two separate pools of sphingomyelin in bhk cells

Biochimica et Biophysics Acta, 1124 (1992) 95-100 0 1992 Elsevier Science Publishers B.V. All rights reserved 0005-2760/92/$05.00

95

BBALIP 53845

Two separate pools of sphingomyelin in BHK cells

Paul Quinn and David Allan

Department of Physiology. Unil’ersity College London, London (U.K.)

(Received 16 July 1991)

Key words: Sphingomyelin; Plasma membrane; Lipid biosynthesis; (BHK cell)

In BHK cells labelled to equilibrium with [3Hlcholine and treated with sphingomyelinase the surface pool of sphingomyelin is degraded very rapidly (half-time 10 min) but the internal pool of sphingomyelin which accounts for about 30% of the total is only degraded slowly (half-time about 80 h) showing that the internal pool does not normally reach the surface. In [3Hlcholine incorporation experiments the internal pool begins to accumulate radioactivity at about the same time as phosphatidylcholine (30 min) but label does not enter the surface pool of sphingomyelin for a further 90 min. The internal and external pools reach the same specific activity only after about 20 h. Pulse-chase analysis with [3H]choline shows that radioactivity in each pool of sphingomyelin continues to increase when the specific radioactivity of phosphatidylcholine is decreasing, consistent with both pools being synthesised from a phosphatidylcholine precursor. The results suggest that sphingomyelin in BHK cells is present not only in the plasma membrane but also in a more rapidly labelling pool which does not mix with the surface pool.

Introduction

Sphingomyelin is the most common phospholipid in

the outer leaflet of the plasma membrane in a variety of different cells [1,21 and a knowledge of its mode of synthesis and introduction into the cell surface is es- sential not only for understanding the way in which plasma membrane phospholipid asymmetry is gener- ated and maintained, but also for elucidating the gen- eral mechanism of intracellular lipid trapsport [3,4]. Most workers have concluded that sphingomyelin is synthesised by transfer of phosphorylcholine from phosphatidylcholine to ceramide but there is still a question regarding the subcellular site of this activity. Although it was originally thought that sphingomyelin is synthesised in situ in the outer leaflet of the plasma membrane [5-121 more recent information suggests that like other sphingolipids [13], sphingomyelin is synthesised in the Golgi apparatus [14-211. However, most of this recent data depends on the incorporation of exogenously added short chain ceramide derivatives which need not necessarily reflect the normal metabolic pathways for de novo synthesis of sphingomyelin. Fur- thermore, some of these experiments have been done with liver and other tissues constitutively synthesising lipoproteins (which are often rich in sphingomyelin [22]

Correspondence: D. Allan, Dept. Physiology, University College

London, University St., London WClE 655, U.K.

and it is not clear whether the ceramide derivatives are incorporated into membranes or into the lipoprotein content of Golgi vesicles. The problem of the precise cellular localisation of sphingomyelin synthesis is there-

fore of continuing interest. In previous work [23] we noted that exposure of

BHK cells to S. aureus sphingomyelinase caused the

rapid breakdown of about two thirds of the cell sphingomyelin, which was consequently assumed to be present on the cell surface. The resulting surface pool of ceramide could be reconverted to sphingomyelin over a period of 2-3 h at the expense of a small pool of phosphatidylcholine. Similar results have been reported by Slotte and co-workers [24]. The remaining one third of total sphingomyelin which was refractory to external sphingomyelinase during a 3 h incubation period appeared to represent an intracellular pool which was not in rapid equilibrium with the surface. In the present paper we show that the two pools of sphingomyelin remain separate for long periods and that the incorporation of [‘HIcholine into surface sphingomyelin is much slower than that of the internal pool. The data has also blen used to obtain an esti- mate of the time taken for newly synthesised sphingomyelin to appear at the cell surface.

