TIIJ.: Joumar. OF Bro~ocrcn~ C~MIBTBY Vol. 247, No. !J, hue of May 10, pp. 2835-2841, 1972

Printed in U.S.A.

Effect of Phospholipase C Hydrolysis of Membrane

Phospholipids on Membranous Enzymes*

(Received for publication, November 22, 1971)

RICHARD D. MAVISJ ROBERTM.BELL,§ AND P. ROY VAGELOS

From the Department of Biological Chemistry, Washington University School of Medicine, Xt. Louis, Missouri 63110

SUMMARY

The response of several Escherichia coli membranous enzymes to hydrolysis of up to 95 % of membrane phospho- lipid has been investigated. Purified phospholipase C of Bacillus cereus was utilized in these studies. The rate and extent of digestion of E. coli phospholipids was independent of whether the lipid was associated with membrane protein or extracted from membranes and sonically dispersed. Phosphatidylethanolamine and phosphatidylglycerol were completely hydrolyzed, while cardiolipin was partially resist- ant to hydrolysis by phospholipase C. Acyl-CoA:glyceroI 3-phosphate acyltransferase and NADH oxidase were in- activated at a rate very similar to the rate of hydrolysis of total lipids. Acyl-CoA: I-acylglycerol 3-phosphate acyl- transferase was inactivated to an extent of 50% during hydrolysis of 50% of membrane phospholipid. The remain- ing activity was stable to continued hydrolysis of phospho- lipid. Glycerol 3-phosphate dehydrogenase and succinic dehydrogenase remained completely active after hydrolysis of 95 % of membrane phospholipids. These results show the heterogeneity of membranous enzymes with respect to their dependence upon the presence of intact membrane phospho- lipids. The lack of effect of membranous lipid-protein inter- actions on the accessibility of phospholipids to phospholipase C hydrolysis was shown in general by the similarity in the rates of hydrolysis by phospholipase C of membranous and isolated phospholipids. More specifically, the similarity in the rate of hydrolysis of phospholipids and the rates of inac- tivation of certain membranous enzymes suggests that these enzymes are dependent on phospholipids whose susceptibility to phospholipase C hydrolysis is similar to the bulk of mem- brane phospholipids.

The apparent independence of the bulk of membrane protein and the bulk of membrane lipid as indicated by the work of

* This investigation was supported in part by Grant GB-5142X from the National Science Foundation and Grant l-ROl-HE-10406 from the National Institutes of Health.

$ National Heart Institute Postdoctoral Fellow I-ROB-HE- 34846-2.

5 National Science Postdoctoral Fellow 40027.

Glaser et al. with spectroscopic probes (1) reveals the need for studying membranous lipid-protein interactions at a more spe- cific level. That such interactions exist is obvious from the phospholipid requirements of several purified membranous en- zymes (2-8) as well as from the dependence of temperature characteristics of various membranous enzyme and transport activities on the phase properties of membrane lipid (9-13). These interactions are not detectable spectroscopically in intact membrane, presumably because they do not represent the bulk of membrane protein or possibly because they are associated with small amounts of phospholipid not detectable by the meth- ods used in such experiments. Therefore a more specific probe, such as enzymic activity, would appear to be useful in studying the role of lipid-protein interaction in membrane function. We have recently reported differences in the response of various membranous enzymes of Escherichia coli to changes in mem- brane fatty acid composition (9). The heterogeneity of lipid- protein interaction revealed by that study suggested that further investigation of the phospholipid requirements of these en- zymes would be of interest. We have recently shown that these enzymes have normal specific activities in membranes of an E. coli mutant with altered phospholipid composition.1

The phospholipid requirements of several membranous en- zymes have been shown by various methods, the most conclu- sive of which is purification of the enzyme to the extent that activation by added phospholipid can be shown. The extreme difficulty encountered in the purification of some membranous enzymes has led to the utilization of more indirect methods. Phospholipase digestions have been utilized with several sys- tems (14-19), but interpretation of these results requires cau- tion. Phospholipase A produces fatty acids and lysophospha- tides, compounds whose surfactant properties are known to inhibit several enzymes nonspecifically. Thus, enzyme inhibi- tion caused by phospholipase A can be due to the surfactant properties of the hydrolysis products rather than to the removal of required phospholipids. The products of phospholipase C hydrolysis, diglyceride and water-soluble phosphoryl com- pounds, do not have surfactant properties. Thus inactivation of a membranous enzyme by phospholipase C can be more safely interpreted as suggesting the requirement of intact phos- pholipid. The purity of phospholipase preparations is critical in such a demonstration. Contaminating proteases or other

1 R. 111. Bell, R. D. Mavis, and P. R. Vagelos, manuscript sub- mitted for publication.

2835

2836 E$ect of Phospholipase C on Membranous Enzymes Vol. 247, No. 9

activities may inactivate membranous enzymes independently of phospholipid hydrolysis. In the present work, a highly pu- rified preparation of phospholipase C from Bacillus cereus (20) was used to investigate the effect of phospholipid hydrolysis on several membranous enzymes of E. coli. The use of membrane preparations from 32P-labeled cells provided a convenient method of following lipid hydrolysis concurrently with the spectropho- tometric determination of several enzymic activities.

