muir, K. J., and Martin, 0. C. (1983) J. ... plus ceramide to DG and SM on the basolateral MDCK .... PA had been converted to C6-NBD-diacylglycero1 (DG), c6-.
Vol. 269, No. 3,h u e of January 21,pp. 1763-1769,1994 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1 W 4 by The American Society for Biochemistry and Molecular Biology, Inc.
Conversion of Diacylglycerol to Phosphatidylcholine on the Basolateral Surface of Epithelial (Madin-Darby Canine Kidney)Cells EVIDENCE FOR THE REVERSE ACTION OF A SPHINGOMYELIN SYNTHASE* (Received for publication, June 2,
1993, and in revised form, August 9,
1993)
Ardy van Helvoort, Wouter van’t Hof$, Tita Ritsemaj, Alex Sandra(, and Gerrit van Meer 11 From the Department of Cell Bwbgy, Medical School, AZU H02.314, University of Utrecht, Heidelberglaan 100,3584 CX Utrecht, The Netherlands and TDepartment of Anatomy, College of Medicine, University of Iowa, Iowa City, Iowa 52242
phosphocholine transferase at the cell surface may When N-6[7-nitro-2,1,3-benzoxadiazol-4-yl]aminohexanoyl-phosphatidic acid (Cs-NBD-PA) is in- have a regulatory role in signal transduction. serted into the plasma membrane of fibroblasts, it is metabolized by the cells to Cs-NBD-diacylglycerol (DG), -triacylglycerol, -phosphatidylcholine (PC), and Phosphatidylcholine (PC)’ is the major phospholipid in -phosphatidylethanolamine (PE) (Pagano,R. E.,Longmuir, K. J., and Martin, 0. C. (1983) J. Biol. Chern. mammalian cells, and generally makes up for more than 50% 268, 2034-2040). In Madin-Darby caninekidney of the cellular phospholipids and more than 30% of the total (MDCK) cells incubated at 10 O C with Cs-NBD-PA, up cellular lipids. The distribution of PC over the various cellular to 70% of the newly synthesized Ce-NBD-PC but no membranes is heterogeneous. The PC content in the plasma Ce-NBD-PE couldbe depleted from the basolateral cell membrane, endosomes and lysosomes is reduced compared to surface by the additionof bovine serum albumin to the the endoplasmic reticulum. A striking example of this hetermedium. Preincubation of the cells with [SH]cholinefor ogeneous distribution is presented by the plasma membrane 2 h at 37 O C prior to Ce-NBD-PA addition at 10 OC of epithelial cells of kidney and intestine, where the apical labeled non-depletable Ca-NBD-PC with a specific ac- domain, lining the body cavities, has a PC content that isfour tivity of >10 times that of the depletable Ce-NBD-PC times lower than that of the basolateral domain, which faces on the basolateral cell surface, indicating that thelat- the neighboring cells and the underlying tissue (1). This ter had not been synthesized by the CDP-choline path- difference in lipid composition may be caused by local synway. Ce-NBD-DG could substitute for Ca-NBD-PA as thesis and degradation. Alternatively, it has been proposed substrate for both intracellular and surfaceCs-NBD- that the low PC content of the apical membrane is due to PC synthesis. In addition, Cs-NBD-PC synthesis on the selectivity in intracellular lipid transport, or, more specificell surface wasindependent of the location of the c e - cally, is a consequence of an enrichment of glycosphingolipids NBD-chain on the 1- or 2-position, indicating that the into theapical transport pathway in the trans-Golgi network reaction occurred by transfer of phosphorylcholine. Using Ce-NBD-ceramide, Ca-NBD-sphingomyelin (SM)(1). To study this question more directly, it would be imporsynthesis also was discovered on the basolateral but tant to determine the site of PC synthesis in epithelial cells not on the apical cell surface. The conversion of PC and the mechanism of its transport to thetwo plasma memplus ceramide toDG and SM on the basolateralMDCK brane domains. In mammalian cells, PC can be synthesized by some five cell surface suggests that thesynthesis of Ce-NBD-PC different pathways (Fig. l), of which the CDP-choline or on this surface occurred via the reverse reaction of a SM synthase. Indeed, the surface Cs-NBD-PC synthe- Kennedy pathway is by far the most significant (2,3). Minor sis was reduced to 40-60% by addition of Ce-NBD- pathways, in most cells, are themethylation of phosphatidylceramide or hydrolysis of cell surface SM by exogenous ethanolamine (PE), base-exchange from phosphatidylserine neutral sphingomyelinase. Since DG activates protein (PS), andacylation of Lyso-PC. Finally, it has been proposed kinase C and ceramide indirectly inhibits this kinase that PCcan also be synthesized by the reverse action of but activates other kinase(s) and phosphatase(s), the sphingomyelin (SM) synthase (4-6). Whereas the former reactions have been assigned to thecytosolic leaflet of mostly * The present work was supported by (senior) fellowships from the the endoplasmic reticulum, SM synthase has been localized Fogarty International Fellowship Program (F06TW 01534)(to A. S.), to theGolgi and, to some extent, to theplasma membrane (7) from the Royal Netherlands Academy of Arts and Sciences (to (see below). G. v. M.), and from the Netherlands Foundation for Chemical ReNewly synthesized PC is thought to reach the plasma search (SON 330-026/D’92/J3’92)with financial aid from the Netherlands Organization for Scientific Research (NWO) (to A.v. H.). membrane mainly by monomeric exchange through the cytoThe costs of publication of this article were defrayed in part by the plasm (8), possibly mediated by a PC transfer protein (9). In payment of page charges. This articlemust therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 Present address: Department of Cell Biologyand Anatomy, Corne11 University Medical College, New York, NY 10021. $ Present address: Institute of Molecular Plant Sciences, Leiden University, Leiden, The Netherlands. 11 To whom correspondence should be addressed. Tel.: 31-30506480;Fax: 31-30- 541797.
