Galactose to ceramide linkage is essential for the binding of a polyclonal antibody to galactosyl ceramide
S h am a Bha t
Department of Neurology, University of Pennsyh,ania Medical Center, Philadelphia, PA, USA
(Received 5 December 1991) (Revision received 12 May 1992)
(Accepted 12 May 1992)
Key words: Glycolipids; Galactosyl ceramide; Galactosyl sulfatide; HIV infection; Oligodendrocyte; Antibody; Human immunodeflciency virus
Summary
Characterization of a polyclonal antibody to galactosyl ceramide (Gal-Cer) which inhibits the internalization and infection of HIV-1 in neural cell lines was carried out. Polyclonal antibody to Gal-Cer was produced by injecting rabbits with Gal-Cer liposomes. The specificity of anti-Gal-Cer binding was studied by high performance thin layer chromatography (HPTLC)-based immunoassay. Using natural and semisynthetic lipids, the specificity of anti-Gal-Cer interaction was studied. The antibody bound to Gal-Cer and its derivatives. The antibody did not bind to glucosyl ceramide or lactosyl ceramide. Glucosyl ceramide differs from Gal-Cer by a hydroxyl group at the fourth carbon and in lactosyl ceramide galactose is linked to ceramide by an intervening glucose molecule. This indicates that D-galactose linked to ceramide is essential for binding. Removal of fatty acid from Gal-Cer, as seen with N-palmitoyl- and N-oleoyl Gal-Cer, had no effect on the binding. It appears that the third carbon of Gal-Cer is not involved in the binding. This is supported by the binding of anti-Gal-Cer to sulfatide or GM4 in which sulfate or sialic acid are added at the third carbon of GaI-Cer, respectively.
Introduction
Galactosyl ceramide (galactocerebroside; Gal-Cer) is a neutral glycosphingolipid located on neural cell membranes. It is a marker for differentiation of oligo- dendrocytes and Schwann cells and becomes highly enriched in myelin (Sternberger, 1982). In adult human brain myelin, Gal-Cer constitutes 22% of total lipids (Norton and Cammer, 1984). Antibodies to Gal-Cer have been used as a marker for oligodendrocytes and Schwann cells (Raft et al., 1978; Mirsky et al., 1980) and have been used as powerful tool in studies of myelination, demyelination and remyelination (Pfeiffer, 1984).
Recently, we reported that anti-Gal-Cer inhibits the internalization and infection of human immunodefi- ciency virus-1 (HIV-1) in the neural cell lines U373-MG
Correspondence to: S. Bhat, Department of Neurology, University of Pennsylvania Medical Center, 3400 Spruce Street, Philadelphia, PA 19104, USA.
and SK-N-MC (Harouse et al., 1991). We also found that gpl20, the envelope glycoprotein of HIV-I, binds specifically to Gal-Cer, sulfatide, and GM4 ganglioside, suggesting that these molecules may serve as alterna- tive receptors for HIV-1 in the brain (Bhat et al., 1991). Since antibodies to glycolipids will serve as an important tool in HIV studies, a detailed study was undertaken to characterize anti-Gal-Cer used in our earlier studies. The data indicate that anti-Gal-Cer binds, very specifically, to Gal-Cer or a molecule de- rived from Gal-Cer, such as sulfatide and the galactose to ceramide linkage is essential for the binding.
Materials and Methods
Antibody production 25 mg of bovine Gal-Cer (Type 1, Sigma Chemicals)
was dissolved in 25 ml of chloroform-methanol (1:1 v /v ) in a round-bottomed flask and evaporated under N 2 to make a thin film on the walls of the flask. The Gal-Cer used was found to be more than 95% pure as
106
judged by high performance thin layer chromatography (HPTLC) analysis. 6 ml of PBS containing 20 mg of BSA was added to the flask and stirred for about 16 h. The mixture was then homogenized in a tight-fitting glass homogenizer. The liposomes were stored at - 20°C. 0.25 ml of liposomes were mixed very well with 0.25 ml complete Freund's adjuvant and injected sub- cutaneously into the backs of rabbits. 1 and 2 weeks later, 0.25 ml of liposomes were mixed with 0.25 ml of incomplete adjuvant and injected subcutaneously. 4 weeks after the first injection, the rabbits were boosted with 0.5 ml of Gal-Cer liposome alone. Seven days after the booster, the rabbits were bled and serum was tested for anti-Gal-Cer activity.
