specific hydrolysis of intact erythrocyte cell-surface glycosphingolipids by endoglycoceramidase :...

Eur. J. Biochem. 218, 637-643 (1993) 0 FEBS 1993

Specific hydrolysis of intact erythrocyte cell-surface glycosphingolipids by endoglycoceramidase Lack of modulation of erythrocyte glucose transporter by endogenous glycosphingolipids

Makoto ITO', Yuko IKEGAMI', Tadashi TA12 and Tatuya YAMAGATA' I Laboratory of Glycoconjugate Research, Mitsubishi Kasei Institute of Life Sciences, Tokyo, Japan ' Department of Tumor Immunology, Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan

(Received June 4/August 30, 1993) - EJB 93 081 9/4

This study represents the specific hydrolysis of cell-surface glycosphingolipids (GSLs) of intact cells by endoglycoceramidase (EGCase; EC.3.2.1.123) which cleaves the linkage between oligosaccharides and ceramides of various GSLs. After a 2-h incubation of horse intact erythrocytes with 20 mU EGCase I1 in the presence of activator at 37"C, 68% of the N-glycolylneuraminic-acid- containing ganglioside GM3(NeuGc) and 70% of 4-0-acetyl GM3(NeuGc) were found to be hydrolyzed without hemolysis, accompanied by a corresponding increase in ceramide but not sphingosine or N,N-dimethylsphingosine. No hydrolysis was observed for sphingomyelin, phosphatidylcholine, phosphatidylethanolamine, cholesterol or membrane proteins. The decrease in immunoreactivity with GMR8 antibody, specific to NeuGca2,3Gal- of GM3(NeuGc), corresponded to that of cell-surface GM3(NeuGc) by the enzyme, and almost no immunoreactivity was found when 70% of the GM3(NeuGc) was hydrolyzed. Besides the cell-surface GM3(NeuGc) of horse erythrocytes, Gg,Cer of guinea pig, GM3(NeuAc) and LcCer of human, and bovine and rabbit erythrocyte IV3Galu-nLc,Cer were found to be efficiently hydrolyzed by EGCase I1 even when present in intact cells, while human erythrocyte Gb,Cer is quite resistant to hydrolysis by the enzyme on the cell surface as well as in detergent micelles. Glucose incorporation via the glucose transporter in erythrocytes was not affected at all by the specific and exhaustive hydrolysis of cell-surface GSLs by EGCase 11. This result strongly suggested that glucose transporter function was not directly modulated by endogenous GSLs. In summary, this paper demonstrates that, together with the assistance of activator protein, EGCase I1 will become a powerful tool for selectively removing sugar chains from cell-surface GSLs without damaging other cell membrane components, and will be useful for describing the biological functions of endogenous GSLs.

Glycosphingolipids (GSLs) are recognized as characteris- tic constituents of the plasma membrane of vertebrates. Since GSL sugar chains are oriented towards the external environ- ment and exhibit structural heterogeneity, GSLs seem to be well suited for taking part in various biological events on the cell surface [l]. During the past decade, the importance of GSLs in cell activities has been increasingly recognized following the demonstration of profound changes in the quantity as well as quality of GSLs accompanying cellular differentiation, oncogenic transformation and the cell cycle [ 11. Furthermore, exogenously added GSLs, especially sialic-

Correspondence to M. Ito, Laboratory of Marine Biochemistry, Faculty of Agriculture, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka, Japan, 812

Abbreviations. GSLs, glycosphingolipids; EGCase, endoglycoceramidase; NaClP,, phosphate-buffered saline; NeuGc, N-glycolylneuraminic acid ; NeuAc, N-acetylneuraminic acid ; LcCer, Gal B1,4Glcpl ,l'Cer; Gg,Cer, GalNAc~l,4Gal~1,4Glc~l,l 'Cer; GaCer, Gal~l,3GalNAc~1,4Gal~l,4Glc~1 ,l'Cer; Gb,Cer, GalNAcDl,3Ga- la l ,4Gal~1,4Glc~1 ,l'Cer; IV3Gala-nLc,Cer, Galal,3Galp1,4Glc- NAc~l,3Gal~1,4GlcBl ,l'Cer; FITC, fluorescein isothiocyanate. Ab- breviations for gangliosides follow the nomenclature system of Svennerholm [33].

