad.
Atomic emission and atomic absorption spectroscopy were carried out in a Perkin Elmer 219 atomic absorption spectrometer and in an ARL spectrographic analyzer, respectively. The formation of valyl-tRNAval and ATP-32PPi exchange were measured according to Yaniv and Gros (7). The concentrations of EDTA, MgC12, and spermine added were varied as indicated in the figures. The incubations were run at 37" for 10 min; the reaction was stopped with 10% cold trichloroacetic acid and the precipitate was col&cted on GF/A filter paper, dried, and counted in a toluene scintillation cocktail. The formation of 132P1ATP was determined by measuring the charcoal-absorbable radiiactivity in a gas flow counter.

RESULTS
LVagnesium Zon Requirement-All of the known amino acidactivating enzymes require divalent cations and are generally studied in the presence of Mg2f. Before examining the effects of polyamines, we studied the dependence upon Mgz+ concentration of both the partial reaction of ATP-PPi exchange, and the net forward reaction, by following the formation of aminoacyl-tRNA. This dependence is shown in Fig. 1. The maximum rates of reaction are observed at different concentrations of added Mg2f.
The difference in response to Mg2+ concentration could indicate different functions for the cation in the two systems or a difference in the effective concentrations in the two reaction mixtures. An indication that the latter consideration may be relevant is seen in the portion of the curve representing low concentrations of metal ion, which shows that the over-all reaction levels off at about 10% of the maximum rate, whereas the exchange reaction drops essentially to zero at lo+ M Mg2+. This suggests that sufficient metal ion to support the over-all reaction is associated with the tRNA that must be included only in the esterification system. The probability that tRNA carries metal M9"lq Molar Cont.
[32P]ATP adsorbed on charcoal and counted with a gas flow counter. The reaction mixture for aminoacylation contained Tris-HCl, ATP, mercaptoethanol, NH&l, and bovine serum albumin for the same concentrations as described above ~111s 10 to 20 pg of tRNAVai, 0.2pmol of [W]valine, and 0.5 unit of enzyme per ml of assay. Incubations were stopped with 10% cold trichloroacetic acid; 0.5 mg of carrier RNA was added for complete precipitation.
The concentrations of Mg*+ ion added were varied as indicated in the figure. C--C, [Wlvalyl-tRNAvai formation; zero on the abscissa represents no added Mg2+.
ions is increased by our observation that the background rate of reaction (in the absence of added &Ig2+) is greater when unfractionated tRNA was used instead of purified tRNAV"i. High background reaction rates were also observed by Igarashi et al. Effects of Spermine and EDNA-To eliminate the effects of contaminating metal ions, Pastuszyn and Loftfield (4) added 10F4 M EDTA to their reaction mixtures; this decreased the observed rates. We first examined the effect of spermine alone; in our reaction mixtures, as shown in Fig. 2, spermine causes a small but consistent stimulation over a broad range of concentrations. However, when EDTA was added in the presence of 1OW M spermine, the reaction was inhibited as shown in Fig. 3. The complete inhibition suggests that the reaction, even in the presence of spermine, depends upon a cation that forms a complex with EDTA. Addition of more spermine does not reverse the inhibition.
An objection to the use of EDTA to demonstrate an obligatory role for metal ions is the hypothetical inhibition of the enzyme  1O70 that of EDTA at this pH (8). The data of Fig. 4 indicate that the rate of aminoacylation of tRNAVal 1s proportional to t.he concentration of MgATP. The absolute activity of valyl-tRNA synthetase at pH 7 is lower than at pH 8. However, at pH 7.0, the relative binding of Mgz+ to ATP is greater and the relative rate of aminoacylation observed in the presence of EDTA is correspondingly increased. Because the concentration of free EDTA is greater at pH 7 than at pH 8, these results show that the reaction in the presence of EDTA is that expected from the amount of MgATP present and that excess free EDTA does not inhibit the reaction.
Metal Ions Bound to &VA-The experiments described above indicated that our tRNA preparations contained activating cations. This appeared to be true even though precautions were taken to remove divalent ion contamination.
All reagents were prepared with glass-distilled water, the enzyme was dialyzed against 1OW M EDTA, and tRNAVa' was treated with 1 mM EDTA and then dialyzed against lo-' M EDTA. Analysis of reagents by atomic emission and atomic adsorption spectroscopy indicated all reagents including spermine contain much less than 10-S M RIg2+. But the EDTA-dialyzed tRNAV"' had small amounts of Mg2+ (--1OW M) along with other cations, Cazf, Mn*+, Znzf, and Co2+. The absolute amounts of these contaminant,s were not determined. Analysis from the laboratory of Kayne and Cohn (5) indicated the cation contaminants to be 0.94 g atom of Zn*+, 0.77 g atom of Ca*+, and small amounts of Mg2+, n4n2+, Fe*+, Co*+, NiQf, Cuz+, and Cd2+ per molecule of EDTAdialyzed enzyme.
