Molecular and Catalytic Properties of Rabbit Testicular Dipeptidyl Carboxypeptidase*

Rabbit testicular dipeptidyl carboxypeptidase activ- ity was purified by a procedure exploiting its affinity for N-a-[l-(S)-carboxy-3-phenylpropyl]-~-lys~l-~- proline. The molecular, catalytic, and immunological properties of the testicular enzyme are presented and compared with the corresponding properties of pulmonary angiotensin-converting enzyme. Although catalytically similar and immunologically related to pul- monary dipeptidyl carboxypeptidase, the testicular enzyme has a molecular weight (100,000) which is lower by a factor of about one-third and differs in its NH2 and COOH termini. Furthermore, we present evidence that the testicular enzyme is not a post-translation product of the pulmonary type enzyme. These data suggest that testicular and pulmonary dipeptidyl carboxypeptidase are two distinct proteins which are catalytically similar and immunologically closely related. the position of catalase, pulmonary angiotensin-converting enzyme, and testicular dipeptidyl car- boxypeptidase.

COOH-terminal dipeptidyl carboxypeptidase activity (angiotensin -converting enzyme, peptidyldipeptide hydrolase, EC 3.4. 15.1) is present in the vascular endothelial cells of virtually all mammalian organs (1). The enzyme plays a key role in the control of blood pressure by converting angiotensin I to angiotensin 11, a powerful vasoconstrictor, and by inactivating bradykinin, a vasodilator (reviewed in ref. 2). The lung is believed to be the most important physiological site of this conversion (3). Dipeptidyl carboxypeptidase activity in the testis, however, differs from that in the lung, first, because it increases dramatically at puberty (1, 4) and is thus probably hormonally controlled, and second, because antipulmonary enzyme antibodies can discriminate between it and the pulmonary enzyme (5, 6).
As a first step toward establishing the physiological function of the testicular enzyme and the biochemical basis for the structural and regulatory differences between it and the pulmonary enzyme, we have now isolated and characterized the pure testicular enzyme. The purification has been facilitated by a step which takes advantage of its high affinity for N-a-[l-(S)-carboxy-3-phenylpropyl]-~-lysyl-~-proline, a recently described (7) potent inhibitor of pulmonary angiotensin-converting enzyme.
* This work was supported by Grant HL21394 from the National Institutes of Health. A preliminary report of some aspects of this study was presented at the 1981 meeting of the American Society of Biological Chemists, St. Louis, MO, June, 1981 (El-Dorry, H., Bull, H., Iwata, K., andThornberry, N., Fed Proc. 40, 1039). T h e costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. This paper presents the molecular and catalytic properties of the purified testicular enzyme and their comparison with the properties of the enzyme from lung.

Methods
Enzyme Assays-Dipeptidyl carboxypeptidase activity was determined using as substrate Hip-His-Leu according to the method of Cushman and Cheung (9). A unit of activity is the amount which catalyzes release of 1.0 pmol of hippuric acid per min at 37 "C under the conditions described by them. Activity on angiotensin I containing a ['4C]leucyl residue at its COOH terminus was quantitated by the release of radioactive His-Leu (8, 10). Products of the enzymatic degradation of bradykinin were identified by paper electrophoresis (11).
Competition Radioimmunoassays-Goat antipulmonary enzyme antibodies were prepared as described previously (5). Antienzyme antibodies were elicited in adult white mice by intraperitoneal injections of pure testicular or pulmonary dipeptidyl Carboxypeptidase (10 pg) contained in 125 p1 of 10 mM Tris-HCI, pH 7.8, and emulsified with an equal volume of complete Freund's adjuvant (Difco). Immunizations were given at intervals of 2 weeks, and maximal titers were usually attained after three injections. The antisera were heated a t 56 "C for 30 min. The radioimmunoassays were carried out as described previously (5) except that the secondary antibody was rabbit antimouse IgG which had been freed of serum angiotensin-converting enzyme by passage over DEAE-cellulose in 15 mM potassium phosphate buffer, p H 8.3 (5). The radioiodinated testicular and pulmonary enzymes used as displaceable antigens (1-3 ng per assay) were labeled as described by Hunter (12) and contained, respectively, 2.6 X 10' and 1.8 X lo4 cpm/ng of protein. T h e final dilutions of mouse antisera against the testicular and lung enzymes were 1:750 and 1:2200, respectively.
