Modification of an arginyl residue in pepsin by 2,3-butanedione.

Abstract 2,3-Butanedione (biacetyl) was found to modify an arginyl residue in porcine pepsin at pH 6.0, 25°. In an inhibition study, half-inactivation occurred after 2½ hours of reaction, and maximum inactivation of 80 to 85% was reached after 24 hours. Amino acid analysis of the modified pepsin revealed that 1 arginine residue remained, indicating that 1 arginine residue had been modified. The loss of activity and loss of arginine occurred concomitantly during the course of reaction. The presence of substrates retarded enzyme inactivation, as well as arginine loss, at corresponding rates. Porcine pepsinogen treated with biacetyl in a pH 7.0 solution did not lose activity significantly, as shown by its proteolytic activity after subsequent activation by acid. Biacetyl-modified pepsin continued to be susceptible to specific esterification reagents: diazoacetyl-dl-norleucine methyl ester, p-bromophenacyl bromide, and 1,2-epoxy-3-(p-nitrophenoxy)propane. Peptides from a chymotryptic digest of pepsin and of biacetyl-treated pepsin were separated by high voltage paper electrophoresis at pH 6.5. Two peptides were isolated, containing one each of the only 2 arginine residues of pepsin. The sequences of these peptides were ascertained from amino acid compositions and amino-terminal determinations. One of the peptides (A1) from biacetyl-pepsin contained almost no free arginine. An additional spot, apparently representing modified arginine, was observed for the digest of Peptide A1 in high voltage electrophoresis. These results indicated that the arginyl residue in Peptide A1, located 12 residues from the carboxyl terminus of pepsin, is the site of biacetyl modification. This arginyl residue apparently does not directly participate in the catalysis of the enzyme. Biacetyl was also found to inactivate other gastric proteases in similar fashion, i.e. human gastricsin, human pepsin, and bovine rennin.

studies of the structure of the active center of pepsin.
Because the reaction of pepsin with diazo compounds results in complete loss of activity, this aspartyl residue is probably involved in catalysis of the enzyme.
The second group of inactivators is made up of the substrate-like epoxides, such as 1,2-epoxy-3+nitrophenoxy)propane reported by Tang et al. (14) and Tang (15). These compounds inactivate pepsin by esterifying a carboxyl group in the active center which differs from the carboxyl groups that react with other types of inactivators (16, 17). Since this reaction completely inactivates the enzyme, the esterified carboxyl group is probably essential for catalysis.
Finally, the phenacyl bromide derivatives, reported by Erlanger et al. (18,19), apparently esterify a carboxyl group not directly involving enzymic catalysis, since the modified enzyme remains partially active.
All three types of inactivators described above, however, act upon carboxyl groups.
A comparison of amino acid sequence of this region in several homologous enzymes, including human pepsin and gastricsin (23) and bovine rennin (24), has revealed the relative positions of the arginyl residues to be identical.
This suggests that the arginyl residues may be important in acidic protease function, and that they have thus been retained during the evolutional process.
To elucidate the role of arginyl residues in peptic activity, we designed a study on the chemical modification of arginyl residues in pepsin.
Although a number of reagents are available for arginine modification (25-28), the useful ones are necessarily reactive in acidic solutions, since pepsin inactivates spontaneously in alkaline solutions.
2,3-Butanedione (biacetyl) was chosen for testing since this reagent has been shown to react with arginyl residues under mild conditions in a near-neutral Moreover, biacetyll has been successfully used to modify arginyl residues in antibodies (32) and carboxypeptidase A (33).
The results of this study indicated that biacetyl modifies t'he arginine 12 residues from the COOH-terminal of pepsin, causing an 85% loss of enzyme activity.

Materials
Porcine pepsin (3 times crystalline) and bovine hemoglobin substrate were obtained from Pentex; aminopeptidase M was purchased from Henley and Co., and crystalline rennin from Nutritional Biochemicals Co. Human gastricsin and pepsin were prepared as described previously (34,35). Diazoacetyl-DL-norleucine methyl ester was a gift from Dr. John N. Mills, Oklahoma Baptist University, Shawnee, Okla. The synthetic procedure used was that described by Rajagopalan L-Alanyl-L-arginine acetate was obtained from Cycle Chemical Co. Pepsinogen was chromatographically pure, as determined by fractionation on a DEAE-cellulose column.
