Diphosphate-Diacylglycerol-Dependent Phosphatidylserine

Vol. 147, No. 2

JOURNAL OF BACTERIOLOGY, Aug. 1981, p. 535-542 0021-9193/81 -0805;35-08$02.00/()0

Characterization of a Membrane-Associated Cytidine Diphosphate-Diacylglycerol-Dependent Phosphatidylserine Synthase in Bacilli ANURADHA DUTT AND WILLIAM DOWHAN* Department of Biochemistry and Molecular Biology, University of Texas Medical School at Houston and the University of Texas Graduate School of Biomedical Sciences, Houston, Texas 77025 Received 23 March 1981/Accepted 15 May 1981

The synthesis of phosphatidylserine in two gram-positive aerobic bacteria has been partially characterized. We have located a cytidine 5'-diphospho-diacylglycerol:L-serine O-phosphatidyltransferase (phosphatidylserine synthase) activity in the membrane fraction of Bacillus licheniformis and Bacillus subtilis. The activity was demonstrated to be membrane associated by differential centrifugation, sucrose gradient centrifugation, and detergent solubilization. The direct involvement of cytidine 5'-diphospho-diacylglycerol in the reaction was demonstrated by the conversion of the liponucleotide phosphatidyl moiety to phosphatidylserine. This activity is dependent on divalent metal ion (manganese being arMd is stimulated by nonionic detergent and its product phosphatidyloptir*l) serine. Based on studies with various combinations of products and substrates, the reaction appears to follow a sequential BiBi kinetic mechanism.

The pathways of phospholipid biosynthesis location. Upon solubilization of the membrane have been well established in several gram-neg- preparation with Triton X-100 and subsequent ative species of bacteria (7), particularly for centrifugation, however, only 18% of the total Escherichia coli (22). In E. coli, many of these activity was detected in the supernatant, with biosynthetic enzymes have been purified either the remainder being unaccounted for. Neither of partially or to homogeneity (22). The CDP-di- these reports ruled out the possibility that CDPacylglycerol-dependent phosphatidylserine syn- diacylglycerol simply stimulated the incorporathase is unique among the phospholipid biosyn- tion of radiolabeled serine into endogenous lipid thetic enzymes in gram-negative organisms in as opposed to participating as a substrate in the that it is associated with ribosomes rather than reaction; precedence for such a stimulatory role the membrane fraction in cell-free extracts (7, has been shown in cardiolipin synthesis in E. 24). Although studies of both the purified en- coli (9). Silber et al. (28), working with another zyme (2) and the enzyme in crude extracts (18) gram-positive anaerobe, Clostridium butyricum, from E. coli suggest a membrane localization for noted the formation of radiolabeled phosphatithis enzyme in vivo, a physiologically important dylserine from phosphatidic acid labeled in the role for the affinity of the enzyme for ribosomes acyl chains in a coupled in vitro assay of CDPhas not been completely ruled out. diacylglycerol synthase and phosphatidylserine Patterson and Lennarz presented evidence for synthase; this observation supports the existence the presence of a phosphatidylserine-synthesiz- of a CDP-diacylglycerol-dependent phosphatiing activity dependent on CDP-diacylglycerol in dylserine synthase in this organism. In addition, membrane ghost preparations of Bacillus sp. this activity was shown to be membrane associstrain PP, an undefined derivative of Bacillus ated as judged by sucrose gradient centrifugamegaterium; this activity, unlike that from tion, thus establishing a membrane location for gram-negative bacteria, was stimulated by the this activity in a gram-positive anaerobe. addition of Mg2+ (21). This communication did There has been a recent report describing in not rigorously exclude the possibility of the as- vivo pulse-chase experiments on B. megaterium sociation of the enzyme with membrane-bound (ATCC 14581) which indicate a precursor-prodribosomes or the absence of the activity from uct relationship between a rapidly metabolizing the supernatant fraction of the cell. Carman and pool of phosphatidylglycerol and phosphatidylWieczorek (3) have reported a phosphatidylser- serine, respectively, suggesting no requirement ine synthase activity in a gram-positive anaer- for a CDP-diacylglycerol-dependent phosphatiobe, Clostridium perfringens, possessing a di- dylserine synthase activity in this organism (17). valent metal ion requirement and a membrane In vitro assays to establish the absence of a 535

