Proc. Nat. Acad. Sci. USA Vol. 69, No. 10, pp. 2793-2797, October 1972

An Osmotically Fragile Mutant of Bacillus subtilis with an Active Membrane-Associated A1 (protoplasts//phospholipase Al inhibitor)

CLAUDIA KENT AND W. J. LENNARZ Department of Physiological Chemistry, The Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, Maryland 21205 Communicated by Albert L. Lehninger, July 24, 1972

ABSTRACT By use of a newly developed procedure for [0.6 M sucrose-0.05 M Na2HP04-0.01 M MgCl2 (pH 7.0)] the isolation of mutants with osmotically fragile proto- followed The cell Was in plasts, a mutant of Bacillus subtilis was isolated that has by centrifugation. pellet suspended a very active system for the catabolism of phos- 15 ml of the same buffer containing 60 mg of lysozyme, then pholipids via the sequential action of a phospholipame Al incubated for 30 min at 300 to convert the cells to protoplasts, (EC 3.1.1.4) and a lysophospholipase (EC 3.1.1.5). The wild- as monitored by phase-contrast microscopy. The protoplasts type bacteria contain no detectable phospholipase Al were added, with stirring, to 450 ml of MP buffer [0.05 M activity, but do contain a protein that specifically inhibits the phospholipase A, in the mutant. This protein may play Na2HPO4-0.01 M MgCl2 (pH 7.0)] at 00 containing 4.5 mg an important role in the control of phospholipid catab- of DNase and 5.0 mg of RNase, then incubated at O0 for olism. 30 min. The suspension was centrifuged for 1 hr at 20,000 X g. The membrane pellet was washed twice by suspension in A screening procedure has been developed in this laboratory 30 ml of MP buffer and centrifugation at 100,000 X g for to isolate mutants of Bacillus subtilis with defective mem- 30 min. The washed membrane pellet was suspended in branes. This procedure involves a comparison of the stability 11.4 ml of MP buffer and stored in 0.5-ml aliquots at -20°. of protoplasts of the mutant with protoplasts of the wild type. For preparation of the supernate, cells were harvested and The protoplast membranes of one mutant isolated by this washed as described for the preparation of membranes. The procedure contained only 30-50% as much phospholipid as washed cell pellet was suspended in 37.5 ml of MP buffer protoplast membranes of the wild type, as a result of active containing 60 mg of lysozyme and incubated at 300 for 30 lipid catabolism in vitro. This paper describes the resolution min. The resulting membrane suspension was centrifuged for and partial characterization of the involved in this 3 hr at 100,000 X g. The supernatant fraction was removed catabolic process: a membrane-bound phospholipase Al (EC and stored in 2-ml aliquots at -200. 3.1.1.4) and a cytoplasmic lysophospholipase (EC 3.1.1.5). The presence in the wild-type cells of a heat-stable protein Preparation of Radioactive Lipids. Wild-type B. subtilis was that is a potent inhibitor of the phospholipase Al in the grown in medium supplemented with 20 MCi of carrier-free mutant is also reported. This inhibitor may be involved in the 32pi or 0.6 1ACi (21 pmol) of [8H]leucine and 0.6 ,uCi (15 pmol) regulation of phospholipid catabolism in vivo. of [3H]isoleucine per ml of culture (these amino acids are incorporated into the branched-chain fatty acyl groups of the MATERIALS AND METHODS lipids). The cells were harvested at late-logarithmic stage by Bacterial Strains. The parent strain used in this study was centrifugation at 8000 X g for 10 min, then washed by suspen- Bacillus subtilis Mu8u5u5, a derivative of Bacillus subtilis sion in 0.05 M Na2HPO4-0.01 M MgCl2 (pH 7.0) and centri- 168 (1). Mutant CMK 33 was obtained by mutagenesis of fuged again. Lipids were extracted (2) from the cell pellet with B. subtilis Mu8u5u5, and was isolated on the basis of the 20 volumes of acidic CHCl3-CH3OH. Techniques for the osmotic fragility of its protoplasts to 0.5 M lactose, which purification of