Mono-sulfated Globopentaosylceramide from Human Kidney*

2 downloads 0 Views 7MB Size Report
Structural Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of ... treatment, DEAE-Sephadex and silicic acid column.
THEJOURNAL OF BIOLOGICAL CHEMISTRY

Val. 264, No. 27, Issue of September 25, pp. 16229-16237,1989 Printed in U.S.A.

Mono-sulfated Globopentaosylceramide from HumanKidney* (Received for publication, November 14, 1988)

Ken-ichi :Nagai,”David D. Roberts,” e Toshihiko Toida,” Hajime Matsumoto,”. Yasunori Kushi,’ Shizuo Handa,e and Ineo Ishizuka”. From the “Departmentof Biochemistry, Teikyo Uniuersity School of Medicine, Ztabashi-ku, Tokyo173, Japan, the bLaboratory of Structural Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Marvland 20892. and the eDeDartmentof Biochemistry, - . Facultyof Medicine, Tokyo Medical and Dental University, Bunkyo-ku, Tokyo 113, Japan

A novel sulfated glycosphingolipid that belongs to various tissues of vertebrates, echinoderms, and microorganisms and their structureswere established (see Refs. 1-3 for the “globo-series” was isolated from human kidney. This lipid was purified froma pooled kidney prepara- review). These sulfated amphiphiles may play roles in the tion by chloroform/~nethanol extraction, mild alkaline maintenance of ionic homeostasis (4-7) and blood coagulation treatment, DEAE-Sephadex and silicicacid column (reviewed in Ref. 3) in mammals. chromatographies, ,and preparative thin layer chroRecently, Roberts et al. (8-10) reported that the adhesion matography. The structure and the properties were glycoprotein laminin, the platelet glycoprotein thrombosponstudiedbyinfrared spectroscopy,two-dimensional din, and von Willebrand factor bind specifically to SM4s.l proton magnetic reslonance spectroscopy, negativesec- Furthermore, using lZ5I-vonWillebrand factor (ll),two unondary ion massspectrometry, solvolysis, composi- known complex sulfated lipids were detected in crude monotional and methylation analyses, monoclonal antibodies, and sulfatide-bimding proteins. Fromthe resultsof sialosyl and monosulfated fractions of human kidney, which the above analyses, the structure of this glycolipid was have been reported to contain only two sulfated glycolipids, SM4s and SM3 (12). This observation prompted us to isolate proposed to be IIS03-3Gal~l-3GalNAc/31-3Galaland characterize the new sulfated lipids in human kidney. 4Ga1~1-4Glc~l-lceramide. The proton resonance at The present studydescribes the structureof a novel sulfated 3.93 ppm of the H43 of the sulfated nonreducing terminalgalactose of this lipid was downfield-shifted glycosphingolipid (designated as SB)belonging to the“globo(A0.48ppm), as compared with H-3of the internal8- series” from human kidney and the reactivity of this new some monoclonal antibodies and the galactose becauseof the electronegativityof the sulfate sulfated amphiphile with ester. This sulfated lipid reacted with a monoclonal sulfatide-binding proteins. anti-SSEA-3 (MC-631) (Kannagi, R., Cochran, N. A., EXPERIMENTAL PROCEDURES’ Ishigami, F., Hakomori, S., Andrews, P. W.,Knowles, B. B., and Solter, D.. (1983)EMBO J. 2,2355-2361), RESULTS on whose epitope is R-3GalNAcB1-3Galal-4GalBl-R’, thinlayerchromatogramsandsolid-phaseradioTLC ofPurifiedAcidic Glycolipid SB-Isolated SB was immunoassay. This lipid also bound to the ‘251-labeled examined by TLC in three solvent systems with various sulfatide-binding protein, thrombospondin. The yield of this sulfated glycolipid was 0.19 nmol/g of tissue, sulfated glycolipids as references (Fig. 3). SB (lune 9 ) had a solvent which was about0.09 and 0.5 mol 7’0 of galactosyl and mobility close to that of SBla (lane 6) in the neutral lactosyl sulfatidesin human kidney. ’ The abbreviations used are: SM4s, galactosylceramide sulfate,

Sulfated glycolipids are aclass of acidic glycolipids containing one or two sulfate esters on their oligosaccharide chains. To date, nearly 20 sulfated glycolipids have been isolated from * This work was supported in part by Grant-in-aid 63780222 for Encouragement of Young Scientists and by Grant-in-aid 61304034 for Co-operative Research (A) from the Ministry of Education, Science and Cultureof Japan. Parts of this study were presented a t the 60th Annual Meeting of the Japanese Biochemical Society (October, 1987, in Kanazawa) and at the29th International Conference on the Biochemistry of Lipids (September, 1988, in Tokyo). 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. Present address: Laboratory of Pathology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892. Present address: Dept.. of Orthopedics, Teikyo University School of Medicine, Itabashi-ku, Tokyo 173, Japan. ‘To whom correspondence should be addressed.

