Biosynthesis of phosphatidylcholine from a phosphocholine precursor pool derived from the late endosomal/lysosomal degradation of sphingomyelin.

Previous studies suggest that the steps of the CDP- choline pathway of phosphatidylcholine synthesis are tightly linked in a so-called metabolon. Evidence has been presented that only choline that enters cells through the choline transporter, and not phosphocholine administered to cells by membrane permeabilization, is incorporated into phosphatidylcholine. Here, we show that [(14)C]phosphocholine derived from the lysosomal degradation of [(14)C]choline-labeled sphingomyelin is incorporated as such into phosphatidylcholine in human and mouse fibroblasts. Low density lipoprotein receptor-mediated endocytosis was used to specifically direct [(14)C]sphingomyelin to the lysosomal degradation pathway. Free labeled choline was not found either intracellularly or in the medium, not even when the cells were energy-depleted. Deficiency of lysosomal acid phosphatases in mouse or alkaline phosphatase in human fibroblasts did not affect the incorporation of lysosomal [(14)C]sphingomyelin-derived [(14)C]phosphocholine into phosphatidylcholine, supporting our finding that phosphocholine is not degraded to choline prior to its incorporation into phosphatidylcholine. Inhibition studies and analysis of molecular species showed that exogenous [(3)H]choline and sphingomyelin-derived [(14)C]phosphocholine are incorporated into phosphatidylcholine via a common pathway of synthesis. Our findings provide evidence that, in fibroblasts, phosphocholine derived from sphingomyelin is transported out of the lysosome and subsequently incorporated into phosphatidylcholine without prior hydrolysis of phosphocholine to choline. The findings do not support the existence of a phosphatidylcholine synthesis metabolon in fibroblasts.

Previous studies suggest that the steps of the CDPcholine pathway of phosphatidylcholine synthesis are tightly linked in a so-called metabolon. Evidence has been presented that only choline that enters cells through the choline transporter, and not phosphocholine administered to cells by membrane permeabilization, is incorporated into phosphatidylcholine. Here, we show that [ 14 C]phosphocholine derived from the lysosomal degradation of [ 14 C]choline-labeled sphingomyelin is incorporated as such into phosphatidylcholine in human and mouse fibroblasts. Low density lipoprotein receptor-mediated endocytosis was used to specifically direct [ 14 C]sphingomyelin to the lysosomal degradation pathway. Free labeled choline was not found either intracellularly or in the medium, not even when the cells were energy-depleted. Deficiency of lysosomal acid phosphatases in mouse or alkaline phosphatase in human fibroblasts did not affect the incorporation of lysosomal [ 14 C]sphingomyelin-derived [ 14 C]phosphocholine into phosphatidylcholine, supporting our finding that phosphocholine is not degraded to choline prior to its incorporation into phosphatidylcholine. Inhibition studies and analysis of molecular species showed that exogenous [ 3 H]choline and sphingomyelin-derived [ 14 C]phosphocholine are incorporated into phosphatidylcholine via a common pathway of synthesis. Our findings provide evidence that, in fibroblasts, phosphocholine derived from sphingomyelin is transported out of the lysosome and subsequently incorporated into phosphatidylcholine without prior hydrolysis of phosphocholine to choline. The findings do not support the existence of a phosphatidylcholine synthesis metabolon in fibroblasts.
De novo synthesis of phosphatidylcholine (PC) 1 predomi-nantly occurs via the CDP-choline pathway (1). Choline can enter the cell via the ATP-dependent choline transporter present in the cell membrane and is subsequently phosphorylated to phosphocholine (PCho) by the cytosolic enzyme choline kinase (EC 2.7.1.32). PC is then synthesized in two steps, catalyzed by the rate-limiting cytosolic and membrane-bound enzyme CTP:phosphocholine cytidylyltransferase (CT; EC 2.7.7.15) and the membrane-bound enzyme CDP-choline:sn-1,2-diacylglycerol cholinephosphotransferase (EC 2.7.8.2). For some cell types, including fibroblasts, it has been reported that the steps of PC synthesis are tightly linked in a so-called metabolon (2)(3)(4). In these cell types only choline that entered the cells through the choline transporter was incorporated into PC. In contrast, PCho administered to cells by partial membrane permeabilization was not incorporated into PC (2,4).