Materials and Methods

BHK21 cells were grown on 3.5 cm plastic dishes in Glasgow MEM buffered with Hepes and supplemented

96

with 10% tryptose phosphate, 2 mM glutamine and 5% foetal calf serum (all obtained from Gibco). [“HICholine (spec. act. 76 Ci/mmoI) was obtained from Amersham International (Amersham, Bucks, U.K.). S. aureus sphingomyelinase was obtained from Sigma Chemical (Poole, Dorset, U.K.) as a solution in 50% glycerol containing 240 pg/ml of protein and with a quoted activity of 160 I.U./mg of protein.

Cells were labelled to equilibrium by incubation for 40 h with [“HIcholine (0.5 yCi in 2 ml of the above growth medium. Glasgow MEM contained 2 pg/ml free choline but the serum added to the growth medium contributed an additional 2 pg,/ml free choline together with about 4 pg/ml of bound choline as phosphatidylcholine present in serum lipoprotein). Cells were washed twice with 2 ml of serum-free medium (2 pg/ml free choline) with the addition of 0.2% bovine serum albumin and 100 pg/ml choline chloride (Sigma) before incubation in the same serum-free medium with and without addition of 5 ~1 of sphingomyelinase. In some experiments, the incubation was carried out in complete growth medium. At various times up to 30 h the medium was removed from duplicate samples of treated and untreated cells and 1.9 ml of methanol/ chloroform (2 : 1) was added to extract Iipids.

For measurement of incorporation of radioactivity into choline lipids, non-confluent cells were washed once with 2 ml of serum-free medium (which contained 2 ,ug/ml choline chloride and 0.2% albumin) and then incubated with 1 ml of the same medium to which was added 0.5 PCi of [‘HIcholine. After various times of incubation 10 ~1 of sphingomyelinase was added to two dishes. 10 min later the medium was removed and lipids were extracted from duplicate control and enzyme-treated samples as above.

For pulse-chase experiments, cells were labelled for 90 min as above using 1 ,uCi/ml of [“HIcholine and then the medium was replaced with 1 ml of serum-free medium containing 100 pg/ml choline chloride. At various times up to 10 h, 10 ~1 of sphingomyelinase was added to duplicate samples and after 10 min incubation, these samples together with matched untreated controls were extracted with chloroform/ methanol as above. Extracts were fractionated by the procedure of Bligh and Dyer [25] and samples of the upper (aqueous) phase were dried and analysed for phospho~lcholine by descending chromatography on Whatman No. 1 paper using ethanol/water (17: 3, v/v> as solvent. SO nmol of standard phospho~lcholine (Sigma) was added to each sample before analysis. The areas containing phosphorylcholine were identi- fied by spraying the chromatogram with sulphosalicylic acid/FeCI, reagent [26]. The spots were excised, eluted into 2 ml of water and were counted after addition of 8 ml of PCS scintillation fluid LAmersham International). From phosphate determinations the pool size of phos-

pho~lchoIine in untreated cells was found to be about 10% of phosphatidylcholine.

In each of the above experiments phosphatidylcholine and sphingomyelin were separated from the lower (chloroform) phase and their radioactivity and phosphorus content determined as previously described [23].

Results

Incubation of cells with sphingomyelinase for as long as 30 h in the absence or presence of serum had no obvious effect on cell morpholo~ and had no apparent effect on cell growth as measured by total lipid phosphorus. Although ceils grew normally with serum added to the medium, removal of serum only caused a significant drop in cell growth 24 h after- wards. The number of dead cells as judged by uptake of Trypan Blue did not exceed 2% at any time.