lmmunostaining of chromatograms Immunostaining of thin-layer chromatograms was
performed as described, with minor modifications (Magnani et al., 1987). In brief, Gal-Cer (type I) and other iipids (Sigma Chemicals) were chromatographed on aluminum-backed silica gel HPTLC plates (E.M. Science) in ch loroform-methanol -H 20 (60:35:8) . The plates were coated with 0.1% poly(iso-butyl methacry- late) (Polysciences Inc.) in hexane for 90 s. The dried plates were incubated with 1% gelatin in 50 mM Tris. HCI, 0.15 M NaC1, pH 7.4 (GTS buffer) for 1 h, followed by anti-GalC in GTS buffer for 1 h at room temperature. The chromatograms were washed three times with GTS buffer containing 0.3 mM NaC1 for 10 min each, followed by incubation with [125I]protein A (2 × 10 ~' c m p /ml in GTS buffer) for 1 h. After wash- ing, as above, the chromatograms were exposed to X-ray film (XAR-5, Eastman Kodak). The autoradio- grams were either scanned by using a scanning densito- meter (Hoefer Scientific), or the chromatograms were exposed to iodine vapors and the lipid spots were scraped and counted in a gamma-spectrometer. The binding was calculated as:
%Binding =
(cpm of t e s t - b l a n k ) / ( c p m of control-blank) × 100
In most cases binding to Glc-Cer was used as blank.
Liposomes Liposomes were prepared by dissolving 5 mg of
lipids in chloroform-methanol (2: 1) and making a thin film on the walls of a test-tube by evaporating under nitrogen. The lipid film was suspended in PBS buffer containing 1 mg/ml of bovine serum albumin followed by sonication or homogenization.
lmmunofluorescence staining of rat oligodendrocytes Enriched oligodendrocytes were cultured on glass
cover slips as described (Bhat and Silberberg, 1985,
1986). 3-5 days after culturing, cells were double- stained with anti-Gal-Cer, polyclonal and monoclonal (Ranscht et al., 1982), as described (Bhat and Silber- berg, 1985, 1986). Normal rabbit serum was used as a control. Monoclonal anti-Gal-Cer was from Barbara Ranscht, La Jolla Cancer Research Foundation.
Purified glycolipids and lipid fractions Total lipids were extracted from human brain (auto-
psy specimens from the Hospital of the University of Pennsylvania). From the total lipids, alkali stable total, neutral and Folch lower phase lipids were extracted as described (Ledeen and Yu, 1982). These lipids were separated by HPTLC followed by immunostaining with anti-Gal-Cer. Various sugars, glycolipids and purified lipids were purchased from Sigma Chemicals.
Results
Antibodies to Gal-Cer bind to Gal-Cer and sulfatide HPTLC was used to study the binding of anti-Gal-
Cer to Gal-Cer and other lipids. The lipids were sepa- rated on silica gel G-60, followed by incubation of the antibody and radiolabelled protein A. The HPTLC autoradiograms were analysed for binding activity. As shown in Fig. 1, anti-GalC bound to Gal-Cer and sulfatide (sulfated Gal-Cer), but did not bind to gluco- syl ceramide (Glc-Cer), gangliosides GM1, GDI~, or neutral lipids from red blood cells. Purified IgG from anti-Gal-Cer serum also showed similar results (data not shown). Normal rabbit serum immunoglobulins did not bind to any of these lipids (Fig. 1C). Other antibod- ies such as anti-NCAM (Bhat and Silberberg, 1986), anti-BSA or AzBs, a monoclonal antibody to ganglio- side G O (Eisenberg et al., 1979) also did not bind to Gal-Cer or sulfatide (data not shown).
