The effects of phosphoglycerides on Escherichia coli cardiolipin synthase

et Biophysics 4cta ELSEVIER

Biochimica et Biophysics Acta 1214 (1994) 323-332

The effects of phosphoglycerides on Escherichia coli cardiolipin synthase Louis Ragolia, Burton E. Tropp *J Queens College, Department of Chemistry and Biochemist!y,

65-30 Kissena Blvd., Flushing, NY11367,

USA

Received 18 March 1994

Abstract Escherichia coli cardiolipin synthase catalyzes the conversion of two phosphatidylglycerol molecules to cardiolipin and glycerol. This enzyme was amplified in strain BL21(DE3) b earing recombinant plasmid pLR3, which was itself constructed by inserting the cls gene downstream from a T7 RNA promoter. Membranes from BL21(DE3)/pLR3 have over 1200 times more cardiolipin synthase activity than do comparable membranes from wild type cells. The enzyme was purified to homogeneity by extraction with Triton X-114 and chromatography on DEAE-cellulose. The purified enzyme migrated as a single band (46 kDa) on SDS-PAGE. This, along with SDS-PAGE analysis of induced protein, supports the notion that cls is the structural gene for cardiolipin synthase. Cardiolipin synthase activity was determined in a mixed micelle assay in which phosphatidyl[2-3H]glycerol was the substrate. The enzyme is inhibited by the product of the reaction, cardiolipin, and by phosphatidate. However, it is not inhibited by two other anionic phosphoglycerides, phosphatidylinositol and bis-phosphatidate. Phosphatidylethanolamine partially offsets inhibition by cardiolipin but not by phosphatidate. Magnesium chloride has the opposite effect. Cardiolipin inhibition of cardiolipin synthase probably plays an important role in regulating cardiolipin synthesis in E. coli. Keywords:

Cardioiipin; Cardiolipin synthase; Phosphoglyceride; Mixed micelle; Phosphatidylglycerol

1. Introduction Cardiolipin (CL), also called diphosphatidylglycerol, is one of three major phosphoglycerides in the Escherichia co/i cell envelope. The enzyme responsible for CL synthesis, CL synthase, catalyzes phosphatidyl group transfer from one phosphatidylglycerol (PG) molecule to another [1,2]. E. coli mutants have been isolated which synthesize only trace amounts of CL [3]. The gene responsible for the defect in CL synthesis, cls, maps at minute 27 of the E. coli genetic map [3,4]. An approx. 2.5fold increase in cls gene expression, as well as an increase in the CL to PG ratio, is observed as E. coli progress from early to late log phase under aerobic conditions [5]. Gene expression is also influenced by the terminal electron acceptor (increasing in the order oxygen, nitrate, fumarate) [5].

Abbreviations: amp, ampicillin; BSA, bovine serum albumin; CL, cardiolipin; IPTG, isopropyl-&Dthiogalactopyranoside; PA, phosphatidic acid; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; TX-100, Triton X-100, 7x-114, Triton X-114. * Corresponding author. Fax: + 1 (718) 997 5531. ’ Ph.D. Program in Biochemistry, The City University Queens College, Flushing, NY, USA.

of New York,

0005-2760/94/$07.00 0 1994 Elsevier Science B.V. All rights reserved SSDI 0005-2760(94)00112-C

The cls gene was originally cloned by Ohta et al. [6] and later by our laboratory [5]. Cells bearing the cls gene in a high copy number plasmid have about ten times more CL synthase than do wild type cells. However, the presence of this additional CL synthase has only a very slight influence on the CL level [6]. These observations raise questions about the physiological significance of changes in cls expression and suggest that CL synthesis is regulated at the enzymatic level. Hiraoka et al. have partially purified the E. coli CL synthase and examined some of its enzymatic properties [7]. However, their results do not explain how CL synthase is regulated. The present study reports the isolation of homogeneous CL synthase and the influence of various phosphoglycerides on the pure enzyme.

2. Materials and methods 2.1. Chemicals isopropyl P-D-thiogalactopyranoside Ampicillin; (IPTG); bovine serum albumin (BSA) (Fraction V powder);

L. Ragolia, B.E. Tropp/Biochimica

324

et Biophysics Acta 1214 (1994) 323-332

Table 1 Plasmid list Plasmid

Relevant properties

Ref.

pET3 pIBI20 pLR3 pLRll1 pPGL2019 pSHll1

Over-expression vector with T7 promotor BamHI site within polylinker cls inserted within overexpression vector. Intermediate plasmid with the cls gene flanked by two BamHI sites pgsA gene, ampicillin resistant. amp’ cls+

181

aProduct

a

This study This study

DO1

151

of IBI, New Haven CT.

BarnHI

J

& Pvull \

Hindlll

EoMl & EcoRV

Pvull/Sma

BamHl \

Fig. 1. Construction of pLR3. Plasmid pLR3 was constructed as described transcription termination sequence and 410, the T7 promoter (4 ). Distances

J

in Section 2. The cls gene is represented are in kilobases.

