39
Biochimica et Biophysics Acta, 574 (1979) @ Elsevier/North-Holland Biomedical Press
39-47
BBA 57403
PLASMA MEMBRANE LIPIDS OF HUMAN DIPLOID FIBROBLASTS NORMAL INDIVIDUALS AND PATIENTS WITH CYSTIC FIBROSIS
JOHN R. RIORDAN,
NOA ALON and MANUEL
BUCHWALD
Research Institute, The Hospital for Sick Children and Department University of Toronto, Toronto M5G 1X8 (Canada) (Received
November
lOth,
FROM
of Clinical Biochemistry,
1978)
Key words: Phospholipid composition; fibroblast, Plasma membrane)
Cystic fibrosis; Fluidity; Lipid analysis; (Human
Summary The lipid composition of isolated plasma membranes of human skin fibroblasts is described for the first time. Plasma membranes from a number of strains of fibroblasts from patients with cystic fibrosis and matched normals were isolated by a recently described procedure and analysed for major phospholipid classes, cholesterol and fatty acids. No differences in the quantities of these compounds were detected between cells of the two different origins. The fetal calf serum used to supplement the growth medium contained relatively more palmitoleate and oleate but less stearate than the membranes. There were also no consistent differences between cystic fibrosis and normal membranes in terms of the fatty acid compositions of their individual phospholipid classes. Consistent with this lack of chemical change in the lipids of membranes of cystic fibrosis cells, the degree of fluorescence polarization of diphenylhexatriene, an index of fluidity, was also unchanged.
Introduction Human diploid fibroblasts grown in culture are widely used in attempts to detect biochemical changes in genetic diseases. We are especially interested in detecting possible alterations in the structure and function of the surface plasma membranes of these ‘cells. We have purified their plasma membranes and partially characterized them [ 11. Currently chemical characterization of their molecular constituents is in progress and will serve as a basis of comparison for different genetic diseases. This paper describes the composition of the major lipid classes of a number of normal cell strains. Cells from age and
40
sex-matched patients with cystic fibrosis were also analysed. There were two main reasons for wanting to determine the membrane lipid composition in cells from patients with cystic fibrosis. First, there is an essential fatty acid deficiency, reflected most consistently as a decreased amount of linoleate in serum adipose tissue [3] and mucus-associated phosphatidylcholine [4] in [21, patients with cystic fibrosis. Although it is most generally believed that this represents decreased absorption of essential fatty acids as part of the general intestinal malabsorption associated with the disease [ 51, some authors [6] have suggested that there may be a basic defect in fatty acid metabolism. Analysis of the fatty acid composition of membrane lipids of cultured cells possessing the cystic fibrosis genotype but not the complicating secondary changes of the disease state should help to distinguish between these two possibilities. Second, a number of changes in membrane function have been proposed in cystic fibrosis [8] even though some of these are not well documented. Thus it seems possible that a feature of membrane structure common to several different functions might be altered. The lipid phase which provides the membranous environment for a variety of functional membrane proteins [9,10] would seem a good candidate for such a common factor. In fact, Christian et al. [ll] recently claimed an increased fluidity of membrane lipids of cystic fibrosis fibroblasts. Methods Cells were obtained and grown and plasma membranes isolated as described previously [ 11. Isolated membrane suspensions were purged with nitrogen or argon and stored at -70°C. Anaerobic conditions were maintained wherever possible in subsequent steps. Membrane protein was measured according to Lowry et al. [ 121. Lipids were extracted from membrane suspensions in CHC13/ CH30H (2 : 1, v/v) essentially according to Folch et al. [ 131. A portion of the extract was used for determination of total lipid phosphorus by the method of Bartlett [ 141, a second portion for separation