536

DUTT AND DOWHAN

CDP -diacylglycerol-dependent phosphatidylserine synthase activity or the presence of an enzyme catalyzing the exchange of the glycerol moiety of phosphatidylglycerol for serine were not performed. Langley et al. (14) have reported the presence of a phosphatidylserine synthase activity dependent on CDP-diacylglycerol in extracts of B. megaterium KM (ATCC 13632) which also showed an in vivo labeling pattern similar to that reported for strain ATCC 14581. This strain is atypical in that phosphatidylserine constitutes an abnormally large fraction of the total phospholipid; they did not mention either the presence or absence of the activity in the bacilli strains studied which contained no detectable phosphatidylserine. The activity was reported to be associated with both the supernatant and pellet fractions separated by centrifugation at 100,000 x g. In addition, the authors did not rule out a possible stimulatory role for CDP-diacylglycerol in the reaction. In mammalian systems, an entirely different pathway for phosphatidylserine synthesis has been established (1), involving a reaction between CDP-ethanolamine and diacylglycerol, followed by the substitution of the head group of the resultant phosphatidylethanolamine by L-serine. In yeast systems, however, in vitro labeling studies on cell-free extracts indicate the simultaneous existence of both the mammalian and bacterial pathways of phosphatidylserine formation (29). Like the enzyme in Bacillus sp. strain PP, C. perfringens, and C. butyricum, the apparent CDP-diacylglycerol-dependent phosphatidylserine synthase activity in yeasts also appears to be particulate and dependent upon divalent metal ion for activity in marked contrast to E. coli enzymatic activity (10). Therefore, both bacteria and yeasts appear to possess a CDP-diacylglycerol-dependent phosphatidylserine synthesizing activity, but this has only been rigorously established for E. coli. The location of this activity, particularly in grampositive organisms, has only been partially investigated. In light of reports that some bacilli may not synthesize phosphatidylserine via a CDP-diacylglycerol-dependent pathway and since these organisms have been used in studies of membrane lipid synthesis and assembly (27), a clearer understanding of the lipid biosynthetic pathways of bacilli seems appropriate. In this communication, we report on a CDP-diacylglycerol:L-serine O-phosphatidyltransferase (EC 2.7.8.8) present in the membrane fraction of the aerobic bacteria B. licheniformis and B. subtilis.

J. BACTERIOL.

serine, [5-H]CMP, sn-[2-:H]glycerol, [; 2P]P, and [lJ'4C]uracil were obtained from Amersham Corp., and sn-[U-'4C]glycero-3-P was obtained from New England Nuclear Corp. Phosphatidylserine, phosphatidylethanolamine, phosphatidylglycerol, cardiolipin, and lysophosphatidic acid were products of Sigma Chemical Co., and lysophosphatidylethanolamine and lysophosphatidylserine were purchased from Avanti Biochemicals. Triton X-100 was obtained from Rohm and Haas, and octyl-,8-D-glucopyranoside (octyl glucoside) was from Calbiochem. Bacterial growth media were supplied by Difco Laboratories. Bacterial strains and media. B. subtilis (ATCC 6051) and B. licheniformis (ATCC 14580) were obtained as freeze-dried cultures from the American Type Culture Collection. The bacteria were grown at 32°C in medium (pH 7.0) consisting of 32 g of tryptone (Difco), 20 g of yeast extract, 5 g of NaCl, 13.6 g of KH2PO4, and 4 g of KOH per liter. Cells were harvested at late log stage of growth by centrifugation. Preparation of cell fractions. All of the following procedures were carried out at 4 to 5°C. Harvested cells (from 1 liter of growth medium) were washed once with 10 mM Tris-maleate (pH 7.0) containing 0.1 mM MgCI2. Cells were suspended in 25 to 30 ml of 50 mM Tris-maleate (pH 7.0) containing 0.1 mM MgCl2 and 5 mM 2-mercaptoethanol and broken by two passes through a French pressure cell. Unbroken cells were removed by a 10-min centrifugation at 3,000 x g. The membrane and supernatant fractions were separated by centrifugation at 100,000 x g for 1 h. The resulting pellet was suspended in the above buffer, recentrifuged, resuspended in buffer and used as the membrane preparation. The membrane fraction was extracted with detergent by including either Triton X100 or octyl glucoside at 2% (wt/vol) in the resuspension buffer. After the suspension was stirred for 1 h, it was centrifuged at 100,000 x g for 2 h; the supernatant constituted the detergent extract. Preparation of substrates. Phosphatidyl-L-[ 1'4C]serine (26), CI)P-1,2-diacyl-sn-glycerol (CDP-diacylglycerol) (16) and [5-'H]CDP-diacylglvcerol (25) were prepared as previously described. Phosphatidic acid used in preparation of phospholipid substrates was prepared from egg phosphatidylcholine (31). Cytidine-P- 2P- 1,2-diacyl-sn-glycerol was synthesized by reacting sn-[;32P]phosphatidic acid with CMPmorpholidate by the method mentioned above (16) for the synthesis of unlabeled CDP-diacylglycerol. The sn-[e32P]phosphatidic acid was prepared as follows. E. coli strain R477-100 (20) was grown at 37°C in 250 ml of LB broth (10 g of tryptone, 5 g of NaCl, and 5 g of yeast extract per liter) containing 5 mCi of [:32p]p,. The cells were harvested in the late log phase of growth, and the total lipid fraction was obtained by the Bligh-Dyer chloroform-methanol extraction procedure (11). The total lipid fraction was treated with crude cabbage phospholipase D (prepared from fresh cabbage) as described by Kates and Sastry (13), except that the reaction was allowed to proceed overnight with constant stirring. A thin-layer chromatogram of the products soluble in chloroform-methanol (1:2) MATERIALS AND METHODS showed 80% of the total radioactivitv comigrating with Chemicals. All chemicals were reagent grade or standard phosphatidic acid. The sn- P]phosphatidic better. The radiochemicals L-[3-3H]serine, L-[1- 4C]- acid was not purified further for use in the chemical