BamHl

by (m). T+ represents

the T7

L. Ragolia, B.E. Tropp/Biochimica

Sepharose CLdB; Triton X-114 (TX-114); DEAE-cellulose; glycerokinase (from E. coli); carbonic anhydrase; egg albumin; N,N,N’,N’-tetramethylethylenediamine; cardiolipin (CL) from E. coli; L-cu-phosphatidyl-m-glycerol (PG) from egg yolk lecithin; L-a-phosphatidylethanolamine (PE) from E. coli; t-a-phosphatidic acids (PA): dipalmitoyl, distearoyl, (cis-91, dioleoyl, and dimyristoyl; L-cY-phosphatidylinositol from soybean; L-cu-phosphatidylL-serine, dipalmitoyl; DNP-aspartate; blue dextran; and bromophenol blue were purchased from the Sigma Chemical, St. Louis, MO. Cytidine .5’-diphosphate-m-dipalmitin (CDP-diacylglycerol) and bis-phosphatidic acid (bis-PA), tetrapalmitoyl were purchased from Serdary Research Laboratories, Ontario, Canada. Bacto-tryptone, Bacto-yeast extract, and Bacto-agar are products of Difco Laboratories, Detroit, MI. Agarose, low melting point agarose, restriction endonucleases and T4 DNA ligase were acquired from BRL, Gaithersburg, MD or IBI, New Haven, CT. Triton X-100 (TX-loo) and calf-intestine alkaline phosphatase were purchased from Boehringer-Mannheim Biochemicals, Indianapolis, IN. GENECLEAN@’ kit was purchased from BIO 101, La Jolla, CA. [3H]Glycerol and [35S]methionine were obtained from New England Nuclear, Wilmington, DE. Carrier-free [ 32PIphosphate is a product of ICN, Irvine, CA. FAST STAIN@ was purchased from Zoion Research, Allston, MA. Polygram Sil G thin-layer chromatography plates were obtained from Brinkmann Instruments, Westbury, NY. All other chemicals were reagent grade or better. 2.2. Media and culture conditions M9ZB broth [8] and LB broth [9] were as described. The media were supplemented with 200 pug/ml of ampicillin (M9ZB-Amp or LB-Amp), and where indicated, IPTG was added to 0.8 mM. Cell growth at 37°C was monitored by measuring turbidity with a Klett-Summerson photometer (red filter). One Klett unit corresponds to approx. 5 . lo6 cells/ml.

325

et Biophysics Acta 1214 (1994) 323-332

by Silhavy et al. [ll]. Plasmid DNA was introduced into bacteria by a low efficiency colony transformation procedure [lo]. Restriction endonucleases, calf-intestine alkaline phosphatase, and T4 DNA ligase were all used in accordance with the distributors instructions. DNA fragments were separated by gel electrophoresis in either agarose (0.8%) or low melting point agarose (1.0%) and purified using a GENECLEAN@ kit according to the manufacturers instructions. 2.5. Plasmid construction Plasmid pSHll1 has a BamHI site on only one side of the cls gene. A BamHI site had to be introduced on the other side before cls could be introduced into pET3. The second BamHI site was derived from the polylinker in pIBI20 (Fig. 1). Plasmid pSHll1 was digested with a combination of PuuII and EcoRI, while pIBI20 was digested with a combination of EcoRV and EcoRI. To prevent pIBI20 self ligation, the digested plasmid was incubated with calf-intestine alkaline phosphatase. The DNA fragments from pSHll1 were separated by gel electrophoresis. The 2.2 kb fragment, containing the cls gene, was excised from the low melting point agarose and purified using the GENECLEAN@’ kit. This 2.2 kb fragment was mixed with digested pIBI20 and the two were ligated with T4 DNA ligase to form pLR111. HBlOl was transformed with pLR111, spread on M9ZB-Amp plates and the correct construction verified by analyzing restriction endonuclease digests. Plasmid pLRll1 was digested with BamHI, and the fragments produced were separated by gel electrophoresis. The 1.8 kb fragment, containing cls, was excised from soft agarose and purified by GENECLEAN @. This 1.8 kb fragment was inserted into a pET3 plasmid which had also been digested with BamHI. Insertion in the correct orientation was verified by analysis of restriction endonuclease digests. 2.6. SDS-polyacrylamide

gel electrophoresis

(PAGE)

2.3. Bacteria and plasmids E. coli HBlOl [F’ recA13 supE44 rpsL20 (sm’) hsds20 (rBW,rn,) arall galK2 lacy1 proA xy115 leu mtll h-1, a highly competent strain [lo], was purchased from Gibco BRL, Grand Island, NY. BL21(DE3) [F - ompT] [8], a lambda lysogen that contains T7 RNA polymerase under the control of a lacUV5 promoter was generously provided by W. Studier. HW55 [HfrC glpR cls-I] was constructed as previously described [4]. Plasmids used are listed in Table 1. 2.4. DNA isolation and manipulations Plasmids were isolated by the alkaline lysis method and plasmid DNA concentrations were determined as described

