Dihydroceramide Biology
STRUCTURE-SPECIFIC METABOLISM AND INTRACELLULAR LOCALIZATION*

(Received for publication, January 10, 1997, and in revised form, May 28, 1997)

Jan Willem Kok Dagger §, Mariana Nikolova-Karakashian , Karin Klappe Dagger , Chris Alexander and Alfred H. Merrill Jr.

From the Dagger  Department of Physiological Chemistry, University of Groningen, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands and the  Department of Biochemistry, Emory University, Atlanta, Georgia 30322

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

This study utilized fluorescent analogs to characterize the intracellular transport and metabolism of dihydroceramide (DH-Cer), an intermediate in de novo sphingolipid biosynthesis. When 6-[N-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]hexanoyl-DH-Cer (C6-NBD-DH-Cer) was incubated with HT29, NRK, BHK, or HL-60 cells, it was efficiently converted to dihydrosphingomyelin and dihydroglucosylceramide, and a number of other sphingolipids, with the nature of the products depending on the cell line. In addition, complex sphingolipids were formed that contained a desaturated (sphingosine) backbone, indicating that DH-Cer (and/or its metabolites) were substrates for the desaturase(s) that introduce the 4,5-trans double bond. Based on the kinetics and inhibitor studies, double bond addition did not appear to occur with the complex sphingolipids directly, but rather, during turnover and resynthesis. The conversion of C6-NBD-DH-Cer to more complex sphingolipids was highly stereoselective for the natural D,erythro isomer of C6-NBD-DH-Cer. Interestingly, the stereochemistry of the sphingoid base backbone also affected the localization of fluorescent sphingolipids: the D,erythro species appeared in the Golgi apparatus, whereas other stereo-isomers accumulated in the endoplasmic reticulum. In addition to C6-NBD-Cer and C6-NBD-DH-Cer, C6-NBD-4-D-hydroxy-DH-Cer gave rise to formation of complex sphingolipids and localized at the Golgi apparatus. These studies indicate that dihydroceramide is used as the initial backbone of complex (glyco)sphingolipids, perhaps to avoid build up of ceramide as an intermediate since this is such a potent bioactive compound. The stereoselectivity in transport and metabolism suggests that trafficking of ceramide is protein-directed rather than simply a consequence of vesicular membrane flow.


INTRODUCTION

Ceramide (Cer)1 is a potent biologically active molecule, which is involved in regulation of important cell biological processes, such as cell growth, cell differentiation, diverse cell functions, and apoptosis (1, 2). For most of these biological activities, ceramide appears to require a "desaturated" sphingoid base backbone (sphingosine; Ref. 3). However, little is known about introduction of the 4-trans double bond into the sphingoid base backbone during de novo biosynthesis of sphingolipids or about the fate of dihydroceramide (DH-Cer) of complex sphingolipids when they are turned over. The latter is particularly relevant given the fact that Cer-mediated signal transduction involves hydrolysis of more complex sphingolipids, such as sphingomyelin (SM).

Ceramide synthesis begins with the condensation of serine and palmitoyl-CoA by serine palmitoyltransferase to form 3-ketosphinganine, which is subsequently reduced to sphinganine. The addition of an amide-linked fatty acid by ceramide synthase yields DH-Cer (4). The introduction of the 4-trans double bond to form Cer has been demonstrated to occur after the synthesis of DH-Cer (4, 5). The initial steps up to the formation of DH-Cer have been shown to take place at the cytosolic surface of the ER (6, 7); however, the biosynthesis of SM (8) and glucosylceramide (GlcCer; Refs. 9 and 10) from Cer occurs at the (early) Golgi apparatus. Thus, similar to glycoprotein processing, most (glyco)sphingolipid biosynthesis appears to be coupled to vectorial transport through the ER/Golgi system. Since GlcCer is synthesized at the cytosolic surface of the Golgi membrane (9, 11, 12), one can envision that Cer, after synthesis at the cytosolic surface of the ER, is transferred to the cytosolic surface of the Golgi, where it can be used as a substrate for GlcCer synthesis. In addition, a fraction of the Cer must be translocated at the Golgi to reach the luminal leaflet for SM biosynthesis (13).

The goal of this study was to clarify the metabolic pathways for conversion of DH-Cer to Cer and more complex sphingolipids, as well as their intracellular localization and transport. For this purpose, we made use of a fluorescent analog of DH-Cer, 6-[N-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]hexanoyl-dihydroceramide (C6-NBD-DH-Cer) in analogy to C6-NBD-Cer (14, 15), which after insertion into the outer leaflet of the plasma membrane undergoes transbilayer movement and diffusion into the cell interior, where it integrates into the membranes of various organelles (16, 17). The major part of the intracellular C6-NBD-Cer pool is located at the ER, by virtue of its large membrane surface, but during prolonged incubation of the cells at 37 °C, becomes concentrated in the Golgi apparatus (16, 17). In the Golgi, conversion of C6-NBD-Cer to C6-NBD-SM and C6-NBD-glycolipids takes place, that subsequently find their way to the plasma membrane (16, 17), at least in part via a vesicular transport system (18). However, it is not known how (C6-NBD-)Cer is translocated from the ER to the Golgi, whether by co-transport with proteins in vesicular shuttles or by (protein-mediated) transport through the cytosol.

The results of our studies show that the incorporation of C6-NBD-DH-Cer into more complex sphingolipids can occur prior to addition of the 4-trans double bond; furthermore, the structural requirements for metabolism and intracellular localization of fluorescent sphingolipids at the Golgi apparatus are highly correlated. The absence of the 4-trans double bond (C6-NBD-DH-Cer) or the presence of an -OH group at the C4 position of the sphingoid backbone (C6-NBD-4-D-hydroxy-DH-Cer) did not interfere with biosynthesis of (glyco)sphingolipids from ceramides, but compounds without the natural (D,erythro) stereochemistry were poorly metabolized and did not concentrate in the Golgi apparatus. It appears that the level of C6-NBD-Cer as an intermediate in biosynthesis of complex desaturated sphingolipids is kept low in cells, which may be related to the potency of this sphingolipid in the regulation of cell function as a second messenger (1, 2).


EXPERIMENTAL PROCEDURES

Materials

The sphingoid bases D,erythro-sphingosine, D,erythro-sphinganine, L,erythro-sphingosine, L,erythro-sphinganine, D,threosphinganine, 1-deoxy-D,erythro-sphinganine, and 1-deoxy-5-hydroxy-D,erythro-sphinganine were synthesized in the laboratory (19). The succinimidyl ester of C6-NBD and propidium iodide were from Molecular Probes (Eugene, OR). Reduced streptolysin O (SL-O) was purchased from Sanofi Diagnostics Pasteur (Marnes-la-coquette, France). 4-D-Hydroxysphinganine, ATP, phosphocreatine, creatine phosphokinase, bovine serum albumin, EGTA, uridine 5'-diphosphoglucose, brefeldin A (BFA), sphingomyelinase (from Staphylococcus aureus), alpha -galactosidase (from Aspergillus niger), beta -galactosidase (from A. niger), alpha -N-acetylgalactosaminidase, beta -N-acetylglucosaminidase (from bovine epididymis), and neuraminidase (type II from Vibrio cholerae and type X from Clostridium perfringens) were purchased from Sigma. NADH and ceramide glycanase were from Boehringer Mannheim. DL-threo-1-Phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP) was from Matreya Inc. (Pleasant Gap, PA).

