A Sulfated Glucosylceramide from Rat Kidney - The Journal of

THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 264, No.10, Issue of April 5, pp. 5974-5969,1989 Printed in U.S. A.

0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

A Sulfated Glucosylceramide fromRat Kidney* (Received for publication, June 15,1988)

Naoko IidaS, Toshihiko ToidaS, Yasunori Kushig,Shizuo Handag, Pam Fredmanll, Lars Svennerholmll, and Ineo IshizukaS From the $Department of Biochemistry, Teikyo University, School of Medicine, Kaga, Ztabashi-ku, Tokyo 173, Japan, the §Department of Biochemistry, Faculty of Medicine, Tokyo Medical and Dental Uniuersity, Yushima, Bunkyo-ku, Tokyo 113, Japan, andthe llDepartment of Psychiatry and Neurochemistry, St. Jorgen’s Hospital, Gothenburg university, S-422 03, Hisings, Backa, Sweden

A novel sulfated glycosphingolipid containing a sulfated glucosyl residue was isolated from rat kidney and purified to homogeneity by column chromatographies with DEAE-Sephadex and silica beads. By compositional analyses, permethylation studies, one- and two-dimensional proton magnetic resonance spectroscopy, infrared spectroscopy, negative secondary ion mass spectrometry, solvolysis, and immunostaining on thin layer chromatogram, the structure of this glycolipid w a ~proposed to beHSOa-3Glcj31-1Cer (where Cer is ceramide). The ceramide portion consisted of 4D-hydroxysphinganineas the sole long chain base, and the fatty acid consisted of predominantly tetracosanoic acid, deduced from both composition analysis and negative secondary ion mass spectrometry. The yield of glucosyl sulfatide was about 5 nmol/g of tissue, being about three times as much as that of lactosylceramide sulfate.

In recent years, sulfated glycolipids of ganglio-series including SM2,’ SB2, SBla, and SMlb from kidneys and intestines of rodents (3) have been isolated in addition to SM4s and SM3 to constituteadistinct group of acidic amphiphiles comparable to gangliosides (4,5). Sulfated glycolipids, as well as cholesterol sulfate, may play roles in the maintenance of ionic homeostasis (6-8) and blood coagulation (3). More detailed studies of the SM4s fractions showed that rat kidneys * This work was supported in part by a grant from the National Swedish Board for Technical Development (Project 84-4667). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Abbreviations of gangliosides and sulfated glycolipids follow the nomenclature system of Svennerholm (1)and Ishizuka (27), respectively, and those of other lipids are according to therecommendation of the Nomenclature Committee, International Union of Pure and Applied Chemistry (2). Theabbreviations and trivial names used are: SM2, gangliotriaosylceramide monosulfate (GgOsesCer I13-sulfate); Lac, D-lactose; Cer, ceramide; SM45, galactosylceramide sulfate (GalCer 13-sulfate); SM3, lactosylceramide sulfate (LacCer 113-sulfate); SM4g, seminolipid (Galpl-3alkylacyl-Gro 13-sulfate); SMlb, gangliotetraosylceramide monosulfate (GgOseXer IV3-sulfate);SB2, gangliotriaosylceramide bis-sulfate (GgOsesCer 113, II13-bis-sulfate); SBla, gangliotetraosylceramide bis-sulfate (GgOserCer I13,1V3-bissulfate); HPTLC, high performance TLC; GLC, gas-liquid chromatography; GC-MS, gas chromatography-mass spectrometry; HPLC, high performance liquid chromatogaphy; BSTFA, bis(trimethy1sily1)trifluoroacetamide Sr sulfated glycolipid X, GM3, N-acetylneuraminyllactosylceramide (I13NeuAc-LacCer);Gm, N-acetylneuraminylgalactosylceramide (13NeuAc-Galcer);Sulph I-monoclonal antibody, an IgG, monoclonal antibody reacting with galactosylceramide sulfate.

contain a novel sulfatide, GlcCer sulfate (sulfated glycolipid X , Sx), which could not be separated from SM4s on TLC unless an acidic solvent system was used. The structure of this Sx is the subject of this report. EXPERIMENTAL PROCEDURES