A possible endogenous source of PCho and choline is provided by the lysosomal degradation of PC and sphingomyelin (SM) by acid sphingomyelinase (5). We hypothesized that if only choline, and not PCho, can enter the PC metabolon and be incorporated into PC in human fibroblasts, then PCho that is formed during the degradation of SM by acid sphingomyelinase in the lysosome must be degraded to choline to be incorporated into PC. Candidate enzymes for intralysosomal hydrolysis of PCho to choline are lysosomal acid phosphatase (LAP) and tartrate-resistant acid phosphatase (TRAP) (6 -8). Their in vivo substrates and functional roles are unclear. LAP is ubiquitously expressed, and TRAP is predominantly expressed in alveolar macrophages and osteoclasts (9), suggesting a specific function of the latter enzyme in these cell types. Another possible route for lysosomal SM-derived PCho is transport across the lysosomal membrane (by an unknown transporter), followed by hydrolysis to choline and subsequent incorporation of choline into the PC synthesis metabolon. A candidate enzyme for the hydrolysis of cytosolic PCho is alkaline phosphatase. In human skin fibroblasts, alkaline phosphatase is predominantly located at the plasma membrane (10,11). Phosphoethanolamine and pyridoxal-5Ј-phosphate have been described as physiological substrates for this enzyme (10). We have investigated the possible role of LAP, TRAP, and alkaline phosphatase in the hydrolysis of PCho derived from the late endosomal/lysosomal degradation of SM.
In this study we compared exogenous choline and late endosome/lysosome-derived PCho as substrates for PC synthesis. Hexadecylphosphocholine (HePC) was used as an inhibitor of PC synthesis (12)(13)(14)(15)(16). The results indicate that PCho derived from lysosomal SM is not degraded to choline but is incorporated into PC as such. We conclude that the PC synthesis metabolon either does not exist in fibroblasts or is not as inaccessible to PCho as indicated by previous studies.
Cell Culture-Human fibroblasts from three healthy subjects and from hypophosphatasia (alkaline phosphatase deficiency), familial hypercholesterolemia (low density lipoprotein (LDL) receptor deficiency), and Niemann-Pick A (NP-A) (acid sphingomyelinase deficiency) patients were cultured as described previously (5). Primary embryonic (embryonic day 12.5) fibroblasts from LAP-deficient (17), TRAP-deficient (18), LAP/TRAP double-deficient, and control mice were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FBS, penicillin (10 IU/ml), and streptomycin (10 IU/ml). LAP/TRAP doubledeficient mice are viable but more severely affected than single-deficient mice. 2 [ 14 C]SM Incubations-Cells were plated in 6-well plates (1.5-2.0 ϫ 10 5 cells/well) and cultured for 4 -5 days until confluence. They were then starved on Ham's F-10 or Dulbecco's modified Eagle's medium with 10% lipoprotein-deficient serum for 24 h to enhance LDL receptor expression. Subsequently, the medium was replaced with medium containing 0.8 nmol of [ 14 C]SM (dissolved in ethanol; final concentration, 0.2%, v/v) in complex with apoE (5 g/ml) ([ 14 C]SM⅐apoE). Cells were incubated for 30 min at 37°C, then the medium was replaced by medium (10% FBS) without [ 14 C]SM⅐apoE, and cells were kept at 18°C for 30 min to enable internalization of the lipid⅐apoE complex. Medium was replaced again with fresh medium (10% FBS) for chase studies at 37°C for various periods of time. In some experiments 50 M HePC was present during the chase. For energy depletion, 10 mM NaN 3 and 10 mM 2-deoxyglucose were dissolved in the chase medium. After incubations cells were trypsinized and pelleted.