VOL. 147, 1981

PHOSPHATIDYLSERINE SYNTHASE OF BACILLI

conversion to cytidine-P-32P- 1,2-diacyl-sn-glycerol. The latter, however, was separated from unconverted sn-[32P]phosphatidic acid and other contaminants on a preparative silica gel, thin-layer plate in the solvent system chloroform-methanol-glacial acetic acid (65:25: 10). The 32P-labeled CDP-diacylglycerol was extracted from the silica gel, using acidic Bligh-Dyer solvent, as previously described (12). Enzyme assays. All assays were carried out at 37°C. One unit of enzymatic activity was defined as the amount of enzyme required to form 1 nmol of product in 1 min under optimal conditions. Specific activity was based on the units per milligram of protein as determined by the method of Lowry et al. (19). Phosphatidylserine synthase activity was followed by the incorporation of L-[3-3H]serine into chloroformsoluble material in the presence of CDP-diacylglycerol (24). The standard assay mixture contained 0.125 M Tris-maleate (pH 7.5), 1% Triton X-100 (wt/vol), 10 mM MnCl2, 10 mM L-[3-3H]serine (400 cpm/nmol), 0.125 mM CDP-diacylglycerol, 0.5 mM phosphatidylserine, and enzyme in a final volume of 0.2 ml; the reaction was terminated after 20 min by the addition of methanol containing 0.1 N HCI. Chloroform-soluble products were separated from water-soluble components by phase partitioning, as previously described (16), and counted for radioactivity. When octyl glucoside (1%) was present in the assay, Triton X-100 was omitted. The hydrolysis of CDP-diacylglycerol was followed by the release of water-soluble [5-3H]CMP

from [5-3H]CDP-diacylglycerol under the conditions described above. Sucrose gradient centrifugation. Either crude cell extracts or cellular fractions were sedimented through 5 to 20% sucrose gradients with a 70% sucrose shelf at the bottom as previously described (7). Centrifugation was carried out at 200,000 x g and 4°C for 2 or 3 h. The distribution of protein and nucleic acid in the gradients was determined by the absorbance at 260 nm (A260). The A280-A26o ratio (30) was used to distinguish between peaks rich in nucleic acid (ribosomal fraction) versus those rich in protein (membrane and supernatant fractions). Alternately, membrane fractions or ribosomal fractions were distinguished, using cells labeled with sn-[2_3H]glycerol or [ U'4C]uracil, respectively. Separation and analysis of lipids. After enzymatic reactions, appropriate carrier lipids (at 1 mg/ ml) were added before the reaction mixture was partitioned between chloroform and acidic (0.1 N HCI) methanol-water. The chloroform-soluble material was analyzed by two-dimensional silica gel, thin-layer chromatography as previously described (20), using chloroform-methanol-water (65:25:4) in the first dimension and chloroform-methanol-acetic acid (65:25: 10) in the second dimension. Carrier lipid and standards were detected by phosphate-positive spray (5) or brief exposure to 12 vapor. Autoradiography was used to locate :12P-labeled lipid. Radiolabeled material was quantitated by scraping and counting the silica gel.