Materials-Wistar rats were obtained from Charles River Japan Inc. Some standard substances purchased were as follows: 4-sphingenine (d181) as thefree base from Serdary Research Laboratories, London, Canada; cholesterol 3-sulfate as a sodium salt from Sigma; nonhydroxy and hydroxy fatty acids from Supelco, Inc., Bellefonte, CA. 4-~-Hydroxysphinganine (t180) was prepared from cerebrosides of rice endosperm in our laboratory. Reference glycosphingolipids including SM4s (9) and SM3 (10) from human kidneys, SM4g from boar testes (11), and GalCer from the white matter of an adult human brain were prepared in our laboratory. LacCer was isolated from human erythrocytes (12) and further purified on an Iatrobeads column according to Ando et at. (13). GlcCer which was prepared from the spleen of a patient suffering from Gaucher’s disease was a gift from K. Ogawa (Department of Pediatrics, Teikyo University). Partially 0-methylated monosaccharides were prepared in our laboratory from the followingsources: 2,3,4,6-tetra-O-methyl-1,5-di-O-acetyland 2,4,6-tri-O-rnethyl-1,3,5-tri-O-acetylglucitol from laminaritetraose; 2,3,4-tri-0-methyl-1,5,6-tri-O-acetylglucitol from gentiobiose; 3,4,6tri-O-methyl-l,2,5-tri-O-acetylglucitol from kojibiose (14); 2,3,6-triO-methyl-1,4,5-tri-O-acetylglucitoland 2,3,4,6-tetra-O-methyl-1,5-di0-acetylgalactitol from lactose. Reagents purchased for derivatization or NMR spectrometry were as follows: dimethyl sulfoxide (Me2SO),anhydrous pyridine, [‘HI Me2S0 (99.95%), and [‘HI20 from Merck, Darmstadt; bis(trimethylsily1)trifluoroacetamide(BSTFA) from Sigma; [‘Hlchloroform (99.9%) and [‘Hlmethanol (99.9%) from Aldrich. Silica Gel 60 HPTLC plates were obtained from Merck. DEAE-Sephadex A-25 and porous silica gel, Iatrobeads (6RS-8060), were supplied by Pharmacia Japan, Tokyo and Iatron, Tokyo, respectively. Three percent OV-101 on GasChrom-Q (100-120-mesh) and 10% EGSS-X onChromosorb AW (80-100-mesh) were purchased from Applied Science Laboratories, and a fused silica capillary column (0.24 mm X 25 m, OV-101 a t 0 , 2 - ~ mthickness) was obtained from Shimadzu, Kyoto. The HPLC prepacked column of Nucleosil silica beads (10 X 300 mm, 5-pm diameter) was the product of Senshu Kagaku, Tokyo. All organic solvents and other compounds wereof analytical or GLC (silylation) grade. Thin Layer Chromatography-TLC was performed on Silica Gel 60 HPTLC plates with the following solvent systems: I, chloroform/ methanol/0.2% CaCL(65:35:6, v/v); 11, chloroform/methanol/acetone/acetic acid/water (82:4:2:1, v/v); 111, chloroform/methanol/3.5 M NHrOH (65:35:6, v/v); IV, chloroform/methanol/water(65:25:4, v/ v); V, l-propanol/l5 N NH,OH/water (80515, v/v). Glycolipids and sulfated glycolipids were visualized with orcinol or resorcinol (19) and azure A reagents (20, 21), respectively. Azure A reagent was prepared by dissolving azure A to saturation in 1 mM H2SO4 at room temperature. After drying, the plate was sprayed evenly with azure A reagent and soaked in 0.04 M HnS04/methanol, 3:l (v/v) several times with gentle agitation to remove unbound dye. The bands of lipids containing sulfate ester were stained blue. Infrared Spectroscopy-For IR spectroscopic analysis, approximately 100 nmol of Sx, which had been thoroughly dried in U ~ C U O

5974

A Sulfated Glucosylceramide Kidney Ratfrom over Pz05,were pressed with 20 mg of dry KBr andscanned in an A302 infrared spectrophotometer (Japan Spectroscopic Co., Tokyo). Gas-Liquid Chromatography-GLC analyses were performed with a Shimadzu GC-7A instrument. The trimethylsilylated methyl glycosides and partially methylated alditol acetates were applied to a fused silica capillary column coated with OV-101 (0.24 mm X 25 m). The temperature was programmed from 160 to 260 "C at the rate of 4 "C/min. For the methyl esters of saturated fatty acids, a capillary column with OV-101 was also used, and the temperature was, after holding a t 150 "C for 2 min initially, linearly raised from 150 to 270 "C at the rateof 4 "C/min. Analyses of the methyl esters of unsaturated fatty acids were achieved by using a glass column (3.2 mm X 1.1 m) of 10% EGSS-X on Chromosorb AW (80-100 mesh) at 165 or 180 "C. N-Acetyl-0-trimethylsilyl derivatives of sphingoid bases were analyzed isothermally at 250 "C by using a capillary column coated with OV-101. Gas Chromatography-Mass Spectrometry-GC-MS was performed by a Shimadzu model 6020 Auto GC-MS apparatus equipped with a SCAP 1123 data system using a glass column (3.2 mm X 1m) of 3% OV-101. Conditions for electron impact mass spectrometry were as follows: ionizing potentials, 20 eV; accelerating voltages, 3.5 kV, ion source temperature, 280 "C. The conditions for gas chromatographic separation were the same as described above, except for sphingoid bases for which the temperature was programmed from 210 to 240 "C at the rateof 2 "C/min and thenheld isothermally at 240 "C. Chemical Analysis-The sugar composition of Sx was determined by GLC as trimethylsilylated methyl glycosides (22). Glycolipids (20 nmol) were dried in uacuo over PZOS and methanolyzed with 0.5 ml of 1.0 M HCl in anhydrous methanol a t 80 "C for 24 h. The reaction mixture was washed twice with n-hexane to extract the methyl esters of fatty acids, and the lower phase was dried and trimethylsilylated by incubation with 20 plof BSTFA/pyridine, 2:l (v/v) at room temperature overnight. The combined hexane layer containing methyl esters of fatty acids was dried up and then subjected to GLC and GC-MS. Methylation and acetolysis of Sx were performed according to Hakomori (23), except for the reduction step, where NaB['HI4 was used instead of NaBHa (15, 24). The acetate of partially methylated alditol was analyzed by GLC and GC-MS as described above. For sphingoid bases, Sx was hydrolyzed in methanol/concentrated HCl/water (82:8.6:9.4, v/v) a t 70 "C for 18 h (25), treated with methanol/acetic anhydride (41, v/v) (261, and then incubated with BSTFA/pyridine (9:1, v/v) to obtain the N-acetyl-0-trimethylsilyl derivatives of sphingoid bases. These derivatives were analyzed by GLC and GC-MS as described above. Quantitative Analyses-Quantification of SM4s, SM3, and Sx in the tissue was achieved by densitometry on TLC. The fractions containing the sulfated glycolipids were developed with solvent system I1 and stained with orcinol reagent. Densitometry was performed using a Shimadzu dual-wavelength TLC scanner CS-930 at 540 nm. The concentration of each sulfated glycolipidwas determined by comparing its absorbance with that of SM4s as a standard. The amountof the sulfate group in Sx was determined by the azure A method of Kean (21,27).About 4 nmol of Sx,which were calibrated previously by GLC using mannitol as an internal standard, were dissolved in 1.5 ml of chloroform/methanol, 1:l (v/v), and then 1.5 ml of 0.025 M HzSO, and 0.3 ml of the dye solution were thoroughly mixed with the lipid solution. The absorbance of the lower phase was measured using a UVIDEC-340 spectrophotometer (Japan Spectroscopic Co., Ltd., Tokyo) a t 635 nm. The sulfate group in Sx was determined using the calibration curve for SM4s. Cholesterol 3-sulfate was estimated by the method of Zak (28) after methanolysis in 40 ml of 1.5 M methanolic HCl at 37 "C for 9 h. A portion of methanolysate containing 0.1-0.3 mg of cholesterol was dried and dissolved in 6 ml of glacial acetic acid. And then FeC13 solution, 0.12% FeC13 in concentrated H,P04/HzS04 (2:23, v/v), was mixed with the lipid solution. After cooling the reaction mixture for 10 min, the absorbance was measured at 550 nm. Soluolysis-Solvolysis of Sx in anhydrous dioxane was performed at 100 "C for 30 min as described previously (29). After removal of the inorganic sulfate released by the solvent partition, the products were analyzed by TLC with the basic solvent system V (1-propanol/ 15 N NH40H/water, 80:5:15, v/v), which was able to separate GlcCer from GalCer (30, 31). Negative Secondary Ion Mass Spectrometry-Secondary ion mass spectra were obtained as previously described (32), with a Hitachi M80A mass spectrometer modified for secondary ion mass spectrometry and a Hitachi 003 data processing system. One pl of Sx solution (1-5