[ 3 H]Choline Labeling-Fibroblasts were plated as described above and incubated with 2 Ci of [ 3 H]choline per well (dissolved in ethanol) in medium (10% FBS) for 2 h. In some experiments 50 M HePC or 10 mM NaN 3 ϩ 10 mM 2-deoxyglucose was present during the incubation.
Digitonin Permeabilization-Digitonin permeabilization was done according to Wanders et al. (19), with slight modifications. Briefly, cells were incubated for 30 min with [ 14 C]SM⅐apoE, washed, trypsinized and taken up in buffer (final pH 7.4) containing 150 mM KCl, 25 mM Tris-HCl, 2 mM EDTA, 10 mM KH 2 PO 4 , 0.1% (w/v) bovine serum albumin, and digitonin at different concentrations. After a 5-min incubation at 37°C, the cells were pelleted by centrifugation. Hexosaminidase activity was measured in the pellets and the supernatant to determine the intactness of the lysosomal membrane (20). The protein concentration was determined according to Lowry et al. (21).
Determination of Degradation Products-The lipids in the pellets, supernatant, or medium were separated from aqueous metabolites with Folch extraction (22). The radiolabeled lipids were separated by HPTLC (5), and aqueous metabolites were separated by ion-exchange chromatography (23). Quantification of each radiolabeled product was achieved by liquid scintillation counting in 10 ml of Ultima Gold (Packard Instrument Co.). 14 C]SM⅐apoE as described above. The fibroblast pellets were homogenized, and lipids were extracted with chloroform:methanol (2:1; v/v) in the presence of 0.01 mM butylated hydroxytoluene and washed according to Folch et al. (22). The lower phase was dried under N 2 , dissolved in 100 l of chloroform:methanol (2:1; v/v), and PC was purified by preparative TLC as described (24). The PC molecular species were then separated by reverse-phase HPLC (25), and their radioactivity was determined by liquid scintillation counting. The identity of the major species was determined based on the retention time of standards, as well as by electrospray mass spectrometry, as described elsewhere (26).  (Fig. 1). During the pulse period 36% of total intracellular SM was degraded to PCho. During the initial phase of the chase, production of labeled PCho was prominent. Thereafter, the amount of radiolabeled PCho gradually decreased followed by the appearance of labeled PC. At all time points the amount of radioactivity in the medium was negligible. The recovery of total radioactivity for all cell lines at each time point was 98.7 Ϯ 3.2% as compared with t ϭ 0. Free choline was not detected during pulse or chase, neither intracellularly nor in the medium.

Validation of the Specific Uptake of [ 14 C]SM⅐apoE by the LDL Receptor and Its Lysosomal Degradation-To
Role of Acid Phosphatases in Incorporation of SM-derived PCho into PC-To determine the role of LAP and TRAP in the incorporation of lysosomal [ 14 C]SM-derived PCho into PC, primary embryonic fibroblasts from normal, LAP-deficient, TRAPdeficient, or LAP/TRAP double-deficient mice were pulsed with [ 14 C]SM⅐apoE and subsequently chased. In each deficient cell type, [ 14 C]SM was similarly degraded compared with cells from normal mice (Fig. 2A). The degradation of [ 14 C]SM produced [ 14 C]PCho in all cell types (Fig. 2B). No labeled free choline was detected. Furthermore, labeled PCho was incorporated into PC in the deficient cells at rates comparable with or even faster than in normal cells (Fig. 2C). Thus, in mouse fibroblasts, acid phosphatase activity was not required for the incorporation of lysosomal [ 14 C]SM-derived PCho into PC.
