THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 250, No. 17, Issue of September 10, pp. 6762-6768, 1975

Printed in U.S.A.

Pulmonary Angiotensin-converting Enzyme STRUCTURAL AND CATALYTIC PROPERTIES*

(Received for publication, March 31, 1975)

MANJUSRI DAS AND RICHARD L. SOFFER~

From the Department of Molecular Biology, Division of Biological Sciences, Albert Einstein College of Medicine, Bronx, New York 10461

Angiotensin-converting enzyme has been solubilized from a particulate fraction of rabbit lung and purified to apparent homogeneity in 11% yield by a procedure including fractionation with DEAE-cel- lulose and calcium phosphate gel, elution from Sephadex G-200, and lectin affinity chromatography. The molecular weight estimated by equilibrium sedimentation was approximately 129,000, either in the absence or presence of 6 M guanidine hydrochloride. A slightly higher value of 140,000 determined for the reduced, denatured protein by gel electrophoresis in the presence of sodium dodecyl sulfate and a much higher figure derived from gel filtration are probably due to the glycoprotein nature of the enzyme. Its oligosaccharide content accounted for 26% of the weight calculated from its amino acid and carbohydrate composition. The estimated content of sugar residues per mole was: galactose, 57; N-acetylglucosamine, 53; mannose, 43; N-acetylneuraminic acid, 19; and fucose, 4. Threonine and alanine were identified, respectively, as NH,-terminal and COOH-terminal residues by the dansylation procedure and by digestion with carboxypeptidase A. The enzyme was found to contain approximately 1 g atom of zinc per mol.

K, values for hydrolysis of hippurylhistidylleucine and angiotensin I were 2.3 and 0.07 mM, and the corresponding turnover numbers were 15,430 and 792 mol/min/mol at 37”. Bradykinin was also a

substrate, and release of its COOH-terminal dipeptide, Phe-Arg, was catalyzed at a comparable rate to that of His-Leu from the COOH terminus of angiotensin I. Enzyme activity required the presence of chloride ions and was inhibited by EDTA and by low concentrations of Bothrops bradykinin-potentiating peptides. In addition, hydrolysis of hippurylhistidylleucine was inhibited competitively by other defined peptides, including di- and tripeptides, which were not substrates.

Angiotensin-converting enzyme (kininase II, EC 3.4.15.1) is a mammalian COOH-terminal dipeptidyl hydrolase (for a recent review see Ref. 1). It was first detected in equine plasma (2) as a chloride ion-dependent, EDTA-sensitive activity catalyzing release of His-Leu from the COOH terminus of angiotensin I and thereby generating angiotensin II, a potent vasopressor octapeptide. Although converting activity is known to be present in many organs (3, 4), the level of angiotensin II in the systemic circulation is thought to be regulated by hydrolysis of angiotensin I in the lungs (5). The rapidity of this hydrolysis in. uiuo (5, 6) and the morphologic appearance of fractionated membranes with converting activ- ity (7, 8) have suggested that the responsible enzyme may be a constituent of the capillary endothelial membrane in direct apposition to the pulmonary circulation.

We recently developed a method for obtaining small

*This work was supported by Grant HL. 17741 from the National Heart and Lung Institute. This paper is the second in a series on angiotensin-converting enzyme and the regulation of vasoactive pep- tides. The first paper in this series is Ref. 9.

$ Faculty Research Associate of the American Cancer Society.

amounts of pure angiotensin-converting enzyme after solubili- zation from rabbit pulmonary membranes (9). It was estab- lished that this enzyme was a glycoprotein and that it could also catalyze release of the COOH-terminal (Phe-Arg) and adjacent (Ser-Pro) dipeptides from bradykinin, a hypotensive nonapeptide. Since bradykinin-potentiating peptides from Bothrops venom (10) were known to inhibit angiotensin conversion in vitro (11) and in uiuo (12), these results indicated

that the same membranous pulmonary enzyme might partici- pate in the maintenance of systemic blood pressure by two mechanisms, i.e. by catalyzipg both the formation of a pressor peptide, angiotensin II, and the inactivation of a depressor peptide, bradykinin.

Our procedure has now been adapted to the isolation of substantial amounts of apparently homogeneous enzyme. In this communication we report data suggesting that the glyco- protein possesses a molecular weight of approximately 129,000, that it contains a single polypeptide chain with threonine and alanine as the NH,-terminal and COOH-terminal residues, and that it is associated with one molar equivalent of bound zinc. Some catalytic properties are also described.

6762

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6763

EXPERIMENTAL PROCEDURE

Materials

Frozen type III rabbit lungs were from Pel-Freeze. Commercial products used in the enzyme purification included Nonidet-P40 (Shell Chemicals), DEAE-cellulose (Bio-Rad) which was cycled in 1 N HCl and 1 N NaOH prior to equilibration in buffer, calcium phosphate gel (Nutritional Biochemicals), and Sephadex G-200 (Pharmacia). Sepha- rose 2B (Pharmacia), containing approximately 4 mg of covalently bound ricin agglutinin per ml bed volume, was prepared by the method of Adair and Kornfeld (13) and was a gift from Dr. Stanley Nathenson of this institution.