RESULTS Subcellular localization of phosphatidylserine synthase. Based on the results of three different experimental approaches, i.e., differ-

537

ential centrifugation, detergent solubilization, and sucrose gradient centrifugation, the phosphatidylserine synthases of B. licheniformis and B. subtilis appear to be tightly membrane associated in crude cell extracts. The majority of the phosphatidylserine synthase activity was recovered in the pellet after centrifugation of crude extracts at 100,000 x g for 2 h (Table 1). The activity not recovered in the membrane fraction was accounted for in the cell supernatant presumably as unsedimented membrane fragments; the increase in specific activity in the membrane fraction is consistent with enrichment of this activity in the membranes. Extraction of the membrane fractions with octyl glucoside resulted in 75 to 80% of the activity no longer being sedimentable with good total recovery of activity. Similar solubilization results were effected by Triton X-100, but two successive solubilizations were required to attain comparable release of activity. Since differential sedimentation of the ribosomally associated phosphatidylserine synthase of E. coli under some conditions could be interpreted as membrane association (24), gradient centrifugation was used to verify the membrane location of this activity in B. licheniformis and B. subtilis; only results for B. licheniformis are shown since the results for B. subtilis were essentially the same. When extracts of these organisms were sedimented through sucrose gradients at low ionic strength, the profiles of A260 and phosphatidylserine synthase activity which were obtained are shown in Fig. 1A. The peak of absorbance near the bottom of the gradient just above the 70% sucrose shelf was due to membrane fragments, and the high absorbance towards the middle and the top of the gradient was due to the ribosomal and cytoplasmic fractions, respectively (24). Nearly all the recovered activity was found to be associated with the TABLE 1. Distribution ofphosphatidylserine synthase activity after differential centrifugation and detergent solubilization Phosphatidylserine synthase

Organism

Fraction

U/mg mg B. licheniformis

B. subtilis

Cell extract Membrane pellet Cell supernatant Detergent extract Detergent pellet

2.3 6.3 0.3 8.2 0.4

Cell extract Membrane pellet Cell supernatant Detergent extract Detergent pellet

1.8 2.5 0.9 3.9 0.8

Yield 100 79 19 61 1

100 62 24 50 10

538

DUTT AND DOWHAN

J. BACTFERIOL.

1.

-j r-

a) _

r

0

5

10

VOLUME FIG. 1. Sedimentation profile (volume in milliliters) of phosphatidylserine synthase after sucrose gradient centrifugation in the presence of 2 mM MgCIl,. The conditions are described in the text, and the bottom of the gradient is at the left. Symbols: O, A26X,; x, units per milliliter of synthase activity. (A) B. licheniformis cell-free extracts; (B) B. licheniformis cell-free extracts incubated with octyl glucoside and sedimented in the presence of 2% detergent.

membrane peak; the total recovery of activity was greater than 90%. This finding is in sharp contrast to the results of identical sucrose gradient centrifugation experiments performed with extracts of E. coli and four other gram-negative organisms, in which the bulk of the phosphatidylserine synthase activity was found to be associated with the ribosomes (7, 24). When the experiment was performed with extracts in the presence of 1 M NH4Cl throughout the gradient (data not shown), the activity remained associated with the membrane fraction, making an ionic interaction between the activity and membranes unlikely. To rule out the possibility of divalent cations being responsible for such an association, a similar experiment was performed with 5 mM Na2EDTA present throughout the gradient; the results were identical to those in Fig. IA. To

further confirm that the association was with membranes, extracts were incubated for 1 h at 4°C with 2% octyl glucoside. Sedimentation of such preparations through a gradient prepared with detergent throughout (Fig. IB) resulted in complete recovery of the activity at the top of the gradient presumably associated with detergent micelles. The absence of the membrane peak indicated complete solubilization of the membrane fraction. Furthermore, since no activity was found coincident with the ribosomal peak, it is unlikely that phosphatidylserine synthase is associated with membrane-bound ribosomes. To confirm that the absorbance peaks did indeed correspond to membranes and ribosomes, extracts of cells labeled either in the membrane fraction (sn-[2-3H]glycerol labeled) or the ribosome fraction ([U-"C]uracil labeled) were subjected to a similar sucrose gradient centrifugation. The radioactive profiles and enzyme activity profiles were wholly consistent with the assignment of absorbance and activity peaks in Fig. 1. The sum total of these results strongly suggests a tight membrane association for the phosphatidylserine synthase activity of B. licheniformis and B. subtilis. Properties of the phosphatidylserine synthase activity. Synthase activity was detectable at a low level under the assay conditions used for Bacillus sp. strain PP (21) extracts mainly due to the strong preference for manganese rather than magnesium as the divalent metal ion in the bacilli studied here (other divalent and monovalent ions were much less effective). The incorporation of radiolabeled serine into chloroform-soluble material in cell-free extracts is dependent on a divalent metal ion, the liponucleotide CDP-diacylglycerol and a nonionic detergent such as Triton X-100 (Table 2); octyl glucoside is also comparably effective at 1%. The increase in phosphatidylserine synthase activity was nearly linear with an increasing Mn2" concentration and reached a maximum around 10 mM. The inclusion of 1% Triton X100 brought about a maximum stimulation of activity of eightfold. Higher Triton X-100 concentrations were not inhibitory (data not shown) in contrast to what has been shown for the phosphatidylserine synthase activities in C. perfringens (3) and E. coli (15). Phosphatidylserine also stimulated the incorporation of serine into lipid product by cell-free extracts, and the combined stimulation by CDP-diacylglycerol and phosphatidylserine was synergistic, indicating that the effects may be mechanistically related. The effect of phosphatidylserine appears not to be protection of the newly formed lipid product from hydrolysis to water-soluble products, since