Titration of anti-Gal-Cer sera binding to Gal-Cer and sulfatide
Gal-Cer and sulfatides were separated on HPTLC and different dilutions of anti-Gal-Cer from different rabbits were incubated, followed by [125I]protein A. The Gal -Cer /su l fa t ide-ant ibody-pro te in A complex was quantitated. The minimum level of detection in this assay was 0.5/zg of Gal-Cer or sulfatide. As shown in Fig. 2, 4BGC-2 had the highest titer whereas 1GC-1 and 1GC-3 had lower titers. With 4BGC-2, in some experiments, sulfatide had lower activity (Fig. 2B) com- pared to Gal-Cer (Fig. 2A) up to 1:200 dilution, be- yond which there was no significant deference in bind- ing.
Immunofluorescence labelling of rat oligodendrocytes Polyclonal and monoclonal anti-Gal-Cer serve as
markers for oligodendrocytes (Sternberger, 1982). Since
107
~'i:iiiii!iii,!~ii!iii!~ii~,
1 2 3 4 5 - 6 1 2 3 4 5 6 1 2 3 4 5 6 Fig. 1. The binding of anti-Gal-Cer to Gal-Cer and other lipids. A. Lipids stained with orcinol. B. Anti-Gal-Cer. C. Normal rabbit serum. Lanes: 1, Gal-Cer (Type I); 2, sulfatide; 3, Glc-Cer; 4, GMI; 5, GDIa; 6, neutral lipids from erythrocytes, which include Glc-Cer, lactosyl ceramide, ceramide trihexoside, globoside, GM3 and GMI. Antibodies were used at 1 : 200 dilution. 5 /xg of each of the lipids were used for the binding
assay.
all the known anti-Gal-Cer stain cultured oligodendro- cytes, we used immunostaining of enriched oligoden- drocytes to characterize anti-Gal-Cer. Cultures of en- riched oligodendrocytes (Bhat and Silberberg, 1985,
Dilution of anti-Gal C Fig. 2. Titration of anti-OaI-Cer sera. A. Anti-Oal-Cer binding to Gal-Cer. B. Anti-GaI-Cer binding to sulfatide. Varying dilutions of anti-Gal-Cer from different rabbits were used for the binding assay, as described in Materials and Methods. 5 /~g of Gal-Cer and sul- fatide were used for the assay, e, 4BGC-2; A, 1GC-1; II, 1GC-3. Binding to Glc-Cer is used as non-specific binding. Data were from one of the three separate experiments. Each point represents aver-
age from duplicate samples, which varied from 1 to 15%.
1986) were double-stained with polyclonal and mono- clonal anti-Gal-Cer. As shown in Fig. 3, only oligoden- drocytes were labelled with anti-Gal-Cer.
Specificity of anti-Gal-Cer binding The specificity of the antibody was determined by
comparing its binding to glycolipids and by the ability of various related lipids and sugars to inhibit the bind- ing. In the first set of experiments, lipids were sepa- rated by HPTLC, followed by binding of anti-Gal-Cer. As shown in Table 1 and Fig. 1, anti-Gal-Cer bound to Gal-Cer and sulfatide. Anti-Gal-Cer did not bind to Glc-Cer, sphingomyelin, GM1, GDI ~ or lactosyl ce- ramide. Anti-Gal-Cer also bound to semisynthetic lipids in which the acyl group of Gal-Cer was modified. These include N-oleoyl Gal-Cer and N-palmitoyl Gal- Cer. Anti-Gal-Cer also bound to GM4 (data not shown).
In the second set of experiments, anti-Gal-Cer was absorbed by incubating with liposomes containing re- lated lipids. As shown in Table 2, liposomes containing Gal-Cer inhibited the binding compared to liposomes containing Glc-Cer. Since anti-Gal-Cer did not bind to Gic-Cer (Fig. 1), Glc-Cer-containing liposomes were used as control. Incubation of antibodies with sugars, which are part of the Gal-Cer or glucosyl ceramide molecule, did not affect the binding of anti-Gal-Cer to Gal-Cer or sulfatide.
Anti-Gal-Cer binding to lipids from human brain Folch lower phase lipids extracted from human brain
were separated by HPTLC. Then lipids were analysed for binding of anti-Gal-Cer. As shown in Fig. 4, in the human brain, anti-Gal-Cer bound to three lipids. Pre- liminary analyses have shown that the three lipids are Gal-Cer, sulfatide and GM4.