The abbreviations used are: PC, phosphatidylcholine; Ce-NBD-, N-6[7-nitro-2,1,3-benzoxadiazol-4-yl]aminohexanoyl-; DG, diacylglycerol; HBSS, Hanks’ balanced salt solution without bicarbonate, 10 mM Hepes, pH 7.35;HBSS + BSA, HBSS containing 1% (w/v) bovine serum albumin; MDCK, Madin-Darby canine kidney; PA, phosphatidic acid; PE, phosphatidylethanolamine; PS, phosphatidylserine; SM, sphingomyelin; BSA, bovine serum albumin; HPLC, high performance liquid chromatography.
1763
PC and SMthe Synthesis at
1764
addition, from the endoplasmic reticulum PC is carried to the plasma membrane by vesicle flow (7). Exchange reactions only introduce PC into thecytoplasmic leaflet of the plasma membrane, and trans-bilayer translocationof PC to theexoplasmic leaflet is thought to be slow (10). Because the difference in PC contentbetween the apical and basolateral plasma membrane domains of epithelial cells resides in the exoplasmic leaflets, an important question remains by what mechanism newly synthesized PC reaches the exoplasmic leaflet of the plasma membrane. This question has been successfully studied for sphingolipids by the use of short chain lipid analogs (1, 7, 11).Similar studies on PC seem feasible from the description of a method to induce cellular synthesis of a short chain PC (12-14). When Pagan0 and co-workers (12-14) added liposomes containing fluorescent, shortchain C6-NBD-phosphatidic acid (PA) to fibroblasts at 2 "C, after 1 h most of the NBD fluorescence was localized in the nuclear membrane, the endoplasmic reticulum, and mitochondria (12). Lipid analysis revealed that, both at 2 "C and 37 "C,a large part of C6-NBDPA had been converted to C6-NBD-diacylglycero1(DG), c6NBD-triacylglycerol and C6-NBD-PC, with C6-NBD-triacylglycerol accumulating in intracellular lipid droplets (13). Insertion of a nonhydrolyzable phosphonate analog of C6-NBDPA resulted in selective staining of the plasma membrane, while treatment of cells with fluorescent PC followed by phospholipase C at 2 "C resulted in extensive labeling of intracellular membranes (14). From these results it was concluded that C6-NBD-PAwas hydrolyzed at the surface to c6NBD-DG, which subsequently translocated intocells followed by intracellular conversion to C~-NBD-triacylglyceroland CgNBD-PC. Because C6-NBD-PC can be quantitativelydepleted from the cell surface by a back-exchange against liposomes or BSA and because C6-NBD-lipids can be depleted from the apical and basolateral cell surface independently (11, 15, 16), the cellular synthesis of Cs-NBD-PC could prove useful in studying PC transport to theepithelial cell surface. So far, the intracellular disposition of the newly synthesized Ca-NBD-PC has not been directly tested. In the present study we describe that in MDCK cells the newly synthesized Cs-NBD-PC consists of two pools, one on the basolateral cell surface, the other intracellular. At 10 "C, up to 70% is depletable from the cell surface by BSA. Only Choline
(7)h
(-CH,)
kl)
Cell Surface
intracellular C6-NBD-PC is synthesized by the CDP-choline pathway as demonstrated by double-labeling with [3H]choline. Evidence for the activity of a phosphocholine transferase that interconverts PC and SM at the basolateral surface is presented. EXPERIMENTALPROCEDURES
Materials-BSA
ps
LFC PE monomethyl
serine
C3-P (2)
CDP-choline + DG
I
-b
(4)
PC
(7) 9 dimethyl PE
(-CH, )
DG
SM FIG. 1. Pathways of PC biosynthesis in mammalian cells. The numbers represent the following enzymes: 1 choline kinase; 2, CTP:phosphocholine cytidylyltransferase; 3, PA phosphohydrolase/ DG kinase; 4, DG cholinephosphotransferase;5, PC:serine O-phosphatidyltransferase; 6, Lyso-PC acyltransferase/phospholipase A; 7, PE N-methyltransferase; 8, SM synthase.