Fig. 1 shows the loss of radioactivity from sphingomyelin in cells labehed to equilibrium with [3H]choline and then incubated with non-radioactive choline with and without addition of sphingomyelinase. These results demonstrated that 70% of total sphingomyclin (presumed to be on the surface or in rapid equilibrium with the celi surface) was rapidly degraded by sphingomyelinase (half-time 10 mitt) but the remaining 30% inside the cells was resistant to attack. This result was not due to the inactivation of the enzyme since addition of further aliquots of fresh enzyme at various times up to 30 h made no difference to the observations. Neither did the addition of larger amounts of enzyme alter the results. The relative proportion of the external pool was a little higher than was indicated in

oo--- Time th)

Fig. 1. Loss of radioactivity from [“H]choline-labelled sphingomyelin

in BHK cells incubated with or without sphingomyelinase. BHK cells

labelled to equilibrium with [“HIcholine were incubated in serum-free

medium with (m) or without (0) addition of sphingomyelinase as

described under Materials and Methods. Points shown are the means of duplicate determinations. This experiment was one of four which

gave similar results.

our previous report [23] presumably because the ex- tended incubation period used in the present work allowed more complete degradation. However, it is clear that the internal pool has little tendency to mix with the pool of sphingomyelin which was accessible to the enzyme, since based on a logarithmic plot of results shown, in Fig. 1, the internal pool breaks down with a half-time of about 80 h or at a rate corresponding to a degradation of about 0.6% + 0.1% per h (mean of four experiments), This is less than the rate of loss of label from sphingomyelin in untreated cells, which was about 1.2% f 0.4% per h (mean of four experiments). Long- term loss of surface sphingomyelin had no apparent effect on cell growth as measured by total lipid phosphorus.

The incorporation of [ “HIcholine into phosphatidylcholine and sphingomyelin over a 44 h period of incubation is shown in Fig. 2 for untreated BHK cells. As noted previously [23], no label appeared in lipid until 20-30 min after addition of the tracer. Thereafter, the label in phosphatidylcholine increased linearly for about 8 h and reached equilibrium after about 25-30 h. Incorporation into sphingomyelin was initially much slower and had not reached equilibrium even after 44 h.

The specific radioactivity of surface and internal pools of sphingomyelin was measured separately fol- lowing a brief treatment with sphingomyelinase (Fig. 3). At all time points this caused breakdown of about 65% of the sphingomyelin measured as inorganic phosphate. However, the proportion of labelled sphingomyelin which was degraded was very small at early time points (Fig. 3) presumably because the externally located sphingomyelin which was susceptible to the en-

Time (h)

Fig. 2. Incorporation of [‘HIcholine into lipids of BHK cells. The

specific radioactivity of phosphatidylcholine ( W) and sphingomyelin

(0) was measured at various times after addition of [3H]choline to

non-confluent BHK cells. Points represent the means of duplicate

determinations which differed by no more than 7% and are derived

from a single experiment which was one of four giving essentially

similar results.

0 2 4 6 8

Time (h)

Fig. 3. Early changes in specific activity of internal and external pools

of sphingomyelin in BHK cells treated with sphingomyelinase. Cells

were incubated with [“HIcholine and treated briefly with sphin-

gomyelinase as described under Methods. The specific radioactivity

of the internal pool of sphingomyelin (W) was taken as dpm in

sphingomyelin/sphingomyelin phosphorus in cells treated with sph-

ingomyelinase for 10 min. The specific activity of external sphingo-

myelin (0) was calculated as:

sphingomyelin dpm (control)

- sphingomyelin dpm (sphingomyehnase-treated)

sphingomyelin phosphorus (control)

- sphingomyelin phosphorus (treated)

The specific activity of phosphatidylcholine is also shown (A ). These

values are means of results from untreated and treated cells which

showed no significant difference from one another. This experiment

was one of four giving similar results.

zyme took longer to acquire radioactivity than the internal pool. Whereas labelling of the inaccessible pool commenced after about 30 min, as did the labelling of phosphatidylcholine, no label was seen in the pool accessible to the enzyme until after 2 h. There- after, the specific activity of the accessible pool increased until it equalled that of the internal pool after about 20 h (not shown). The brief sphingomyelinase treatment used here had no significant effect on labelling of phosphatidylcholine at any time point.