Enzyme. Endoglycoceramidase (EC 3.2.1.123).

acid-containing GSLs (gangliosides), have been found to evoke cell growth arrest [2-41, cell differentiation [3] and neurite extension [ 5 ] . Exogenously added gangliosides might interact with the growth factor receptors [2, 41 and certain protein kinases on the cell surface [6]. It is also possible that gangliosides bind to calmodulin and regulate the activity of calmodulin-dependent enzymes such as CAMP phosphodiest- erase [7, 81. However, the actual roles of endogenous GSLs in biological processes need to be further clarified. So far several different approaches have been taken to characterize the function of endogenous GSLs. One simple approach is to use a specific inhibitor of GSL synthesis. The ceramide ana- logue l -phenyl-2-decanoylamino-3-morpholino-l -propano1 is a potent inhibitor of the enzyme UDP-glucose :ceramide glucosyltransferase and is used to deplete GSLs from various mammalian cells [9]. The second method is to add ligands specific for GSLs, such as cholera toxin B subunits specific for GM1 gangliosides [lo, 111 and several antibodies specific to GSL species [12]. Another technique is to treat cells with glycosidases. Although sialidase has been used frequently for this purpose [13], sialidase is not specific for GSLs. It acts on both glycoproteins and GSLs. On the other hand, we have proposed a new approach to assess the biological signifi-

638

cance of endogenous GSLs using the GSL-specific enzyme endoglycoceramidase (EGCase) and its protein activator [14]. The possible effects of ceramides remaining on the cell surface should be kept in mind when GSLfunctions in situ are proposed by this method.

EGCase is the enzyme capable of cleaving the glycosidic linkage between oligosaccharides and ceramides of various GSLs, except cerebosides and sulfatides [15, 161. It consists of three different molecular species, each with its own specificities (EGCases I, I1 and 111) [17]. EGCases have not been used for research on living cells under physiological conditions since they require a detergent to express full activity, possibly due to their hydrophobic properties [17]. Recently, protein activators stimulating EGCase activity in the absence of detergents have been purified from the culture supernatant of Rhodococcus sp. [18] that had been isolated as an EGCase producer [ 15 - 171. Two molecular species of activator proteins were found [18], one of which stimulated specifically the activity of EGCase I (activator I) and the other that of EGCase I1 (activator 11). In this paper, we used EGCase I1 and activator I1 as the tool for cleaving sugar chains from cell-surface GSLs, since purified activator I is not yet avail- able.

The first purpose of this paper is to demonstrate that EGCase I1 can hydrolyze cell-surface GSLs without damaging other cell membrane components. It was reported that exogenously added GM3, but not a bovine brain ganglioside mixture, increased the activity of glucose transporter of human erythrocytes [19, 201. The second purpose is thus to examine whether cell-surface endogenous GSLs may directly affect glucose transporter function using EGCase I1 as the tool for removing cell-surface GSLs. The kinetics of EGCase I1 for cell-surface GSLs of erythrocytes is reported in the following paper in this journal [21].

MATERIALS AND METHODS

Materials

GM3(NeuGc),4-O-acetyl-GM3(NeuGc), and IV’Galu- nLc,Cer were kindly provided by Dr H. Higashi at our laboratory while Gg,Cer was applied by Dr Y. Hirabayashi (Riken, Japan). NeuAc-lactose, sialidase (neuraminidase type I1 from Vibrio cholera), sphingomyelinase (from Staphylo- coccus aureus), cellotriose, cytochalasine B, glucose oxidase and molecular standard proteins for SDSPAGE were obtained from Sigma; GM3(NeuAc) was from Bachem, Triton X-100 from Pierce, precoated silica gel 60 HPTLC and TLC plates from Merck, 2-deoxy[’H]glucose from Amersham, sphingosine and N,N-dimethylsphingosine from Matrea Inc., fresh difibrinated horse, guinea pig, rabbit and bovine blood from Nippon Bio-Test Laboratories (Japan), Hanks’ solution from Nissui Co. (Japan), goat F(AB’)2 fragment to mouse IgM, conjugated to fluorescein isothiocyanate (FITC), from Cappel.

EGCase and activator assay EGCase activity was assayed using purified Gg,Cer as

the substrate in the presence of Triton X-100 as described [17]. One milliunit (mu) of enzyme was defined as that amount capable of catalyzing the hydrolysis of 1 nmol substrate/min. In this paper, the amount of activator in the reaction mixture is indicated by the concentration of the purified activator protein.

Isolation of EGCase I1 and activator I1 from Rhodococcus sp. and preparation of 27.9-kDa activator

EGCase was isolated from the culture supernatant of a Rhodococcus sp. M-750 as described [ 171. Activator I1 specifically stimulating the activity of EGCase I1 was isolated from a Rhodococcus sp. M-777 as described [18]. The 27.9- kDa polypeptide possessing activity identical to that of native activator I1 (69.2 kDa) was prepared from activator I1 by trypsin treatment followed by trypsin inhibitor column appli- cation as described [22]. It is of interest that activator I1 was completely converted to 27.9-kDa polypeptides while full activity was retained after exhaustive digestion with trypsin. This is in contrast to ECCase, whose activity was totally abolished after trypsin treatment. Thus, using this method, activator without any EGCase contamination was obtained [22]. This 27.9-kDa polypeptide was used instead of the native form for all experiments in this study, and is referred to simply as ‘activator’.