PC, phospholipase C (Bacillus
from Sigma. Phospholipase D from cabbagewas from Boehringer Mannheim, Germany. Sphingomyelinase C (Staphylococcus aureus) was a kind gift of Dr. Ben Roelofsen, Utrecht. Ca-NBD-PC and -PE were from Avanti Polar Lipids Inc., Pelham, AL. C6-NBD-ceramide was obtained from Molecular Probes, Eugene, OR, and [ m e t h ~ l - ~ H ] choline chloride from DuPont NEN. Cells-Confluent monolayers of Madin-Darby canine kidney (MDCK) strain I1 cells were grownon 0.4-pm pore size, 4.7-cmZfilters glued to the base of a plastic ring so as tobe able to separate apical and basal medium (Transwells; Costar, Cambridge, MA), all as described before, and contained 4.7 X lo6cells/filter (11,16). Two filters (-10' cells) were pooled for each lipid analysis. Preparation of c6-NBD-Lipid Precursors-C6-NBD-DG was prepared from Ce-NBD-PC using phospholipase C. For this, 250 nmol of Ce-NBD-PC were dissolved per 1ml of Hanks' balanced salt solution without bicarbonate, containing 10 mM Hepes, pH 7.35 (HBSS), 62 pg/ml taurocholate, and 10 units/ml phospholipase C and incubated for 2-4 h a t 37 "C. Ce-NBD-PAwas prepared from C6-NBD-PCusing phospholipase D. 2.3 pmol of CB-NBD-PCwere suspended into 2 ml of 100 mM CaCIZ,100 mM sodium acetate, pH 6, containing 3 units/ ml phospholipase D, to which was added 0.5 ml of diethyl ether. The mixture was incubated for 0.5 h at room temperature under continuous stirring. Products were purified by preparative TLC in the acidic direction (16), located by UV detection, and identified by comparison to authentic standards. After scraping, the products were extracted from the silica and quantitated by measurement of the NBD fluorescence (16). Assay for the site of Synthesis of Ce-NBD-PC and of C6-NBDSM-Cs-NBD-DG or C,-NBD-PA in Cs-NBD-PC experiments and CB-NBD-ceramidein C6-NBD-sphingomyelin(SM) experiments were added to cells at 10 "C by exchange from BSA. NBD-lipid complexes (16). For this, 1 ml of NBD-lipid (routinely 5 PM) in HBSS BSA (1%w/v) was added to the apical and 2 ml to the basal side of the filter. The presence of NBD-lipid products at the apical surface was assayed by continuous depletion from the surface by the BSA present in the apical medium and of NBD-lipids on the basolateral cell surface by the BSA in the basal medium (11).Each incubation was followed by a wash in HBSS BSA for 0.5 h a t 10 "C to complete the depletion, which is not included in the incubation times. Although originally no BSA (14) or low concentrations ofBSA (0.03% w/v) (16) were used to insert C6-NBD-PA or Ce-NBD-ceramide into the plasma membrane, BSA a t a concentration of 1%did not reduce the PE efficiency of incorporation (not shown). Still, at 1%,or 145 nmol/ml, the molar excess of BSA over the lipid products to be depleted was far higher than the450-foldthat was found to be required for complete depletion of C6-NBD-PC in a model membrane study (17). Fluorescent lipids were extracted from the pooled apical media, the pooled basal media, and from the cells on the filter. They were quantitatively analyzed by two-dimensional TLC, and NBD fluorescence was measured in the individual lipids as before (16).Average values are followed by +S.D.(number of measurements). Incorporation of fH]Choline into C6-NBD-PC"Cell monolayers were preincubated in HBSSfor 1h a t 37 "C to reduce cellular choline pools and labeled for 2 h at 37 "C with 100 kBq/ml [3H]choline in HBSS before incubation with C6-NBD-PA for 3 h at 10 'C as described above. After lipid extraction from media and cells and TLC analysis, the newly synthesized Cs-NBD-PC was further purified by HPLC to fully separate it from native PC with which Cs-NBD-PC partially overlaps on theTLC plate. Where indicated, the lipid extracts from apical media and basal media were pooled before HPLC because of the small signals in the apical medium. NBD-fluorescence and 3H radioactivity were measured in the HPLC eluate. Reuersed-phase HPLC-Cs-NBD-lipids were separated by reversed-phase HPLC on a 4 x 250-mm Spherisorb ODs-2 (5 pm) column (Pharmacia LKB Biotechnology Inc.) using CH3OH/HzO/ H,PO, (850/250/1 v/v), buffered to pH 7 with triethylamine, as the
choli;j(7k(.cH31
1
P-choline
fraction V,egg
cereus), CDP-choline, and phosphorylcholine chloride were purchased
+
+
PC and SM Synthesis at the Cell Surface mobile phase. Samples were applied in 0.1 ml, eluted at a flow rate of 1.0 ml/min, and NBD fluorescence in the eluate was measuredin 0.5= 535 nm, using an SFM ml samples at A+= = 470 nm and fluorimeter (Kontron Instruments, Zurich, Switzerland). Alternatively, the fluorescence was measured on line with a Jasco 821-FP fluorescencedetector (Japan Spectroscopic Co., Hachioji City, Japan). This procedure was used to separate the [1-C6-NBD-,2-palmiof C6-NBD-PA and Cetoyll-, and [ l-palmitoyl,2-C~-NBD]-isomers NBD-PC (18).In addition, it was used in the double-labelexperiments with ['Hlcholine. In contrast to Ce-NBD-PC, native PC was retained on the column. After each run the column was cleaned by running a gradient of 0-100% CHC13 in CHaOH. Radioactivity was determined by liquid scintillation counting (16). Background values per 0.5-ml fraction after 2- and 20-h prelabeling with [3H]choline were 100 and 400 dpm for the cells and 40 and 160 dpm for the media, respectively. Hydrolysis of SM by Exogenous Sphingomyelinase-Sphingomyelinase (2.5 units/ml of phosphate-bufferedsaline) was added to both surfaces of a cell monolayer on a filter for 1 h at 37 "C, while another filter was mock-treated. SM hydrolysis was assayed in a parallel set of filters. For this, the phospholipids were extracted and separated by two-dimensional TLC, after which the loss of SM phosphate in the sample that had been treated with sphingomyelinase was calculated by comparingthe phosphate content of the individual phospholipid spots to that in mock-treated cells (11). Higher sphingomyelinase concentrations did not enhance SM hydrolysis. RESULTS AND DISCUSSION