In cells which were labelled with [“Hlcholine for 90 min (at which time no label had reached the external pool of sphingomyelin) and then chased with unlabelled choline (Fig. 41, specific radioactivity of phosphatidylcholine continued to increase for the next 2 h and then slowly decreased with a half-time of about 30 h (Fig. 4a). In contrast, the specific radioactivity of the surface pool of sphingomyelin was very low initially but increased steadily throughout the period of the experiment even while the specific radioactivity of phosphatidylcholine was falling (Fig. 4b). The specific radioactivity of the internal (enzyme-resistant) pool also increased during the period when that of phosphatidylcholine was falling, but did eventually reach a maximum value which approached the specific activity

98

Time (h)

; s

0 0 2 4 6 8 10

Time (h)

Time (h)

Fig. 4. Pulse-chase kinetics of labelling of phosphatidylchohne, sphingomyelin and phosphorylcholine with [‘HIcholine. Cells were la-

belled with [‘HIcholine, chased with medium containing unlabelled

choline and then exposed to sphingomyelinase as described under

Materials and Methods. The results represent the time-course of

changes in specific radioactivity of: (a) phosphatidylcholine; (b) exter-

nal and internal sphingomyelin (measured as in Fig. 3); and (c)

phosphorylcholine. Closely similar results were seen in two further experiments. In (a) filled symbols represent experimental points

obtained in the presence of sphingomyelinase. Hollow symbols represent untreated controls. In Fig. 4b circles represent the internal

sphingomyelin pool and squares represent the external pool.

of phosphatidylcholine 8-10 h after addition of non-radioactive choline. The specific radioactivity of phosphorylcholine declined continuously from the initiation of the chase conditions (Fig. 4~).

Discussion

As has been shown for a number of other cell types, incorporation of [ ‘HIcholine into total cellular sphingomyelin in BHK cells lags 3-6 h behind incorporation into phosphatidylcholine (Fig. 2). This is one element of the evidence which has been used to argue that synthesis of sphingomyelin depends on prior synthesis of phosphatidylcholine. However, from the data presented here it is clearly simplistic to think of just a single pool of sphingomyelin in BHK cells, since in terms of accessibility to external sphingomyelinase (Fig. l>, on the basis of differential rates of labelling with [3H]choline (Fig. 3) and from the results of pulse-chase experiments (Fig. 4) sphingomyelin exists as two distinct pools accounting, respectively, for 70% and 30% of the total. Based on accessibility to sphingomyelinase the larger of these is assumed to be in the external face of the plasma membrane but may also include any endosomal membranes which are in rapid equilibrium with the cell surface. However, previous results with glutaraldehyde-fixed cells treated with sphingomyelinase suggest that this endosomal pool must be relatively small compared with that on the surface membranes [23]. These results with BHK cells differ from those obtained using human skin fibroblasts where 90% of total sphingomyelin is degradable by external

sphingomyelinase [27,28]. However, the human cells have a larger proportion of plasma membrane than BHK cells (50% of total cell phospholipids vs. 20% [23]) so that any internal sphingomyelin would be less apparent in the former case.

It is clear from Fig. 1 that since the internal pool of labelled sphingomyelin remains largely inaccessible to exogenous sphingomyelinase for a long period, it does not easily reach the surface and therefore cannot be the precursor of plasma membrane outer leaflet sphingomyelin. We know from previous work [23,24] that

cells pretreated with sphingomyelinase and then rein- cubated in sphingomyelinase-free medium are capable of replacing their surface sphingomyelin within 3 h, whereas our observations show that the internal pool could be arriving at the surface no faster than a half- time of 80 h. Perhaps surprisingly, protracted exposure to sphingomyelinase seemed to have no apparent effect on cell growth even though this treatment has been reported to cause a pronounced redistribution of cholesterol in human skin fibroblasts and a marked inhibition of Nat/K+-ATPase [24].