5975

pglpl) dissolved in chloroform/methanol (2:1,v/v) was loaded on a silver plate with 1p1 of triethanolamine and mixed. The excess solvent was evaporated in the air. The accelerating voltage was 3 kV. A xenon ion beam of 5 X lo-' A bombarded the target at 5 keV. Proton Magnetic Resonance CH NMR)-For 'H NMR spectroscopy, about 400 nmol of purified Sx were treated repeatedly with 0.5ml portions of C[2H]Cl~/CH30[2H], 4:l (v/v), followed by desiccation over Pz05in Vacuo to exchange the labile protons with deuteron. Then thoroughly dried Sx was redissolved in 0.5 ml of [2H]MeZSO/ ['H],O (982, v/v, 0.8 mM in the final concentration of Sx). The spectra were obtained by a GX-400 spectrometer of Japan Electron Optical Laboratory (JEOL) equipped with a JEOL PLEXUS data system and a RSX-11M computer system operated in the Fourier transform mode. The operation conditions for one-dimensional spectra were as follows: frequency, 400 MHz; sweep width, 4 kHz; flip angle, 90" (7.4-ps pulse); sampling points, 16,000; accumulation, 6000 times; temperature, 30 "C. Two-dimensional multiple relayed COSY spectra were obtained by using the pulse sequences of Bax of Drobny (33). The delay time in multiple relayed COSY spectra were set a t 31 ms.All the twodimensional spectra were measured with 512 X 2048 data points and a spectral width of2.5 kHz. One hundred twenty-eight scans were accumulated for each tl, and the total measurement time for a twodimensional spectrum was about 24 h. An apodization factor for t, and tz was sine-bell. The contour display of the spectrum was presented inthe absolute-value mode. Chemical shifts were indicated by parts/million from the signal of tetramethylsilane as an internal standard. Immunological Staining Assay-Radioimmunodetection of the binding of the Sulph I-monoclonal antibody to the Sx preparation was performed. Sulph I-monoclonal antibody was produced by immunizing BALB/c mice with SM4s from human kidneys coated on Salmonella minnesota bacterial membranes. Immunological staining assay to galactosyl or glucosyl sulfatides as well as the epitope specificity of the antibody has been described recently (34). The chromatogram was developed with solvent system 11. RESULTS