EXPERIMENTAL PROCEDURE

!~ateriaZs-Glycerol-3-P, 3-(4,5-dimethyl thiazolyl-2)-2,5- diphenyl tetrazolium bromide, phenazine methosulfate, 5,5’- dithiobis(2-nitrobenzoic acid), deoxyribonuclease, phospholi- pase D, NADH, NAD+, glucose-6-P, and glucose-6-P dehy- drogenase were purchased from the Sigma Chemical Co., St. Louis, MO. Palmityl-CoA and oleyl-CoA were from P-L Bio- chemicals, Milwaukee, Wis. 1-Acylglycerol 3-phosphate was prepared by treatment of 1-acylglycerol 3-phosphorylcholine with phospholipase D (21). 1-Acylglycerol 3-phosphorylcho- line (prepared by treatment of egg lecithin with phospholipase ‘4) was purchased from Cycle Chemical, Los Angeles, Calif. [32P]Phosphoric acid was purchased from New England Nuclear.

Growth of E. coli and Preparation of 32P-Labeled Membrane Fraction-E. coli strain H139 was grown at 37” in 1% Difco Bacto Tryptone, 0.5% sodium chloride, 0.2% glycerol, 1.7 pM

thiamine, 2.0 pM pantothenate, 0.2% sodium succinate, and 2 to 5 mCi of [32P]phosphate per liter. Cells were harvested in late exponential phase, washed once with cold 0.02 M potassium phosphate buffer, pH 7.0, resuspended in one-fiftieth the origi- nal culture volume of the same buffer and frozen. The suspen- sion was thawed and the cells were broken in a French pressure cell previously cooled in ice. The cell extract was made 2.5 mM in MgClz and a few crystals of deoxyribonuclease were added. The extract was incubated on ice for 30 min and then centrifuged at 6,000 x g for 10 min to remove unbroken cells. The membrane fraction was then collected by centrifugation for 45 min at 50,000 x g. After washing twice by resuspension and homogenization with a Teflon-glass homogenizer in cold 0.02 M potassium phosphate buffer, pH 7.0, the membrane fraction was resuspended in the same buffer at a protein concentration of 10 mg per ml and stored frozen. Protein was determined by a microbiuret procedure (22). Specific activities of 32P-labeled phospholipid in these preparations ranged from 0.2 to 1.0 Ci per mole. Sonicated membrane fraction was prepared by three 30-s periods of sonication on ice with a Branson Sonifier model s75.

Phospholipid Extraction and An&y&-Samples of 32P-labeled membrane fraction were diluted with water to 0.8 ml for extrac- tion of phospholipids by the method of Bligh and Dyer (23). The chloroform phase of the extraction, containing the phos- pholipids, was washed twice with water and then counted di- rectly to determine total phospholipid content or analyzed for relative amounts of individual phospholipids by thin layer chro- matography on acetone-activated 250 p Analtech Silica Gel G plates with chloroform-methanol-acetic acid (65: 25 :S) as de- veloping solvent (24).

The identity of the phospholipids was established by com- parison of Rp values with those of appropriate standards (ob- tained from Supelco, Inc., Bellefonte, Penn.). The radioactive lipids were located by autoradiography of the chromatograms with Kodak “noscreen” x-ray film. Appropriate areas of sil- ica gel were scraped into scintillation vials and counted. Phos-

pholipid compositions were determined as percentage of 32P counts per min in each lipid class. All counting of radioactivity was performed in a Packard Tri-Carb scintillation counter with Bray’s scintillation solution (25).

For dispersion of extracted [32P]phospholipid in buffer, a chloroform solution of lipids was evaporated to dryness in a Py- rex tube under a stream of nitrogen. Buffer was added and the lipid dispersed by several 30-s periods of sonic oscillation in an ice bath with a Branson Sonifier model S75. Undispersed lipid was removed by centrifugation. For phospholipase C assays, the dry phospholipids were dissolved directly in diethyl ether.

