Direct interaction, instrumental for signaling processes, between LacCer and Lyn in the lipid rafts of neutrophil-like cells.

Lactosylceramide [LacCer; β-Gal-(1-4)-β-Glc-(1-1)-Cer] has been shown to contain very long fatty acids that specifically modulate neutrophil properties. The interactions between LacCer and proteins and their role in cell signaling processes were assessed by synthesizing two molecular species of azide-photoactivable tritium-labeled LacCer having acyl chains of different lengths. The lengths of the two acyl chains corresponded to those of a short/medium and very long fatty acid, comparable to the lengths of stearic and lignoceric acids, respectively. These derivatives, designated C18-[(3)H]LacCer-(N3) and C24-[(3)H]LacCer-(N3), were incorporated into the lipid rafts of plasma membranes of neutrophilic differentiated HL-60 (D-HL-60) cells. C24-[(3)H]LacCer-(N3), but not C18-[(3)H]LacCer-(N3), induced the phosphorylation of Lyn and promoted phagocytosis. Incorporation of C24-[(3)H]LacCer-(N3) into plasma membranes, followed by illumination, resulted in the formation of several tritium-labeled LacCer-protein complexes, including the LacCer-Lyn complex, into plasma membrane lipid rafts. Administration of C18-[(3)H]LacCer-(N3) to cells, however, did not result in the formation of the LacCer-Lyn complex. These results suggest that LacCer derivatives mimic the biological properties of natural LacCer species and can be utilized as tools to study LacCer-protein interactions, and confirm a specific direct interaction between LacCer species containing very long fatty acids, and Lyn protein, associated with the cytoplasmic layer via myristic/palmitic chains.

Under illumination, nitrophenylazide generates the highly reactive nitrene moiety, which can covalently bind to nearby molecules.
HL-60 cells were maintained in culture in RPMI-1640 medium supplemented with 10% FBS. To induce differentiation into D-HL-60 cells, HL-60 cells were cultured for 8 days in the presence of 1.3% DMSO, with differentiation confi rmed by CD11b expression using fl ow cytometry analysis ( 21,22 ).
Synthesis of 18-amino-octadecanoic acid (see Fig. 2). To 8 ml of a suspension of 0.75 M octadecanedioic acid (compound 7 ) in CH 3 OH was added 104 l of 98% H 2 SO 4 , and the reaction mixture was stirred overnight under refl ux. The mixture was concentrated under vacuum, diluted with ethyl ether, and washed three times with saturated Na 2 CO 3 solution. The organic phase was dried over Na 2 SO 4 end evaporated under reduced pressure, resulting in a 97% yield of compound 8 ( 23 ) .
Compound 8 (5.8 mmol) was added to a solution of 0.07 M Ba(OH) 2 in dry CH 3 OH (85 ml) and the mixture was stirred for 24 h at 40°C and fi ltered under reduced pressure. The precipitate was washed with ethyl ether and dissolved in water (H 2 O). The solution was acidifi ed with HCl to pH 3, and extracted twice with CH 2 Cl 2 . The solvent was evaporated under vacuum and the residue purifi ed by fl ash chromatography and eluted with an 8:2 mixture (v/v) of petroleum ether:ethyl acetate, resulting in compound 9 , with a 52% yield. BH 3 /THF complex (10 ml, 10 mmol, 1.0 M solution in THF) was added drop wise over 20 min to a solution of compound 9 (3.7 mmol) in anhydrous THF (20 ml) at Ϫ 20°C. The solution was stirred for 10 min, allowed to warm to room temperature, stirred for 24 h, and quenched with H 2 O (50 ml) at 0°C. One gram of solid K 2 CO 3 was added after 10 min and the mixture was extracted with ethyl ether (3 × 50 ml). The combined organic layers were dried over Na 2 SO 4 and concentrated under reduced pressure, resulting in compound 10 , with a yield of 81% ( 24 ).
