Membrane-bound Angiotensin-converting Enzyme from Rat Lung*

in rat lung pre-dominantly in association with membranous subcellular enzyme from a particulate fraction of rat between deoxycholate purified chromatography. The fraction with converting enzyme activity Sephadex G-200 chromatography an estimated molecular weight of 270,000 observed gel


SUMMARY
Angiotensin-converting enzyme was found in rat lung predominantly in association with membranous subcellular particles.
The converting enzyme was solubilized from a particulate fraction of rat lung that sedimented between 775 and 54,000 x g with sodium deoxycholate and was subsequently purified with DEAE-cellulose and Sephadex G-200 chromatography.
The fraction with converting enzyme activity obtained from Sephadex G-200 chromatography had an estimated molecular weight of 270,000 as observed by gel filtration.
Analytical disc gel electrophoresis of this preparation showed a single band.
The specific converting enzyme activity of the purified preparation was 17.6 pmoles per min per mg with hippurylhistidylleucine as substrate and 1.8 pmoles per min per mg with angiotensin 1 as substrate. This represented a greater than loo-fold purification of the activity of subcellular particles. Sodium chloride was required for activation of the enzyme, and it was inhibited by EDTA, bradykinin potentiating factor-nonapeptide, and angiotensin 2, but not by histidylleucine. The purified preparation was free of carboxypeptidase activity on angiotensin 1. The distribution of converting enzyme activity in subcellular particles from rat lung paralleled that of .5'-AMPase activity which has been associated with pinocytotic vesicles of endothelial membranes. Little or no converting enzyme activity was present in particles obtained from lung lavage.
Angiotensin-converting enzyme which cleaves the terminal dipeptide from angiotensin 1 to form vasoactive angiotensin 2 ( Fig. 1) was first identified in serum by Skeggs ant1 associates in 1954 (1). Recently, it was found that considerably more COILversion of angiotensin 1 takes place during its passage through the pulmonary circulation than occurs in the blood alone (2-4). The enzyme responsible for this conversion has been found in several tissues but appears to be particularly concentrated in the lung (5-9).
This enzyme also has been reported to cleave the * This investigation was sunnorted by Grants HI,06924 and HL14456 from the National Heart and L;ng Institute, National Institutes of Health, United States Public Health Service.
$ Recipient of Research Career Development Award, United States Public Health Service.
terminal dipcptide from bradykinin and has been referred to as a dipeptidyl hytlrolase (6, 10). Purification of the enzyme has been attempted from rabbit, calf, hog, and human lung (11-15); from bovine nrd rat kidney (5, 7); and from hog and human plasma (6,13,16). Although there have been indications that the enzyme is associated with subcellular particles in the lung (17)(18)(19), there have bee:1 few attempts to extract and purify the enzyme tlirertly from particulate elements of the lung. lgic el al. (13) used 0.26ci; deosycl~olate and Sephadex G-W0 clirornato~ral~li~ to purify converting enzyme from a particulate fraction of hog lung. Ilowcver, they did not provide any data for this purifictl prcparatioll, nor did they use it in their reported esperiments.
I)orer et al. (14) also attempted to solubilize convertiur enzyme from a particulate fraction of hog lung by using 0.40/, Triton X-100.
They concluded that this procedure, which provided 4.6 units at a specific activity of 0.64 pmoles per min per mg, was not as successful as their procedure using a supernatant fraction which provided 40 units at a specific activity of 0.75 pmoles per min per mg from the same amount of starting material.
The present study was undertaken to determine the subcellular distribution of this ellzymc in the rat lung and to solubilize and isolate the enzyme from particulate structures of the lung. It was found that the predominant amount of converting enzyme in the rat lung is associated with structural elements, probably of membranous origirl, and that the enzyme can be solubilized from these particles ant1 highly purified by a combination of DEAEcellulose and Scphadex G-200 chromatography.
Furthermore, it n-as found that the distribut,ioll of converting enzyme activity corresponded closely to that of 5'-AMPase activity in fractionated subcellular particles, and that no converting enzyme activity coultl be recovered from cells and particles obtained by lung lavage. The molecular weight of converting enzyme was estimatctl to be about 270,000 by gel filtration. Washes from the animals were pooled. An aliquot of the pool was kept and the remainder was centrifuged at 54,000 X g for 1 hour. The precipitate was resuspended in a small volume of 0.02 M potassium phosphate buffer, pH 8.3. The suspension was homogenized by hand with 30 strokes of a Thomas B hand homogenizer, and all samples were stored at 2" prior to the determination of converting enzyme activity. A Wright-Giemsa stain of a microscopic slide from the suspended particles prior to homogenization showed the presence of mononuclear cells; after hand homogenization, a similarly prepared slide showed only nonspecific particles and nuclei, but no intact cells.