CL synthase purity and molecular weight were determined using the discontinuous buffer system described by Laemmli [12]. The gel was stained with FASTSTAIN@ in accordance with the manufacturers instructions and destained with multiple washings of 10% (v/v) aqueous glacial acetic acid. 2.7. Analysis of proteins labeled with [“‘S]methionine BL21(DE3)/pLR3 was incubated with [ 35S]methionine to label overproduced proteins so that their molecular weight could be determined. Five ml cultures of BL21(DE3)/pLR3 and BL21(DE3)/pET3 in M9ZB-Amp broth were incubated at 37°C with shaking at 250 rpm. When the cultures reached a turbidity of 200 Klett units,

326

L. Ragolia, B.E. Tropp/Biochimica

they were treated with IPTG for 30 min to induce T7 RNA polymerase. Then, rifampicin was added at a concentration of 200 kg/ml to block normal E. coli mRNA synthesis. After a 5 min incubation, 200 ~1 of cells were removed from the culture, and transferred to a sterile microfuge tube. These cells were labeled for 5 min with 4.0 PCi of [35S]methionine (1180 Ci/mmol), centrifuged for 30 s, and resuspended in 200 ~1 of PAGE buffer (50 mM Tris-HCl (pH 6.5), 2 mM EDTA, 1% /3-mercaptoethanol, 1% SDS, 8% glycerol, 0.025% bromophenol blue). This suspension was then placed in a boiling water bath for 5 min, and loaded on an SDS-polyacrylamide gel. A sheet of Kodak@ XAR film was placed over the dried gel and exposed for 1 week. 2.8. Lipid analysis Lipid distributions in cells harboring plasmids pLR3 or pET3 were determined with or without IPTG induction. Five ml cultures of either BL21(DE3)/pLR3 or BL21(DE3)/pET3 in M9ZB-Amp broth were incubated at 37°C with shaking. At approx. 40 Klett units, IPTG and 5 &i/ml of carrier-free [32P]phosphate were added. Then the cultures were incubated for a further 2 h. Cells were harvested by centrifugation and resuspended in 1.0 ml of distilled water. Phospholipids were isolated and chromatographed as previously described [4]. Radioactive lipids were detected by autoradiography, and quantified by cutting radioactive spots from the TLC plate and counting them in a liquid scintillation counter. 2.9. Amplification

of cardiolipin

synthase

BL21(DE3)/pLR3 was cultured in 400 ml of M9ZBAmp broth shaking (250 rpm) at 37°C. When the culture reached a turbidity of approx. 150 Klett units, IPTG was added and the cells were incubated for an additional 3 h. Unless otherwise stated all further procedures were performed at 4°C. Cells were harvested by centrifugation at 5000 X g, washed once with 0.1 M Tris-HCl (pH 7.8) containing 5.0 mM EDTA and 10 mM P-mercaptoethanol, and resuspended in 200 ml of the same buffer. The cells were then disrupted by sonication using a Model-W140 (Heat Systems-Ultra Sonics, Plainview, NY) sonicator with six 30 s pulses, at a setting of 8, pausing 15 s between pulses. Unbroken cells and cell debris were removed by centrifugation at 5000 X g for 5 min. The supernatant was collected and centrifuged in a Sorvall RC80 ultracentrifuge fitted with a T865 rotor, at 150000 X g for 1 h. The pellet, crude membrane, which contained approx. 36 mg of protein, was stored frozen at -70°C. No appreciable loss of enzymatic activity was observed over a 6 month period. 2.10. Purification

scheme

Extensive attempts were made to reproduce the binding of cardiolipin synthase to Whatman Pll phosphocellulose

et Biophysics Acta 1214 (1994) 323-332

as described by Hiraoka et al. [7]. Exact conditions were duplicated, including the use of Triton X-100 (TX-1001 for enzyme solubilization. The cardiolipin synthase produced by induced BL21(DE3)/pLR3 failed to bind to the column. It was therefore necessary to try another approach. Approx. 36 mg of thawed crude membrane was resuspended with a Teflon@ homogenizer in 12 ml Buffer I, which contained 100 mM potassium phosphate (pH 7.5) and 10 mM P-mercaptoethanol. The final protein concentration was adjusted to 3.0 mg/ml. The suspension, was stirred gently for 15 min, centrifuged at 150000 X g for 1 h, and resuspended in an equal volume of Buffer I containing 1.0% TX-114. After 1 h of mixing, the insoluble material was removed by centrifugation at 150000 X g and the supernatant (TX-114 extract) was carefully layered onto 12 ml of a buffered sucrose solution which had been incubated previously at 30°C. The buffered sucrose solution contained 6.0% sucrose, 0.06% TX-114, 100 mM potassium phosphate (pH 7.51, and 10 mM /3-mercaptoethanol. TX-114 aggregates, which included integral membrane proteins, were formed by placing the tube in a 30°C water bath for 5 min. Then, the tube was centrifuged at 300 X g for 5 min at room temperature. The supernatant was removed, and the oily droplet was washed with 2 ml of Buffer I at 30°C. Proteins were solubilized by diluting the oily droplet, with 12 ml of buffer containing 10 mM Tris-HCl (pH 7.5), and 10 mM P-mercaptoethanol at 4°C. A DEAE-cellulose (fine) column (1 cm X 8 cm) was equilibrated with Buffer A, which contained 10 mM TrisHCl (pH 7.51, 10 mM /3-mercaptoethanol, 20% sucrose, and 1.0% TX-100. The column was then charged with 12 ml of solubilized protein (0.17 mg/ml) and washed with four solutions (12 ml each): (i> Buffer A without sucrose; (ii) Buffer A without TX-100 or sucrose; (iii) Buffer I, and (iv) 5.0 mM potassium phosphate (pH 7.5) containing, 10 mM P-mercaptoethanol. Enzyme activity was eluted using a linear gradient containing 0.2% TX-100, 10 mM /3-mercaptoethanol and potassium phosphate (pH 7.5) from 5.0 mM to 400 mM. The eluate was collected in 3 ml fractions and assayed for protein, and enzymatic activity. Active fractions were pooled, mixed with an equal volume of 40% sucrose, and stored as frozen aliquots at -70°C. No appreciable loss of enzymatic activity was observed over a 6 month period. 2.11. Phosphatidyl[2-