Isolation of Sx from Kidneys-570 g of kidneys from 8week-old Wistar rats were extracted in three steps with chloroform/methanol(2:1, v/v), chloroform/methanol/0.88% KC1 (60:120:9, v/v), and chloroform/methanol/0.4 M CH3COONa (30:60:8, v/v) (17). Acidic lipids were separated from the lipid extract according to themethod previously reported (17) with some modifications. Briefly, the combined crude extracts were concentrated to dryness. After partition in a Folch system (18), solvents of the lower phase were evaporated and then the residue was methanolyzed with 0.2 M NaOH in anhydrous methanol for 3 h at 37 "C. After neutralization, the reaction mixture was adjusted to chloroform/methanol/water,8614:l (v/v), and partitioned together with the upper phase. The second upper phase was dialyzed against water and lyophilized. The lyophilisate was combined with the second lower phase to obtain the crude alkali-resistant lipids. The alkali-resistant lipids were made up to 1.5 liters of chloroform/methanol/water, 5:lO:l (v/v), and applied to a column (125 ml, 2.2 X 33 cm) of DEAE-Sephadex A-25 (acetate form). After the neutral glycolipids werewashed out from the column with 5 volumes of the same solvent mixture, the acidic lipids were separated by a linear gradient of 4.0 liters of chloroform/methanol aqueous ammonium acetate (0.03-1.0 M, 5:10:1, v/v). Eluted glycolipids were monitored on TLC with solvent system I1 (Fig. 1). Glycolipids and sulfated glycolipids werevisualized with orcinol (Fig. 1A ) and azure A reagents (Fig. 1B). An unidentified sulfated glycolipid (Sx), which was sensitive to both orcinol and azure A reagents, is pointed out by arrows in Fig. 1. Sx was coeluted with SM4scontaining hydroxy fatty acids from DEAE-Sephadex but migrated much faster than S M 4 s on TLC in solvent system 11, suggesting that Sx may be a glycolipid monosulfate which is not any one

A Sulfated Glucosylceramidefrom Rat Kidney

5976

sm4g

SM4s SM3

sm2 30605040

70 Fraction No.

80

90

100

110

120

FIG. 2. Thinlayerchromatography of purified Sx. The plates were developed with: A, chloroform/methanol/0.2W CaCh (65:35:6, v/v); B, chloroform/methanol/acetone/acetic acid/water (8:2:4:2:1,v/v); C, chloroform/methanol/3.5 M NH,OH (65:35:6), and stained with orcinol reagent. Lane I , standard cholesterol 3-sulfate; lane 2, SM4g from boar testis; lane 3, purified Sx; lane 4, SM4s from human kidney; lane 5, SM3 from human kidney; lane 6, neutral glycosphingolipid standards: GalCer from human brain and LacCer from human erythrocytes, from top to bottom.

suggesting that Sx may not be any one of the molecular species of SM4s. Solvent system11, which has been known to SM4g be useful to separate sulfated glycolipids from each other (29), was also found to be convenient in separatingSx from SM4s. SM45 The alkali-resistant neutral lipids eluted from the DEAESM3 Sephadex column were subjected to column chromatography with Iatrobeads using a gradient system (chloroform/methaSM2 nol/water, 95:5:0 to 70302, v/v) to procure the glucosylceramide (GlcCer-IV)which hasthe ceramidemoiety corre30 5040 60 70 80 90 100 110 120 sponding to that of Sx. Purity of isolatedGlcCer-IVwas Fraction No. confirmed on TLC in solvent system V (Fig. 7) and the solvent FIG. 1. Elution profile of acidic glycolipids from a DEAE- system containing borate(40) (data not shown). Sephadex column. Alkali-resistant and acidic lipids were eluted Infrared Spectroscopy-Both the absorptions at1230 cm" from the DEAE-Sephadex column with a linear gradient of ammo- (due to S=O stretching) and 815 cm" (due to C-0-S vibranium acetate in chloroform/methanol/water,5:101 (v/v). Portions tion) were observed in the IR spectrum of Sx, indicating the from every fivetubes, each 20 ml, were monitored by TLC. The plates were developed with solvent system I1 and stained with orcinol presence of an equatorial sulfate ester of hexose (11, 12, 15). reagent ( A ) and azure A reagent ( B ) . The amounts of the eluates The absence of an ester carbonyl absorption at 1740 cm" applied were: A , 50 pl each from Fractions 25-100 and 150 pl from (11) supported the glycosphingolipid nature of Sx. Further106-150; B, 100 p1 from Fractions 25-75,50 pl from 80-100, and 150 more, amideabsorptions were observed at 1540 and 1635 pl from 105-150. The arrow points to Sr.*, G Mwas ~ negative to azure cm". The above results suggested that Sx was a sulfated A immeidately after staining until a few days later the brownish band glycosphingolipid. ~ weakly stained to blue due to its high was visible; **, G M was Chemical Analysis-GLC analysis of the trimethylsilylated concentration, but the color changed gradually to brownin a few days; ***, this bandappearedblue immediately after staining and methyl glycosides established thatSx had only glucose as the carbohydrate component,definitely showing that Sx was difturned yellow in a few days. ferent from SM4s in the carbohydrate moiety. Sx was subof the molecular species of SM4s. In the fraction 115-120, a jected to methylationanalysis. By GLC of the partially methband migrating faster than SM4g was stained with azure A. ylated alditol acetates, Sx yielded only one peak. The retenof this peak completely This compound was identified to be cholesterol 3-sulfate by tion time and the mass spectrum the R,value and detectionof cholesterol in theacid hydrolysis coincided with that of the standard 2,4,6-tri-O-methyl-l,3,5tri-0-acetylglucitol preparedfrom laminaritetraose. From the products. The fractions containingSx were combined, dialyzed, and result of quantification of sulfate groups inSx by the azureA lyophilized. The further purification of Sx was achieved by method, the ratio of the number of the sulfate group to the the column chromatography using Iatrobeads (1.2 x 46 cm, glucosyl residue was shown to be 1:1. Taking the R p value 52 ml) and a linear gradient with 2.0 liters of chloroform/ into consideration, it was suggested that Sx may probably methanol/water, 190:101 to 70:30:2 (v/v). The fractions con- contain a glucose sulfated at the C-3 position. Table I shows the composition of the lipophilic moiety of taining Sx were combined, concentrated, and then subjected to thefinal procedure for purification, which was achieved by Sx obtained by GLC and GC-MS analysis. The major nonHPLC on a Nucleosil silica gel column (1.0 x 30 cm, Senshu hydroxy fatty acids of Sx were 24:O (47.5%), 23:O (12.6%), and Kagaku, Tokyo) with a linear gradient of chloroform/metha- 22:O (15.6%), in sharp contrast to the predominance of 16:O no1 (1O:O to 7:3, v/v, 600 ml). This step was duplicated to (37.0%) inSM4s (15). The 2-hydroxy fatty acids,which constituted 76.2% of the total fatty acids in SM4s(15), were completely remove SM4s. Isolated Sx was subjected to TLC with solvent systems I, not detected in Sx. Sx did not contain hphingenine (d181) 11, and 111. Sx migrated as a singleband in all solvent systems. at all, but 4-~-hydroxysphinganine (t180) as the sole long On TLC in the neutral solvent system I or the basic solvent chain base. system 111, Sx comigrated with SM4s containingnonhydroxy 'HNMR Spectroscopy-The spectra of the whole and amfatty acids (Fig. 2, A and C). In the acidic solvent system 11, plified ring proton region (inset) of Sx are shown in Fig. 3, (Fig. 2B), and the double relayed COSY spectrum of Sx in Fig. 4. The however, Sx migratedmuch fasterthanSM4s