C6-NBD-sphingolipid Synthesis and Incubation Conditions

C6-NBD-ceramides were synthesized from various sphingoid bases and the succinimidyl ester of C6-NBD according to Kok et al. (20). Synthesized C6-NBD-sphingolipids were purified by HPTLC and checked for purity by HPLC, before and after acid hydrolysis (see below). Purity was >= 99%. C6-NBD-DH-GlcCer and C6-NBD-DH-SM were synthesized in HT29 cells from C6-NBD-DH-Cer. These products were purified by HPTLC, followed by HPLC, until homogeneity. For cell incubations, the C6-NBD-sphingolipids in ethanol were injected into serum-free culture medium (0.5%, v/v, final ethanol concentration) at concentrations of C6-NBD-sphingolipid up to 5 µM (see also Ref. 21).

Lipid Extraction and Analysis

Following these incubations, equal amounts of cells for each condition were subjected to lipid extraction by the procedure of Bligh and Dyer (22). Extracted lipids were analyzed by HPTLC, HPLC, or both. For HPTLC, the plates were developed in two dimensions, employing solvent system A (CH3Cl/CH3OH/20% (w/v) NH4OH, 14:6:1, v/v/v) in the first dimension and solvent sytem B (CH3Cl/CH3OH/86% (v/v) CH3COOH, 45:20:7, v/v/v) in the second. This procedure allowed separation of most C6-NBD-sphingolipid species, but does not clearly separate C6-NBD-GlcCer from C6-NBD-GalCer (cf. Ref. 23). For further analysis, individual C6-NBD-sphingolipids were scraped from the HPTLC plates and eluted from the silica by washing with 10 ml of CH3Cl/CH3OH (1:1, v/v), followed by 10 ml of CH3OH. The individual C6-NBD-sphingolipid species were quantified by measuring the NBD fluorescence in CH3OH in a fluorimeter at excitation wavelength 465 nm and emission wavelength 530 nm. Thereafter, C6-NBD-GlcCer and C6-NBD-GalCer were separated on HPTLC plates, sprayed with boric acid (1.25 g/50 ml CH3OH), using solvent system C (CH3Cl/CH3OH/25% (w/v) NH4OH, 13:7:1, v/v/v), followed by quantification as described.

In some cases, individual biosynthesized C6-NBD-sphingolipids were subjected to enzymatic analysis for identification purposes. After extraction from cells, HPTLC separation, and elution from the silica, these assays were usually performed in a 50 mM CH3COONa solution at pH 4.5 at 37 °C, in the presence of 0.1% of Triton X-100 or 150 µg of taurodeoxycholate. The resulting products were analyzed either by HPTLC or HPLC.

During incubation of HT29 G+ cells with C6-NBD-DH-Cer, a substantial amount of a C6-NBD-lipid was synthesized that did not migrate with the standards (see also Ref. 23). An attempt was made to identify this NBD-lipid by enzymatic analysis. It was completely resistant to (bacterial) sphingomyelinase in the presence of Triton X-100 and Mg2+ at pH 7.4, but was almost completely (circa 95%) degraded to ceramide by ceramide glycanase, when incubated for 2 h at 37 °C in the presence of taurodeoxycholate at pH 5.0. Various glycosidases were tested for their capacity to degrade this C6-NBD-glycolipid. alpha - and beta -galactosidase, beta -N-acetylglucosaminidase, and neuraminidase (either from V. cholerae or C. perfringens) were ineffective; however, the C6-NBD-glycolipid was converted to C6-NBD-LacCer by alpha -N-acetylgalactosaminidase, which suggests it is a ceramide trihexoside (CTH), containing a terminal N-acetylgalactosamine linked in alpha  configuration to the galactose of C6-NBD-LacCer. Further structure identification was not performed, due to the small amounts of C6-NBD-lipid available and contamination of this C6-NBD-lipid by endogenous cell lipids comigrating on HPTLC.

HPTLC does not separate C6-NBD-DH-sphingolipids from their desaturated counterparts; therefore, HPLC analyses were employed using a system described by Pagano and Martin (24) that allows separation of C6-NBD-DH-Cer from C6-NBD-Cer. We found that not only the two ceramide species, but virtually all C6-NBD-sphingolipids synthesized from C6-NBD-DH-Cer were resolved by this system. The fluorescent sphingolipids were separated on an analytical reverse-phase C18 column (Waters, Milford, MA) with CH3OH/H2O/H3PO4 (850:150:1.5, v/v/v) as the mobile phase (24) and detected with a Shimadzu RF-535 fluorescence detector and a C-R5A plotter/integrator. Total lipid extracts or individual C6-NBD-sphingolipid species, previously separated on HPTLC, were dissolved in 500 µl of mobile phase, and an aliquot was loaded on the HPLC column, followed by elution at a flow rate of 1 ml/min. Fluorescent sphingolipids were detected at 530 nm by excitation at 465 nm, and quantified by means of the peak area (3270 arbitrary units/pmol of C6-NBD-sphingolipid). The peak area was proportional to the injected amount of fluorescent lipid at least up to 300 pmol.

When a total lipid extract from HT29 G+ cells, incubated with C6-NBD-DH-Cer for 24 h, was analyzed by HPLC, a number of peaks appeared, as shown in Fig. 1. These peaks were identified as follows. 1) Some peaks could be directly identified by comparison to synthesized standards (C6-NBD derivatives of DH-Cer/DH-LacCer/Cer/GlcCer/GalCer/SM). 2) Further identification was performed by separation of a total C6-NBD-sphingolipid extract on two-dimensional HPTLC, as described above, followed by separation of C6-NBD-GlcCer from C6-NBD-GalCer. The resulting individual sphingolipid species, which were in fact mixtures of the desaturated and saturated counterpart, were then separated on HPLC. This allowed identification of all C6-NBD-sphingolipid peaks in the HPLC profile. All quantitative measurements of C6-NBD-(DH-)-sphingolipids were performed according to this methodology. 3) Confirmation of identification was obtained in the case of C6-NBD derivatives of (DH-)SM, (DH-)-LacCer, and (DH-)CTH, by enzymatic degradation of these C6-NBD-sphingolipids by sphingomyelinase and ceramide glycanase, respectively. The products were identified as C6-NBD-(DH-)Cer by HPLC. In the HPLC profile of a total C6-NBD-sphingolipid extract, two series of sphingolipid species could be discerned, one containing DH-Cer as the backbone and the other Cer. In each series the elution order is C6-NBD-CTH right-arrow LacCer right-arrow GlcCer + GalCer right-arrow SM right-arrow Cer (C6-NBD-GlcCer and C6-NBD-GalCer have the same retention times). The desaturated and saturated form of each sphingolipid species were separated by 6.7 min, on average.


