AtX)Nts 0300947591(X)O53P .Scand. J. Immunol. 33,435-440. 1991

Glycosylation is Required for CoronavirusTGEV to Induce an Efficient Production ofIFNa by Blood Mononuciear Cells

B. CHARLEY. L. LAVENANT & B. DELMASLaboratoire de Virologie et d'Immunologic Moleculaires, LN.R.A., Centre de Reehcrehes de Jouy-en-Josas, Domaine de Vilvert. Jouy-cn-Josas. France

Charley. B. Lavenant. L. & Delma.s. B. Glycosylation is required for Coronavinis TGEV toInduce an Bllicient Production oflFNx by Blood Mononuciear Cells. Seimtf J. Immunot. 33,435440. 1991.

Porcine peripheral blood mononuciear cells (PBMCl are induced to produce interferon alpha(IFNa) following in vitro exposure to eoronavirus TGEV (transmissible gastroenteritis virus)-infected glutaraldehyde-fixed cell monolayers or lo TGEV virions. In the present reporl. weexamined the possibility that glycosylation of viral proteins could play a major role in interactionswith PBMC leading to the production ol" IFN-i. Con A pretreatment of TGEV-infected cellmonolayers before fixation with glutaraldehyde and exposure to PBMC caused adosc-dcpendentinhibition of IFN:e induction, implying that masking of carbohydrates at the surface of infectedcells lowered IFN-a-induciion. Similarly, inhibition of N-linked glycosylation by tunicamycinduring viral infection of cell monolayers altered their ability to induce IFNa. In addition,complete cleavage of'complex type' oligosaccharides by pcptide-N-glycohydrolase F lowered thecapacity of TGEV virions to induce IFNx. Thus, these tindings strongly suggest thatglycosylation ofthe vira! proteins, and more precisely the presence of complex-type oligosacehar-ides, is an important requirement for a completely ellicient interaction with PBMC leading to theproduction of IFN-a,

B. Ctjartey. Lahoratnirc de V'irotogie et dlmmunotogte Moteeutaires. I.N.R.A.. Centre deReetiereties de Jouv-en-Jti.\as. Domtiine de Vitven. 7fi..15O Jiiii\-en-Josas. France

Various stimuli such as viruses, bacterial pro-ducts or tumour cells can induce leucocytes toprodtice interferon alpha (IKNx)[7]. The demon-stration thai inactivated vira! particles or gluiar-aldchyde-fi.xcd virus-infected cells could induceIFNx [2. 4. 10. 12. 13, 17] led to the suggestionthat the IFN-inducingcapacity wasnot related tovirus replication but could rather reflect directmembrane interactions between leucocytes andviral proteins present at the surface of viralparticles or virus-infected cells [7. 17].

Studies conducted to characterize the nature ofIFNx-produccr cells (IPC) in response to suchstimuli, indicated that IPC were infrequent buthighly efficient mononuciear cells [3, II. 23]. IPCcould be characterized in several mammalianspecies as non-adherent, non-T. non-B cells.expressing MHC class II and CD4 molecules [5.9,18.21.22.24].

In contrast, few reports arc available about the

nature of the viral component involved in mem-brane interactions with lymphocytes leading tothe activation of IFN-a-coding genes. Thus, inthe course of studies on IFNa induction by herpessimplex virus (HSV). Lebon [19] showed thaimonoclonal antibodies (McAbs) directed at themajor viral glycoprotein D could inhibit HSV-induced IF'ND: production. These data suggestedthat a defined glycoprotein could play a crucialrole in IFNsi-induction. ln the case of transmiss-ible gastroenteritis virus (TGEV). a coronaviruswhich induces acute diarrhoea and intense IFNaproduction in newborn piglets [14] we conductedexperitncnts in which McAbs directed at the threedifferent viral structural proteins were addedduring in vitro exposure of leucocytes to infeetedcells. It appeared that IFNx-induction could beblocked hy two of four McAbs directed at theIransmetiibraneglycoprotein M (formerly namedEl) whereas McAbs specific for the other strue-

