o-linked glcnac in serotype-2 adenovirus fibre

Eur. J. Biochem. 184,205-211 (1989) FEBS 1989

@Linked GlcNAc in serotype-2 adenovirus fibre Marie-Laure CAILLET-BOUDIN ', Gerard STRECKER' and Jean-Claude MICHALSKI *

' Laboratoire de Chimie Biologique et Unite de Recherche Associte du Centre National de la Recherche Scientifique no. 217, Unite de Virologie Moltculaire de I'lnstitut National de la Sante et de la Recherche Medicale (Unite 233), Lille

Universite des Sciences et Tcchniques de Lille Flandres-Artois

(Received March 29/May 26, 1989) - EJB 89 0380

Serotype-2 adenovirus fibre is shown to possess an 0-linked GlcNAc residue and to have affinity for wheat germ agglutinin. The cytoplasmic and nuclear fibres are both glycosylated. Glycosylation seems to take place in the cytoplasm since most of the [ ''C]ClcN-labelled fibre is found in this compartment, little label being associated with the microsomes. Glycosylation of the fibre was not affected by inhibitors of N- and 0-glycosylation.

A variation in fibre glycosylation is observed among adenovirus. Among the serotypes tested, only serotype- 5 adenovirus (another subgroup C virus) also incorporated [14C]GlcN into its fibre, but did not possess affinity for wheat-germ agglutinin.

The GlcNAc is located in the N-terminal two-thirds of the fibre and more probably in the N-terminal one- third. The free or penton-base-associated fibres are similarly glycosylated. These results suggest that glycosylation is not involved in viral adsorption and in assembly with the capsid penton base. Thus, glycosylation might be a characteristic feature of subgroup C viruses.

Adenovirus is a well-characterized double-stranded DNA virus encapsulated in a icosahedral capsid. The latter is composed of at least nine structural proteins. Ishibashi and Maize1 [l] have shown that one adenovirus type-2 early protein and one late protein were glycosylated, i.e. proteins synthesized respectively before or after DNA viral replication. The glycosylated early-protein gene is mapped in the early region E3 [2]. This protein has an apparent molecular mass of 19 kDa. It is synthesized and glycosylated in the endoplasmic reticulum, processed from a precursor and kept in the endoplasmic reticulum [ 3 - 51. Its glycan is linked to the protein by a classical GlcNAc-Asn linkage [6].

The glycosylated late protein of the adenovirus was identified as being the fibre [l]. This protein is anchored in the penton base at the capsid vertices by its shaft part, the knob being needed for viral adsorption onto the cell receptors (reviewed in [7]). ['4C]GlcNH, was incorporated into the fibre polypeptide. The sugar-protein linkage is alkali sensitive [l].

GlcNAc residues occur both in N-glycosidic-type and 0- glycosidic-type glycoproteins. In the N-glycans, GlcNAc is involved in the N-glycosidic bond with the amide nitrogen of asparagine. In 0-linked glycans, GlcNAc residues most commonly occur in oligosaccharides linked to polypeptide by a GalNAc residue (reviewed in [8, 91). Recently, evidence for an 0-linked GlcNAc was reported [lo, 1 I]. Since then several proteins with such an 0-linked GlcNac were described, such as the proteins of the nuclear pore complex [12- 141, a human erythrocyte cytoplasmic protein [15], nuclear proteins [16, 171 and the cytomegalovirus 149-kDa protein [ 181.

In this paper, we studied adenovirus type-2 fibre and identified 0-linked GlcNAc residues. Glycosylation probably takes

Corrrspondence to M. L. Caillet-Boudin, Unite de Virologie Moleculaire de I'INSERM [U. 2331, 2 place de Verdun, F-59045 Lille Ccdex, France

Ahhreviations. WGA, Wheat-germ agglutinin; Con A, concanavalin A; GalNAc, N-acetyl-P-D-galactosamine.

place in the cytoplasm and was not affected by either N- or 0-glycosylation inhibitors. Fibre glycosylation was shown to vary among different adenovirus serotypes and GlcNAc might be a characteristic feature of subgroup-C adenoviruses. The role of this 0-linked GlcNAc is discussed.

