Carboxypeptidase from Yeast

In order to render the carboxypeptidase from yeast more readily available for use in sequence studies in protein chemistry, the procedure of Hata, Hayashi, and associates for the preparation of the enzyme has been adapted to a larger scale through cooperation with the New England Enzyme Center. From 100 pounds of bakers’ yeast, the yield was about 1.4 g of the chromatographically purified, electrophoretically homogeneous enzyme, free of other proteinase activity. The carboxypeptidase from the yeast of different manufacture has properties corresponding very closely to those observed initially by Hayashi and Hata; the glycoprotein has a molecular weight of about 61,000, has a nitrogen content of 12.74%, yields 15.1 g of hexose per 100 g, and consists of a single polypeptide chain of 442 residues with 16 residues of glucosamine in the carbohydrate moiety. Lysine is at the NH2 terminus and -Asp-Ser-Thr-Leu is the COOH-terminal sequence. The broad carboxypeptidase activity at pH 5, which includes the liberation of proline, has been tested with the present preparation on glucagon, the B chain of insulin, and reduced and carboxymethylated pancreatic ribonuclease. The amino acids in these sequences released most slowly were glycine and aspartic acid. The carboxypeptidase retains activity in 6 M urea and can be used for the study of proteins that may have inaccessible COOH termini in the usual aqueous solution; native pancreatic ribonuclease is one example.


SUMMARY
In order to render the carboxypeptidase from yeast more readily available for use in sequence studies in protein chemistry, the procedure of Hata, Hayashi, and associates for the preparation of the enzyme has been adapted to a larger scale through cooperation with the New England Enzyme Center. From 100 pounds of bakers' yeast, the yield was about 1.4 g of the chromatographically purified, electrophoretically homogeneous enzyme, free of other proteinase activity.
The carboxypeptidase from the yeast of different manufacture has properties corresponding very closely to those observed initially by Hayashi and Hata; the glycoprotein has a molecular weight of about 61,000, has a nitrogen content of 12.74%, yields 15.1 g of hexose per 100 g, and consists of a single polypeptide chain of 442 residues with 16 residues of glucosamine in the carbohydrate moiety. Lysine is at the NH2 terminus and -Asp-Ser-Thr-Leu is the COOH-terminal sequence. The broad carboxypeptidase activity at pH 5, which includes the liberation of proline, has been tested with the present preparation on glucagon, the B chain of insulin, and reduced and carboxymethylated pancreatic ribonuclease. The amino acids in these sequences released most slowly were glycine and aspartic acid. The carboxypeptidase retains activity in 6 M urea and can be used for the study of proteins that may have inaccessible COOH termini in the usual aqueous solution; native pancreatic ribonuclease is one example.
Several carbosypeptidascs from nonmammalian sources are known to have the ability to remove most amino acid residues, including that of proline, from the COOH termini of proteins. The list includes enzymes from yeast (l-6), citrus peel (7), or citrus leaves (8,9), bean leaves (10, 1 l), Aspergillus (12), and cotyledons of germinating cotton seedlings (13,14). The carboxypeptidases from most of these sources have the property of being inhibited by diisopropyl phosphofluoridate, as is also true of a carboxypeptidase from germinating barley (15), an enzyme which is not reported to release proline.
In each instance, removal of possible contaminating traces of proteinases is a key aspect of the method of preparation.
The enzyme from yeast has properties which facilitate the scaling-up of the preparation and the isolation of a homogeneous enzyme.
The present experiments are based upon the procedures of Hata, Hayashi,16,17) for the isolation from bakers' yeast of a purified glycoprotein (originally termed proteinase C) with carboxypeptidase activity. The enzyme, after formation from its precursor (16), is similar in its chromatographic properties to peptidase ,f3 isolated from brewers' yeast by F6lix and Brouillet (I).
These peptidases also act on terminal ester groups (as in N-acetyl-n-tyrosine ethyl ester (2)) or on terminal amide groups (as in benzyloxycarbonyl-glycyl-L-I,henylalanine amide (I, 4)), but the rate of hydrolysis is much slower for the amides than for terminal groups with the -COOH fret. Since the principal action of these enzymes is in the removal of COOHterminal residues from polypeptide chains, it is practical to term them carboxypeptidases; in an analogous sense, chymotrypsin is generally classified as a proteinase rather than as an esterasc. The designation carboxypeptidase Y can serve to distinguish the yeast enzyme from the similar enzymes from other sources.