the lipids (3) and the measurement of radio- stabilizes protoplasts of the wild type. (Kent, Krag, and activity in lipid samples (2) have been described. Lennarz, manuscript in preparation). Chromatography of Lipids. Plates for thin-layer chromatog- Preparation of Membrane and Supernatant Fractions for raphy of lipids were prepared as described (4). Lipids were also Enzyme Studies. Cells were grown at 300 in a medium con- chromatographed on paper impregnated with silicic acid. The taining (in 1.0 liter): 10.0 g of Bacto peptone, 10.0 g of Yeast paper chromatograms were developed in diisobutylketone- Extract (Difco), 5.0 g of NaCl, and 0.4 g of Na2HPO4, adjusted acetic acid-H20 40:25:5 for 15--2ahr. to pH 7.0 with 2 N NaOH. Wild-type and mutant 33 cells Materials. Twice crystallized lysozyme was purchased from (1.5-liter culture) in late-logarithmic phase were collected by Worthington Biochemical Corp., Viperus russelli phos- centrifugation and washed by suspension in 90 ml of buffer pholipase A2 from Sigma Chemical Co., Rhizopus delamar from Miles Lab., and Cutscum from Fisher Scientific Abbreviations: PG, phosphatidylglycerol; lyso PG, lyso- Co. 82pi (carrier-free), - [U-"4C]leucine, i- [4,5-3H ]leucine, and phosphatidylglycerol; PE, phosphatidylethanolamine; lyso PE, Na[1-'4C]acetate were obtained from New England Nuclear lysophosphatidylethanolamine. Corp., and i-[5-3H]isoleucine from Schwarz-Mann. Spe- 2793 Downloaded by guest on September 26, 2021 2794 Biochemistry: Kent and Lennarz Proc. Nat. Acad. Sci. USA 69 (1972)

100 of ['Hileucine and [sH]isoleucine. As shown in Table 1, the distribution of label was similar in lipids from the wild-type cells and protoplasts, and mutant cells, but in the lipids from the mutant protoplasts there was a striking increase in free 0- r_ 60 fatty acids, with a concomitant decrease in both polar lipids 0. -_ and diglyceride. -J\ When lysed protoplasts of the mutant were separated into 5. 40 - membrane and supernatant fractions, extensive breakdown of FZ exogenous 82P-labeled lipid was observed only upon recom- 2 20 bination of the membrane and supernatant fractions (Table 2, Exp. 1). Neither fraction alone catalyzed extensive hydrolysis of 82P-labeled lipid to water-soluble 30' 60' 120' products containing 82P. INCUBATION TIME When wild-type crude extracts were similarly prepared and fractionated, no lipolytic activity was observed, even with the FIG. 1. Lipid breakdown during incubation with lysozyme. recombined fractions. In addition, when membranes from the Wild-type and mutant cells were grown in 90 ml of medium wild type were combined with the supernate from the mutant, containing 4.4 uCi of 82Pi. Cultures were harvested at 2.0 OD or vice versa, there was still very little hydrolysis of "2p- units (mid-log) or 3.2 OD units (late-log). The cells were washed labeled lipid to water-soluble products containing "2P. As with 0.6 M sucrose-0.05 M Na2HPO4-0.01 M MgCl2 (pH 7.0), then suspended in 6.0 ml of the same buffer containing 3.6 mg of shown in Table 2, Exp. 2, when either the supernate or the lysozyme (mid-log) or 9.4 mg of lysozyme (late-log). The cell membranes of the mutant were boiled before recombination, suspensions were incubated at 300 and 1.0-ml aliquots were re- no breakdown was observed. moved at various times to determine the amount of 32P-lipid re- To determine if two enzymes were necessary for the maining. The aliquots were centrifuged at 100,000 X g for 1 hr; hydrolysis of "2P-labeled lipid to water-soluble products 4.0 ml of acidic CHC13-CU30H was added to the resulting pellet, containing "2p, purified [(2P]phosphatidylglycerol (PG) was and the lipids were extracted and counted (2). 