Biochimica et Biophysica Acta, 439 (1976) 125-132

© Elsevier Scientific Publishing Company, Amsterdam -- Printed in The Netherlands BBA 37389 THE ESTIMATION AND COMPARISON OF M O L E C U L A R WEIGHT Ob ANGIOTENSIN I CONVERTING ENZYME BY SODIUM DODECYL SULFATE-POLYACRYLAMIDE GEL ELECTROPHORESIS J. J. LANZILLO and B. L. FANBURG New England Medical Center Hospital and the Department of Medicine, Tufts University School oJ Medicine, Boston, Mass. 02111 (U.S.A.)

(Received January 5th, 1976)

SUMMARY The angiotensin I converting enzyme from rat lung was observed to be a glycoprotein containing 8.3 ~o carbohydrate and consisting of a single polypeptide chain with an estimated molecular weight of 139000 as determined by sodium dodecylsulfate-polyacrylamide gel electrophoresis and 150 000 by sucrose density gradient sedimentation analysis. A comparison of the mobility of angiotensin I converting enzyme from rat lung, rabbit lung, and two hog lung sources on sodium dodecyl sulfate-polyacrylamide gels indicates that all four enzymes have very similar molecular weights and subunit structures. Some previously reported molecular weight discrepancies appear to be due to anomalous behavior of the enzyme on gel filtration.

INTRODUCTION Although the angiotensin I converting enzyme (kinase II, peptidyl dipeptide hydrolase, peptidyl dipeptidase, EC 3.4.15.1) has been purified from several sources and some of its physical properties determined [1], there is still much controversy with respect to its molecular weight and subunit structure. Molecular weights ranging from 129 000 to 480 000 have been reported (Table I). These values may reflect differences related to techniques used to estimate molecular weight, differences related to purification schemes, or differences in species specific enzyme structure. In at least one case, technique and species specific differences do not explain the observed results. The values reported for hog lung angiotensin I converting enzyme as determined by Sephadex gel filtration were 300 000 [3], 206 000 [6] and 150 000 [10]. The recently reported discovery that some angiotensin I converting enzymes are glycoproteins [7, 12] offers an explanation for some of the molecular weight discrepancies, since glycoproteins often show anomalous behavior when gel filtration is used to determine molecular weight [13]. The sodium dodecyl sulfate (SDS)-polyacrylamide gel technique of Segrest and Jackson [14] minimizes the influence of sugar moieties on molecular weight determiAbbreviation: SDS, sodium dodecyl sulfate.

126 nation by utilizing a range of acrylamide concentrations for electrophoresis. In this study, we have estimated the molecular weight of angiotensin I converting enzyme from rat lung using this technique as well as sucrose density gradient sedimentation. This value is compared with the molecular weight as determined by Sephadex G-2t30 gel filtration. We also have compared angiotensin I converting enzyme from lungs of other species with that of the rat using SDS-polyacrylamide gel electrophoresis. METHODS

Preparation of angiotensin I converting enzyme Rat lungs were prepared and subcellular fractions were separated as previously reported [5]. For each preparation, combined fractions which sedimented between 3100 and 54 000 × g were collected from 40 animals and were homogenized with a VirTis homogenizer in 250 ml of 0.02 M potassium phosphate buffer, pH 8.3. The homogenate was allowed to stand at 2 °C for 20 days since this procedure was found to yield as much converting enzyme from subcellular particles as did treatment with deoxycholate [5]. Further purification was carried out as previously described [5] except that the concentrate was pumped onto the Sephadex G-200 column as a single charge rather than as a series of small portions. Marker proteins were used as previously described to calibrate the Sephadex G-200 column [5].

Assay for angiotensin I converting enzyme activity Converting enzyme activity was determined using hippuryl-L-histidyl-L-leucine (Hip-His-Leu) (Vega-Fox Chemical Co.) as substrate as previously described [5]. Protein was determined by the method of Lowry et al. [15].