of individual phospholipid classes by two-dimensional thin-layer chromatography according to Rouser et al. [ 151 and a third portion for determination of fatty acids and cholesterol by gasliquid chromatography. Phospholipid spots on developed thin-layer chromatograms were visualized by staining with I2 vapour. The silica gel in these areas as well as blank areas was scraped from the plates and digested in hot H2S04 and H202 prior to measurement of Pi. Fatty acid analyses were also performed on the individual phospholipids eluted from the scraped spots. Summation of the amounts of each fatty acid present was also used to quantitate the phospholipids. Fatty acids in the lipid extract were saponified with 0.1 N KOH in ethanol/ diethyl ether (3 : 1, v/v) at 100°C for 2 h. Hz0 was added to reduce the C,H,OH concentration to 50%. Cholesterol was extracted with an equal volume of petroleum ether. The ethanol/water phase containing the potassium salts of the fatty acids was back-washed 3 times with light petroleum followed by hydrolysis with 0.25 N HCl. Fatty acids were methylated by boiling in BF3/ CH30H (14X, w/v) for 5 min and the esters together with any aldehyde dimethyl acetals which may be formed from plasmalogens extracted with light
41
petroleum. C,,: 0 as internal standard was added to the original lipid extract. The methyl esters of fatty acids were resolved by gas-liquid chromatography on 10% DEGS (stabilised) on W, AW (100/200 mesh) at 185°C. Flame ionization detector responses were determined for each fatty acid relative to the internal standard and used for quantitation. Variability among replicate determinations was of the order of 2% in the case of arachidonic acid and less for the more prevalent fatty acids. Cholesterol was quantitated relative to cholestane as internal standard from gas-liquid chromatography on 3% QF-1 on W, AW-DHCS (100/200 mesh) according to van Hoevan and Emmelot [ 161. Detector responses over a wide range of amounts of cholesterol and a fixed amount of cholestane were also determined and used in the quantitation. The fluorescence polarization of diphenylhexatriene in the plasma membranes from normal and cystic fibrosis was determined according to the methods of Shinitzky and Inbar [17]. Results Phospholip
ids
The plasma membranes from both normal and cystic fibrosis cells contained about 0.45 mol total phospholipids/mg protein (range of 0.40-0.50 for 4 pairs of membrane preparations). Table I shows the proportions of individual phospholipid classes present in plasma membranes from the two sources. In both cases phosphatidylcholine is the dominant class, accounting for about 40% of the total. The other choline-containing phospholipid, sphingomyelin makes up at least 10% of the total so that these two constitute one-half of the total phospholipid. Approx. 20% of the total is phosphatidyl ethanolamine. Considerable amounts (nearly 10%) of the lyso forms of phosphatidylcholine and phosphatidylethanolamine are also present. The next most prevalent compound is phosphatidylserine, accounting for about 10%. Small amounts of phosphatidylinositol and phosphatidic acid were also present in both membranes as were some unidentified compounds.
TABLE
I
PHOSPHOLIPID
COMPOSITION
OF FIBROBLASTS
PLASMA
MEMBRANES
(% TOTAL
LIPID P)
Spots containing the compounds indicated were scraped from silica gel TLC plates following their separation by twodimensional chromatography [151. Total phosphorus in each was determined according to Bartlett [ 141 on four pairs of membranes; mean values ? S.D. are given. All were rounded off to the nearest whole number. Phospholipid
Normal
Cystic
Phosphatidylethanolamine Phosphatidylcholine Sphingomyelin Phosphatidylserine Phosphatidylinositol Phosphatidicacid Lysophosphatidylethanolamine Lysophosphatidylcholine Unidentified
21 ? 38 f lo? 9?1 621 3k 7t1 6kl 2+1
23 + 3 42 ? 2 9fl 7fl 4?1 1+1 6kl 7?1 2f 1
2 3 1
1
fibrosis
42 TABLE
II
FATTY
ACID
The
total
COMPOSITION
lipid
methyl&ion. Fatty
extracts
The
of
methyl
OF
FIBROBLAST
the
plasma
esters
of the
acid
PLASMA
membranes fatty
acids
were were
Myristate
(14
: 0)
2.1
PaImitate
(16
: 0)
24.6
: 1)
Cystic
to
saponification.
and quantitated
by
acid
hydrolysis
and
GLC.