The kinetics of sphingomyelin labelling with [‘HIcholine (Fig. 3) shows that radioactivity appears in

99

the internal pool much more rapidly than it enters the external pool. Thus, 2 h after addition of labelled

choline, when the specific radioactivity of the internal pool has reached about a third of the level of the phosphatidylcholine, essentially no radioactivity has reached the external pool. This is inconsistent with a mechanism whereby the internal pool is derived from the external pool, (for example by endocytosis of plasma membrane and deposition of membrane lipid in lyso- somes). Thus, the two pools are not only topologically

separate but metabolically distinct. The precise cellular localisation of the internal pool

is unknown; it seems unlikely to be present on the cytosolic face of any organelle because sphingomyelin breakdown in cells exposed to sphingomyelinase is no greater when the cells are permeabilised [23]. Neither is the sphingomyelin of isolated microsomes available to this enzyme [29,30].

Furthermore, membrane sphingomyelin of en- veloped virus grown in these cells is found exclusively on the outer lipid leaflet [31]. Possibly, in view of its apparent failure to mix rapidly with the surface pool, the internal pool could be present in a part of the cell which rarely shares material with the cell surface, e.g., the nuclear membrane [32]. Such an explanation might be consistent with the observation that the internal

pool labels relatively quickly (Figs. 3, 4) since the nuclear membrane is proximal to the endoplasmic reticulum which is the major site of phopholipid biosynthesis. However, nuclear membranes generally appear to contain no more sphingomyelin than endoplasmic reticulum or mitochondrial outer membrane [33] so that it seems more likely that the endoplasmic reticulum/ Golgi complex, which accounts for about 40% of total BHK cell phospholipid, is the repository of most of the internal pool of sphingomyelin. This suggests that the endoplasmic reticulum/ Golgi would have to contain about 5% sphingomyelin as a proportion of its total phospholipids. It is significant in this connection that a detailed subcellular fractionation study of BHK cells has shown that sphingomyelin accounts for as much as 10% of endoplasmic reticulum phospholipids [34]; even allowing for some contamina- tion with fragments of plasma membrane, this implies that a substantial proportion of the total cellular sphingomyelin must reside in the endoplasmic reticulum. A recent report concluded that a significant portion of the cellular sphingomyelin in glioma cells was internal 1161 and supports our finding (Fig. 3) that the internal pool of sphingomyelin labelled initially much faster than the plasma membrane pool. There has recently been some dispute regarding the proportion of total cellular sphingomyelin and cholesterol which is located at the plasma membrane [28,36]; it is possible that different cell types vary in their content of internally located sphingomyelin.

The results of pulse-chase experiments (Fig. 4) ac-

cord with previous work [7,10-121 and suggest that

surface and internal sphingomyelin are synthesised from pre-existing phosphatidylcholine. The specific activity of both pools of sphingomyelin continues to rise

for many hours after the specific activity of phosphatidylcholine has begun to fall, although the internal pool reaches its maximum specific activity before the external pool. As we showed above that the internal sphingomyelin is unlikely to be be the precursor of the

surface pool, it is reasonable to suppose that labelled sphingomyelin appears at the surface at the expense of a fraction of the internal labelled phosphatidylcholine which is translocated to a site (perhaps the Golgi), where it can donate its phosphocholine headgroup to ceramide. The lag period before choline is incorporated into the surface pool of sphingomyelin (Fig. 3) would then be related to the distance in both topologi- cal and metabolic terms between the outer face of the plasma membrane and the cytosolic face of the endoplasmic reticulum which is the primary site of choline incorporation into phosphatidylcholine. The results shown in Fig. 3 suggest that there is a delay of about 90 min between the onset of choline incorporation into endoplasmic reticulum phosphatidylcholine and the first appearance of label in plasma membrane sphingomyelin. Previous estimates of the time taken to transport phospholipid from the endoplasmic reticulum to the plasma membrane have varied from a few minutes [37,38,39] to hours [40,41]. The period of about 90 min derived in the current work is consistent with the longer of these estimates. This delay could reflect the time required for transport of the phosphorylcholine moiety from phosphatidylcholine in the endoplasmic reticulum to the Golgi, for transfer of phosphoryl-