MATERIALS AND METHODS

Virus and cells To purify viruses, human adenovirus types 2 , 3, 4, 5 , 7 , 9

and wild type were grown on KB cells maintained in suspen- sion culture at 4 x l o 5 cells/ml in Eagle's minimum essential medium supplemented with 5% horse serum. Cells were infected at a multiplicity of infection of 10 - 20 plaque-forming units/cell.

To test glycosylation inhibitors and to prepare microsomes, adenovirus type-2 infection was performed on HeLa cell monolayers in Dulbecco's modified Eagle's medium.

PuriJicution of adenovirus type-2jibre and penfon

Soluble fibre and penton were purified from adenovirus- type-2-infected KB cells by a four-step procedure, namely fluorocarbon extraction, ammonium sulfate precipitation, DEAE-Sephadex and hydroxyapatite chromatography [19].

Labelling conditions

To label glycoproteins, D-U-['~C]G~CN (300 Ci/mol, Com- missariat a 1'Energie Atomique, Saclay, France) was added in a normal culture medium while ~ - [ 2 - ~ H ] M a n (16 Ci/mmol, New England Nuclear, USA) was added after 20-fold concentration of the cells in glucose-depleted medium.

To label proteins, met (600 - 700 Ci/mmol, Amer- sham, UK), ~- [3 ,4-~H~]Val (35 Ci/mmol, France) or [I4C]-

206

formic acid (sodium salt; 60 Ci/mol, Commissariat a 1'Energie Atomique, Saclay, France) were added in appropriate amino- acid-depleted medium.

To purify the fibre from labelled infected cells, radioactivity was added at 1 pCi for labelled amino acids and at 0.125 pCi/ml for ['4C]GlcN. To obtain a highly labelled cell extract, radioactivity was used at 15 pCi/mmol for a shorter time (1 h or 6 h).

Analysis of sugar composition

The fiber was washed by three precipitation rounds in 50% ethanol/water (by vol.), was then lyophilized and submitted to acidic methanolysis (methanol/0.5 M HCl, 80"C, 24 h). Monosaccharides were analyzed by GLC as trifluoroacetate derivatives according to Zanetta et al. 1201.

Acid hydrolysis of [ ''C]GlcN-labelled fiber was performed in 4 M trifluoroacetic acid at 100°C for 24 h. The hydrolysate was dried by evaporation under reduced pressure, resuspended in water and analyzed by thin-layer chromatography (Silica gel 60, Merck, Darmstadt, FRG) with n- butanol/acetic acid/water (2: 1 : 1 ; by vol.) as a solvent. Galactosamine and glucosamine (chloride salts, Sigma) were used as standards. For determination of radioactivity, the plate was scraped every 0.5 cm, the silica was resuspended in water (500 pl) and 4.5 ml scintillation cocktail was added.

Release qf oligosaccharides from the j h r hy reductive ulkaline treatment or hydrazinolysis

[14C]GlcN-radiolabelled fiber was treated with 2 mlO.1 M NaOH, 1 M KBH, at 37°C for 24 h. The reaction was stopped by adding Dowex 50x8 (20 - 50 mesh, H + form). Borate was eliminated as methylborate by repeated evaporation under reduced pressure with methanol.

The reaction mixture was chromatographed on a Bio-Gel P-4 column (200-400 mesh; 1 x 150 cm), in 50 mM acetic acid. Radioactive material was pooled and further analyzed by descending paper chromatography on Whatman 3 paper with rz-butanollacetic acid/water (4: 1 : 5 ; by vol.) as solvent. Glucosaminitol was identified after de-N-acetylation with 6 M HCI (IOOT, 4 h) by ion-exchange chromatography on a column of AG 50 WX8 (Hi form; Bio-Rad; 0.5 x 20 cm), equilibrated and eluted with 0.3 M HCI 1211.