L-Leucine and sodium cholate were from Fisher Scientific Co. Whatman DE52 microgranular DEAE-cellulose was obtained from Reeve-Angel. Sephadex G-200 and the molecular weight marker kit were obtained from Pharmacia Fine Chemicals, Inc. Triton X-100 and acrylamide were from Eastman Kodak Co. Sodium dodecyl sulfate was from Sigma Chemical Co.

Subcellular Distribution
of Converting Enzyme Activity in Rat Lung-Subcellular fractionation of the lung showed that the major portion of converting enzyme activity was associated with particulate elements of the lung (Table I). Only 23% of the total enzymatic activity of the homogenate remained in the supernatant after centrifugation at 54,500 x g for 1 hour, while about 50% of the total protein of the homogenate was in the supernatant.
With some preparations, enzymatic activity in the supernatant was so low that it could not be measured (e.g. Table IV).
The highest specific activity in the subcellular fractions (0.2 pmoles per min per mg) was associated with the P3 fraction.
Examination of this fraction by electron microscopy (Fig. 2) revealed that it contained vesicular membranous structures of various sizes; no mitochondria, nuclei, or other cellular organelles were identified.
Purijication of Converting Enzyme jrom Combined Fractions P2 and PS-For purification procedures, combined Fractions P2 and P3 were used as the source of enzyme.
Passage of the dialyzed sodium deoxycholate extract of these fractions through a column of DEAE-cellulose resulted in the recovery of several protein fractions (Fig. 3). Converting enzyme activity was eluted at 39 to 104 mM KC1 with a continuous 0 to 200 mM KC1 gradient in buffer. At this stage of purification, converting enzyme was stable for 14 days at room temperature and for at least 30 days at 2". The enzyme could withstand repeated freezing and thawing without loss of activity.
Lyophilization, however, resulted in considerable loss of enzymatic activity. With Sephadex G-200 column chromatography a single active peak was eluted shortly after the void volume, at 176 to 200 ml (Fig. 4). At this stage of purification, converting enzyme was stable for at least 3 months at 2". Disc gel electrophoresis of the FIG. 2. Electron microscopic appearance of the 3,500 to 54,500 X g (P3) particulate fraction of homogenized lung. Pellets were fixed in 3% glutaraldehyde, buffered with 0.1 M sodium cacodylate in 5% sucrose, and postfixed in 2y0 aqueous osmium tetroxide, treated with 1% aqueous uranyl nitrate, and placed in 70% alcohol. After rapid dehydration, blocks were embedded in epoxy resin and sections were obtained with a Porter-Blum MT-2 microtome and examined and photographed unstained with a Philips EM-200 microscope.
X 34,000. The fraction was subjected to electrophoresis as described under "Methods." The gel was sliced along its long axis into two pieces. One-half was stained in Coomassie brilliant blue while the other half was cut into thin slices and incubated with Hp-His-Leu under standard assay conditions. Only the slices corresponding to Ihe single heavy band showed converting enzyme activity. fraction with converting enzyme activity obtained from Sephadex G-200 chromatography showed the presence of a single heavy band over a 4-fold concentration range. Incubation of gel slices under standard assay conditions showed that only the slices corresponding to the single heavy band were enzymatically active (Fig. 5).
An assessment of the procedure for purification (Table II) showed that a greater than loo-fold purification of enzymatic activity from combined Fractions P2 and 1'3 had been achieved, and that the specific activity of converting enzyme with Hp-His-Leu as substrate was 17.6 pmoles per min per mg. The final pooled fraction from the Sephadex G-200 column contained about 0.2y0 of the original protein from the combined 1'2 and 1'3 fractions from-lung homogenates (Table II).
Only 25.6% of the total converting enzyme activity was recovered with the purification procedure.
About one-half of the loss occurred with sodium deosycholate extraction.
It appeared that the loss represented failure of extraction rather than denaturation.
Any loss 011 column chromatography could not be accounted for specifically. Since the enzyme was able to withstand exposure to room tem- The following marker proteins (K,,) were used to calibrate the Sephadex G-200 column: ribonuclease A (0.685), chymotrypsinogen A (0.590), ovalbumin (0.422), and aldolase (0.228). Converting enzyme had an elution volume of 185 ml corresponding to I<,, value of 0.123. perature readily, it did not seem that this was a factor in this loss. Small losses did occur in the BioMed UF-1 concentration.