435

436 B. Charter. L. Lavenant & B. Dehiias

tural proteins (N and S) had no effects [4]. Morerecently, a series of epitope virus mutants wereselected on Ihc basis of their resistance to theIFN-induction blocking McAbs. Al! the relevantatnino :ieid substitutions identified hy sequencingthe mutant M genes were localized within the first22 N"terminal residues of the molecule. Further-more, two mutations were shown to markedlylower ihe IFNa-inducing ability of the corres-ponding viral mutants (Laudc ct at., submittedfor pulication). All these findings strongly impliedtbat IFNa-induction by TGEV could result frominteractions between leucocytes and a short N-terminal domain of the viral glycoprotein M.Interestingly, one ofthe mutations altering lFNx-induction affected the unique functional N-glyco-sylation site ofthe glycoprotein M (Laude et at..submitted for publication). It was thereforetempting lo postulate that carbohydrate moietieson ihe viral protein M could pUty a crucial role forinducing IFNa. This question was addressed byusing three different approaches. (1) Because ofits high capacity to bind to carbohydrates [I]Concanavalin A (Con A) was added to infectedcell monolayers before addition of leucocytes. (2)TGEV-infecied cell monolayers used to induceIFNof were pretreated with tunicamyein, an anti-biotic which inhibits N-glycosylation [8]- (3) Puri-fied viral particles were treated with glycosidaseswhich cleaved N-linked oligosaccharides from theviral protein M, before addition to leucocytes forinducing IFNa. This report shows that the pres-enee of 'complex type" carbohydrates on TGEVpolypeptides is required for efficiently inducingleucocytes to secrete !FNa.

MATERIALS AND METHODS

Preparation of PBMC. Porcine periphenil bloodmononuciear cells (PBMC) were prepared from hepari-ni/cd hlood collected from 2 4-month old animals, bycentrifugation on Ficoll (MSL. density 1.077, Eurobio.Paris. France). The PBMC were rcsuspended in RPMI-1640 medium supplemented with 10"" heat-inactivatedfetal calf serum. 2 niM L-g!uiamine and antibiotics.

Viru.s. AsasourceofTGi-V. we used thecell-adapicdPurdue 115 strain. Methods for propagation andtitration were as described by Laude et at. []5\.

IFN induction. PBMC were indueed to produce IFNshy overnight incubation on TGEV-infecled. glutaralde-hyde-fixed cell monolayers as described previously [4]:brielly. pig kidney cells (PD5) were plated in 96-wellmicroplates. infected by the coronavirus TGEV for18h. then fixed with 0.25% glutaraldehyde (1 h at 4 C)and stored with 3% giycine. Monolayers were washed

before addition of PBMC {100 fi\ per well at 5 x 10''. ml).Supernatants were colleeted after 18 h of incubation at37 C and assayed for IFN activity.

Con .4 treatment tif TGEV-infected cett monotayers.Pig kidney cell (PD5) monolayers. infected for 18 h.were washed once with PBS, ihen incubated for I h al37 C with various amounts of Con A (Miles-Yeda,Israel) before being fixed with glularaldehyde as de-scribed above.

Tuntiamycin treatment of 'tGEV-infected cett mono-tayers. At I h post-infectton, tunicamycin {Serva.Heidelberg. FRG) was added to the pig kidney celltnonoiayers al a dosage of 2 /ig/ml for 18 h. untilfixation by glutaraldehyde.

Gtyco.sida.se treatments. Endo-/J-N-acetylglLicosami-nidase H (endo H). endo-/f-N-acetylgltieosaminidase F{endo F)and peptide-s-glycosidase F(PNGase F) werepurchased from Boehringer-Mannheim-F""ranee (Mey-lan. France}. Aliquots of 20 /J1 of purified TGEVsuspensions (500 /ig/ml) prepared as described byLaude et at. 115], with the addition o\' 3 /d of "S-labellcdTGEV [6]. were incubaled with 20 ;jl of glycosidase and20 ;il of 20 mM sodium phosphate (pH 7.2). 50 msiEDTA solution, for 18 h at 37 C. The final concentra-tions of enzymes were 60 U/ml for glyeosidase F, 15 U/ml for endo F and 0.3 U/ml for endo H. The resultingpreparations were used to induce PBMC as follows:serial 10-fold dilutions of 10 ii\ aliquots were added toPBMCCi X IO'vml) in a final volumeofO.2 ml for IS hat37 C. In order to ascertain the en-^ymatic cleavage. 6 ;iialiquots of each preparation were subjected to electro-phoresis on ]O"'li polyacrylamide gels which were thenprocessed for Huorography.

IEN hioa.ssav. Log3 dilutions of PBMC supernatantswere assayed for IFN on bovine MDBK cells usingvesicular stomatitis virus as a challenge 114|. A standardporcine IFNa was included in each assay. This standardwas ealibraled on MDBK cells with the human inter-national reference IFN B69/19 (NIH. Bethesda. MD.USA). In our results. I U is equivalent to 1 IU of humanIFN.

RESULTS

Effects of Con A treatment of TGEV-infectedeetls on IENct induction

PBMC exposed lo TGEV-infected glutaralde-hyde-fixed cell monolayers are induced lo secreteIFNa: in the results shown in Table I. 5x 10''PBMC per ml produced 1000 U/ml IFN follow-ing 18 h incubation at 37 C. In order to evaluatethe role played by viral glycoproteins present atihe surface of infected cells for the induction ofIFNof, Con A was added to TGEV-infected celltnonoiayers before fixalion wiih gtutaraldchyde.Table I shows that Con A treatment causes adose-dependent inhibition o\' IFNx-induction.Control experiments showed that uninfected cellmonolayers. with or without Con A treatment.

