Identification of histidine-119 as the target in the site-specific inactivation of ribonuclease A by ferrate ion.

Ferrate ion, a powerful oxidant which is an analog of orthophosphate ion, has previously shown some promise as a site-specific probe of enzymes which interact with phosphate compounds. In order to explore the general applicability of this reagent, it has been tested against ribonuclease A, an enzyme whose structure and active center have been well described. Treatment with a molar ratio of ferrate to enzyme of less than 20 leads to a loss of 87% of the activity. The known competitive inhibitors, 2'-cytidylic acid, inorganic pyrophosphate, and orthophosphate all protect the enzyme from inactivation. Inactivation is accompanied by a loss of the capacity to bind 2'-cytidylic acid. Ferrate inactivation at pH 5.0 is accompanied by the modification of only one amino acid. The amino acid which was identified by amino acid and sequence analyses of peptide fragments obtained by cyanogen bromide treatment and selective proteolysis proved to be histidine-119, whose essential role at the active center has long been established.

Ferrate ion, a powerful oxidant which is an analog of orthophosphate ion, has previously shown some promise as a site-specific probe of enzymes which interact with phosphate compounds.
In order to explore the general applicability of this reagent., it has been tested against ribonuclease A, an enzyme whose structure and active center have been well described.
Treatment with a molar ratio of ferrate to enzyme of less than 20 leads to a loss of 87% of the activity.
The known competitive inhibitors, 2'-cytidylic acid, inorganic pyrophosphate, and orthophosphate all protect the enzyme from inactivation.
Inactivation is accompanied by a loss of the capacity to bind 2'-cytidylic acid. Ferrate inactivation at pH 5.0 is accompanied by the modification of only one amino acid. The amino acid which was identified by amino acid and sequence analyses of peptide fragments obtained by cyanogen bromide treatment and selective proteolysis proved to be hi&dine-119, whose essential role at the active center has long been established. Lee and Benisek (1) showed in a pioneering study that ferrate ion, a powerful oxidant and an analog of phosphate ion, inactivates rabbit muscle phosphorylase b by abolishing its capacity to bind the modulator 5'-AMP. They deduced from the protective action of 5'-AMP that ferrate ion was directed to the phosphate binding site. Further work from Benisek's laboratory has established tyrosine-75 as the amino acid which is oxidized (2). Rajababu and Axelrod (3) found that all phosphatases of a large number tested are inactivated by ferrate ion in a sitespecific manner. Those enzymes tested which do not utilize phosphate were not appreciably affected by the reagent. The present study was undertaken to locate precisely the region modified by the reagent. Pancreatic RNase was an obvious choice due to the detailed information on its structure and function and its ready availability.
Despite amino acid residues 20 and 21. The smaller fragment is the "S-peptide." The larger is the "S-protein" (4).  (16), care must be taken to utilize the optimal pH conditions for the inactivation procedure.
An advantage to using ferrate as a site-specific oxidation probe is that at lower pH ferrate is reduced by HZ0 with the release of O2 (17) and thus there is no residual ferrate ion to oxidize amino acid residues randomly. The importance of pH control in preventing overoxidation became evident when preliminary inactivations were carried out at pH 7.0. Amino acid analyses of ferrate-inactivated RNase A at this pH revealed that approximately 2 tyrosines, 2 lysines, and 1 histidine were destroyed. When  ing, although it did not affect the ability of the enzyme to bind to Cibacron Blue F3G-A. Phosphorylase b lost the ability to bind 5'-AMP following ferrate inactivation (1). To examine the binding of 2'CMP to ferrate-inactivated RNase A the gel filtration method of Hummel and Dreyer (12) was employed. As may be seen in Fig. 2, the ferrate-treated enzyme, unlike the untreated enzyme did not retard the elution of 2'CMP. This result suggests that ferrate is acting at or near the active site of ribonuclease in a site-specific manner. The enzyme and ferrate were incubated for 10 min at 25°C after which an aliquot was removed for enzymatic assay.   (Table III), thus implicating either His-105 or His-119 as the residue oxidized by ferrate.
To establish the identity of the oxidized residue sequential analysis was performed on the tryptic peptide 105-124 which was obtained from the CNBr Fragment 80-124. The sequence analysis was precisely as expected except that His-119 did not appear. Histidine-105 was recovered quantitatively. Thus, it may be concluded that histidine-119 is the amino acid residue which is oxidized by ferrate.
To substantiate this finding, ferrate-oxidized RNase A was hydrolyzed with S. aureus V8 protease in acetate buffer pH 4.0. Under these conditions the enzyme is specific for peptide bonds on the carboxyl side of glutamic acid residues (14). Since the RNase was not subjected to disulfide reduction, one can expect three peptides, l-2, 3-9, and 112-124, which are readily resolved by gel filtration on Sephadex G-50 from the larger peptide 10-111. The individual small peptides which were poorly separated were pooled and analyzed. The amino acid analysis corresponded to the expected mixture. Only 0.22 residues of histidine was found, confirming the above results since the only histidine present in the intact enzyme in Fragments l-2,3-9, and 112-124 is at 119. Effect of Ferrate Treatment on the Histidine Content of RNase as Determined with Diethylpyrocarbonate-Diethylpyrocarbonate is a fairly specific modifying reagent for histidine and is analytically useful in its determination owing to the increase in absorption of the ethoxy-formylated histidine at 240 nm (19). Melchior and Fahrney have used this '%labeled reagent on RNase A and found that only 3 of the 4 residues are modified (20). They attribute their results to the inaccessibility of one of the histidines. Roosemont (21) has recently shown that the inaccessible histidine residues of  native proteins can be made quantitatively reactive with the reagent by recourse to higher ratios of the reagent in the presence of denaturants such as urea or sodium dodecyl sulfate. We confirmed the results of Melchior and Fahrney that native RNase has only 3 accessible histidine residues when tested with the normal concentration of the reagent. The ferrate-incubated enzyme, we found, contained only 2 reactive histidines. When the experiment was repeated with a lo-fold increase in the diethylpyrocarbonate in the presence of urea, the native enzyme showed 4 reactive residues, while the ferrate-treated enzyme showed the expected 3 (Table IV).