Polyamide layers were obtained from Gallard-Schlesinger Chemical Corp. Other reagents were of analytical grade or the highest quality available commercially.

Methods
Proteolytic Activity---The proteolytic activity was measured with bovine hemoglobin as substrate.
The procedure was essentially that of Anson and Mirsky (37).
Milk-clotting Activity-The procedure of Berridge was followed for the determination of milk-clotting activity (38). Inactivation &u&es-A typical reaction mixture consisted of 1 ml of a 0.1% solution of pepsin in 0.2 M potassium phosphate buffer, pH 6.0. To this solution, about 10 ~1 of biacetyl were added. The reaction was carried out at 25 f I". Aliquots of 10 ~1 were withdrawn at different time intervals for the assay of enzyme activity.
An enzyme solution containing no biacetyl was incubated along with the reaction mixtures to serve as a cqntrol.
Preparation of Biacetyl-Pepsin-Crystalline pepsin (30 mg) was allowed to react with 0.3 ml of biacetyl in 30 ml of 0.2 M potassium phosphate buffer, pH 6.0. After standing for 24 hours at room temperature, the solution was dialyzed in a 4" cold room for 40 hours against several changes of distilled water, after which it was lyophilized.  (3). At the end of reaction, the mixture was dialyzed against distilled water at 4". The resulting pepsin had no proteolytic activity and contained 1.3 residues of norleucine according to amino acid analysis.
Reaction of Biacetyl-treated Pepsin with p-Bromophenaeyl Bromide-The reaction of pepsin or biacetyl-treated pepsin with p-bromophenacyl bromide was carried out according to the method of Erlanger et al. (18). The reacted pepsin was dialyzed at 4" against distilled water.
Biacetyl ModQication of Oxidized B Chain of Bovine lnsulin-The oxidized B chain of bovine insulin (3 mg) was suspended in 1 ml of 0.02 M potassium phosphate buffer, pH 6.0. Biacetyl, 10 ~1, was added to this solution, which was then incubated with occasional shaking at 25" for 24 hours. The product was lyophilized thoroughly, redissolved in small amounts of water, and lyophilization was repeated.
The aminopeptidase M hydrolysate of the modified B chain of insulin contained 1 residue of lysine (the same as that in the unmodified B chain) and no arginine, indicating the complete modification of arginine by biacetyl.
Biacetyl ModiJication of Dipeptide L-Alanyl-L-arginine-Dipeptide (6 mg) was dissolved in 1 ml of 0.02 M potassium phosphate buffer, pH 6.0, to which 10 ~1 of biacetyl were added. The solution was left standing at room temperature for 24 hours and then lyophilized.
The resulting material was subjected to high voltage electrophoresis as a 20-cm band on Whatman 3MM paper at pH 3.5 (60 volts per cm for 1 hour).
A guide strip was cut and treated with cadmium-ninhydrin reagent to reveal the peptide bands. Only trace amounts of the original dlpert de were present (relative mobility to free arginine was 0.95). A strong peptide band with a mobility of 0.80 (i.e. relative to free arginine, 1.0) was present, apparently representing the biacetyl-modified dipeptide.
This peptide band was cut out, and the material was recovered by elution.
The hydrolysate obtained with methanesulfonic acid contained free alanine, but no free arginine.
Amino Acid Analysis--Amino acids were quantitated with a Spinco model 120B amino acid analyzer having a modified "range card" in the recorder to permit analysis in the range of 0.001 to 0.1 mole; the procedure of Spackman (39) was followed.
For quantitation of biacetyl-modified arginine, several hydrolytic methods were tried. Yankeelov (31) reported that about 12% of mod;fied arginine was regenerated to free arginine in hydrolysis with 6 N HCI. We found that the hydrollsis of biacetyl-modified alanylarginine by 6 N HCl resulted in about 20% of the modified arginine being regenerated back to free arginine.
(In some earlier experiments, when HCI hJ-drolpsis was used, the analytical data for arginine were corrected for regeneration.) The digest with aminopeptidase M did not regenerate free arginine from either modified alanylargin:ne or the B chain of insulin.
However, this enzymic digest of pepsin could not fully release the arginine content of 2 residues, and thus was unsuitable for amino acid analysis of the modified pepsin. A third method, hydrolysis with methanesulfonic acid, devised by Liu and Chang (40), was most successful. The hydrolysis of biacetyl-modified alanylarginine produced no free arginine.