Glycosvhtion for lENi Induction 437

TABLE I. Effect of Con A treaimcni of TCJEV-infcctedcell monolayers on their ability to induce PBMC to

ConO'P •irN

A concentrat ion*nl)titres (U ml ) '

IIKKHI

(1.5 S

M)50

>3500>3

1. Dala represeniative of Iwo experiments.2. Various amounts of Con A were added lo TGEV-

infected pig kidney cell monotayers 1 h before glutaral-dehyde fixation.

?. IFN produced by PBMC at 5 x IO*;ml exposed tocell monolayers for 18 h.

did not induce PBMC to produce IFNx (notshown).

Effects of tunicamycin treatment ofTGEV-infected ceils on lENoL-imluction

Pig kidney cell monolayers were ireaied bytunieamyein for 18 h following initiation of viralinfeciion. then fixed with gttitaraldehyde. TunJea-myin-treated eell monotayers indueed PBMC toproduee 10 IOO times less IKN than tnoek-trcated eeil monolayers (Table 11; four indepen-dent experiments). It was shown previously thattunicamycin (at 2/ig/ml) inhibited N-tinked glyeo-sylation with slight eflects on ihe overall eellprotein synthesis (Dehiias and Laude. submittedfor publieation). Our present dala indicate thattunicamycin markedly affects the ability ofTGEV-infected cells to induce IFNa production.

II. Effeet of tunicamyein irealment ofTGEV-infected cell tnonolayors on their capa-city to induce PBMC to secrete Il-N^'

Tuniciimycin(2 ;ig/ml)treatment*

P B M C conccnt ra l ion (lO^/ml)

2.5 10

1000^100

9000 90001000

1. OHVJ representative of four experiments.2. TGEV-infected cell monolayers were

treated or not with 2 >ig.nil lunicamycin duringIS h, then fixed with glut;ir;ildchydc.

3. IEN (U ml) produLxd hy PBMC exposedto cell monolayers for IK h.

Effects of virus digestion by glycosidases onIENa induction

In order lo further analyse the role played byoligosaccharide moieties of TCiRV proteins forIFNa induclion. we examined the elTects ofpretreating virus particles by gtycosidascs beforeincubation with PBMC. Three endoglyeosidasesable to cleave N-linked oligosaccharides wereused: endo H and endo F eleave predominantlyhigh mannose oligosaccharide chains, whereasPNGase K cleaves both high mannose and eom-ptcx-lype saccharides [27]. Virus particles pre-treated with endo H or endo F induced the sameamounis of IFNa as control virions (Fig. I). Incontrast, PNGase F markedly decreased thevirions* ability to induce PBMC production ofIFN'DC Fig. I shows that the same amount ofPNGase F-treated virus could induce 10-100times less IFNi than other vira! preparations.SDS-polyacrylamide ge! electrophoresis controlanalysis, performed on the mixture of cold and.15S labelled TGEV. showed that endo H andendo Kcatised a Mr shift ofthe major 29 kDa Mprotein species to 26 kDa. while 30 36 kDaspeeies remained unaffected (Fig. 2). In contrast.PNGase F caused a complete carbohydrate clea-vage leading to the presence ol' only one 26 kDamolecular species. The nucleoprotcin N was notaffected by digestion which indicates that proteo-lytic cleavages of viral polypeptides were unlikelyto occur during glycosidase treatments. When

10" F

I 2 3Log virus dilution

F I G . L ElTects of glycosidase treatment of T G E virionson lKN^ induct ion. Aliquots of virus particles werepret rea ted with P N G a s e E ( • ) . endo I (-f-). endo H ( x )or control medium ( • ) for IU h. P B M C (2.5 x lU'Vnil)were induced to produce IENa by ii jcubanun (IU h)with U)-fnld dilutions of eaeh virus prepara t ion . (Datarepresentat ive of four experiments.)

438 B. Charley. L. Lavenant & B. Delmas

oocHZou

U-(U(ft

roOza.

oDZLJ

I

oOzUJ

N 4 7 K _

M 29-36K

M 2 6 K _

Eiu. 2. SDS-polyacrylamide gel electrophoresis of con-trol or glycosidase-treated virus particles.

ciliquots of virus preparations were incubatedwith PBMC. a residual PNGase F activity mighthave directly acted on PBMC. thereby alteringtheir ability to secrete IFNa. This did not appearto be the case since the induction of IFNaremained unchanged in a control experiment inwhich PNGase F was added to the virus justbefore incubation with PBMC (Fig. 3).