Unlabeled angiotensin I and dithiothreitol were Calbiochem prod- ucts. Angiotensin I labeled with carbon-14 in the COOH-terminal leucine residue (333 nCi/pmol) was from New England Nuclear. Gua- nidine hydrochloride, (l-aspartate, 5isoleucine) angiotensin II and Pyr- Lys-Trp-Ala-Pro (“Bradykinin-potentiating factor”) were products of Schwarz/Mann. Pyr-Trp-Pro-Arg-Pro-Gln-Ile-Pro-Pro (SQ 20881) was a gift from Dr. Z. P. Horovitz of Squibb. Hippurylhistidylleucine (Hip-His-Leu) was from Research Plus. A standard kit of dansylated amino acids and bradykinin triacetate were purchased from Sigma. Other peptides were obtained from Bachem or Cycle Chemicals. Materials for disc gel electrophoresis were Fisher products and carboxypeptidase A was from Worthington.

Methods

Enzyme Assays and Kinetic Analyses-The standard assay was the spectrophotometric procedure described by Cushman and Cheung (14) using Hip-His-Leu as substrate. A unit of enzyme activity is the amount required to catalyze formation of 1.0 wmol of hippuric acid per min at 37” under their standard conditions.

Conversion of angiotensin I labeled with carbon-14 in the COOH-ter- minal leucine residue was estimated by the release of radioactive His-Leu as described previously (9).

Reaction mixtures containing bradykinin and other possible peptide substrates were examined by paper electrophoresis with a potential gradient of 50 volts per cm for 120 min in pyridine/acetic acid/water (l/10/89), pH 3.5. The dried strips were developed with 0.3% ninhy- drin.

Gel Electrophoresis-Gel electrophoresis under denaturing condi- tions was performed on 5 to 20% acrylamide gradients with electrode buffer containing 0.1% sodium dodecyl sulfate in 0.025 M TrisJ0.192 M

glycine, pH 8.3 (15). The proteins were initially reduced and denatured by boiling for 2 min in 2% sodium dodecyl sulfate and 5% 2-mercapto- ethanol. Slabs were stained for protein with Coomassie blue (16). For standard electrophoresis of native proteins the conditions were identi- cal except for the absence of sodium dodecyl sulfate.

Sedimentation Equilibrium-Molecular weights were estimated at 20” by the meniscus depletion method of Yphantis (17) in a Beckman model E ultracentrifuge equipped with interference optics and a temperature control unit. A 3-mm column of solution containing 0.3 mg of protein per ml was used in a 12.mm double sector aluminum- coated Epon cell. FC-43 fluorocarbon was placed in the bottom of the cell. For determining the subunit value the dissociating conditions described by Marshall and Cohen (18) were employed.

Amino Acid Analysis-Amino acid analyses were performed accord- ing to the method of Spackman et al. (19). Aliquots of enzyme (80 pg) were hydrolyzed with constant boiling 5.7 N HCl in evacuated, sealed tubes for 24 and 72 hours at 110”. Methionine and half-cystine were estimated as methionine sulfone and cysteic acid, respectively, on a sample of the enzyme that had been oxidized with performic acid (20). Tryptophan was determined by a minor modification of the method of Liu and Chang (21). The enzyme was hydrolyzed with 4 N methanesul- fonic acid and the hydrolysate was applied to the short column of the amino acid analyzer. The usual procedure for short column runs was modified to achieve a good separation of tryptophan and glucosamine. Elution was carried out with 0.21 M sodium citrate, pH 5.4, at 40” and no buffer change was made.’

Carbohydrate Composition-Analysis of the carbohydrate residues was carried out by gas-liquid chromatography as described by Coding- ton et al. (22).

COOH-terminal Analysis-A sample of enzyme (4.5 mg) in 2.2 ml of 0.2 M N-ethylmorpholine acetate, pH 8.5, was digested at 25” with 15 ~1 of a solution containing 6 mg of carboxypeptidase A per ml in 2 M LiCl,

‘C. Y. Lai, unpublished procedure.

0.2 M N-ethylmorpholine acetate, pH 8.5 (23). Aliquots (0.5 ml) were removed at various time intervals, treated immediately with 0.13 ml of 1 N HCl to stop the reaction, and subjected to amino acid analysis.

NH,-terminal Analysis-The dansylationZ procedure was carried out as described by Gray (24). After acid hydrolysis the dansyl derivatives were analyzed by poly amide thin layer chromatography.

Metal Analysis-The metal analyses were performed with a model 303 Perkin-Elmer atomic absorption spectrophotometer. Before analy- sis the enzyme was dialyzed extensively against 20 mM Tris-HCl, pH 7.13. Determinations of zinc, manganese, and cobalt content were made at 215.1, 280.9 and 241.0 nm, respectively, using a 10x expansion scale. The absorbances were compared with a blank containing the dialysis buffer.

Protein Determination-The method of Lowry et al. (25) was employed using crystalline bovine serum albumin as the standard.

RESULTS

Enzyme Purification

A preparation from 7.8 kg of starting material is described below and summarized in Table I. Six batches of lungs, each approximately 1.3 kg, were processed separately through the DEAE-cellulose step and then combined prior to subsequent procedures. All operations were performed at O-4”. Essentially identical results were obtained with two large scale prepara- tions.