0 0, wild- incubated sequentially with the membranes and supernate type, late-log phase cells; [- - -5, wild-type, mid-log phase from the mutant (Table 3, Exp. 1). Upon incubation with the cells; 0 * mutant, late-log phase cells; M- - -U, mutant, membranes, a considerable proportion of the PG was con- mid-log phase cells. verted to lyso phosphatidylglycerol (lyso PG), but there was little degradation to water soluble products containing 32p. cifically labeled [1-14C]fatty acyl- and [2-14C]fatty acyl phos- When the lipids from the first incubation were extracted and phatidylethanolamine were prepared by Dr. Fred Albright in incubated with the supernate, the. lyso PG was degraded, this laboratory. with a net loss of ["2P]phospholipid. In Exp. 2 the incubations were performed in reverse order and lyso PG was not RESULTS formed until the second incubation. Therefore, two enzymes The screening procedure used to isolate mutant CMK 33 are necessary for the hydrolysis of "2P-containing products, a involves conversion of mutagenized cells to protoplasts by treatment with lysozyme, followed by comparison of the stability of mutant and parent protoplasts. This screening TABLE 1. Distribution of ['H]fatty acids in wild-type procedure and the properties of some of the other mutants will and mutant B. subtilis be published elsewhere (Kent, Krag, and Lennarz, manuscript in preparation). Protoplasts of mutant CMK 33 contained (% of total) only 30-50% as much phospholipid and diglyceride as the Polar Di- Free fatty protoplasts of the wild type, although intact cells of the lipids glyceride acids mutant and wild type contained identical amounts of these lipids. The decline in lipid content during the conversion of WT cells 64 28 1 of CMK 33 cells 67 27 2 cells CMK 33 to protoplasts was followed by measurement WT protoplasts 73 22 1 of the decrease in endogenous "2P-labeled lipids. As shown in CMK 33 protoplasts 45 12 33 Fig. 1, 70% of the total phospholipid was degraded after 120 min. In contrast, in the wild type only 20% of the phospho- Cells were grown in 30 ml of medium containing 0.12 mCi lipid was hydrolyzed; in most experiments hydrolysis was less (each) of ['H]leucine and ['H]isoleucine. The cells were har- than 5%. As shown, lipid degradation in the mutant was vested in late-logarithmic phase (3.5 OD units at 660 nm), washed independent of the phase of growth. with 0.7 M sucrose buffer [0.7 M sucrose-0.05 M Na2HP04-0.01 To determine if one specific phospholipid class was hy- M MgC12 (pH 7.0)], and divided into two equal portions. The drolyzed in CMK 33, 32P-labeled lipids of the mutant and lipids from one portion (cells) were extracted immediately. The wild-type intact cells and protoplasts were analyzed by other portion was incubated with 0.84 mg of lysozyme in 1 ml of chromatography on silicic acid-impregnated paper. Although 0.7 M sucrose buffer for 1.5 hr before extraction. The incubation the total amount of phospholipid was much lower in proto- mixtures were extracted with 24 ml of acidic CHClr-CH3OH, plasts of the mutant than in those of the wild type, the per- as described in Fig. 1, and the isolated lipids were chromato- graphed on a thin-layer plate with petroleum ether-diethyl centage of each phospholipid did not change, indicating that ether-acetic acid 80:75:1.5. Silica gel zones containing radio- the hydrolysis was not specific for one class of phospholipids. active lipid were scraped from the plate, transferred to a vial, Lipids of the wild type and mutant were also labeled in the and counted after addition of 0.5 ml of CHOH-H20 1:1 and 15 branched fatty acyl chains by growing the cells in the presence ml of Triton scintillation fluid. Downloaded by guest on September 26, 2021 Proc. Nat. Acad. Sci. USA 69 (1972) in a Mutant of Bacillus subtilis 2795

phospholipase A in the membranes that converts PG to SUPCMK 33 lyso PG, and a lysophospholipase in the supernate that hydrolyzes lyso PG. By use of purified phosphatidylethanolamine (PE) labeled in the branched fatty acyl groups as , the phospho- a lipase A in the mutant membranes was further characterized.