EXPERIMENTAL PROCEDURE lllaterialsFleischmann's compressed bakers' yeast was purchased as l-pound cakes from Standard Brand's Inc., New York. Bovine pancreatic RNase A (RAF grade) and carboxypeptidase A (COADFP grade) were obtained from Worthington; crystalline porcine glucagon was obtained from Lilly and the U chain of RCMl bovine insulin was purchased from Schwarz-Mann. The RCM RNase used was a sample prepared by Crestfield et al. (18). Benzyloxgcarbonyl dipeptides were purchased from Cycle; a-Nacetyl-n-tyrosine ethyl ester and a-N-benzoyl-r-tyrosine p-nitroanilide from Sigma. DFP was purchased from Aldrich.
Assay of Enzymatic Activity-The activity of carboxypeptidase Y was assayed in three ways. In the course of purification of the enzyme, activity against ol-N-acetyl-r-tyrosine ethyl ester (2) was measured with a Radiometer pH-stat.
Location of the enzyme in chromatographic eluates was conveniently performed by measurement at 410 nm of the product of the hydrolysis of ol-N-benzoyl-L-tyrosine p-nitroanilide (17). Carboxypeptidase activity of the purified enzyme was determined by measuring the rate of release of free amino acid from benzyloxycarbonyl-n-glutamyl-L-tyrosine or benzyloxycarbonyl-L-phenylalanyl-n-leucine. The reaction mixture was 3.3 mM in substrate and 50 mM in sodium acetate buffer or pyridine acetate buffer, at pH 5.5 in a total volume of 1 ml. The reaction mixture wTas incubated at 25" for 10 or 20 min. After incubation, the reaction was terminated by the addition of ninhydrin reagent (19), and the color development (20) was performed immediately.
Potential activity of carboxypeptidase Y in the initial extract was estimated after activation of the precursor (16) in 337, diosanc.
The proteinase activity of yeast and subsequent fractions was determined against casein (Pentes) by a modification (2) of the method of Kunitz.
Xeasurements of Carboxypeptidase Action-With glucagon and the U chain of RCM insulin, 0.8 mg of substrate were dissolved in 200 ~1 of 0.1 M pyridine acetate buffer (pH 5.5), and a few ~1 of solution containing 8 pg of carboxypeptidase Y were added. Incubation was at 35" for 2 hours. The hydrolysis was stopped by heating the mixture in a boiling water bath for a few minutes. The pyridine acetate buffer was removed in a desiccator; the samples were dissolved in pH 2.2 buffer and applied to an amino acid analyzer after the removal of any insoluble material by centrifugation.
With RCM RNase as substrate, the sample was 1.35 mg and 1 y0 of that amount of enzyme was used for digestion at 25" for 15 hours.
rnmodified RNase (10 mg) was digested at 25" by 45 pg of carbosypeptidase Y in 1 ml of the pH 5.5 buffer which was made 6 M in urea and contained 200 nmoles of norleucine as internal standard.. The reaction was terminated by adding Dowex 50-X12 (I~+ form, 120 mesh) until the pH of the supernatant, after shaking, was 2.5 to 3. The adsorbed amirlo acids were eluted by 5 s Nl-I,OH according to the procedure of hmblcr (21) and measured on an automatic analyzer.
COOKterminal and NtI&erminal Analysis of Carboxypeptidase Y-COOII-terminal analysis of (Larboxypeptidase Y was performed with carboxypeptidase A because denatured carboxypeptidase Y was insoluble at pI1 5.5 but soluble at about pH 8. Freshly preljared carboxypeptidase Y was inactivated by incubation at room temperature for 30 min with 100 m&f sodium phosphate (pH 7.0), 1 mM in DFP.
The reaction mixture was desalted on a column of Sephadex G-25 with 307, acetic acid as the eluent, and the protein fraction was lyophilized.
The protein (1.2 mg) was dissolved in 1 ml of 100 mM N-ethylmorpholine acetate (pH 7.6) and heated in a boiling water bath for 5 min. Carbosypeptidase A (20 pg) was added to the cooled solution and the mixture was incubated at 25". At appropriate time intervals, 200.~1 portions of the mixture were withdrawn and treated as described for the carboxypeptidase Y digestion of RCA1 RNase.
?;I&-terminal residues were estimated by the cyanatc method (28), modified for application to about 1 mg of carbamylated protein.