“HIglycerol

preparation

Phosphatidyl[2-3H]glycerol was synthesized enzymatitally from [2-3H]glycerol. Glycerokinase was used to convert [2-3H]glycerol into [2-3H]glycerol 3-phosphate. Then crude membranes prepared from HW55/pPGL2019 were used to convert [2-3H]glycerol 3-phosphate and CDP-diacylglycerol to phosphatidyl[2-3H]glycerol as described by Ohta et al. [6]. Phosphatidyl[2-3H]glycerol was eluted from the Polygram Sil G thin-layer chromatography plate by the method of Kates [13]. The specific activity was adjusted by

327

L. Ragolia, B.E. Tropp / Biochimica et Biophysics Acta 1214 (1994) 323-332

the addition of unlabeled lipid to approx. 20 000 dpm/nmol of r_-cY-phosphatidyl-DL-glycerol. 2.12. CL synthase assay E. coli CL synthase was assayed at 37°C by following [2-3H]glycerol release as described by Hiraoka et al. [7] with one important modification. Unreacted phosphatidyl[Z3H]glycerol was removed by TCA precipitation instead of chloroform extraction. This modification resulted in a lower background, increasing the sensitivity of the assay. The standard assay mixture (50 ~1) contained 320 mM potassium phosphate (pH 7.1) 10 mM P-mercaptoethanol, 40 PM phosphatidyl[2- 3H]glycerol (20 000 dpm/nmol) and 0.03% TX-100. The tubes were preincubated at 37°C for 5 min and the reaction started by adding enzyme. Typically, the purified enzyme was between 2-32 ng, and the crude envelope was between 30-50 pg per assay. Reactions were terminated by adding 1 vol. of a solution which contained 20 mg/ml BSA and 10 mg/ml glycerol, followed by the quick addition of 2 vol. of 10% TCA. After brief agitation, the material was centrifuged at 12 000 rpm for 5 min in a microfuge at room temperature and 150 ~1 of the supematant counted. The results reported reflect incorporation by 75% of the assay mixture. One unit of enzyme activity is defined as 1 nmol of [2-3H]glycerol released per min [7]. When lipids were analyzed, the reaction was scaled up 20-fold and terminated by the addition of chloroform/methanol (1:2). The lipids were extracted as previously described [4] and chromatographed by thin-layer chromatography on Polygram Sil G plates in a solvent system of tetrahydrofuran/methylal/methanol/4 N ammonium hydroxide (50:25:25:5) [2]. 2.13. Mixed micelle formation The Sepharose CL-6B chromatographic technique of Carman and Dowhan [14] was used to show that the lipids under study form mixed micelles with TX-100. Sepharose CL-6B was packed in a glass column (1.5 cm X 35 cm) and washed with equilibration buffer containing 350 mM potassium phosphate (pH 7.0), 0.02% TX-loo, and 10 mM P-mercaptoethanol. The void volume and inclusion volume were determined with Blue Dextran and DNP-aspartate, respectively. The flow rate through the column was approx. 0.5 ml/min and fractions of about 0.8 ml were collected. Each of the following were chromatographed alone on a Sepharose CL-6B column in a volume of 0.5 ml: 1.0 mM TX-100, 0.125 mM phosphatidyl[23H]glycerol (650 dpm/nmol), and 0.075 mM cardiolipin (diphosphatidyl[Z 3H]glycerol) (650 dpm/nmol). A combination of TX-100, phosphatidyl[2-3H]glycerol and cardiolipin and a combination of TX-100, phosphatidylglycerol and diphosphatidyl[2- 3H]glycerol, at the above concentrations, were also analyzed by gel filtration. TX-100 elution

was monitored by absorbance at 275 nm. Radioactive phosphatidylglycerol and cardiolipin elution were followed by counting 200 ~1 of each fraction. Protein concentrations were determined by the method of Lowry et al. [15] as modified by Peterson [16]. Radioactivity was counted in an Isocap 300 scintillation counter with 6 ml of Ecoscint @-A scintillation liquid.