VOL. 147, 1981


TABLE 2. Requirements for phosphatidylserine synthase activity Product formed (nmol/mg of protein) b

Component(s)a B. IiB subchenifortilis mis

78 48 Complete system Minus MnCl2, plus 10 mM 6.1 2.2 EDTA 6.7 4.7 Minus CDP-diacylglycerol 6.5 5.6 Minus Triton X-100 12.0 8.2 Minus phosphatidylserine Minus CDP-diacylglycerol and 2.4 1.2 phosphatidylserine a The complete system contained the components of the standard assay, using radiolabeled serine as described in the text. b Based on incorporation of L-[3-3H]serine into chloroform-soluble material after a 25-min incubation with cell-free extracts.

no such hydrolase reaction could be detected with lipid labeled in the serine moiety under standard assay conditions; such a reaction does take place in the absence of divalent metal ions. Stimulation by phosphatidylserine is also not due to a carrier effect aiding the extraction of the labeled product by chloroform; adding phosphatidylserine just before termination of a carrier lipid-free assay mixture does not cause an increase in the amount of labeled lipid recovered. Other lipids, such as phosphatidylglyc-

erol, phosphatidylcholine, phosphatidylethanolamine, cardiolipin, and several lysophospholipids, including lysophosphatidylserine, had no effect on serine incorporation, thereby ruling out a generalized lipid effect. The variation in specific activity measurements between preparations of cell-free extracts was 30 to 40%, whereas variation within any single preparation was not signiflcant. The phosphatidylserine synthase activity from B. licheniformis was also measured by the L-serine-dependent release of [5-3H]CMP from [5-3H]CDP-diacylglycerol (Table 3). This extract also contained an apparent CDP-diacylglycerol hydrolase activity (indicated by serineindependent release of CMP) as do extracts from gram-negative bacteria (7, 23). The release of CMP stimulated by L-serine is consistent with the presence of a CDP-diacylglycerol-dependent phosphatidylserine synthase activity. The addition of phosphatidylserine had no significant effect on this serine-dependent release. After corrections were made for the hydrolase activity, the rate of CMP release was very comparable to the rate of serine incorporation in the absence

539

of phosphatidylserine (Table 2). Finally, the incorporation of radiolabeled L-serine into phosphatidylserine could be shown to be highly dependent on CMP in the presence of phosphatidylserine when CDP-diacylglycerol was absent (Table 4). Labeled serine incorporated into chloroformsoluble material was found exclusively in phosphatidylserine (no phosphatidylethanolamine was formed) in both bacilli as evidenced by silica gel thin-layer chromatography in chloroformmethanol-water (65:25:4) versus appropriate standards. The absence of detectable labeled phosphatidylethanolamine suggests the lack of measurable phosphatidylserine decarboxylase activity under the standard assay conditions. Radiolabel released from [5-3H]CDP-diacylglycerol was confirmed to be CMP by ascending paper chromatography versus standard CMP in saturated (NH4)2SO4 in 0.1 M potassium phosphate (pH 6.8)-isopropanol (100:2). To definitively establish the direct involvement (as opposed to a stimulatory role) for CDPdiacylglycerol in phosphatidylserine formation, the reaction of cytidine-P-32P-diacylglycerol (i.e., liponucleotide labeled in the phosphatidyl moiTABLE 3. L-Serine-dependent CMP-releasing activity in B. licheniformis Product formed

(nmol/

Component(s)

mg of

protein)' 18 Complete system' 16 Minus phosphatidylserine 5.6 Minus L-serine 5.6 Minus L-serine and phosphatidylserine aThe complete system was constituted as described for the standard assay mixture described in the text, except the liponucleotide as [5-3H]CDP-diacylglycerol was radiolabeled rather than the L-serine. Based on the formation of water-soluble [53H]CMP after a 25-min incubation with cell-free ex-

tracts.