5977

A Sulfated Glucosylceramide from Rat Kidney H6a

TABLE I Composition of fatty acid and sphingoid base i n Sx Analyses were performed by GLC as described in the text. The values were obtained by synthetically judging the results with both OV-101 and 10% EGSS-X to rule out the interference by dibutylphthalate and other contaminants.

2.0

Sphingoid base

Weight % of total

Fatty acid

Weight % of total

14:O 2.2 d181" 0 16:O 9.4 t180" 100 18:O 4.0 20:o 220 15.6 23:O 12.6 24:O 47.5 24:l Traceb 25:O Traceb d181, 4-sphingenine; t180, 4-D-hydroxysphinganine. * Peaks of fatty acids smaller than 2.0% were designated as trace.

1

~

"

4.2

4.0

1

"

~

3.0

I

~

3.b

"

~

"

3.4

~ 3.2

~

~

PPM ' " 7.8

"

3.0

r

i'

.. N -

O D

4'

I

1

.I1 ." l

6

'

'

'

'

,

5

'

'

"

l

"

d

~

'

l

'

3

'

~

'

l

2

"

'

'

/

'

1

"

FIG. 4. Two-dimensional double relayed COSY spectrum of Sz.The conditions for the analysis were described in the text. 1,2 etc. means the cross-peak of e.g. H-1 and H-2protons.

'

FIG. 3. One-dimensional 'H NMR spectrum of Sx. Inthe inset, the abscissa scale of the region between 3.0 and 4.4 ppm was amplified 3-fold. The conditions for the analysis were described in the text. The symbols, H-I to H-6b, correspond to the ring protons of the glucosyl residue.

TABLEI1 Chemical shifts and coupling constants for ring protons of Sx and GlcCer Chemical shift

H-5

H-4 H-3 H-2 H-l

H-6b H-6a

PPm

assignments of the signals shown in Table I1 are based on the results obtained by the multiple relayed COSY spectra and irradiation studies of each ring proton in turn. A doublet of H-1 resonance at 4.222 ppm (JI,z= 7.8 Hz) coupled to theH2 triplet at 3.094 ppm could be ascribed to P-glucose (35, 39). The resonance of H-2 ring proton in S x was greatly shifted to a higher field (3.094 ppm) than H-2 in SM4s (3.532 ppm, data not shown) (38) by the difference of 0.438 ppm. The similar behavior of the resonance of H-2 in GlcCer has been well known (35, 39). The difference in the chemical shifts between the protons of Sx and glucosylceramide (GlcCer-IV; see the following section, "Solvolysis of Sx") was largest with the H-3 ring protons (0.835 ppm) (Table 11),mainly due to thedeshielding effect by the electronegative sulfate group to the geminal proton (36, 37), supporting that the location of the sulfate group is at the C-3 position. In Table 11, the downfield shifts of the H-2 and H-4 resonances of Sx against the corresponding signals of GlcCer were 0.141 and 0.200 ppm, respectively. In contrast, thedownfield shifts of the H-2 and H-4 resonances of SM4s from the corresponding resonances of galactosylceramide (GalCer) were 0.207 and 0.317 ppm, respectively (38). The large downfield shift of the H-4 ring proton of the galactosyl residue in SM4s (0.317 ppm) may be explained as follows. H-4 in SM4s is equatorial to the ring and receives a

Sx GlcCer" Difference ( S X- Glc Cer)

4.222 3.094 3.976 3.228 3.166 3.437 3.657 4.100 2.953 3.141 3.028 3.086 3.4* 3.663 0.122 0.141 0.835 0.200 0.080 0.0 -0.006

Coupling constant

Hz 9.3 5.9 2.0 8.8 9.3 5.9 2.0 8.8 ____ GlcCer consisting from t180 and nonhydroxy fatty acids (GlcCerIV, see the text) was prepared from rat kidney. Because the signals of H-6a proton were covered by the large signal of ['HIHO, the chemical shift of H-6a proton could be read down to only the first decimal place. Sn GlcCer"

