414
Biochimica @ Elsevier
et Biophysics Acta, Scientific Publishing
380 (1975) 414-420 Company, Amsterdam
-
Printed
in The Netherlands
BBA 56557
POLYGLYCEROPHOSPHATIDE
A. AUDET,
R. COLE
Department (Received
IN ESCHERZCHZA
COLZ
and P. PROULX
of Biochemistry, August
METABOLISM
22nd,
University
of Ottawa
(Canada)
1974)
Summary When Escherichiu coli B cells were labelled with [’ 4 C] glycerol and chased, there was a marked sparing of the phosphatidyl moiety compared to the nonacylated glycerol moiety of phosphatidylglycerol. When energydepleted cells were restored to an energy-rich medium there resulted a conversion of 32 P-labelled cardiolipin to phosphatidylglycerol, a lack of phosphatidic acid accumulation and no loss in total polyglycerophosphatide counts. In cellfree extracts, phosphatidic acid produced from 3 * P-labelled cardiolipin by the action of Escherichiu coli phosphalipase D, was readily recycled to form polyglycerophosphatide. In the presence of glycerol, such extracts displayed transphosphatidylase activity by degrading cardiolipin to phosphatidylglycerol mainly. The results as a whole indicate that the enzyme synthesizing cardiolipin together with cardiolipin-hydrolyzing phospholipase D constitute a cycle which is normally involved in the turnover of polyglycerophosphatides in Escherichia coli.
Introduction Evidence reported by Hirschberg and Kennedy [l] and by Hostetler et al. [2] indicates that cardiolipin synthesis in Escherichia co/i can proceed via a transphosphatidylation mechanism as in reaction 1. Very recently Cole et al. [3] reported the presence in this organism of a phospholipase D catalyzing reaction 2. 2 phosphatidylglycerol cardiolipin
s’
s’
phosphatidic
glycerol
+ cardiolipin
acid + phosphatidyl
The combined activities of reactions lipase D attack on phosphatidylglycerol.
(1) glycerol
(2)
1 and 2 are equivalent to a phosphoYet to be verified however is the
415
possibility that E. coli phospholipase D can also catalyze a transphosphatidylation reaction with glycerol and thus cause the reverse of reaction 1. The combined synthesis and breakdown reactions in this case would constitute an exchange mechanism between the nonacylated glycerol of phosphatidyl glycerol and free glycerol. In either case, the pooled activities can be tentatively designated as the polyglycerophosphatide cycle pending further characterization of the enzymes involved. Since phospholipase D activity is stimulated by ATP [ 31, the cycle is under energy control and this would explain the responsiveness of phosphatidylglycerol and cardiolipin levels to energy supply in the cell [4,5] . One could predict that the activity of the cycle in intact cells would result in a sparing of the phosphatidyl moiety of phosphatidylglycerol and a faster rate of turnover of the non-acylated glycerol moiety. Also, if either reaction 2 or the reverse of reaction 1 occurs in vivo one should be able to show under appropriate conditions a conversion of cardiolipin to phosphatidic acid and/or phosphatidylglycerol. This investigation is a study of some of the enzymes concerned with the polyglycerophosphatide metabolism of E. coli. Evidence is presented which indicates a functional phospholipase D acting on cardiolipin in intact cells. Methods Chase studies with 1 4 C-labelled cells E. coli B (ATCC 11303) and K19 were grown at 37°C to the early log phase in 300 ml of Medium 56 [6] containing 50 PCi [1,3-l 4 C] glycerol (spec. (act 20-25 Ci/mol). The cells were harvested at room temperature by centrifugation for 10 min at 3000 X g, transferred to Medium 56 containing 0.75 M glycerol and incubated at 37°C. At different times, 50-ml aliquots were removed for lipid analysis. Mild alkaline hydrolysis of phosphatidylglycerol isolated at 0 time indicated that 96118% of the label was in the glycerophosphorylglycerol product, acetolysis of this intact lipid revealed an equal distribution of label (within 2-3s) in both of its glycerol moieties. Studies on the turnover of cardiolipin in vivo E. coli B was cultured for 16 h at 37°C in energy rich medium containing per 1 15 g of bactopeptone, 1 g of yeast extract, 5 g of NaCl and 20 g of glucose autoclaved separately and 2 mCi [ 3*PI orthophosp,hate. The cells were then harvested and transferred to 0.1 M sodium chloride adjusted to pH 7.4 (energy free medium) and incubated at 37°C for 3 h. The cells were then collected by centrifugation and resuspended in complex medium to yield an absorbance of 0.2. At different times during this last incubation, 10 ml aliquots were removed for lipid analysis. Studies with cell-free extracts Cell-free extracts of E. coli B were prepared as described previously [3] and were used directly or following dialysis against 3 1 of 0.1 M phosphate buffer pH 7.3 at 4°C for a total of 19 h with 2 changes of buffer. A cytosol fraction was prepared from the undialysed cell-free extract, by centrifuging at