Enzyme Assays-Phospholipase C was assayed by digestion of 32P-labeled E. coli phospholipids in a two-phase system simi- lar to that used by Zwaal et al. (20). Two milliliters of di- ethyl ether containing 2.4 pmoles of 32P-lipid phosphorus were added to a screw cap culture tube containing 0.5 ml of 0.05 M

potassium phosphate buffer, pH 7.0. The assay was initiated by pipetting the phospholipase directly into the water phase, capping the tube tightly, and placing it on a Vortex mixer at room temperature. The assay was terminated by removing the tube from the mixer and allowing the phases to separate. The ether phase, 0.5 ml, was then counted in Bray’s solution. Since phospholipase C hydrolysis releases lipid phosphorus in water-soluble form, the loss of radioactivity from the ether phase is a measure of phospholipase C activity. One unit of activity is the amount of enzyme catalyzing hydrolysis of 1 pmole of lipid per min.

The spectrophotometric assays for both the acyl-CoA:glyc- erol-3-P acyltransferase and the acyl-CoA: I-acylglycerol-3-P acyltransferase have been described previously (26). Glyc- erol-3-P dehydrogenase was assayed according to the method of Lin et al. (27) by measuring the rate of phenazine methosulfate- mediated reduction of the tetrazolium dye 3-(4,5-dimethyl thiazolyl-2)-2,5-diphenyl tetrazolium bromide to its formazan which absorbs at 550 nm. A millimolar extinction coefficient of 8.75 was used for the formazan. NADH oxidase activity was measured spectrophotometrically by following a decrease in absorbance at 340 nm. Incubation mixtures contained 2.2 pmoles of histidine buffer, pH 6.5, 0.4 pmole of MgS04 and 0.078 pmole of NADH in a final volume of 0.2 ml. The reac- tion was initiated by addition of membrane fraction. Succinic dehydrogenase activity was measured as the phenazine metho- sulfate-mediated reduction by succinate of the tetrazolium dye used in the glycerol-3-P dehydrogenase assay. Incubation mixtures contained 1 pmole of disodium succinate, 2.25 pmoles of KCN, 6 pg of 3-(4,5-dimethyl thiazolyl-2) -2,5-diphenyl tetrazolium bromide, 20 pg of phenazine methosulfate, 0.4 pmole of MgS04, 2.2 pmoles of histidine buffer, pH 6.5, in a final volume of 0.2 ml. Reduction of the tetrazolium dye was followed as in the glycerol-3-P dehydrogenase assay. Glucose- 6-P dehydrogenase was assayed in a final volume of 0.6 ml con- taining 2 pmoles of glucose-6-P, 15 nmoles of NADP, 1.78 pmoles of MgC&, and 30 pmoles of Tris-HCl buffer, pH 7.8. Reduction of NADP was followed at 340 nm. All spectro- photometric assays were performed on a Gilford model 2400 recording spectrophot.ometer at 25”.

RESULTS

The assay described here, with radioactively labeled phos- pholipids extracted from E. coli grown in the presence of [““PI- phosphate, provides a convenient and reliable method of meas-

Issue of May 10, 1972 R. D. Davis, R. AI. Bell, and P. R. Vagelos

uring phospholipase C activity. The use of the two phases, ether and water, allowed ready separation of unhydrolyzed lipid from phosphoryl moieties released by phospholipase activ- ity. As shown in Fig. 1, the rate of phospholipid degradation was linear up to 25 min and proportional to the amount of en- zyme added. Linearity with time was observed during degra- dation of up to 80% of the lipid in the assay. No activation by Ca++ or inhibition by EDTB was observed in this assay.

Miith this assay, phospholipase C was purified from the cul- ture medium of B. cereus essentially according to the procedure of Zwaal et al. (20). The resulting preparation, which had a specific activity of 560 pmoles of lipid phosphorus released per min per unit of absorbance at 280 nm, migrated as a single major band on disc gel electrophoresis in sodium dodecyl sulfate. In order to test for possible proteolytic activity, an incubation of 1 mg of commercial, soluble glucose-6-P dehydrogenase with 1.2 enzyme units of the purified phospholipase C was performed for 1 hour at room temperature. No loss of dehydrogenase activity resulted.

The susceptibility to hydrolysis of E. coli membranous phos- pholipids by phospholipase C was compared to that of phospho- lipids extracted from the E. coli membrane by following their rate:; of hydrolysis under identical conditions. The effect of sonication of the membrane on susceptibility of phospholipids to hydrolysis was similarly studied. As shown in Fig. 2, the rate of hydrolysis of E. coli membrane phospholipids by purified phospholipase C was very similar in intact membrane fraction, sonicated membrane fraction, and a suspension of sonically dis- persed lipids isolated from the membrane fraction, under the conditions used. This similarity was confirmed in an experi- ment (not shown) with one-third the concentration of 32P-lipid utilized in the experiment of Fig. 2.