7.8 7.8

8.8 8.8

stronger anisotropic effect from the equatorial sulfate group at theC-3 position than theaxial H-2 proton, while in Sx H4 as well as H-2 of the glucosyl residue is axial to thering and receives a smaller anisotropic effect from the sulfate group (36). The values of J3,4and J4,5 of S x were 8.8 and 9.3 Hz, respectively, in good agreement with those reported for GlcCer ~ (Table 11).Dabrowski et al. (39) have explained that J s , and J4,5 of GalCer were smaller (3.0 and 1.2 Hz)because of gauche

'

~

'


5978

positions of the pairs H-3/H-4 and H-4/H-5, respectively. In contrast, the conformations of the pairs H-3/H-4 and H-41 H-5 in a glucosyl residue are both antiperiplanar. No resonances were observed between 5.4 and 5.6 ppm, suggesting the absence of a trans-double bond of 4-sphingenine and the cis-protons of unsaturated fatty acids in Sx, consistent with the compositional analyses. Immunological Studies Using Sulph I-Antibody-As verified by the radioimmunoassay on HPTLC plates (Fig. 5), there was no detectable binding of the Sulph I-antibody to the preparation of Sx, although 3.5 nmol were applied. Sulph I-antibody was the monoclonal antibody specific to SM4s, SM3, and SM4g (34). This result is one piece of evidence that Sx was glucosylceramide sulfate and not galactosylceramide sulfate. Sulph I-antibodyseems to recognize the configuration of the equatorial H and/or the axial OH group at the C-4 position of the galactose as the structureof the epitope. This result also means that less than 0.07% of SM4s, if any, contaminated this preparation, because only 5 pmol of standard SM4s, that were divided into two bands, were visible. Negative Secondary Ion Mass Spectrometry-In the spectrum (Fig. 6), the most intense molecular ion species ((MH)-) detected at m/z 908 was formed by the loss of a proton

1 2 3 4 5 6

FIG. 7. Solvolysis of Sx. The products obtained by desulfation of purified Sx were separated on a HPTLC plate with l-propanol/l5 N NH.OH/water (805:15, v/v) and stained with orcinol reagent. Lane I, intact Sx; lane 2, Sx after solvolysis; lane 3, GlcCer-IV, which has the ceramide moiety corresponding to that of Sx, prepared from rat kidney as described in the text; lane 4, GlcCer from the spleen of a Gaucher patient; lane 5, GalCer from humanbrain; lane 6, the ceramide monohexosides from rat kidney.

a

1

->os&$+, CH2

OH

0,

c,,c\

H N ’,c OH

/

C

M

.c\

OH

FIG. 8. The structure of Sx. The ceramide moiety was shown as N-(n-tetracosanoyl)-4-hydroxysphinganinesince this structure represents the major component. For variations of the fatty acids, see Table I.

from the sulfated glycosylceramide with 24:O and t180as the fatty acid and the sphingoid base, respectively (32). Furthermore, the peaks of mlz 880 and 894 were also very intense. These ions corresponded to (M-H)- of GlcCer sulfate with 22:O and 23:O fatty acids, respectively, and t180 sphingoid base. This result was consistent with those obtained by the GLC analysis. In negative secondary ion mass spectra of Sx, however, it was deduced that the longer the fatty acid chain SX 1.25 2.5 5.0 10 20 40 80 pmol lengths were, the more intense were the molecular ions FIG. 5. Radioimmunodetection of the binding of the Sulph formed, in comparison to the fattyacid composition obtained I-antibody t o Sx on HPTLC. A series of various concentrations, from 1.25 to 80 pmol, of standard galactosylsulfatide and purified Sx, by the GLC analysis. Actually the (M-H)- ion corresponding 3.5 nmol, was applied to the aluminum-backed HPTLC plate. The to the ceramide with 25:O fatty acid and t18:O sphingoid base chromatogram was developed with solvent system I1 and the radio- detected a t m/z 922 was as intense as the molecular species immunodetection assay performed as described under “Experimental with 200 and t180. Procedures.” Solvolysis of Sx-In order to provide a standard for desulfated Sx on TLC, GlcCer-IV, which has theceramide moiety Fatty acid similar to that of Sx, was prepared from rat kidneys as described under “Isolation of Sx from Kidneys.” GlcCer-IV 20:O 22:O 24:O 23:O 25:O migrated in a single band, both in solvent system V (Fig. 7) andthe solventsystemcontaining borate (40) (datanot shown). In order to confirm that thecomponent carbohydrate was exclusively glucose, GlcCer-IV was subjected to GLC analysis as thetrimethylsilylated methylglycosides. In the’H 908 NMR spectrum of GlcCer-IV, no resonances were observed between 5.4 and 5.6 ppm similar to Sx, suggesting that GlcCerIV may also lack a trans-double bond of 4-sphingenine and the cis-protons of unsaturated fattyacids. Furthermore, GLC analyses showed that GlcCer-IV had a ceramide similar to that of Sx, i.e. t180 andnonhydroxy fatty acids. The chemical shifts and thecoupling constants of GlcCer-IV were used for comparative studies withthose of Sx (Table 11). : 1: ‘jI: Ij The desulfated Sx prepared by solvolysis comigrated with GlcCer-IV and much faster than GalCer from human brain M I2 FIG. 6. Negative secondary ion mass spectrum of Sx. The (d18:1,24:0) in basic solvent system V(Fig. 7). In thissolvent table shows the peaks from the anionized molecular ion species system, it had been reported that GalCer migrated slower more hydroxyl groups the molecular corresponding to each ceramide with fatty acids combined with t180 than GlcCer and that the sphingoid. species contained, the slower these migrated (30, 31). Con-