We attempted to remove these trace amounts of metal ions by passage through Cheiex 100, whieh is known to be a very strong metal binder. Unfortunately, atomic emission spectroscopy indicated that the samples passed through such columns contain in presence of spermine and EDTA Reaction rates in the presence of EDTA and spermine with crude and purified tRNA. Reactions were carried out as described in the legend to Fig. 1 with 3.7 x 10-&M tRNAval or 10m4 M unfractionated Escherichia coli tRNA.
EDTA, MgC&, and spermine were added as shown. Both tRNA preparations were dialyzed against 10e4 M EDTA as described in the text. Dashes indicate no addition of that compound. The introduction of metal by tRNA is seen more easily when crude tRNA is used instead of the purified acceptor for a given amino acid. In Table I are shown rates of esterification of tRNAVal with crude and purified preparations.
The amount of crude tRNA used was 27 times that of the purified material, in order to provide equivalent amounts of valine acceptance. The rate of reaction was approximately 1 order of magnitude greater when crude tRNA was used in the absence of added metal ions. In the presence of lo-" M EDTA, spermine stimulated the crude system to one-third the maximum rate obtained with Mg*+, whereas in the purified system only about 370 of the maximum activity was found in the presence of spermine, and even this reaction was essentially eliminated by 10ea M EDTA.
Studies of Robison and Zimmerman (9) on the cation dependence of the transfer reaction of tRNAPhe from bakers' yeast were interpreted as showing that spermine is more effective t.han h4g2+ in transferring phenylalanine from isolated Phe-AMP-enzyme complex to tRNA Phe although the initial rate of transfer is higher with Mg*+. The amount of spermine required to catalyze this reaction maximally was only 6 PM, in comparison to the 0.3 rnbi ;LIg2+ needed. This indicates a very high binding constant for spermine with transfer RNA. Spermine also affected the intensity of fluorescence more effectively than Mg2+ (9).
These observations led us to study the effect of spermine on the requirement for Mg *+ in the transfer reaction. The data of Fig. 5 show that spermine causes a significant stimulation of the rate of esterification when suboptimal concentrations of Mg*+ were added. The stimulation shotvn for 10e4 M spermine was identical with that obtained with lo+ M spermine. However, as the Mg2+ concentration is increased, the stimulation decreases and the maximum velocity is not changed by the presence or absence of spermine.
PPi Exchange and Pyrophosphorolysis of Val-AMP in Presence of Sperm&e--In the experiments described above spermine FIG. 6 (right). Effect of spermine on pyrophosphorolysis of Val-AMP. The reaction was routinely performed in duplicate and started by addition to an otherwise complete reaction mixture of Val-AMP synthesized and freshly purified on a Dowex 1 column according to Berg (10). The mixture, 0.2m1, of 0.01% bovine serum albumin and 5 mM sodium cacodylate (pH 6.0) also contained 6 J/ml of mercaptoethanol, 0.2 pmol of Na432PZ07 (3.7 to 3.9 x lo5 cpm/pmol), 0.1 to 0.2 pmol of Val-AMP, magnesium acetate, spermine (neutralized to pH 6.6) and EDTA (pH 7.0) as indicated, and enzyme. After 10 min at 20", the reaction was stopped by adding 0.5 ml of ice-cold 0.2 M Na4P20?, 7% HClOd. Subsequently 0.5 ml of a 30 mg/ml acid-washed charcoal suspension was added and the solution passed through a GF/A glass fiber paper disc which waswashed withan additional 50ml of ice-cold water. Radioactivity of discs dried under infrared irradiation was determined in a G-

DISCUSSION
Theoretically, the possible roles of cations in aminoacyl-tRNA synthetase activity include the formation of a complex with ATP, conferring a necessary structure on tRNA and binding to the enzyme. In general, ATP serves as a substrate for all enzymes only as a metal complex, usually with Mg2+, although in many cases other divalent metal cations can be substituted. Holler (11) has shown that the complex of ATP with spermine does not react with the cnzymc. Indications that polyamines can support aminoacyiation of tRNA, thus, led to proposals of radically different mechanisms than had previously been accepted. The finding that all of the stimulation caused by spermine is eliminated by EDTA reduces the strength of these proposals; it is necessary to consider a fundamental role for divalent metal ions even in the presence of spermine.