The specificity of the purified phospholipase C toward in- dividual phospholipids was investigated by analyzing the [““PI- phospholipids extracted at various times during the phospho- lipase digestions described in Fig. 2. As shown in Fig. 3A, the rates of hydrolysis varied for the three major classes of phos- pholipids in the E. coli membrane fraction. Phosphatidyleth- anolamine, the major phospholipid of E. co& comprising 75 to 8004 of the total membrane phospholipids, was hydrolyzed most rapidly. After 80 min of digestion, only 20% of the phosphati- dylethanolamine remained. Phosphatidylglycerol, comprising about 15O’ 0 of E. coli phospholipids, was hydrolyzed at a very similar rate. Cardiolipin, a 5 to 10% component of E. coli phospholipids, was hydrolyzed slower than phosphatidyleth- anolamine or phosphatidylglycerol. During the first 5 min of digestion, cardiolipin appeared to increase. After 10 min, cardiolipin began to decrease and hydrolysis of cardiolipin con- tinued until 60% of the initial cardiolipin remained at 60 min. Very little further decrease in cardiolipin occurred in the next hour of incubation during which phosphatidylethanol- amine and phosphatidylglycerol decreased at a constant rate.

Similar specificity with respect to rates of hydrolysis of in- dividual phospholipids was observed in sonicated membrane fraction (Fig. 3B) or in phospholipids extracted from mem- brane fraction and dispersed in buffer by sonic oscillation (Fig. 3C). In both preparations, phosphatidylethanolamine and phosphatidylglycerol were hydrolyzed at similar rates, phos- phatidylethanolamine decreasing to 11% in 80 min of digestion during which phosphatidylglycerol decreased to about 20%. Cardiolipin was hydrolyzed more slowly than phosphatidyl- glycerol or phosphatidylethanolamine. An initial hydrolysis

MINUTES pt ENZYME ADDED

FIG. 1. Assay of phospholipase C. Time and enzyme depend- ence. A, time course of assay. 0, no enzyme; A, 5 pl of enzyme ; 0, 15 ~1 of enzyme. B, enzyme dependence of assay. Time of assay was 10 min.

I I I 40 80 120

MINUTES

FIG. 2. Hydrolysis of membrane phospholipids by purified phospholipase C. 0, intact membrane fraction; l , sonicated membrane fraction; A, isolated lipid sonically dispersed in buffer. All incubations were performed at room temperature in 0.24 ml of 0.02 M potassium phosphate buffer, pII 7.0 containing 0.2 unit of purified phospholipase C. Each incubat.ion contained an amount of [azP]lipid corresponding to 2 mg of intact membrane protein. Samples of 20~1 were withdrawn for phospholipid analy- sis (described under “Experimental Procedure”) at the times indicated.

of about 20 to 30% of the cardiolipin in the first 10 to 20 min of digestion was followed by a virtual cessation of further hydroly- sis of cardiolipin. Incubation for an additional 100 min pro- duced very little loss of cardiolipin while a further decrease in both phosphatidylethanolamine and phosphatidylglycerol of 30 to 40% occurred. No initial increase in cardiolipin, as seen in intact membrane fraction, was observed in sonicated membrane fraction or in isolated, sonically dispersed phospholipid from the membrane fraction.

The effect of hydrolysis of membrane phospholipids on sev- eral membranous enzyme activities of I<. coli was investigated. As shown in Fig. 4, membranous palmityl-CoA:glyccrol-3-P acyltransferase activity was inactivated by treatment with purified phospholipase C at a gradually decreasing rate very similar to the rate of hydrolysis of membrane phospholipids.

2838 Effect of Phospholipase C on Membranous Enzymes Vol. 247, No, 9

I I 61 40 40

MINUTE:’ MINUTE:’ 120 120

FIG. 3. Hydrolysis of individual phospholipids by purified phospholipase C. Phospholipid analysis was performed on the samples from the incubations in Fig. 2. A, intact membrane fraction; B, sonicated membrane fraction; C, isolated lipid sonically dispersed in buffer. 0, phosphatidylethanolamine; 0, phosphahidylglycerol; A, cardiolipin.

80 120

FIG. 4.‘Effect of phospholipase C hydrolysis of membrane phospholipids on membranous palmityl-CoA:glycerol-3-P acyl- transferase activity. Membrane fraction (10 mg of protein) was incubated at room temperature in 0.9 ml of 0.02 M potassium phos- phate buffer wit,h (closed symbols) and without (open symbols) 0.6 unit of purified phospholipase C. Samples of 10 ~1 were with- drawn at times indicated for enzyme assay or phospholipid analy- sis. A, A, percentage of [a*P]lipid remaining; 0, 0, percentage of palmityl-CoA:glycerol-3-P acyltransferase activity remaining. Specific activity at zero time was 1.2 nmoles per min per mg of protein.