The amount of modified arginine could not be analyzed directly in the amino acid analyzer.
We confirmed the report of Yankeelov (31) that several overlapping peaks appearing in the regions were basic amino acids normally eluted. Therefore, the Modijication of an Arginine in Pepsin Vol. 247, No. 9 amount of modification of arginine was determined by the difference in arginine content before and after modification. The hydrolysis with methanesulfonic acid was carried out by the method of Liu and Chang (40). Protein or peptide (0.02 to 0.05 pmole) was hydrolyzed with 0.2 ml of 3 N methanesulfonic acid containing 0.2% 3-(2-aminoethyl)indole hydrochloride at 110" for 24 hours.
Amino-terminal Residue Determination-The NHa-terminal residues of the peptides were determined after reaction with dansyl chloride according to the method of Gray and Hartley (41). The dansyl amino acids were identified by two-dimensional thin layer chromatography on polyamide layers (42). High Voltage Paper Electrophoresis-The commercial apparatus produced by Savant Instruments, Inc. was used with the following buffer systems: pyridine-acetic acid-water (25: 1:225 by volume), pH 6.5; formic acid-acetic acid-water   I3ecause the reaction apparently proceeded more rapidly at higher pH values, pH 6.0 was chosen for further experiments.
Varying the biacetyl concentration from 5 to 20 ~1 per ml of incubation mixture failed to alter the course of inactivation appreciably. The use of citrate or phosphate buffer had no apparent effect. iZnother arginine-modifying reagent, benzil (27), incubated as a saturate solution in the same buffer, had no measurable inactivation ability. Phenylglyoxal at 100 mM inactivated about 50y0 of enzyme activity after 4 hours of reaction.
At 10 mM, however, no inactivation was observed.
After 20 hours of reaction of biacetyl with pepsin, the mixture was dialyzed in the cold against distilled water. The dialyzed sample regained no act'ivity, indicating that the inactivation was irreversible and probably due to the modification of arginyl residues in pepsin molecules.
This was confirmed by amino acid analysis, which showed a loss of about 1 residue of arginine in biacetyl-treated pepsin ( Table I). The same result was obtained whether the hydrolysis was carried out with HCl or methanesulfonic acid.
Porcine pepsinogen was incubated with biacetyl under the same conditions used in the inactivation of pepsin, except that the pH of the solution was 7.0. A considerably slower inactivation rate was observed, approximately one-tenth of that of pepsin in pH 6.0 solution.
With these substrates, the specific activity of biacetyl-treated enzyme was only 15% of that of native enzyme, confirming the results obtained with the hemoglobin assay (Fig. 1). The activity of biacetyl-treated pepsin was also measured in solutions ranging from pH 1.0 to 5.0, with hemoglobin as substraDe. optimum.
The treated enzyme showed a marked increase in absorbance in the wave length region 235 to 280 nm over that of untreated pepsin (Fig. 2). Although Yankeelov reported that biacetyl-modified arginine can be partially regenerated by salt (30), we observed no recovery of lost, enzyme activity up.on incubation of biacetyl-treated pepsin in 1 M NaCl, pH 6.0. Biacetyl-treated pepsin was further incubated with three other known inactivators of pepsin. The first, diazoacetyl norleucine methyl ester (3), completely inactivated the biacetyl-treated pepsin (Table II).
Amino acid analysis showed that 1.3 residues of norleucine were incorporated.
The third, pbromophenacyl bromide (18), reduced the remaining enzymic activity of biacetyl-treated pepsin, but (as in the inactivation of native enzyme) failed to abolish it altogether (Table II)   acetyl-To determine whether the biacetyl-modified arginyl residue is located in or near the active center of pepsin, inactivation was carried out in the presence of a number of known synthetic enzyme substrates.
In consequence, the rate of inactivation was found to be significantly reduced by N-acetyl-L-phenylalanyl-L-diiodotyrosine and N-acetyl-tityrosyl-L-tyrosine (Table  III).
Experiments carried out in pH 6 and 3 had the similar results.
N-Carbobenzoxyglycyl-L-phenylalanine and its amide, two poor substrates, and glycylglycine had no protecting effect when present at 1 mM.