The source of divalent cations in our system is primarily the tRNA. It is not clear why dialysis against EDTA is relatively ineffective in removing these ions. It can be estimated, however, that the quantity of purified tRNA used in esterification reactions, at least 10-S M, would carry sufficient metal to support a significant rate of reaction. Increased amounts of crude tRNA necessary to provide amino acid acceptance equivalent to the purified species would provide much more metal (6). Since most of this bound metal can be displaced by lop4 M spermine, the formation of metal-ATP can account for all of the stimulation observed on the addition of polyamines.
did not stimulate the ATP-PPi exchange reaction whereas in another study (4) it actually inhibited the residual react,ion still observed in the absence of added divalent metal ions. To explore further the role of spermine in the ATP-PPi exchange reaction it was desirable to study its effect on both the formation of the enzyme-Val-AMP complex from ATP and valine and on the reversal of this reaction. An attempt to study the forward reaction under the conditions of the over-all exchange but excluding the binding of nonreacting substrates met with considerable technical difficulties. On the other hand, pyrophosphorolysis of Val-AMP in the presence of spermine could be followed readily. Due to the instability of Val-AMP in neutral or alkaline solution, our experiments were performed at pH 6.0. As seen in Fig. 6, compiled from several separate experiments, spermine also inhibited the pyrophosphorolysis of Val-AMP. Inhibition was observed also in a similar experiment but in the presence of 1 mM EDTA (not shown). No effect of spermine on the reaction was observed in the presence of a considerably suboptimal or a usual Mg2+ concentration.
The significance of the small but reproducible stimulation in 13 mM Mg is unclear. Together with the observation that pyrophosphorolysis proceeded at an appreciable rate even in the absence of added magnesium ions this pattern of behavior resembles that of the over-all valyl-tRNA synthetase-catalyzed pyrophosphate exchange observed by Pastuszyn and Loftfield (4).

Independent
evidence for a role for polyamines in the transfer of amino acid from an aminoacyl-adenylate-enzyme complex to tRKA has been reported by Robison and Zimmerman (9). They correlated the rate of transfer with changes produced in the physical properties of the tRNA. It should be noted, however, t'hat the rates of transfer reported are approximately 2 orders of magnitude less than that calculated from the turnover of the enzyme used. This discrepancy suggests that their observations did not deal with the rate of transfer but with another property of the system. In an unpublished study' preliminary evidence was obtained for slow attainment of equilibria in the transfer reaction. It is possible that the effects of various cations may be attributed to the displacement of equilibria.
The small differences observed in the concentration dependence of the partial and over-all reactions can be understood in terms of the distribution of cations in the different reaction mixtures. It is clear that a metal ion comples of ATP is a necessary reagent in all cases; it is not yet clear whether metal ions play another role, associating with enzyme or tRKA. Our failure to observe stimulation of ATP-1'Pi exchange reactions by polyamines can be explained by the absence of divalent metal ions from the reaction mixtures. From Pastuszyn and Loftficld's data (4) it appears that spermine actually inhibited this reaction as it inhibited the pyrophosphorolysis of Val-AMP in our study. This could be due either to a decrease in the effective concentration of pyrophosphate or to a direct effect on the enzyme, as observed with the isoleucyl synthetase (9). It appears, therefore, that the residual PPi exchange activity still observed in the presence of spermine (4) is more likely to represent incomplete inhibition (as observed by Holler for isoleucyl-AMP formation (11)) rather than weak catalysis of the fnrward reaction.
The finding that stimulation of aminoacylation of tRNA b3 polyamines is an artifact eliminates the necessity to propose a concerted reaction mechanism for the aminoacyl-tRNA synthetase (2). This finding, however, does not distinguish between ahernative mechanisms and a more positive analysis by kinetic methods has been carried out to establish the role of aminoacyl-AJIP as an intermediate (12). The experiments of others that indicate a role for spermine as a substitute for Xg2+ were carried out with a variety of preparations under various conditions. In view of our results, we must conclude that each of these experiments was carried out under conditions that permitted the artifact described in this article to give a false impression that Mg2+ is not required by the synthetase studied. Since submission of this article we have been informed of the independent studies of Santi and Webster (13) that corroborate our conclusions. Also, Kayne and Cohn," stated that spermidine will not stimulate the aminoacylation of tRNA"? by Escherichia eoEi isoleucyl-tRNA synthetase when the h'Ig2+ concentration is less than 5 PM.