416
100 000 X g for 1 h. The preparations were used to assay phospholipase D [3] or to study incorporation of phosphatidic acid into more complex lipids under conditions specified in the text. Protein concentrations were determined according to the method of Lowry et al. (71. Preparation of lipid substrates 3 2 P-labelled cardiolipin (spec. act. 1.62-3.2 Ci/mol) was obtained as described previously [ 31. Labelled phosphatidic acid (spec. act. 0.92 Ci/mol) was prepared by the action of plant phosphalipase D IS] on E. coli phosphatidyleth~ol~ine labelled with [1,3-l 4 C] glycerol. The product was purified by thin-layer chromatography on silica gel G with choloroform/methanol/water (65 : 25 : 4, by vol.) and yielded only labelled glycerophosphate after mild alkaline hydrolysis. The glycerophosphate, Rf 0.23, was identified by paper chromatography with phenol/water (5 : 2, w/w) as solvent. Lipid phosphorus was determined by the method of Bartlett [lo]. Lipid analyses Lipids were extracted by the method of Bligh and Dyer [9] and separated on silica gel G plates with ~hlorofo~/meth~ol~acetic acid (65 : 25 : 8, by vol.} or chloroform/meth~ol~water (65 : 25 : 4, by vol.) as solvents. Components were revealed and counted as stated previously [3,11]. When necessary, phosphatidylglycerol was eluted from silica gel G by extraction according to the method of Bligh and Dyer [9] and subject to acetolysis [12] for at least 6 h. Labelled triacetin was trapped by adding 2 mg of unlabelled glycerol to the initial acetolysis mixture, The acetolysis products were separated by thin-layer chromatography on silica gel G with petroleum ether (b.p. 60-80”)/ ether/formic acid (55 : 45 : 1.5, by vol.) as solvent. Results and Discussion Results shown in Fig. 1 indicate that the non-acylated glycerol moiety of phosphatidylglycerol turns over at an initial rate about twice that of the acylated glycerol moiety in growing cultures of E. coEi B and E. coli K19. These data are in accord with a functional polyglycerophosphatide cycle in this organism. To further test the involvement of c~diolip~-specific phospholip~e D in intact E. eoli, cells of strain B were first cultured in complex medium containing [ 32 P] orthophosphate. They were then transferred to an energy-free medium, a condition which is known to increase the level of cardiolipin at the expense of phosphatidylglycerol according to reaction 1 [ 51. When these deprived cells were restored to their original energy-rich medium (Fig. 2) a marked reconversion of cardiolipin to phosphatidylglycerol occured. There was however no accumulation of phophatidic acid and phosphatidyleth~ol~ine levels like the total polyglycerophosphatide counts remained essentially constant throughout. When this experiment was repeated in the presence of pentachlorophenol an inhibitor of phosphatidic acid metabolism in Hemophilus parainfluenzae [ 131 there was again no accumulation of phosphatidic acid (re-
417
Fig. 1. (A) Turnover of non-acylated glycerol (~2.containing 135 000 dpm at 0 time) and acylated glycerol (0. containing 143 000 dpm at 0 time) of phosphatidylglycerol in E. coli B. (B) Turnover of non-acylated glycerol (Cl, containing 104 800 dpm at 0 time) and acylated glycerol (0. containing 108 900 dpm at 0 time) of phosphatidylglycerol in E. coliK 19. Both strains were labelled with [1.3-14Clglycerol. The non-acylated and acylated moieties were analysed following acetolysis of phosphatidyl glycerol.
suits not shown). This inhibitor was effective under the conditions used for E. coli B, by blocking the conversion of cardiolipin to phosphatidylglycerol and by causing breakdown of phosphatidylethanolamine as well as cessation of growth. The results at this point implied a complete incorporation of any phosphatidic acid formed into phosphatidylglycerol and/or a transphosphatidylation with cardiolipin rather than just hydrolysis of this lipid. When labelled cardiolipin was incubated for 30 min with cell-free extracts of E. coli in the absence of detergent (Table I) it was degraded to approxi-
T
15
TIME
*
30
45
(‘N~UTES)
Turnover of 32P-labelled phosphoglycerides of E. coli B cells transferred from an energy-free to an energy-rich medium, Phosphatidylethanolamine (Lx); phosphatidyl glycerol (0): cardiolipin (0); phosphatidic acid (a). Lipids were analysed as their mild alkaline hydrolysis products. The counts in each fraction are expressed as percent of total phosphoglyceride counts (700 000 dpm) recovered at the onset of the chase. Fig.