Determination of GSL hydrolysis of erythrocytes

Determination of GSL hydrolysis of erythrocytes was conducted using by high-performance thin-layer chromatog- raphy (HPTLC) of GSLs remaining in cells after enzyme treatment. Following treatment of intact erythrocytes with enzyme, GSLs were extracted from cells by sonication for 30 min with 500 pl chloroform/propan-2-ol/water (7/11/2, by vol.). Cell extracts were centrifuged at 16000 rpm for 5 min and supernatant was withdrawn. Cells were extracted with the same solvent three times. Supernatants from each extraction were combined and dried under N, gas. GSLs were analyzed by HPTLC using developing solvent system I (chloroform/methanol/water, 65/25/4, by vol.) for Gg,Cer, system I1 (chlorofordmethanol/water, 60/40/10, by vol.) for Gb,Cer and IV’Galu-nLc,Cer, and system I11 (chlorofordmethanol/ 0.25% KCl, 5/4/1, by vol.) for acidic GSLs. GSLs were visualized by spraying the HPTLC plates either with orcinol/ H,SO, reagent for neutral GSLs or resorcinol/HCl reagent for acidic GSLs, and were determined by a Shimadzu CS- 9000 chromatoscanner with the reflectance mode set at 540 nm for orcinol/H,SO, staining and at 580 nm for resorcinol/HCl staining. The extent of hydrolysis of GSLs by the enzyme was calculated from the decrease in GSLs after treatment of erythrocytes with the enzyme.

Measurement of hemolysis

following paper [21]. This was determined by the method described in the

Determination of phospholipids, cholesterol, ceramides and sphingosines

Phospholipids and GSLs were extracted from erythrocytes using the same procedure for GSL extraction as described above. Cholesterol, ceramides and sphingosines were extracted with chlorofordmethanol (211, by vol.). They were separated on TLC aluminum plates coated with silica gel 60 using developing solvent system IV (chloroform/methanol/ 0.02% CaCl,, 8013515, by vol.) for phospholipids and GSLs, and system V (chloroform/methanol/28% NH,OH, 80/10/2, by vol.) for cholesterol, sphingosines, N,N-dimethylsphin- gosines and ceramides, respectively. These lipids and GSLs were visualized by dipping the TLC plate into a staining

639

solution consisting of 0.03% Coomassie brilliant blue R in 20% methanol containing 0.1 M NaCl for 30 rnin with gentle shaking [23]. The plate was removed from the staining solution, immersed in 20% methanol containing 0.1 M NaCl and allowed to stand for 3 min. The plate was air-dried and scanned with a Shimadzu CS-9000 chromatoscanner with the reflectance mode set at 580 nm.

SDSPAGE of membrane proteins

Packed erythrocytes (5 pl) were incubated with 10 mU EGCase I1 in the presence of 5 nmol activator in 50 p1 20 mM potassium phosphate pH 7.0 made isotonic with 0.85% NaCl (phosphate-buffered saline, NaCl/P,) at pH 7.0 for 5 h. For sialidase treatment, 20 mU Vibrio sialidase was used instead of EGCase I1 without activator. After incubation with enzymes, cells were collected by centrifugation (2000 rpmX5 rnin); 500 pl water was added to hemolyze the erythrocytes. Erythrocyte membranes were collected by ul- tracentrifugation (50000 rpmX30 min). They were washed with 500 p1 distilled water twice more and dissolved in 8 pl SDS/sample buffer and heated at 100°C for 6 min. SDS/ PAGE was conducted on a slab gel with a 10-20% gradient acrylamide according to the method of Laemmli [24]. The proteins were stained by Coomassie brilliant blue.

Staining of horse erythrocyte cell-surface GM3 with monoclonal antibody GMRS

Intact horse erythrocytes (20 pl packed cells) were incubated with 100 pl monoclonal antibody GMR8 [25] specific to GM3(NeuGc) at 5°C for 1 h. After being washed twice with 1 ml NaCUP,, the erythrocytes were treated with 100 pl second antibody (FITC-conjugated goat anti-mouse IgM) at 5°C for 30 min and then analyzed by flow cytometry.

Determination of glucose uptake by erythrocytes

Glucose uptake by erythrocytes was determined by two different methods. Method I measures glucose consumption in the reaction mixture and method I1 measures the incorporation of 2-deo~y[~H]glucose in cells.

Method I: intact erythrocytes (20 pl packed cells) were incubated with 20 mU EGCase I1 in the presence of activator (10 nmol) in 120 pl Hanks’ solution (glucose, 1 mg/ml, pH 7.0) at 37°C. After the incubation period indicated in Fig. 3 (A, B), 10 pl supernatant was withdrawn and glucose content was determined by the glucose oxidase method [26].