Preparation of Homogenate and 16,000 x g Fraction- Frozen rabbit lungs (1.3 kg) were chopped into small pieces and suspended in 2600 ml of a buffer containing 20 mM Tris-HCl, pH 7.8, 30 ITIM KCl, 5 mM magnesium acetate, and 0.25 M sucrose. The suspension was homogenized 3 times for 45 s in a Waring Blendor with l-min cooling intervals in a water-ice bath and then filtered through two layers of cheese- cloth. The homogenate (3500 ml) was centrifuged at 16,000 x g in a GSA rotor for 1 hour and the pellet was dispersed in 10 mM

Tris-HCl, pH 7.8. The dispersed pellet is called the 16,000 x g fraction.

Preparation of Nonidet-P40 Extract-The 16,000 x g frac- tion (2100 ml) was made 0.5% with respect to Nonidet-P40, agitated vigorously with a magnetic stirrer for 3 hours, and then centrifuged at 16,000 x g for 1 hour. The supernatant solution is referred to as the Nonidet-P40 extract.

DEAE-cellulose Fractionation-The NonidetP40 extract (1700 ml, 22.8 g of protein) was mixed with a slurry (4500 ml) of DEAE-cellulose (packed volume 1900 ml) equilibrated in 10 mM potassium phosphate buffer, pH 6.5, containing 0.5% Nonidet-P40. After stirring for 45 min the suspension was filtered through 2 sheets of Whatman No. 1 filter paper on a Buchner funnel. The filter cake was suspended in 4 liters of the same phosphate buffer containing 0.025% Nonidet-P40, stirred for 30 min and then refiltered. Enzyme was then eluted with 0.2 M KC1 by dispersing the filter cake in 2.5 liters of phosphate buffer containing 0.025% Nonidet-P40 and 0.35 M KCl. After stirring for 45 min the mixture was filtered and elution was repeated with 2 liters of buffer containing 0.2 M KCl. The combined filtrates were concentrated approximately 8-fold by ultrafiltration in a Diaflo apparatus with a PM 10 filter and then dialyzed extensively against 1 mM potassium phosphate, pH 7.5, containing 0.025% Nonidet-P40.

Calcium Phosphate Gel Fractionation-The DEAE-cellulose eluate was fractionated with calcium phosphate gel in two steps. The procedure depends on the fact that activity is not adsorbed to the gel in the first step, whereas a large fraction of

2The abbreviation used is: dansyl, 5-dimethylaminonaphthalene-l- sulfonyl.

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6764

TABLE I

Purification of angiotensin-cowerting enzyme from 7.8 kg of rabbit lungs

Fraction

Homogenate 16,000 x g fraction

Nonidet-P40 extract DFAE-cellulose eluate

(cone) First calcium phosphate

gel supernatant (cone) Second calcium phosphate

gel eluate (cone) Sephadex G-ZOO, peak II

(cone) Ricin-agglutinin-Sepharose

eluate (cone)

Volume

1 x 10-2

210 126 102

33.6

3.00

0.455

0.400

0.240

-

m

1

I

Protein

g x 10-3

721 331 137

48.7

4.30

1.04

0.118

0.080

Specific activity

nitslmg protein

0.09 0.18 0.42

0.84

7.7

21

67

89

contaminating proteins are. After removal of these extraneous proteins in the first step, the enzyme is adsorbed to the gel in the second step. In each step different batches of gels behaved somewhat differently, and careful pilot experiments were critical in optimizing yield and purification.

Six batches of enzyme from the DEAE-cellulose step were combined (3360 ml, 48.7 g) and treated with a slurry (4640 ml) containing 375 g (dry weight as solids) of calcium phosphate gel in 1 mM potassium phosphate, pH 7.5. The mixture was stirred for 1 hour and then centrifuged for 30 min at 10,000 x g. The supernatant solution containing the enzyme was concentrated by ultrafiltration. This first calcium phosphate gel supernatant fraction (300 ml, 4.30 g) was then treated with a second slurry (900 ml) containing 75 g of gel in water. The suspension was stirred for 1 hour, then centrifuged at 16,000 x g for 30 min and the pellet was extracted similarly with 1200 ml of water. Activity was eluted from the washed pellet by extracting twice with 1000 ml of 10 mM potassium phosphate, pH 7.5. The active extracts were combined and concentrated to 200 ml by ultrafiltration through a PM 10 membrane. They were then diluted to 2000 ml with 10 mM Tris-HCl, pH 7.8, and reconcentrated lo-fold. This procedure was used to eliminate residual detergent and was performed a total of 6 times, until the filtrate no longer contained material absorbing at 230 nm when read against the Tris buffer. The eluate was finally concentrated to 46 ml and centrifuged at 16,000 x g for 1 hour to remove a small, loosely packing precipitate.

Gel Filtration on Sephadex G-200-The concentrated second calcium phosphate gel eluate was applied to a column (105 x 9 cm) of Sephadex G-200 which was equilibrated and developed with 10 mM Tris-HCl, pH 7.8. Fractions (20 ml) were collected at a flow rate of 40 ml per hour. The recovery of activity was quantitative; however, it was approximately equally divided into two well resolved protein components (Fig. 1). The first of these (Peak I) was present in the void volume of the column and was not significantly purified with respect to the input material. The second component, Peak II (2820 to 3220 ml, VeIV, - 1.3) was concentrated by ultrafiltration and dialyzed against 10 mM Tris-HCl, pH 7.4, containing 0.15 M NaCl. The elution volume of this fraction corresponded to that of a protein of molecular weight approximately 300,000, as estimated by the method of Andrews (26).