Preparation of Inhibitor-Sepharose Gel-Sepharose CL-4B was activated and coupled to the amino group of the inhibitor through a bisoxirane spacer arm as described by Sundberg and Porath (13). Washed Sepharose CL-4B (100-ml packed gel, approximately 67 g, suction-dried cake) was stirred for 8 h at 23 "C in a solution made from 100 ml of 1,4-butanediol diglycidyl ether (approximately 70%') and 100 ml of 0.6 M NaOH containing 2 mg of sodium borohydride per ml. The activated Sepharose was then washed on a sintered glass funnel with 10 liters of water, and the filter cake, approximately 67 g, was stirred for 3 days a t 37 "C with 100 ml of 0.5 M KLCO.,(pH 11) containing 2.2 mM inhibitor. Residual epoxide was then blocked by treatment with 1 M glycine, pH 10, for 12 h at 37 "C. The gel was washed thoroughly with 0.5 M NaCl and water and was finally stored at 4 "C in 0.1 M NaHC0:r (pH 8.0) containing 1 M NaC1. The concentration of covalently bound inhibitor, estimated using the radioactive compound, was 1.0 mM in the packed gel.
Preparation of Radioactive Pulmonary Angiotensin-converting Enzyme-Fresh rabbit lung (4.4 g) was minced and shaken in 6 ml of Ringer bicarbonate (14) which contained 2 mg of glucose per m l , 300 pCi of L-["S]methionine, and 40 p~ of the other L-amino acids. Incubation was for 6 h at 37 "C in an atmosphere of 5% C02 and 958 02, after which the tissue was homogenized and its angiotensinconverting enzyme was solubilized by treatment with Nonidet P-40 (8). The solubilized enzyme was subjected to affinity chromatography as described below for testicular dipeptidyl carboxypeptidase. The resulting eluate (70 munits/pg) contained 1.95 X 10' cpm/pg and exhibited a single radioactive and Coomassie blue-reactive band after slab gel electrophoresis in the reduced denatured state.
Amino Acid Analyses-Amino acid analyses were performed by the method of Spackman et al. (15) on samples of the enzyme which had been hydrolyzed with constant boiling 5.7 N HCI in evacuated sealed tubes for 24 and 72 h in 110 "C. Methionine and half-cystine were estimated as methionine sulfone and cysteic acid on an aliquot of enzyme that had been oxidized with performic acid (16). Tryptophan was measured after hydrolysis in 4 N methanesulfonic acid (17).
Carbohydrate Analysis-The content of N-acetylglucosamine and N-acetylgalactosamine was determined on the amino acid analyzer after hydrolysis of the enzyme with 4 N HC1 at 100 "C for 5 h. The amino sugars were separated by elution at a flow rate of 44 ml/h with sodium citrate buffer, pH 5. 26 (0.158 N Na') at 70 "C (18). N-Acetylneuraminic acid was estimated after hydrolysis with 0.05 N HzSOi at 80 "C for 1 h as described by Aminoff (19). Other carbohydrate residues were quantitated by gas-liquid chromatography (20).
Protein Determinations-Protein concentrations were determined by the method of Lowry et al. (21) using bovine serum albumin as the standard.