Disc gel electrophoresis (a) Polyacrylamide gels. Electrophoresis was performed according to Maurer and Allen (Allen No. 1 system) using 8 ~ non-gradient polyacrylamide gels cast in Tris-sulfuric acid, pH 9.0 [16]. Fractions from the Sephadex G-200 column were used without further concentration. 200/~1 (5 to 20 #g of protein) of each fraction in 0.02 M potassium phosphate buffer, pH 8.3, and approximately 10 ~ in glycerin were routinely applied to the gels. Gift samples of angiotensin I converting enzyme from rabbit lung (a gift from Dr. D. W. Cushman, Squibb Institute for Medical Research, New Brunswick, N.J.) and two hog lung sources (gifts from Dr. F. E. Dorer, the Veterans Administration Hospital, Cleveland, Ohio, and from Dr. E. G. Erd0s, University of Texas, Southwestern Medical School, Dallas, Texas, respectively) were diluted with 0.02 M potassium phosphate buffer, pH 8.3, so that approximately 10/~g of protein could be applied to each gel in a 200/zl volume. The power setting was 1 mA/tube for 15 min and then 3 mA/tube for 90 min. Following electrophoresis, gels were stained for glycoproteins with periodic acid-Schiff reagent using a modification of the method of Segrest and Jackson [14]. For this procedure, the gels were removed from the running tubes and immersed overnight in 12.5 % trichloroacetic acid. The gels were rinsed once with 7 ~ acetic acid, oxidized for 1 h with 1 ~ periodic acid in 3 ~ acetic acid, rinsed with 7 ~ acetic acid, immersed for 30 min in 0.5 ~ potassium metabisulfite in 3 ~ acetic acid, rinsed with 7 ~ acetic acid, and stained for 2 h at 5 °C with Schiff reagent (Fisher Scientific

127 Co.). Glycoprotein bands appeared pink against a clear background; the patterns were recorded immediately. Periodic acid-Schiff staining of converting enzyme samples was always performed in the presence of control gels subjected to the same electrophoretic treatment as the gels which contained converting enzyme. Each control gel contained 50 #g of protein and was stained in both the oxidized and unoxidized modes. Four different control proteins were used including two glycoproteins, ovalbumin (Worthington Biochemical Corp.) and human glycoprotein fraction IV (Schwarz/Mann); and two non-carbohydrate proteins, aldolase and chymotrypsinogen A (Worthington Biochemical Corp.). Coomassie Brilliant Blue R-250 was used according to Maurer and Allen to detect non-carbohydrate proteins [16]. Gels were destained with a Bio-Rad Model No. 172 diffusion destainer in H20/ethanol/acetic acid (65:25:8, v/v). (b) Sodium dodecyl sulfate (SDS)-polyacrylamide gels. 5 ~ SDS-polyacrylamide gels were prepared according to Weber et al. [17]. Angiotensin I converting enzyme samples were prepared from aliquots of the Sephadex G-200 fraction and gift samples. Gift samples of converting enzyme were diluted as described above. To each 200/zl aliquot was added 2 mg SDS, 48 mg urea, and 2/~12-mercaptoethanol. Each sample was heated at 60 °C for 1 h. The power setting was 8 mA/tube for 3.5 h. Gels were stained with periodic-Schiff reagent and Coomassie Brilliant Blue as described above. (c) Molecular weight studies on SDS-polyacrylamide gels. 4 to 8 ~ SDSpolyacrylamide gels were prepared according to Weber et al. [17]. 25 #g per gel of the following marker proteins were used: myosin (heavy chain) (a gift from Dr. Y. Yazaki, University of Tokyo, Faculty of Medicine, Hongo, Tokyo, Japan), phosphorylase A (Boehringer-Mannheim), catalase (P.L. Biochemicals, Inc.), y-globulin and serum albumin (Schwarz-Mann), aldolase, and ovalbumin (Worthington Biochemical Corp.). Marker proteins were dissolved in 0.02 M potassium phosphate buffer, pH 8.3. Angiotensin I converting enzyme samples and marker protein samples were treated with SDS, urea, and 2-mercaptoethanol as described above. Electrophoresis was performed at room temperature with a power setting of 8 mA/tube for 3.5 h for the 4 - 6 ~ gels, for 4 and 6 h for the 7 ~ gels, and for 6 and 24 h for the 8 ~o gels. The data were treated as specified by Segrest and Jackson [14]. The mobility was plotted against log molecular weight for the marker proteins in the 4-8 ~o gels. Observed angiotensin I converting enzyme molecular weights were obtained from these curves and plotted against acrylamide concentrations.

Sucrose density gradient sedimentation of angiotensin I converting enzyme Sucrose density gradient sedimentation was carried out on 5-20~ sucrose density gradients in 0.05 M Tris. HC1 buffer, pH 7.2. Ovalbumin and aldolase were used as marker proteins, and calculation of molecular weight was performed as described by Martin and Ames [18].

Analysis of angiotensin I converting enzyme for neutral sugar A 40 ml sample of angiotensin I converting enzyme from the Sephadex G-200 column was concentrated to 5 ml (Millipore 25 mm U-F cell with a PSED Pellicon

128

i9

18

t6

Io I 4

I 5

I 6

I 7

I 8

ACRYLAMIDECONCENTRATION(%) Fig. 1. Plot of observed molecular weights of angiotensin I converting enzyme fractions from rat lung vs. acrylamide concentration on 4-8 % SDS-polyacrylamide gels. Each point is the mean of 4 determinations. Calibration of gels was carried out as described in Methods.

membrane). The concentrate was pumped onto a 2.5 × 75 cm column of Bio-Gel P-300 (Bio-Rad Laboratories) equilibrated with 0.02 M potassium phosphate, pH 8.3. Elution was performed at 9 ml/h with this same buffer. Enzymatically active samples from the column were pooled, lyophilized, and analyzed for total neutral sugar by the anthrone method of Roe [19]. Converting Enzyme Activity o---.o Ovolbumin

k..