To a solution of compound 10 (3 mmol) in dry CH 2 Cl 2 (15 ml) was added 850 l of triethylamine (Et 3 N), followed by the dropwise addition of methane sulfonyl chloride (MsCl; 4.2 mmol, C24:0 and C24:1 fatty acids, or C24-LacCer, are components of plasma membrane lipid rafts in human neutrophils and act as pattern recognition receptors responsible for chemotaxis, phagocytosis, and superoxide generation ( 5,7,8 ). These functions are highly dependent on the Src family kinase, Lyn. The interaction of microorganisms with LacCer activates Lyn, resulting in its phosphorylation, a reaction considered the fi rst step in the Lyn-mediated immunological functions of human neutrophils ( 9 ). These events can be easily reproduced experimentally by treating neutrophils with anti-LacCer antibodies.
The LacCer contents in lipid rafts of human promyelocytic leukemia HL-60 cells induced to differentiate into neutrophilic cells by DMSO are similar to those of human neutrophils ( 7 ). However, the LacCer on plasma membranes of these DMSO-treated human promyelocytic leukemia cells (D-HL-60 cells) is more enriched in shorter fatty acids than the LacCer of neutrophils ( 7 ). Treatment of D-HL-60 cells with formyl peptide (fMLP; N -formylmethionine-leucinephenylalanine), but not anti-LacCer monoclonal antibody, was found to induce Lyn phosphorylation and chemotactic and superoxide-generating activities. Nevertheless, treatment of D-HL-60 cells, which consisted mainly of C16:0-LacCer and had quite low contents of C24-fatty acid chains, with anti-LacCer antibodies induced chemotactic and superoxide generating activities following cell loading with exogenous C24-LacCer ( 7 ). Lyn knockdown by siRNA completely abolished the effect of C24:1-LacCer loading on the LacCermediated functions of D-HL-60 cells. These observations suggested that C24-LacCer is specifi cally required for neutrophil properties and that long hydrophobic chains on LacCer may be responsible for their direct interaction with Lyn, via interdigitation with the acyl chains that allow the association of Lyn with the cytosolic membrane layer.
Photolabeling experiments have been performed using radiolabeled and photoactivable lipids ( 10 ). A GM1 ganglioside containing tritium-labeled sialic acid and with an azide at the end of the ceramide moiety was the fi rst radiolabeled ganglioside ( 11 ), followed by the synthesis of other photoactivable gangliosides ( 12 ). Upon addition to cultured cells, these exogenous ganglioside derivatives become associated with the cells and undergo metabolic processing in ways that closely resemble those observed for natural compounds, suggesting that they may be utilized as tools to study the interactions between GSLs and proteins. Photolabeling with photoactivable GM3 provided evidence for the specifi c role of GM3 in reducing insulin receptor-mediated processes in adipocytes ( 13 ) and showed that, following its interaction with CD9, GM3 downregulated tumor cell motility and malignancy ( 14 ). GM1 was found to interact with several membrane and cytosolic proteins (15)(16)(17)(18), two of which were characterized as caveolin-1 ( 16 ) and tubulin ( 17 ), and that GM3, GM1, and GD1b interacted with TAG-1 in the plasma membrane lipid rafts of neuronal cells ( 19,20 ).
This study was designed to show a direct interaction between C24-LacCer and Lyn. Cells were photolabeled with tritium-labeled analogs of LacCer having the photoactivable nitrophenylazide group at the end of the acyl chain. ethyl ether (3 × 50 ml). The combined organic phases were dried over Na 2 SO 4 and concentrated, resulting in compound 13 , at a yield of 71%.
Compound 13 (2 mmol) was dissolved in 5% CH 3 COOH in CH 3 OH (5 ml) and a catalytic amount of palladium on charcoal (6 mg) was added in an argon atmosphere. The reaction mixture was hydrogenated at atmospheric pressure for 20 h. The mixture was fi ltered and the catalyst washed with 5% CH 3 COOH in CH 3 OH. The solvent was evaporated, resulting in compound 14 , at a yield of 80%.
Protection and activation of 12-aminododecanoic acid and of 18-amino-octadecanoic acid (compound 14 ) . Protection of ( 15 ) and activation of compound 16 of -amino acids were as previously described ( 12 ). 326 l) at 0°C. After allowing it to stand for 30 min, the mixture was stirred for 4 h at room temperature, diluted in 50 ml CH 2 Cl 2 , and washed three times with 50 ml H 2 O. The organic layer was dried over Na 2 SO 4 and evaporated under reduced pressure, resulting in compound 11 , with a yield of 93%.