It required Ca++ (1.0 mM) and was inhibited by EDTA. 40 Optimal activity was observed in the presence of the detergent

Cutscum. The apparent Km for PE was 1 mM. The products of a the membrane phospholipase A reaction with PE as substrate U.e on were lysophosphatidylethanolamine (lyso PE) and free fatty *acid, as determined by thin-layer chromatography in four MembCMK 33 + SUP WT solvent systems. The lysophospholipase(s) present in the MembcMK 33 + MembWT supernate of the mutant has not been characterized in detail, but it is clear that it is present in the wild type as well as in the mutant. Minutes at 300 To determine the positional specificity of the phospholipase A, mutant membranes were incubated with PE containing FIG. 2. Rate of phospholipase Al activity. Mutant or wild- 14C-labeled fatty acids at the 1 or 2 position of the glycerol type membranes (21 ,ug of protein), mutant supernate (126 lug moiety. The positional distribution of 14C-labeled fatty acids of protein), or wild-type supernate (160 /ug of protein) were in- in the substrate was determined by incubation of the [14C]PE cubated with 40 nmol of [3H]PE (225 cpm/nmol); 0.05 M Tris- with Viperus russelli . When PE containing maleate (pH 8.0); 1.0 mM CaCI2; and 0.6% Cutscum, in a volume of 0.05 ml, at 300. The reactions were terminated by the addition 94% of the 14C label in the 2-fatty acyl group was used as of 3.0 ml of isopropyl alcohol-heptane-1 N H2S04 40: 10:1 and substrate for the phospholipase A in the mutant membranes, 0.35 ml of H20, and the 3H-labeled fatty acids were extracted the lyso PE and free fatty acid products contained 92 and 8% (85% yield) (5). of the 14C label, respectively. When the PE contained 78% of the 14C label in the 1-fatty acyl group, the lyso PE and free pholipase A1 activity was observed with crude extracts or fatty acid products contained 11 and 89% of the label, respec- membranes of the wild type. Most interesting, however, was tively. The enzyme in the mutant membranes is, therefore, a phospholipase Al. the observation that when either membranes or supernate from wild were added to the The rate of hydrolysis of PE by membranes of the mutant the type membrane-associated are shown in Fig. 2. The stimulation observed when mutant supernate was added to mutant membranes results from the TABLE 3. Sequential incubations of [82P]PG action of the lysophospholipase in the mutant supernate, with mutant membranes and supernate which releases additional fatty acid from the lyso PE produced by the membrane phospholipase. As noted above, no phos- First incubation Second incubation TABLE 2. Incubation of mutant and wild-type membranes Exp. 1 and supernate with 82P-labeled lipids Membranes plus [32P]PG Supernate plus lipid of first incubation % Breakdown 32P-Lipid Lyso PG 32P-Lipid Lyso PG (% Loss) (% of total) (% Loss) (% of total) Exp. 1 MembranescMK 33 10.5 1.7 29.5 20.8 0.2 SUPCMK33 7.3 Exp. 2 MembranescMK 33 + SUPCMK 33 43.8 Supernate plus [32P]PG Membrane plus lipid product MembraneswT 0 of first incubation SUPWT 1.6 32P-Lipid Lyso PG 32P-Lipid Lyso PG MembranesWT + SUPWT 3.0 (% Loss) (% of total) (% Loss) (% of total) MembraneSWT + SUPCMK 33 5.9 MembranescMK 3l + SUPWT 4.9 1.1 0.1 8.9 24.2 Exp. 2 MembranescMK 33 + SUPCMK 33 30.5 The first incubation mixture in each experiment contained: Boiled MembranescMK 33 + SUPCMK 33 2.5 membranes (460 Mg of protein) or supernate (230 Mg of protein); MembranescmK 33 + Boiled SUPCMK 33 0.3 0.05 M Na2HP04 (pH 7.0); 0.01 M MgCl2; 0.25% Cutscum; 20 nmol of [32PJPG (5400 cpm), in a total volume of 0.3 ml. Incubation mixtures in Exp. 1 contained: membranes (300 After incubation of the mixture at 300 for 90 min, the lipids were Mug of protein) and/or supernate (560 Mg of protein); 0.25% extracted as described in Fig. 1, but were washed with acidic Cutscum; 0.05 M Na2HP04 (pH 7.0); 0.01 M MgCh2; 12 nmol of saline only