Affinity Chromatography on Immobilized Ricin Ag- glutinin-The concentrated Peak II Sephadex filtrate was subjected to chromatography on a column (1 x 13 cm) of Sepharose 2B containing covalently bound ricin agglutinin and

EFFLUENT VOLUME (ml)

FIG. 1. Sephadex G-ZOO gel filtration of the second calcium phos- phate gel eluate. Details of the procedure are described in the text.

equilibrated in 10 mM Tris-HCl, pH 7.8/0.15 M NaCl. This lectin specifically recognizes galactose residues (27). Because the capacity of the column was only 800 enzyme units, chromatography was carried out 10 times with 4 ml of enzyme solution. After application the sample was allowed to equili- brate for 1 hour in the column which was then washed with 60 ml of starting buffer. Elution was performed with 0.1 M lactose in the Tris-NaCl solution. About 6 ml of eluting buffer were washed into the column and allowed to equilibrate for 2 hours to ensure complete dissociation of the enzyme-lectin complex. The enzyme was then eluted with an additional 30 ml of the lactose-containing buffer. The combined lactose eluates were concentrated by ultrafiltration and dialyzed extensively against 10 mM Tris-acetate, pH 7.4. The concentrated ricin- agglutinin-Sepharose eluate appeared faintly yellow and pos- sessed a specific activity similar to that of homogeneous enzyme obtained by our previously published procedure (9). It retained its full activity after storage at -20” for 6 months.

Physical and Chemical Characteristics of Purified Enzyme

Gel Electrophoresis-The purified enzyme exhibited a single protein band after gel electrophoresis under standard condi- tions (Fig. 2). Activities for the hydrolysis of Hip-His-Leu, angiotensin I, and bradykinin are known to reside in this protein, since they have previously been shown to be co-eluted from the gel slice in which it was present (9). The enzyme preparation also showed a single major protein component after denaturation and electrophoresis in the presence of sodium dodecyl sulfate. In confirmation of our previous results (9), the molecular weight of the reduced, denatured enzyme was estimated to be approximately 140,000 by this technique.

The behavior of the enzyme in the excluded Sephadex G-200 fraction (Peak I) was also examined (Fig. 2). It failed to enter the separating gel under standard conditions. When it was compared with an equal number of units of purified enzyme under reducing, denaturing conditions, its principal polypep- tide possessed the same mobility, shape, and staining intensity as the enzyme protein. Only minor amounts of other Coomassie blue-reactive components were visible despite the fact that the specific activity of this fraction was only 20% that of the purified enzyme. A possible explanation for these results may be that this fraction contains the same enzyme glycopolypep- tide complexed to other glycoproteins which fail to stain with

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

92,000 67,000 45,000

24,000

FIG. 2. Polyacrylamide disc gel electrophoresis under standard and denaturing conditions. Lanes A and B contained, respectively, 5 fig (445 milliunits) of purified enzyme and 25 fig (450 milliunits) of the Sephadex G-200 Peak I fraction which were subjected to electrophore- sis on a 5 to 20% acrylamide gradient slab under standard conditions. Lanes C, D, and E contained, respectively, 3 fig (267 milliunits) of purified enzyme, 15 pg (270 milliunits) of the Sephadex G-200 Peak I fraction and 2 to 4 pg of molecular weight markers (phosphorylase a, albumin, ovalbumin, fl-casein, and cytochrome c). Electrophoresis of the reduced, denatured proteins was carried out on a 5 to 20% acrylamide gradient in the presence of 0.1% sodium dodecyl sulfate.

Coomassie blue (28). Consistent with this interpretation is the fact that all of the activity in the fraction and more than 80% of its protein could be bound by Sepharose containing covalently bound ricin agglutinin or concanavalin A. Like the purified

enzyme, the excluded Sephadex fraction was active on angi- otensin I and bradykinin in addition to Hip-His-Leu. Further- more, antibody raised in goats against the purified enzyme displayed quantitatively and qualitatively identical inhibitory kinetics with both enzyme preparations. Attempts to further purify activity from this fraction have been unsuccessful.

Optical Characteristics-The enzyme exhibited a typical protein spectrum in the ultraviolet range (not illustrated) with an absorption maximum at 280 and minimum at 250 nm. A solution containing 1 mg of enzyme protein,per ml possessed optical densities of 8.78, 1.10, and 1.95 at 230, 260, and 280 nm, respectively. The high latter value presumably reflects the large content of tryptophan residues in this protein (Table II). Under the assumptions stated in Table II, the molar extinction coefficient of the enzyme at 280 nm can be calculated as 2.3 x 105. The enzyme had no detectable absorbance in the visible region at a concentration of 1 mg/ml.

Molecular Weight-The plots of log fringe displacement against r2 from the sedimentation equilibrium analyses under native or dissociating conditions (Fig. 3) showed good linearity, indicating a high degree of mass homogeneity. The molecular weights derived from these data were 129,000 and 128,500 under the native and denaturing conditions, respectively. These values were obtained using a partial specific volume of 0.705 ml/g determined (29) from the amino acid and carbohy- drate composition. They agree well with our original estimate of 136,000 for the native enzyme derived from a sedimentation constant of 7.9 S during glycerol gradient centrifugation (9).