Hydrazinolysis was according to [22].

Electrophoretic techniques

Analytical SDSIPAGE. Polypeptide analysis was performed in a SDS-containing 15% polyacrylamide slab gel (acrylamide/bisacrylamide, 50 : 0.235) overlaid by a 5% spacer gel (acrylamide/bisacrylamide, 50: 1.33) in the discontinuous buffer system described by Laemmli [23]. Samples were de- natured by heating for 2 min at 100 "C in 0.0625 M Tris/HCl, pH 6.8, 4% SDS, 6 M urea, 10% 2-mercaptoethanol, 0.002% bromophenol blue.

A,lffi'nity electrophoresis. To test adenovirus type-2 native fibre with different lectins, immunoelectrophoresis-like affinity electrophoresis was used (as described by Brig-Hansen et al. [24]). Lectins (concanavalin A, wheat-germ agglutinin, Lens culinzaris agglutinin up to 1 mg/ml) were incorporated into the first dimension agarose gel. Anti-(adenovirus type-2 fibre) serum (15 pl/ml) was included in the second dimension agarose gel.

Immuno-blotting. This was achieved according to Towbin et al. [25]. Western blots were incubated successively with wheat-germ agglutinin (WGA; 30 pg/ml) and avidin-peroxidase (25 pg/ml). In the HCI-hydrolyzed-fibre study, biotinylated WGA (Vecor Laboratories, Burlingham, USA; 10 p/ml) was used to intensify the reaction and then streptavidin-peroxidase was added.

Chemical hydrolysis of the f ibre

This was as described by Caillet-Boudin et al. [26].

Cel1,fractionation

Rough microsomes were isolated using a modification of the procedure described in 1271. During purification, nuclei and cytoplasm aliquots were stored to test fibre glycosylation in these two cellular compartments. The post-mitochondria1 supernatant was adjusted to 1.8 M sucrose by addition of 2.5 M sucrose in SO mM Tris/HCl, pH 7.4, 50 mM KC1, 5 mM MgClz and laid on 0.5 ml 2 M sucrose. Then, 1 ml 1.55 M sucrose, 7 ml 1.35 M sucrose and 0.5 nil 1 M sucrose were successively laid on. Centrifugation was performed for 15 h at 41 000 rpm in a tft 65 rotor.

Ainino mid analysis

Samples of adenovirus fibre corresponding to 0.225 mg protein were hydrolyzed for 24 h with 4 M HCl in sealed tubes under nitrogen and analyzed in a Beckman multichrom amino acid autoanalyser.

RESULTS

Giycm characterization

Purified fibre was subjected to hydrolysis in 4 M HC1 for 24 h at 110°C and the hydrolysate was analyzed on an amino acid analyzer. Glucosamine was detected at approximately one residue/polypeptide. Galactosamine was absent. We then made an attempt to label the fibre with ['4C]GlcN and [3H]Man as described in Material and Methods. Label was found only when ['4C]GlcN was used.

[14C]GlcN was incorporated without metabolic conver- sion because only GlcN was identified by GLC analysis of monosaccharide released after acid methanolysis and by thin- layer chromatography analysis of the monosaccharides released after acid hydrolysis of the ['4C]GlcN-labelled fibre.

The glycan-protein linkage was alkali labile. Identification of the group liberated by mild alkaline hydrolysis indicated that more than 90% of the radioactivity was present in a fraction comigrating with an acetylglucosaminitol standard. Ion-exchange chromatography after strong acid hydrolysis confirmed the occurrence of glucosaminitol (Fig. 1). N o trace of N-acetylgalactosaminitol was detected after alkaline elimination. This excluded the most common 0-glycosylation species : the 0-GalNAc linkage. Experimental conditions used for the p-elimination reaction excluded the possibility of GlcNAc linked to amino acids such as OH-Lys or OH-Pro, these types of linkages not being hydrolyzed in the alkaline conditions used in this study.