Isolation of mRNA and Cell-free Protein Synthesis-Male New Zealand White rabbits were killed by air embolism, and total RNA was isolated from the excised testes using the guanidine thiocyanate procedure described by Chirgwin et al. (22). RNA prepared in this manner had an AZ~Y,/AXHII value of 1.8-2.0. Poly(A)-containing RNA was then isolated by chromatography on oligo(dT)-cellulose (Type T-3, Collaborative Research, Waltham, MA) (23) and was estimated by using an A'"?,, value of 25.0. This RNA was translated for 90 min at 30 "C in a micrococcal nuclease-treated rabbit reticulocyte lysate system (24). The assay mixture (180 pl, pH 7.4) contained 60 ~1 of lysate (Bethesda Research Laboratories), 25 mM Hepes,' 10 mM creatine phosphate, 48 mM KC1, 87 mM potassium acetate, 1.2 mM MgCI,, 15 pg/ml of creatine kinase, a mixture of unlabeled amino acids (without methionine), 50 FM each, 200 pl of ['%]methionine (1100 Ci/mmol, New England Nuclear), and 30 pg of poly(A)-containing RNA per ml. After cell-free translation, the reaction mixture was adjusted to 2% SDS, boiled for 3 min, and diluted 4-fold with 2.5% Triton X-100, 190 mM NaCI, 50 mM Tris-HC1, pH 7.5, and 6 mM EDTA. Goat anti-pulmonary angiotensin-converting enzyme antiserum (15 pl) was then added, followed by 100 p1 of protein A-Sepharose. The total mixture was incubated overnight with end-to-end rotation. The immune complex-protein A-Sepharose was collected by centrifugation and washed with 20 mM sodium phosphate buffer, pH 7.4, containing 150 mM NaCI, 0.1%' SDS, and 0.5% Nonidet P-40. The antigen was eluted by boiling in 5 mM Tris-HC1, pH 6.8, containing 20% glycerol, 10 mM dithiothreitol, 2% 2-mercaptoethanol, and 5% SDS. The Sepharose was removed by centrifugation, and the supernatant was analyzed by electrophoresis on 7.5% polyacrylamide slab gels in the presence of SDS (25) and by fluorography using E:N.'HANCE (New England Nuclear).

Enzyme Purification-One of three similar preparations
which yielded comparable results is described below and is summarized in Table I  layers of cheesecloth, the homogenate (276 ml) was centrifuged at 16,000 X g for 1 h. The resulting pellet was suspended in 200 ml of Tris buffer, and Nonidet P-40 was added to a final concentration of 0.5% (v/v). The suspension was stirred for 4 h and centrifuged a t 16,000 X g for 1 h. T o the supernatant solution (Nonideet P-40 extract, 200 ml) was added l g of streptomycin sulfate. This treatment was found to be essential for removing nucleic acid which was present in the testicular Nonidet P-40 extract in high quantity and interfered in the subsequent affinity column procedure. After stirring for 20 min, the turbid solution was centrifuged at 16,000 X g, and the streptomycin supernatant (190 ml) was adjusted to 0.1 M Hepes, pH 7.5, 0.3 M KC1, and 0.1 mM ZnC1.. It was then applied at a flow rate of 20 ml per h to an affinity column (1.5 X 60 cm) containing 1 pmol of bound inhibitor per ml of packed volume and equilibrated in the same Hepes buffer. The column was then washed first with 300 ml of buffer containing 0.5% Nonidet P-40 and then with buffer lacking detergent until the absorbance at 280 nm was below 0.01.
Enzyme was finally eluted with buffer containing 0.1 mM inhibitor. Fractions were monitored by their absorbance at 280 nm, and those containing protein were pooled, concentrated by ultrafiltration through a PM-10 filter, and dialyzed for 12-h intervals, first against 3 changes of 4 liters of 10 mM Tris-HC1, p H 7.5, containing 1 mM EDTA, and then against 4 changes of the same buffer lacking EDTA but supplemented with 100 PM ZnC12. After these dialyses, the affinity eluate was concentrated by ultrafiltration to a final volume of 1.0 ml and was subjected to gel filtration through a column (1.0 X 100 cm) of Sephadex G-200 which was equilibrated and developed with 10 mM Tris-HC1, p H 7.5. Fractions were collected at a flow rate of 20 ml per h, and those having a specific activity greater than 160 were combined, concentrated, and stored frozen at -20 "C.