6765

TABLE II Amino acid and carbohydrate composition of angiotensin-converting

enzyme

Amino acid or carbohydrate .g per mg proteina

Lysine Histidine Arginine Aspartic acid ThreonineC SerineC Glutamic acid Proline Glycine Alanine Half-cystined Valine- Methioninee Isoleucine Leucine Tyrosine Phenylalanine Tryptophanf

Total FUCOSe

Mannose

Galactose N-acetylglucosamine N-acetylneuraminic acid

Total

38.1 30.1 55.6 83.7 33.1 35.8 95.8 41.0 22.7 42.7 11.7 41.0 28.5 31.9 80.5 51.6 48.4 35.9

808 4.8

58.4 78.3 91.0 46.6

279

Nearest integer residues per mleb

35 26 42 a6 38 49 88 50 47 71 13 49 26 33 a4 38 39 23

a37 4

43 57 53 19

176

o Values for amino acids are those obtained after 72 hours hydrolysis. b Calculated using a molecular weight of 129,000 and the weight sum

of amino acid and carbohydrate residues. If these residues account for the entire weight, then 1 mg of protein measured by the Lowry procedure corresponds to 1.087 mg of enzyme.

c Corrected for decomposition by extrapolation of 72. and 24-hour values to zero time.

d Determined as cysteic acid. ‘Determined as methionine sulfone. ‘Determined after hydrolysis with 4 N methanesulfonic acid.

r2km2) FIG. 3. Sedimentation equilibrium of the purified enzyme at 20”

under standard and dissociating conditions. A, centrifugation was carried out for 20 hours at 16,200 rpm in 0.1 M sodium phosphate, pH 7.1; B, centrifugation was carried out for 20 hours at 19,160 rpm in 20 mM sodium phosphate, pH 7.0, containing 6 M guanidine hydrochloride and 1 mM dithiothreitol. Further details are described under “Meth- ods.”

This,agreement provides additional evidence for the conclusion (9) that the V,/V,, value of 1.3 observed with gel filtration on Sephadex G-200 is influenced by the glycoprotein nature of the

enzyme and yields a spuriously high estimate of molecular weight.

Amino Acid and Carbohydrate Composition (Table IO- Aromatic residues (phenylalanine, tyrosine, tryptophan) ac-

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6766

counted for 11.9% of the total amino acids, a value 45% higher than the average derived from a large number of sequenced proteins (30). The fraction of hydrophobic residues calculated according to Heller (31) was 44%, while the ratio of polar to apolar residues (32) was 1.2. These results indicate that this membrane-bound enzyme possesses a moderately high degree of hydrophobicity.

directly determined K, and V,,, values of 0.85 PM and 1.4 pmol/min/mg of protein for bradykinin using a highly purified enzyme preparation from porcine lungs. The corresponding figures reported by them for angiotensin I were 30 FM and 2.1 /Imol/min/mg.

Sugar residues were found to constitute approximately 26% of the total weight of amino acid and carbohydrate. This oligosaccharide content is considerably higher than that which we reported originally (9). Part of the discrepancy is probably due to the fact that our early value was based on the dry weight of a very small sample of the glycoprotein. However, the content of N-acetylneuraminic acid relative to the other sugars is clearly higher in the present preparation than in that obtained by our previous procedure. It is not unlikely that further study will disclose some heterogeneity in the oligosac- charide moiety of the enzyme.

COOH-terminal Analysis-The kinetics of amino acid ap- pearance during digestion with carboxypeptidase A (Fig. 4) indicated the presence of a COOH-terminal alanine residue. Serine and leucine were found in the next two positions; however, the results did not permit an unequivocal assignment of their sequence. No additional amino acids were found when carboxypeptidase B was included in the reaction mixture. For the data shown in Fig. 4, the amount of alanine after digestion for 1 hour corresponded to only 0.63 residue per mol of enzyme, as estimated using the assumptions described in Table II. However, when the digestion was performed for the same time period at 37” in buffer containing sodium dodecyl sulfate (33), the release of alanine, serine, and leucine corresponded, respectively, to 1.09, 1.02, and 1.16 residues per mol. Under these conditions several other amino acids were found in smaller quantities.

Activity with Hip-His-Leu as substrate was more than 90% dependent upon the presence of NaCl and was markedly inhibited by 0.1 mM EDTA and by low concentrations of bradykinin-potentiating peptides (Table III). The NaCl re- quirement could be met with M&l, but not with Na,SO,. p-Hydroxymercuribenzoate (0.1 mM) did not alter the reaction velocity. Similar data have been described for angiotensin-con- verting enzyme from many sources (1); however, in contrast to the results of Cheung and Cushman (36), we found the nonapeptide to be slightly more inhibitory than the pentapep- tide. Since the enzyme contains 13 half-cystine residues, its diminished activity in the presence of sulfhydryl compounds may be due to reduction of an important intrachain disulfide bridge rather than to heavy metal chelation, as has previously been suggested (14).

Phe-Arg, Ser-Pro, and angiotensin II, which are produced during the hydrolysis of bradykinin and angiotensin I, were found to inhibit the enzyme at fairly low concentrations

6 t

NH,-terminal Analysis-Dansyl-threonine, N’-dansyl-lysine and o-dansyl-tyrosine were the only derivatized amino acids detected on polyamide thin layer chromatograms after hydrol- ysis of the dansylated protein.