No GlcNAc was found after the hydrazine treatment. After re-N-acetylation and chromatography on Bio-Gel P-4 all radioactivity was co-eluted as micromolecular material (co-eluting with 3 H 2 0 ) . We assume that, like most 0-linked

207

A Table 1 Glycosylation in the presence of glycosylation inhibitors Thc ['4C]GlcN and [35S]Met labelling were performed separately

GlcNAc-ol

0

Distance, c m

GIcNHz-oI GalNH2-ol r I i '

f r t 10 20 LO 60

rnl

Fig. 1. Liberation of N-acetylfilucosaminitol by alkulinelreductive tr-mtment. (A) Paper chromatography. (B) The material migrating like the N-acetyl-glucosaminitol (GlcNAC-ol) by paper chromatography was cluted, hydrolyzed with 6 M HCI for 4 h at 100'C and applied on an AG50 WX8 column in 0.3 hl HCI. Glucosaminitol (GlcNH2-ol) and galiictosaminitol (GalNH2-oI) wcrc added as internal standards

monosaccharides, GlcNAc was released and largely degraded during hydrazinolysis as suggested by Takasaki et al. [28]. This result together with the observed release of N-acetylglucosaminitol after the /?-elimination reaction, confirms the presence of an 0-linked GlcNAc in the adenovirus type-2 fibre.

The adenovirus tpe-2 fibre shows affinity for WGA, similar to the nuclear-pore protein [12, 141. WGA was able to slow down the fibre electrophoretic migration when incorporated in the agarose gel while neither concanavalin A (Con A) nor Lens culinaris agglutinin affected it. The WGA affinity was also confirmed by Western blotting. GlcNAc (0.1 M) in the incubation buffer inhibited the affinity for by WGA.

Effect of N- and 0-glycosylujion inhibitors Since the 0-GlcNAc linkage was unusual and only re-

cently described, specific glycosylation inhibitors were tested. Tunicamycin and 2-deoxy-D-glucose, which are both inhibitors of N-linked glycosylations did not affect ['4C]GlcN incorporation at the usual concentrations (0.5 pg/ml and 10 mM, respectively). At higher concentrations, [14C]GlcN incorporation decreased in the same way as [35S]Met incorporation. Under these conditions, the synthesis of all proteins was decreased as seen in SDS/polyacrylamide gels (not shown) but the ratio of ["SlMet incorporated to incorporated [14C]GlcN was constant (Table 1).

Monensin, which inhibits protein transfer from the endoplasmic reticulum to the Golgi where 0-glycosylation usually

Inhibitor [Inhibitor] 1 4 ~ ~ 3 5 5 s incorporated

Tunicamycin 0 PglassaY 0.45 0.1 pg/assay 0.53 0.5 kg/assay 0.47 5 pg/assay 0.48

10 pg/assay 0.52

2-Deoxy-~-glucose 0 mM 0.057 10 mM 0.042

Monensine 0 PM 0.079 1 PM 0.074 5 pM 0.05

10 pM 0.04

takes place 129, 301, does not inhibit fibre glycosylation (Ta- ble 1). The fibre of all samples had incorporated [14C]GlcNAc as seen by autoradiography of the SDS polyacrylamide gel (not shown). At high concentrations, the same inhibitory effect on protein synthesis was also observed (not shown).

Intracellulur localization offibre glycosylation

The fibre is synthesized in the infected cell cytoplasm and quickly transferred to the nucleus [31]. To compare fibre glycosylation in the cytoplasm and nucleus, cellular fractionation was performed as described in Material and Methods. The fibres found in the cytoplasm and in the nucleus showed the same affinity for WGA (not shown).

As N-glycosylation usually takes place in the endoplasmic reticulum, the endoplasmic reticulum of infected cells was prepared by sucrose gradient centrifugation as described in [27]. Viral proteins were labelled by ['4C]GlcN or [35S]Met incorporation for 18-24 h post infection. Most of the [14C]GlcN radioactivity remained in the bottom of the centri- fuge tube and corresponded to the fibre seen in a SDS] polyacrylamide gel. In addition, two small peaks of [14C]GlcN were found in the endoplasmic reticulum region (Fig. 2). With the [35S]Met-labelled extract, we did not see any preferential fibre accumulation through the gradient with regard to the other viral proteins. Most of the glycosylated fibre seems to be in the cytoplasm.