D i s c u s s i o n
Polyclonal anti-Gal-Cer was produced in rabbits us- ing bovine Gai-Cer. The antibody was characterized by its ability to bind to oligodendrocytes, Gal-Cer and by
108
Fig. 3. lmmunofluorescence staining of rat oligodendrocytes by anti- GalC. Cultures of enriched oligodendrocytes were double-stained with polyclonal anti-GaI-Cer (4BGC-2; 1:50 dilution) and mono- clonal anti-Gal-Cer (culture supernatant, 1:10 dilution). Rhod- amine-conjugated goat anti-rabbit IgG (1 : 100 dilution) and fluores- cein-conjugated anti-mouse lgG (1:50 dilution) were used as second antibodies. A. 4BGC-2. B. Monoclonal anti-Gal-Cer. C. Phase con- trast. Please note only 3 out of 12 cells were stained with anti-Gal-Cer.
Rest of the unstained cells are presumably astrocytes.
a b s o r p t i o n wi th G a l - C e r l i posomes . T h e speci f ic i ty was
e l u c i d a t e d by us ing a va r i e ty o f s e m i - s y n t h e t i c a n d
na tu r a l l ipids, wh ich s h o w e d tha t D-ga lac tose l i nkage to
c e r a m i d e is e ssen t ia l .
A n t i - G a l - C e r b o u n d speci f ica l ly to G a I - C e r o r its
de r iva t ives such as su l fa t ide o r GM4. Al l t h r e e l ip ids
have D-ga lac tose l i nked to c e r a m i d e ( T a b l e 1). A n t i -
G a l - C e r d id no t b ind to G l c - C e r , GM1, o r GDla , in
wh ich t h e g lucose is l i nked to c e r a m i d e . A d d i t i o n o f
GDla Galfl 1-3GalNAcfl l-4Galfl l-4Glc-Cer 20-+ 11 I
NeuNAc Neul NAc Sphingo- 5 + 3
myelin c Psychosine d + + G i n 4 NeuAca2-3GaI-Cer + +
Anti-Gal-Cer (4BGC-2) was used at 1:200 dilution. a Fatty acid in Gal-Cer is palmitic acid. b Fatty acid in Gal-Cer is oleic acid. c Galactose of GaI-Cer replaced by phosphocholine. d Gal-Cer without fatty acid. + +, Anti-Gal-Cer bound to psychosine and GM4 but was not quantitated. 5 tzg of lipids were used for the binding assay, as described in Materials and Methods. Values are means -+SD from three experiments.
su l fa te , as s e e n wi th su l fa t ide , o r s ial ic acid, as s e e n
wi th GM4 , d id no t a f fec t t h e b i n d i n g s ignif icant ly . R e -
m o v a l o f g a l a c t o s e as s e e n wi th s p h i n g o m y e l i n a f f e c t e d
t h e b ind ing , sugges t i ng tha t D-ga lac tose is e s sen t i a l for
t h e b ind ing , o - / 3 - G a l a c t o s e d id no t inh ib i t t he b ind ing ,
sugges t i ng tha t g a l a c t o s e a l o n e is no t su f f i c i en t for t he
b ind ing . A n t i - G a l - C e r d id no t b i n d to lac tosyl ce r -
a m i d e . In t he lactosyl c e r a m i d e m o l e c u l e , g a l a c t o s e is
l i nked to c e r a m i d e wi th an i n t e r v e n i n g g lucose
TABLE 2
Inhibition of anti-GaI-Cer binding to Gal-Cer by liposomes and sugars
Effectors Concentration % Binding
Glc-Cer liposomes a 500/~g 100 Gal-Cer liposomes a 500/zg 13_+ 3
Methyl /3-D- galactopyranoside b 500 mM 105 _+ 15
Methyl /3-D- glucopyranoside b 500 mM 107 _4- 11
a Anti-Gal-Cer (1 : 200 dilution) was incubated with Gal-Cer or Glc- Cer liposomes for 16 h at 23°C, followed by centrifugation. The supernatant was assayed for binding.
b Sugars were added with anti-Gal-Cer during the binding assay. For experimental details see Materials and Methods. Values are means _+ SD from three experiments.