Two protein components were bound by the affinity column and specifically eluted by the inhibitor (Fig.   1). The minor protein (5%) probably corresponds to the pulmonary type of dipeptidyl carboxypeptidase (angiotensin-converting enzyme), which is found in vascular endothelial cells throughout the body (26,271, since it migrated identically with that enzyme during gel electrophoresis in the reduced denatured form with a mobility suggesting a molecular weight of 140,000 and since (not shown) both it and the major protein component were immunospecifically removed from solution by treatment with goat antipulmonary enzyme antibodies and donkey antigoat globulin. When radioactive pulmonary enzyme, which had been purified by the same affinity chromatographic procedure used for testicular dipeptidyl carboxypeptidase, was added to a crude testicular homogenate and the preparation was subjected to the standard isolation procedure, the radioactive protein was not converted to a smaller species (Fig. 1).
This suggests that the testicular type of dipeptidyl carboxy- peptidase does not arise by post-translation proteolytic cleavage of the larger molecule (see below).
During affinity chromatography, approximately 25% of the applied enzyme activity was recovered. The loss of activity was almost certainly not due to residual inhibitor in the eluted enzyme since the specific activity was so high and since pilot experiments indicated that radioactive inhibitor could be completely removed by the dialysis procedure described above. Furthermore, increasing the concentration of inhibitor 5-fold during the elution procedure did not augment the recovery of activity. The affinity column, which had a capacity of 7 units/ ml of packed volume, could be reused after extensive washing with 0.1 M Hepes, pH 7.5, and 1.0 M KC1 and then reequilibrating in the same buffer but with a concentration of 0.3 M KCl.
Physiochemical Properties-The testicular dipeptidyl carboxypeptidase exhibited a mobility corresponding to a molecular weight of 100,000 after slab gel electrophoresis in the reduced denatured state (Fig. 1). Under the same conditions, the molecular weight of the pulmonary enzyme was calculated to be 140,000. The molecular weight of native pulmonary and testicular dipeptidyl carboxypeptidases was estimated by centrifugation of a mixture of the two enzymes on a glycerol gradient using the method of Martin and Ames (28) (Fig. 2). Molecular weight values of 145,000 and 94,000 were calculated Density gradient centrifugation of purified testicular dipeptidyl carboxypeptidase. After determining its approximate sedimentation properties in the absence of other proteins, a sample (0.5 unit) of purified testicular enzyme was mixed with 0.7 unit of purified pulmonary angiotensin-converting enzyme and 0.5 mg of bovine liver catalase in a final volume of 0.2 ml. The mixture was subjected to centrifugation for 18 h at 39,000 rpm a t 2 O C through a 5-20% glycerol gradient (13 ml) in 10 mM Tris-HCI, pH 7.5. using an SW 41 rotor. The tube was then punctured at the bottom, and 26 fractions (0.5 ml) were collected and monitored for enzyme activity with Hip-His-Leu as substrate (0) and for absorbance at 404 nm (0). A, B, and C denote, respectively, the position of catalase, pulmonary angiotensin-converting enzyme, and testicular dipeptidyl carboxypeptidase.
for the native pulmonary and testicular enzymes, respectively.
Arginine (0.6 mol/mol of enzyme) was identified as the NHp-terminal residue of testicular dipeptidyl carboxypeptidase by the Edman procedure (29) followed by acid hydrolysis of the anilinothiazolinone (30). Serine (0.6 mol/mol of enzyme) was identified as the COOH-terminal residue by hydrazinolysis (29). Analysis also yielded the expected corresponding threonyl and alanyl residues from pulmonary angiotensinconverting enzyme, confirming results obtained previously by the dansylation technique and by digestion with carboxypeptidase A (8).
The amino acid compositions of the testicular and pulmonary enzymes are presented in Table 11. The higher content of tryptophanyl residue (68 pg/mg) probably accounts for the extraordinarily high absorbance value a t 280 nm of 2.9 displayed by a solution containing 1.0 mg of enzyme protein per ml.