GalCer 13-sulfate; GalNAc, N-acetylgalactosamine; Cer, ceramide; GalCer, Galpl-1Cer; LacCer, lactosylceramide, Galpl-4Glcpl-lCer; GbsCer, globotriaosylceramide, Galal-4Ga1/31-4Glc~l-1Cer;Gb4Cer, globotetraosylceramide, GalNAc~l-3Galal-4Gal~l-4Glc~l-lCer; GbsCer, globopentaosylceramide, Gal~l-3GalNAc~l-3Galal-4Gal~l4Glcpl-lCer; Gg3Cer, asialo GM2, gangliotriaosylceramide, Gal NAc/3l-4Gal~l-4Glc~l-1Cer; Gg4Cer, asialo GM1, gangliotetraosylceramide, Gal~l-3GalNAc~l-4Gal~l-4Glc~l-lCer; SM3, lactosylceramide sulfate, LacCer 113-sulfate;GM3 (NeuAc), I13aNeuAc-LacCer; GD3 (NeuAc-NeuAc), I13a(NeuAca2-8NeuAc)-LacCer; SM2, gangliotriaosylceramide sulfate, Gg3Cer 113-sulfate;SB2, gangliotriaosylceramide bis-sulfate, GgSCer 113,1113-bis-sulfate;SBla, gangliotetraosylceramide bis-sulfate, Gg4Cer I13,1V3-bis-sulfate;SMla, gangliotetraosylceramide sulfate, Gg,Cer 113-sulfate; SB, globopentaosylceramide sulfate, GbsCer V3-sulfate; Hex, hexose; HexNAc, N-acetylhexosamine; d18:1, 4-sphingenine; t180, 4-hydroxysphinganine; Me2S0, dimethyl sulfoxide; HPTLC, high performance TLC; GLC, gas liquid chromatography; GC-MS, gas chromatography-mass spectrometry; COSY, chemical shift correlated spectroscopy. Portions of this paper (including “Experimental Procedures” and Figs. 1 and 2) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

16229

Gb5Cer V-sulfateof Human Kidney

16230

S B were hexadecanoic (28.6 weight %), tetracosanoic (28.0%), docosanoic (25.5%),and octadecanoic (11.1%). Hydroxy fatty acids could not be detected in this lipid by either GLC and GC-MS. SB contained 4-sphingenine (68.7%) and 4-hydroxysphinganine (31.3%) as thesphingoid bases. From the chemical shifts and coupling constants of onedimensional 'H NMR spectrum of SB at 60 "C (see Fig. 6A and Table I), it could be deduced that SB contains one aA 1 2 3 4 5 6 7 0B9 l 2 3 4 5 6 7 0C9 1 2 3 4 5 6 7 0 9 hexose, three @-hexoses,and one @-N-acetylhexosamine.SB may belong to globo-series, judging from the triplet signal of FIG.3. TLC of the purified SB. Plates were developed with a neutral solvent system VI (chloroform/methanol/water,603523, v/v) H-5 of a-galactose (4.12 ppm) (24, 25). The @-glucoseof SB (panel A ) , a basic solvent system 111 (chloroform/methanol/3.5M may be bound directly to the ceramide as supported by the triplet signal of H-2 of @-glucoseat 3.047 ppm.The peak area NH,OH,60:40:9,v/v) ( p a n e l B ) , andanacidicsolvent system V (chloroform/methanol/acetone/CH~COOH/water, 62:42:1, of R5 (H-5 proton) of the sphingenine in olefinic region was v/v) (panel C). The bands were stained by spraying with orcinol approximately 66% of those of the anomeric protons inagreereagent. Lane I, shrew cerebralacidic lipids; lane 2, SM4s from human ment with the value of 4-sphingenine obtained by GLC. kidneys; lane 3, SM3 from human kidneys; lane 4, SM2 from rat The content of sulfate ester on SB was estimated by the kidneys; lane 5, SB2 from rat kidneys; lane 6,SBla from rat kidneys; method of Tadano-Aritomi and Ishizuka (35). The color yield lane 7, GM3 (NeuAc) from human kidneys; lane 8, CD3(NeuAc-, NeuAc) from dolphin kidneys; lane 9, SB. of peracetylated SB (9.3/100 nmol) wasclose to that of peracetylated SM4s (10.1/100 nmol), indicating that SB is a system (Fig. 3A), whereas this lipid migrated slower than GD3 monosulfated glycolipid. This conclusion is also supported by ( l a n e 8) in both basic (Fig. 3B) and acidic (Fig. 3C) solvent the coelution of SB with other monosulfated glycolipids from systems. SB was not stained by Dittmer and Lester reagent the DEAE-Sephadex column (Fig. 1). Negative Secondary Ion-Mass Spectrometry-Negative sec(data notshown) (51), indicating the absence of phosphate in ondary ion-mass spectrometry analysis of SB was performed the molecule. IR Spectroscopy-The infrared spectrum of SB was similar (Fig. 4). Intense pseudomolecular ions (M minus H)- were to that of SM2 (52), except for the much larger absorptions detected at m/z 1551, 1579, 1593, and 1611 corresponding to of the OH (3100-3600 cm") and C-OH (1060 cm") relative the major ceramide species: 22:0, 24:0, 25:O plus 4-sphingento -CH,- (2920 and 2840 cm"). The absorptions at 1240 (S ine, and 250 plus 4-hydroxysphinganine. Among them, the combination of tetracosanoic acid and 4-sphingenine gave the = 0 stretching) and 810 cm" (C-0-S vibration) couldbe ascribed to anequatorial sulfate ester of hexose (52,53). The most prominent peak (m/z 1579). Furthermore, characteristic absence of absorptions at 850 and 1740 cm" suggested that fragment ions, which were cleaved sequentially at the glycoS B lacks in axial sulfate esters and carboxyl esters (32). The sidic linkage, could be obtained. The ion arising from the absorption at 890 cm" indicated the presence of at least one terminal sulfated hexose was observedat m/z 241 (HS03-Hex @-anomeric configuration of hexopyranoside (32,53, 54). minus 2H)" The small peak at m/z 444 wasderived from the These results, the staining characteristic on TLC, and the nonreducing terminal disaccharide with a sulfate ester, anionic nature strongly suggested that SB is a sulfated gly- (HS03-Hex-O-HexNAcminus 2H)-. Similarly, the ions at m/ z 606 and 622 (sulfated trisaccharide), m/z 768 and 784 cosphingolipid. Compositional Analysis-SB contained galactose, glucose, (sulfated tetrasaccharide), and m/z 946 (sulfated pentasacand N-acetylgalactosamine in a molar ratio of 3.0:l.Ol.l. No charide) were observed. sialic acid was detected in this lipid. The major fatty acids of The above data suggest that SB has a nonbranching sac*241 243- 2 2437