The course of inactivation, as well as the loss of arginine by the reaction of biacetyl, was measured in both the presence and the absence of N-acetyl+tyrosyl-L-tyrosine at pH 6.0. In the absence of substrate, enzymic activity disappeared concomitantly with arginine loss. As described above, at the maximum inactivation (85%), 1 residue of arginine reacted. However, this relationship remained fairly consistent during the course of reaction.
In the presence of synthetic substrate, the rates of loss of enzyme activity and arginine residue were slower than in control experiments.
When the data obtained at 3, 6, and 9 hours were averaged (Fig. 3), N-acetyl-L-tyrosyl-L-tyrosine ModiJication of an Arginine in Pepsin Vol. 247,No. 9 was found to slow the inactivation by 22.6% and the loss of arginine by 23.6%.
These results indicated that substrates of pepsin protected the modification of arginine, and that loss of arginine brought about the loss in enzymic activity.

Identijicatim of Modi$ed Arginyl
Residue-Since the amino acid sequence around the 2 arginyl residues in porcine pepsin is known (20-22), experiments were designed to determine which of the arginines was modified by biacetyl.
Our approach was similar to that used to determine the arginine sequences in human pepsin (23). Pepsin and biacetyl-modified pepsin in 0.25 % concentrations were digested separately by ar-chymotrypsin.
The peptide mixtures were subjected to high voltage electrophoresis at pH 6.0. In the native pepsin digest, only two peptide spots, A1 and AZ, migrated toward the cathode (Fig. 4). These two peptides were then recovered from preparative high voltage electrophoresis, and their amino acid compositions as well as their NH*terminal residues were determined (Table IV).
Peptide AZ apparently contained two peptides, which were separated by prolonging the preparative electro- The position of peptides are indicated as dark areas.
The circled areas indicate that only faint spots were found for Peptide Al from biacetyl-reacted pepsin. An amino acid mixture (AA) was also run to give the position of neutral amino acids (AT) and arginine (Arg).
The electrophoresis was carried out in pH 6.5 at 70 volts per cm for 60 min. The acidic peptides which moved toward the anode were omitted from the figure. phoresis to 90 min. One of the peptides in A1 was derived from the 18th and 21st residues from COOH terminus and had the sequence of Ile-Arg-Glu-Tyr (17). The second peptide contained 9 residues from the carboxyl terminus of pepsin and had the sequence Asn-Lys-Val-Gly-Leu-Ala-Pro-Val-Ala.
It contained no arginine.
The electrophoretic pattern of peptides derived from 9-and 24hour digests of 0.25% biacetyl-treated pepsin showed a nearabsence of Peptide A1 (Fig. 4), apparently due to the mod;fication of arginine C-12 (12th residue from COOH terminus of pepsin (20)). The electrophoretic pattern of Peptide At did not differ appreciably from that of biacetyl-modified pepsin (Fig. 4). However, Peptides A1 and AZ manifested the same color intensity either after prolonged digestion with cY-chymotrypsin or when the concentration of the modified pepsin was 1% instead of 0.25%. Thus, the hydrolysis by or-chymotrypsin to form modified Peptide A1 was slower in modified pepsin, where arginine was 1 residue away from the bond hydrolyzed.
Since the modified arginine is known to retain its positive charge (32)) modified Peptide A1 would migrate to the same position as the unmodified peptide.
These inferences led us to perform experiments to identify the biacetyl-modified arginine in Peptides A1 and AZ obtained from the digest of modified pepsin recovered from preparative high voltage electrophoresis.
Peptides A1 and A2 from biacetyl-pepsin were digested separately with aminopeptidase M. The resulting amino acid mixtures were subjected to high voltage paper electrophoresis at pH 3.5 (Fig. 5). Peptide Az from either native or biacetyl-modified pepsin showed an arginine spot, with no additional spot near the region where the basic amino acids are located.
The digest of Peptide A1 from native contained free arginine, as expected. However, the digest of Peptide A1 from biacetyl-treated pepsin contained lysine and only trace amounts of arginine.
Additionally, this digest contained a spot with a relative mobility of 0.74 (the mobility of arginine being defined as 1.0) ; this was assumed to be the biacetyl-modified arginine.