2.
418 TABLE
I
CONVERSION FREE (A)
Initially,
MM
ATP,
of
PHOSPHATIDIC
the
all the
l?.
added
to
The
incubation
ACID
tube.
2 mg
30
constitute
0.4
protein)
mM
and
contained tivated
19
2 mg enzyme.
mM
of Triton the
incubation
MgC12.
mixtures
coli B (4 mg protein)
control
incubation
(B)
OF
TO
POLYGLYCEROPHOSPHATIDES
BY
E.
COLI
CELL-
EXTRACTS
EDTA
X-100.
complete
mM
L-serine,
dpm
detergent.
Lipids
were
Conditions
added.
Incubations
were
The
complete
in 2 ml.
Incubations
mM
0.1
acid mild
Phosphatidic
L-swine
for
and
system
was
M phosphate
for
solubilized 30
alkaline
min
30
E. 1%
37°C.
hydrolysis
min
7.3,
10 mM
M&12,
in buffer.
at 37OC.
2.8
In the case
After
30
min
DLdi-glycerolphosphate
incubated
buffer in
at
pH
son&ted
10 mM
DLiu-glycerophosphate.
were
as their
buffer
cardiolipin
20 mM
10
phosphate
“2P-labelled
[14Clphosphatidic analysed
in 2 ml.
dmp
CTP,
system.
20
500
700
was also 0.4
containined
mM
contained
18
mixture
CTP. of
and
pH
another 7.3,
coli
Triton The
2.8
30 mM
B cell-free X-100.
control
The
min
at 37’C.
ATP, sonicate final
contained
of
were
10
mM
(8 mg mixture
heat-inac-
products.
Counts
recovered
in each
lipid
acid Phosphatidyl
Cardiolipin
glycerol
(A) ATP,
Mg2+,
ATP,
Mg2+
EDTA
700 5720
Complete
Heat-inactivated Active
preparation
600
preparation
710
16780
6370
4720
8160
4890
7180
6260
19000 1780
mately equal amounts of phosphatidic acid and phoshatidylglycerol whether or not cofactors such as CTP, serine and a-glycerophosphate were added to the medium. Under these conditions phosphatidic acid could not serve for synthesis of more complex lipid. However, when Triton X-100 was added in low concentrations to a mixture of co-factors, labelled cardiolipin and cell-free extract which had been preincubated together for 30 min, (complete system) there was no further breakdown of cardiolipin but an appreciable conversion of phosphatidic acid to phosphatidylglycerol did take place. In the presence of detergent but not appreciably in its absence, counts were lost to the water-soluble fraction which was not further analysed. Again when labelled phosphatidic acid was the substrate in the presence of detergent and the same cofactors, but in absence of cardiolipin all the product counts were recovered as polyglycerophosphatides and water-soluble material. Under these conditions there was no apparent conversion of phosphatidic acid to phosphatidylethanolamine. These results with cell-free extracts clearly indicate however that phosphatidic acid produced from cardiolipin breakdown or added directly to the medium, can be recycled for the synthesis of polyglycerophosphatides. This fact could in itself explain the lack of phosphatidate accumulation when cardiolipin is converted to phosphatidylglycerol in vivo but other explanations are possible. Evidence summarized in Table II reveals that cardiolipin breakdown occuring in dialysed cell-free homogenates or in the 100 000 X g (1 h) supernatant of the cell, greatly favors phosphatidylglycerol production over that of phosphatidic acid when glycerol is added to the medium. Thus in E. coli extracts breakdown of cardiolipin can proceed via a transphosphatidylation
419 TABLE
II
EFFECT OF GLYCEROL PREPARATIONS
ON CARDIOLIPIN-SPECIFIC
PHOSPHOLIPASE
D OF E. COLA CELL-FREE
(A) The incubation mixture contained in 2 ml, 2.8 mMATP, 10 mM Mg%,dialysed cell-free E. coli B sonicate (2.8 mg protein) 0.1 M phosphate buffer pH 7.3, 32P-labelled cardiolipin (6700 dpm) and glycerol as specified. Incubations were for 30 min at 37’. (B) The conditions were similar to those in (A) except that the 100 000 X g supernatant of E. coli B (3.4 mg protein) and 9700 dpm 32P-labelled cardiolipin were used. Conditions
Counts recovered as products
Ratio of product counts phosphatidylglvcerol
ATP 1% glycerol 4% glycerol 10% glycerol
1070 1100 1230 1690
1.4 1.6 2.2 2.7
Mg%, ATP 4% glycerol
5840 5500
1.5 2.9
phosphatidic acid
(A) Mg*,
(R)