Role of Alkaline Phosphatase in Incorporation of SM-derived PCho into PC-The role of alkaline phosphatase in the incorporation of lysosomal [ 14 C]SM-derived PCho into PC was determined with alkaline phosphatase-deficient fibroblasts. These cells were pulsed with [ 14 C]SM⅐apoE and subsequently chased. [ 14 C]SM was degraded to PCho, which was then incorporated into PC (Fig. 3). Labeled free choline was found neither in the cellular fraction nor in the medium. These results were similar to the degradation of [ 14 C]SM in normal fibroblasts (Fig. 1). This indicated that alkaline phosphatase did not play a role in the incorporation of lysosomal [ 14 C]SM-derived PCho into PC.

Effect of HePC and Energy Depletion on the Incorporation of Exogenous Choline and SM-derived PCho into PC-Adminis-
tration of radiolabeled choline to normal fibroblasts via the culture medium resulted in its incorporation in intracellular PCho and PC (Fig. 4). The presence of the PC synthesis inhibitor HePC caused accumulation of PCho and decreased PC synthesis, in agreement with the reported specific inhibition by HePC of CT, the rate-limiting enzyme of PC synthesis (12)(13)(14)16). Energy depletion, accomplished by the addition of NaN 3 and 2-deoxyglucose, caused accumulation of intracellular choline and decreased the amounts of labeled PCho and PC (Fig.  4). The total choline uptake in the presence of NaN 3 and 2-deoxyglucose was ϳ25% of the uptake in the absence of these additives (with or without HePC), showing that choline uptake is energy-dependent.
Addition of HePC after the incubation with [ 14 C]SM⅐apoE had no effect on the degradation of SM, but it inhibited PC synthesis from the endogenous PCho (Fig. 5). Energy depletion by the addition of NaN 3 and 2-deoxyglucose during the chase resulted in slower degradation of SM and slower synthesis of labeled PC (Fig. 6). Thus, the incorporation of SM-derived PCho into PC is energy-dependent. Under conditions of energy depletion labeled free choline was detected neither intracellularly nor in the medium.  treated with increasing concentrations of digitonin to selectively permeabilize the plasma membrane. Leakage of hexosaminidase activity to the permeabilization buffer was used as an indicator of the integrity of the lysosomal membrane. At low digitonin concentrations (5-10 g/ml), ϳ88% of [ 14 C]PCho, but only a small percent of the cellular hexosaminidase, was found in the permeabilization buffer (Fig. 7). Increasing the digitonin concentration above 10 g/ml markedly enhanced the release of hexosaminidase, and it was complete at a concentration of 160 g/ml digitonin. These data indicate that most of the [ 14 C]SMderived PCho was present in the cytoplasm. Notably, the addition of 5 g/ml digitonin or more caused leakage of cytoplasmic proteins to the permeabilization buffer (40 -60% of total cell-associated protein). This indicates that the absence of hexosaminidase activity in the permeabilization buffer of cells treated with digitonin concentrations of 10 g/ml or less was indeed because of intactness of the lysosomal membrane, rather than inability of the hexosaminidase molecules to pass through the plasma membrane.

Analysis of the Molecular Species of PC, Synthesized from [ 3 H]Choline and [ 14 C]SM-To investigate further whether SM-
derived PCho could enter the PC synthesis metabolon, the molecular species of PC synthesized from [ 3 H]choline or [ 14 C]SM were analyzed. As can be seen in Fig. 8, the distribution of the radioactivity among PC species synthesized from both precursors is nearly identical, suggesting a common biosynthetic pathway. DISCUSSION In the present study we describe the incorporation of labeled PCho derived from the late endosomal/lysosomal degradation of SM into PC via the CDP-choline pathway of PC synthesis (1). We have shown that the uptake of [ 14 C]SM⅐apoE was LDL receptor-mediated and that [ 14 C]SM was exclusively metabolized in the late endosomal/lysosomal compartment, because NP-A fibroblasts, which are deficient in acid sphingomyelinase, did not metabolize [ 14 C]SM. Also, in this cell line no evidence was found for the incorporation of SM-derived [ 14 C]choline into PC, even after prolonged incubation. This finding excludes the In these in situ experiments, we did not observe production of radiolabeled free choline from SM, indicating that SM-derived PCho was not hydrolyzed to choline prior to its incorporation into PC. Even under conditions of energy depletion, when resynthesis of PCho from choline would be inhibited, no radiolabeled free choline was detected. These results corroborate our previous in vitro studies with fibroblast homogenates at acidic pH, where we found PCho, and not choline, as an end product of PC degradation. PCho, which was predominantly generated by acid sphingomyelinase activity, was apparently not further degraded by acid phosphatases (5). Furthermore, in another in vitro study the incubation of purified rat liver lysosomes with PC or PCho at acidic pH levels produced virtually no free choline (28).