Since it is known that the However, the spot apparently representing the biacetyl-arginine was clearly present. From these results, we concluded that the biacetyl reacted with the native pepsin at the arginine residue which is in the sequence of Peptide A1: Asp-Arg-Ala-Asn-Asn-Lys-Val-Gly-Leu-Ala-Pro-Val-Ala. The COOH-terminal alanine is the COOH-terminal of pepsin. Therefore, the modified arginine must be located 12 residues from the carboxyl terminus of the enzyme.

Reaction of Biacetyl with Other Gastric
Proteases-Several other gastric proteases, human gastricsin, human pepsin, and bovine rennin, were treated with biacetyl.
As shown in Table I, enzymic inactivation of about 75 to 80% was observed in all three cases. The amino acid analysis of human enzymes showed a loss of about 1 residue of arginine.

DISCUSSION
The inactivation of pepsin by biacetyl is apparently due to the modification of arginine located 12 residues from the carboxyl terminus (C-12) in the enzyme molecule.
Several pieces of evidence support this conclusion: (a) biacetyl-treated pepsin lost 1 residue of arginine, as determined by amino acid analysis, (b) the relationship between the loss of activity and loss of arginine remained about the same during the course of inactivation (Fig. 3), (c) the peptide containing arginine C-12 was nearly absent in the or-chymotryptic digest of biacetyl-treated pepsin (0.25% pepsin) while the peptide containing an arginine located 20 residues from the carboxyl terminus (C-20) remained essentially unchanged, (d) an enzymic hydrolysate of Peptide A1, containing arginine C-12, was almost devoid of free arginine and contained a ninhydrin-positive spot, apparently representing biacetyl-modified arginine (Fig. 5).
The modification of arginine C-12 caused a similar degree of decrease in the hydrolysis rate of both protein and dipeptide substrates. The dipeptide substrates contain two side chains which may participate in the binding to the active center of the enzyme. However, pepsin is known to contain not only two major side chain-binding sites, but also additional hydrophobic binding sites which interact with the side chains of amino acids that are 1 or 2 residues removed from the 2 major binding residues (46)(47)(48). The results of this study suggest that the loss of activity is due to an impairment of the primary binding or catalytic site of the enzyme.
The role of arginine C-12 in the catalysis of pepsin is uncertain. It is unlikely, however, that this arginine side chain participates directly in enzymic hydrolysis, since the enzyme remains partially active after modification.
They also,reported a shift of PH-L,,~ curve (on the low pH side) of about half-pH unit. They suggested that the arginine modified by phenylglyoxal (probably also arginine C-12) may form a salt linkage to a carboxyl group in the active center of the enzyme. Thus the modification would cause a shift in the pK of the carboxyl group and in the pH-k,,t curve. However, a change of pK in an active center residue mediated indirectly by the conformational change after arginine modification has not been ruled out. We feel that if arginine C-12 plays a role as suggested by either Hartsuck and Tang (17) or Kitson and Knowles (49), a more complete inactivation would be expected after the modification.
The presence of substrates of pepsin slows the rate of inactivation. This protecting effect appears to be due to the binding of a substrate molecule in the active center of the enzyme. The fact that poor substrate and nonsubstrate peptides do not protect and that protection by substrates is observed at both pH 3 and 6 supports this view. However, the protection experiments do not establish the location of arginine C-12 in the three-dimensional structure of the enzyme. It is still possible, that the effect observed is due to a substrate-induced change of conformation in the pepsin molecule.
The absence of significant modification of arginine C-20 is probably due to a steric hindrance.
Since biacetyl is not an "active center-directed reagent," the reactivity of two arginines should not differ significantly, if they are equally accessible. Therefore, it is possible that the side chain of arginine C-20 is relatively "buried" in the tertiary structure of the enzyme. Yankeelov (31) showed that biacetyl reacts with amino acid imidazole groups at a much slower rate. These minor reactions are probably not responsible for the inactivation of pepsin. No significant loss of lysine and histidine was evident after mod.fication. Additionally, a number of studies on the modification of amino groups in pepsin have indicated that no loss of enzymic activity results from N-acetylation (50), from loss of amino groups (51), or from N-ethoxyformylation (52). All of these results argue against the modXcation of the amino terminus as the basis for inactivation.
Aclcnowledgments-We wish to thank Dr. Jean Hartsuck for the initial suggestions of arginine modification experiments, Dr. John Mills for a sample of diazoacetyl-m-norleucine methyl ester, and Ms. Barbara Cox for valuable assistance in the preparation of this paper.