Presence of Newly Synthesized c6-NBD-PC on the Cell ,S'urfuce"In a first series of experiments, C,-NBD-PA was added to MDCK cells for 3 h at 10 "C as described under "Experimental Procedures." This resulted in the synthesis of 23 f 7 pmol of Cs-NBD-PC/107 cells ( n = 13). Intracellular Ca-NBD-PC can be discriminated from Ca-NBD-PC present on the cell surface by the fact that the latter is available for depletion by liposomes or BSA in the medium (11,15,16). In the above experiments, 70 f 11%of newly synthesized c6NBD-PC could be depleted from the cell surface (n = 13). It has been demonstrated that externally added Cs-NBD-PA is rapidly degraded in the plasma membrane to Cs-NBD-DG, which subsequently translocates into cells (14). When the 10 "c experiment was repeated with externally added cg' NBD-DG, this resulted in synthesis of 4.3 f 0.3 pmol of cgNBD-PC/lO' cells, of which 70 f 2% was present at thecell surface (n = 2). Because synthesis of Cs-NBD-PC from c6NBD-DG was about 6 times lower than that from C6-NBDPA ( n = 6), most likely due to the reduced exchange rate of Cs-NBD-DG with the plasma membrane (14), C6-NBD-PA was used in subsequent experiments to increase the signal. The presence of Cs-NBD-PC at theouter cell surface after synthesis at 10 "C could be the consequence of either of two events: ( a ) transport from some intracellular siteof synthesis to the cell surface or ( b ) synthesis in the exoplasmic leaflet of the plasma membrane itself. To discriminate between these possibilities, we first performed prolonged incubations at 10 'c. In the case of intracellular synthesis of this Cs-NBDPC and subsequent translocation to the cell surface, all cg' NBD-PC should eventually become accessible to depletion by BSA. However, no significant increase in the depletable fraction of Cs-NBD-PC was observed over a period of 4 h (not shown). This implies that Cs-NBD-PC at thecell surface had been synthesized there and had not reached this location by transport, consistent with the fact that, of the two possible pathways, (i) vesicular transport is blocked at 10 "C (19) and (ii) translocation of PC across the plasma membrane (flipflop) isslow with a tlIP of hours even at 37 "C (10, 15). Incorporation of fH]Choline into Newly Synthesized C6NBD-PC-TO further investigate the biosynthetic relationship between intracellular and cell surface (2,-NBD-PC, the 10 "cexperiments were combined with [3H]choline incuba-
1765
tions. We argued that since the major pathway by which choline is incorporated into PC, theCDP-choline pathway, is cytosolic (Z), the incorporation of [3H]cholinemight reveal a difference in metabolic origin of the two C,-NBD-PC pools. When [3H]cholinewas added simultaneously with Cs-NBDPA, no 3H incorporation into C6-NBD-PC was detected after 3 h at 10 "C. For this reason, cellular choline levels were first reduced by a 1-h incubation at 37 "C in choline-free buffer (HBSS), and the cellular choline pools werelabeled with [3H] choline for 2 hat 37 "C before addition of Cs-NBD-PA. Under these conditions, [3H]cholinewas incorporated into Cs-NBDPC during the subsequent incubation for 3 hat 10 "C. Fig.2A shows a representative experiment. Analysis of Cs-NBD-PC after TLC by reversed-phase HPLC resulted in one major peak of fluorescence at 25 ml, containing about 95% of the total signal, and a minor peak at 22.5ml. Since the peaks coeluted with the major and minor peak of a Cs-NBD-PC standard, they were taken to represent the sn-2 and sn-1 isoforms of Cs-NBD-PC, respectively (18). Measurement of 3H label in intracellular Cs-NBD-PC (Fig. 2 A ) showed two peaks of radioactivity that coeluted with the fluorescence. However, the radioactivity at 22.5 ml was due to some nonNBD-lipid contamination. It was present in an elution pattern (Fig. 2C) of a mixture of the lipid extracts of two cell populations, one of which had been incubated with Cs-NBD-PA and the other with [3H]choline.The mixture can containonly non-radioactive Cs-NBD-PC and non-NBD radioactive products. In a subsequent experiment we observed that native Lyso-PC eluted at position 22.5. Since only the Cs-NBD-PC spot of the TLC plate had been applied to the HPLC, and since this spot partially overlaps with native PC but not LysoPC, the 3H-labeled Lyso-PC must have been formed during extraction of the [3H]PC from the silica. Indeed, the radioactivity measured in the peak at 22.5ml consistently was about 0.5% of the signal for [3H]PCthat was measured in the subsequent CHC13 wash of the column. Also, the signal is absent from the medium samples (Fig. 2, B and D ) which did not contain [3H]PC, as native PC is not extracted into the medium byBSA. In conclusion, the radiolabel at a 25-ml elution volume was present as C6-NBD-PC, more specifically as [l-palmitoyl,2-C6-NBD]-PC,while the label at 22.5 ml was due to Lyso-PC and most likely also to thesn-1 isomer of csNBD-PC. The Cs-NBD-PC at the cell surface, which was recovered from the BSA-containing medium, contained little radioactivity, showing that thesurface pool is not synthesized by direct transfer of [3H]choline itself. It contained less than 20% of the total 3H-labeled Cs-NBD-PC radioactivity (cf. Fig. 2, B with A ) , while in five experiments the relative amount of csNBD-PC fluorescence on the cell surface was 54 f 10% (n = 9). The specific activity of the surface Cs-NBD-PC (dpm/ pmol NBD) was 0.20 f 0.03 times that of the intracellular pool ( n = 7). This difference in specific activities shows that the two pools are of different metabolic origin. The choline donor of the surface Ca-NBD-PC synthesis had not equilibrated with the intracellular CDP-[3H]choline, the precursor for the synthesis of the intracellular Cs-NBD-PC. To demonstrate that[3H]cholinecould be introduced into cell surface Cs-NBD-PC, we performed prolonged prelabeling with [3H] choline for 20 h at 37 "C, preceding the 10 "C experiment. Under these conditions, the surface pool of C,-NBD-PC did become radiolabeled (Fig. 3, A and B), which confirms that indeed Cs-NBD-PC on the surface couldbe metabolically labeled. Complete equilibrium labeling was not reached in all experiments.