Fig. 1. HPLC profile of C6-NBD-(glyco)sphingolipids synthesized from C6-NBD-DH-Cer in HT29 cells. Approximately 107 HT29 G+ cells were incubated with 2.5 µM C6-NBD-D,erythro-DH-Cer for 24 h at 37 °C. A total lipid extract was made and run on an analytical reverse-phase HPLC column, employing CH3OH/H2O/H3PO4 (850:150:1.5, v/v/v) as the mobile phase, at a flow rate of 1 ml/min. Fluorescent C6-NBD-lipids were detected at 530 nm by excitation at 465 nm. A number of peaks were detected, which were identified by means of standards and comparison to HPTLC analysis, combined with enzymatic degradation (see "Experimental Procedures").
[View Larger Version of this Image (13K GIF file)]

The sphingoid base composition of endogenous sphingolipids was analyzed as follows. Total lipid extracts were subjected to base hydrolysis in 0.1 N KOH in CH3OH for 90 min at 37 °C. After neutralization and reextraction, the sphingoid base composition of either the total sphingolipid pool or of individual lipid species was analyzed. In the latter case, base hydrolyzed extracts were separated on HPTLC, employing solvent system D (CH3Cl/CH3OH/67% (v/v) CH3COOH, 28:15:3, v/v/v), followed by elution of lipid species from the silica. Either a total sphingolipid extract or individual sphingolipid species were subsequently acid hydrolyzed in 0.5 N HCl in CH3OH for 16 h at 63-64 °C, followed by neutralization and reextraction. Released sphingoid bases were derivatized with ortho-phthaldialdehyde and analyzed by HPLC, as described previously (25). To correct for possible losses during the acid hydrolysis and extraction procedures, samples were spiked with 100 pmol of N-acetyl-C20-sphinganine as an internal standard. This was done before the lipid extraction in the case of total sphingolipid analysis and before lipid elution from the silica, in the case of analysis of the sphingoid base composition of individual sphingolipid species.

Cell Culture

Monocultures of HT29 G+ cells were grown in Dulbecco's modified Eagle's medium (DMEM, containing 25 mM glucose), supplemented with 10% (v/v) decomplemented (56 °C, 30 min) fetal calf serum (FCS), in a water-saturated atmosphere of 5% CO2/95% air. During the exponential phase of growth, the culture medium was changed every 48 h. NRK cells were grown in DMEM with 5% FCS, BHK cells in Glasgow minimum essential medium supplemented with 5% FCS and 10% tryptose phosphate broth, and HL-60 cells in RPMI 1640 medium containing 10% FCS.

Preparation of Subcellular Fractions from Rat Liver or HT29 Cells and Cell Permeabilization

Cells were permeabilized with SL-O according to the procedure of Tan et al. (26), using 1.3 units/ml SL-O. Efficiency of permeabilization (>90%) was checked in every experiment by counting propidium iodide-labeled (permeabilized) and total cells. Experiments with SL-O-permeabilized cells were performed in the presence of an ATP-regenerating system (18). Cell homogenates were prepared from approximately 5 × 107 HT29 cells. Cells were washed two times with Hank's buffer, followed by scraping and centrifugation. The cell pellet was frozen at -80 °C for 30 min and thawed. The pellet was then resuspended in buffer (140 mM KCl, 0.5 mM KH2PO4, and 20 mM HEPES, pH 7.3) and homogenized by 100 strokes in a tight Dounce homogenizer. Cytosol was prepared from HT29 G+ cells as described before (18). Rat liver ER membranes were isolated as described by Croze and Morré (27). Rat liver Golgi membranes were isolated as described previously (28). For in vitro C6-NBD-DH-GlcCer desaturation studies, ER and/or Golgi membranes were incubated in the presence of 1 mM NADH.

Fluorescence Microscopy

Cells were grown on glass coverslips, placed in 35-mm diameter plastic Petri dishes. Experiments were carried out 72 h after passage. Cells were incubated with C6-NBD-(DH-)- Cer (2.5-5 µM) in culture medium without FCS for 1 or 24 h at 37 °C, followed by extensive washing. After the 1-h incubation, a "back-exchange" was performed by incubating the cells with 5% bovine serum albumin in cold (2 °C) medium for 30 min, followed by extensive washing. Fluorescence microscopy was performed with a Leitz Orthoplan microscope equipped with a Leitz Vario Orthomat 2 photography system. Photomicrographs were taken at 10-s exposure times using Illford HP5 film that was processed at 3200 ASA.

Protein Determination

Protein was determined by the Lowry procedure, modified as described by Peterson (29), with bovine serum albumin as the standard.


RESULTS

C6-NBD-DH-Cer Is Converted to Complex Sphingolipids, Both Saturated and Desaturated

C6-NBD-Cer, a short-chain fluorescent analog of ceramide, is a well established precursor for (glyco)sphingolipid biosynthesis (17, 23, 30). In HT29 G+ cells, a number of products are synthesized during a 24-h incubation with C6-NBD-Cer, the main products being C6-NBD-SM and C6-NBD-GlcCer, plus lesser amounts of C6-NBD-GalCer, C6-NBD-LacCer and C6-NBD-CTH (see "Experimental Procedures"; cf. Ref. 23). In this study, we examined the fate of a C6-NBD-analog of dihydroceramide (C6-NBD-DH-Cer, Fig. 2C), which should, in principle, serve as a precursor for the synthesis of C6-NBD-Cer via double bond introduction between C4 and C5 of the sphinganine backbone. When HT29 G+ cells were incubated with 2.5 µM C6-NBD-DH-Cer for 24 h at 37 °C, the product profile was qualitatively comparable to that obtained after C6-NBD-Cer incubation; however, there were quantitative differences (Fig. 1, Table I; cf. Ref. 23).



Fig. 2. Structures of sphingoid bases. A, sphingosine and deoxy- and hydroxysphinganines; B, stereoisomers of sphinganine; C, C6-NBD-derivatized sphinganine. Note that 1-deoxy-5-hydroxysphinganine has an additional -OH group at C5, compensating for the increased hydrophobicity due to the loss of the -OH group at C1.
[View Larger Versions of these Images (14 + 11K GIF file)]

Table I. Biosynthesis and desaturation of C6-NBD-sphingolipid products from C6-NBD-DH-Cer in different cell types

Approximately 107 HT29 G+, NRK or HL-60 cells were incubated with 2-4 µM C6-NBD-DH-Cer for 24 h at 37 °C, followed by extraction of total lipid. C6-NBD-sphingolipids were separated by two-dimensional HPTLC, followed by further separation of C6-NBD-GlcCer from C6-NBD-GalCer, as described under "Experimental Procedures." Individual C6-NBD-species were quantified by fluorescence after scraping from the TLC plate and elution from the silica, as described. After solvent evaporation, each C6-NBD-sphingolipid species was analyzed by HPLC, separating the dihydrosphingolipid from the desaturated counterpart. The amounts of the latter two forms of the three major sphingolipid species (Cer, GlcCer, SM) were used to calculate the percent of desaturation of each species, according to the formula {pmol desaturated/(pmol desaturated + pmol saturated)} × 100%. Each C6-NBD-(dihydro)sphingolipid species is expressed as the percentage of total C6-NBD-lipid (ND, not detectable). The total amount of C6-NBD-lipid was 1 nmol/mg of cellular protein, on average, and similar for the three cell types. Values are the means (±S.D.) of three independent experiments.