After 100 min of incubation, 75% of the phospholipids were hydrolyzed and 90% of the glycerol-3-P acyltransferase activ- ity was lost. Similarly, membranous NADH oxidase was in- activated by phospholipase C treatment at a rate comparable to that of phospholipid hydrolysis (Fig. 5). Inactivation of

\

\ \ \

I 40 80 120

MINUTES

FIG. 5. Effect of phospholipase C hydrolysis of membrane phos- pholipids on membranous NADH oxidase. An incubation was performed as in Fig. 4. NADH oxidase activity was assayed at times indicated. Specific activity at zero time was 27 nmoles per min per mg of protein.

membranous oleyl-CoA:l-acylglycerol-3-P acyltransferase (Fig. 6) proceeded at a rate identical with phospholipid hydrolysis until 450; of the activity remained, at which point inactivation ceased. The residual activity remained constant during an additional hour of treatment resulting in continued hydrolysis of phospholipids. Glycerol-3-P dehydrogenase and succinic dehydrogenase activities in membrane fractions were unaffected by hydrolysis of over 80% of membrane phospholipids as shown in Figs. 7 and 8.

Issue of Alsy 10, 1972 R. D. illavis, R. ill. Bell, and P, IL Vagelos 2839

1 I I 0 A0 so 120

MINUTES

Fro. 6. Effert of phospholipase C hydrolysis of membrane phos- pholipids on membranous oleyl-CoA:l-arylglycerol-3-P acyl- transferase activity. An incubation was performed as in Fig. 4. Oleyl-CoA:l-acylglycerol-3-P acyltransferase activity was as- sayed at times indicated. Specific activity at zero time was 2.3 nmoles per min per mg of protein.

0 A0 80 120 MINUTES

Fra. 7. Effect of phospholipase C hydrolysis of membrane phos- pholipids on membranous glycerol-3-P dehydrogenase. An in- cubation was performed as in Fig. 4. Glycerol-3-P dehydrogenase was assayed at Grnes indicated. Specific activity at zero time was 11.6 nmoles per min per mg of protein.

i\ttempts to reactivate palmityl-CoA:glycerol-3-P acyltrans- frrase or Nr\DII oxidaxe activities of phospholipase C-treated membranes by addition of suspensions of phosphatidylglycerol or phosphatidylethanolamine purified from lipid extracts of E. coli were unsuccessful. The glycerol-3-P acyltransferase and SADII oxidase activities of partially inactivated membrane fraction washed free of phospholipase were stable to incubation at room temperature for at least 15 min or freezing at - 15”.

I I I I 40 80 120

MINUTES

Fra. 8. Effect of phospholipase C hydrolysis of membrane phospholipids on membranous succinic dehydrogenase. An in- cubation was performed as in Fig. 4. Succinic dehydrogenase activity was assayed at times indicated. Specific activity at zero time was 29.4 nmoles per min per mg of protein.

Since the &ability of the glycerol-3-P dehydrogenase and suc- cinic dehydrogenase during digestion of 800/E, of total membrane phospholipid could be attributed to preferential association of these enzymes with the 20% of phospholipid which remained intact, an attempt was made to completely hydrolyze the mem- brane phospholipids. Incubation of membrane fraction with 10 times the concentration of purified phospholipase C used in Figs. 2 to 7 resulted in hydrolysis of 96% of the phospholipids in 155 min. Bnalysis of the unhydrolyzed phospholipid re- vealed that 2y0 of the original phosphatidylethanolamine, 4.54; of the original phosphatidylglycerol, and 329; of the original cardiolipin remained. Ko loss of glycerol-3-P dehydrogenase or succinic dehydrogenase activities over that observed in con- trol incubations without phospholipase C was observed in these membranes.

DISCUSSION

The purity of the phospholipase C used here is comparable to that obtained by Zwaal et al. (20). Since the assay u,sed here was performed at room temperature, the specific activity reported here (560 units per absorbance unit at 280 nm) com- pares favorably with that reported by %waal et al. (20), who assayed the enzyme at 37”. The stability of the soluble en- zyme, glucose-6-P dehydrogenase, in an incubation with 1.2 units of phospholipase per mg of dehydrogenase argued against the presence of any significant amounts of proteolytic activity in our preparation of phospholipase C. This preparation there- fore appeared suitable for a study of the specific effects of hy- drolysis of membrane phospholipids to diglyrerides.

The similarities of rates and specificity of phospholipase C hydrolysis of intact E. coli membrane, sonicated membrane, and isolated E. coli phospholipid suggests that the physical state of membranous phospholipids is similar to the physical state of sonically dispersed phospholipids, and that no protec- tion from hydrolysis is afforded the bulk of membrane phospho-

&,Tecl of Phospholipasc C on Membranous En,xyrri,es Vol. 247, No. 9

lipids by membrane protein. This suggestion is in agreement with recent evidence that the bulk of membranous phospho- lipids exist. in a physical state independent of membrane pro- teins (1, 28, 29).