: : Converting Enzyme Activity o.--.o Aldolose

P. 1.0

0.4 \

0.8

Air t

0.6 0.4

it/

0.2 0

I

1 Bottom

I

I

5

I

I

9

I

I

13

I

i

0.3

0.2

t I

17

~

0.1 1

I

21 Top

I I I Bottom

I 5

I

I 9

I

I I 13

I I 17

I 21 Top

TUBE NUMBER Fig. 2. Sucrose density gradient sedimentation analysis of the heavy fraction of rat lung angiotensin I converting enzyme performed as described by Martin and Ames [18]. An average molecular weight of 150 000 was observed using ovalbumin (left) and aldolase (right) as marker proteins. Converting enzyme activity is expressed as absorbance at 228 nm following the assay with Hip-His-Leu as described in Methods.

129

Fig. 3. Gel electrophoresis of angiotensin I converting enzyme samples as described in Methods. Top, 8 ~ polyacrylamide gels; bottom, 5 % SDS-polyacrylamide gels. From left to right: rabbit lung, hog lung (Dorer), hog lung (Erd6s), rat lung. The arrow points to precipitated SDS. RESULTS

Molecular weight determination by SDS-polyacrylamide gel electrophoresis T h e m o l e c u l a r weight o f angiotensin I converting enzyme varied f r o m 185 000

130 in a 4~o gel to 139 000 in an 8~o gel (Fig. 1). The value on the 8~o gel was essentially unchanged from that on the 7 ~o one. The angiotensin I converting enzyme behaved like other glycoproteins in this system by showing an anomalously high molecular weight value that decreased significantly as the percentage of acrylamide in the gel increased.

Molecular weight by sucrose density gradient sedimentation The molecular weight of angiotensin I converting enzyme as determined by sucrose density gradient sedimentation was observed to be 150 000 (Fig. 2).

Sugar analysis Angiotensin 1 converting enzyme from rat lung contains 0.083 mg carbohydrate per mg protein, or 8.3 ~o. This value agrees well with the 8 ~ sugar content for hog lung angiotensin I converting enzyme previously reported by Oshima et al. [6].

Electrophoresis of angiotensin I converting enzymes from rat, rabbit and hog lung A comparison of the mobilities of the angiotensin I converting enzymes from rat, rabbit, and two sources of hog lung (Fig. 3) indicates that all of these angiotensin I converting enzymes are similar in molecular weight and subunit structure. Angiotensin I converting enzyme from rat lung, as well as the two hog lung angiotensin I converting enzymes, stained with periodic acid-Schiff reagent indicating that all are glycoproteins; however, the rabbit lung angiotensin I converting enzyme did not take the periodic acid-Schiff stain. In view of the observation that rabbit lung angiotensin I converting enzyme is a glycoprotein with 26~ sugar [12], we assume that our negative results were due to an insufficient amount of protein on the gel since the Coomassie Blue band for protein also was very faint. DISCUSSION Our results indicate that angiotensin I converting enzymes in lungs from several animal sources including rat, rabbit, and hog, have similar molecular weights. The enzyme is a glycoprotein consisting of either a single polypeptide chain or of subunits that are not readily dissociable. Anomalous behavior on gel filtration and on SDS-polyacrylamide gel electrophoresis at a single acrylamide concentration could account for the discrepancies in molecular weight previously reported (Table I), including our finding of a 270 000 molecular weight for rat lung angiotensin I converting enzyme as determined by Sephadex G-200 gel filtration [5]. That the behavior on Sephadex G-200 is indeed anomalous is indicated by our molecular weight value of 150 000 for the native rat lung angiotensin I converting enzyme obtained by sucrose density gradient sedimentation. Softer et al. [12] have made a similar observation with angiotensin I converting enzyme from rabbit lung. Their determination of molecular weight by SDS-polyacrylamide gel electrophoresis agreed closely with that by sedimentation equilibrium analysis and with our value for angiotensin I converting enzyme from rat lung. Although the possibility that angiotensin I converting enzyme appears as a polymer on Sephadex G-200 cannot be entirely eliminated, our experience suggests that this is not the case. We have re-chromatographed unconcentrated samples of rat lung