1.9
(16
3.2
3.8
Stearate
: 0)
26.9
25.8
1)
27.5
31.6
(18
:
(18
: 2)
(18
Arachidonate
(20
Lignocerate
(24
: 4) : 0)
fibrosis
25.1
Palmitoleate Oleate
subjected
separated
Percent
Normal
Linoleate
MEMBRANES
2.0
2.5
4.7
4.2
8.8
5.2
Cholesterol The cholesterol contents of membranes from a number of matched normal and cystic fibrosis cell strains were measured and the mean value for four normal membrane8 is 140 + 10 (S.D.) ,ug/mg membrane protein (0.37 pmol/ mg). The four matched cystic fibrosis membranes contained a mean value of 150 f 12 (S.D.) ,ug/mg protein (0.38 pmol/mg). Hence, the cholesterol/phospholipid ratios were 0.82 for the normals and 0.84 for cystic fibrosis membranes. Total fatty acids The plasma membranes of human fibroblasts appear to have fatty acid composition (Table II). By far the predominant species stearate and oleate, each making up approx. 25% of the total mitoleate, linoleate, arachidonate and lignocerate contribute
TABLE
III
FATTY
ACID
The
total
were
grown
Fatty
acid
lipid
COMPOSITION extracts
was analysed
of
the
in the
:0 :1
16
: 0
24.1
16
:1
11.2
18
: 0
18.9
18
:1 :2
36.9
18
same
CALF serum
way
SERUM
used
as were
the
to
supplement
cell membranes
1.1
3.4 1.7 0.7
20
:0 :4
24
: 0
0.6
* Not
calf
1.5
18:3+20:0* 22
FETAL
Percent
14
14
OF fetal
7.0
resolved
from
each
other
on the
10%
DEGS
a fairly simple are palmitate, Myristate, palless than 5%
column.
the
medium
(Table
II).
in which
the
fibroblasts
ACID
2
:1 :
:4 :0
: 0
:
20
22
24
24
1
18 2 18:3or20:0
18
t
-
-
1.5
0.5
22
30
1
2
:1
: 0
18
resolved
and
t
1.5
0.5
24
28
4
39
1
2
12
9
-
7
2 -
16
31
3
16
2
3
10
8
6
3 -
17
30
4
17
2
3
CF
quantitated
TLC
CLASSES
dimensional
N
CF
were
two
N
esters
by
PHOSPHOLIPID
ethanolamine
41
16
OF
Phosphatidyl-
: 0
14
(a)
separated
choline
methyl
were
Phosphatidyl-
acid
16
fatty
:0 :1
acid
The
clauses
COMPOSITION
14
Fatty
thylated.
phospholipid
FATTY
The
IV
TABLE
GLC
the as with
2
3
2
2
6
1 -
18
19
16
27
N
spots
FIBROBLAST
2
2
5
1 -
16
19
18
25
2
3
CF
SphingomyeIin
by
[201,
OF
2
4
t
t
-
2
2
11
52
7
20
N
seline
t
t
2
1
10
51
7
21
3
5
CF
membrane
plates, acids.
1
3
t
t
-
1
1 -
11
16
9
60
N
inositol
t t
-
0
1 -
12
19
8
61
1
2
CF
‘t’.
phospbolipids
Phosphatidyl-
fatty
the
MEMBRANES
Phosphatidyl-
the total
from
PLASMA scraped trace
-,
2
6
-
2
1
1 3
25
18
13
29
N
choline
phatidyl
-
1
1
1 4
25
17
11
31
2
7
CF
none
saponified.