PC and SM Synthesis at the Cell Surface
1766
n:
Cellular
Medium
5.0
2.5
E g o z
5
Q
-0
v
t
m
L!
5'0 2.5
Elution 3 0 volume 20 (mi)
020 25
30
25
FIG. 2. Specific incorporation of [aH]cholineinto intracellularCe-NBD-PC. lo7 MDCK cells were prelabeled for 2 h at 37 "C with ['Hlcholine in HBSS, followed by incubation with Cs-NBD-PA for 3 h at 10 "C as described under "Experimental Procedures." NBD fluorescence (0) and 'H radioactivity (0)were measured in the HPLC eluate of TLC purified Cs-NBD-PC. Intracellular lipids (Panel A ) were discriminated from cell surface lipids by depletion of the latter into the HBSS + BSA of the 10 "C incubation (Panel B ) . Apical and basal HBSS + BSA were pooled. ['Hlcholine radioactivity is recovered in the intracellular C6-NBD-PC(Panel A ) , but not in the cell surface Ce-NBD-PC that is recovered in the BSA-medium (Panel B ) . In the experiment of Panels C and D, ['Hlcholine (2 h 37 "C) and Cs-NBD-PA (3 h 10 "C) were added to separate filters (lo7 cells for each label). Subsequently, the radioactive lipid extracts of cells and medium were pooled with the fluorescent extracts, and TLC and HPLCwere performed. No radioactivity is recovered at theCs-NBD-PC position in the pooled extracts (Panels C and D ) .
FIG. 3. Incorporation of['Hlcho-
line into intracellular and cell surface Ce-NBD-PC after prolonged labeling times. MDCK cells were preincubated for 20 h at 37 "C with 100 kBq/ml PHIcholine in minimum essential mehum containing 20 mM Hepes and 0.25% fetal calf serum, before the addition of Cs-NBD-PA for 3 h at 10 "C. Panels and symbols are as in Fig. 2. In Panel A , radioactivity at positions 20-22 ml is due to 'H-labeled Lyso-PC (see text).
-
-
P
Cellular 4
Ei
I _
Medium E
T
P
B
'0 4 0 1
v
0
m =
X
2
20
Polarity of Ce-NBD-PC on the MDCK Cell Surface-Next, we addressed the question of whether the synthesis of c6NBD-PC would occur preferentially on the apical or the basolateral surface of the epithelial MDCKcells. Earlier studies have demonstrated that C6-NBD-lipids on the basolateral cell surface are not accessible to BSA in the apical medium and vice versa (ll),implying that apical and basolateral C6-NBD-lipids can be depleted independently. When the apical andbasal media were analyzed separately, the apical/basolateral polarity of the Cs-NBD-PC on the cell surface was 0.07k 0.03 ( n = 34);Le. 14 times more Cs-NBDPC was synthesized on the basolateral than on the apical cell surface domain. Separate analysis of the specific activities of the apical and the basolateral surface Cs-NBD-PC in two double-label experiments (Table I) demonstrated that the specific activity of the Ca-NBD-PCon the basolateral surface was 10 times lower than the specific activity of intracellular Ce-NBD-PC. The higher specific activity of the (minor amount of) Cs-NBD-PC on the apical surface, which approaches that of the intracellular pool, suggests that it is
25
25
30 2 0
21 I
3d
Elution volume (ml)
TABLE I Incorporation of r H l c h ~ l i n einto C6-NBD-Pcsynthesized intracellularly or on either cell surface Data represent the mean of two experiments carried out in duplicate following the protocol of Fig. 2 and areexpressed per lo7cells (n = 4). The fraction of total Ca-NBD-PC on the apical and basolateral cell surface was 4. f 1% and 27 f 2%, respectively. The surface polarity of Ca-NBD-PC, i.e. the ratio of apical Cs-NBD-PC over basolateral Ce-NBD-PC, was 0.16 f 0.03. Sample
C6-NBD-PC [aH]Ch&ne Pmol
dprn
activity
%Iative specific activity
dprnfpmol
Cellular 3.89 k 0.25 5195 f 413 1337 f 78 974 k 64 0.23 f 0.30 222 k 26 Apical 124 k 27 Basolateral 1.45 f 0.12 181 f 43
1 0.73 0.09
derived from the intracellular C6-NBD-PC. Mechanism of C6-NBD-Pc Synthesis on the Cell SurfaceAfter addition of Cs-NBD-PA, we noticed that C6-NBD-PE was synthesized at levels up to half of total Cs-NBD-PC. At 10 "C, 97 f 2% ( n= 4) of the newly synthesized Cs-NBD-PE
PC and SM Synthesis theat was protected against depletion by BSA (cf. Table 11). The absence of C,-NBD-PE from the cell surface makes it unlikely that Cs-NBD-PC synthesis at the surface is caused by the stepwise methylation of C6-NBD-PE, which in any case is normally a minor pathway for PC production (2). The sn-1 and sn-2 C6-NBDpositional isomers of C6-NBDPC were produced in thesame ratio on thecell surface (Figs. 2, B and D, and 3E) as intracellularly (Figs. 2, A and C, and 3A).This implies that the &glyceride backbone of the C6NBD-PA was incorporated intact into Ce-NBD-PC in the surface reaction as it is in the intracellular CDP-choline pathway. This was independently confirmed by the addition of the purified sn-1 or the purified sn-2 isomer of Cs-NBDPA to MDCK cells at 10 "C, which resulted in the exclusive production of the corresponding isomer of Ce-NBD-PC both intracellularly and on the cell surface (not shown). While this could still imply a CDP-choline pathway, it excludes the introduction of a C6-NBD-acyl chain by transacylation or deacylation/reacylation