Chromatographic analysis was performed on the nanomole scale. Values were corrected by t.he recovery factors given by Stark and Smyth (29). Analysis of fraction C was omitted.
Edman degradation by the subtractive method was performed as described by Salnikow et al. (30).
Amino Acid AnalysesAnalyses of acid hydrolysates (22) were performed with a Durrum I-500 amino acid analyzer (range 1 to 10 nmoles per amino acid), and those of enzymic digests were made with a Beckman model 120 analyzer (range 5 to 30 nmoles) or in the 1-nmole range on an instrument of the Spackman et al. (23,24) type modified for use with 0.2%mm diameter columns by Dr. Ta-hsiu Liao of this laboratory. Tryptophan content was determined by the method of Hugli and Moore (25). Halfcystine and methionine were determined as cysteic acid and methionine sulfone, respectively, after performic acid oxidation as described by Him (26). Hexosamine content was estimated as glucosamine on an amino acid analyzer after the glycoprotein was hydrolyzed with 5.7 N HCl at 100" for 6 hours according to Johansen et al. (27).
Addilional d[ethodsDisc electrophoresis (31) was performed with 7% polyacrylamide gel at $-I 8.9 without a stacking gel. Molecular weight was estimated by disc electrophoresis in the presence of sodium dodecyl sulfate by the method of Weber and Osborn (32).
Total hexose was determined by the orcinol method of Tsugita and Akabori (33). Protein and nucleic acid content of the fractions obtained during the purification of carboxypeptidase Y were estimated by the spectrophotometric method of Warburg and Christian (34). Concentration of purified carboxypeptidase Y was determined spectrophotometrically by use of E:gnrn = 15.0; this value was obtained with a sample of the purified enzyme after correction for moisture content (a separate sample was dried at 105" to constant weight).
In the use of the enzyme on proteins, the substrate concentration was usually determined by amino acid analysis after acid hydrolysis of an aliquot of the protein solution.
Preparation of the E'nzyTne-An outline of the purification steps, which represent a scale-up and modification of those developed by Hayashi et al. (16) and Aibara et al. (17) is shown in Table I; Steps 1 to 5 wele designed and conducted in cooperation with Mr. Henry Blair and Dr. Stanley E. Charm of the New England Enzyme Center, Tufts Iinivcrsity School of Medicine, Roston. In Step 1, 100 ml of chloroform per pound of yeast (half of the volume used previously (16)) were found to be enough to cause solubilization of all of the potential enzyme activity. This change facilitated the ccntrifugation by reducing the amount of chloroform in suspension and decreased the extraction of nucleic acid.
The removal of cell debris and the fractionation by ammonium sulfate were critical steps in the large scale preparation because of the volumes that needed to be centrifuged at high speed. These steps were successfully accomplished through use of a Sharples centrifuge, the center cylinder of which was previously cooled to 5". Antifoam (Dow Corning AF) was added for prevention of surface inactivation of the enzyme through foaming of the solution.
The bottom part of the settled autolysate, which contained more chloroform than the top part, was centrifuged, separately, batchwise, in a bucket-type centrifuge.
For activation of the inactive precursor by proteolysis, the (NH,)2S04 precipitate obtained in Step 3 was dissolved in 3 liters of 50 mM sodium acetate buffer (pH 5.0) per kg of wet precipitate.
Incubation at 25" and pH 5.0 for 20 hours yielded maximum activity (36). There is extensive digestion of undesired proteins in Step 4 by yeast proteinases as well as by the carboxypeptidase during the activation step, and the peptides are removed by the dialysis in Step 5.
For removal of nucleic acids and basic proteins, stepwise chromatography on diethylamino ethyl cellulose (DE-52, Whatman) was employed.
The fractions containing enzyme coincided with the elution of a darkly colored eluate; the first eluate, light yellow in color, was discarded, and the next eluate, dark Flow rate, 100 ml per hour (See Fig. 1).
Step 7. Rechromatography Pooled active fractions (Fig. 1 In this step, nucleic acids were almost completely removed and the yield of the enzyme was 80 '%. On the large scale, precipitation of the enzyme by dialysis against saturated (NH,)S04 was used. On a small scale preparation, ultrafiltration at 4" was used to concentrate the enzyme without detectable inactivation.