Lysophos-
quality:
eluted,
bydroIysed
and
1
1
t
t
2
3 -
14
48
5
16
N
ethanolamine
phatidyl
Lysophos-
detected.
t
t
1
2 -
12
52
6
18
1
2
CF
me-
44
each. In no case is there a significant difference between normal and cystic fibrosis membranes. The fatty acid composition of the fetal calf serum which is the only source of fatty acids in the growth medium is shown in Table III. It can be seen from this that the fibroblast membrane fatty acids are significantly different from those of the growth medium. Fatty acids of individual phospholipids
Table IV shows the fatty acid compositions of each of the major phospholipid classes. Phosphatidylcholine contains about 40% palmitate, 30% stearate, 25% oleate, less than 5% palmitoleate, 4% arachidonate, plus 2% myristate, 1% myristoleate and traces of other 18 and 20 carbon acids. Phosphatidylethanolamine has a markedly different composition with 15% palmitate, 30% stearate, 15% oleate, 17% arachidonate, about 10% each of Cz4 and C,,:, acids, l-3% myristate, myristoleate and palmitoleate and very small amounts of other 18 and 20 carbon acids. The third major phospholipid, sphingomyelin, is distinctly different from either of the first two with 25% palmitate and 15-20% each of palmitoleate, stearate and oleate. There is also 2 -3% myristate and myristoleate, 5-770 of the CzzzO acid and nearly 10% of a longer chain compound not unequivocally identified. Phosphatidylserine also exhibited its own unique fatty acid profile of 15% palmitate, 5% palmitoleate, nearly 50% stearate, 15% oleate, 3% myristate, 1% myristoleate, and about of C22:0. There is a notable absence of arachidonic acid. Phosphatidylinositol is distinguished by its exceptionally high palmitate content (60%) and also contains nearly 10% palmitoleate, 15-20% stearate, 10% oleate, 3% myristate, 1% myristoleate and only slight traces of compounds of longer chain length or greater degree of unsaturation. Lysophosphatidylcholine differs from phosphatidylcholine primarily in its palmitoleate content whereas lysophosphatidylethanolamine is distinguished from phosphatidylethanolamine by the very different proportion of stearate. 0.5 0.4 1
P
0.3 0.2
8%
I
oo,@Q
Q8aa@@
.
o.IJ-_____ 3.0
3.1
3.2
3.3
l/T
3.4
3.5
3.6
(OK-‘)x103
Fig. 1. Arrhenius plots of degree of fluorescence polarization (P) of diphenylhexatriene in isolated plasma membranes from CF (0) and normal (0) fibroblasts at temperatures ranging from 16 to 5O’C. Polarization measurements were made using an Aminco-Bowman spectrophotofluorometer with Glan prism polarizers. Values of P were calculated from the equation ‘w - Ivh p = _._~ IV, + fvh where I,, and f,h are emission intensities polarized on excitation with a vertically polarized beam.
in the vertical
and horizontal
directions
respectively
In none of these profiles are there significant between cystic fibrosis and normal membranes.
quantitative
differences
Lipid fluidity
As one measure of fluidity the fluorescence polarization of diphenylhexatriene in the lipid phase of cystic fibrosis and normal membranes was measured as a function of temperature and the results expressed in the form of Arrhenius plots (Fig. 1). As expected the polarization increased as the temperature was lowered reflecting the decreased lipid fluidity. The gently sloping curves are nearly linear and show no inflection points over the range of 50.-15°C. At all temperatures there were no significant differences in the polarization of the probe in the normal and cystic fibrosis membranes. Similar results were obtained with membranes from three additional paired strains of normal and cystic fibrosis fibroblasts. Discussion Little information is yet available on the lipid composition of the plasma membrane (or other subcellular membranes) of human fibroblasts grown in culture. However, relative amounts of the different phospholipid classes in the whole cells have been estimated from 32P-labelling experiments [ 181 and found to be 50% phosphatidylcholine and 25% phosphatidylethanolamine with sphingomyelin and phosphatidylserine contributing about 6% each and phosphatidylinositol 8%. Hence a comparison of the plasma membrane and whole cell compositions reveals an enrichment of sphingomyelin in plasma membranes as in the case with other cell types [ 19,201. There is a reduction of phosphatidylcholine in going from cell homogenates to plasma