choline to ceramide and for vesicular transport of sphingomyelin to the surface. The phosphocholine- transferase activity itself is not limiting, since previous

work has shown that after degradation of surface sphingomyelin to ceramide and subsequent removal of the sphingomyelinase, resynthesis of sphingomyelin pro- ceeds immediately and is complete within 2-3 h [23,24]. This is a far higher rate of sphingomyelin synthesis (about 40% per h) than is normally seen in growing cells (8% per h).

The pattern of labelling of the internal pool of sphingomyelin was intermediate between that of the external pool and that of phosphatidylcholine. In incorporation experiments, labelling of the internal pool proceeded linearly after a lag period, of 30 min (like phosphatidylcholine) although the specific radioactivity increased at only about half the rate of phosphatidylcholine (Fig. 3), perhaps because it was synthesised from a pool of phosphatidylcholine which was of relatively low specific activity. Under pulse-chase conditions (Fig. 4) the internal sphingomyelin pool reached

100

its maximum specific activity before the external pool but after phosphatidylcholine. Thus, the internal pool may be physically and metabolically closer to its precursor pool of phosphatidylcholine than is the surface sphingomyelin.

The presence of two distinct pools of sphingomyelin in BHK cells suggests that there may be two separate sites of synthesis of this lipid. It is not clear if other cell types also possess a pool of internal sphingomyelin, although preliminary results with rat liver cells show

that about 25% of sphingomyelin is inaccessible to sphingomyelinase (unpublished results). Our findings do raise the possibility that in at least some of the

recent reports [14-211 regarding the localisation of sphingomyelin biosynthesis the authors may have been measuring synthesis of internal sphingomyelin and not necessarily the pool which eventually reaches the cell surface. This point is particularly relevant when cells are radioactively labelled for periods of less than 2 h which in our system would only label internal sphingomyelin. However, our results do not exclude the possibility of a single site for sphingomyelin synthesis (e.g., the cis Golgi), but in this case it would mean that 30% of total sphingomyelin synthesised would be taking a retrograde path back to the endoplasmic reticulum instead of going to the plasma membrane.

Acknowledgements

We would like to acknowledge the interesting and helpful discussions with Karl-Josef Kallen which have contributed to the preparation of this manuscript. This work was funded by The Wellcome Trust.

References

I Zwaal, R.F.A., Roelofsen, B.. Comfurius. P. and Van Deenen,

L.L.M. (1975) Biochim. Biophys. Acta 406, 83-96.

2 Op den Kamp, J.A.F. (1979) Annu. Rev. Biochem. 48, 47-71.

3 Dawidowicz, E.A. (1987) Annu. Rev. Biochem. 56, 43-61.

4 Van Meer, G. (1989) Annu. Rev. Cell Biol. 5, 247-275.

S Marggraf, W.D., Diringer, H., Koch, M.D. and Anderer, F.A.

(1972) Hoppe Seyler’s Z. Physiol. Chem. 353, 1761-176X.

6 Ullman, M.D. and Radin, N.S. (1974) J. Biol. Chem. 249, 1506-

1512.

7 Diringer, H.. Margraf, W.D., Koch, M.A. and Anderer, F.A.

(1972) Biochem. Biophys. Res. Commun. 47, 1345-1352.

8 Bernert, J.T. Jr and Ullman, M.D. (1981) Biochim. Biophys. Acta

666, 99-109. 9 Voelker. D.R. and Kennedy, E.P. (1982) Biochemistry 21, 2753-

275’).

10 Eppler, CM., Malewicz, B., Jenkin, H.M. and Baumann,

(1987) Lipids 22, 351-357.