Glycosylution localization on the fibre polypeptide

Under mild conditions, HC1 cleaves aspartyl-prolyl bonds. The HC1-hydrolysed-fibre products were described and identified in [26]. The 44-kDa and 34-kDa bands correspond respectively to the N-terminal two-thirds and C-terminal two- thirds of the fibre, the 15-kDa band corresponds to the C- terminal one-third and the 17-kDa band could not be identified (Fig. 3 A). When ['4C]G1CN-labelled fibre was hydrolysed, the 44-kDa and 17-kDa bands were strongly labelled whereas the 34-kDa band was never evident and the 15-kDa band was poorly revealed after a long exposure (Fig. 3B). Both the 44-kDa and 17-kDa bands were revealed by WGA after Western blotting (Fig. 3B). Thus, the GlcNAc residue was located in the N-terminal two-thirds of the fibre, and probably in the N-terminal one-third because the 34-kDa band was not revealed. In the model of Green et al. 1321, the C-terminal one-third of the fibre constitutes the fibre knob

208

t

. . ; ( . . . . . . . . . . . . . . : :

: : : :

i : . . . : i, j

0 10 20 30 Fraction number

C Bottom .-> Top

Fig. 2. Post-mitochondria1 supernatant fractionation into rough microsomes. Adenovirus-type-2-infected HeLa cells (about 40 x lo6) were labelled with ['4C]GlcN (15 pCi/ml) or [35S]Met (15 pCi/ml) for 18-24 h post infection. Rough microsomes were prepared as described in Materials and Methods. (A) SDS/polyacrylamide gel electrophoretic analysis of the ['4C]GlcN-labclled gradient extract. Fraction numbers arc indicated; V, virus control. Polypeptide I1 corresponds to hexon, polypeptide I11 to penton base, polypeptide IV to fibre. (B) SDS! polyacrylamide gel electrophoretic analysis of the [35S]Met labelled extract. (C) Radioactivity through the sucrose gradient: ( 0 . . 0 ) ["S]Mct; (0-0) ['4C]GlcN. The bottom is on the left. The cpm were determined on 5O-pl samples, precipitated by 10% trifluoroacctic acid. The gradient of sucrose molarity is drawn

(Fig. 3 C). Thus, glycosylation does not seem to be located in this knob, but in the shaft or in the first 44 amino acids of the fibre. A spontaneously hydrolysed fibre which has lost its first 17 amino acids, as described in [33], keeps its affinity for WGA (Fig. 3 B). Then, the GlcNAc residue is not located in the first 17 amino acids of the polypeptide but between residues 18 and 406 (corresponding to the 44-kDa band) and most probably between residues 18 and 194.

Glycosylution variution umong the fibres of dilfirent serotype adenoviruses

Fibre length, its hemagglutination capacity and antigenicity are known to vary among the different serotypes. These properties are some of the criteria used for adenovirus

classification (reviewed in [7]). Is fibre glycosylation a stable characteristic or does it vary among the different serotypes?

Various serotypes belonging to different subgroups were tested by [14C]GlcN incorporation (serotypes 3 and 7, subgroup B; serotypes 2 and 5, subgroup C; serotype 9, subgroup D; serotype 4, subgroup E). Only subgroup-C viruses (serotypes 2 and 5) had incorporated radioactivity (Fig. 4A). No label was detected in the fibre region for the other serotypes. Yet, WGA failed to reveal the adenovirus type-5 fibre after transfer (Fig. 4C). This might mean that it is another glycan, unless it is more buried in the protein structure and therefore inaccessible to WGA. Indeed, except for the first 50 residues, there is only 69% similarity between the type- 5 and type-2 adenovirus fibres [36]. The cytomegalovirus 149- kDa glycoprotein has no affinity for WGA either [18, 371. Thus, a great variation in fibre glycosylation is observed

