Ribonuclease Inhibitor from Human Placenta : Interaction with Derivatives of Ribonuclease A *

Several specific modifications, both proteolytic and chemical, have been performed on RNase A. The ability of each of these RNase A derivatives to bind the human placental RNase inhibitor has been determined in competition binding assays. The interaction depends upon the native conformation of the enzyme. Loss of active site residues His-12 and His-119, along with auxiliary residues Lys-7, Phe-120, Asp-121, and Ser-123 in RNase S-protein, des-(121-124)-RNase, and des-(119-124)RNase, demonstrated that these residues are not essential for RNase-inhibitor binding. Carboxymethyl-Hisla-and carboxymethyl-His-119-RNase As interacted 3 times more strongly with the RNase inhibitor than did RNase A. Although ionic interactions between the basic RNase A (p1 = 9.4) and the acidic RNase inhibitor (p1 = 4.7) are involved, modification of the four arginine residues of RNase A by a&diketones does not alter the binding. Complete modification of the lysine residues of RNase A, with retention of positive charge through amidination, decreased but did not abolish the interaction. However, loss of positive charges through carbamylation of lysine residues decreased the interaction by 90% after only 3 lysine residues per molecule of enzyme were modified. Bound RNase inhibitor protected RNase A from inactivation by reagents which are known to inactivate the enzyme by reaction at Lys-41, a residue which appears to participate in the RNase-inhibitor binding and in the resultant inactivation of the enzyme.

Several specific modifications, both proteolytic and chemical, have been performed on RNase A. The ability of each of these RNase A derivatives to bind the human placental RNase inhibitor has been determined in competition binding assays. The interaction depends upon the native conformation of the enzyme. Loss of active site residues His-12 and His-119, along with auxiliary residues Lys-7, Phe-120,  in RNase S-protein, des-(121-124)-RNase, and des-(119-124)-RNase, demonstrated that these residues are not essential for RNase-inhibitor binding.
Carboxymethyl-Hisla-and carboxymethyl-His-119-RNase As interacted 3 times more strongly with the RNase inhibitor than did RNase A.
Although ionic interactions between the basic RNase A (p1 = 9.4) and the acidic RNase inhibitor (p1 = 4.7) are involved, modification of the four arginine residues of RNase A by a&diketones does not alter the binding. Complete modification of the lysine residues of RNase A, with retention of positive charge through amidination, decreased but did not abolish the interaction. However, loss of positive charges through carbamylation of lysine residues decreased the interaction by 90% after only 3 lysine residues per molecule of enzyme were modified.
Bound RNase inhibitor protected RNase A from inactivation by reagents which are known to inactivate the enzyme by reaction at Lys-41, a residue which appears to participate in the RNase-inhibitor binding and in the resultant inactivation of the enzyme.
A number of properties of partially purified RNase A inhibitors from mammalian tissues, in particular from rat liver, have been described (l-4). Primarily as a consequence of the difficulties encountered in the extraction and purification of the protein, there have been very few studies on the structural details of the interaction of the inhibitor with the enzyme. With a crude preparation of the RNase inhibitor from rat liver (a high speed supernatant fraction), Roth and Hurley (5) examined the effects of a number of modifications of RNase A on its interaction with their inhibitor preparation. Previously we have reported the isolation of purified human placental RNase inhibitor and described a number of its properties (6). This protein has a molecular weight near 50,000; it forms a 1:l complex with bovine pancreatic RNase A and is a noncompetitive inhibitor of this enzyme, with a Ki of 3 X 10-l' M. Amino acid analyses indicated that the molecule has 11 disulfide bridges and 8 free -SH groups. The -SH groups are essential for maintenance of the activity of the inhibitor.
In the accompanying communication, procedures have been reported which have simplified the preparation and the assay of the placental inhibitor (7). In this report we examine the interaction of RNase A with human placental RNase inhibitor; two general approaches have been undertaken.
Specific proteolytic cleavage of the enzyme molecule designed to remove parts of the active site were performed to evaluate the contribution of this domain for binding of the RNase inhibitor.
Secondly, specific chemical modifications of groups on the enzyme were performed to assess their role in the inhibitor-RNase interaction.
The derivatives of RNase A were then assayed for inhibitor binding by competition with RNase A in the cyclic 2',3'-CMP assay described in the preceding report (7) Crestfield et al. (11). To 1 mg of RNase A in 0.9 ml of 0.2 M Na acetate buffer, pH 5.5, there was added 0.1 ml of 0.8 M iodoacetate dissolved in 0.2 M Na acetate buffer, adjusted to pH 5.5 with sodium hydroxide.