Compound 11 (4.8 mmol) was dissolved in 15 ml anhydrous dimethylformamide (DMF), to which 15 mmol of NaN 3 were added. The solution was stirred overnight at 45°C, cooled at room temperature, and concentrated. The crude material was dissolved in 50 ml CH 2 Cl 2 and washed three times with 50 ml H 2 O. The organic phase was dried over Na 2 SO 4 and evaporated under reduced pressure, resulting in compound 12 , with a yield of 58% ( 25 ).
Compound 12 (2.8 mmol) was dissolved in 0.5 M KOH in CH 3 OH (45 ml) and stirred for 48 h. Its pH was adjusted to pH 3 with 1 M HCl, after which it was concentrated and extracted with Deprotection . Tritiated compound 4A or B (1.7 mol) was solubilized in 32% aqueous ammonia (2 ml) in a screw-cap fl ask for 24 h at room temperature with vigorous stirring. The reaction mixture was dried and the residue was purifi ed by column chromatography with CH 2 Cl 2 -MeOH-2 N NH 3 , 60:35:4 by volume, resulting in compound 5A or 5B , at a yield of 80%.
Azide labeling . This fi nal reaction was conducted in a dark room. Radioactive compound 5A or 5B (10 mol/ml) was dissolved in anhydrous DMF. An equimolar quantity of Et 3 N and a 2-fold molar quantity of 4-F-3-NO 2 -phenylazide, dissolved in ethanol, were added and the mixture was stirred overnight at 80°C. The reaction mixture was concentrated and compound 6A or compound 6B was purifi ed by fl ash chromatography with CH 2 Cl 2 -MeOH-2-propanol-H 2 O, 70:10:20:3 by volume. Fractions containing the homogeneous tritiated and photoactivable LacCer were immediately solubilized in MeOH and stored at 4°C in a dark glass bottle.

Analysis of protein patterns
The concentrated DRM fractions prepared from illuminated C18-[ Proteins cross-linked to tritium-labeled LacCer derivative were detected by digital autoradiography. The presence of Lyn and G ␣ i was assessed by immunoblotting with specifi c antibodies, followed by reaction with secondary HRP-conjugated antibody ( 11,12,15,18,19,27,30,33 ). 1 H and 13 C NMR spectra were recorded with a Bruker AVANCE-500 spectrometer at a sample temperature of 298 K. NMR spectra were recorded in CDCl 3 or CD 3 OD and calibrated using the TMS signal as internal reference.

Other experimental procedures
Mass spectrometric analyses were performed in positive or negative ESI-MS. MS spectra were recorded on a Hewlett-Packard HP-5988-A or a Thermo Quest Finnigan LCQ™ DECA ion trap mass spectrometer, equipped with a Finnigan ESI interface; data were processed by Finnigan Xcalibur software system. All reactions were monitored by TLC on silica gel 60 F-254 plates (Merck).
Radioactivity associated with cells and cell fractions was determined by liquid scintillation counting. Digital autoradiography of the PVDF membranes was performed with a Beta-Imager 2000 (Biospace, Paris).

Determination of Lyn phosphorylation
All procedures before illumination were performed under red safelight. Culture dishes (10 cm) were coated with 5 ml of 10 g/ml anti-LacCer mAb T5A7 or normal IgM solutions as described ( 5 ) After ultracentrifugation, 12 fractions were collected, starting at the top of the tube. The light-scattering band, corresponding to detergent-resistant membrane (DRM), was located at the interface between 5% and 30% sucrose and corresponded to fraction 5. The entire procedure was performed at 0-4°C in ice immersion ( 7,29,31 ). In some experiments, the DRM fractions were diluted with 10 vol of buffer A, and then concentrated at 200,000 g for 1 h at 4°C. The resulting pellets were subjected to SDS-PAGE and Western blotting analysis.