209

Asp-Pro Asp-Pro NH2 19L 195 LO6 LO7 COOH

NH2 L3 LOO COOH . -

C i ta i l i shaft knob

Fig. 3. Localisation ofglycosylution on the fibre polypeptide. (A) Sketch illustrating the HCI hydrolysis of the fibre. (B) HC1 hydrolysis of ['4C]GlcN- or ['4C]formic-acid-labelled fibres. The fibres were incubated with 0 mM (lane a), 10 mM (lanes b and c), 25 mM (lanes d and e) or 50 mM (lanes f and g) HCI overnight at 56°C. ['4C]GlcN-labelled fibre was used in lanes a, c, e and g; ['4C]formic-acid-labelled fibre was used in lanes b, d and f ; WGA blotting of HCl hydrolyzed fibre, lane i; and spontaneously hydrolyzed fibre, lane k. Lanes a, h and j, fibre control; V, virus control. (C) Sketch showing the different structural parts of the fibre according to the model of Green et al. [32]

among the different adenovirus subgroups and serotypes. Nevertheless, we cannot say now if these other adenovirus fibres are glycosylated or if they do have GlcNAc in their structure.

Glycosylation and penton formation

Soon after its synthesis, the fibre is assembled with the penton base [31]. Glycosylation could thus be involved in penton baselfibre assembly. To test this hypothesis, we have compared the glycosylation rate of the free fibre with that of the assembled fibre. The quantity of fibre in a fibre or penton preparation was determined by the peak obtained by the rocket technique [34] or by [3H]Val incorporation. In the latter case, the fibre dissociated from penton base by sodium deoxycholate treatment (as described in [351) was purified by chromatography. The glycosylation rate of the free or assembled fibre was always determined by [14C]GlcN incorporation. In all experiments, the ratio of glycosylation of free and assembled fibre varied over 0.86- 1.3, the mean value being 1.36. Consequently, free or assembled fibres seem to be glycosylated at the same rate.

DISCUSSION

In this paper, the 0-GlcNAc linkage in the adenovirus type-2 fibre first suggested by Ishibashi and Maize1 [I] was definitely shown. The 0-GlcNAc linkage to the protein is confirmed by the absence of effect due to inhibitors of N- glycosylation (tunicamycin, 2-deoxy-~-glucose) on [14C]GlcN incorporation into the fibre. Moreover, monensin which inhibits classical 0-glycosylation by impeding protein transfer from the endoplasmic reticulum to the Golgi is also without effect on fibre glycosylation.

Thus, glycosylation of adenovirus type-2 fibre was similar to that of the nuclear-pore-complex proteins recently described [12- 141, although these latter contain many GlcNAc residues/polypeptide unlike the adenovirus type-2 fibre which has only one. Several characteristics are common to these proteins : (a) the sugar composition structure and 0-glycosidic linkage, (b) the same affinity for WGA, (c) the transfer towards the cell nucleus, the adenovirus assembly taking place in the nucleus and (d) the cellular location of glycosylation in the cytoplasm on the outside of the endoplasmic reticulum, as seen by sucrose gradient centrifugation.

21 0

Fig. 4. Glycosylation ofthe fibres qf the dgfereni udenovirus serotypes. (A) SDS/polyacrylamide gel of [’4C]GlcN-labelled cells infectcd by diffcrent serotypes. (B) Coomassie blue staining after blotting. Lanes a and b, two concentrations (20 p1 and 40 pl) of adenovirus type-5; lanes c and d, adenovirus serotype 2 (5 1.11 and 1 Opl). (C) WGA binding to the blot