4444 446-2

*606

608-2

*768 770- 2

930 932-2

770-1

g321

HSO 9 - O - H e x1 1 0 ~ H e x N A c ~ l H e xI i - 0, ~ H e Ix 1 0' 1 H eI x - O ' lCer I

FIG.4. Negative secondary ion-

mass spectrometry of underivatized 462-2 sulfated glycolipid SB. Hex, hexose; i2C HexNAc, N-acetylhexosamine; Cer, cer. 241'255 "'

'

R 259 -

I

~

462 786-2 460

259-2 257

amide; dl8:1,4-sphingenine; t18:0,4-hydroxysphinganine;M - H , the pseudomolecularion (the molecular minus 1). Mass numbers indicated with open stars in the formula and larger figures in the spectrum are intense signals in the spectrum. The peaks in the mass ranges higher than m/z 500 and m/z 700 were amplified 50- and 200-fold, respectively.

624624-2 *622

786' *784

948A

4!&2 i

I

1228

::22

M - H 1551 = d 1 8 : l + C22:O 1579 = d 1 8 : l + C24:O 1593 = d 1 8 : l

+

C25:O

1611 = t 18:O + C25:O :123

:222

!122

lL22

:s2z

Gb5Cer V-sulfate of Human Kidney

16231

the latter ion corresponds to the di-0-methyl-mono-0-acetylated fragment ion arising from C-4 to C-6 of hexose and the secondary ion formed by a loss of CH30H, respectively. The characteristicions were observed onboth the peak of 4substituted galactitol and glucitol acetates at m/z 233 and 162, which were derived from the structurecontaining c - 3 to C-6 of the 4-substituted hexoses and the structure containing C-1 toC-3 of the hexoses, respectively (23) (see Fig. 5A,peaks 2 and 3). The peak of 4,6-di-O-methyl-2-N-methylacetamiTI1 dogalactitol acetate contained the characteristicfragment ions, m/z 159 (or 117), 171, and 275 (or 243), which arose from the structures containing C-1 to C-2 of hexosaminitol 1181 0.70 (or minus ‘CH2C=O), C-1 to C-3 of the 3-substituted hexosaminitol with a loss of CH3COOH,and C-1 to C-4 of the 3substituted hexosaminitol (orminus CH30H), respectively (23). From these results, the structure of SB could be tentatively assigned as HS03-3Gal~l-3GalNAc@l-3Galal-4Gal~l4Glcpl-lCer. Methylation-Solvolysis-Remethylation Analysis-The substitution of the sulfate group at C-3 of the terminalgalactose of SB was further confirmed by a methylation-solvolysisremethylation technique (22). Permethylated SB was solvolyzed to remove the sulfate ester and thenremethylated using C[’HI3I to label the position of sulfate substitution on galactose (23).The desulfation and remethylation of permethylated SE resulted in disappearance of 1 mol of the acetateof 2,4,6tri-0-methylgalactitol (compare the size of peak 1 in Fig. 5,A and E ) and appearance of 1 mol of the acetate of 2,3,4,6110 1 0 tetra-0-methylgalactitol (Fig. 5B, peak 4 ) . The characteristic fragment ions, m/z 162,205, and 145 (205 minus CH&OOH), which were derived from the structurescontaining C-1 to C1 3 of the hexose and C-3 to C-6 of the hexose, respectively, shifted upward by three mass units to m/z 165, 208, and 148 (208 minus CH3COOH) (Fig.5, B, peak 4 , and C ) . The appearance of 2,4,6-tri-O-methyl-3-O-deuteriomethylgalacti1 0 ma to1 acetate established the attachment of the sulfate group at FIG. 5. Mass spectrum and mass chromatogram of acetates C-3 of the nonreducing terminal galactose. of partially methylated-deuteriomethylated alditols derived Proton-NMR Spectroscopy-The spectrum of the whole from desulfated SB. S13 was permethylated and divided in half. and amplified anomeric proton region (inset) of sulfated Half was acetolyzed and reduced by NaB[*HI4,while the other half was solvolyzed, and then remethylated with C[’H]J and reduced as glycolipid SB are shown in Fig. 6A. Fig. 6B shows a doublerelayed COSY spectrum of SB. The assignments of the signals described under “Experimontal Procedures.” The acetates of partially methylated alditols were determined by GC-MS (3% SP-2340 col- shown in Table I are based on the results obtained by the umn). Panel A , mass chromatogram (before solvolysis);panel B, mass multiple-relayed COSY (COSY, relayed COSY, and doublechromatogram (after solvolysis);panel C, mass spectrum of 2,4,6-tri- relayed COSY) spectra and the systematicstudies of the 0-methyl-3-mono-0-deuteriomethylgalactitol acetate. The peaks identified were acetates of 1, 2,4,6-tri-O-rnethylgalactitol; 2, 2,3,6-tri- globo-series glycosphingolipids from human teratocarcinoma (24), mouse kidney (25), and of the various series of glycolipids 0-methylgalactitol; 3, 2,3,6-tri-O-methylglucitol; 4, 2,4,6-tri-0methyl-3-mono-0-deuteriomethylgalactitol. Arrows indicate the peak including gangliosides (55). which was derived from a non-sugar contaminant. Mass chromatoThe well-separated spectral region of the anomeric protons grams were obtained by collecting the ions of rn/z 118, 145,148,161, of SB showed the presence of five monosaccharides. A doublet 162, 165, 208, 233 (panel .€I), and also 236 (panel A ) . Closed circles = 7.3 Hz) coupled to H-2 of H-1 resonance at 4.184 ppm (J1,2 indicate the ions shifted by three mass units after trideuteriomethy(J2.3 = 7.8 Hz), anda doublet at 4.26 resonances at 3.047 ppm lation. Ordinate, ion intensity (figures preceded by an asterisk indicate = 6.8 Hz) in two-dimensional COSY spectrum (Fig. the amplification of the intensity); abscissa, retention time; TII, total ppm (J1,2 ion intensity. 6B) can unequivocally be ascribed to a @-glucoseand a Pgalactose of the lactose structure next to ceramide (22, 24, 56). In the one-dimensional spectra (Fig. 6, A and E (lower charide core with one sulfate ester on the hexose at the panel)), the signal of @-glucosewas smaller than the other nonreducing terminal. Based ontheseresults, the partial structure of SB could ble assigned as HSO3-(2, 3 or 6)GalP- anomeric peaks. In addition, another small doublet was de= 7.8 Hz) tected at a slightdownfield position (4.194 ppm; J1,2 GalNAcP-Gala-GalP-GlcP-Cer. Methylation Analysis--Methylated SB yielded the acetates (Table I). Itmay be probable that thesignal of P-glucose was of 2,4,6-tri-O-methylgalactitol,2,3,6-tri-O-methylgalactitol, split, reflecting the combination of sphingoid bases to which 2,3,6-tri-O-methylglucitol, and 4,6-di-O-methyl-2-deoxy-2-N-the glucose was bound (see “Discussion”).A pair of signals of methyl-2-acetamidogalactitol in a molar ratio of 2:l:l:l. = 8.3 Hz) was a proof for H-1 resonance at 4.584 ppm (J1,2 The identification of these peaks was also achieved by GC- the presence of a P-N-acetylgalactosamine (23). The resoMS analysis. The mass chromatogram indicated the coinci- nance at 4.319 ppm (J1,* = 7.8 Hz) was assigned to a 3-sulfated dence of the specific ions at m/z 161 and 129 with the peak P-galactose linked to C-3 of N-acetylgalactosamine (23). The of 2,4,6-tri-O-methylgal.actitol acetate (22). The former and signal of H-3 of this galactose was greatly shifted to the