TABLE 4. CMP-dependent incorporation of L-serine into lipid Product formed (nmol/ mg of protein)' Organism Minus CMP

Plus CMP

49 9.4 B. licheniformis 81 11 B. subtilis a The assay conditions were as described in Table 2, except that CDP-diacylglycerol was omitted from all reactions and 0.25 mM CMP was included where indicated. The reactions were carried out for 30 min.

540

DUTT AND DOWHAN

ety) with unlabeled L-serine was followed in cellfree extracts of B. licheniformis. The conversion of label in the phosphatidyl moiety to phosphatidylserine (no phosphatidylethanolamine was formed) was almost totally dependent on added serine in cell-free extracts (Table 5); the low level of conversion in the absence of serine as well as the formation of Px (unknown spot with an Rf similar to that of phosphatidylglycerol) was significantly reduced by using washed membranes (data not shown). The reaction is highly dependent on divalent metal ion and is not stimulated by lysophosphatidylserine. The divalent metal ion effect on the loss of CDP-diacylglycerol confirms the dependence of the phosphatidylserine synthase on manganese and rules out the possible role of the divalent metal ion in protecting the product against hydrolysis (Table 5). The two- to threefold stimulation by phosphatidylserine is consistent with the exchange of 32P-labeled phosphatidyl backbone between CDP-diacylglycerol and phosphatidylserine in the presence of L-serine. In summary, radiolabeled serine was incorporated into phosphatidylserine dependent on CDP-diacylglycerol and stimulated by phosphatidylserine. Release of label from cytidine-labeled CDP-diacylglycerol was stimulated by serine but not phosphatidylserine. The incorporation of radiolabeled serine into phosphatidylserine was also shown to be dependent on CMP in the absence of liponucleotide. In addition, the phosphatidyl moiety of the liponucleotide was shown to be directly converted to phosphatidylserine dependent on serine; this conversion was

J. BACTERIOL.

stimulated by phosphatidylserine. These results are consistent with an activity in membrane preparations of B. licheniformis, which catalyzes the formation of phosphatidylserine by a CDPdiacylglycerol-dependent pathway similar to that found in gram-negative bacteria (7). Similar results were seen in B. subtilis, which was not as extensively studied. The stimulatory effects of phosphatidylserine and CMP on the various reactions studied in this report suggest that the phosphatidylserine synthase in these organisms catalyzes a sequential BiBi reaction. It is known that in such mechanisms, the presence of one of the products in a reaction involving two reactants and two products can enhance an exchange reaction between a product and one of the reactants in the presence of the other reactant (4). Therefore, the incorporation of L-serine into lipid should be stimulated by addition of phosphatidylserine to serine plus CDP-diacylglycerol and by the addition of CMP to phosphatidylserine and serine.

DISCUSSION The subcellular distribution experiments reported here markedly distinguish phosphatidylserine synthases in several gram-negative organisms from these enzymes in the bacilli studied here as well as from yeasts (29), C. perfringens (3), and C. butyricum (28), which have been previously studied. These two types of enzymes are also distinguished by their dependence on divalent metal ions for activity. The membraneassociated activities studied in this communication appear not to follow a BiBi ping-pong kinetic mechanism as does the enzyme from E. coli (26), but rather a BiBi sequential reaction B. TABLE 5. Distribution of products of kinetic mechanism similar to the E. coli memlicheniformis phosphatidylserine synthase, using brane-associated phosphatidylglycero-P syncytidine-P-32P-diacylglycerol as substrate thase (8), which catalyzes a similar displacement % Distribution of 2P of CMP from liponucleotide. The physiological Component(s) CDP-DG PA PS PX" significance, if any, for these different mechanisms remains unknown. Now that the reaction 9 32 19 40 Complete systemb 15' 85 Minus enzyme catalyzed has been identified and the conditions 8 3 19 70 Minus L-serine and PS solubilization have been established, it for 6 14 17 63 Minus PS should be possible to purify these enzyme activ5 14 17 64 Minus PS, plus lyso-PS ities from gram-positive bacteria so that they 13 87 Minus MnCl2, plus EDTA in more detail and compared both PA (phos- can be studied 'Abbreviations: CDP-DG (CDP-diacylglycerol); and enzymologically to the E. coli physically PX (unknown spot phatidic acid); PS (phosphatidylserine); enzyme. with Rf of phosphatidylglycerol). bThe complete system was composed of the standard assay The presence of a CDP-diacylglycerol-demixture as described in the text, except the radiolabel as pendent pathway for the formation of phospharather in the was liponucleotide cytidine-P-_2P-diacylglycerol as than L-serine. The source of enzyme was a cell-free extract, tidylserine in B. licheniformis and B. subtilis, and the reaction was allowed to proceed for 60 min. The well as in the previously studied gram-positive chloroform-soluble material was subjected to chromatography bacteria, indicates considerable similarity beas described in the text. At least 50,000 dpm were analyzed in tween gram-negative and gram-positive bacteria each experiment. metabolism. 'Equal to the level of the phosphatidic acid in the labeled in their pathways for phospholipid On the other hand, the data of Lombardi et al. liponucleotide used in these experiments.