causing an apparent reversal of reaction 1 as well as via reaction 2. Although the conditions used in these experiments also favor the conversion of glycerol to sn-3glycerophosphate, it is clear that glycerol itself can serve as phosphatidyl acceptor. Otherwise, because the proper phosphatase is lacking in cytosol [4], phosphatidylglycerophosphate should have been one product formed in this fraction but it was not. Whether this transphosphatidylase activity proceeds with stereospecificity in the case of glycerol and whether other acceptors are possible remains to be determined. On the basis of these results, the lack of phosphatidic acid production noted in Fig. 2 can be tentatively explained by a reversal of reaction 1. The results of this study together with other data pertaining to cardiolipin synthesis in E. coli [1,2,5] do strongly suggest the functioning of a polyglycerophatide cycle in intact cells. Our own data underline phospholipase D involvement. There is at the moment no definite evidence indicating whether this cycle is catalysed by a single enzyme responsive to energy levels in the cell or whether separate enzymes are involved in the synthesis and degradation of cardiolipin. As this study was in progress, Ballesta et al. [14] also observed unequal rates of turnover in the glycerol moieties of E. coli phosphatidylglycerol. On the basis of genetic and biochemical evidence they proposed a mechanism involving an exchange between the non-acylated glycerol of phosphatidylglycerol and free glycerol. Their results however, are completely compatible with a functional polyglycerophosphatide cycle in E. coli. With other types of chase studies, it can be readily shown that polyglycerophosphate metabolism proceeds with a net loss of counts in 3 2 P-labelled and [’ 4 C] acetate-labelled cells [ 11,151. The proposed cycle could not account for such loss since it spares phosphatidyl groups. To explain this loss, alterna-
420
tive pathways based on enzyme activities discovered in cell-free extracts [16181 can be proposed but whether such pathways act on endogenous lipids of normal, growing cells remains to be proved [ll].The ultimate metabolic fate of polyglycerophophatides in vivo appears to involve formation of complex water-solube products; however the metabolic route for this conversion is not yet elucidated [19]. Acknowledgement This investigation Canada.
was supported
by the Medical
Research
Council
of
References 1
Hirschberg,
2
Hostetler,
C.B.
and
K.Y..
Kennedy,
Van
den
E.P.
Bosch,
1972) H.
Proc.
Natl.
Van
Deenen,
and
Acad.
Sci.
U.S.
L.L.M.
69,
648451
(19’72)
Biochim.
Biophys.
Acta
260,
507-513 3
Cole.
4
Cronan,
R.,
Berms, J.E.
G. and
and
5
Rampini,
6
Monad,
7
Lowry,
8
Yang,
S.F.
(1969)
Bligh,
E.G.
and
9 10
C..
Barbu
O.H.,
and
A.,
11
Audet,
12
Renkonen,
13
One,
14
Ballesta,
15
Bright-Gaertner.
Y.
16
Okuyama,
17
Proulx,
18
Albright,
19
Van
Dyer,
Nantel, 0.
and
G. and D.C.
E. and
H. and P. and F.R.,
White,
L.M.G..
and
14,
208-211
(1971)
J. Bacterial.
D.A.
Schaechter,
S. (1969) (1969) and
Schulman,
H.
108, M.
and
W.J.
Biophys.
Paris Ser. Biophys. (1951) 37,
D. 270. Acta.
882-885
7, 585-599
J. Biol.
Chem.
193,
265-275
911-917 Acta
348,
334-343
1065-1071 (1973)
Biophys.
Acta
116,
Acta
270,
176.
12+124
210-214
40-49
47.1125-1128
(1973)
Kennedy
J. Bacterial.
Biophys.
J. Biochem.
Lennan.
325-332
25-60
42.298-304
Biochim.
Biochim. Can.
Physiol.
Biochim.
Sot.
P. (1972)
R.S.
337,
265,
466-468
P. (1974)
and
Proulx,
C.K.
J. Biochem.
234.
Sci.
Biochim.
Randall,
Acta.
Acta.
Acad.
(1951)
A.L.
Oil Chem.
C.L.
Biophys.
Biophys. C.R.
M.
Can.
Chem.
Proulx,
Nojima,
Fung,
Farr,
(1959)
J. Am.
De Garcia,
J. (1970)
Enzymol.
W.J.
Biochim.
Biochim.
Cohen,
N.J.,
J. Biol.
(1965)
White,
J.P..
Golde.
G. and
Methods
(1959)
(1972)
Polonovski,
Rosebrough,
G.R.
P. (1974)
P.R.
J.. Cohen-Bazier,
Bartlett,
1372
Proulx.
Vagelos,
E.P.
J. Biol. (1973)
Chem. Proc.
248, Natl.
3968-3977 Acad.
Sci.
U.S.
70.
1368-