A B
I
Fig. 4. Binding of anti-Gal-Cer to neutral lipids from human brain. Neutral lipids extracted from human brain were separated on silica gel G-60, followed by the binding assay as described in Materials and Methods. A. Anti-Gal-Cer (4BGC-2). B. Anti-BSA. 1, Gal-Cer (up- per band) and sulfatide (lower band); 2, Glc-Cer; 3, neutral lipids from the brain. For the relative positions of the lipids, please see Fig. 1A. 5/zg of Gal-Cer, sulfatide and Glc-Cer were used for the assay.
Antibodies were used at 1 : 200 dilution.
molecule (Table 1). This suggests that, for binding of anti-Gal-Cer, galactose must be attached to ceramide. Experiments with semisynthetic cerebrosides, in which the acyl groups are modified with different fatty acids, show that the acyl groups are not the contributing factor in anti-Gal-Cer binding.
Removal of anti-BSA from anti-Gal-Cer serum (since anti-Gal-Cer was raised by using Gal-Cer lipo- somes with BSA) by using a BSA-column did not decrease the binding, showing that anti-BSA in anti- Gal-Cer did not bind to Gal-Cer (data not shown). This is also supported by the fact that anti-BSA did not bind to Gal-Cer or sulfatide (Fig. 4).
Benjamins et al. (1987) have raised polyclonal anti- body to Gal-Cer by including keyhole limpet hemo- cyanin (KLH). Their antibody reacted with Gal-Cer and psychosine but did not react with Glc-Cer, sul- fatide and gangliosides. Moreover, their antibody was inhibited with higher concentration of galactose. This suggests that different polyclonal anti-Gal-Cers may have different specificities. Bansal et al. (1989) also characterized the lipid binding specificity of three monoclonal antibodies, 01, 04 (Sommer and Schachner, 1982) and R-mAb (Ranscht et al., 1982), which react with Gal-Cer. These monoclonal antibodies were raised against brain membranes. They found that 01 reacted with Gal-Cer, monogalactosyl-diglyceride and psycho- sine, whereas R-mAb reacted with Gal-Cer, mono- galactosyl-diglyceride, sulfatide, seminolipid and psy- chosine. 04 reacted with sulfatide, seminolipid, and cholesterol. Goujet-Zalc et al. (1986) reported the characterization of a monoclonal antibody which reacts
109
with sulfatide and seminolipid. The availability of anti- bodies with different specificities will help in under- standing more about the receptors for HIV in brain.
Our recent findings showed that anti-Gal-Cer in- hibits the uptake of HIV in neural cell lines, U373-MG and SK-N-MC (Harouse et al., 1991). Our studies also indicated that gpl20, the envelope glycoprotein of HIV-1, binds to GalC, suggesting that Gal-Cer may serve as an alternate receptor for HIV in the brain (Bhat et al., 1991). The specificity of gpl20 binding was similar to the binding of anti-GalC to GalC. GalC and sulfatides are specific lipids of myelin and myelin- forming cells: oligodendrocytes and Schwann cells (Norton and Cammer, 1984; Sternberger, 1982). GM4 is also found in human brain white matter and myelin (Ledeen and Yu, 1982). These studies indicate that anti-Gal-Cer may serve as another tool in the study of acquired immunodeficiency syndrome (AIDS).
Acknowledgements
I thank Laura Lynch for technical assistance, Dr. Steve Spitalnik for helpful discussions during the course of this project, Dr. B. Ranscht for monoclonal anti- Gal-Cer, Dr. Robert Yu for GM4. I also thank Drs. Donald Silberberg and David Pleasure for their helpful comments on the manuscript.
References
Bansal, R., Warrington, A.E., Gard, A.L., Ranscht, B. and Pfeiffer, S.E. (1989) Multiple and novel specificities of monoclonal anti- bodies 01, 04, and R-mAb used in the analysis of oligodendrocyte development. J. Neurosci. Res. 24, 548-557.