VOL. 147, 1981


(17) suggests that such a pathway for phosphatidylserine synthesis may not be necessary in the strain of B. megaterium which they investigated. Langley et al. (14) reported results similar to those of Lombardi et al. (17), but concluded that their results were due to a large, slowly metabolized pool of phosphatidylserine. We were able to detect a membrane-associated phosphatidylserine synthase activity in B. megaterium (ATCC 14581), as were Langley et al. (14), with similar properties to the B. licheniformis activity, but due to its low level we were unable to characterize it fully (data not shown). In none of these bacilli were we able to demonstrate an exchange of serine for the glycerol moiety of phosphatidylglycerol as should occur based on the labeling data of Lombardi et al. (17). Further studies, most likely using defined mutants in lipid metabolism, will be required to establish which pathways predominate in vivo. Finally, when assayed under the conditions of the enzyme from E. coli (6), we have noted in vitro levels of phosphatidylserine decarboxylase activity in B. licheniformis and B. subtilis comparable to the levels of synthase activity; similar results have been reported for other gram-positive bacteria (3, 14, 28). This activity is usually in excess relative to phosphatidylserine synthase in gram-negative organisms (7). This result may simply represent suboptimal assay conditions for the decarboxylase in gram-positive bacteria or may reflect a real difference which may affect the dynamics of phosphatidylserine metabolism in vivo. ACKNOWLEDGMENTS This work was supported in part by Public Health Service grant GM 20478 from the National Institutes of General Medical Sciences. In addition, A.D. was supported by a Molecular Basis of Cell Function grant GM 07542 from the National Institutes of Health.

1.

2.

3.

4. 5.

6.

LITERATURE CITED Borkenhagen, L. F., E. P. Kennedy, and L. Fielding. 1961. Enzymatic formation and decarboxylation of phosphatidylserine. J. Biol. Chem. 236:28-30. Carman, G. M., and W. Dowhan. 1979. Phosphatidylserine synthase from Escherichia coli. The role of Triton X-100 in catalysis. J. Biol. Chem. 254:8391-8397. Carman, G. M., and D. S. Wieczorek. 1980. Phosphatidylglycerophosphate synthase and phosphatidylserine synthase activities in Clostridium perfringens. J. Bacteriol. 142:262-267. Cleland, W. W. 1970. Steady state kinetics, p. 1-65. In P. D. Boyer (ed.), The enzymes, vol. 2. Academic Press, Inc., New York. Dittmer, J. C., and R. L. Lester. 1964. A simple specific spray for the detection of phospholipids on thin layer chromatograms. J. Lipid Res. 5:126-127. Dowhan, W., W. T. Wickner, and E. P. Kennedy. 1974. Purification and properties of phosphatidylserine decarboxylase of Escherichia coli. J. Biol. Chem. 249:

541

3079-3084. 7. Dutt, A., and Dowhan, W. 1977. Intracellular distribution of enzymes of phospholipid metabolism in several gram-negative bacteria. J. Bacteriol. 132:159-165. 8. Hirabayashi, T., T. J. Larson, and W. Dowhan. 1976.

9. 10.

11.

12.

13. 14. 15.

16.

17.

18.

19. 20.

21. 22. 23.

24.

25.