cr““

W

I

-

L

U

L

16232

Gb&er vj-sulfateof Human Kidney

A

4113

(HSO,) -3Ga131-3

TMS

1i

I

1

1

i'~""~/""l~"'l''''~'~~'l'r 4 . 9 4.8 1 . 7 4.6 4.5 4.4 4 . 3 4 . 2

N~acetyl Gal.? flb5

4.1

Anomeric regio region

FIG. 6. Proton magnetic resonance spectrum of sulfated glycolipid SB from human kidney. The purified S B (340 nmol) were treated repeatedly with 0.5-ml portions of CH30['H], followed by desiccation over P20,in uacuo to exchange the labile protons with deuteron. Then thoroughly dried S B was redissolved in 0.5 ml of a mixture of [2H]-Me2S0/[2H]z0, 98:2, v/ v (final concentration, 0.68 mM). Chemical shifts were indicated by ppm from the signal of tetramethylsilane (2"s) as an internal standard. Panel A, onedimensional spectrum of glycolipid SB. Inset, the signal intensities of the region between 4.08 to 4.92 ppm were amplified 3-fold.Panel B, two-dimensional doublerelayed COSY spectrum of SB. Arabic numerals ( 1 - 4 ) indicate the ring protons of monosaccharides. The solid, broken, and dotted circles indicate the crosspeaks between H-4 and H-5 of G a h l - 4 residue, between H-3 and H-4, and H-1 and H-2 of GlcP1-1 residue, respectively (these peaks couldbe observed only when the threshold of the intensity was set at a lower level) (data not shown).

B

4 Q

k

downfield (3.93 ppm) because of the electronegativity of the sulfate ester (Fig. 6 B , upper panel) (57, 58). The additional doublet at theextreme downfield position (4.804 ppm) and a = 3.4 Hz)were compatible with vicinal coupling constant (Jl,z an internal a-galactose linked to C-4 position of a galactose, which resonated at 4.791-4.88 ppm (24,25,55).These protonNMR parameters of the anomeric region of SB well coincided with those of the Gb&er and the V3aNeuAc-Gb&er (24)

2.