cells, or DRM fractions from D-HL-60 cells loaded with C18-[ 3 H]LacCer-(N 3 ) and
All together, the yield of the GSL synthetic process was around 50%, which was very high considering the fi nal specifi c radioactivity of 1.5 Ci/mmol. Figure 3 shows the analytical controls as MS and NMR spectra of the synthesized 18-amino-octadecanoic acid ( Fig. 2 ), confi rming its structure. Figure 3 also

H]LacCer-(N )
Cells were administered the photoactivable species of LacCer together with the natural species, for a total of 0.5 g/ml of cells. Photoactivable LacCer was diluted with the natural compound to reduce self-quenching at the moment of illumination. Trypan blue assays showed that the viability of untreated cells was 96.33 ± 1.6%; when loaded with C18-[ 3 H]LacCer-(N 3 ) and C24-[ 3 H]LacCer-(N 3 ), they had a viability of 98.33 ± 0.3% and 96.00 ± 1.5%, respectively.

5, 7 ). After administration of exogenous C18-[ 3 H]LacCer-(N 3 ) and C24-[
3 H]LacCer-(N 3 ), the cells were illuminated under UV light, harvested, and lysed in lysis buffer containing Triton X-100. The lysates were loaded onto discontinuous sucrose gradients and ultracentrifuged; 12 fractions were collected from the top of each tube and their amounts of radioactivity analyzed. Figure 4 shows that, following loading with both C18-[ 3 H] LacCer-(N 3 ) and C24-[ 3 H]LacCer-(N 3 ), part of the radioactivity was associated with the low density fraction (DRM), corresponding to fraction 5, which is resistant to detergent solubilization. Radioactivity associated with high density fractions 9-12 was due to the photoactivable LacCer molecules peripherally associated with membrane components, but not correctly inserted into the membrane bilayer, or to their catabolites formed after internalization and degradation ( 12 ). Thus, this radioactivity does not refl ect a physiological interaction.
To determine whether the distribution of radioactivity was dependent on the process of illumination, gradient fractions were also prepared from nonilluminated cells. Figure 4 shows that the illumination process did not signifi cantly alter the distribution of radioactivity.
To verify whether the two LacCer analogs maintain the biological properties of the corresponding natural LacCer species, we assessed the phagocytosis activity of D-HL-60 cells loaded with C18-   Figure 7 shows that similar amounts of radioactivity were present in the immunoprecipitates from D-HL-60 cells loaded with the C24 and C18 derivatives ( Fig. 7A ). However, Lyn was present in immunoprecipitates from cells loaded with C24-[ 3 H]LacCer-(N 3 ), but not from cells loaded with C18-[ 3 H]LacCer-(N 3 ) ( Fig. 7B ). These results indicate that only LacCer containing very long fatty acids is present in the same membrane domain as Lyn.
To confi rm these fi ndings, we performed immunoprecipitation from DRM fractions with anti-Lyn and under disaggregating conditions. Figure 8 shows that Lyn was immunoprecipitates were separated by SDS-PAGE and blotted onto PVDF membranes, which were probed with rabbit anti-pY369-Lyn IgG (P-Lyn) and rabbit anti-Lyn IgG. Figure 5 clearly shows that Lyn was phosphorylated only in cells loaded with C24-[ 3 H]LacCer-(N 3 ), whereas Fig. 6 shows that D-HL-60 cells loaded with C24-[ 3 H]Lac-Cer-(N 3 ), but not illuminated, maintained the property of cells loaded with natural C24-LacCer and were capable of phagocytosing microorganisms. In contrast, phagocytosis did not occur when the cells were loaded with C18-[ 3 H] LacCer-(N 3 ), similar to fi ndings with natural C19-LacCer ( 7 ). These results confi rm that LacCer requires a long fatty acyl chain to activate Lyn and induce the phagocytic process and suggest that the photoactivable LacCer derivative is suitable for studying the properties of the natural compound.