3. Results and discussion 3.1. CL synthase induction The plasmid pLR3 was constructed by inserting the cls gene into the BamHI site in the pET3 vector, just downstream from the T7 promoter (see Section 2). The plasmids pLR3 and pET3 were introduced into BL21(DE3), a strain with a T7 RNA polymerase gene under the control of the lacUV5 promoter. T7 RNA polymerase is induced when cells are incubated with IPTG. The T7 promoter’s influence on cls was tested by comparing induced and uninduced BL21(DE3)/pLR3 and BL21(DE3)/pET3 for lipid distribution, proteins formed, and CL synthase activity. The optimum IPTG induction time was determined to be approx. 2 h. Induction for longer periods produced no further increase in the specific activity of CL synthase from crude cell extracts (data not shown). As shown in Table 2, incubation with IPTG has a marked influence on the lipid composition of BL21(DE3)/pLR3 but not on that of BL21(DE3)/pET3. The CL level present in membranes from IPTG induced BL21(DE3)/pLR3 was over four times that of BL21(DE3)/pET3. The increased CL level was at the expense of PG. The PE levels were nearly the same in all cultures. These results indicate that the CL synthase activity in pLR3 is under the control of the T7 promoter. Ohta et al. have demonstrated that cls codes for a protein with a molecular weight of 46 kDa [6]. Therefore, induced BL21(DE3)/pLR3 should also code for a protein of this size but induced BL21(DE3)/pET3 should not. As described in Section 2, cells were incubated in the presence of IPTG, rifampicin, and [35S]methionine to label

Table 2 Phospholipid

distribution

Strain

BL21(DE3)/pET3 BL21(DE3)/pET3 BL21(DE3l/pLR3 BL21(DE3)/pLR3

Mel%

(0.8 mM IPTG) (0.8 mM WIG)

PE

PG

CL

75.8 77.6 77.2 71.9

19.1 18.0 16.2 6.3

5.1 4.4 6.6 21.8

Bacteria were cultured in 5.0 ml of M9ZB-Amp, shaking at 37°C and 250 ‘pm. At a turbidity of approx. 40 Klett units, IPTG was added, where indicated to 0.8 mM, along with 5 &i/ml of carrier-free [32P]phosphate, and the culture was incubated for another 2 h. Phospholipids were isolated and analyzed as described in Section 2.

328

L. Ragolia, B.E. Tropp/Biochimica

et Biophysics Acta 1214 (1994) 323-332 Table 3 Solubilization

and purification

of cardiolipin

synthase

Fraction purification

Total activity (units)

Protein (mgl

Sp. activity (units/mg protein)

Yield (o/o)

(-fold)

Crude membrane TX-114 DEAE

21000 17100 3400

30 2.2 0.8

507 2380 8400

100 81 16

1 4.7 16.6

Strain BL21(DE3)/pLR3 was induced with 0.8 mM IPTG and CL synthase was purified as described in Section 2. The crude membrane, TX-114, and DEAE-cellulose fractions correspond to the first high speed pellet, the TX-114 oil droplet dissolved in buffer (10 mM Tris-HCl (pH 7.5) and 10 mM P-mercaptoethanol), and active fractions from the DEAE-cellulose column, respectively.

1

2

Fig. 2. SDS-PAGE on whole-cells. Strain BL21(DE3)/pLR3 was cultured in M9ZB-Amp medium shaking at 37°C. At a turbidity of 200 Klett units, cells were induced with 0.8 mM IPTG, followed by the addition of 200 pg/mI rifampicin and 4.0 /.Ki of [35S]methionine (1180 Ci/mmol). Radioactive membrane proteins were extracted and analyzed by SDSPAGE as described in Section 2. Lane 1, corresponds to BL21(DE3)/pET3 and lane 2, corresponds to BL21(DE3)/pLR3.

proteins under the control of the T7 promoter. BL21(DE3)/pLR3 synthesizes one labeled protein (46 kDa) which is not also synthesized by BL21(DE3)/pET3 (Fig. 2). Virtually all of this protein is present in the cell membrane fraction (data not shown). Furthermore, its molecular weight is the same as that of purified CL synthase (see below), supporting the notion that cls is the structural gene for CL synthase. IPTG increased the specific activity of CL synthase in BL21(DE3)/pLR3 but not in BL21(DE3)/pET3 (data not shown). Upon analysis of crude membranes, the CL synthase activity from IPTG-induced BL21(DE3)/pLR3 was nearly 30 times higher than that from uninduced BL21(DE3)/pLR3 and over 1200 times higher than that from induced BL21(DE3)/pET3 (data not shown). 3.2. CL synthase purification Attempts to purify CL synthase from BL21(DE3)/pLR3 by the procedure of Hiraoka and coworkers [7] were unsuccessful because CL synthase did not bind to What-

man Pll phosphocellulose. The reason for the inability of the BL21(DE3)/pLR3 CL synthase to bind to phosphocellulose is not clear. Perhaps the CL synthase produced by BL21(DE3)/pLR3 has a slightly different amino acid sequence from the enzyme isolated by Hiraoka and coworkers. Because CL synthase did not bind to phosphocellulose, an alternate purification procedure was devised. The purification steps for CL synthase are summarized in Table 3. TX-114 was used to solubilize CL synthase. When the TX-114 extract was incubated at 30°C ten degrees above TX-114’s cloud point, an oil droplet containing CL synthase was formed. This cloud out step resulted in a nearly 5-fold purification (Table 3). Proteins associated with the oil droplet were purified by DEAE-cellulose chromatography. Elution with a linear phosphate gradient yielded a homogeneous protein in fractions 6-12 (Fig. 3). Earlier fractions had some enzyme activity but also contained other proteins (Figs. 3 and 4). At each stage of the purification, proteins were analyzed by SDS-PAGE and the results are shown in Fig. 4. As summarized in Table 3, CL synthase was purified nearly 17-fold by this purification scheme, corresponding to an approx. 40 OOO-fold increase in the purity of the