The limited hydrolysis of cardiolipin showed here is in con- currence with the observation of Zwaal et al. (20) that this en- zyme is sensitive to the chemical nature of the lipid substrate. They reported that red blood cell sphingomyelin was not hy- drolyzed by phospholipase C, thereby explaining the previously observed maximum hydrolysis of 70% of red blood cell mem- brane phospholipids (1). The present work showed that up to 50% of E. coli cardiolipin was hydrolyzed by phospholipase C at a rate comparable to the rates of hydrolysis of phosphatidyl- ethanolamine and phosphatidylglycerol, while the remaining cardiolipin was stable lo further incubation with the phospho- lipase.

The resistance of a major fraction of cardiolipin to hydrolysis by phospholipase C may be explained in a number of ways. One possibility is that two classes of cardiolipin exist, differing in physical state or chemical composition. A heterogeneity of bacterial cardiolipin has been previously suggested (30, 31). The occurrence of partial hydrolysis of cardiolipin in isolated lipid as well as in the intact membrane fraction would seem to eliminate a role of the membrane protein in causing hetero- geneity of the physical state of E. coli cardiolipin, or at least require a separate explanation for partial hydrolysis of cardio- lipin in isolated lipid. Thus a chemical heterogeneity, most probably due to differences in the nature of the fatty acyl res- idues is indicated. A decided preference of phospholipase C for phospholipid substrate containing unsaturated acyl residues has been shown (32). Another explanation for partial hydroly- sis might be an association of the less polar cardiolipin with diglyceride produced by phospholipase C hydrolysis. Bangham and Dawson have attributed a cessation of phospholipase C hy- drolysis to the accumulation of diglyceride which was shown to affect the charge properties of lipid micelles (33). The small amount of cardiolipin present in E. coli relative to the other major phospholipids might also be the reason for its limited hy- drolysis. It should be noted, however, that treatment of intact membranes with 10 times the concentration of phospholipase C used in the experiment of Fig. 2 resulted in the hydrolysis of over 95% of both phosphatidylethanolamine and phosphatidyl- glycerol, while 32y0 of the original cardiolipin remained unhy- drolyzed; i.e. the fmal concentration of cardiolipin was twice that of unhydrolyzed phosphatidylglycerol or phosphatidyl- ethanolamine. Thus a lesser affinity of the enzyme for cardio- lipin must be postulated if concentration effects are to be invoked. It is of interest that cardiolipin was resistant to phos- pholipase C hydrolysis in membrane preparations derived from mitochondria (18, 19). In contrast, synthetic cardiolipin, or cardiolipin isolated from beef heart, was degraded by phospho- lipase C in a two-phase system of water and diethyl ether (34).

The apparent initial increase in cardiolipin observed during incubation of intact membranes with phospholipase C may be due to conversion of phosphatidylglycerol to cardiolipin by en- zymes present in the membrane. This conversion has been shown previously in vivo and in vitro (24, 35, 36). Sonication apparently destroyed this activity in the present work, since no increase in cardiolipin occurred when sonicated membrane frac- tions were studied.

The inactivation of acyl-CoA:glycerol-3-P acyltransferase at

a rate very similar to that of total phospholipid hydrolysis sug- gcsts that either this enzyme is dcpendenl on phospholipid whose accessibility to phospholipase C is similar to that of the bulk of membrane lipids or that intact membrane structure is necessary for the activity of this enzyme. Previous work from this laboratory (9) showed that the presence of trans-unsaturated fatty acids in membrane phospholipids perturbed the tempera- ture characteristics of this enzyme, implying an association with phospholipid. However, increasing the degree of unsaturation of membrane lipids had no effect on the Arrhenius plot of this activity, suggesting its independence of phase changes in mem- brane lipids.

The 50% inactivation of acyl-CoA: 1-acylglycerol-3-P acyl- transferase observed during phospholipid hydrolysis is similar to the partial inactivation of this enzyme previously observed upon treatment of the E. coli membrane fraction with maleic anhydride (26). Thus it appears that either two enzymes, differing in their phospholipid dependence and response to maleylation, catalyze this reaction, or, if a single enzyme is re- sponsible, it is converted to a half-active form by either maley- lation or hydrolysis of membrane phospholipids. Since inac- tivation of the labile portion of this activity is complete after the initial hydrolysis of 50% of the phospholipid, it would ap- pear that this enzyme is dependent upon phospholipid which is part of a more accessible fraction of membrane phospholipids. Alternatively, destruction of membrane integrity sufficient to maximally inactivate this enzyme could result from hydrolysis of 50% of the membrane lipid. This is in contrast to the inac- tivation of acyl-CoA:glycerol%P acyltransferase and NADH oxidase which appears to require total hydrolysis of membrane phospholipid.