once. The extracted lipid was dispersed by sonication 32P-labeled total lipid from B. subtilis (3000 cpm); in a total in 0.2 ml of 0.05 M Na2HP04- 0.01 M MgC12 (pH 7.0), and was volume of 0.3 ml. In Exp. 2, the conditions were identical except then incubated with supernate or membrane as described above, that the amount of protein and substrate were decreased by one- except that no additional Cutscum was added. All incubations half, and the final volume was 0.25 ml. After incubation of the were performed in duplicate and the lipids from each incubation mixture at 300 for 2 hr, the remaining lipids were extracted with mixture were analyzed by thin-layer chromatography in CHCl- 6 ml of acidic CHC13-CH30H as described in Fig. 1. CH30H-H20 90:40:5. Downloaded by guest on September 26, 2021 2796 Biochemistry: Kent and Lennarz Proc. Nat. Acad. Sci. USA 69 (1972)

WT 6C .4W Phospholipase A, OH Lysophospho- r OH -OPOXipa - HO - OPOX OP lIipose [Opox Blocked by Presence of 4C In hibitor CMK 33 0-

H OH Phospholipase Al r: OH Lysophospho- 2C lip ose oI Inhibitor Absent or Defective FIG. 4. Possible mechanism for control of phospholipase a. A, 50 100 150 200 250 300 activity in B. subtilis. M'g SUPWT FIG. 3. Inhibition of phospholipase Al activity in CMK 33 by wild-type supernate. Conditions were as described in Fig. 2; phospholipids to water-soluble products containing 32p, with CMK 33 membranes (21 /Ag of protein) were incubated for 20 min concomitant release of free fatty acids. When mutant extracts at 300. were separated into membrane and supernatant fractions, neither fraction alone catalyzed the hydrolysis of 82P-labeled lipid to water-soluble products containing 32p. Sequential phospholipase Al of the mutant there was marked inhibition of incubation of 32P-labeled PG with the membrane and its activity (Fig. 2). The effect of increasing concentrations of supernatant fraction revealed that a phospholipase A in the the wild-type supernate on the phospholipase Al is shown in membranes hydrolyzed the PG to lyso PG, and a lyso- Fig. 3. phospholipase in the supernate hydrolyzed the lyso PG to a The phospholipase Al inhibitor in the supernate of the wild water-soluble phosphate compound. Incubation of the mem- type is a protein; it is sensitive to trypsin, nondialyzable, and brane phospholipase with labeled PE preparations with known is not inactivated after incubation for 10 min at 1000. Its positional distribution of the labeled fatty acyl groups showed molecular weight is about 30,000, as estimated by sucrose that the membrane-associated enzyme is a phospholipase A1. gradient centrifugation. It has been purified 4-fold by precipi- Recently, the presence of a phospholipase Al unique to the tation with 65-100% saturated ammonium sulfate. The spores and sporangia of Bacillus megaterium has been reported inhibitor in the membranes of the wild type is also sensitive to (8). trypsin and is heat stable, but is has not been established that Studies of lipid catabolism in the osmotically fragile mutant the supernatant and membrane-associated inhibitors are resulted in the novel observation that the wild type contains a identical. potent inhibitor of the phospholipase A in the mutant. The The detailed nature of the mechanism of inhibition of the inhibitor, which was found both in the cytoplasm and asso- phospholipase is not known. Initial experiments indicated that ciated with the membrane, is a heat-stable, trypsin-sensitive the inhibitor was competitive with the substrate PE. However, protein. The molecular weight of the inhibitor in the cyto- the extent of inhibition was increased after incubation of the plasm is about 30,000, as determined by sucrose gradient enzyme with inhibitor, suggesting that a time-dependent centrifugation. The inhibitor does not act in a reversible inactivation of the enzyme by inhibitor occurs. Attempts to manner. It is