HT29 NRK HL-60

C6-NBD-DH-Cer 38.0  ± 2.0 47.3  ± 6.4 66.7  ± 12.3
Cer 0.7  ± 0.1 1.0  ± 0.1 0.2  ± 0.04
Desaturation (%) 1.7  ± 0.4 2.0  ± 0.9 0.3  ± 0.4
DH-SM 20.0  ± 5.3 43.0  ± 5.3 9.4  ± 3.3
SM 0.4  ± 0.1 6.7  ± 0.8 1.4  ± 0.5
Desaturation (%) 2.1  ± 1.1 13.4  ± 9.1 12.9  ± 0.4
DH-GlcCer 30.8  ± 4.1 1.1  ± 0.3 9.5  ± 4.2
GlcCer 4.1  ± 0.5 0.6  ± 0.2 3.7  ± 1.6
Desaturation (%) 11.8  ± 5.0 35.5  ± 12.6 27.8  ± 3.6
DH-LacCer + LacCer 1.0  ± 0.4 0.3  ± 0.1 8.9  ± 2.6
DH-CTH + CTH 3.7  ± 0.9 ND 0.2  ± 0.1
DH-GalCer + GalCer 1.3  ± 0.2 ND ND

The (glyco)sphingolipids synthesized form C6-NBD-DH-Cer were analyzed for the presence of a backbone double bond by reverse-phase HPLC (Fig. 1, Table I; see "Experimental Procedures"). All of the C6-NBD-sphingolipid species of HT29 cells were composed of a saturated (DH-Cer-containing) and a desaturated (Cer-containing) form after 24 h (Fig. 1). In all C6-NBD-sphingolipid species, sphinganine predominated as the sphingoid backbone and the percentages of desaturation differed (Table I) in the order C6-NBD-GlcCer >>  C6-NBD-SM > C6-NBD-Cer.

Direct Desaturation of C6-NBD-GlcCer Does Not Occur in HT29 Cells

The results presented show that C6-NBD-DH-Cer is used by cells as a substrate for headgroup addition. Since desaturated complex sphingolipids are also formed, this opens the possibility that the double bond is introduced at the level of ceramide and/or more complex dihydro-sphingolipids, as depicted in Fig. 3. To discriminate among these alternatives, the following experiments were performed. 1) Kinetic analysis of GlcCer formation showed that desaturated GlcCer was not found in HT29 cells during the first 6 h of incubation with C6-NBD-DH-Cer, when a considerable pool of DH-GlcCer had already build up (Table II). This suggested that a direct precursor-product relation between DH-GlcCer and GlcCer was unlikely. 2) When HT29 cells were first incubated with C6-NBD-DH-Cer for 6 h at 37 °C (allowing the formation of C6-NBD-DH-GlcCer, but not GlcCer), and then incubated in the presence of PDMP (in a concentration that is known to inhibit the formation of C6-NBD-GlcCer from C6-NBD-Cer by >90%), no synthesis of GlcCer occurred (Table II). The amount of C6-NBD-DH-GlcCer decreased, showing that there is considerable turnover of this sphingolipid. 3) C6-NBD-DH-GlcCer was purified from HT29 cells by HPLC (see "Experimental Procedures"). The lipid was, either in the presence or absence of PDMP, added to intact or streptolysin O-permeabilized HT29 cells, or to homogenates or cytosol from HT29 cells, potentially allowing it to reach the desaturase more efficiently. Alternatively, isolated liver ER or Golgi membranes or a combination of these two membrane preparations were incubated with C6-NBD-DH-GlcCer, in the presence of 1 mM NADH. Formation of C6-NBD-GlcCer only occurred in intact cells and was inhibited by PDMP, again indicating that C6-NBD-GlcCer biosynthesis is preceded by the formation of C6-NBD-Cer from C6-NBD-DH-Cer. The latter, i.e. desaturation of C6-NBD-DH-Cer, was observed to occur in ER membranes (data not shown). 4) When HT29 cells were incubated with C6-NBD-DH-Cer for 24 h at 37 °C in the presence of BFA, the formation of C6-NBD-DH-GlcCer increased 3.49-fold (n = 2), but the formation of C6-NBD-GlcCer decreased (0.57-fold; n = 2). In other words, more C6-NBD-DH-GlcCer as a potential substrate for desaturation was available, but less desaturated C6-NBD-GlcCer was found. It should be noted that desaturase activity per se was hardly affected by BFA, since formation of C6-NBD-Cer decreased only slightly (0.83-fold; n = 2).


Fig. 3. Potential metabolic pathways involved in C6-NBD-GlcCer (SM) biosynthesis from C6-NBD-DH-Cer. Biosynthesis of C6-NBD-GlcCer and SM from C6-NBD-Cer is well established (17, 23, 30). Apparently, as shown in this study, (C6-NBD-)DH-Cer, which is the proposed precursor for (C6-NBD-)Cer biosynthesis (desaturation), can serve as a direct precursor for (C6-NBD-)-GlcCer and SM biosynthesis, without preceding desaturation. The question arises whether (C6-NBD-)DH-GlcCer and DH-SM can serve as precursors for GlcCer and SM formation, involving direct desaturation of headgroup containing sphingolipids. Alternatively, GlcCer and SM are formed as a result of DH-GlcCer (DH-SM) hydrolysis, followed by desaturation of DH-Cer to Cer and subsequent headgroup addition. Given experimental results, the latter appears more likely see "Results").
[View Larger Version of this Image (15K GIF file)]

Table II. C6-NBD-GlcCer formation in the presence of PDMP

Approximately 107 HT29 G+ cells were incubated with 2.5 µM C6-NBD-DH-Cer for 6 or 24 h at 37 °C. Alternatively, the cells were first incubated with C6-NBD-DH-Cer alone for 6 h, followed by addition of PDMP (100 µM) and continued incubation with both C6-NBD-DH-Cer and PDMP for 18 h. After lipid extraction, individual C6-NBD-sphingolipid species were separated by two-dimensional HPTLC, followed by separation of C6-NBD-GlcCer from C6-NBD-GalCer, as described under "Experimental Procedures." C6-NBD-GlcCer was recovered from the silica and analyzed by HPLC, separating C6-NBD-DH-GlcCer from C6-NBD-GlcCer. The amounts of the latter two forms of C6-NBD-glucosylceramide are given in pmol/mg of cellular protein. Data are the means of two independent experiments (ND, not detectable).

C6-NBD-DH-Cer incubation schedule C6-NBD-DH-GlcCer C6-NBD-GlcCer

pmol/mg protein
6 h 27.5 ND
24 h 119.5 20.9
6 h + 18 h (PDMP) 5.8 ND

Since direct desaturation of C6-NBD-DH-GlcCer was not observed in HT29 cells, the alternative metabolic route, wherein C6-NBD-GlcCer is hydrolyzed to C6-NBD-DH-Cer which is desaturated to C6-NBD-Cer followed by reglycosylation (Fig. 3), is more likely. The observation that C6-NBD-Cer levels are kept low (Table I) suggests that shuttling of C6-NBD-Cer into C6-NBD-sphingolipid products is a very efficient process.

C6-NBD-DH-Cer Metabolism (Desaturation and Conversion to Complex Sphingolipids) Is Not Restricted to HT29 Cells

Conversion of C6-NBD-DH-Cer to C6-NBD-SM and C6-NBD-glycolipids occurred in a variety of cell types (Table I, III, and IV), such as NRK and BHK fibroblasts, HL-60 cells, HepG2 cells, oligodendrocyte-like 36C2.21 cells, and neuro-2A cells. The composition of the C6-NBD-sphingolipid products differed, however, with C6-NBD-SM being the major product in most cell types except HT29 G+ and HL-60 cells, where C6-NBD-GlcCer predominated. The composition of complex C6-NBD-glycolipids was cell type-specific. C6-NBD-CTH (see "Experimental Procedures") was found in substantial amounts only in HT29 G+ cells, which expressed a relatively low level of C6-NBD-LacCer, the presumed precursor for C6-NBD-CTH. In contrast, HL-60 cells, with low amounts of C6-NBD-CTH, express relatively high amounts of C6-NBD-LacCer.

When the degree of desaturation in individual C6-NBD-sphingolipid species was analyzed in HT29 G+ cells, NRK fibroblasts, and HL-60 cells, the extent of desaturation was cell type-dependent, with HT29 G+ cells showing the lowest degree of desaturation (Table I).