Alkylation was carried out for 3 h at 37°C. The protein was desalted and lyophilized as described above.

Modification of Arginine Residues-The reaction
with butanedione was performed under the conditions described by Riordan (12). A fresh solution of the reagent was prepared by the addition of 10 ~1 of 2,3-butanedione to 1.0 ml of 0.1 M Na borate buffer, nH 9.0. The resulting solution had a pH of 8.3 and was adjusted to pH 7.8 with HCl. To 1 ma of RNase A in 0.1 ml of 0.1 M Na borate buffer. DH 7.8. IA , there was added 0.9 ml of the reagent solution.
The reaction was carried out overnight at 37°C. The mixture was acidified by the addition of 0.1 ml of 15% acetic acid. Samples of 50 ~1 were removed for amino acid analysis.
Modification of arginine residues with cyclohexanedione was performed according to Patthy and Smith (13,14). To 1 mg of RNase A in 0.1 ml of 0.1 M Na borate buffer, pH 9.0, there was added 0.9 ml of a freshly prepared solution of reagent that contained 10 mg of 1,2cyclohexanedione dissolved in 1.0 ml of the pH 9.0 borate buffer. No release of free lysine was observed upon amino acid analysis of methyl p-hydroxybenzimidate-modified RNase hydrolyzed for up to 96 h in 6 N HCl. No new ninhydrin-positive peak which would correspond to l -p-hydroxybenzimidyllysine or its breakdown products was eluted from the column. The basic and aromatic e-p-hydroxybenzimidyllysine would not be expected to elute from the column of the Durrum D-500 as used under the routine conditions for analysis of protein hydrolysates.

RESULTS
Interaction of Protease-modified RNase A with RNase Inhibitor-A number of RNase A derivatives, derived by specific proteolytic modifications which remove residues at the NH2-or COOH-terminal regions of the enzyme, were studied in order to examine the contribution of the residues to the binding of RNase inhibitor.
RNase A is selectively cleaved after alanine-by subtilisin (25, 26) to yield RNase S, which can be dissociated into the NH*-terminal S-peptide, residues (1 to 20), which contains His-12 of the active center, and the S-protein, residues (21 to 124), which contains His-119. The results of competition binding assays, performed as described under "Experimental Procedures," with various concentrations of these derivatives are shown in Fig. 1. The preparation of RNase S, which retained 99% of the activity of RNase A (Table I) interacted with placental RNase inhibitor to the same extent as RNase A. RNase S-protein, which had only 0.4% of the enzymic activity of RNase A toward cyclic 2',3'-CMP, in competition assays, had an RSO value of 1.0, and thus interacted with the RNase inhibitor to the same extent as does RNase A. No interaction of RNase S-peptide with the RNase inhibitor was detectable up to a loo-fold molar excess of S-peptide over RNase A. The RNase inhibitor binding domain therefore resides in the S-protein portion of the RNase A molecule; the catalytic site residue His-12 is not essential for binding of the RNase inhibitor.
Proteolysis of RNase A with pepsin removes four amino acids from the COOH terminus of the molecule (8); the product des-(121-124)-RNase retains about 1% the activity of RNase A toward cyclic 2',3'-CMP (9). Further removal of Phe-120 and the catalytic site residue His-119 with carboxypeptidase A yields des- (119-124) Interaction of Ribonuclease A and Its Inhibitor with RSO values of 1.0 (Table I) Table II were obtained from the graphical data in Fig. 2. Details of the extent of modification and the residual enzymic activity of the derivatives are also included in the table.
Fully denatured and reduced RNase A was carboxamidomethylated with iodoacetamide so as not to alter the charge on the opened molecule. This derivative did not interact with the RNase inhibitor.
The intact three-dimensional structure of the enzyme molecule is thus required for binding to RNase inhibitor, as to be expected for a speG!& interaction between the two molecules.
Chemical modification of the guanidino groups of arginine residues was performed with 2,3-butanedione and its higher structural analogue 1,2-cyclohexanedione.
Extensive modification of the enzyme was accompanied in both cases by loss of more than 95% of the activity toward cyclic 2',3'-CMP. The inactivation of RNase A upon modification of its arginine residues is consistent with the observations of Patthy and Smith (14). Yankeelov (27), reported that 45% of the enzymic activity of RNase A was retained upon modification of the enzyme with oligomers of 2,3-butanedione.