The carbohydrate component of the testicular enzyme (Table 11) accounted for about 20% of the weight of its combined aminoacyl and sugar residues. The composition differed from that of the pulmonary glycoprotein particularly in its relatively large content of galactosamine and in the absence of fucose.
Immunological studies indicate a high degree of similarity between the two protein chains despite the structural differences described above. In the case of studies using antipulmonary enzyme antibodies, immune preparations from both goat and mouse were employed and yielded essentially identical results. The data obtained with mouse preparations (Figs.  3 and 4) were chosen for presentation since that species was the only one immunized with the pure rabbit testicular enzyme. In a competition radioimmunoassay using antipulmonary enzyme antibodies and radioiodinated pulmonary enzyme,  (1:2200). Those in B represent experiments using the radioiodinated testicular enzyme as the displaceable antigen and antitesticular enzyme antiserum (1:750). Competing antigens were the purified testicular (0) and pulmonary (0) enzymes. the pure pulmonary glycoprotein was much more effective than the testicular enzyme as a competing antigen (Fig. 3A). As expected, in the control experiment carried out identically except that the testicular protein was used as displaceable antigen, the pulmonary and the testicular preparations yielded superimposable displacement curves (not shown). These results indicated that the antipulmonary enzyme antibodies recognized determinants on the pulmonary protein which milliunits) of purified testicular dipeptidyl carboxypeptidase (0) or pulmonary angiotensin-converting enzyme (0) were preincubated at 37 "C in 150 pl of 10 mM Tris-HC1, pH 7.5, containing 0.15 M NaCl, 1 mg of bovine serum albumin per ml, and the indicated amounts of heat-inactivated (56 "C, 30 min) antiserum from a mouse which had been immunized with the pure pulmonary enzyme. After 1 h of preincubation, residual enzyme activity was determined in the standard assay (9) by addition of 100 pl of a solution containing 12.5 mM NaCl and incubation for an additional 30 min.

TABLE I11
Catalytic properties ofpulmonary a n d testicular dipeptidyl carboxypeptidase Pulmonary Testicular -

792
'I Based on a molecular weight of 140,000 and 100,OOO and a weight of 1.08 and 1.21 mg of enzyme per mg of protein measured by the Lowry procedure for pulmonary and testicular enzyme, respectively.
were not present in the testicular polypeptide. In striking contrast (Fig. 3B), when the radioimmunoassay was performed using antitesticular enzyme antibodies and radioiodinated testicular enzyme, identical displacement curves were generated with the preparations from lung and testis, ie. the antitesticular enzyme antibodies did not recognize any determinants of the testicular protein which were not also present in the pulmonary species. These data suggest that the enzyme from testis is immunologically closely related to the pulmonary polypeptide.
Catalytic Properties-The basic catalytic properties of the pulmonary and the testicular enzyme are presented in Table  111. Although the kinetic parameters of the pulmonary enzyme have been reported previously (8), they were redetermined in this study so that meaningful comparisons of the two enzymes, assayed under identical conditions, could be made. The data obtained for both enzymes plotted according to the method of Lineweaver and Burk are presented in Fig. 5. Both enzymes were found to have the same K , for Hip-His-Leu and angiotensin I. Although the testicular enzyme displayed much higher V,,, values for both substrates, the corresponding turnover numbers, based on a molecular weight of 100,000 and a weight of 1.21 mg of enzyme per mg of protein measured by the Lowry procedure, are quite comparable to the pulmonary glycoprotein (see Table 111). Braydkinin (80 nmol) was completely degraded by incubation with 20 milliunits of enzyme at 37 "C for 1 h. Phe-Arg and Ser-Pro, the COOH-terminal and penultimate dipeptidyl residues, were identified as reaction products. Activity of the enzyme with 5 mM Hip-His-Leu was reduced more than 90% by omission of NaCl or by addition of 0.1 mM EDTA and 6.6 milliunits of enzyme activity and was inhibited 50% by the presence of 5.1 x 10"" M N-a-[ l-(S)-carboxy-3-phenylpropylJ-~-lysyl-~-proline (7) or 3.7 X 10"" captopril (31). In addition, the testicular and pulmonary enzymes were inhibited comparably by antipulmonary enzyme antibodies (Fig. 4).