C6-NBD-DH-Cer Metabolism Requires Correct Stereochemistry

The presence of the 4-trans double bond is not required for conversion of C6-NBD-Cer to complex sphingolipids, as shown above. C6-NBD-4-D-hydroxy-DH-Cer (Fig. 2A), containing 4-D-hydroxysphinganine (also called "phytosphingosine") as the backbone, functioned as a precursor for the biosynthesis of (glyco)sphingolipids as well (Table III), showing that the 4-OH group does not interfere with head-group addition. However, when HT29 G+ cells were incubated for 24 h at 37 °C with C6-NBD-L,erythro-DH-Cer (see Fig. 2B), essentially none was converted to more complex C6-NBD-(glyco)sphingolipids (Table III). Another C6-NBD-DH-Cer stereoisomer (containing D,threo-sphinganine), also yielded very low amounts of products after a 24-h incubation in HT29 G+ cells. A similar result was obtained when D,erythro- and L,erythro-C6-NBD-Cer were compared (Table III). In conclusion, the enzymes involved in the biosynthesis of GlcCer and SM are highly selective with regard to the stereochemistry of the sphingoid base in (DH-)Cer (cf. Fig. 2B). As shown in Table IV, this stereoselective conversion of C6-NBD-DH-Cer was not limited to HT29 cells, since it was also observed in NRK, BHK, and HL-60 cells.

Table III. C6-NBD-(glyco)sphingolipid biosynthesis from different C6-NBD-ceramide precursors

Approximately 107 HT29 G+ cells were incubated with one from a series of C6-NBD-ceramides, based on various different sphingoid backbones. For the analysis of biosynthetic conversion to (glyco)sphingolipid products, cells were incubated with 2 µM C6-NBD-ceramide for 24 h at 37 °C, followed by lipid extraction and analysis by HPLC, as described under "Experimental Procedures." BFA-treated cells were preincubated with the drug for 45 min, and BFA was present (5 µg/ml) during the entire 24-h incubation with C6-NBD-ceramides. C6-NBD-sphingolipid products synthesized from the C6-NBD-ceramide precursor were pooled and expressed as the percentage of total C6-NBD-lipid. Data are the means of two independent experiments.

C6-NBD-derivatized Total fluorescence in (glyco)sphingolipid products
Control BFA

%
D,erythro-Sphingosine 79.7 87.3
L,erythro-Sphingosine 4.2 8.5
D,erythro-Sphinganine 71.9 87.3
L,erythro-Sphinganine 1.5 2.4
D,threo-Sphinganine 1.4 2.6
4-D-Hydroxysphinganine 86.3 87.9
1-Deoxysphinganine 11.8 10.5
1-Deoxy-5-hydroxysphinganine 8.0 3.5

Table IV. Structure-specific C6-NBD-(glyco)sphingolipid biosynthesis in different cell types

Approximately 107 NRK, BHK, or HL-60 cells or 50 µg of Golgi membranes isolated from rat liver were incubated with one from a series of C6-NBD-ceramides, based on different sphingoid backbones. For the analysis of biosynthetic conversion to (glyco)sphingolipid products, cells were incubated with 2.5 µM C6-NBD-ceramide for 24 h at 37 °C, followed by lipid extraction and analysis by HPTLC, as described under "Experimental Procedures." Golgi membranes were incubated with 4 µM C6-NBD-ceramide and 10 mM uridine 5'-diphosphoglucose for 4 h at 37 °C. C6-NBD-sphingolipid products synthesized from the C6-NBD-ceramide precursor were pooled and expressed as the percentage of total C6-NBD-lipid. Data are the means of two independent experiments.

C6-NBD-derivatized Total fluorescence in (glyco)sphingolipid products
NRK cells BHK cells HL-60 cells Liver Golgi

%
D,erythro-Sphinganine 47.2 48.8 24.4 22.0
L,erythro-Sphinganine 2.7 3.0 0.8 0.6
D,erythro-Sphingosine 70.6 53.8 47.8 41.3
L,erythro-Sphingosine 4.3 6.8 3.5 0.4

Two Cer analogs synthesized from sphingoid bases lacking the -OH group at C1 (C6-NBD-1-deoxy-D,erythro-DH-Cer and C6-NBD-1-deoxy-5-hydroxy-D,erythro-DH-Cer; Fig. 2A) were also analyzed as negative controls in metabolic headgroup addition and transport studies. These 1-deoxy compounds are not converted to complex (glyco)sphingolipids, which can be concluded from 1) the nature of the compounds, lacking the 1-OH group necessary for headgroup addition and 2) TLC analysis confirming the absence of complex (glyco)sphingolipids (data not shown). Some conversion of these derivatives did occur, nonetheless, as shown in Table III. The product was not positively identified, but might represent double bond introduction into the sphinganine backbone.

C6-NBD-DH-Cer Metabolism Is Strongly Correlated with Localization of Fluorescent Sphingolipids at the Golgi Apparatus and Plasma Membrane

C6-NBD-Cer is a well known Golgi marker (14, 15). In accordance with the resemblance between C6-NBD-Cer and C6-NBD-DH-Cer concerning their metabolic fate in cells, C6-NBD-DH-Cer was also translocated to the Golgi apparatus, as can be seen in Fig. 4. After a 1-h incubation at 37 °C, C6-NBD-D,erythro-DH-Cer was found in the Golgi apparatus of HT29 G+ cells (Fig. 4A), very similar to C6-NBD-D,erythro-Cer (Fig. 4B). When the intracellular pool of C6-NBD-sphingolipids, extracted from cells that had been incubated with C6-NBD-D,erythro-DH-Cer for 1 h and subsequently subjected to a back-exchange procedure, was analyzed by HPLC, 45.6% (n = 2) was found as C6-NBD-(glyco)sphingolipid products (C6-NBD-DH-GlcCer and C6-NBD-DH-SM).


Fig. 4. Intracellular localization of different stereoisomers of C6-NBD-(DH-)Cer in HT29 cells. HT29 G+ cells, grown on coverslips, were incubated with one of a series of C6-NBD-ceramides (5 µM) for 1 h at 37 °C, followed by a back-exchange procedure to remove the plasma membrane C6-NBD-lipid pool. When C6-NBD-ceramide was based on the D,erythro isomer of either sphinganine (A) or sphingosine (B), Golgi staining (arrowheads) was observed. In the case of the L-erythro isomer of sphinganine (C) or 1-deoxy-5-hydroxysphinganine (D; see Fig. 2 for structures), no Golgi staining was observed. A control experiment was performed involving co-incubation of cells with both D,erythro- and L,erythro-sphinganine based C6-NBD-DH-Cer (E), showing that Golgi structures (arrowheads) are not disrupted in cells incubated with the L,erythro isomer of C6-NBD-DH-Cer (cf. C). Bar, 10 µm.
[View Larger Version of this Image (84K GIF file)]

Following prolonged incubation at 37 °C (24 h) with either C6-NBD-D,erythro-DH-Cer (Fig. 5A) or C6-NBD-D,erythro-Cer (Fig. 5B), an intense plasma membrane fluorescence was observed. When the plasma membrane pool of C6-NBD-sphingolipids, obtained after a back-exchange performed on cells that were incubated for 24 h with C6-NBD-D,erythro-DH-Cer, was analyzed by HPLC, both saturated and desaturated forms of C6-NBD-sphingolipid species were recovered. 66% (n = 2) of the total cell-associated C6-NBD-DH-GlcCer pool was found in the plasma membrane, comparable to the percentage for C6-NBD-GlcCer (61.5%; n = 2; cf. Ref. 23).