3

e @

I-1

(Table I). Additionally, the presence of H-5 proton of Galal4 at 4.117 ppm (see the cross-section, designated as a solid circle, of H-4 of Gala residue in Fig. 6B, upperpanel)indicates that SB has the structure belonging to globo-series, as mentioned by Kannagi et al. (24) and Sekine et al. (25). TLC-Immunostaining of Intact and Desulfated SB and Partially Purified AcidicLipids from Human Kidneys-SB has an internal sugar sequence (R-3GalNAc@l-3Galal-4Gal@l-R')

Gb&er V-sulfate Kidney of Human

16233

TABLEI Chemical shifts (6) and coupling constants (J) of H-I to H-4 protons ofSB One-dimensional (Fig. 6A) and two-dimensional (Fig. 6B) spectra of SB were obtained at 60 and 30 "C, respectively. Chemical shift Glycolipids Galpl-4 Galal-4

GalNAcpl-3 Galpl-3

Glcpl-1Cer ppm

GL-5" GL-7" SB

H-1 H-1 H-lb

4.20 4.24 4.325

4.61 4.57 4.615

4.80 4.81 4.828

4.26 4.26 4.27

H-1'

4.319

4.584

4.804

4.26

H-2' H-3' H-4'

3.48 3.93 3.974

3.88 3.69 3.861

3.78 3.62 3.974

3.30 3.45 3.818

3.047 3.37 3.30

Hz 3.9 2.9 3.9

7.3 7.8 7.3

7.3 8.3 7.8

3.4

6.8

7.3 (7.8)d

7.1 3.7

8.6 2.0

Glycolipids

4.184 (4.194)d

Coupling constant

GL-5" GL-7"

J,,*

SB

J1,ze

7.3 7.8 7.8

J1.J

7.8

52.1

9.2 3.9

51.2

8.3

4.17 4.19 4.183

J3,b

8.3 8.3 8.3

8.8 3.9

7.8 6.6

"Data (at 29 "C)on GL-5 (GbsCer)and GL-7 (V3aNeuAc-Gb5Cer)were taken from Kannagi et al. (24). Data obtained at 60 'C, accurate to fO.001 ppm. e Data obtained at 30 "C, accurate to fO.OO1 ppm. H-1 signal of 0-glucose attached to 4-hydroxysphinganine. e Data obt,ained at 60 ' C , accurate to f0.49 Hz. 'Data obtained at 30 "C, accurate to f0.49 Hz.

that is recognized by an. anti-SSEA-3 antibody (MC-631) (45, 46) in the carbohydrate structure. Thus,we examined whether this sulfated lipid reacts with this monoclonal antibody. In Fig. 7B (lane b), 50 pmlol of the lipid was strongly stained on TLC-autoradiogram. The lipid could also be detected inacidic lipid extracts from two' individuals (Fig. 7 B , lunes c and d ) and a pooled kidney pre:paration (Fig. 7 B , lane e). Thepurified lipid did not bind anti-SSEA-4 which requires sialic acid for binding (46) (data notshown). Gangliosides which may belong to globo-series and react with this antibody were present in two of the kidney preparations (Fig. 7 B , lanes c and d ) . We have found that these gangliosides also bind with the antiSSEA-4.3Similar reactivity of S B with the monoclonal antiSSEA-3 antibodywas observed by solid-phase radioimmunoassay. S B (6 ng/well) bound anti-SSEA-3 strongly, whereas other sulfated glycolipids (SM4s and SM3) did not bind the antibody at levels up to 50 ng/well. Reactivity of solvolyzled S B with anti-SSEA-3antibody was studied by using TLC-enzymoimmunostaining (Fig. 7C). The faint band (lune 5, stained by orcinol) that comigrated with mouse kidney Gb5Cer (lane 3 ) was clearly detected by antiSSEA-3 antibody (lunes 8-10). Gb5Cer has a similar mobility to Gg4Cer in this neutral solvent system. However, the solvolyzed S B was not stained by polyclonal anti-Gg4Cer antibody. In addition,a small amount of Gb3Cerwas in the solvolysate detected by monoclonal anti-Gb3Cer antibody on TLC-enzymoimmunostainingas one of the degradation products from Gb5Cer (data notshown). These results further confirmed the tentatively assigned structure for SB, which was deduced from the data of the methylation, the negative

secondary ion-mass spectrometry and the 'H NMR analyses. In conclusion, we propose thestructure of S B is HSO33Ga1/31-3GalNAc@l-3Galal-4Gal~l-4Glc/3l-lCer (Fig. 8). Interaction with the Sulfatide-binding Proteins-In a previous study using von Willebrand factor ( l l ) , SM4s, SM3, and a putative trihexosyl-sulfatedglycolipid were detected in human kidney. We therefore reexamined the binding of several sulfatide-binding proteins to kidney lipids and to the purified SB. Thrombospondin bound to S B on TLC(Fig. 9B, lane d ) , although the binding was weaker than to a 25-fold lower concentration of SM4s. Using high concentrations of kidney acidic lipid extracts, several slow migrating components were stained with thrombospondin (Fig. 9B, lanes e-g). The mobilities of these components appear to be retarded relative to thepurified lipids due to thelarge amount of total lipid applied to each lane.The multiple active lipids, however, are consistent with the presence of four minor sulfated glycolipids SA, S B , X,and Y in these preparations.Weak binding of laminin and von Willebrand factor to S B was also detected on TLC (data notshown). DISCUSSION

A novel monosulfated glycosphingolipid S B belonging to globo-series was isolated from human kidney, and the structure was fully characterized. On the basis of experimental results from IR spectroscopy, 'H NMR, negative secondary ion-massspectrometry, GLC, methylationanalysis, solvolysis, and TLC-immunostaining, the complete structure of S B is proposed to be HS03-3Gal/31-3GalNAc/31-3Galal-4Gal~l4Glc@l-lCer(Gb5CerV3-sulfate) (Fig. 8). The yield of Gb5Cer

GbBCerV-sulfate of Human Kidney

16234

human kidney (12, 21). To date, there have been no reports of sulfated glycolipids containing the globo-series carbohydrate structure. On two-dimensional COSY spectrum, the signal (3.93 ppm) of the H-3 proton of the nonreducing terminal @-galactose, which has a sulfateester at C-3 hydroxy group, was shifted to the downfield (A0.48 ppm), as compared with H-3 proton of the internal @-galactose(Fig. 6B and Table I). This observation is in good agreement with the results reported by Harris and Turvey (57) on several sulfated glycosides and by Gasa et al. (58) on SM4s. The downfield shift of H-3 protonshould be explained in terms of a deshielding effect of sulfate ester

a

b

c d

e

f a

b

c

d

1

-.