into a preexisting PC (Fig. 1) as a possible mechanism for the synthesis of C6-NBD-PC at the cell surface. This agrees with the observation that free cgNBD-fatty acid cannot be reutilized by the cell (13). Ce-NBD-DG is the immediate precursor for the CDP-choline pathway, while both C6-NBD-PAand Ce-NBD-DG could be the direct precursor for the surface PC synthesis. The ratio of surface C,-NBD-PC to intracellular C6-NBD-PC synthesized from externally added Ca-NBD-PA was the same as that from Ce-NBD-DG (see above). This makes it very likely that Ce-NBD-PA wasfirst converted to C6-NBD-DG by a plasma membrane phosphohydrolase (13) (Fig. 1) after which it was utilized for both the intracellular andthe surface synthesis of Ce-NBD-PC in the same ratio as externally added Ca-NBDDG. Phosphorylcholine Donor for the Synthesis of Ce-NED-PC on the Cell Surface-With C6-NBD-DGas thedirect precursor for the surface synthesis of C6-NBD-PC,the reaction requires the transfer of a phosphorylcholine moiety onto the DG backbone. The phosphorylcholine could be derived either from pools of free phosphorylcholine or CDP-choline, or from the phospholipids PC or SM (or their corresponding lysophospholipids). Free phosphorylcholine and CDP-choline are absent from the incubation media, although the former could be released to some extent into themedium upon PC hydrolysis during signal transduction events (20, 21). Addition of phosphorylcholine and CDP-choline in the mM range in the external medium did not stimulate the synthesis of cell surface Ce-NBD-PC. Nor did it decrease the specific activity of cell surface Ce-NBD-PC in a 20 h double-label experiment (not
Cell Surface
1767
shown). Thus, PC and SM are the major candidates as the head group donor. It has been suggested that PCcan be synthesized from SM by a phosphocholine transferase, that is, by the reverse action of SM synthase (4-6), an enzyme which was originally assigned to theplasma membrane (6,22). It is now known that most SM synthase activity is localized to the Golgi (23-261, although still some 13% has been localized unequivocally to the plasma membrane (25). To determine whether SM synthase activity in MDCK cells is present at thecell surface, we applied a protocol similar to that used to study cell surface synthesis of PC. The short chain SM-precursor Cg-NBDceramide was added to MDCK cells on filters at 10 "C in the presence of BSA to deplete newly synthesized surface (26NBD-SM into the medium.Fig. 4 shows that part of the newly synthesized C6-NBD-SMcould indeed be depleted into the medium. The presence of C6-NBD-SMon the surface was not due to leakage from damaged cells, as less than 2% of newly synthesized C6-NBD-glucosylceramide,the other major product from C6-NBD-ceramide(11,16), was detected on the cell surface (Table 11). The fraction of the Ce-NBD-SM that was synthesized on the cell surface decreased with higher precursor concentration and temperature (Fig. 4). The concomitant relative increase in the intracellular synthesis suggests that at 0 "C and low ceramide concentrations diffusion of C6-NBD-ceramideinto the cell is a rate limiting factor for intracellular C6-NBD-SM synthesis. Synthesis at thebasolateral membrane levels off at about 10%of total synthesis of C6-NBD-SM.This is in good agreement with cell fractionation databy Futerman etal. (25) on rat liver. The real ratio between the SM synthaseactivities on the surface and in the cellis only reached when both reactions are saturated. From the absolute amounts synthesized in Fig. 4 saturation of both reactions is approached at 65 ~ L MC6-NBD-ceramide and 10 "C. SM synthesis occurred preferentially distal to the(medial) Golgi when plasma membrane SM had been hydrolyzed by exogenous sphingomyelinase to produce ceramide (27). Obviously, the contribution of surface SM synthesis to total cellular SM synthesis depends on how and where ceramide is generated. The apical/basolateral polarity of C6-NBD-SM that was synthesized at 10 "C on the cell surface was 0.14 f 0.04 ( n = 20), colocalizing SM synthesis and PCsynthesis on the basolateral plasma membrane (Fig, 4, TablesIand 11). This colocalization was the first piece of evidence to substantiate the idea that SM synthase was responsible for the conversion of C6-NBD-DG to Cs-NBD-PC on the cell surface. Independent evidence was obtained from testing the prediction that ceramide would inhibit PC synthesis, as DG had been found
TABLEI1 Inhibition of the synthesis of C8-NBD-PC on the cell surface by C6-NBD-ceramide MDCK cells were incubated with 5 p~ Cs-NBD-PA at 10 "Cas described under "Experimental Procedures" with simultaneous addition of CB-NBD-ceramideat the concentrations indicated. Numbers (pmol/107 cells) represent the mean of two experiments performed in duplicate ( n = 4). ND, not detectable ( e l pmol). Compared to the intracellular CB-NBD-PCsynthesis, that detected on the basolateral surface was reduced to 45 5% ( n = 4) in the presence of 25 p~ Cg-NBD-ceramide. In the control, 59% of newly synthesized Cs-NBD-PC was on the basolateral surface (apical/basolateral synthesis: 0.06). Of the newly synthesized C6-NBD-SM,this was 19% (5 W M C6-NBD-ceramide) and 12% (25 p M CB-NBD-ceramide) of the total, with apical/basolateral ratios of0.12 and 0.17, respectively. Cs-NBDceramide Basolateral Surface Cellular Apical Cellular Basolateral Surface Cellular Apical Cellular