Fig. 5. Staining of the plasma membrane of HT29 cells after incubation with C6-NBD-(DH-)Cer. HT29 G+ cells, grown on coverslips, were incubated with one of a series of C6-NBD-ceramides (2.5 µM) for 24 h at 37 °C. When C6-NBD-ceramide was based on the D,erythro isomer of either sphinganine (A) or sphingosine (B), plasma membrane staining (arrows) was observed. In the case of the L-erythro isomer of sphinganine (C) or 1-deoxy-5-hydroxysphinganine (D; see Fig. 2 for structures), no plasma membrane staining was observed. Note that even after 24 h, Golgi staining did not occur in the cells incubated with C6-NBD-ceramide analogs that are not metabolized (C and D; cf. Tables III and V). In the case of C6-NBD-D,erythro-DH-Cer, the plasma membrane fraction, which could be recovered by a back-exchange procedure, contained both saturated and desaturated forms of C6-NBD-sphingolipid products (see "Results"). Bar, 10 µm.
[View Larger Version of this Image (176K GIF file)]

As shown in Fig. 4C, C6-NBD-L,erythro-DH-Cer did not label the Golgi apparatus after a 1-h incubation at 37 °C, but was, rather, associated with the ER compartment. No C6-NBD-(glyco)sphingolipid products were detectable (n = 2), when the intracellular pool of C6-NBD-sphingolipids, extracted from cells that had been incubated with C6-NBD-L,erythro-DH-Cer for 1 h and subsequently subjected to a back-exchange procedure, was analyzed by HPLC. To certify that C6-NBD-L,erythro-DH-Cer had not disrupted the Golgi structure, HT29 G+ cells were incubated with a mixture of both D,erythro- and L,erythro-C6-NBD-DH-Cer. Fluorescence was apparent in the Golgi compartment and the ER, as shown in Fig. 4E. Therefore, the ability of the D,erythro isomer to reach the Golgi is not blocked by the L-erythro isomer. Prolonged incubation at 37 °C (24 h) with C6-NBD-L,erythro-DH-Cer did not result in plasma membrane staining (Fig. 5C). Similar to C6-NBD-L,erythro-DH-Cer and in accordance with their metabolic fate, C6-NBD-D,threo-DH-Cer and C6-NBD-L,erythro-Cer did not label the Golgi compartment, nor the plasma membrane (Table V).

Table V. C6-NBD-(glyco)sphingolipid biosynthesis and localization in Golgi and plasma membrane after incubation with different C6-NBD-ceramide precursors

HT29 G+ cells were incubated with one from a series of C6-NBD-ceramides, based on various different sphingoid backbones. For the analysis of biosynthetic conversion to (glyco)sphingolipid products, cells were incubated with 2.5 µM C6-NBD-ceramide for 24 h at 37 °C, followed by lipid extraction and analysis by HPLC, as described under "Experimental Procedures." For microscopic visualization of the (intra)cellular staining pattern, cells on coverslips were incubated with 5 or 2.5 µM C6-NBD-ceramide for 1 or 24 h, respectively. For additional documentation of results, see Table III and Figs. 4 and 5.

C6-NBD-derivatized (Glyco)sphingolipid biosynthesis Golgi/plasma membrane staining

D,erythro-Sphingosine + +
L,erythro-Sphingosine  -  -
D,erythro-Sphinganine + +
L,erythro-Sphinganine  -  -
D,threo-Sphinganine  -  -
4-D-hydroxysphinganine + +
1-Deoxysphinganine  -  -
1-Deoxy-5-hydroxy-sphinganine  -  -

The correlation between C6-NBD-(DH-)Cer labeling of the Golgi apparatus and its conversion to (glyco)sphingolipids was further extended by observations on the fate of C6-NBD-derivatized 4-D-hydroxysphinganine and 1-deoxy compounds (Table V). C6-NBD-4-D-hydroxy-DH-Cer (D,erythro) resulted in (glyco)sphingolipid biosynthesis as well as Golgi and plasma membrane staining. On the other hand, C6-NBD-derivatized 1-deoxysphinganine compounds did not yield biosynthetic complex sphingolipid products and did not label the Golgi apparatus and plasma membrane, even after prolonged incubation at 37 °C (Figs. 4D and 5D). The latter result is in agreement with results obtained by Pütz and Schwarzmann (31), using a 1-O-methyl analog of C6-NBD-Cer.

Stereoselectivity of metabolism was observed in four cell types tested: HT29 epithelial cells, NRK and BHK fibroblasts, and HL-60 cells (Tables III and IV). In accordance with these observations, Golgi labeling in NRK cells showed a similar dependence on C6-NBD-ceramide structure; only D,erythro isomers of C6-NBD-DH-Cer (Fig. 6A) and C6-NBD-Cer (Fig. 6C) gave rise to Golgi labeling, in contrast to C6-NBD-L,erythro-DH-Cer (Fig. 6B) and C6-NBD-1-deoxy-5-hydroxy-D,erythro-DH-Cer (Fig. 6D). Incubation of NRK cells with C6-NBD-4-D-hydroxy-DH-Cer again resulted in Golgi staining (Fig. 6E). Similarly, in BHK cells, D,erythro isomers of C6-NBD-DH-Cer (Fig. 7A) and C6-NBD-Cer (Fig. 7B) gave rise to prominent Golgi labeling, while L,erythro isomers of both C6-NBD-lipids did not (Fig. 7, C and D). In accordance with these results, the cell-associated fluorescence after a 1 h incubation consisted almost entirely of C6-NBD-DH-Cer (98.5%; n = 2) or C6-NBD-Cer (97.0%, n = 2) in the case of the L,erythro isomers, while 19.9% (n = 2) of C6-NBD-D,erythro-DH-Cer and 43.9% (n = 2) of C6-NBD-D,erythro-Cer had been converted to (glyco)sphingolipid products.


Fig. 6. Intracellular localization of different stereoisomers of C6-NBD-(DH-)Cer in NRK cells. NRK fibroblasts, grown on coverslips, were incubated with one of a series of C6-NBD-ceramides (5 µM) for 1 h at 37 °C, followed by a back-exchange procedure to remove the plasma membrane C6-NBD-lipid pool. Identical to the results obtained in HT29 cells (Fig. 4), C6-NBD-derivatives of the D,erythro isomers of either sphinganine (A) or sphingosine (C) resulted in Golgi staining (arrowheads), in contrast to L,erythro-sphinganine (B) or 1-deoxy-5-hydroxysphinganine (D). 4-D-Hydroxysphinganine-based C6-NBD-ceramide also resulted in Golgi staining (E; see Fig. 2 for structures). Bar, 10 µm.
[View Larger Version of this Image (152K GIF file)]


Fig. 7. Intracellular localization of different stereoisomers of C6-NBD-(DH-)Cer in BHK cells. BHK fibroblasts, grown on coverslips, were incubated with one of a series of C6-NBD-ceramides (5 µM) for 1 h at 37 °C, followed by a back-exchange procedure to remove the plasma membrane C6-NBD-lipid pool. Similar to the results obtained in HT29 and NRK cells (Figs. 4 and 6), C6-NBD-derivatives of the D,erythro isomers of either sphinganine (A) or sphingosine (B) resulted in Golgi staining (arrowheads), in contrast to L,erythro isomers of sphinganine (C) or sphingosine (D; see Fig. 2 for structures). Bar, 10 µm.
[View Larger Version of this Image (125K GIF file)]

Absence of Metabolic Conversion of (DH-)Cer Isomers Is Not a Consequence of Intracellular Localization

From the foregoing, one might conclude that unnatural stereoisomers of C6-NBD-(DH-)Cer are not metabolized because these isomers cannot reach the Golgi complex. However, this hypothesis was disproven by several types of experiments.