Method 11: intact erythrocytes (5 p1 packed cells) were incubated with 5 mU EGCase I1 in the presence of activator (2.5 nmol) or with 20 pM GM3 in 27 pl NaCIP, pH 7.0 at 37°C. After 2 h, 3 pl 2-deoxy[’H]glucose (0.75 pCi) was added to start the measurement of glucose uptake. For control experiments, EGCase 11, activator and GM3 were omit- ted from the incubation. After the incubation period indicated in Fig. 3 (C, D), the incorporation of 2-deoxy[’H]glucose was stopped by adding 500 pl terminating buffer (NaCl/P, containing 40 pM cytochalasin B). The red cells were washed three times with 500 p1 of the same buffer, dried on the filter paper and the radioactivity measured with a scintillation counter using a 2,5-diphenyloxazole/toluene (5 g/l) scintillation cocktail.

RESULTS

Effects of EGCase I1 treatment on cell-surface molecules

EGCase I1 hydrolyzes various GSLs (except cerebrosides and sulfatides) but does not degrade glycoproteins or sphingomyelin [17]. In this study, we examined the specificity of the enzyme towards these molecules present in intact cells. After a 2-h incubation of intact horse erythrocytes with 20mU EGCase I1 in the presence of activator, 68% of N- glycolylneuraminic-acid-containing GM3 (indicated as GM3 in this study) and 70% of 4-0-acetyl GM3 were found to be hydrolyzed (lane 4, Fig. 1A; values are means of three independent experiments). Virtually no hydrolysis was observed for sphingomyelin, phosphatidylcholine or phosphatidylethanolamine by EGCase I1 treatment (lane 4), i.e. 96.5 % sphingomyelin, 97.2% phosphatidylcholine and 99.7% phosphatidylethanolamine were found to remain after EG- Case I1 treatment (means of three independent experiments). Fig. 1A also shows the hydrolysis of sphingomyelin by a sphingomyelinase from Staphylococcus aureus (lane 2) and that of GM3 by a sialidase from Vibrio cholera (lane 3), which selectively hydrolyzed GM3 but not 4-0-acetyl-GM3 at all, and the quantitative production of LcCer after sialidase treatment (lane 3). Cholesterol and ceramides migrated at the front of the solvent in developing system IV and were thus analyzed using developing solvent V. Free ceramides were produced by the action of the enzyme (lane 3, Fig. 1B). The amount of ceramides produced was 6.1 nmol, accounting for about 65% of the erythrocyte GM3 used. This result indicates that the ceramides released from horse erythrocyte GM3 re- mained virtually in the intact form, since 68% of GM3 and 70% of 4-0-acetyl GM3 were found to be hydrolyzed (lane 4, Fig. 1 A). It should be noted that neither sphingosine nor N,N-dimethylsphingosine was produced after a 2-h (lane 3, Fig. lB), or even a 16-h incubation (data not shown). The same result was obtained when guinea pig erythrocytes were used instead of horse erythrocytes (data not shown). No cholesterol hydrolysis occurred (lane 3, Fig. 1 B). EGCase I1 also did not affect the SDSPAGE pattern of horse erythrocyte membrane proteins (lane 3, Fig. lC) , while sialidase did to some extent (lane 4). As shown in Fig. l C , neither EGCase I1 (lane 5), activator (lane 6) nor sialidase (lane 7) used in this study was incorporated into the membrane fractions (lanes 3 and 4). In summary, these results demonstrate that EGCase I1 can specifically remove GSL sugar chains from the cell surface of erythrocytes without affecting other membrane molecules in intact cells.

GMRS antibody reactivity in relation to the quantity of cell-surface GM3 of intact horse erythrocytes

Immunoreactivity of intact horse erythrocyte cell-surface GM3 with monoclonal antibody GMR8, specific to NeuGc- a2,3Gal- [25], was examined before and after treatment with EGCase 11. Before EGCase I1 treatment, intact horse erythrocytes were found to be strongly stained with this monoclonal antibody, as revealed by cytofluorometric analysis (Fig. 2 A, a). Immunoreactivity decreased over time of incubation with the enzyme, and after 15 min a significant change was observed (Fig. 2A, b-e). The decrease in immunoreactivity with GMR8 corresponded to the decrease in GM3 of cells by the enzyme as shown in Fig. 2B. When 70% of GM3 was hydrolyzed by the enzyme, virtually no immunoreactivity was found (Fig. 2A, e) and there were no further changes in