The role of the glycan residue is unknown. Since fibre glycosylation of other adenovirus serotypes is not a constant characteristic, dcspite an analogous capside structure, glycosylation does not seem to be implicated either in viral adsorption, in penton assembly or in the fibre oligomerisation mechanism. This last point was confirmed by the study of the adenovirus type-5 mutant, the H5 ts 142, which does not glycosylate its fibre at 39.5”C and produces a 6s fibre similar to the trimeric wild-type fibre 1381. Thus, GlcNAc might be a characteristic feature of subgroup-C adenoviruses, the other fibre properties being shared with the fibres of the different adcnovirus serotypes. Moreover, the glycosylation site between residues 18 and 406 of the fibre outside the knob part in the model of Green et al. [32] and the identical glycosylation rate in free or assembled fibres also preclude a role for glycosylation viral in adsorption and penton assembly. It was also described in 1351 that glycosylation occurs at the same site in free or assembled fibres.

Recently, DNA-binding proteins were described to have an 0-GlcNAC linkage and it was suggested that 0-linked GlcNAc residues may play a role in the mechanism or regu- lation of transcription activation of RNA polymerase I1 [17]. Twenty years ago, Levine and Ginsberg [39] showed that purified adenovirus type-5 fibre at a high concentration inhibited the multiplication of type-5 adenovirus, probably by blocking the synthesis of DNA, RNA and proteins. But the fibre concentration used in these experiments was higher than that of the synthesized fibre in an infected cell. In vitro, fibre affinity for DNA was demonstrated [40]. However, adenovirus type-5 fibre did not show affinity for WGA but incorporated [14C]GlcN; these observations are confusing and glycosylation might have a role in these activities.

We wish to thank Dr G . Tamura (Tokyo) for the gift ol‘ tunicamycin and Dr Debray (Villeneuve d’Ascq) for the lectins. We are grateful to Mr P.A. Boulanger who advised us to study the glycosylation fibre, to Prof. A. Verbert and J. C. D’Halluin for valuable criticism during preparation of the manuscript. We also thank Mrs C. Cousin for virus stock, Prof. P. Degand for amino acid analysis, Mrs C. Alonso for her skilful technical assistance in the glycan characterization, and Mrs V. Delecroy for secretarial assistance. and Mr S. Ball for English improvement. This work was supported by the Institut National de la SuntC et de f a Recherche MCdicaIc, by the UniversitC du droit et de 1u suntC de Lille (UERIII), and in part by thc Centre National de la Recherche Scientifique (Unit6 AssociPe N o . 217: Relations structure;fonction des constituants membranuirc.s; Director: Professor J. Montreuil), by the UnivcwitC des Sciences et technique.^ de Lillc Flandres-Artois and by the MinistPre de I’Educution Nationale.

REFERENCES

1. 2.

3.

4.

5.

6. 7.

8.

9.

10. 11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

21. 22.

23. 24.

25.

26.

27.

28.

29.

Ishibashi, M., & Maizel, J . V. (1 974) Virology 58, 345 - 361. Jeng, Y. H., Wold, W. S. M., Sugawara, K. & Green, M. (1978)

Wold, W. S. M., Cladaras, C., Deutscher, S. L. & Kapoor, S. D.

Persson, H., Jornvall, H. & Zabielski, J. (1980) Proc. Nut/ Acud.

Paabo, S., Bhat, B. M., Wold, W. S. M. & Peterson, P. A. (1987)

Kornfeld, R. &Wold, W. S. M. (1981) J . Virol. 40, 440-449. I’ettersson, U. (1984) in The adenoviruses (Ginsberg, H . S.. ed.)

Kornfeld, R. & Kornfeld, S. (1976) Annu. Rev. Biochem. 217,

Montreuil, J. (1 982) in Glycopruteins. Comprehensive hiochemistrj~ (Neurberger A. & Van Deenen L., eds) vol. 19B (1 I ) . pp. 1 - 188, Elsevier, Amsterdam.

Torres, C. R. &Hart , G. (1984) J . Biol. Chmm. 255). 3308-3317. Holt, G. D. & Hart, G. W. (1986) J . Biol. Chem. 261, 8049-

Hanover, J. A,, Cohen, C. K., Willingham, M. C. & Park, M. K.