Coimmunoprecipitation of C24-[ 3 H]LacCer-(N 3 ) and Lyn
C24-LacCer and Lyn have been shown to belong to the same membrane domain, with both immunoprecipitated

Identifi cation of LacCer-protein complexes
LacCer-enriched lipid raft-mediated neutrophil activities have been shown to involve Src family kinase-, PI-3 kinase-, and heterotrimeric G-protein-mediated signal transduction pathways ( 5,8 ). To better identify the radioactive LacCerprotein complexes, proteins from DRM fractions were concentrated, separated by SDS-PAGE, blotted onto PVDF membranes, and analyzed by digital radioimaging ( Fig. 9 ). Assays of D-HL-60 cells loaded with C24-[ 3 H]LacCer-(N 3 ) showed a few radioactive bands within a wide range of molecular masses, ranging from 20 kDa to 75 kDa, and included bands of 41 kDa and 56 kDa ( Fig. 9A ), suggesting that specifi c proteins are present in the membrane environment of LacCer species containing very long fatty acids. After radioimaging, the PVDF membranes were incubated with antibodies to Lyn and G ␣ i. These two antibodies bound specifi cally to the bands at 56 kDa and 41 kDa, respectively, suggesting that LacCer may directly interact with Lyn and G ␣ i. Indeed   the hydrolysis of GSLs at the cell surface by GSL hydrolases are involved in the regulation of apoptosis ( 39,40 ).
In mammals, ceramides are synthesized by a family of six enzymes, ceramide synthase (CerS)1-6, each of which uses a relatively restricted subset of fatty acyl-CoAs to N-acylate the sphingoid long-chain base ( 41 ). The levels of expression of each CerS encoding gene differ among several tissues, suggesting that molecular variations and expression patterns of fatty acid chains of GSLs refl ect the functions of these cells. CerS2 is the predominant CerS of C24 ceramides ( 42,43 ). CerS2 knockdown mice exhibited severe hepatopathy, myelin sheath defects, and cerebellar degeneration. The availability of very long-chain fatty acids requires ELOVL enzymes in mammals. ELOVL1 shows high activity toward saturated and monounsaturated C20and C22-CoAs, and has been shown to be essential for the production of C24 fatty acid containing sphingolipids. ELOVL1 activity is regulated by CerS2. Knockdown of ELOVL1 or CerS2 markedly reduced Lyn activation in HeLa cells ( 44 ). These results are consistent with our observation that C24-LacCer molecules are indispensable for Lyn-coupled LacCer-enriched lipid raft-mediated neutrophil functions ( 7 ).
LacCer-enriched lipid rafts mediate neutrophil chemotaxis, phagocytosis, and superoxide generation of neutrophils ( 5,8,45 ). DMSO-differentiated HL-60 cells are commonly used for analyzing neutrophil functions, including migration, phagocytosis of opsonized microorganisms, superoxide generation, and degranulation ( 5,8,(46)(47)(48)(49). However, those HL-60 cells do not possess LacCermediated neutrophil functions, including migration, superoxide generation, or phagocytosis, because the presence of C24-LacCer in plasma membranes is insuffi cient to form Lyn-coupled LacCer-enriched lipid rafts ( 7 ). In contrast, C24-LacCer is present in granules of DMSOdifferentiated HL-60 cells, indicating that biosynthesis of C24-LacCer in itself is normal. It seems, therefore, that the delivery system of C24-LacCer from Golgi apparatus to plasma membranes in DMSO-differentiated HL-60 cells is abnormal. Although it remains unclear why C24-LacCer is not selectively delivered to plasma membranes from Golgi apparatus in DMSO-differentiated HL-60 cells, these cells are a good model to analyze the organization and signal transduction mechanisms of LacCer-enriched lipid rafts. LacCer is located on the outer layer of plasma membranes, whereas Lyn is present at their cytosolic face via myristic/ palmitic acyl chains. It has been suggested that the long fatty acids of LacCer may interdigitate ( 50,51 ) with the acyl chains of the cytosolic layer, reducing membrane thickness ( 52,53 ). This may allow for direct interactions between LacCer and Lyn, which may be instrumental for Lyn phosphorylation, considered the starting process necessary for the immunological functions of neutrophils ( 7 ). To date, however, direct interactions between LacCer and Lyn through the membrane layer have not been observed experimentally.