11 Maagraf, W.D. and Kanfer, J.N. (1984) Biochim. Biophys.

793, 3466353.

I2 Gallaher. W.R. and Blough. H.A. (1975) Arch. Biochem.

phys. 168, 104- 114.

W.J.

Acta

Bio-

13 Schwarzmann, G. and Sandhoff, K. (1990) Biochemistry 2Y.

l0865-10871.

14 Lipsky, N.G. and Pagano, R.E. (lY85) J. Cell Biol. 100, 27-34.

15 Pagano, R.E. (1988) TIBS 13. 2022205.

16 Cook, H.W., Palmer, F.B.St.C., Byers, D.M. and Spence. M.W.

(1988) Anal. Biochem. 174. 552-560.

17 Kobayashi, T. and Pagano. R.E. (19x0) J. Biol. Chem. 264.

5966-5973.

18 Van Meer. G.. Stelzer, E.H.K, Wijnaedents-van-Resandt, R.W.

and Simons, K. (1987) J. Cell Biol. 1051623-1635.

19 Futerman. A.H., Stieger, B., Hubbard, A.L. and Pagano, R.E.

(1990) J. Biol. Chem. 265. X650-8657.

20 Jackel, D., Karrenbauer, A., Birk, R., Schmidt, R.R. and Wieland,

F. (1990) FEBS Lett. 261. 155157.

21 Van’t Hoff. W. and Van Meer. G. (1990) J. Cell Biol. I II,

977-986.

22 Skipski. V.P. tlY72) in Blood Lipids and Lipoproteins (Nelson,

G.J., ed.). pp. 471-583, Wiley-Interscience, New York.

23 Allan, D. and Quinn, P. (1988) Biochem. J. 254, 765-771.

24 Slotte. J.P.. Harmala, A.-S., Jansson, C. and Porn, M.I. t IYYO)

Biochim. Biophys. Acta 1030, 251-257.

25 Bligh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37,

911-917.

26 Handbook of Chromatography t 1972) (Zweig, G. and Sherma. J..

eds.). p. 143.

27 Slotte, J.P. and Bierman, E.L. (1988) Biochem. J. 250, 65336%.

2X Lange, Y., Swaisgood, M.H., Ramos, B.V. and Steck, T.L. (19X0)

J. Biol. Chem. 264. 3786-3793.

29 Nilsson, O.S. and Dallner. G. (1977) J. Cell Biol. 72, 568~583.

30 Nilsson, OS. and Dallner. G. (1977) Biochim. Biophys. Acta 464.

453-45x.

31 Allan, D. and Quinn, P. (I9891 Biochim. Biophya. Acta YX7.

190-204.

32 Allan, D. and Raval, P.J. (1987) Biochim. Biophys. Acta 897.

3555363.

33 White. D.A. (1073) in Form and Function of Phospholipida

(Ansell. G.B., Hawthorne, J.N. and Dawson. R.M.C.. eds.). pp.

341-482, Elsevier, Amsterdam.

34 Renkonen. O., Gahmberg, C.G.. Simons, K. and Kaartamen. L.

(1972) Biochim. Biophys. Acta 255, 66-7X.

35 Reference deleted.

36 Van Meer. G. (1987) Trends Biochem. Sci. 12, 375376.

37 Kaplan, M.R. and Simoni, R.D. (1985) J. Cell Biol. 101, 446-453.

38 Kaplan. M.R. and Simoni, R.D. (1985) J. Cell Biol. 101, 441-445.

39 Pagano, R.E. and Sleight, R.G. (1985) Science 229, 1051-1057.

40 Thompson, G.A. Jr. and Nozawa, Y. (1977) Biochim. Biophys.

Acta 472, 55-92.

41 Lange, Y. and Mathies, H.J.G. (1984) J. Biol. Chem. 259. 14624-

14630.

two separate pools of sphingomyelin in bhk cells

Documents