Holt, G . D., Snow, C. M., Senior, A., Haltiwanger, S . , Gerrace,

Davis, L. I. & Blobel, G. (1987) Proc. Natl Acad. Sci. LISA 84,

Holt, G. D., Haltiwanger, R. S., Torres, R. & Hart. G. W. (1987)

Schindler, M., Hogan, M., Miller, R., & De Gaetano, D. (1987)

Jackson, S. P. & Tjian, R. (I 988) Cell 55, 125 - 133. Bcnko, D. M., Haltiwanger, R. S., Hart, G. W. & Gibson, W.

(1988) Proc. Natl. Acud. Sci.USA 85,2573 -2577. Boulanger, P. A. & Puvion, F. (1973) Rur. J . Biocluw. 39. 37-

42. Zanetta, J . P., Breiickenridge, W. C . & Vinccndon, G. (1972) J .

Chrornatogr. 69, 291 - 301. Gardell, S. (1953) Acta Chem. Scand. 7, 207-218. Michalski, J. C., Peter-Katalini, J . P., Egge, H., Paz-Parent, J . ,

Montreuil, J. & Streker, G. (1984) Carhohydr. Rcs. 134. 177- 189.

J . Virol. 28, 314-323.

(1985) J . Bid. Chem. 260, 2424-2431.

Sci. USA 77,6349 - 6353.

Cell50, 311 -317.

pp. 208 - 270, Plcnum Press, NY.

217-237.

8057.

(1987) J . Bid. Chem. 262, 9887-9894.

L. &Hart , G . W. (1987) J . Cell. B i d . 104, 14847-14850.

7552 - 7556.

J . Biol. Chem. 262, 14847-14850.

J . Biol. Chem. 262, 1254-1260.

Laemmli, U. K. (1970) Naturr 227,680-685. Bsg-Hansen, T. C., Bjerrum, 0. T., & Brogren, C. H. (1 977) A m / .

Towbin, H., Staehelin, T. & Gordon, J . (1979) Pror. Nuri. Artrd.

Caillet-Boudin, M. L., Novelli, A., Gesquiere, J. C. & Lemay, P.

Gaetani, S., Smith, J . A., Feldman, R. A. & Morimoto, T. (1983)

Takasaki, S., Mizuochi, T. & Kobata, A. (1982) Methods

Smilowitz, H. (1979) Mol. Pharrnacol. 16, 202-214.

Biochem. 81, 78 - 87.

Sci. USA 76, 4350-4354.

(1988) Ann. Inst. PusteurlVirol. 139, 141 -156.

Methods Enzymol. 96, 3 -24.

Enzyrnol. 83, 263-268.

21 1

30. Tarlakoff, A,, Vassalli, P. & Detraz, E. (1979) J . Cell. Biol. 79,

31. Velicer. L. F. & Ginsberg, H. S. (1970) J . Vrrol. 5, 338-352. 32. Green, N. M., Wrigley, N. G., Russel, W. C., Martin, S. R. &

Mc Lacklan, A. D . (1983) EMBO 1. 2, 13S7- 1365. 33. Devaux, C., Caillet-Boudin, M. L., Jacrot, B. & Boulanger, P.

(1987) Virology 161, 121 -128. 34. Weeke, B. (1973) Scand. J . Immunol. 2, 41- 56.

694-707. 35. Boudin, M. L. & Boulanger, P. A. (1982) Virology 116, 589-604. 36. Chroboczek, J. & Jacrot, B. (1987) Virology 161, 549-5554. 37. Benko, D. M. & Gibson, W. (1986) J . Virol. 59, 703 -71 3. 38. Cheng Chee-Seung, C. & Ginsberg, H. S. (1982) J . Virol. 42,

39. Levine, A. J. & Ginsberg, H. S. (1967) J . Virol. I , 741 - 751. 40. Levine, A. J. & Ginsberg, H. S. (1968) .I. Virol. 2, 430-439.

932-950.

o-linked glcnac in serotype-2 adenovirus fibre

Documents