Photoactivable GSLs have been shown to be good tools to study GSL-protein interactions ( 13,14,16,17 ). To verify the direct interaction between C24-LacCer and Lyn DISCUSSION GSLs are components of the outer layer of plasma membranes, with considerable heterogeneity, due to heterogeneity of both the oligosaccharide and ceramide moieties . They are enriched with respect to glycerophospholipids in limited areas containing a few selected proteins, many of which are associated with cell signaling ( 28,34 ). GSLs have been found to modulate cell signaling processes through specifi c interactions between their oligosaccharide chains and proteins/glycoproteins in their environment. In contrast, it is unclear whether ceramides play specifi c roles in such processes, although changes in the structures of both the long-chain base and the fatty acids have been associated with the differentiation and aging processes (35)(36)(37)(38). In addition, ceramides resulting from These results are similar to those observed when D-HL-60 cells were loaded with natural LacCer molecules containing very long and short/medium fatty acids, suggesting that the photoactivable LacCer derivatives mimic the natural molecules.
After cell loading and illumination, the patterns of the radioactive proteins were analyzed. Illumination rapidly transforms the azide to nitrene, a very reactive group capable of covalently binding to molecules in their nearest environment. Thus any interaction of tritium-labeled LacCer derivative with proteins would yield tritium-labeled LacCer-protein complexes that can be analyzed by PAGE, followed by radioimaging of the blotted material.
through the membrane layers, we synthesized a photoactivable LacCer containing a reactive azide group at the end of the acyl chain. This chain was of a length similar to that of C24 fatty acids, as shown by molecular modeling. The photoactivable C24 fatty acid containing LacCer was labeled with tritium at position 3 of sphingosine, yielding C24-[ 3 H]LacCer-(N 3 ). As a control, we also synthesized a tritium-labeled photoactivable LacCer containing a shorter acyl chain, C18-[ 3 H]LacCer-(N 3 ).
In this study, we showed that C24-[ 3 H]LacCer-(N 3 ) covalently binds to Lyn in LacCer-enriched lipid rafts. D-HL-60 cells were loaded with these two types of [ 3 H] LacCer-(N 3 ), with percentages of each found to be present in the DRM fractions prepared from cell lysates. Lyn phosphorylation and phagocytosis were observed in cells loaded The immunoprecipitates were separated by SDS-PAGE and blotted onto PVDF membranes, which were probed with rabbit anti-Lyn IgG. The immunoblot shown is representative of three different experiments. B: The radioactivity associated with the immunoprecipitates was determined by liquid scintillation counting. Data are expressed as percentages of total radioactivity. Data are the mean of three different experiments. *** P < 0.0005. necessary for transfer of information through the membrane. However, it is diffi cult to imagine that these interactions are associated with specifi c hydrophobic moieties . Rather, LacCer containing C24 fatty acids may promote the organization of specifi c lipid rafts necessary to allow the interaction of several proteins instrumental to the cell signaling processing. Without C24-LacCer molecules, this microenvironment would not be available, Lyn would not be recruited and activated by phosphorylation, and the signaling process would not be induced.
The administration of C24-[ 3 H]LacCer-(N 3 ) to the cells followed by illumination resulted in the association of radioactivity with the plasma membrane lipid rafts. Most of the radioactivity was linked to lipids, which was not unexpected, because lipids are the main components of lipid rafts. Some of the radioactivity, however, was linked to proteins, but only a few proteins were labeled, suggesting that their interactions with LacCer are specifi c. One of the cross-linked proteins was identifi ed as Lyn, whereas a second protein was stained with anti-G ␣ i. Because the azide moiety is at the end of the LacCer fatty acid, which is inserted into the outer membrane layer, and Lyn is associated with the cytosolic membrane layer by insertion of its acyl chains, the only mechanism by which a LacCer-Lyn complex can form is by the acyl chain of LacCer contacting with the acyl chain(s) of Lyn. In contrast, when C18-[ 3 H]LacCer-(N 3 ) was administered to cells, little radioactivity was associated with the lipid raft proteins, and a LacCer-Lyn complex could not be detected.
These fi ndings confi rm that LacCer containing C24 fatty acid is in contact with the acyl chains of Lyn and that this is necessary for the fi nal activation of Lyn, through phosphorylation, followed by the cascade of reactions