140,

i

,

"

,

1

I

,350

2wo

7

120 1500 f

g

60

.c B g n

60

s

E

low +

40

500

e

i

20

100 0 I 0

2

4

6

6

10

Fraction Fig. 3. DEAE-cellulose elution profile. The TX-114 droplet was dissolved in buffer (10 mM Tris-HCl (pH 7.5) and 10 mM P-mercaptoethanol) and the solution was chromatographed on a DEAE-cellulose column (1 X 8 cm) as described in Section 2. Enzyme was eluted with a linear gradient from 5.0 mM to 400 mM potassium phosphate (pH 7.51, containing 0.2% TX-100, and 10 mM pmercaptoethanol. The eluate was collected in 3 ml fractions and subjected to assays for protein (01, cardiolipin synthase activity ( w), and phosphate (0).

329

L. Ragolia, B.E. Tropp/Biochimica et Biophysics Acta 1214 (1994) 323-332

enzyme as compared to unamplified whole cells (see below). When lipids produced by the CL synthase assay were extracted and chromatographed as described in Section 2, two radioactive lipids were observed. One of these was unreacted PG and the other co-chromatographed with CL (data not shown).

200 3

150

2

100

b

50 0

3.3. CL synthase characterization The enzyme obtained was homogeneous, yielding a single electrophoretic band at 46 kDa when tested by SDS-PAGE, and had a specific activity of 8400 units/mg. The intact protein was subjected to Edman degredation. Trace quantities of amino acid derivatives were detected, suggesting that the amino terminus is blocked. Hiraoka et al. [7] reported that their partially purified CL synthase had an optimal activity at pH 7 in the presence of 0.015% TX-100, 400 mM potassium phosphate, and 1.0 mg/ml BSA. Although, their CL synthase binds to phosphocellulose and the CL synthase from BL21(DE3)/pLR3 does not, the optimal assay conditions for the two preparations were nearly the same. Two slight differences were observed. The enzyme isolated from BL21(DE3)/pLR3 has optimal activity at 320 mM potassium phosphate and 1.5 mg/ml BSA, while the corresponding values for the preparation of Hiroaka et al. were 400 mM and 1.0 mg/ml, respectively. (data not shown). The salt stimulation appears to be specific for the phosphate ion. Sodium phosphate also stimulates activity but potassium chloride and sodium sulfate at the same ionic strength do not (data not shown).

200 2

150

8

loo

E

50 0

m C

25

30

35

40 45 Fraction

50

7 E 5

s

3

&

2 0

55

60

Fig. 5. Sepharose CG6B chromatography of TX-loo, CL, and phosphatidyl[2-3H]glycerol. The column (1.5 X 35 cm) was equilibrated and eluted at room temperature (0.8 ml/min) with buffer consisting of 350 mM potassium phosphate (pH 7.0), 10 mM S-mercaptoethanol, 0.02% sodium azide, and 0.02% TX-100 in A and C, and no TX-100 in B. The following samples (0.5 ml) were applied: A, 1.0 mM TX-100, B, 0.125 mM phosphatidyl[2-3H]glycerol (650 dpm/nmol); and C, 1.0 mM TX100 and 0.125 mM phosphatidyl[Z3H]glycerol. The concentration of TX-100 was determined from its molar extinction coefficient (1.41. lo3 at 275 nm), and the concentration of both phosphatidyl[Z3H]glycerol was determined by liquid scintillation counting. The symbols used are (0) TX-100, and ( n ) PG.

3.4. Mixed micelle formation

3

4

5

Fig. 4. SDS-PAGE analysis of fractions during purification. At each stage of purification, a sample containing between lo-100 pg was analyzed by SDS-PAGE. Lane 1, corresponds to the crude membrane fraction; lane 2, to the proteins after the potassium phosphate wash; lane 3, to the TX-114 solubilized protein; lane 4, to the oil droplet dissolved in 10 mM Tris-HCI (pH 7.5) and 10 mM B-mercaptoethanol; and lane 5, to the pure DEAE-cellulose fraction. The arrow indicates the cls gene product of 46 kDa.