I

1

I

, : I


5979

cosy1 sulfatide corresponds to about 2.7% of the concentration of SM4s and about 3.5-fold of SM3. The yields of SM4s and SM3 were close to the values in a previous report (15), i.e. 182 and 1.7 nmol/g of tissue, respectively. The amount of cholesterol 3-sulfate was about a half of SM4s. This result DISCUSSION was consistent with the value reported by Iwamori et al. (45). Glucosyl sulfatide, HS030-3Glc@1-1Cer,is a relatively miIn thisstudy, a novel sulfated glycosphingolipidcontaining nor component of the sulfated glycosphingolipids in rat kida sulfated glucosyl residue ( S x ) was isolated from rat kidney. On the basis of experimental results described in this paper, ney. The composition of the sphingoid base of glucosyl sulthe structure of Sx was proposed to be HSO~-3Glc~l-lCerfatide, however, wascharacteristic. Accordingly, it is expected (Fig. 8). There have been no previous reports of an existence that glucosyl sulfatide may play roles in the function of the of this sulfated glycosphingolipid in the biosphere (7). We kidney. designate this novel sulfatide as glucosyl sulfatide. REFERENCES Because of possession of a glucosyl residue, glucosyl sulfa1. Svennerholm, L. (1963) J. Neurochem. 10,613-623 tide showed some properties distinct from SM4s containing a 2. IUPAC-IUB Commission on Biochemical Nomenclature (1977) galactosyl residue. On TLC, glucosyl sulfatide moved much Lipids 12,455-468 faster than SM4s in the acidic solvent system, similar to the 3. Roberts, D. D. (1987) Methods Enzymol. 138,473-483 4. Yamakawa, T. (1984) in Ganglioside Structure,Function and fact that GlcCer moved much faster than GalCer in the basic Biomedical Potential (Ledeen, R. W., Yu, R. K., Rapport, M. solvent systems. In the'H NMR spectrum, the characteristic M., and Suzuki, K., eds) pp. 3-13, Plenum Publishing Corp., upfield shift of the resonance of H-2 ring proton in glucosyl New York sulfatide from that in SM4s was observed. Because the con5. Radin, N. S. (1983) in Handbook of Neurochemistry (Lajtha, A., figuration of the proton at theC-4 position in glucosyl sulfaed) pp. 163-177, Plenum Publishing Corp., New York tide was opposite to that in SM4s, the coupling constants of 6. Ishizuka, I., and Tadano, K. (1982) in New Vistas in Glycolipid and J4,J and the chemical Research (Makita, A., Handa, S., Taketomi, T., and Nagai, Y., H-4 with the vicinal protons (JB,~ eds) pp. 195-206, Plenum Publishing Corp., New York shift for H-4 of glucosyl sulfatide were distinct from those of 7. Ishizuka, I., and Yamakawa, T. (1985) in Glycolipids (Wiegandt, SM4s. For a similar reason, Sulph I-antibody, which bound H., ed) pp. 101-196, Elsevier Science Publishers B. V., Amsterto SM4s, did not bind to glucosyl sulfatide. dam The structure of the major ceramide moiety of glucosyl 8. Niimura, Y., and Ishizuka, I. (1986) J. Biochem. (Tokyo) 100, sulfatide was shown in Fig. 8. Glucosyl sulfatide did not 825-835 contain 4-sphingenine (d18:l) at all but contained 4-D-hy9. Makita, A. (1964) J. Biochem. (Tokyo) 55,269-276 droxysphinganine (t180). By contrast, SM4s from rat kidney 10. Mirtensson, E. (1966) Biochim. Biophys. Acta 116,521-531 contained predominantly d18:l (85.2%) (15). The trihydroxy 11. Ishizuka, I., Suzuki, M., and Yamakawa, T. (1973) J. Biochem. (Tokyo) 73,77-87 bases are the characteristic bases in the glycolipids from the 12. Yamakawa, T., Yokoyama, S., and Handa, N. (1963) J. Biochem. kidney of mammals such as equine (42) and dolphin (43), as (Tokyo) 63, 28-36 well as the small intestine (41). Other sulfated glycosphingo- 13. Ando, S., Isobe, M., and Nagai, Y. (1976) Biochim. Biophys. Acta lipids isolated from rat kidneys except for SM4s, such as SM2, 424,98-105 SB2, and SBla, also contained t180 as the major sphingoid 14. Ishizuka, I., and Yamakawa, T. (1968) J. Biochem. (Tokyo) 6 4 , 13-23 base (15-17). The characteristic ceramide structure for galacK., and Ishizuka, I. (1982) J. Biol. Chem. 257, 1482tosyl versus glucosyl sulfatides is in accordance with the 15. Tadano, 1490 different metabolic pathways for galactosylceramide and gan- 16. Tadano, K., and Ishizuka, I. (1982) J. Biol. Chem. 257, 9294glio-series glycosphingolipids. On the other hand, in a recent 9299 study which compared glycosphingolipidsof human epithelial 17. Tadano, K., Ishizuka, I., Matsuo, M., and Matsumoto, S. (1982) J. Biol. Chem. 2 5 7 , 13413-13420 and nonepithelial cells of ureter (44), it was shown that the epithelial mono- and diglycosylceramideshad ceramides com- 18. Folch, J., Lees, M., and Sloane Stanley, G . H. (1957) J. Biol. Chem. 226,497-509 posed mainly of C18 and C20 trihydroxy long chain bases in 19. Svennerholm, L. (1957) Biochim. Biophys. Acta 24,604-611 contrast to the nonepithelial glycolipids with predominantly 20. Green, J. P., and Robinson, J. D., Jr. (1960) J. Biol. Chem. 235, C18 dihydroxy long chain bases. This example suggests the 1621-1624 possibility that glucosyl sulfatide may belocalized in the 21. Kean, E. L. (1968) J. Lipid Res. 9, 319-327 22. Nagai, K., Ishizuka, I., and Oda, S. (1984) J. Biochem. (Tokyo) epithelial side. 95,1501-1511 Glucosylceramide (GlcCer-IV), which appeared to be the precursor of glucosyl sulfatide, i.e. GlcCer with 4-D-hydrox- 23. Hakomori, S. (1964) J. Biochem. (Tokyo) 55, 205-208 S. B., and Hakomori, S. (1987) Methods Enzymol. 138, ysphinganine and nonhydroxy fatty acids, was isolated from 24. Levery, 13-25 the neutral glycolipid fraction of rat kidney. The structureof 25. Sweeley, C. C.,and Moscatelli, E. A. (1959) J. Lipid Res. 1 , 40GlcCer-IV was determined by analytical TLC (44), composi47 tional analysis (data not shown), and 'H NMR spectroscopy. 26. Carter, H.E., and Gaver, R.C. (1967) J. Lipid Res. 8. 391-395 Among the glucosylceramides from rat kidney, the molecular 27. Tadano-Aritomi, K., and Ishizuka, I. (1983) J. Lipid Res. 24, 1368-1375 species with 4-~-hydroxysphinganine and hydroxy fatty acids 28. Zak, B. (1957) Am. J. Clin. Pathol. 27,583-588 were the major component, but theamount of GlcCer-IV was 29. Ishizuka, I., Inomata, M., Ueno, K., and Yamakawa, T. (1978) J. relatively small. Also in human ureter, the major monoglycoBiol. Chem. 253,898-907 sylceramide was GlcCer with 4-~-hydroxysphinganine and 30. Kean, E. L. (1966) J . Lipid Res. 7, 449-452 hydroxy fatty acids in both epithelial and nonepithelial cells 31.Ogawa, K., Fujiwara, Y., Sugamata, K., and Abe, T. (1988) J. Chromatogr. 426, 188-193 (44), in agreement with the results of this study. In the present study,the approximate yields (nmol/g of wet 32. Kushi, Y., Handa, S., and Ishizuka, I. (1985) J. Biochem. (Tokyo) 97,419-428 tissue) of SM4s, SM3, glucosyl sulfatide, and cholesterol 3- 33. Bax, A., and Drobny, G. (1985) J. Mugn. Reson. 61,306-320 sulfate from 60-day-old rat kidneys were201,1.6,5.5, and 34. Fredman, P., Mattsson, L., Andersson, K., Davidsson, P., Ishi89.3, respectively. In other words, the concentration of gluzuka, I., Jeansson, S., Minsson, J.-E., and Svennerholm, L.