Final purification was performed by column chromatography on DEAF-Sephadex A-50 at 4" as described previously (2). -After rechromatography, the enzyme was reprecipitated by dialysis against saturated (NH&S04. The thick suspension can be stored at -20" indefinitely. For USC, a portion corresponding to about 100 mg of protein was dissolved in a minimum volurne of water and dialyzed against H20 or 0.01 M phosphate buffer (pH 7) to give 10 ml of an approximately 1% aqueous solution of the enzyme.
The frozen solutions showed no loss of activity in sis months.
Lyophilization causes loss of more than 207" of the activity.

Purification
and Characterization of the &ry?ne-The elution pattern for the first chromatography at 5" on DEAE-Sephadex A-50 is shown in Fig. 1. The content of proteinase A in the extract of Fleischmami's yeast was about half that obtained earlier with Oriental yeast (see Fig. 3 in Reference (2))) whereas the content of carboxypeptidase Y was about the same with the two commercial yeasts. Upon rechromatography, the carboxypeptidase yielded a single peak, and the specific activity was constant across the peak. From 96 pounds of yeast, 1.4 g of carbosypeptidase Y are obtained (yield, 5770). In laboratory scale experiments starting with 5 pounds of yeast, 45 mg of the purified enzyme were obtained (recovery, 307,).
The degree of 1)urification (about 450.fold) and the yield are calculated from activities measured after an aliquot from Step 2 is treated to 2299 convert fully the precursor to enzyme; otherwise, the calculation of the degree of purification comes out incorrectly high. Purified carboxypeptidase Y yielded a single band on disc electrophoresis both in the absence and in the presence of sodium dodecyl sulfate.
The enzyme after reduction and carbosymethylation also showed no evidence of heterogeneity upon gel electrophoresis.
The molecular weight of the enzyme estimated by gel electrophoresis was 61,000 by comparison with standard proteins (Fig. 2), which agrees with the ultracentrifugally determined value obtained with earlier preparations (17). The purified enzyme was nearly homogeneous by end group analysis.
The cyanate method, applied to the enzyme on the ultramicro scale, yielded 0.75 residue of lysine as an NHz-terminal residue, and a trace (below 0.02 residue) of aspartic acid, serine, and alanine.
The COOH-terminal sequence was deduced as -Asp-Phe-Ser-Leu from the data in Fig. 3. These experiments indicate that the enzyme consists of a single polypeptide chain.
The nitrogen content of carboxypeptidase Y was 12.747,, on an ash-and moisture-free basis. No nucleotide component is present, as judged from ultraviolet absorption (R*&A26" = 1.76). The amino acid composition was estimated from duplicate analyses of 5.7 s HCl hydrolysates at 110' for 22 hours (Table  II) ; the nitrogen recovery (including NH8 and hexosamine) was 103%.
The enzyme yielded 15.1 g of hexose per 100 g.
H~J&O~JS~S of Small Peptide Substrates-The specific activities of the enzyme against benzyloxycarbonyl-L-glutamyl-L-tyrosine and benzyloxycarbonyl-r,-phenylalanyl-L-leucine were 15.4 and 122 pmoles mini mggi, respectively. The latter peptide was the best substrate so far tested for the enzyme (K, 5 lop4 M). This enzyme hydrolyzed a-N-acetyl-L-tyrosine ethyl ester with a specific activity of 90 pmoles mini mg-i.
All of these activities were completely inhibited by lop4 M IJFl' or p-HMl{.
Action on Larger Peptides and ProteinsThe enzyme hydrolyzed bovine glucagon, the 1s chain of RCM insulin, and RCM RNase releasing only the amino acids which arc to be expected  3. Rate of release of amino acids from 11FP-inactivated carboxypeptidase Y by digestion with pancreatic carboxypeptidase A. Protein concentration was 0.1% in 0.1 M N-ethylmorpholine-acetate (pH 7.6), and the reaction was carried out at 35" with a substrate to enzyme ratio of 50. Other details are given in the text.
from the COOH-terminal sequences, as shown iu Table III. Neither peptides nor other amino acids were detectable on the amino acid analyzer.
The results with insulin and RNase show that the glycine residues in these sequences are only slowly released. Aspartic acid was rapidly released from RCM RNase, whereas its release was slow from glucagon.
With RCM RNase, hydrolysis of peptide bonds on both sides of a proline residue was achieved.
The above results, as well as the following observation, show that proteinase action, which might hydrolyze internal bonds in a protein to yield large peptide fragments, is absent in the preparation of carboxypeptidase Y. After an extensive digestion of RCbl RNase by the enzyme (under the same conditions as used for the data in Table III), one aliquot of the digestion mixture was subjected to determination of the released amino acids and the other to one cycle of subtractive Edman degradation.