Heavy Metal Analyses-The zinc content of the enzyme, as determined by atomic absorption spectroscopy, was 10.0 nmol per mg of protein, corresponding to 1.19 g atoms per mol of enzyme. No manganese or cobalt was detected.

MINUTES

FIG. 4. Release of amino acids from purified angiotensin-converting enzyme by digestion with carboxypeptidase A. The procedure was carried out at 25” as described under “Methods.” O-O, alanine; O-O, serine; A-A, leucine.

Catalytic Properties of Purified Enzyme TABLE III

K, values for Hip-His-Leu and angiotensin I were 2.3 and 0.07 mM. The corresponding V,,,,, values were 130 and 6.67 pmol/min/mg of protein. These results correspond to h,,, values of 15,430 and 792 mol/min/mol of enzyme using the assumptions described in Table II. The paper electrophore>is technique for detecting ninhydrin-reactive products of brady- kinin degradation (9) is not sufficiently quantitative to permit precise measurement of these parameters. However, a peptide analyzer was employed to establish that the enzyme liberated 4.6 pmol of the COOH-terminal dipeptide Phe-Arg/min/mg of protein from 1.6 mM bradykinin.3 Bradykinin is thus hydro- lyzed almost as rapidly as angiotensin and its k,,, value can be calculated as 546 mol/min/mol of enzyme assuming the determination to represent a maximal rate. Both bradykinin and angiotensin I were found to inhibit enzymatic release of hippuric acid from Hip-His-Leu. Their respective concentra- tions required to yield a 50% decrease in the rate of hydrolysis of 5 mM Hip-His-Leu were 9 and 50 FM. Dorer et al. (35) have

Catalytic properties of arzgiotensin-conuerting enzyme

Addition or deletion Inhibit iona

%

- NaCl + 0.1 @ EDTA + 1.0 @ Bradykinin-potentiating

pentapeptideb

92 96

+ 0.75 clM_ Bradykinin-potentiating nonapept ideC

+ 1.0 I+ Dithiothreitol + 0.04 W- Angiotensin II + 1.0 ti- His-Leu + 1.0 @ Phe-Arg + 0.1 mM_ Ser-Pro

61

66 75 49 19 56 44

3 L. J. Greene, unpublished procedure.

“The complete system contained 0.1 M potassium phosphate, pH 8.3, 0.3 M NaCl, 5.0 mM Hip-His-Leu, and 8.6 milliunits of enzyme per 250 ~1 of reaction mixture. Incubation was at 37” for 30 min.

b Pyr-Lys-Trp-Ala-Pro. c Pyr-Trp-Pro-Arg-Pro-Gln-Ile-Pro-Pro.

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6767

whereas His-Leu, the COOH-terminal dipeptide of angiotensin I, was less effective. A more systematic investigation of the action of small peptides was carried out with the compounds shown in Table IV. Kinetic analyses indicated that all of these peptides were competitive inhibitors of enzyme activity. In the series Ala-Ala, Ala-Ala-Ala, Ala-Ala-Ala-Ala there was no alteration in the K, value with increasing chain length. However, only the tetrapeptide was found to be substantially hydrolyzed when 250 nmol of peptide were incubated with 20 milliunits of enzyme for 30 min and examined by paper electrophoresis. Acetylation of the (r-NH, group of the peptide decreased its inhibitory activity while simultaneously render- ing it more susceptible to the action of the enzyme. For example, N-acetyl-Ala-Ala-Ala was extensively hydrolyzed, yet possessed a considerably higher K, value than the unsub- stituted tripeptide. Amongst the dipeptides the lower K, values for Lys-Ala, Arg-Ala, and Tyr-Ala compared to those for Asp-Ala, Glu-Ala, and Leu-Ala suggest that the affinity of the enzyme for basic and aromatic peptides may be higher than that for peptides with acidic and branched side chains.

DISCUSSION

The present procedure for isolating angiotensin-converting enzyme after solubilization from lung particles has provided us with substantial amounts of material which is sufficiently pure for structural studies. The molecular weight of 129,000 esti- mated for the native enzyme by equilibrium sedimentation agrees well with the value obtained by glycerol gradient centrifugation (9) and provides additional evidence that the higher figure derived by gel filtration is due to the large oligosaccharide content of this protein (9). Our results are therefore consistent with those of others who have determined molecular weights for the lung enzyme of 270,000 to 300,000 based on elution volumes from Sephadex G-200 (8, 39, 40).