The butanedionemodified RNase -A interacted with RNase inhibitor to the same extent as RNase A. Cyclohexanedione-modified RNase A still interacted with the RNase inhibitor, but to a lesser degree (RSO = 2.0) than does RNase A. The four guanidino groups of RNase A, in arginine residues 10,33,39, and 85 (28), are thus not essential for the binding of RNase A to the inhibitor. Carboxymethylation of histidine residues of RNase A with iodoacetate at pH 5.5 results in the modification of N-3 of His-12 or N-l of His-119 in a ratio of approximat,ely 1:s (29 Reagent was added as the solid, to give a final concentration of 0.1 M, to 1 mg of RNase A in 1 ml of 0.2 M Na bicarbonate buffer, pH 8.5, 1 mu in EDTA. The reaction was performed at 37°C. Samples of 50 ~1 were removed at the indicated times and each was acidified with 20 ~1 of 15% acetic acid to terminate the reaction. Enzymic activity was then determined by the cyclic 2',3'-CMP assay (7). Amino acid analyses were performed as described under "Experimental Procedures." FIG. 5 (center).
Effect of amidination with methyl acetimidate on the activities of RNase A at pH 7.5 (0) and pH 8.5 (A), RNase A complexed with RNase inhibitor at pH 7.5 (0) and the activity of RNase inhibitor complexed with RNase A at pH 7.5 (X). Incubations were carried out at 37°C in 1 ml of either 0.2 M Hepes buffer, pH 7.5, or 0.2 M Na bicarbonate buffer, pH 8.5, both 1 mM in EDTA, 5 mM in dithiothreitol, and 0.1% in bovine serum instances, complete amidination of RNase A led to a decrease, but did not abolish, the interaction of the derivative with RNase inhibitor.
Methyl acetimidate-and methyl p-hydroxybenzimidate-modified RNase As had similar R,s values of 3.7 and 4.0, respectively.
Carbamylation of RNase A with cyanate, however, completely abolished the interaction with RNase inhibitor after 6 lysine residues were modified. Partial carbamylation of native RNase A generates derivatives with decreased enzymic activity dependent upon the extent of modification.
The results of competition assays with these derivatives are difficult to interpret, since more than one species of enzyme contributes to the hydrolysis of the cyclic 2',3'-CMP substrate. To study the effect of carbamylation of RNase A on its ability to bind to RNase inhibitor, carbamylation of carboxymethyl-His-RNase was chosen for study. The carboxymethyl derivative binds to the RNase inhibitor, is nearly enzymically inactive, and thus facilitates the interpretation of competition assay results. The effect of progressive carbamylation of this derivative on its ability to interact with RNase inhibitor was determined by competition binding assays. The change in Rso value with extent of carbamylation is plotted in Fig. 3. After only 3 lysine residues were carbamylated, 90% of the ability of the derivative to bind to RNase inhibitor was lost. We have noted that upon storage of the carbamylated derivatives there was a slow (days) decrease in their ability to interact with RNase inhibitor.
The loss of RNase activity upon amidination of its lysine residues is shown in Fig. 4. In 0.2 M Na bicarbonate buffer, pH 8.5, with 0.1 M methyl acetimidate, a rapid inactivation of RNase A occurs. The rate of inactivation exceeds the rate of overall modification of lysine residues; amidination of 2 to 3 lysine residues per molecule of enzyme was accompanied by a 70 to 80% loss of activity toward cyclic 2',3'-CMP. This result is consistent with the observation by Heinrikson (30), that alkylation of Lys-41, situated near the active site, proceeds preferentially and results in the inactivation of the Hours albumin. Reagent was added as the solid at the times indicated by the arrows to give a final concentration of 0.1 M each. Samples of 10 ~1 were removed at the indicated times and diluted into 1 ml of 20 mM Tris. HCl buffer, pH 7.5, 1 rnM in EDTA, 5 mM in dithiothreitol and 0.1% in bovine serum albumin and then assayed with yeast ribosomal RNA (6). Latent RNase activity (0) was measured after inactivation of the inhibitor by p-hydroxymercuribenzoate.  5). The rate of inactivation of RNase A is less at pH 7.5; hydrolysis of the reagent at this pH limited the rate and extent of enzyme modification.