640

1 2 3 4 5 6 7

4-GM3 a m

SM

GM3 -

Me2-SP

SP - 20

- - _-- -

m m - - 1 2 3

Fig. 1. Effects of EGCase I1 treatment on erythrocyte cell-surface components. (A) TLC showing the contents of GSLs and phospholipids of horse erythrocytes after EGCase I1 treatment. Lane 1, control (no enzyme treatment); lane 2, sphingomyelinase treatment; lane 3, sialidase treatment; lane 4, EGCase I1 treatment. PE, phosphatidylethanolamine; PC, phosphatidylcholine ; SM, sphingomyelin; GM3, GM3(NeuGc) ; 4-GM3, 4-0-acetyl GM3(NeuGc); LC, lactosylceramide. Intact erythrocytes (10 p1 packed cells) were incubated either with EGCase I1 (20 mu) , sialidase (20 m u ) or sphingomyelinase (500 mu) in 100 pl NaCIP,, pH 7.0, for 2 h at 37°C; 10 nmol activator for EGCase I1 and 10 mM MgCI, for sphingomyelinase were included in the reaction mixtures. (B) TLC showing horse erythrocyte ceramide and cholesterol contents after EGCase 11 treatment. Lane 1, standard; CH, cholesterol; CER, ceramide; SP, sphingosine; Me2-SP, N,N- dimethylsphingosine. lane 2, control (no enzyme treatment); lane 3, EGCase treatment. (C) SDSPAGE showing the membrane proteins after EGCase I1 treatment. Lane 1, molecular standard proteins (94 kDa, phospholipase b ; 67 kDa, bovine serum albumin; 43 kDa, oval- bumin; 30 kDa, carbonic anhydrase; 20 kDa, soybean trypsin inhibitor; 14 kDa, P-lactoalbumin); lane 2, control (no enzyme); lane 3, EGCase treatment; lane 4, sialidase treatment; lane 5 , EGCase 11 (5 pg); lane 6, activator (2.8 pg); lane 7, sialidase (1.8 pg). See text for details.

1 2 3 4

the antibody reactivity of cell-surface GM3 with the enzyme until 16 h (data not shown).

Extent of hydrolysis of cell-surface GSLs of various erythrocytes by EGCase I1

Under physiological conditions, about 5% of GM3 was found to be resistant to EGCase I1 as well as Vibrio sialidase after exhaustive digestion. On the other hand, Triton X-100 hemolyzed the erythrocytes completely and all the GM3 was found to be hydrolyzed completely by both enzymes. These results indicate that some horse erythrocyte GM3 on the cell surface may be cryptic toward exogenous enzymes or present inside the cell. This phenomenon was also observed using EGCase II on various intact erythrocytes as shown in Table 1. After incubation with 20 m U EGCase I1 in the presence of activator (60 pM) at 37°C for 16 h, 94.7% of horse erythrocyte GM3 and 97.6% of horse erythrocyte 4-0-acetyl-GM3, 98.2% of guinea pig erythrocyte Gg,Cer, 79.8% and 74.4% of rabbit and bovine erythrocyte IV’Cala-nLc,Cer, and 91.1% of human erythrocyte LcCer and 83.4% of human erythrocyte GM3(NeuAc) were found to be hydrolyzed. In the presence of Triton X-100 instead of the activator, all these GSLs were found to be hydrolyzed completely by EG- Case I1 under these conditions except Gb,Cer (Table 1). Gb,Cer, which is the predominant human erythrocyte GSL, is quite resistant to hydrolysis by EGCase I1 in intact cells

Table 1. Hydrolysis extent of various erythrocyte cell-surface GSLs by EGCase 11. Intact erythrocytes (10 p1 packed cells for horse and guinea pig, 20 p1 for rabbit, bovine and human) were incubated with 20 mU EGCase I1 in the presence of 10 nmol activator in 100 p1 NaC1/Pi (pH 7.0) at 37°C for 16 h. See text for details.

Erythrocytes GSLs Hydrolysis ~~

with with activator Triton

x-100

% ~~

Horse GM3(NeuGc) 94.7 100 4-0-Ac-GM.,(NeuGc) 97.6 100

Guinea pig Gg,Cer 98.2 100 Rabbit lV’Gala-nLc,Cer 79.8 100 Bovine WGala-nLc,Cer 74.4 100 Human LcCer 91.1 100

GM,(NeuAc) 83.4 100 Gb,Cer 7.0 12.9

as well as in Triton X-100 micelles (Table 1). Erythrocytes were found to be fully active after exhaustive digestion with the enzyme based on glucose transporter function as described below.

641

3 QO

G I

7 c10

]b 2 00

Q I

r I

0- 30 60 90 120

Time (min)

J

Fluorescence Intensity

Fig. 2. Immunoreactivity of intact erythrocyte cell-surface GM3 before and after EGCase I1 treatment. (A) Cytofluorometric analysis of cell-surface GM3 before and after EGCase I1 treatment: (a) before enzyme treatment (positive control); (b-e) incubation with EGCase I1 for (b) 15 min, (c) 30 min, (d) 60 min and (e) 120 min; (9 negative control without first antibody. (B) Decrease in cell-surface GM3 by EGCase 11. Intact horse erythrocytes (20 p1 packed cells) were incubated with 20 mU EGCase I1 in the presence of 10 nmol activator in 100 jd NaCIP,, pH 7.0, at 37°C for the time indicated in the figure. Cell-staining with monoclonal antibody GMR8 is described in the text.