Molecular Weight Determination--By calibration of the Sephadex G-200 column with proteins of known molecular weights, it was determined that the molecular weight of converting enzyme was approximately 270,000 (Fig. 6). Carboxgpeptidase Activity of Fractions with Converting Enzyme Activity during Pzkjkation-The sodium deoxycholate extract of combined Fractions P2 and 1'3 and the DEAE-cellulose and Sephadex G-200 chromatography fractions containing converting enzyme activity were checked for carboxypeptidase activity by incubation with angiotensin Al labeled either in the terminal leucine position with %I or in the 8-Phe position with 14C (see Fig. 1). With [3H]leucine-labeled angiotensin 1, radioactivity was found largely in the His-Leu position after incubation with the sodium deoxycholate extract and chromatographic separation (Fig. 7A). Radioactivity also was found in the leucine position with the sodium deoxycholate extract, indicating the presence of carboxypeptidase activity. However, there was little cleavage of the terminal leucine after the DEAE-cellulose step of purification and no cleavage by the fraction obtained from Sephadex G-200. A similar removal of carboxypeptidase activity with purification was observed when [YJphenylalanine-labeled angiotensin 1 was used as substrate (Fig. 7B). It was presumed, in this case, that radioactivity occurring in the phenylalanine position on paper chromatography was that from the cleaved terminal amino acid of angiotensin 2, phenylalanine, after angiotensin 1 was hydrolyzed.
This does not exclude the possibility that a larger peptide cleavage product separated in the same position as phenylalanine.
Although radioactivity predominated in this position on paper chromatography following incubation with the sodium deoxycholate extract and was present with the DEAEcellulose-separated product containing converting enzyme activity, incubation with the Sephadex G-200 fraction which clearly had converting enzyme activity failed to produce any radioactivity in the phenylalanine position. Angiotensin-converting Enzyme Activity oj Sephadex G-200 Fraction with Angiotensin 1 as Substrate-Angiotensin 1, radioactively labeled in the terminal leucine position, was tested to determine the substrate saturation characteristics of converting enzyme. Maximum specific activity for the enzyme with angiotensin 1 as substrate was about 1.8 pmoles per min per mg (Fig. 8). This was about one-tenth the maximum activity when Hp-His-Leu was used as substrate with the Sephadex G-200 fraction (Table II). Activation and Inhibition of Angiotensin-converting Enzyme-Converting enzyme from rat lung required NaCl for activation and the activity was inhibited in the presence of 1OW RI EDTA (Table III).
Similar results have been observed for the enzyme from rabbit (1 l), calf (la), dog (24), hog (25), and guinea pig lung (25) ; from horse (I), hog (26), and human (16) plasma; and from bovine kidney cortex (7). Converting enzyme also was inhibited by angiotensin 2, and by a bradykinin potentiating peptide from the venom of Bothrops jararaca (Table III) concentrations required for inhibition of 50% of the enzymatic activity (Is,) were 6 X lop5 M for angiotensin 2, and 2 X lo-' M for BP1 ga. Converting enzyme was not inhibited by His-Leu with concentrations up to lop3 M.
Location of Angiotensin-converting Enzyme-The precise location of converting enzyme in the lung, as well as in other tissues, remains unknown. 5'-AMPase activity has been identified by electron microscopy as a marker associated with endothelial cell membranes and with pinocytes of these membranes (27,28). For this reason, we thought it would be of interest to study the distribution of 5'-AMPase activity in the subcellular fractions obtained from the lung and to determine how closely this corresponds to the distribution of converting enzyme activity. Table  IV demonstrates that there is a very close correspondence of converting enzyme and 5'-AMPase activities.
The highest specific activity for both of these enzymes occurred in the particulate fraction which sedimented between 3,100 and 54,500 x g, and Activity--Material obtained from lung lavage was tested for converting enzyme activity to determine if particles containing this enzyme could be removed from the lung in this manner. iMononuclear cells were obtained by this procedure and particles obtained from disruption of these cells and other components of the lavaged material failed to demonstrate the presence of cow verting enzyme activity (Table V). DISCUSSIOS The general concept that many compounds including polypeptidcs, prostaglandins, amines, and nucleotides are altered during passage through the pulmonary circulation is now well established.
The mechanisms for their alterations are variable. In the case of angiotensin 1, a significant alteration occurs through hydrolysis of the terminal dipeptide with the forrnation of vasoactive angiotelisin 2.
The location of the converting enzyme within the lung is not known.
Our studies, as well as those of others (17, 29), indicate that the enzyme is primarily bound to subcellular rnembranous structures.