However, at pH 7.5, no appreciable loss of RNase activity occurred when the enzyrne was bound to the RNase inhibitor.
Also, no inactivation of the inhibitor occurred under these conditions. In 0.2 M bicarbonate, pH 8.5, it was not possible to study the protective effect of the inhibitor on the amidination of the enzyme because, under these conditions, the RNase inhibitor was rapidly inactivated prior to the addition of reagent.
Amidination of RNase A with methyl p-hydroxybenzimidate in 0.2 M Hepes buffer, at pH 7.5 and pH 8.5, with resultant loss of enzymic activity along with the effect of this imidoester on the RNase . inhibitor complex is shown in Fig. 6 The loss of Asp-121 in des-(121-124)-RNase is thought to raise the pK, of His-119 and results in the almost complete inactivation of the enzyme (10). The K, for cyclic 2',3'-CMP of this derivative is only slightly altered; however, there is a 13-fold decrease in the ability to bind 2'-CMP, possibly as a result of the loss of Ser-123 which may participate in a hydrogen bond with the phosphate group of the substrate (32). Removal of Phe-120 results in the complete loss of enzymic activity. Phe-120 is thought to maintain the correct orientation of His-119 at the catalytic center of the enzyme (33). None of these key residues, nor those of the S-peptide, which are involved in maintaining the correct orientation and charge distribution of the catalytic site histidine residues, are essential for binding of the RNase inhibitor.
It is difficult to compare our results with those of Roth and Hurley (5) because of the different assay conditions and the absence of analytical data on some of their derivatives. The present study clearly confirms their qualitative observation that carbamylation of RNase A interferes with its combination with the RNase inhibitor. They also indicated that performic acid oxidized RNase A, and reduced and carboxymethylated RNase did not combine with RNase inhibitor.
Of the modifications which are thought to retain the native conformation of the enzyme, the cyanate reaction, which neutralizes the charges on lysine residues, yields the most marked decrease in the inhibitor-enzyme interaction. Derivatizations of the lysine.residues by reactions that do not alter the charge have a much smaller effect. Since RNase A is a basic protein (p1 9.4, Richards and Wyckoff (34)) and the RNase inhibitor is an acidic protein (p1 4.7, Blackburn et al. (6)), charge effects are likely to be involved. Roth and Hurley (5) found that RNase A, deaminated by treatment with nitrous acid, interacted with their RNase inhibitor, a finding somewhat contradictory to their findings upon carbamylation.
They proposed that arginine residues were involved in the interaction; the present study indicates that the guanidino groups of arginine residues are not essential for RNase inhibitor binding. Modification of the amino groups of RNase A inactivates the enzyme primarily as a result of modification of [35][36][37][38][39][40]. Modification of Lys-7 can also lead to a decrease in enzymic activity (38,39,41,42). Lys-41 is most reactive to a number of reagents as a result of its lower pK, estimated to be 8.8 (40). The precise role that Lys-41 plays in the catalytic mechanism is unknown. It is apparently not directly involved in binding substrate (34,43). After binding of substrate, however, it may have a catalytic role as a result of a changed conformation at the active site of the enzyme (44, 45). Since we obtain a major decrease in RNase inhibitor binding with the carbamylation of only 3 out of the 10 lysine residues, and since Lys-41 is known to be one of the most reactive lysine residues, we conclude that Lys-41 may be involved in the RNase inhibitor binding and results in the inactivation of the enzyme. This hypothesis is strengthened by the finding that the presence of the RNase inhibitor blocks reactions that inactivate the free enzyme at Lys-41. In addition, Lys-7 and Lys-66 are known to be normally readily available residues (43), but Lys-7 can be ruled out in this case because of the data with S-peptide and S-protein.
The unexpected finding that carboxymethylation of His-119 or His-12 strengthens the binding to inhibitor may indicate that carboxymethylation alters the region of the active site in such a way as to render the nearby Lys-41 and neighboring residues more readily available for participation in the binding process. Yang and Hummel (46) noted that carboxymethyl-His-119-and His-12-RNase A are denatured much more slowly than RNase A by urea; the pH dependence of the denaturation of carboxymethyl-His-119-RNase A was almost identical to that of RNase A in the presence of the substrate analogue pyrophosphate, and suggested that similar stabilizing forces were involved.
Achnomledgments-We are indebted to Stanford Moore for counsel in the course of this research, to Paul Montalban for skilled technical assistance, and to Barbara Curtopelle for preparation of the typescript.