Cs-NBD-PC
Ce-NBD-PE
CM
0 5 25
Ca-NBD-SM
Ce-NBD-gl~co~4ceramide
PWl
6.4 f 0.6 5.2 f 1.0 7.3 f 3 . 6
0.6 & 0.1 0.6 f 0.3 0.2 f 0 . 2
10.0 f 1.9 7.3 f 2.3 5.1 f 2 . 7
3.5 f 0.2 4.4 f 0.5 5.5 fO.l
ND ND ND
0 138 f 19 399 f 25
4 .+ 1 34 f 2 9 f5 34 & 9
0 13 f 1 52 & 5
ND ND
-
PC and SMthe Synthesis at
1768
Cell Surface
pase-generated DG stimulates protein kinase C. Its removal by the phosphocholine transferase would down-regulate the activity of the kinase (35). In addition, the newly produced 0 ceramide molecule has been implicated in the activation of other protein kinases and phosphatases and, afterthe action v 20 of ceramidase, in the inhibition of protein kinase C (29, 33, I 34). Thus, the ceramide-DG interconversion could be crucial 4 in regulating the relative concentration of the two lipid second messengers and thereby the overall signal transduction pathways mediated by these lipids. It will be important to find out whether the affinity of the phosphocholine transferase at the 0 10 20 65 cell surface for DG and ceramide is itself subject to regulation. Implications of SM-PC Interconversion on the Cell Surface C6-NBD-ceramide (pM) FIG.4. Synthesis of Cs-NBD-SM on the cellsurface. Varying for Epithelial Lipid Polarity-Conversion of SM to PCon the concentrations of C6-NBD-ceramidewere added to both sides of a basolateral surface of epithelial cells would yield a relative monolayer ofMDCK cells and the cells were incubatedin the presence enrichment of SM on the apical and of PC on the basolateral of BSA for 3 h as described under “Experimental Procedures.” Ex- surface, basolateral conversion of PC to SM would create the periments were performed completely at 0 “C or 10 “C. Cs-NBD-SM opposite situation. In both cases an epithelial lipid polarity was harvested from the apical and basal medium and fromthe cells. would be generated. In contrast, the basolateral enrichment The percentage of total C6-NBD-SM present in the medium from each side of the filter is plotted against the concentration C6-NBD- of both PC and SM, which occurs for example in intestinal ceramide applied. Standard deviations (n = 4) fall within the size of epithelia ( l ) , could be caused by selective apical hydrolysis of the symbols. C6-NBD-SM synthesis at 0 “C: 2, 9,and 44 pmol/107 both PC and SM butnot by the polarized disposition of the cells at 1, 5, and 25 p~ ceramide; at 10 ‘C: 15, 61, 206, 333, and 426 phosphocholine transferase as the transferase can only expmol/lO’ cells at 1, 5, 15, 25, and 65 p~ ceramide. A similar experi- change the one lipid for the other. However, evidence for a ment for 3 h at 10 “C using Cs-NBD-PA as a precursor showed that very different scenario for the generation of epithelial lipid C6-NBD-PC synthesis on the basolateral surface decreased from 67 f 1 % of total synthesis at 1 pM Cs-NBD-PA to 55 & 2% at 25 p M polarity has been found. Intracellularly synthesized lipids are Ce-NBD-PA (n = 2), total synthesis being 5 and 46 pmol/lO’ cells, sorted before they reach the epithelial cell surface: C6-NBDglucosylceramide displayed a 2-7-fold higher apical/basolatrespectively. era1 polarity of delivery than Cs-NBD-SM (11, 16, 36). This to inhibit SM synthesis (6, 28). The addition of Cs-NBD- is not caused by the selective synthesis of Cs-NBD-SM on the ceramide indeed resulted in a concentration dependent inhi- basolateral surface by the phosphocholine transferase activity bition of Cs-NBD-PC synthesis from Cs-NBD-PA on the shown in Fig. 4, as in these experiments any NBD-lipids basolateral cell surface, but not of intracellular Cs-NBD-PC synthesized at thecell surface were removed beforethe actual synthesis (Table 11).The synthesis of intracellular C6-NBD- transport measurements. Also, the contribution of surface synthesis to total SM biosynthesis at 37 “C (-lo%, Fig. 4) PE in fact was stimulated. A similar inhibition of surface Cs-NBD-PC synthesis was wouldbe fartoo low to explain the observed differences. observed after 61 f 3% ( n = 4) of the cellular SM had been Finally, selective hydrolysis of SM at theapical surface could hydrolyzed by exogenous sphingomyelinase as was determined be excluded by the observation that no such hydrolysis was in parallel samples (see “Experimental Procedures”). Synthe- observed in experiments where the C6-NBD-SMwas allowed sis of Cs-NBD-PC at the basolateral cell surface, assayed at to remain in the apical membrane for prolonged periods of 10 “C as under ‘‘Experimental Procedures,” was inhibited to time in the absence of BSA from the medium (11,36). The