The TX-100 optima at 0.015% suggests that PG and TX-100 form mixed micelles. This possibility was tested by chromatographing TX-loo, PG, and a mixture of the two on Sepharose CL-6B. As shown in Fig. 5, PG alone elutes in the void volume and PG in 0.03% TX-100 elutes in the same position as TX-100. Since PG is converted to CL, CL’s influence on mixed micelle formation was also studied. CL alone elutes in the void volume and CL in 0.03% TX-100 elutes in the same position as TX-100 (data not shown). Furthermore, CL, PG, and TX-100 form a mixed micelle when present in the same mixture (data not shown). These results indicate that CL synthase can be studied under the defined conditions of a mixed micelle assay [17]. In addition, mixed micelles have been reported to mimic the biological membrane surface [18,19]. Although BSA stimulates CL synthase activity, it was omitted from the standard assay because it might interfere with the interpretation of the mixed micelle data. 3.5. Effects of enzyme concentration

and incubation

CL formation was followed as a function three different enzyme concentrations (Fig.

time

of time at 6). In the

330

L. Ragolia, B.E. Tropp/Biochimica 5000

4000

et Biophysics Acta 1214 (1994) 323-332

_~

3500

I 2500

tg

3000

v

ri.. 2000

1000

-

0 ,;:I:__:~ 0

120

_1 40

60

80

100

0’

Time (min) Fig. 6. Effect of enzyme concentration on CL formation. CL synthase assays were performed as described in Section 2, with enzyme concentrations of 40 ng/ml (01, 80 ng/ml (m), or 160 ng/ml (+). Glycerol release was measured after 5 min at 37°C.

reaction catalyzed by 80 ng/ml of CL synthase, the standard reaction condition, CL formation was linear for about 10 min. However, all three reactions came to a halt well before all of the PG was converted to CL. At each enzyme concentration, the reaction ceases when approx. 3.5 . lo3 mol of CL are formed per mol of enzyme. 3.6. Effects of phosphoglycerides

on CL synthase activity

The CL synthase reaction produces two products, CL and glycerol. Either could be responsible for the enzyme inhibition noted in Fig. 6. Therefore, CL synthase was assayed in the presence of various concentrations of glycerol and CL. Glycerol has no effect on CL synthase even when concentrations as high as 160 mM are used (data not shown). In contrast, CL has a marked inhibitory effect (Fig. 7) when present at a concentration of 25 /_LM(3.6 . lo3 mol of CL/mol CL synthase). This clearly shows that CL synthase is inhibited by its product CL. The fact that CL inhibits CL synthase means that estimates of enzyme activity for crude membranes given in Table 3 are probably too low because the high level of CL inhibits enzyme activity and unlabeled PG lowers the specific activity of the added phosphatidyl[2-3H]glycerol. Other anionic phosphoglycerides were also tested to determine whether they can inhibit CL synthase. As shown in Fig. 7, dipalmitoyl PA also has a strong inhibitory effect. Nearly identical inhibition was observed when dimyristoyl, dioleoyl, or distearoyl PA were tested (data not shown). bis-PA and CDP-diacylglycerol have virtually no inhibitory effect, and phosphatidylinositol has a slight stimulatory effect. Zwitterionic lipids were also tested. PE and phosphatidylserine, at a concentration of 25 PM, stimulate activity by approx. 20% (data not shown).

” 0

” 5

” 10

I

” 15

20

25

30

Phospholipid (PM) Fig. 7. assays CL ( + added after 5

Effect of phosphoglycerides on CL synthase activity. CL synthase were performed as described in Section 2, except that additional ), PA (a), CDP-diacylglycerol (A ), PI ( n ), or bis-PA (0) were at the indicated concentrations. Glycerol release was measured min at 37°C with a purified enzyme concentration of 320 ng/ml.

Since E. coli membranes have a fairly high PE concentration, the effect of this lipid on CL synthase activity was studied in more detail. As shown in Fig. 8, PE not only influences the rate of CL formation but also the extent of CL formation. PE has a stimulatory effect when added at the start of the reaction but has no effect once the reaction is complete (Fig. 9). /I-Mercaptoethanol and BSA also do not stimulate the completed reaction (data not shown) and PG has virtually no effect (Fig. 9). The addition of another 2 ng of enzyme does, however, result in more CL synthesis (Fig. 9).’ This, along with analysis of lipids in the reaction mixture (data not shown), demonstrates that the reaction mixture still contains unreacted PG.

o

y,

0

.-.

20

40

60

80

100

Time (min) Fig. 8. Effect of PE on CL synthase activity. Purified CL synthase (2 ng) was assayed at 37°C as described in Section 2. Tubes were pre-incubated either in the absence of PE (0) or the presence of 24 PM PE ( n ), and glycerol release followed with time.

L. Ragolia, B.E. Tropp/Biochimica

CL and PA inhibition appear to be different. PE offsets CL inhibition but has no effect on PA inhibition (Fig. lOA). Magnesium chloride has the opposite effect; it reverses PA inhibition but not CL inhibition (Fig. 10B).

800 600 -

3.7. Effects of nucleotides

00

20

40

60

80

100

Time (min) Fig. 9. Reversal of CL synthase inhibition. Purified CL synthase was assayed as described in Section 2. At 60 min, when the reaction was complete, the reaction mixture was divided into four equal portions. Then, either distilled water (a), (2.0 nmol PE (A), 2.0 nmol phosphatidyl[2-‘H]glycerol (+), or 2.0 ng CL synthase (w) were added and glycerol release determined at the indicated times.