sistent with the compositional analyses, the desulfated Sx migrated slower than GlcCer from the spleen of a Gaucher patient, which consisted of nonhydroxy fatty acids and predominantly d18:l.

5980

A Sulfated Glucosylceramide from Rat Kidney

(1988) Biochem. J . 252,17-22 35. Koerner, T. A. W., Prestegard, J. H., Demou, P. C., and Yu, R. K. (1983) Biochemistry 2 2 , 2676-2687 36. Harris, M. J., and Turvey, J. R. (1970) Carbohydr. Res. 15,5763 37. Gasa, S., Nakamura, M., Makita, A,, Ikura, M., and Hikichi, K. (1985) Carbohydr. Res. 137,244-252 38. Nagai, K., Toida, T., Iida, N., and Ishizuka, I. (1987) in Carbohydrates (Lichtenthaler, F. W., and Neff, K. H., eds) p. C-59,

Gesellschaft Deutscher Chemiker, Frankfurt, Federal Republic of Germany 39. Dabrowski, J., Egge, H., and Hanfland, P. (1980) Chem.Phys. Lipids 26,187-196

40. Sugita, M., Yamamura, T., Takamiya, Y., and Hori, T. (1984) Proc. Jpn. Conf Biochem. Lipids (in Japanese)26, 104-107 41. Karlsson, K.-A. (1982) in Biological Membranes (Chapman, D., ed) vol. 4, pp. 1-74, Academic Press, New York 42. Gasa, S., and Makita, A. (1980) J . Biochem. (Tokyo) 8 8 , 11191128

Nagai, K., Tadano-Aritomi, K., Kawaguchi, K., and Ishizuka, I. (1985) J . Biochem. (Tokyo) 98,545-559 44. Breimer, M. E., Hansson, G . C., and Leffler, H. (1985) J . Biochern. (Tokyo)98,1169-1180 45. Iwamori, M.,Moser, H. W., and Kishimoto, Y. (1976) Biochim. Biophys. Acta 441,268-279 43'