The molecular weight of the enzyme under denaturing conditions determined by equilibrium sedimentation or disc gel electrophoresis was found to be comparable to that of the native protein. The slightly higher figure with gel electrophore- sis is probably due to the oligosaccharide moiety, since this technique (38), like gel filtration (26), is known to yield anomalously high values for glycoproteins. Only threonine and alanine were detected as NH,-terminal and COOH-terminal residues, respectively, by dansylation and by digestion with carboxypeptidase A. These data are consistent with the results

TABLE IV Peptide inhibitors of angiotensin-converting enzyme

Peptide Kia

mM -

Ala-Ala 0.08 Ala-Ala-Ala 0.08 Ala-Ala-Ala-Ala 0.08 N-acetyl-Ala-Ala 4.0 N-acetyl-Ala-Ala-Ala 0.30 N-acetyl-Ala-Ala-Ala-Ala 0.13 Lys-Ala 0.05 Arg-Ala 0.06 Asp-Ala 0.50 Glu-Ala 1.30 Leu-Ala 0.40 Tyr-Ala 0.06

I “Values were determined by the method of Dixon (37) using

Hip-His-Leu as substrate.

of molecular weight determinations in suggesting that the enzyme contains a single, large poiypeptide chain. It thus appears to differ from a soluble porcine pulmonary angioten- sin-converting enzyme which was reported to be composed of three polypeptide subunits, each of molecular weight 70,000 (41), but to resemble somewhat a membranous converting enzyme from porcine kidneys thought to possess a molecular weight of 195,000 and to contain a single polypeptide chain with about 8% neutral sugars (42).

Angiotensin-converting enzyme has been suspected to be a metalloenzyme since its discovery (2). Our finding that it contains a stoichiometric amount of tightly bound zinc is consistent with the observation that manganous, cobaltous, and zinc ions, respectively, restored 40, 160, and 100% of activity to a rabbit lung acetone powder extract which had been inactivated by dialysis against EDTA (14). This COOH- terminal dipeptidyl hydrolase thus resembles other exopepti- dases, such as carboxypeptidase A (43), carboxypeptidase B (44), and leucine aminopeptidase (45), in its content of zinc.

When relatively crude preparations of angiotensin-convert- ing activity have been subjected to gel filtration, two or more peaks of activity, one of which is in the void volume, have often been observed (41, 46, 47). In at least one instance, enzyme in the excluded fraction has differed from the retained activity in not catalyzing bradykinin degradation (48). In our preparation the two activity peaks in the gel filtration step do not appear to represent different polypeptide chains. Enzyme in the ex- cluded fraction, like the purified enzyme, was active on bradykinin. It was inhibited by antibody prepared against the purified enzyme and it migrated on gels in the reduced, denatured form identically to the purified enzyme. However, it failed to penetrate standard gels. A possible explanation for this activity is that it is due to the same glycopolypeptide enzyme, which in this fraction is tightly bound to one or more other proteins. Conceivably such a complex might derive from topographical relationships in the original membrane structure and its nonenzymatic components might play an important role in regulating activity in uiuo.

Little is known about the substrate specificity of angioten- sin-converting enzyme other than the fact that it does not act upon peptides with protected COOH termini and does not catalyze hydrolysis of peptide bonds containing the imino group of a prolyl residue (1). Our studies indicate that a large number of peptides which need not necessarily be substrates can act as competitive inhibitors of the enzyme. The results suggest a low affinity for peptides containing acidic or branched chain amino acids. The lower affinity for angiotensin I as compared with bradykinin (35), which does not contain ctcidic or branched chain amino acids, may be explicable on this basis. Similarly, we have found that human fibrinopeptide A, which contains a high proportion of acidic amino acids (49), inhibits the enzyme only at very high concentrations (K,-

1.1 mM) and is not a substrate. Bothrops peptide inhibitors of angiotensin-converting en-

zyme have been shown to transiently alleviate experimental hypertension in rats thought to be due to renin-dependent overproduction of angiotensin I (50). Encouraging clinical results have recently been obtained with the nonapeptide in human disease (51). Since the physiologically important activ- ity is thought to be luminal in the pulmonary vascular bed (5, 6), it should be accessible to the action of circulating mac- romolecules. In this context it is worth noting that we have obtained antiserum with anticatalytic activity from goats

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6768

immunized with the purified enzyme. It will be interesting to determine whether this antiserum can protect rabbits against the hypertensive effect of intravenously administered angio- tensin I.

Acknowledgments-We are grateful to Dr. John Codington, Department of Biological Chemistry, Harvard Medical School, for performing the carbohydrate analysis and to Dr. Lewis Greene, Brookhaven Laboratories, for determining the specific activity of the enzyme with bradykinin as substrate. In addition, we thank Dr. Robin Briehl of this institution, Dr. C. Y. Lai, Roche Institute of Molecular Biology, and Dr. Roberta Colman, Department of Chemistry, University of Delaware, for helping us in their laboratories with equilibrium sedimenta- tion, amino acid analysis, and determination of heavy metal content, respectively.

1.

2.

3. 4.

5. 6.

7.

8.

9.

10. 11. 12. 13.

14.

15.

16.

17. 18.

19.

20.

REFERENCES

Bakhle, Y. S. (1974) in Angiotensin (Page, I. H., and Bumpus, F. M., eds) pp. 41-80, Springer-Verlag, New York

Skeggs, L. T., Kahn, J. R., and Shumway, N. P. (1956) J. Exp. Med. 103,295-299

Huggins, C. G., and Thampi, N. S. (1968) Life Sci. 7,633-639 Cushman, D. W., and Cheung, H. S. (1971) Biochim. Biophys.

Acta 250, 261-265 Ng, K. K. F., and Vane, J. R. (1967) Nature 216,762-766 Ryan, J. W., Smith, U., and Niemeyer, R. S. (1972) Science 176,

64-66 Sander, G. E., and Huggins, C. G. (1971) Nature New Biol. 230,

27-29 Lanzillo, J. J., and Fanburg, B. L. (1974) J. Biol. Chem. 249,

2312-2318 Soffer, R. L., Reza, R., and Caldwell, P. R. B. (1974) Proc. Natl.