2500

331

et Biophysics Acta 1214 (1994) 323-332

,

I

0’ 5

0

’

”

10

15

on cardiolipin

synthase activity

Because the energy state of the bacteria seems to influence CL levels [2], the effects of various nucleotides on CL synthase activity were examined. ATP, ADP, and AMP have no effect on CL synthase activity when added at 10 mM in the presence of 320 mM phosphate. However, 10 mM ATP had a slight stimulatory effect on CL synthase activity in the absence of phosphate (data not shown). These experiments show that CL, the product of the CL synthase catalyzed reaction, inhibits the E. coli enzyme. CL has also been reported to be an inhibitor of Micrococcus lysodeikticus CL synthase [21]. However, the two enzymes appear to be different since Micrococcus lysodeikticus CL synthase is also inhibited by glycerol, the other end product of the reaction, as well as by PE and phosphatidylinositol. CL inhibition of the E. coli enzyme probably helps to regulate the PG to CL ratio in vivo. In this regard, it is interesting to note that IPTG induced BL21(DE3)/pLR3 have a very high CL to PG ratio. This is probably due to the fact that a very high CL concentration is required to inhibit the extraordinary high CL synthase level present in the induced cells. Of course, other factors such as CL location in the cell envelope may also play an important role in CL regulation.

’ 20

25

Acknowledgments Phosphatidylethanolamine

2500

(pM)

,

I

Magnesium

chloride

We wish to thank William Studier and William Dowhan for kindly providing strains and plasmids. We thank Sheldon Heber and Yu-wen Hwang for helpful discussions. The data have been taken in part from a dissertation submitted to the Faculty of Biochemistry of the City University of New York in partial fulfillment of the requirements for the PhD degree. This research was supported by NIH (General Medical Science Award GM 34688) and a BHE/PSC Grant from the City University of New York.

(mM)

Fig. 10. (A) Effect of PE on the anionic phosphoglyceride inhibition of CL synthase. Purified CL synthase (320 ng/ml) was assayed as described in Section 2. Tubes were pre-incubated with 12 PM CL (0) or 12 PM PA (m) and the indicated concentration of PE. Glycerol release was measured after 5 min at 37’C. (B) Effect of magnesium on the inhibition of CL synthase by anionic phosphoglycerides. Purified CL synthase (320 ng/ml) was assayed as described in Section 2. Tubes were pre-incubated with 12 PM CL (0) or 12 PM PA ( n) and the indicated concentration of magnesium chloride. Glycerol release was measured after 5 min at 37°C.

References [II Hirschberg,

C.B. and Kennedy, E.P. (1972) Proc. Natl. Acad. Sci. USA 69, 648-651. ]2] Tunaitus, E. and Cronan, J.E. Jr. (1973) Arch. B&hem. Biophys. 155, 420-427. [3] Pluschke, G., Hirota, Y. and Overath, P. (1978) J. Biol. Chem. 253, 5048-5055. ]4] Hwang, Y.W., Engel, R. and Tropp, B.E. (1984) J. Bacterial. 157, 846-856.

332

L. Ragolia, B.E. Tropp/Biochimica

(51 Heber, S. and Tropp, B.E. (1991) Biochim. Biophys. Acta 1129, 1-12. [6] Ohta, A., Obara, T., Asami, Y. and Shibuya, I. (1985) J. Bacterial. 163, 5066.514. [7] Hiraoka, S., Kazuki, N., Uetake, N., Ohta, A. and Shibuya, I. (1991) J. Biochem. 110, 443-449. [8] Studier, F.W. and Moffatt, B.A. (1986) J. Mol. Biol. 189, 113-130. [9] Miller, J.H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. [lo] Hanahan, D. (1985) in DNA cloning Vol. 1, a practical approach (Glover, D.M., ed.), pp. 109-135, IRL Press, Oxford. [ll] Silhavy, T.J., Berman, M.L. and Enquist, L.W. (1984) Experiments with gene fusions, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. [12] Laemmli, U.K. (1970) Nature 227, 680.

et

BiophysicsActa 1214 (1994) 323-332 [13] Kates, M. (1986) Techniques of Lipidology (Burden, R.H. and van Knippenberg, P., ed.), Elsevier Science, Amsterdam. [14] Carman, G.M. and Dowhan, W. (1979) J. Biol. Chem. 254, 83918397. [15] Lowry, O.H., Rosenbrough, N.J., Farr, A.L. Znd Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. [16] Peterson, G.L. (1977) Anal. Biochem. 83, 346-356. [17] Lin, Y.P. and Carman, G.M. (1990) J. Biol. Chem. 265, 166-170. [18] Dennis, E.A. (1983) in The Enzymes (Boyer, P.D., ed.), 16, pp. 307-353, Academic Press, New York. [19] Hannun, Y.A., Loomis, C.R. and Bell, R.M. (1985) J. Biol. Chem. 260, 10039-10043. [20] Ohta, A., Waggoner, K., Radominska, A. and Dowhan, W. (1981) J. Bacterial. 147, 552-562. [21] De Siervo, A. (1975) Can. J. Biochem. 53, 1031-1034.