Cell-free Synthesis of Testicular Dipeptidyl Carhoxypeptidase-Polyadenylated mRNA isolated from mature rabbits was translated in a mRNA-dependent rabbit reticulocyte lysate system in the presence of [""S]methionine, and the products were immunoprecipitated using antiserum raised against pulmonary angiotensin-converting enzyme (shown in Fig. 3 to cross-react with testicular enzyme). The immunoprecipitate was analyzed by electrophoresis on polyacrylamide gels in the presence of SDS, and the protein pattern was visualized using fluorography. The immunospecific protein was observed to migrate faster than the corresponding authentic enzyme (Fig.  6) due to the absence of a glycosylation system in the in vitro protein system (32. 33). An apparent M , of 83,000 was calculated for the testicular enzyme synthesized in uitro. This M , is 17,000 less than the value determined using the authentic glycosylated enzyme. The immunospecific radioactive band was found to be competitively eliminated when excess authentic unlabeled testicular enzyme was added to the translation mixture prior to immunoprecipitation. In addition, when serum from nonimmune goat was used for immunoprecipitation of the in vitro translation mixture, no specific bands were observed (data not shown).

DISCUSSION
Dipeptidyl carboxypeptidase activity in the male reproductive system, unlike that elsewhere in the body, appears to be hormonally regulated (1, 4). The catalytic properties of a partially purified fraction of this exopeptidase from semen were found to be similar to those of pulmonary angiotensinconverting enzyme (34). This observation provided the rationale for adapting the recently developed competitive inhibitor of converting enzyme, N-a-[ l-(S)-carboxy-3-phenylpropyl]-~lysyl-L-proline (7), to purification of the testicular dipeptidyl carboxypeptidase, and this strategy has enabled us to obtain the enzyme in pure form.
The polypeptide responsible for dipeptidyl carboxypeptidase activity in most organs and body fluids is structurally and catalytically similar to pulmonary angiotensin-converting enzyme (2). However, our results indicate that the responsible testicular polypeptide is much shorter and contains different NH2-and COOH-terminal residues than the pulmonary enzyme. In view of these structural differences and because there is no known function of angiotensin I1 in the male reproductive system, it seems appropriate to refer to this enzyme as testicular dipeptidyl carboxypeptidase rather than angiotensin-con- verting enzyme.
Despite these structural differences, our data obtained with antibodies against the two pure glycoproteins indicate that they are closely related. Thus their active sites are probably very similar, since antipulmonary enzyme antibodies inhibit the activity of both enzymes identically and since they exhibit indistinguishable requirements, substrate specificities, kinetic properties, and inhibitor profiles. That antipulmonary enzyme antibodies can recognize determinants of the pulmonary protein which are not present in the testicular enzyme is to be expected since the pulmonary polypeptide is so much larger. However, antitesticular enzyme antibodies failed to identify determinants on the testicular enzyme which are absent from the pulmonary glycoprotein. The data from the in vitro translation experiment suggests that the testicular enzyme is not generated post-translationally by proteolysis of pulmonary type angiotensin-converting enzyme.
Furthermore, mRNA isolated from lung and translated in the same reticulocyte lysate system used for the in vitro synthesis of the testicular enzyme, produced an immunoprecipitable translation product with an M , of 129,000 (35). Thus it is likely that the difference in size of the two enzymes may be attributed to the existence of two related genes, a difference in the transcription of the same gene or a difference in the processing of the RNA transcript. It will be especially interesting to determine whether the biosynthetic mechanism responsible for the difference in structure of the testicular dipeptidyl carboxypeptidase also mediates its atypical regulatory properties.