Benjamins, J.A., Callahan, R.E., Montgomery, I.N., Studzinski, D.M. and Dyer, C.A. (1987) Production and characterization of high titer antibodies to galactocerebroside. J. Neuroimmunol. 14, 325- 338.
Bhat, S. and Silberberg, D.H. (1985) Rat oligodendrocytes have cell adhesion molecules. Dev. Brain Res. 19, 139-t45.
Bhat, S. and Silberberg, D.H. (1986) Oligodendrocyte cell adhesion molecules are related to neural cell adhesion molecule (N-CAM). J. Neurosci. 6, 3348-3354.
Bhat, S., Spitalnik, S.L., Gonzalez-Scarano, F. and Silberberg, D.H. (1991) Galactosyl ceramide or a molecule derived from it is an essential component of the neural receptor for HIV-1 envelope glycoprotein gpl20. Proc. Natl. Acad. Sci. USA 88, 7131-7134.
Eisenbarth, G., Walsh, F. and Nirenberg, M. (1979) Monoclonal antibody to a plasma membrane antigen of neurons. Proc. Natl. Acad. Sci. USA 76, 4913-4917.
Fredman, P., Mattsson, L., Andersson, K., Davidsson, P., Ishzuka, I., Jeansson, S., Mansson, J.-E. and Svennerholm, L. (1988) Charac- terization of the binding epitope of a monoclonal antibody to sulfatide. Biochem. J. 251, 17-22.
Goujet-Zalc, C., Guerci, A. and Zalc, B. (1986) Schwann cell marker defined by a monoclonal antibody (224-58) with species cross-re- activity. II. Molecular characterization of the epitope. J. Neu- rochem. 46, 435-439.
110
Harouse, J., Bhat, S., Laughlin, M., Stefano, K., Spitalnik, S.L., Silberberg, D.H. and Gonzalez-Scarano, F. (1991) Inhibition of HIV-I by antibodies to galactosyl ceramide in neural cell lines. Science 253, 320-323.
Ledeen, R.W. and Yu, R.K. (1982) Gangliosides: Structure, isolation and analysis. Methods Enzymol. 83, 139-191.
Magnani, J.L., Spitalnik, S.L. and Ginsburg, V. (1987) Antibodies against cell surface carbohydrates: Determination of antigen structure. Methods Enzymol. 138, 195-207.
Mirsk'y, R., Winter, J., Abney, E.R., Pruss, R.M. Gavrilovic, J. and Raft, M.C. (1980) Myelin-specific proteins and galactolipids in rat Schwann cells and oligodendrocytes in culture. J. Cell Biol. 84, 483-494.
Norton, W.T. and Cammer, W. (1984) Isolation and characterization of myelin. In: P. Morell (Ed.), Myelin, Plenum Press, New York, NY, pp. 147-195.
Pfeiffer, S.E. (1984) Oligodendrocyte development in culture sys-
terns. In: W.T. Norton, (Ed.), Oligodendroglia, Plenum Press, New York, NY, lap 233-298.
Raft, M.C., Mirsky, R., Feilds, K.L., Lisak, R.P., Dorfman, S.H., Silberberg, D.H., Gregson, N.A., Liebovitz, S. and Kennedy, M.C. (1978) Galactocerebroside is a specific cell surface antigenic marker for oligodendrocytes in culture. Nature 274, 813-816.
Ranscht, B., Claspshow, P.A., Price, J., Noble, M. and Seifert, W. (1982) Development of oligodendrocytes and Schwann cells stud- ied with a monoclonal antibody against galactocerebroside. Proc. Natl. Acad. Sci. USA 79, 2709-2713.
Sommer, I. and Schachner, M. (1982) Monoclonal antibodies (01 to 04) to oligodendrocyte cell surfaces. An immunocytological study in the central nervous system. Dev. Biol. 83, 311-327.
Sternberger, N.H. (1982) Pattern of oligodendrocyte function seen by immunocytochemistry. In: W.T. Norton (Ed.), Oligodendroglia, Plenum Press, New York, NY, pp. 125-173.