!

I

e .-%

b

rrcpnr

’”

on the geminal equatorial proton. The anomeric proton of this sulfated galactose also shifted to a slightly downfield position (A0.125 ppm) as compared with Gb5Cer (Table I) probably by the same mechanism to theeffect of 3-sulfates of SM2 (21), SB2 (22), SBla (23), and gangliotetraosylceramide sulfate (Gg,Cer IV3-sulfate) (59). Splitting of the H-1 signal (4.184 and 4.194 ppm) of @-glucosethat is linked to the sphingoid base was observed (Fig. 6 and Table I). Yamada et al. (60) reported that H-1 proton of galactosylceramide with 2-hydroxy fatty acid shifted to downfield by 0.04 ppm, as compared with that of galactosylceramide with nonhydroxy fatty acid. They suggested that the downfield shift of the anomeric proton of galactose arose from a “through-space interaction” with the neighboring ceramide fragments, namely the 2-hydroxy group of fatty acid. By analogy, an additional hydroxy group of 4-hydroxysphinganine might have deshielded the anomeric proton of @-glucoseanisotropically, although the downfield shift (0.01 ppm) was not so large. This supposition also coincides with the fact that the ratio (40 and 60%) of the intensities of two doublets (4.194 and 4.184 ppm) for &glucose is reasonable for the composition of the sphingoid bases of SB (t18:0,31% and d18:1, 69%). Gb5CerV3-sulfate reacts with the monoclonal anti-SSEA-3 antibody (MC631) (45,46) thatrecognizes the internal strucof antigens on TLCture, R-3GalNAc@l-3Galal-4Gal@l-R’, immunostaining (Fig. 7) and solid-phase radioimmunoassay. The structure of “ R moiety, HSO3-3GalD1- of Gb5Cer V3sulfate appeared not to interfere with the recognition by this antibody. On the other hand, the monoclonal anti-SSEA-4 antibody (MC813-70) (45,46) whose epitope is reported to be NeuAca2-3Gal/31-3GalNAc@l-R did not bind Gb5Cer V3-sulfate on TLC. This may indicate that the presence of NeuAc

i

1 2 3 4 5 6 7 8 910 FIG.7. TLC-immunostaining of intact and desulfated sulfated glycolipid SB and partially purified kidney lipids. Panel A, TLC of S B and kidney acidic lipids developed in solvent system IV (chloroform/methanol/0.2% CaC12, 60:35:7,v/v) and stained with orcinol reagent: lane a, rat brain acidic lipidssupplemented with GM3 and GM2; lane b, GhlCer (1 nmol); lane c, SB (0.5 nmol); lanes d-f, partially purified acidic lipids from three humankidney preparations (100 mg wet weight). On lane a, two bands that migrated between Gb4Cer and S B were GM3 (upper) andGM2 (lower). Panel B, autoradiogram of TLC-immunostaining(anti-SSEA-3)insolvent system IV; lane a, Gb4Cer (0.2 nmol); lane b, SB (0.05 nmol); lanes ce, partially purified acidic lipids from three human kidney preparations (100 mg wet weight) (corresponding to lanes d-f in panel a ) . The preparations in lanes c and d are from kidneys from single patients andwere eluted from DEAE-Sepharose using step gradients and also containing monosialosyl gangliosides reacting with antiSSEA-3. These were removed from the thirdpooled kidney preparation by gradientelutionfromDEAE-Sephadex (lane e ) . Panel C, TLC-enzymoimmunostaining (anti-SSEA-3) of the solvolyzed S B in solvent system VI (chloroform/methanol/water, 603523, v/v): lanes 1-5, stained by orcinol reagent; lanes 6-10, enzymoimmunostaining: lane I, cerebral total acidic lipidsof rat; lane 2, human kidney GbsCer (1.2 nmol); lane 3, mouse kidney GbsCer (0.5 nmol); lane 4 , intact SB (0.5 nmol); lane 5 , desulfated SB (0.1 nmol); lane 6, mouse kidney GhsCer (0.13 nmol); lane 7, intact SB (0.1 nmol); lane 8, desulfated SB (0.1 nmol); lane 9, desulfated S B (0.06 nmol); lane I O , desulfated SB (0.02 nmol).

a b c d ef g h a b c de f g FIG. 9. Autoradiogram of thrombospondin binding to SB and kidney lipids. Purified lipids or partially purified acidic lipids from human kidneys were chromatographed on aluminum-backed HPTLCplates in solvent system IV (chloroform/methanol/0.2% CaC12, 6035:7,v/v) andstained withorcinol (panel A ) or ‘“1thrombospondinas described under“Experimental Procedures” (panel B ) . Panel A : lane a, rat brain acidic lipids supplemented with GM3 and GM2; lane b, SM4s (2 nmol);lane c, SM3 (1nmol); lane d, SM2 (1 nmol); lane e, SB (0.5 nmol); lanes f-h, partially purified acidiclipids from three human kidney preparations (100 mg wet weight) (same as lanes d-f in Fig. 7A). On lane a, two bands that migrated between SM2 and SB were GM3 (upper) and GM2 (lower). Panel B: lane a, SM4s (0.02 nmol); lane b, SM3 (0.44 nmol); lane c, SM2 (0.5 nmol); lane d , SB (0.5 nmol); lane e-g, acidic lipids from three humankidney preparations (150 mg wet weight) (corresponding to lanes f-h in panel A).