NADH oxidase, like acyl-CoA : glycerol-3-P acyltransferase, was inactivated by phospholipase C at a rate very similar to that of phospholipid hydrolysis. Thus the complex process of electron transport appears to require phospholipid with an ac- cessibility to phospholipase digestion equal to that of the bulk of membrane lipid or at least require intact membrane. The role of phospholipid in mitochondrial electron transport has been indicated in a number of ways among which was inactiva- tion by phospholipase C (18). In contrast, membranous NADH oxidase of Bacillus megaterium is reportedly stable to phospho- lipase C treatment (37).

The two dehydrogenases studied here, which are not depend- ent on electron transport systems but simply catalyze the direct reduction of phenazine methosulfate by substrate, remain ac- tive after hydrolysis of over 95Y0 of membranous phospholipid. This observation together with our previous demonstration that Arrhenius plots of membranous glycerol-3-P dehydrogen- ase are independent of membrane fatty acid composition (9), strongly suggests the autonomy of this enzyme with respect to membrane phospholipids.

Mitochondrial succinic dehydrogenase is reportedly stable to phospholipase C hydrolysis (19) although an earlier report is in conflict with this observation (38). Succinic dehydrogenase has been solubilized and purified from mitochondria, and while lipids appeared to stabilize the enzyme under certain conditions (39), an absolute requirement for phospholipid was not evident. Arrhenius plots of mitochondrial succinic dehydrogenase from rat liver and sweet potato have been shown to be discontinuous at temperatures coinciding with phase changes in membrane lipids detected by spin label probes (40, 41). Disruption of t,h+

Issue of May 10, 1972 R. D. Mavis, R. M. Bell, and P. R. Vagelos 2841

mitochondria with a nonionic detergent eliminated the break in the Arrhenius plots but apparently did not inactivate the enzyme. Thus an enzyme-lipid interaction resulting in a re- flection of the phase properties of the lipid in the activity of the enzyme does not necessarily imply an absolute dependence upon phospholipid for activity. Conversely, the enzyme acyl-CoA: glycerol-3-P acyltransferase, has an apparent dependence on

intact phospholipid but does not reflect the phase properties of membrane phospholipids in its temperature characteristics.

All of the enzymes studied here have been shown to be a part of the cytoplasmic membrane (42-44). Thus the heterogeneity of these enzymes with respect to phospholipid dependence is characteristic of a single membrane structure. Currently pop- ular models of membrane structure show various degrees of intercalation of proteins by lipid within a single membrane (45). The bulk of the lipid, however, is thought to exist in a bilayer structure independent of protein. Thus only a small fraction of the lipid is in actual contact with protein. The results re- ported here are consistent with such a model. Membranous enzymic activities showed varying degrees of dependence on in- tact phospholipids, while the bulk of phospholipid appears in a physical state similar to isolated phospholipid.

Acknowledgments-We thank Dr. Michael Glaser for helpful discussions. We are indebted to Dr. Ray Fall for performing the gel electrophoresis. The expert technical assistance of Mr. Walter Nulty, Mrs. Karol Mavis, and Miss Linda Loney is

gratefully acknowledged.

REFERENCES

1. GLBSER, M., SIMPKINS, H., SINGER, S. J., SHEETZ, M., AND CHAN, S. I. (1970) Proc. Nat. Acad. Sci. U.S.A. 65, 721

2. ROTHFI~:LD, L., AND FINKELSTEIN, A. (1968) Annu. Rev. Bio- them. 37,‘463’

I

3. ROTHFIELD, L., AND ROMEO, D. (1971) Bacterial. Rev. 35, 14 4. KUNDIG. W.. AND ROSEM~IN. S. (1971) J. Biol. Chem. 246,

1407-1418 5. CUNNINGHBM, C. C., AND HAGER, L. P. (1971) J. Biol. Chem.

246, 1575-1582 6. CUNNINGHAM, C. C., AND HAGER, L. P. (1971) J. Biol. Chem.

246, 1583-1559 7. SANDERMANN, H., JR., AND STROMINGER, J. L. (1971) Proc. Nat.

Acad. Sci. U. S. A. 68, 2441 8. HIG~SHI, Y., AND STROMINGER, J. L. (1970) J. Biol. Chem. 245,

3691-3696 9. M.IVIS, R. D., AND VAGELOS, P. R. (1972) J. Biol. Chem. 247,

652 10. PCHAIRER, H. U., AND OVERATH, P. (1969) J. Mol. Biol. 44, 209 11. WILSON, G., ROSE, S. P., AND Fox, C. F. (1970) Biochem.

Biophys. Res. Commun. 38, 617 12. OVERATH, P., SCHAIRER, H. U., AND STOFFEL, W. (1970) Proc.