131 TABLE I REPORTED MOLECULAR WEIGHTS OF ANGIOTENSIN I CONVERTING ENZYME Source [ref.]

Molecular weight

Human lung [2] Hog lung [3] Calf lung [4] Rat lung [5] Hog lung [6]

480 000 300 000 300 000 270 000 206 000

Bovine kidney cortex [7] Hog plasma [8] Human plasma [9] Hog and guinea pig lung [10] Guinea pig plasma [11] Rabbit lung [12]

195 000 155 000 150 000 150 000 150 000 140 000 129 000

Subunits

m

m

3 × 70 000 none

none none

Method for molecular weight determination Sephadex G-200 Sephadex G-200 Sephadex G-200 Sephadex G-200 Sephadex G-200 5 ~ SDS-polyacrylamide 5 ~ SDS-polyacrylamide Sucrose density gradient Sucrose density gradient Sephadex G-150, sucrose density gradient Sephadex G-150 Gradient SDS-polyacrylamide Sedimentation equilibrium analysis

angiotensin I converting enzyme with concentrations of NaC1 up to 1 M with no observable change in its apparent molecular weight of 270 000. Our results with SDS-polyacrylamide gel electrophoresis of hog lung angiotensin I converting enzyme indicate that the differences in reported molecular weights are not related to idiosyncrasies in the purification schemes. It is interesting to note that our results for rat lung angiotensin I converting enzyme on a 5 ~o SDS-polyacrylamide gel showed an apparent molecular weight of 180 000, similar to that of 206 000 observed by Nakajima et al. [6] for the enzyme from hog lung on a 5 ~o SDS-polyacrylamide gel. The variation in molecular weights at different acrylamide concentrations indicates that molecular weight observations of glycoproteins on SDSpolyacrylamide gel at a single acrylamide concentration are potentially inaccurate. We were unable to obtain the subunit pattern for this enzyme as reported by Nakajima et al., who observed a molecular weight of 206 000 and a subunit structure of three equivalent polypeptide chains of approximately 70 000 molecular weight each. Oshima et al. [7], however, did not observe any subunit structure for the angiotensin I converting enzyme from bovine kidney cortex although they reported its molecular weight to be similar to that of the hog l u n g enzyme. ACKNOWLEDGMENT

We wish to thank Ms. Leslie Wells for her skillful technical assistance with these experiments. Sugar analysis was performed through the courtesy of Dr. Morris C y n k i n by Ms. M a r g a r e t K n o w l t o n . REFERENCES 1 Erd6s, E. G. (1975) Circ. Res. 36, 247-255 2 Fitz, A. and Overturf, M. (1972) J. Biol. Chem. 247, 581-584 3 Dorer, F. E., Kahn, J. R., Lentz, K. E., Levine, M. and Skeggs, L. T. (1972) Circ. Res. 31,356-366

132 4 Stevens, R. L., Micalizzi, E. R., Fessler, D. C. and Pals, D. T. (1972) Biochemistry 11, 2999-3007 5 Lanzillo, J. J. and Fanburg, B. L. (1974) J. Biol. Chem. 249, 2312-2318 6 Nakajima, T., Oshima, G., Yeh, H. S. J., Igic, R. and Erd/Ss, E. G. (1973) Biochim. Biophys. Acta 315, 430-438 7 0 s h i m a , G., Gecse, A. and Erd6s, E. G. (1974) Biochim. Biophys. Acta 350, 26-37 8 Lee, H. J., LaRue, J. N. and Wilson, I. B. (1971) Biochim. Biophys. Acta 235, 521-528 9 Lee, H. J., LaRue, J. N. and Wilson, I. B. (1971) Arch. Biochem. Biophys. 142, 548-551 l0 Lee, H. J., LaRue, J. N. and Wilson, I. B. (1971) Biochim. Biophys. Acta 250, 549-557 11 Grandino, A. and Paiva, A. C. M. (1974) Biochim. Biophys. Acta 364, 113-119 12 Das, M. and Softer, R. L. (1975) J. Biol. Chem. 250, 6762-6768 13 Andrews, P. (1965) Biochem. J. 96, 595-606 14 Segrest, J. P. and Jackson, R. L. (1972) Methods Enzymol. 28, 54-63 15 Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951) J. Biol. Chem. 193. 265-275 16 Maurer, H. R. and Allen, R. C. (1972) Z. Klin. Chem. Klin. Biochem. 10, 220-225 17 Weber, K., Pringle, J. R. and Osborn, M. (1972) Methods Enzymol. 26, 3-27 18 Martin, R. G. and Ames, R. N. (1961) J. Biol. Chem. 236, 1372-1379 19 Roe, J. H. (1955) J. Biol. Chem. 212, 335-343