The possible role of acid phosphatases in the degradation of PCho in the present in situ study was investigated with embryonic fibroblasts from LAP and/or TRAP (double)-deficient mice. These mice have been developed to study the physiological role of the acid phosphatases (17,18). 2 The natural substrates for these enzymes have not been elucidated. However, an early study reported that PCho was a poor substrate for a purified rat liver acid phosphatase in vitro (29). Our study shows that PCho is not an in situ substrate for acid phosphatases present in mouse fibroblasts.
Because we have not found any indication that SM-derived PCho was hydrolyzed to free choline, it is likely that PCho is an end product of lysosomal degradation. However, most of the intracellular SM-derived PCho was present in the cytoplasm, not in the lysosomes, as was shown by digitonin-induced membrane permeabilization. This suggests that PCho must be transported out of the lysosomes by an as yet unknown transporter. Subsequently, it can serve as a substrate for PC synthesis. The experiments with alkaline phosphatase-deficient human fibroblasts indicate that PCho is not hydrolyzed in the cytosol by alkaline phosphatase. In conclusion, all results indicate that PCho is not hydrolyzed prior to its incorporation into PC.
Our data point to a direct incorporation of cytosolic PCho, endogenously produced by the late endosomal/lysosomal breakdown of SM, into PC. This is not compatible with previous reports that concluded that in C6 rat glioma cells, C 3 H 10 T 1 ⁄2 fibroblasts, and rat hepatocytes the intermediates of PC synthesis are channeled to the enzymes involved, defining a socalled PC synthesis metabolon (2,4). This conclusion was based on experiments in which PCho, which was delivered to the cytosol via partial membrane permeabilization, was not used for PC synthesis. However, our results show that PCho, which is derived from the lysosomal degradation of SM, is directly incorporated into PC via the CDP-choline pathway of PC synthesis. The latter is supported by the fact that HePC, a specific inhibitor of this pathway, inhibited the incorporation of both exogenous choline and endogenous PCho into PC. HePC has been shown by several groups to inhibit PC synthesis in different cell types (12)(13)(14)(15)(16), including human breast fibroblasts (30). HePC inhibits CT, the rate-limiting enzyme of PC synthesis, which results in the accumulation of PCho. In those previous studies with HePC, the incorporation of exogenous choline into PC was investigated. In the present study with human skin fibroblasts, the inhibition of PC synthesis from exogenous radiolabeled choline by HePC and the subsequent accumulation of PCho were confirmed. HePC also inhibited PC synthesis from lysosomal SM-derived PCho in these cells, suggesting that this PCho, like exogenous choline, entered the CDP-choline pathway of PC synthesis that is regulated by CT. The fact that the molecular species of PC synthesized from exogenous [ 3 H]choline and of PC synthesized from late endosomal/lysosomal [ 14 C]PCho were nearly identical indicates a common diacylglycerol precursor pool. This supports the idea that [ 14 C]SMderived PCho enters the normal PC biosynthetic pathway, rather than incorporating to PC via another route.
In conclusion, our results clearly indicate that in human skin fibroblasts, PCho that is produced by acid sphingomyelinase activity is transported intact from the late endosomal/lysosomal compartment to the cytoplasm and then enters the CDPcholine pathway of PC synthesis. This means that if the proposed PC synthesis metabolon exists, it is not as impermeable to intermediates of PC synthesis as indicated by previous studies (2,4).