Two lines of indirect evidence suggest that the enzyme is associated with membranes from endothelial cells. Firstly, when angiotensin 1 is perfused in the pulmonary circulation, the volume of distribution of it and its product, angiotensin 2, is that of the intravascular or extracellular compartrncnt (29, 30). Secondly, 5'-AMPase activity has been found to be associated with pinocytotic vesicles of endothclial cells as determined by histochemical techniques (27)(28)(29), and our studies show that the subcellular distribution of 5'-A;14Pase activity following fractionation of lung tissue parallels that of angiotensin-converting 2317 enzyme activity.
Also, the enzyme was not recovered in lung washes, suggesting that it is not associated with cells free in the alveolar or airway spaces, such as macrophages.
We utilized several methods in attempts to solubilize the enzyme from the combined P2 and 1'3 fractions.
In addition to sodium deosycholate extraction, these included the use of the bile salt sodium cholate, the ionic detergent sodium dodecyl sulfate, and the nonionic detergent Triton X-100.
Although sodium cholate was as effective a solubilizer as sodium deoxycholate, its USC resulted in some loss of enzymatic activity. Sodium dodecyl sulfate at 0.1% resulted in complete loss of enzymatic activity.
Trit>on X-100 at 0.2% was an effective solubilizer and did not inactivate the enzyme; however, adequate removal of Triton X-100 could not be accomplished readily. Converting enzyme was found to leach into 0.02 M phosphate buffer at pH 8.3 when the combined P2 and P3 fractions were left standing at 2" for several days. Storing of the homogenized pellets at 2" for 30 days resulted in the leaching of as much converting enzyme and protein into the phosphate buffer as could be obtained with sodium deoxycholate treatment of fresh preparations.
The converting enzyme obtained by DEAE-cellulose and Sephadcx G-200 chromatography appears to be highly purified as determined by gel electrophorcsis.
Carboxypeptidase activity, which was associated with the crude lung subcellular particles, was removed during the purification process.
Our final preparation had a specific activity of 17.6 pmoles per min per mg. This approximates the value of 22 pmoles per min per mg reported for a preparation from rabbit lung by Cushman and Chcung (9) and is somewhat higher than the value of 10 pmoles per min per mg reported by Darer et al. (14) for converting enzyme from hog lung. Converting enzyme from rat lung has an activity with Hp-His-Leu as substrate about 10 times higher than that with angiotensin 1 as substrate.
This is qualitatively in agreement with the Hp-His-Leu to angiotensin 1 ratio of 18 reported for the rabbit lung enzyme (9) and 2.7 reported for the calf lung enzyme (12), but differs from the value of 0.11 reported for the converting enzyme from dog lung (31).
The molecular weight of the converting enzyme from rat lung was estimated to be 270,000. This compares with the following values obtained from other species: hog lung, 300,000 (14) ; calf lung, 300,000 to 330,000 (12) ; human lung, 480,000 (15) ; human plasma, 150,000 (16); hog and guinea pig plasma and lung, 150,000 (25); and porcine plasma, 155,000 (26). The estent of molecular similarity of the enzyme from these various sources remains to be determined.
Our findings of activation of the enzyme by chloride and inhibition by ElYl'A are similar to those observed for the enzyme from rabbit (11)) calf (12)) dog (24)) hog (25)) and guinea pig (25) lung; from horse (l), hog (26), and human (16) plasma; and from bovine kidney cortex (7). The inhibit.ion data for angiotensin 2 and UPI-',, cannot be directly compared with values obtained for other converting enzymes, except as discussed below, because of differences in assay conditions, substrates, and substrate concentrations.
Nevertheless, there seems to be a close correlation between the various reported values. The concentration required for inhibition of 5094 of rat lung converting enzyme (ZbO) by anpiotensin 2 was 2 X 1OW M. This is qualitatively similar to the f4,, value of lop4 M obtained with [Asp',Valj]angiotensin 2 on the enzyme from porcine plasma with angiotensin 1 as substrate reported by Lee et al. (26). These workers also observed that His-Leu at 1O-3 M did not inhibit the porcine plasma enzyme. The Is,, for BPP9, with rat lung converting enzyme was 2 X IO-' M. This value can be directly compared with an 1~~ value of 5.6 x lo-' III observed for the enzyme from rabbit lurlg, since the assay systems were identical (32). An Iso of 1.43 x lo-' also has been reported for the enzyme from an extract of dog lung using angiotensin 1 as substrate (33). BPPs, inhibits the con version of angiotensin 1 to angiotensin 2 in the isolated lungs of guinea pig, dog, and rat (34).