When cells were treated with BFA, the Golgi merged with the ER and all C6-NBD-ceramide analogs should have access to (glyco)sphingolipid synthases. Under these conditions, there remained a clear difference between the metabolic capacity of the various C6-NBD-ceramide analogs (Table III). In addition, when C6-NBD-ceramide isomers were incubated with isolated liver Golgi membranes (Table IV), only the D,erythro isomers of C6-NBD-(DH-)Cer were metabolized to (glyco)sphingolipids. Therefore, one can conclude that accessibility of the C6-NBD-ceramide analogs to the biosynthetic enzymes does not play a role in structure-specific conversion of ceramide to (glyco)sphingolipids.


DISCUSSION

C6-NBD-DH-Cer Is at the Basis of Various Metabolic Pathways; Cell Biological Consequences

Based on the results of this study, we conclude that C6-NBD-DH-Cer can be converted in the Golgi apparatus to various dihydro(glyco)sphingolipids that are transported from the Golgi to the plasma membrane, similar to the (glyco)sphingolipids synthesized from C6-NBD-Cer. The formation of dihydro forms of complex sphingolipids was not an artifact of the use of C6-NBD analogs, because sphingolipids typically contain at least some sphinganine in the backbone. We have found this to be the case for HT29 G+ cells as well, because HPLC analysis of endogenous sphingolipids, as described under "Experimental Procedures," showed that 25% (± 5%; n = 4) of the total sphingoid base was sphinganine, indicating that dihydro-sphingolipids account for this percentage of the total sphingolipid (analysis of the sphingoid base composition of individual sphingolipid species revealed that GlcCer and SM indeed partly contain DH-Cer as the backbone; data not shown). Studies of de novo sphingolipid synthesis in J774 cells using [14C]serine (32) have also found that DH-Cer initially constitutes a substantial portion of the label in complex sphingolipids, but over time the backbones became primarily Cer. Therefore, we conclude that relatively little DH-Cer is converted to Cer during the initial synthesis of sphingolipids de novo.

Apparently, saturated (glyco)sphingolipids are enzymatically hydrolyzed to dihydroceramides, which can in turn be converted to ceramides by double bond addition. As shown in this study, this appears to be the exclusive pathway for metabolic conversion of DH-GlcCer to GlcCer, since direct desaturation of DH-GlcCer was not observed (Fig. 3). This observation is corroborated by in vitro assays of dihydroceramide desaturase activity, showing that N-octanoyl-DH-GlcCer was not metabolized, in contrast to N-octanoyl-derivatives of DH-Cer and (at a lower rate) DH-SM.2 The presence of a double bond in the Cer backbone of sphingolipids is apparently a function of several processes: the rate of de novo DH-Cer synthesis, the rate at which DH-(glyco)sphingolipids are turned over to DH-Cer, and the efficiency of DH-Cer desaturation prior to (re)utilization for complex sphingolipid synthesis. The rationale for this seemingly circuitous pathway may be to minimize the accumulation of Cer as an intermediate of the de novo biosynthetic pathway, presumably because Cer is more potent than DH-Cer in triggering diverse cell responses, including apoptosis. Indeed, our results (Table I) show that in all cell types studied upon incubation with C6-NBD-DH-Cer, low amounts of C6-NBD-Cer were found.

Structural Requirements for Metabolic Conversion of Ceramides

In our studies in four different cell types, the natural stereoisomer of the sphingoid base backbone structure was required for metabolism. This structural selectivity resided at the level of the biosynthetic enzymes, rather than different intracellular localization of enzyme and substrate (C6-NBD-(DH-)ceramide). As discussed above, when isolated liver Golgi membranes were incubated with D or L isomers of C6-NBD-erythro-(DH-)Cer, metabolism to C6-NBD-(DH-)GlcCer and C6-NBD-(DH-)SM was stereoselective (Table IV). Our results do not fully correspond to those of Pagano and Martin (24), who studied metabolism (and intracellular localization) of various stereoisomers of C6-NBD-(DH-)Cer in BHK and V79 fibroblasts. They observed conversion to SM of all four stereoisomers of C6-NBD-Cer, as well as D,erythro and L,threo isomers of C6-NBD-DH-Cer. On the other hand, both D- and L-erythro isomers of C6-NBD-Cer were converted to GlcCer in their studies, while all other C6-NBD-(DH-)Cer stereoisomers tested, including C6-NBD-D,erythro-DH-Cer did not lead to GlcCer formation. Our results obtained after 24-h incubations with different stereoisomers in BHK cells are similar to those obtained in other cell types, showing little metabolism of L,erythro isomers and extensive conversion of D,erythro isomers (Tables III and IV). D,erythro isomers were converted mostly to SM (41.5% (n = 2) for C6-NBD-D,erythro-Cer; 41.9% (n = 2) for C6-NBD-D,erythro-DH-Cer; cf. Table IV), but also to GlcCer. The residual metabolism of L,erythro isomers in BHK cells was mostly to SM as well (6.1% (n = 2) for C6-NBD-L,erythro-Cer; 2.5% (n = 2) for C6-NBD-L,erythro-DH-Cer; cf. Table IV). In contrast, in HT29 G+ cells residual formation of GlcCer occurred most prominently (also in the case of the D,threo isomer of C6-NBD-DH-Cer; data not shown), while the D,erythro isomers gave rise to extensive formation of this glycolipid (Table I and Ref. 23). Thus, the fate of the L-erythro isomers is a good reflection of the preferred metabolic route of the D-erythro isomers in both cell types. In conclusion, in our opinion, D,erythro isomers are good substrates for both GlcCer (prominent in HT29) and SM biosynthesis (prominent in fibroblasts, such as NRK and BHK), while unnatural stereoisomers are generally poor substrates, but not specifically for either GlcCer or SM biosynthesis.

Unlike unnatural stereochemistry, the absence of the 4-trans double bond (as in C6-NBD-DH-Cer) or the presence of an additional -OH group at the C4 position of the sphingoid backbone (as in C6-NBD-4-D-hydroxy-DH-Cer) did not interfere with headgroup addition. The SM synthase and glycolipid synthases, such as GlcCer, GalCer, and LacCer synthases, apparently recognize C6-NBD-derivatives of all three, naturally occurring sphingoid backbones.