membranes so that the total amount of choline phospholipid is unchanged at about one-half of the total. The proportions of the other phospholipids determined are not markedly different in cell homogenates and plasma membranes. Estimation of the lyso forms in the whole cells were not made [IS]. The overall phospholipid compositions of our human fibroblast plasma membranes are remarkably similar to those reported by a number of laboratories for rat liver plasma membranes [21-231. However, relative to the values of van Hoeven and Emmelot [16] in a carefully performed analysis of liver membranes, the proportions of sphingomyelin and phosphatidylserine are lower in fibroblast membranes while considerable lysophosphatidylcholine and lysophosphatidylethanolamine are present (possibly in part reflecting hydrolysis of plasmalogen forms of phosphatidylcholine and phosphatidylethanolamine). Nevertheless, the strong similarity with the majority of reports is striking and somewhat surprising to us considering the species, tissue and growth conditions differences between these two types of plasma membrane. The relatively high levels of cholesterol are certainly characteristic of all plasma membranes which have been analysed [ 161 and together with sphingomyelin provide a useful plasma membrane marker. Palmitate and stearate comprised virtually all of the saturated fatty acids in the total lipid fraction of human fibroblast plasma membranes as is the case in most mammalian cell lipids [ 16,241; a trace of lignocerate is present. Oleate
46
contributes the major portion of unsaturated fatty acids with only small amounts of palmitoleate, linoleate, and arachidonate. More arachidonate and other polyunsaturates are present in homogenates of a mouse fibroblast cell strain, LM cells [ 241 and in rat liver plasma membranes [ 161. The comparison of the lipid compositions of fibroblast plasma membranes of cells derived from normal individuals and patients with cystic fibrosis failed to show any significant differences in phospholipid or fatty acid compositions. The latter observation is pertinent to the altered handling of essential fatty acids by patients with cystic fibrosis [ 251. Our findings would suggest that cultured cells containing the cystic fibrosis genotype [26] but not the secondary complications of the disease do not express any alteration in fatty acid metabolism. However, the fibroblasts were entirely dependent upon the serum supplement in the growth medium as a source of fatty acids. Therefore, it is possible that growth of these cells in serum-free medium with added selected essential fatty acids [27], such as linoleate which is markedly altered in lipids of patients with cystic fibrosis [ 28,291, might reveal an altered capacity of the cystic fibrosis cells to deal with them. Having first established the compositions of membranes derived from cells grown in serum containing medium, we are beginning such experiments. The lack of differences in the diphenylhexatriene polarization between cystic fibrosis and control membranes over a wide temperature range is consistent with the very similar chemical compositions of the membrane lipids. This finding also makes it unlikely that an altered fluidity of the bulk membrane lipid can account for the several changes in membrane function reported in cystic fibrosis [8]. Christian et al. [ 111 recently reported a decreased fluorescence polarization of a fluorophor associated with a phospholipid incorporated into cystic fibrosis fibroblast homogenates. However, the compound used, l-acyl-2-(N-4-nitrobenzo-2-oxa-l,3diazole)aminocaproyl phosphatidylcholine might be expected to report on a different location in the lipid bilayer than diphenylhexatriene which is oriented between and parallel to the fatty acid acyl side chains [ 301. Furthermore, the experiments of Christian et al. [ 111 were done on crude homogenates of only one cystic fibrosis fibroblast strain rather than on purified plasma membranes of several matched pairs of cystic fibrosis and normal cells. Acknowledgements This work was supported the Sellers Foundation.
by the Canadian
Cystic Fibrosis
Foundation
References 1
Kartner,
2
Kuo.
N..
P.T.,
3
Caren.
4
Sanjurjo.
Alon,
Huang,
R. and
Corbo.
P.. Allue.
5
Rosenlund.
6
Rivers,
7
Elliot,
M.L.,
J.P.W. R.B.
N.,
and
(1976)
Buchwald.
H.N.
and
L. (1966) X.
and
M. and
Basset,
D.R.