16235

Gb5Cer V-sulfate of Human Kidney residue on the terminal Gal of GbsCer must be essential for antibody binding and that a sulfate ester could not replace the sialic acid. Gb5CerV3-sulfatebinds to the cell adhesive protein, thrombospondin (Fig. 9). Binding to Gb5Cer V3-sulfate (0.5 nmol) was similar to binding to the same level of SM2. However, the affinity of thrombo3pondin for Gb&er V3-sulfate is much weaker than for SM4s .and SM3. The weak binding observed here is consistent with the previous finding that thrombospondin strongly prefiers sulfated glycosphingolipids with short oligosaccharide chains (11)and may explain why this lipid was not detected previously. Gb&er V3-sulfate could be detected in all the renal acidic lipid extracts testedwith the anti-SSEA-3antibody (Fig. 7 B , lunes e"). The concentrations of this sulfated glycolipid in kidney may vary with the individual, deduced from the staining intensities. Therefore, the monoclonal anti-SSEA-3 antibody (MC-631) may be useful for surveying the contents of this sulfated glycolipid in various tissues. Further investigations should be necessary to establisha biosynthetic and biodegradative pathwa:ys and to clarify a biological significance of this novel sulfated amphiphile in human kidney.

16. 17. 18. 19. 20. 21. 22. 23.

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

42.

REFERENCES

47.

1. Radin, N. S. (1983) in Handbook of Neurochemistry (Lajtha, A., ed) 2nd Ed., Val. 3, pp. 163-177, Plenum Publishing Co., New York 2. Makita, A., and Taniguchi,,N. (1985) in Glycoli ids (Wiegandt, H., ed) Vol. 10, pp. 1-99, Elsevler Sclentlfic Publishing lo., Amsterdam 3. Roberts, D. D. (1987) Methods Enzymol. 138,473-483 4. Karlsson, K.-A. (1982) in Biological Membranes (Chapman, D., ed) Vol. 4, pp. 1-74, Academic Press, London 5. Ishizuka, I., and Tadano, K. (1982) Adu. Exp. Med. Biol. 162,195-206 6. Ishizuka, I., and Yamakalwa, T. (1985) in Glycolipids (Wiegandt, H., ed) Vol. 10, pp. 101-196, El.sevier Scientific Publishing Co., Amsterdam 7. Niimura, Y., and Ishizuka, I. (1986) J . Biochem. (Tokyo) 100,825-835 8. Roberts,.D. D., Rao, C. hi., Magnani, J. L., Spitalnik, S. L., Liotta, L. A,, and Glnsburg, V. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 1306-1310 9. Roberts, D. D., Haverstick, D. M., Dixit, V. M., Frazier, W. A., Santoro, S. A., and Ginsburg, V. (1'985)J. Biol. Chem. 260,9405-9411 10. Roberts, D. D., Williams, S. B., Gralnick, H. R., and Ginsburg, V. (1986) J. Biol. Chem. 261. 3306-3309 11. Roberts, D. D., RaO,C. N., Liotta, L. A., Gralnick, H. R., and Ginsburg, V. $1986)J.Biol. Chem. 261,,6872-6877 12. Martensson, E. (1966) Blochrm. Biophys. Acta 1 1 6 , 521-531 13. Nagai, K.-i., Ishizuka, I., and Oda, S. (1984) J. Biochem. (Tokyo) 9 6 , 15011511 14. Makita, A. (1964 J. Biochem. (Tokyo) 66,269-276 15. Rauvala, H. (1976) Biochim. Biophys. Acta 424,284-295

48.

~~

~

Y

~~

24.

Acknowledgments-We are grateful to Drs. S. Hakomori (The Biomembrane Institute and theUniversity of Washington, Seattle), R. Kannagi (KyotoUniversity,Kyoto), P. W. Andrews, B. B. Knowles, and D. Solter(TheWistarInstitute, Philadelphia) for kindly supplying the monoclonal antibodies. Special thanks are due to Drs. M. Miura and S. Rdoteki of Kosei General Hospital, Nakanoku, Tokyo, for supplying kidney tissues. We are obliged to Drs. A. Suzuki and M. Sekine (The TokyoMetropolitan Institute of Medical Science, Tokyo) for valuable advice for TLC-immunostaining andfor kindly supplying Gb&er of mouse kidney.