Metabolism of C6-NBD-(DH-)ceramide Is Correlated with Intracellular Localization at the Golgi Apparatus

As shown in Tables III, IV, V and Figs. 4, 5, 6, 7, there is a clear correlation between the capacity of a fluorescent ceramide analog to be metabolized to complex sphingolipids and its intracellular localization at the Golgi apparatus. The question that remains is how the two phenomena are related. There are three possibilities, one of which can be ruled out by experimental evidence presented in this paper. 1) Metabolism depends on whether the C6-NBD-ceramide analog reaches the Golgi apparatus. This was disproven by BFA experiments (Table III) and metabolic studies in isolated Golgi membranes (Table IV), as discussed above. 2) Golgi localization is a consequence of metabolic trapping of ceramides. According to this hypothesis, the lack of accumulation of C6-NBD-L,erythro-Cer in the Golgi is due to its inability to be metabolized (cf. Ref. 31). Unfortunately, we were not able to completely block C6-NBD-D,erythro-(DH-)Cer metabolism to test this hypothesis (PDMP efficiently blocks GlcCer biosynthesis, but SM (and GalCer) formation are not inhibited). 3) Metabolic conversion and transport of (DH-)Cer to the Golgi are both selective for the correct (DH-)Cer stereoisomer. According to this hypothesis, C6-NBD-(DH-)Cer is specifically transported to and/or retained in the Golgi apparatus. A specific interaction with factors governing transport to or retention in the Golgi apparatus would not occur for unnatural isomers of C6-NBD-(DH-)Cer or C6-NBD-(DH-)Cer analogs with modifications of the -OH group at the C1 position of the sphingoid base. Future studies will focus on the molecular mechanisms involved in the transport of Cer from the ER compartment to the Golgi apparatus.


FOOTNOTES

*   The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The syntheses of the sphingoid base stereoisomers and analogs were conducted under Grant GM 46368 (to A. H. M.) from the National Institutes of Health.
§   Supported by a fellowship of the Royal Netherlands Academy of Arts and Sciences and by the Netherlands Organization for Scientific Research. To whom all correspondence should be addressed. Tel.: 31-50-3632725; Fax: 31-50-3632728; E-mail: j.w.kok{at}med.rug.nl.
1   The abbreviations used are: Cer, ceramide; BFA, brefeldin A; C6-NBD, 6-[N-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]hexanoyl or hexanoic acid; CTH, ceramide trihexoside; DH-Cer, dihydroceramide; DMEM, Dulbecco's modified Eagle's medium; ER, endoplasmic reticulum; FCS, fetal calf serum; GalCer, galactosylceramide; GlcCer, glucosylceramide; HPTLC, high performance thin layer chromatography; HPLC, high pressure liquid chromatography; LacCer, lactosylceramide; PDMP, DL-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol; SL-O, streptolysin O; SM, sphingomyelin.
2   Michel, C., van Echten-Deckert, G., Rother, J., Sandhoff, K., Wang, E., and Merrill, A. M., Jr. (1997) J. Biol. Chem. 272, in press.

REFERENCES

  1. Hannun, Y. A. (1994) J. Biol. Chem. 269, 3125-3128 [Free Full Text]
  2. Mathias, S., and Kolesnick, R. (1993) Adv. Lipid Res. 25, 65-90 [Medline] [Order article via Infotrieve]
  3. Bielawska, A., Crane, H. M., Liotta, D., Obeid, L. M., and Hannun, Y. A. (1993) J. Biol. Chem. 268, 26226-26232 [Abstract/Free Full Text]
  4. Merrill, A. H., Jr., and Wang, E. (1986) J. Biol. Chem. 261, 3764-3769 [Abstract/Free Full Text]
  5. Rother, J., van Echten, G., Schwarzmann, G., and Sandhoff, K. (1992) Biochem. Biophys. Res. Commun. 189, 14-20 [Medline] [Order article via Infotrieve]
  6. Mandon, E. C., Ehses, I., Rother, J., van Echten, G., and Sandhoff, K. (1992) J. Biol. Chem. 267, 11144-11148 [Abstract/Free Full Text]
  7. Hirschberg, K., Rodger, J., and Futerman, A. H. (1993) Biochem. J. 290, 751-757 [Medline] [Order article via Infotrieve]
  8. Futerman, A. H., Stieger, B., Hubbard, A. L., and Pagano, R. E. (1990) J. Biol. Chem. 265, 8650-8657 [Abstract/Free Full Text]
  9. Futerman, A. H., and Pagano, R. E. (1991) Biochem. J. 280, 295-302 [Medline] [Order article via Infotrieve]
  10. Trinchera, M., Fabbri, M., and Ghidoni, R. (1991) J. Biol. Chem. 266, 20907-20912 [Abstract/Free Full Text]
  11. Jeckel, D., Karrenbauer, A., Burger, K. J. N., van Meer, G., and Wieland, F. (1992) J. Cell Biol. 117, 259-267 [Abstract]
  12. Coste, H., Martel, M. B., and Got, R. (1986) Biochim. Biophys. Acta 858, 6-12 [Medline] [Order article via Infotrieve]
  13. Hoekstra, D., and Kok, J. W. (1992) Biochim. Biophys. Acta 1113, 277-294 [Medline] [Order article via Infotrieve]
  14. Lipsky, N. G., and Pagano, R. E. (1985) Science 228, 745-747 [Medline] [Order article via Infotrieve]
  15. Pagano, R. E. (1989) Methods Cell Biol. 29, 75-85 [Medline] [Order article via Infotrieve]
  16. Lipsky, N. G., and Pagano, R. E. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 2608-2612 [Abstract]
  17. Lipsky, N. G., and Pagano, R. E. (1985) J. Cell Biol. 100, 27-34 [Abstract]
  18. Babia, T., Kok, J. W., van der Haar, M., Kalicharan, R., and Hoekstra, D. (1994) Eur. J. Cell Biol. 63, 172-181 [Medline] [Order article via Infotrieve]
  19. Nimkar, S., Menaldino, D., Merrill, A. H., and Liotta, D. C. (1988) Tetrahedron Lett. 29, 3037-3040 [CrossRef]
  20. Kok, J. W., and Hoekstra, D. (1993) in Fluorescent Probes for Biological Function of Living Cells: A Practical Guide (Mason, W. T., and Relf, G., eds), pp. 100-119, Academic Press, London
  21. Kok, J. W., Hoekstra, K., Eskelinen, S., and Hoekstra, D. (1992) J. Cell Sci. 103, 1139-1152 [Abstract/Free Full Text]
  22. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917
  23. Babia, T., Kok, J. W., Hulstaert, C., de Weerd, H., and Hoekstra, D. (1993) Int. J. Cancer 54, 1-7 [Medline] [Order article via Infotrieve]
  24. Pagano, R. E., and Martin, O. C. (1988) Biochemistry 27, 4439-4445 [Medline] [Order article via Infotrieve]
  25. Merrill, A. H., Jr., Wang, E., Mullins, R. E., Jamison, W. C. L., Nimkar, S., and Liotta, D. C. (1988) Anal. Biochem. 171, 373-381 [Medline] [Order article via Infotrieve]
  26. Tan, A., Bolscher, J., and Ploegh, H. (1992) J. Cell Biol. 116, 1357-1367 [Abstract]
  27. Croze, E. M., and Morré, D. J. (1984) J. Cell. Physiol. 119, 46-57 [Medline] [Order article via Infotrieve]
  28. Kok, J. W., Babia, T., Klappe, K., and Hoekstra, D. (1995) Biochem. J. 309, 905-912 [Medline] [Order article via Infotrieve]
  29. Peterson, G. L. (1977) Anal. Biochem. 83, 346-356 [Medline] [Order article via Infotrieve]
  30. van Meer, G., Stelzer, E. H., Wijnandts-van-Resandt, R. W., and Simons, K. (1987) J. Cell Biol. 105, 1623-1635 [Abstract]
  31. Pütz, U., and Schwarzmann, G (1995) Eur. J. Cell Biol. 68, 113-121 [Medline] [Order article via Infotrieve]
  32. Smith, E. R., and Merrill, A. H., Jr. (1995) J. Biol. Chem. 270, 18749-18758 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.