Effect of EGCase I1 on glucose transporter function It was reported that exogenously added GM3 but not a

bovine brain ganglioside mixture increased the activity of glucose transporter of human erythrocytes [19, 201. We thus examined whether cell-surface endogenous GSLs may directly affect glucose transport function using EGCase I1 as the tool for removing cell-surface GSLs as described above. Interestingly, the consumption of glucose from the incubation mixture by guinea pig and horse erythrocytes was not affected at all by EGCase treatment (Fig. 3 A, B). Cytochalasin B, a specific inhibitor of glucose transporter, at the concentration of 40 pM was found to greatly reduce the incorporation of glucose by erythrocytes regardless of the presence or absence of the EGCase I1 (Fig. 3 A, B), indicating that glucose incorporation of erythrocytes observed in this study was actually mediated by the glucose transporter. The hydrolysis of cell-surface GSLs of both erythrocytes by EGCase I1 in

this condition was simultaneously examined. It was found that after incubation with the enzyme for 1 h 45% of the guinea pig erythrocyte Gg,Cer and 52% of the horse erythrocyte GM3 were hydrolyzed, and the extent of hydrolysis increased over time of incubation and attained 92% and 93%, respectively, after 5 h without any damage to other cell membrane components. No differences in glucose consumption between enzyme-treated cells and control cells were observed even after a 16-h incubation when 98% of the guinea pig erythrocyte Gg,Cer and 97% of the horse erythrocyte GM3 were found to be hydrolyzed (data not shown). Glucose incorporation by erythrocytes was also examined by uptake of 2-deo~y[~H]glucose, an analog of glucose that is not me- tabolized and thus accumulates in cells. As shown in Fig. 3C and D, 2-deo~y[~H]glucose incorporation by both guinea pig and horse erythrocytes was not changed at all by cell-surface GSL hydrolysis by the enzyme. Exogenous GM3 added at

642

0.40 I

u 0.30' ' ' ' ' ' ' ' ' ' ' ' 0.25 II(.I.I.I.I. 0 1 2 3 4 5 6 0 1 2 3 4 5 6

Time (h)

5000 6ooo T----7

Time (min)

Time (h)

4000 (D)

0 30 60 90

Time (min)

Fig. 3. Glucose uptake by erythrocytes after EGCase I1 treatment. Glucose consumption from incubation mixture by guinea pig (A) and horse (B) erythrocytes; 2-Deoxy['H]glucose uptake by guinea pig (C) and horse (D) erythrocytes. All results are means of duplicate determinations. (0) Control; (0) EGCase I1 treatment; (0) control with cytochalasin B (40 pM); (W) EGCase 11 treatment with cytochalasin B (40 pM); (A) GM3 (40 pM); (A) GM3 (40 pM) with cytochalasin B (40 pM). See text for details.

the concentration of 20 pM had no effect on 2-de0xy[~H]glucose incorporation by horse erythrocytes (Fig. 3 C), but slightly stimulated that by guinea pig erythrocytes which had no endogenous GM3 (Fig. 3D). It should be noted that 2- deoxy['H]glucose incorporation was also specifically inhib- ited by the addition of cytochalasin B (Fig. 3C, D). In summary, it is strongly suggested that erythrocyte glucose transporter function was not directly modulated by endogenous GSLs including GM3.

DISCUSSION

The cellular uptake of glucose is mediated by an integral membrane glycoprotein, glucose transporter, that binds glucose and transfers it across the lipid bilayer [27]. In 1985, cDNA cloning of erythroid-type glucose transporter was achieved; at least six isoforms of glucose transporter have now been cloned [28]. The erythroid-type glucose transporter consists of three parts: a transmembrane domain which is made of 12 putative membrane-spanning domains, a cyto- plasmic domain, and an exoplasmic domain containing N- linked oligosaccharides which is an essential part for expres- sion of glucose transport activity [29]. The exoplasmic domain may interact with certain cellular components and may play an important role in tissue-specific regulation of glucose transporter function. It was reported that exogenously added GM3 but not a bovine brain ganglioside mixture increased

carboxyl methylation of the glucose transporter of human erythrocytes [19] and that the increase in glucose transporter methylation sites enhanced its activity [20]. In order to examine the possibility that endogenous GSLs can regulate glucose transporter function, we have treated the erythrocytes with EGCase I1 in the presence of activator and examined glucose incorporation via the glucose transporter as well as the extent of hydrolysis of endogenous GSLs by the enzyme. Interestingly, it was clearly shown in this study that glucose incorporation via the glucose transporter in erythrocytes was not affected at all by the specific and exhaustive hydrolysis of cell-surface GSLs including GM3 by EGCase I1 (Fig. 3). This result strongly suggested that erythrocyte glucose transporter function was not directly modulated by endogenous GSLs. This paper also suggests that the function (or beha- vior) of endogenous GSLs is somehow different from that of exogenously added GSLs.