contribution of the hydrolysis of PC and SMor their 37 f 7%, while apical and intracellular synthesesof Cs-NBDinterconversion on the cell surface to the steady-state lipid PC were 98 f 6% and 83 f 3% of the control, respectively ( n composition of the plasma membrane (domains) may be sig= 4). Incomplete inhibition of surface PC synthesis may have been due to incomplete hydrolysis of surface SM. Synthesis nificant. Further insight into this problem awaits work supof Cs-NBD-PE in these experiments was reduced by 75%. porting the general nature of the lipid-mediated signal transBecause a reduction of Cs-NBD-PE was not observed after duction processes and determining their frequency and the level of surface lipid metabolism during such events. Finally, ceramide addition (Table 11), it must have been an effect of a better understanding is required not only of the interconthe loss of SM for which we presently have no explanation. version of lipid second messengers as shown in the present Implications of SM-PC Interconversion on the Cell Surface study but, perhaps more importantly, of the mechanisms by for Regulation of Signal Transduction-The colocalization of which cells remove the various lipid second messengers from PC and SM synthesis on the basolateral plasma membrane the plasma membrane to restore the initial equilibrium. and the inhibition of the surface PC synthesis by ceramide support the role of SM asthe phosphorylcholine donor in the Acknowledgment-We are grateful to Marion Thielemans for exsurface PC synthesis in MDCK cells. SM and DG are con- pert technical help. verted to PC and ceramide and vice versa by a phosphocholine REFERENCES transferase that no longer by definition functions strictly as 1. Simons, K., and van Meer, G. (1988) Biochemistry 27,6197-6202 a “SM synthase’’ (6). In fact, this enzymemaybe partly 2. Kent, 6. (1990) Prog. Lipid Res. 29,87-105 responsible for the hydrolysis of SM to ceramide during signal 3. Hjelmstad, R. H., and Bell, R. M. (1991) Biochemistry 30,. 1731-1740 4. Kudoh, T.,Velkoff, M. A,, and Wenger, D.A. (1983) Bmhrm. Bmphys. transduction, which has been attributedtoa cell surface Acta 754,82-92 sphingomyelinase (20, 29, 30). The activity of the surface 5. Spence, M. W., Clarke, J. T.R., and Cook, H. W. (1983) J. Biol. Chem. 268,859558600 phosphocholine transferase, which may bea different enzyme 6. Marmraf, W.-D., and Kanfer, J. N. (1984) Biochim. Biophys. Acta 7 9 3 , 343=353 from the Golgi SM synthase,is intriguing in relation to those 7. van Meer, G. (1993) Curr. @in. Cell Biol. 6,661-673 types of signal transduction whereby DG and ceramide are 8. Kaplan, M. R.,and Simoni, R.D.(1985) J. Cell Biol. 101,441-445 9. Wirtz, K. W. A. (1991) Annu.Reo. Biochem. 60, 73-99 produced in the plasma membrane (29, 31-34). Phospholi40
-
basolateral0”C basolateral 10°C apical 0°C apical 10°C
PC and SM Synthesis at theCell Surface 10. Devaux, P. F. (1991)Biochemistry 30,1163-1173 11. van M e r , G., Stelzer, E. H. K., Wijnaendts-van-Resandt, R. W., and Simons, K. (1987)J. Cell BWL 106,1623-1635 0,R.E., Longmuir, K. J., Martin, 0. C., and Struck, D. K. (1981)J. 12. ell Bwl. ' 91,872-877 13. Papno,R. E., Longmuir. K. J.. and Martin. 0.C. (1983)J. Biol. Chem.
pY
"- - _"
n n ~m u - 3 M -_ n "_,
14. Pagano, R.E., and Longmuir, K. J. (1985)J. Bwl. Chem. 260,1909-1916 16. Sleiqht, R. G., and Pagano, R. E. (1984)J. CeU Biol. 99,742-751 16. van t Hof, W., Silvius, J., Wieland, F., and van Meer, G. (1992)Biochem. J. 283,913-917 17. Dao, H.-N. T.,McIntyre, J. C., and Sleight, R. G. (1991)Anal. Biochem. 196.4653 ~, 18. M a i n , 0.C., and Pagano, R. E. (1986)A d Biochem. 169,101-108 19. Fries, E., and Lindstrom, I. (1986)Biochem. J. 237,33-39, 20. Okazaki. T..Bell. R. M.. and Hannun. Y. A. (1989) . , J. Bwl. Chem. 264. 1907619680 21. van Blitterswijk, W. J., Hilkmann, H., de Widt, J., and van der Bend, R. L. (1991)J. Biol. Chem. 266,10344-10350 ~~
~~
'
1769
22. Voelker, D. R., and Kennedy, E. P. (1982)Biochemistry 21,2753-2759 23. Lipsky, N. G., and Pagano, R. E. (1985)J. CeU Biol. 100,27-34 24. Malgat, M., Maurice, A., and Baraud, J. (1986)J. Lipid Res. 27, 251-260 25. Futerman, A. H., Stieger, B., Hubbard, A. L., and Pagano, R. E. (1990)J. Biol. Chem. 266,8650-8657 26. Jeckel. D.. Karrenbauer. A,. Birk. R.. Schmidt. R. R.. and Wieland.F. (1990)@EBS Lett. 261,155-157 ' 27. Kallen, K.-J., Quinn, P., and Allan, D. (1993)Biochim. Biophys. Acta 1166,305-308 28. Ullman, M. D., and Radin, N. S. (1974)J. Biol. Chem. 249,1506-1512 29. Kolesnick, R. (1992)Trends Cell Biol. 2,232-236 30. Koval, M., and Pagano, R. E. (1991)Biochim. Biophys. Acta 1082, 113125 31. Licovitch, M. (1992)Trends Biochem. Sci. 17,393-399 32. Nishizuka, Y. (1992)Science 268,607-614 33. Hannun, Y. A., and Bell, R. M. (1987)Science 236,670-674 34. Merrill, A. H., Jr., (1991)J. Bknerg. Biomembr. 23, 83-104 35. Hampton, R.Y., and Morand, 0.H. (1989)Science 246,1050 36. van't Hof, W., and van Meer, G . (1990)J. Cell BioL 111,977-986