It was found that within experimental error no amino acids other than NH*-terminal lysine were lost by Edman degradation. In addition, no peptides of an intermediate size were found when the digestion mixture was subjected to gel filtration chromatography on Sephadex G-75 using 50% acetic acid as eluant.

Use for Analysis of COOH-terminal
SequencesIn order to test the use of carboxypeptidase Y for sequence determination, the rates of release of amino acids from RCM RNase by the enzyme were followed quantitatively.
The result is summarized in Fig.  4. In the first 10 min, five amino acids were liberated rapidly; thereafter histidine and proline appeared, accompanied by further release of valine.
The results are consistent with the known structure of RNase (39,40). Carboxypeptidase Y could not release any measurable amino acids from native RNase after a period of 30 min of incubation.
However, the carboxypeptidase is still active in the presence of 6 M urea (6) and in such a solution released seven amino acids from RNase (Fig. 5). The rate of the release of amino acids was about half of that observed in the RCM RNase experiments.
This method provides an approach to the determination of the COOH-terminal sequence of proteins that are difficult to denature.
In these experiments,

DISCUSSION
Yeast proteinase A is an enzyme which may contaminate carboxypeptidase Y, because both enzymes have similar properties (sugar content, isoelectric point, and molecular weight) (2). A specific inhibitor for proteinase A has not yet been found.
However, the content of proteinase A in the extract of Fleischmann's yeast used here was remarkably smaller than that found with the Oriental yeast used previously (2). This fact facilitated the isolation of carboxypeptidase Y in pure form as judged by the absence of measurable proteinase activity when tested against glucagon, B chain of insulin, and RCM RNase.
Hugli (41) has used carboxypeptidase Y to remove five amino acid residues from DNase A after an internal cleavage of one bond by chymotrypsin; the resulting product, DNase (1 to 173, 179 to 259), retained nearly full DNase activity, and this is a very sensitive test for The substrates, at a concentration of abollt 0.5c/, were digested at pH 5.5 at an enzyme to substrate ratio of 1: 100. Glncagon and the B chain of insulin were incubated at 35" for 2 hours; RCM RNase was held at 25" for 15 hollrs. The released amino acids are expressed as mole per mole of srtbstrate.  (pH 5.5), and the reaction was performed at 25" with a substrate to enzyme ratio of 280. The results indicate the sequence -(Val,Pro)-His-Phe-Asp-Ala-Ser-Val (39,40). the presence of proteolytic enzymes since D?r'ase is so suscept.ible to proteolysis (42).
Repeated freezing and thawing of solutions of carboxypeptidase Y, or prolonged storage of the enzyme at room temperature, can lead to autodigestion with the liberation of free amino acids, primarily the amino acids in the COOH-terminal portion of the protein.
Estremely dilute solutions of the enzyme lose activity.
Carboxypeptidase Y is relatively stable in the presence of protein denaturants (6); about 20% of the activity was lost after incubation Q-ith 6 M urea at 25" for 1 hour.
The specificity of carboxypeptidase Y has been examined with synthetic peptides (3). The enzyme released most types of amino acids, including proline, from the COOH termini of the substrates tested; however, the release of glycine and sometimes aspartic acid was slow. The latter observation is in accord with the result that poly-L-aspartic acid is a poor substrate for the enzyme, whereas poly+glutamic acid is rapidly hydrolyzed by the enzyme (4). The slow release of glycine is a property shared by pancreatic carboxypeptidase A (al), but the chemical properties of the yeast enzyme are markedly different from those of the pancreatic enzyme since carboxypeptidase Y is inhibited by l>F1',2 p-HX'IB, and Hg2f.
The broad specificity and the stability in urea of carboxypeptidase Y make the enzyme applicable to some sequence determinations which are not possible with other carboxypeptidases. Furthermore, as a result of its amidase action, this enzyme might be applied to the sequence analysis of peptides having amidated COOH-terminal groups such as oxytocin and vasopressin. The preparation described in this communication has been used by Liao et al. (44) in sequence studies on peptides from pancreatic deoxyribonuclease (peptides number C4, Th20, Th8a, and T12-ThSc).
Homoserine and carboxymethyl cysteine were released by carbosypeptidase Y, and the release of prolinc was helpful in two of the instances.