Lampio et al. examined the action of galactose oxidase on both galactose and N-acetylgalactosamine residues of erythrocyte cell-surface GSLs [30]. They found that 60- 70% of human and porcine erythrocyte Gb,Cer were oxi- dized by the enzyme, while only 4-12% of guinea pig erythrocyte Gg,Cer was accessible. In contrast, almost all Gg,Cer but only a small amount of Gb,Cer on the cell surface was found to be hydrolyzed by EGCase I1 in this study. This discrepancy may come from differences in the specificities of the enzymes used. EGCase I1 hydrolyzes globo-series GSLs very slowly [17] and galactose oxidase acts on galactose residues (in the case of Gb,Cer) much faster than on N-acetylgalactosamine residues (in the case of Gg,Cer).

Endo-,!?-galactosidase has been successfully used for cell- surface modification [31]. The changes after endo-P-galactosidase treatment occurred mainly in lacto-N-glycosyl series GSLs with long carbohydrate chains and glycoproteins, while EGCases work specifically on GSL molecules, except cerebrosides and sulfatides.

Nores et al. [32] described the threshold density of cell- surface GM3(NeuAc) of B16 melanoma recognized by the monoclonal antibody M-2590 in an all-or-none fashion. We also studied the threshold density of horse erythrocyte GM3 using the monoclonal antibody GMR8, which is specific for NeuGcu2,3Gal- [25]. As shown in Fig. 2, almost no immunoreactivity was found when 70% of GM3 was hydrolyzed by EGCase 11, and the remaining GM3 was still susceptible to the enzyme.

We thank Drs Y. Hirabayashi and H. Higashi for providing GSL samples. This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (052741 06) from the Ministry of Education, Science and Culture of Japan.

REFERENCES 1. Hakomori, S. (1 981) Glycosphingolipids in cellular interaction,

differentiation, and oncogenesis, Annu. Rev. Biochern. 50, 733-764.

2. Bremer, E. G., Schlessinger, J. & Hakomori, S. (1986) Ganglio- side-mediated modulation of cell growth: specific effects of GM3 on tyrosine phosphorylation of the epidermal growth factor receptor, J. Bid. Chem. 261, 2434-2440.

3. Nojiri, H., Takaku, F., Terui, Y., Miura, Y. & Saito, M. (1986) Ganglioside GM3: an acidic membrane component that in- creases during macrophage-like cell differentiation can induce monocytic differentiation of human myeloid and monocytic leukemic cell lines HL60 U937, Pro. Nuti Acud. Sci. USA 83, 782-786.

643

4. Nojiri, H., Stroud, M. & Hakomori, S. (1991) A specific type of ganglioside as a modulator of insulin-dependent cell growth and insulin receptor tyrosine kinase activity, J. Biol. Chem. 266, 4531 -4537.

5. Tsuji, S., Arita, M. & Nagai, Y. (1983) GQlb, a bioactive ganglioside that exhibits novel nerve growth factor (NGF)- like activities in the two neuroblastoma cell lines, J. Biochem. (Tokyo) 94, 303-306.

6. Tsuji, S., Nakajima, J., Sasaki, T. & Nagai, Y. (1985) Bioactive gangliosides. IV. Ganglioside GQI b/Ca2 ' dependent protein kinase activity exists in the plasma membrane fraction of neuroblastoma cell line, GOTO, J. Biochem. (Tokyo) 97, 969- 972.

7. Higashi, H., Oomori, A. & Yamagata, T. (1992) Calmodulin, a ganglioside-binding protein: binding of gangliosides to calmodulin in the presence of calcium, J. Biol. Chem. 267, 9831 -9838.

8. Higashi, H. & Yamagata, T. (1992) Mechanism for ganglioside- mediated modulation of a calmodulin-dependent enzyme, J. Biol. Chem. 267, 9839-9843.

9. Radin, N. S. & Inokuchi, J. (1991) Use of PDMP for the study of glycosphingolipid functions, Trends Glycosci. Glycotech- no/. 3, 200-213.

10. Spiegel, S. & Fishman, P. H. (1985) Direct evidence that endogenous GM1 ganglioside can mediate thymocyte proliferation, Science 230, 1285-1287.

11. Spiegel, S. (1989) Inhibition of protein kinase C- dependent cellular proliferation by interaction of endogenous ganglioside GMI with the B subunit of cholera toxin, J. Biol. Chem. 264,

12. Dyer, C. A. & Benjamins, J. A. (1990) Glycolipids and transmembrane signaling: antibodies to galactocerebroside cause an influx of calcium in oligodendrocytes, J. Cell. Biol. 111, 625-633.