Membrane-associated phosphatidylglycerophosphate synthetase from Escherichia coli: purification by substrate affinity chromatography on cytidine 5'-diphospho-1,2-diacyl-sn glycerol Sepharose. Biochemistry 15: 5205-5211. Hirschberg, C. B., and E. P. Kennedy. 1972. Mechanism of the enzymatic synthesis of cardiolipin in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 69:648-651. Kanfer, J., and E. P. Kennedy. 1964. Metabolism and function of bacterial lipids. II. Biosynthesis of phospholipids in Escherichia coli. J. Biol. Chem. 239:17201726. Kates, M. 1972. Lipid extraction procedures, p. 347-353. In T. S. Work and E. Work (ed.), Laboratory techniques in biochemistry and molecular biology, vol. 3. NorthHolland Publishing Co., Amsterdam. Kates, M. 1972. Separation of lipid mixtures, p. 393-469. In T. S. Work and E. Work (ed.), Laboratory techniques in biochemistry and molecular biology, vol. 3. NorthHolland Publishing Co., Amsterdam. Kates, M., and P. S. Sastry. 1969. Phospholipase D. Methods Enzymol. 14:197-203. Langley, K. E., M. P. Yaffe, and E. P. Kennedy. 1979. Biosynthesis of phospholipids in Bacillus megaterium. J. Bacteriol. 140:996-1007. Larson, T. J., and W. Dowhan. 1976. Ribosomal-associated phosphatidylserine synthetase from Escherichia coli: purification by substrate specific elution from phosphocellulose using cytidine 5'-diphospho- 1,2-diacyl-sn-glycerol. Biochemistry 15:5212-5218. Larson, T. J., T. Hirabayashi, and W. Dowhan. 1976. Phosphatidylglycerol biosynthesis in Bacillus licheniformis. Resolution of membrane-bound enzymes by affinity chromatography on cytidine diphospho-sn-1,2-diacylglycerol Sepharose. Biochemistry 15:974-979. Lombardi, F. J., S. L. Chen, and A. J. Fulco. 1980. A rapidly metabolizing pool of phosphatidylglycerol as a precursor for phosphatidylethanolamine and diglyceride in Bacillis megaterium. J. Bacter,ol. 141:626-634. Louie, K., and W. Dowhan. 1980. Investigations on the association of phosphatidylserine synthase with the ribosomal component from Escherichia coli. J. Biol. Chem. 255:1124-1127. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. Nishijima, M., and C. R. H. Raetz. 1979. Membrane lipid biogenesis in Escherichia coli: identification of genetic loci for phosphatidylglycerophosphate synthetase and construction of mutants lacking phosphatidylglycerol. J. Biol. Chem. 254:7837-7844. Patterson, P. H., and W. J. Lennarz. 1971. Studies on the membranes of Bacilli: phospholipid biosynthesis. J. Biol. Chem. 246:1062-1072. Raetz, C. R. H. 1978. Enzymology, genetics, and regulation of membrane phospholipid synthesis in Escherichia coli. Microbiol. Rev. 42:614-659. Raetz, C. R. H., C. B. Hirschberg, W. Dowhan, W. R. Wickner, and E. P. Kennedy. 1972. A membranebound pyrophosphatase in Escherichia coli catalyzing the hydrolysis of cytidine diphosphate-diglyceride. J. Biol. Chem. 247:2245-2247. Raetz, C. R. H., and E. P. Kennedy. 1972. The association of phosphatidylserine synthetase with ribosomes in extracts of Escherichia coli. J. Biol. Chem. 247: 2008-2014. Raetz, C. R. H., and E. P. Kennedy. 1973. Function of

542

DUTT AND DOWHAN cytidine diphosphate-diglyceride and deoxycytidine diphosphate-diglyceride in the biogenesis of membrane lipids in Escherichia coli. J. Biol. Chem. 248:1098-

1105. 26. Raetz, C. R. H., and E. P. Kennedy. 1974. Partial purification and properties of phosphatidylserine synthetase from Escherichia coli. J. Biol. Chem. 249:50385045. 27. Rothman, J. E., and E. P. Kennedy. 1977. Rapid transmembrane movement of newly synthesized phospholipids during membrane assembly. Proc. Natl. Acad. Sci. U.S.A. 74:1821-1825.

J. BACTERIOL. 28. Silber, P., R. P. Borie, and H. Goldfine. 1980. The enzymes of phospholipid synthesis in Clostridium butyricum. J. Lipid Res. 21:1022-1031. 29. Steiner, S., and R. L. Lester. 1972. In uitro studies of phospholipid biosynthesis in Saccharomyces cerevisiae. Biochim. Biophys. Acta 260:222-243. 30. Warburg, O., and W. Christian. 1942. Isolierung and Kristallization des Garungs Ferments Enolase. Biochem. Z. 310:284-421. 31. Wells, M. A., and D. J. Hanahan. 1969. Phospholipase A from Crotalus adamantens venom. Methods Enzymol. 14:178-184.