Acad. Sci. U. S. A. 71, 1720-1724 Ferreira, S. H. (1965) Brit. J. Pharmacol. Chemother. 24, 1633169 Bakhle, Y. S. (1968) Nature 220,919-921 Ng. K. K. F., and Vane, J. R. (1970) Nature 225,1142-1144 Adair. W. L.. and Kornfeld. S. (1974) J. Biol. Chem. 249.

4696-4704 Cushman, D. W., and Cheung, H. S. (1971) Biochem. Pharmacol.

20,1637-1648 Baum, S. G., Horwitz, M. S., and Maizel, Jr., J. V. (1972) J.

Viral. 10,211-219 Weber, K., and Osborn, M. J. (1969) J. Biol. Chem. 244,

4406-4412 Yphantis, D. A. (1964) Biochemistry 3,297-317 Marshall, M., and Cohen, P. P. (1972) J. Biol. Chem. 247,

1641-1653 Spackman, D. H., Stein, W. H., and Moore, S. (1958) Anal.

Chem. 30,1190-1206 Moore, S. (1963) J. Biol. Chem. 238,235-237

21.

22.

23. 24. 25.

26. 27.

28.

29.

30.

31. 32. 33. 34.

35.

36.

37. 38.

39.

40.

41.

42.

43.

44.

45.

46.

47. 48.

49.

50.

51.

Liu, T. Y., and Chang, Y. H. (1971) J. Biol. Chem. 246, 284222848

Codington, J. T., Sanford, B. H., and Jeanloz, R. W. (1972) Biochemistry 11, 2559-2564

Ambler, R. P. (1967) Methods Enzymol. 11,436-444 Gray, W. R. (1972) Methods Enzymol. 25, 121-138 Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

(1951) J. Biol. Chem. 193,265-275 Andrews, P. (1965) Biochem. J. 96,595-606 Nicolson, G. L., Blaustein, J., and Etzler, M. E. (1974) Bio-

chemistry 13,196-204 Fairbanks, G., Steck, T. L., and Wallach, D. F. H. (1971)

Biochemistry 10,2606-2616 Gibbons, R. A. (1966) in Glycoproteins (Gottschalk, A., ed) pp.

29-95, Elsevier, Amsterdam Davhoff. M. 0. (1972) Atlas of Protein Seauence and Structure

i, 355-359 Heller, J. (1966) Biochemistry 7,2906-2913 Hatch, F. T. (1965) Nature 206,777-779 Buhler, D. R. (1963) J. Biol. Chem. 238, 1665-1674 Lineweaver. H.. and Burk. D. (1934) J. Am. Chem. Sot. 56,

658-666 Dorer, F. E., Kahn, J. I-I., Lentz, K. E., Levine, M., and Skeggs,

L. T. (1974) Circulation Res. 34.824-847 Cheung: H. S., and Cushman, D. W. (1973) Biochim. Biophys.

Acta 293,451-463 Dixon, M. (1953) Biochem. J. 35,170-171 Allen, R. H., and Mehlman, C. S. (1973) J. Biol. Chem. 248,

3660-3669 Dorer, F. E., Kahn, J. R., Lentz, K. E., Levine, M., and Skeggs,

L. T. (1972) Circulation Res. 31,356-366 Stevens, R. L., Micalizzi, E. R., and Pals, D. T. (1972) Biochemistry

11,2999-3007 Nakajima, T., Oshima, G., Yeh, H. S. J., Igic, R., and ErdSs,

E. G. (1973) Biochim. Biophys. Acta 315,430-438 Oshima. G.. Geese. A.. and ErdGs, E. G. (1974) Biochim. Biophys.

Acta 350,‘26-37 Vallce. B. L.. and Neurath. H. (1954) J. Am. Chem. Sot. 76,

5006-5007 Folk, J. E., Piez, K. A., Carroll, W. R., and Gladner, J. A. (1960)

J. Biol. Chem. 235,2272-2277 Carpenter, F. H., and Vahl, J. M. (1973) J. Biol. Chem. 248,

294-304 Cushman, D. W., and Cheung, H. S. (1972) in Hypertension ‘72

(Genes& J., and Koiw, E., eds) pp. 532-541, Springer-Verlag, New York

Fits, A., and Overturf, M. (1972) J. Biol. Chem. 247,581-584 Grandino, A., and Paiva, A. C. M. (1974) Biochim. Biophys. Acta

364,113-119 Blomblick, B., Blombiick, M., Grondahl, N. J., Guthrie, C., and

Hixton, M. (1965) Acta Chem. Stand. 19,1788-1789 Krieger, E. M., Salgado, H. C., Assan, C. J., Greene, L. J., and

Ferreira. S. H. (1971) Lancet 1.269-271 Gavras, H., Brunner, H. R., Laragh, J. H., Sealey, J. E., Gavras,

I., and Vukovich, R. A. (1974) New Engl. J. Med. 291,817-821

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

M Das and R L SofferPulmonary angiotensin-converting enzyme. Structural and catalytic properties.

1975, 250:6762-6768.J. Biol. Chem. 

  http://www.jbc.org/content/250/17/6762Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/250/17/6762.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on May 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from