I

Rauvala, H. (1976) J. Biol. Chem. 2 6 1 , 7517-7520 Makita, A,, and Yamakawa, T. (1964) J. Biochem. (Tokyo)66,365-370 Kawanami, J. (1968) J. Biochem. (Tokyo)64,625-633 Makita, A., and Yamakawa, T. (1963) Jpn. J. Exp. Med. 33,361-368 Naeai. K.-i.. Tadano-Aritomi. K.. Kawaeuchi. K.. and Ishizuka. I. (1985) J. i?iochern.'( Tokyo) 98,545-559 Tadano, K., and Ishizuka, I. (1982) J. Biol. Chem. 2 6 7 , 1482-1490 Tadano. K.. and Ishizuka. I. (1982) J. Biol. Chem. 267.9294-9299 Tadano; K.; Ishizuka, L,'Matsuo,'M., and Matsumoto, S. (1982) J. Biol. Chem. 257.13413-13420 , Kannagi,R., Levery, S. B., Ishigami, F., Hakomori, S., Shevinsky, L. H., Knowles, B., and Solter, D.(1983) J. Biol. Chem. 268,8934-8942 Sekine, M., Suzuki, M., Inagaki, F., Suzuki, A., and Yamakawa, T. (1987) J. Biochem. (Tokyo) 101,553-562 Tadano, K., and Ishizuka, I. (1979) Biochim. Biophys. Acta 676,421-430 Iida, N., Toida, T., Kushi, Y., Handa, S., Fredman, P., Svennerholm, L., and Ishizuka, I. (1989) J. Biol. Chem. 264,5974-5980 Svennerholm, L. (1957) J. Neurochem. 1,42-53 Svennerholm, L. (1957) Biochim. Biophys. Acta 24,604-611 Burunngraber, E.G., Tettamanti, G., and Berra, B. (1976) in Glycolipid Methodology (Wittig, L. A., ed) pp. 159-186, American Oil Chemists' Society, Champaign, Illinois Folch, J., Lees, M., and Sloane Stanley, G. H. (1957) J. Biol. Chem. 2 2 6 , 497-509 Ishizuka, I., Suzuki, M., and Yamakawa, T. (1973) J. Biochem. (Tokyo)7 3 , 77-87 Bax, A,, and Drobny, G. (1985) J. Magn. Reson. 61,306-320 Abe, K., and Tamai, Y. (1982) J. Chromatogr. 232,400-405 Tadano-Aritomi, K., and Iahizuka, I. (1983) J. Lipid Res. 24,1368-1375 Handa, S., Kushi, Y., Kambara, H., and Shizukuishi, K. (1983) J . Biochem. (Tokyo) 93,315-318 Kushi, Y., Handa, S., Kambara, H., and Shizukuishi, K. (1983) J. Biochem. (Tokyo) 94,184-1850 Kushi, Y., Handa, S., and Ishizuka, I. (1985) J. Biochem. (Tokyo) 97,419428 Hakomori, S. (1964) J. Biochem. (Tokyo) 66,205-208 Stellner, K., Levery, B., and Hakomori, S. (1987) Methods Enzymol. 1 3 8 , 13-25 Nakamura, K., Hashimoto, Y., Suzuki, M., Suzuki, A,, and Yamakawa, T. (1984) J. Biochem. (Tokyo)96,949-957 Ishizuka, I., and Tadano-Antoml, K. (1984) J. Biochem. (Tokyo) 96,829839 Stoffyn, P., and Stoffyn, A. (1963) Biochim. Bio hys Acta 70, 107-108 Chou, D. K. H., Ilyas, A. A,, Evans, J. E., CosteEo, C., Quarles, R. H., and Jungalwara, F. B. (1986) J. Biol. Chem. 2 6 1 , 11717-11725 Shevinsky, L. H., Knowles, B. B., Damjanov, I., and Solter, D. (1982) Cell 30,697-704 Kannagi, R. Cochran, N. A,, Ishigami, F., Hakomori, S. Andrews, P. W., Knowles, B. B., and Solter, D. (1983) EMBO J. 2,235b-2361 Magnani, J. L., Smith, D. F., and Ginsburg, V. (1980) Anal. Biochem. 1 0 9 , 399-402 Hawkes, R., Niday, E., and Gordon, J. (1982) Anal. Biochem. 1 1 9 , 142147 Higashi, H., Fukushi, Y., Ueda, S., Kato, S., Hirabayashi, Y., Matsumoto, M., and Naiki, M. (1984) J. Biochem. (Tokyo)9 6 , 1517-1520 Magnani, J. L., Nilsson, B., Brockhaus, M.,Zopf,D., Steplewski, Z., Koprowski, H., and Ginsburg, V. (1982) J. Biol. Chem. 2 6 7 , 1436514369 Dittmer, J. D., and Lester, R. L. (1964) J . Lipid Res. 6,126-128 Tadano, K., and Ishizuka, I. (1980) Biochem. Biophys. Res. Commun. 9 7 , 126-132 Yamakawa, T., Kiso, N., Handa, S., Makita, A., and Yokoyama, S. (1962) J. Biochem. (Tokyo) 62,226-227 Hakomori, S., Siddiqui, B., Li, Y-T., Li, S-C., and Hellerqvist, C. G. (1971) J. Bwl. Chem. 246,2271-2277 Gasa, S., Mitsuyama, T., and Makita, A. (1983) J. Lipid Res. 2 4 , 174-182 Herlant-Peers, M.C., Montreuil, J., andStrecker, G. (1981) Eur. J. Biochem. 117,291-300 Harris M. J., and Turvey, J. R. (1970) Carbohydr. Res. 15,57-63 Gasa, h., Nakamura, M., Makita, A., Ikura, M., and Hikichi, K. (1985) Carbohydr. Res. 137,244-252 Leffler, H., Hansson, G. C., and Stromberg, N. (1986) J. Biol. Chem. 2 6 1 , 1440-1444 Yamada, A., Dabrowski, J., Hanfland, P., and Egge, H. (1980) Biochim. Bwphys. Acta 618,473-479

39. 40. 41.

43. 44. 45. 46.

49. 50. 51. 52. 53. 54. 55. 56.

~~~~

57. 58. 59. 60.

I

I

I

.

~~~

Continued on next page.

16236

Gb5Cer V'-sulfate of Human Kidney

GM3 GM1 GDla GDlb GTla

sm4.

A

-X

'SA CY CSB

CCD3

.

.

. q

GbaCer Vi-sulfate 0: Human Kidney