Nat. Acad. Sci. U. S. A. 67, 606 13. WILSON, G., AND Fox, C. F. (1971) J. Mol. Biol. 55, 49

14. 15.

16.

17.

18.

19.

20.

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.

ZAKIM, D. (1970) J. Biol. Chem. 245, 4953-4961 MART~NOSI, A., DONLEY, J. R., PUCNLL, A. G., AND HALPIN,

R. A. (1971) Arch. Biochem. Biowh?rs. 144. 529 ATTWOO;, D., GRAHAM, A. B., AND”WOOD; GI C. (1971) Bio-

them. J. 123, 875 BURSTEIN, C., LOYTER, A., AND RACKER, E. (1971) J. Biol.

Chem. 246, 4075-4082 BURSTEIN, C., KANDRACH, A., AND RACKER, E. (1971) J. Biol.

Chem . 246, 4083-4089 AWASTHI, Y. C., RUZICKA, F. J., AND CRANE, F. L. (1970)

Biochim. Biophys. Acta 203, 233 ZWAAL, R. F. A., ROELOFSON, B., COMFURIUS, P., AND VAN

DEENEN, L. L. M. (1971) Biochim. Biophys. Acta 233, 474 LONG, C., ODAVIC, R., AND SARGENT, E. J. (1967) Biochem. J.

102, 216 MUNKRES, K. D., AND RICHARDS, F. M. (1965) Arch. Biochem.

Biophys. 109, 466 BLIGH, E. G., AND DYER, W. J. (1959) Can. J. Biochem. Physiol.

37, 911 AMES, G. F. (1968) J. Bacterial. 95, 833-843 BRAY, G. A. (1960) Anal. Biochem. 1, 279 RAY, T. K., CRONAN, J. E., JR., MAVIS, R. D., AND VAGELOS,

P. R. (1970) J. Biol. Chem. 245, 6442-6448 LIN, E. C. C., KOCH, J. P., CHUSED, T. M., AND JORGENSEN,

S. E. (1962) Proc. Nat. Acad. Sci. U. S. A. 48. 2145 STEIM, j. M.; TOURTELLOTTE, M. E., REINERT,‘J. C., McEL-

HANEY, R. N., AND RADER, R. L. (1969) Proc. Nat. Acad. Sci. U. S. A. 63, 104

ENGLEMAN, D. M. (1971) J. Mol. Biol. 58, 153 SHORT, S. A., AND WHITE, D. C. (1970) J. Bacterial. 104, 126-132 TUCKER, A. N., AND WHITE, D. C. (1970) J. Bacterial. 102,

506-513 VAN DEENFN, L. L. M., DE HAAS, G. H., HEEMSIIERK, C. H.

T., AND MEDUSKI, J. (1961) Biochem. Biophys. Res. Commun. 4, 183

BANGHAM, A. D., AND DAWSON, R. M. C. (1962) Biochim. Bio- phys. Acta 59, 103

DE HAAS, G. H., BONSEN, P. P. M., AND VAN DEENEN, L. L. M., (1966) Biochim. Biophys. Acta 116, 114

CRONAN, J. E., JR. (1968) J. Bacterial. 95, 2054-2061 STANACEV, N. Z., CHANG, Y. Y., AND KENNEDY, E. P. (1967) J.

Biol. Chem. 242, 3018-3019 EISENBERG, R. C., Yu, L., AND WOLIN, M. J. (1970) J. Bac-

teriol. 102, 161-171 CERLETTI, P., CAIAFA, P., GIORDANO, M. G., AND GIOVENCO,

111. (1969) Biochim. Biophys. Acta 191, 502 CERLETTI, P., CAIAFA, P., GIORDANO, M. G., AND TESTOLINI,

G. (1970) Lipids 5, 953 RAISON, J. K, LYONS, J. M., AND THOMSON, W. W. (1971)

Arch. Biochem. BioDhus. 142, 83 RAISON, J. K., LYON;, j. M., MEHLHORN, R. J., AND KEITH,

A. D. (1971) J. BioZ. Chem. 246, 4036-4040 OSBORN, M. J., GANDER, J. E., PARISI, E., AND CARSON, J.

(1972) J. BioZ. Chem. 247, in press BELL, R. M., MAVIS, 12. D., OSBORN, M. J., AND VAGELOS, P.

R. (1971) Biochim. BioDhus. Acta 249, 628 WHITE, D.‘A., ALBRIGHT~F~R., LENNA~Z, W. J., AND SCHNAIT-

MAN, C. A. (1971) Biochim. Biophys. Acta 249, 636 SINGER, S. J. (1971) in L. ROTHFIELD (Editor), Membrane

structure and function, Academic Press, New York