DISCUSSION
The present results clearly demonstrate that the inactivation of RNase A by ferrate ion is specifically targeted at the active site. The only amino acid modified in this reaction, His-119, has long been implicated in the active center of this enzyme (see review by Richards and Wyckoff (18) Examples are nucleoside diphosphokinase (28,29), ATP-citrate lyase (30), phosphoglycerate mutase (31) bacterial phosphotransferase systems (32,33), phosphoenolpyruvate synthetase (34), and succinyl-CoA synthetase (35,36).
In previous work in this laboratory, it was shown that all of a large number of phosphatases examined were inactivated by low concentrations of ferrate ion. In those cases examined, the inactivation was found to occur in a site-specific manner, although no effort was made to identify the modified amino acid residues. Recently, Lee and Benesik have reported that tyrosine-75 is oxidized when rabbit muscle phosphorylase b is inactivated by ferrate (2). In this instance the enzyme loses is its activity because its capacity to interact with its activator, 5'-AMP, is lost. We have also found that dehydrogenases utilizing pyridine nucleotide were also inactivated in a sitespecific manner (15). These results encourage the view that ferrate inactivation may be a useful tool in probing active centers of enzyme which utilize phosphates as substrates or effecters.
The utility of ferrate ion for such studies is contingent on the presence of an easily oxidizable amino acid closely positioned to the ferrate ion, which presumably binds at the phosphate site. It is known that the reagent readily oxidizes amines and alcohols (37). The potentially vulnerable amino acid residues include tyrosine, serine, threonine, histidine, tryptophan, methionine, lysine, cysteine, and possibly cystine. In preliminary studies, the three amino acids we have examined, lysine, histidine, and tyrosine, were found to be oxidized when their carboxyl and a-amino groups are protected. Despite the fact that we found but one amino acid was modified in RNase, it must be recognized that had methionine been oxidized to a sulfoxide, its reaction would not have been detected in the amino acid analysis following HCl hydrolysis. Ray and Koshland (38) discovered that under these conditions 85% of the sulfoxide is converted back to methionine.
It was possible, however, to rule out any substantial formation of the sulfoxide on the basis of the CNBr fragmentation pattern and sequence analysis. CNBr cannot cleave at a methionine sulfoxide (39).
Ferrate, unlike other site-specific reagents which consist of a chemically reactive group attached to a substrate-site-seeking group, is both the reactive group and binding group. The preferred site of action is presumably in the area where its effective concentration is high, namely in the binding area. However other vulnerable sites, which abound in proteins, might also be slowly attacked in a nonspecific manner. Fortunately the ferrate ion has only an ephemeral existence, and thus, its time-averaged concentration must be considerably less than the nominal value based on its initial concentration. Experience has shown us that at pH 5.0 nonspecific oxidation is at a minimum. The oxidation-reduction potential of ferrate ion is known to increase with [H'] (16), while the rate of reduction by water increases (17). As a consequence there is little residual ferrate to act nonspecifically on the sites not favored by the concentrating mechanisms of the substrate binding site. Earlier work at pH 7.0 resulted in the conversion of about 2 residues of lysine to lysine &semialdehyde.
Nonspecific oxidation is minimized by working at pH 5.0 and titrating the enzyme by repeated addition of small amounts of ferrate to about 90 to 95% inactivation.
RNase is, of course, among the best described and understood enzymes. The present work was not undertaken to elucidate the behavior of this enzyme but to examine the action and ability of ferrate ion as a site-specific probe for enzymes which utilize phosphate compounds.
Our results encourage the belief that ferrate ion will prove generally useful for the exploration of many of the enzymes belonging to this class, as was suggested by the original work of Lee and Benisek (1).