not simply a nonspecific protease, since the reverse the inhibition by washing the inactivated mutant partially purified inhibitor has no effect on the activity of membranes or by treating them with excess substrate were several enzymes, including three from other unsuccessful. To determine if the inhibitory protein was sources. Moreover, the fact that the heat-stable protein in the simply a nonspecific protease, it was assayed for inhibitory wild type has no effect on the activity of these phospholipases activity with several other enzymes. The inhibitor had no indicates that it is not an enzyme that catalyzes the reacyla- effect on the lysophospholipase in the supernate of mutant tion of lysophospholipids formed by the action of the phos- CMK 33 or on Viperus russelli phospholipase A2, Rhizopus pholipases. Further studies on the mechanism of action of the delamar lipase (which acts as a phospholipase Al) (6), Escher- inhibitor await its purification and determination of the ichia coli phospholipase Al (7), Bacillus subtilis phosphatidic relationship between the supernatant and membrane- acid and UDP-glucose: diglyceride glucosyl associated inhibitors. , and glucose-6-phosphate dehydrogenase. From these studies it is clear that phospholipase A1 activity is not detectable in the wild type, but that the wild type con- DISCUSSION tains an inhibitor of the active phospholipase A1 in CMK 33. Presumably, the wild type also contains a phospholipase A1, A mutant of Bacillus subtilis has been isolated by a screening but its activity is masked by the inhibitor. Assuming that the procedure that detects mutants with fragile membranes. This defect in CMK 33 results from a single mutation, the simplest mutant contains a highly active system for phospholipid hypothesis to explain these observations is that the defect in catabolism in vitro. During the conversion of mutant cells to the mutant is in the inhibitor, which may normally function as protoplasts there is a general hydrolysis of the 32P-labeled a regulator of phospholipase Al activity (Fig. 4). An alterna- Downloaded by guest on September 26, 2021 Proc. Nat. Acad. Sci. USA 69 (1972) Phospholipase Al in a Mutant of Bacillus subtilis 2797

tive hypothesis is that the wild type, rather than containing a 1. Yoshikawa, H. & Sueoka, N. (1963) Proc. Nat. Acad. Sci. a fatty acyl USA 49, 806-813. masked phospholipase Al, contains phospholipid 2. Gould, R. M. & Lennarz, W. J. (1970) J. Bacteriol. 104, group transacylase. In the mutant this enzyme may be 1135-1144. defective, so that instead of functioning as a transacylase it 3. Hopfer, U., Lehninger, A. L. & Lennarz, W. J. (1970) J. acts as a . In any event, little is known about Membrane Biol. 2, 41-58. regulatory mechanisms involved in lipid degradation in bac- 4. Skipski, U. P., Peterson, R. F. & Barclay, M. (1964) Biochem. teria; mutant CMK 33 and the wild-type phospholipase in- J. 90, 374-378. S. Y. in Methods in Enzymology, may prove to valuable tools to study such regula- 5. Gatt, & Barenholz, (1969) hibitor be ed. (Lowenstein, J. (Academic Press, New York), Vol. 14, tion of lipid catabolism. pp. 167-170. It is a pleasure to acknowledge the expert technical assistance 6. Slotboom, A. J., DeHaas, G. H., Bonsen, P. P. M., Burback- of Janet Kennedy and Blakelyn Albright. This work is supported Wisterhuis, G. J. & VanDeenen, L. L. M. (1970) Chem. Phys. by Public Health Service Grant A106888 07. Claudia Kent was a Lipids 4, 15-29. trainee supported by Public Health Service Grant GB 31121; 7. Scandella, C. J. & Kornberg, H. (1971) Biochemistry 10, a portion of the work reported is from a thesis submitted in 4447-4456. partial fulfillment of the requirements for the degree of Doctor of 8. Raybin, D. M., Bertach, L. L. & Kornberg, A. (1972) Philosophy at The Johns Hopkins University. Biochemistry 11, 1754-1760. Downloaded by guest on September 26, 2021