J. Clin.
Riordan, (1972)
Endocrinol.
Rodriquez-Soriano,
Selekman, Hassam, Pediatrics
J.A., A.G. 57,
Kim, (1975) 474-479
J.R.
H.J.
Metabol.
J. (1977) and
Lancet
(1977)
J. Pediatr.
26,
Lancet
Kritchevsky, 2, 642
J. Memb.
Biol.
36.191-211
60.394-403 470-477 1, 752 D. (1977)
Pediatrics
59.
428-432
and
47
6
Di Sant’Agnese,
9
Far&,
P.A.
R.N.,
Bloj,
and
B.,
Davis,
P.B.
Morero.
(1976)
R.D.,
N. Engl.
Sin&z,
F. and
J. Med. Trucco,
395,
534-541
R.E.
(1975)
Biochim.
Biophys.
Acta
415.
231-251 10
Lucy,
11
Christian.
J.A.
12
Lowry.
13
Folch,
14
Bartlett,
15
Rouser,
(1974) S.T.
O.H.,
Lett.
wanti,
M. and
G.R.
(1959)
ed.),
Vol.
16
Van
Hoeven,
17
Shinitsky,
J. Biol.
18
Malkiewicz-Wasowicz,
R.P.
Farr,
A.L.
and
and
234.
Emmelot,
Inbar.
B.,
P. (1972) 0.
Commun. (1951)
J. Biol.
A.
(1967)
New
York,
J. Memb.
Biochim.
Gamst.
R.J.
79,
J. Biol.
Chem.
226,
966-972 Chem.
193,
265-275
497-509
466-468
Dekker,
M. (1976)
Res.
Randall,
(1957)
Yamamoto.
Marcel
Biophys.
and
G.H.
Chem.
G.,
1,5105-5111
Biochem.
Sloane-Stanley.
1, 99-162,
M. and
Suppl.
(1977)
N.J.,
Kritchevsky,
G.V.,
40,
J.A.
Rosebrough,
J., Lees. G.,
FEBS
and
and
in Lipid Biol.
Biophys.
9.
Analysis
(Marinetti,
105.-126
Acta
Str$mme,
Chromatographic
NY
J.H.
433.
133-149
(1977)
Biochim.
Biophys.
Acta
482,
358-
369 19
Benedetti,
E.L.
33-120,
and
Academic
20
Patton.
S. (1970)
21
Pfleger,
R.C..
22
Wood.
23
Henning,
R.,
24
Horwitz.
A.F.,
R.
25
Campbell.
26
Changus,
27
Schroeder,
28
Sahu,
29
McEvoy,
30
Dale.
I.M., J.E.
R.E.,
F., F.A.
H.D.
Pitot, W.S.
(1975)
Chen,
L.A.
Stoffel, and
(1976)
J.F..
Lancet and
Membranes
(Dalton,
A.J.
F. (1968) (1970)
(1974)
R.B.
(1976)
Arch.
Am.
Biochemistry
Haguenau.
F..
7. 2826-2833
Hoppe-Seyler’s
M.M.
Caton.
Glaser,
Pathol.
M. and Rev.
Resp.
Proc.
Z. Physiol.
Natl.
Pediatrics Lab.
Vagelos, Dis.
Med. P.R.
Acad. 57,
100,
233-239
L. (1977)
J. Biol.
Chem.
252,
Chem. U.S.
351,
71,
1191-1199
3115-3119
480-486 7-11
(1976)
115,
Sci.
J. Biol.
2, 236 Brand,
and
141,174-182
W.
Burger,
D.N.
The
489-491
Biophys.
and
(1977)
in
NY
Snyder,
H.C.
Perlmutter,
Lynn,
29,
and
and
M.E.
Crazier, and
Biol.
Biochem.
Hatten,
(1968)
York,
N.G..
Arch.
Kaulen,
P.
New
J. Theor.
Anderson,
(1970)
S. and
Emmelot, Press,
7500-7510
Chem.
251,
5015-5026
eds.).