13. Wu, G. & Ledeen, R. W. (1991) Stimulation of neurite out- growth in neuroblastoma cells by neuraminidase: putative role of GM1 ganglioside in differentiation, J. Neurochem. 56, 95- 104.

14. Ito, M., Ikegami, Y., & Yamagata, T. (1991) Elucidation of biological functions of cell-surface glycosphingolipids in situ by endoglycoceramidase made possible using their activator proteins, Glycoconjugate J. 8, 157.

15. Ito, M. & Yamagata, T. (1986) A novel glycosphingolipid-de- grading enzyme cleaves the linkage between the oligosaccha- ride and ceramide of neutral and acidic glycosphingolipids, J. Biol. Chem. 261, 14278-14282.

16. Ito, M. & Yamagata, T. (1990) Endoglycoceramidase from Rho- dococcus sp. G-74-2, Methods Enzymol. 179, 488-496.

17. Ito, M. & Yamagata, T. (1989) Purification and characterization of gl ycosphingolipid-specific endoglycosiddse (endoglycoceramidase) from a mutant strain of Rhodococcus sp., J. Biol. Chem. 264, 9510-9519.

18. Ito, M., Ikegami, Y. & Yamagata, T. (1991) Activator proteins for gl ycosphingolipid hydrolysis by endogl ycoceramidases, J. Biol. Chem. 266, 7919-7926.

16512-16517.

19. Yoon, H.-J., Park, H.-J. & Jhon, G.-J. (1992) GM3 affects carboxyl methylation in the human erythrocyte membrane, Bio- chem. Int. 28, 393-399.

20. Jhon, G.-J. & Hah, J.-S. (1991) The increase of activity of erythrocyte glucose transporter by carboxyl methylation, Korean Biochem. J. 24, 189-192.

21. Ito, M., Ikegami, Y. & Yamagata, T. (1993) Kinetics of endoglycoceramidase action toward cell-surface glycosphingolipids of erythrocytes, Eur: J. Biochem. 218, 645-649.

22. Ito, M., Ikegami, Y., Omori, A. & Yamagata, T. (1991) Conver- sion of endoglycoceramidase-activator I1 by trypsin to the 27.9 kDa polypeptide possessing full activity, J. Biochem. (Tokyo) 110, 328-332.

23. Nakamura, K. & Handa, S. (1984) Coomassie brilliant blue staining of lipids on thin-layer plates, Anal. Biochem. 142, 406-41 0.

24. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T 4, Nature 227, 680-685.

25. Ozawa, H., Kawashima, I. & Tai, T. (1992) Generation of mu- nne monoclonal antibodies specific for N- glycolylneuraminic acid-containing gangliosides, Arch. Biochem. Biophys. 294,

26. Henry, J. B. (1974) in Clinical chemistry, Todd-Sanford clinical diagnosis laboratory methods (Davidson, I. H., ed.) 15th edn, pp. 601 -61 2, Saunders, Philadelphia, London.

27. Ohta, T. (1992) Molecular biology of mammalian glucose trans- porters, Trends Glycosci. Glycotechnol. 4 , 99- 105.

28. Mueckler, M., Baldwin, S. A,, Panico, M., Blench, I., Moms, H. R., Allard, W. J., Lienhard, G. E. & Lodsh, H. F. (1985) Sequence and structure of a human glucose transporter, Sci- ence 229, 941 -945.

29. Feugeas, J.-P., Neel, D., Goussault, Y. & Derappe, C. (1991) Glycosylation of the human erythrocyte glucose transporter: a minimum structure is required for glucose transport activity, Biochim. Biophys. Acta 1006, 59-62.

30. Lampio, A,, Finne, J., Homer, D. & Gahmberg, C. G. (1984) Exposure of the major human red-cell glycolipid, globoside, to galactose oxidase, Eur: J. Biochem. 145, 77-82.

31. Fukuda, M. N., Fukuda, M. & Hakomori, S. (1979) Cell surface modulation by endo-P-galactosidase, J. Biol. Chem. 254,

32. Nores, G. A,, Dohi, T., Taniguchi, M. & Hakomori, %-I. (1987) Density-dependent recognition of cell surface GM3 by a certain anti-melanoma antibody, and GM3 lactone as a possible immunogen: Requirements for tumor-associated antigen and immunogen, J. lmmunol. 139, 3171 -3176.

33. Svennerholm, L. (1 963) Chromatographic separation of human brain gangliosides, J . Neurochem. 10, 61 3 -623.

427 -433.

5458 -5465.

specific hydrolysis of intact erythrocyte cell-surface glycosphingolipids by endoglycoceramidase :...

Documents