102

10'

2 3 4 5 6Log virus dilution

EiG. 3. Direct cfTect of PNGase E on virus-inducedPBMC. PNGase E was incubaled with virus particlesfor 18 hr at 37 C (•) or just before addition to PBMC(O), Controls consisted of untreated virus particles {•).

DISCUSSION

The resutts of the present investigation indicatethat glycosytation of coronavirus TGEV proteinsis of major importance for induction of IFNaproduction by blood mononuclcar cells.

The fact that potent IFNa induction isachieved by lixed cells, as already described [2, 4,13. 17], strongly suggested that TFNx induction isindependent of virus penetration into PBMC butcould rather reflect direct inleraetion betweenmembrane-associated viral antigens and PBMCmembranes [7, 19]. In a lirst set of experiments,TGEV-infected cells were treated with Con Abefore fixation with glutaraldehyde: such treat-ments caused a dosc-dependcnt inhibition in theability of cell tnonoiayers to induce PBMC toproduce IFNa (Table I). Beeause Con A has ahigh capacity to bind to complex-type carbo-hydrates [26]. these data imply that rnasking ofcarbohydrates present at the surface of infectedcells will impede adequate interaction withPBMC leading to IFNa-production. However,the possibility that the Con A inhibitory effectmight be due to steric hindrance of the relevantstructure could not be excluded. In addition. ConA could bind to cell membrane glyeoproteins aswell as viral glyeoproteins.

A more precise experimental approach con-sisted, therefore, in treating TGEV-infeeted cellswith tunieamyein, an antibiotic which blocks N-linked glycosylation [8]. Such a treatment waspreviously shown to have slight effects on iheoverall eell protein synthesis and M and Spolypeptides were produced with Mr correspond-ing to those predicted for the apoproteins (Del-mas and Latide. submitted for publication}. Inour experiments, tunicamycin redueed theabilityof infected cell monolayers to Induce I FNa (TableII), which implies that inhibition of glycosylationin cell monolayers has negative effects on theexpression of an "IFNa-inducing-signal* at theeell surface. However, since tunicamyein eanaffect glycosylation ofcell membrane proteins inaddition to virus-eoded proteins, it is not possibleto determine from these data whether reductionof IFN-induetion is related to the absence ofcarbohydrate moieties on viral or on cellularproteins. Moreover, tunicamycin could alsoblock the transport ofthe viral apoproleins to thecell surface (see for example. Ref. 20).

For these reasons a last set of experiments wasconducted in which virus particles were treated

GIvcosvlation for lENi Induction

with glycosidases before being incubated withPBMC in order to induce IFNx. Three enzymeswere used: two glycosidases (endo H and endo F)eleave predominantly high mannose oligosae-eharide chains whereas peplide-N-glycohydrolaseV (PNGase F) cleaves both high mannose and'complex-type' oligosaccharides [27]. Our resultsshow thai PNGase F was the only enzyme whichcould reduce TGEV virions ability to induceIFNa (Fig. I). A control experiment showed thatthis rcduciion was nol due lo a direct efiecl ofPNGase F on PBMC (Fig. 3). When treatedvirionaliquois were tritratedon PBMC (Fig. I)itappeared that a reduced IFNa-induetion wasobtained with diluted PNGase F-treatcd viralpreparations whereas high concentrations ofPNGase F-treated virus (about 1.5 /ig virions for5 X 10̂ PBMC) induced as much IFNo: as control,endo H-or endo F-treated virus. We have alreadyreported strong evidence for the role of glycopro-tein M as the effector viral molecule for IFNx-induction ([4]; Laude ct at., submitted for publica-tion). It is therefore interesting to notice thatPNGase F was also the only glycosidase testedwhieh could cleave completely oligosaecharidcsfrom the viral protein M. leading to the produe-tion of a 26 kDa molecule (Fig. 2). correspondingto the size predicted for the apoprotein M [ 16] andfound after tunicamycin treatment (Delmas andLaude. submitted for publication). Moreover,since PNGase F is able to cleave 'complex-type'oligosaccharides in contrast to endo H or endo F[27]. our results indieate that only a completecleavage of'complex-type' oligosaeeharides fromthe viral protein M alters the virus" ability toinduce PBMC to produee IFNa.

What is the role(s) ofearbohydrates, and moreprecisely eomplex-type oligosaeeharides. presentat the surface of virus particles or virus-infectedcells, in the induction of IFNa? It is unlikely thatcarbohydrates alone are the 'IFNa-inducing sig-nal': in fact, we show here that deglyeosylatedvirus particles do induce IFNa when exposed athigh concentrations to PBMC (Fig. I). Onepossible explanation is that glycosylation of viralprotein M is an itnportant structural character-istic creating Ihe adequate conformation requiredfor it to interact with a putative receptor onPBMC membrane, leading to activation of IFNagenes. The need for adequate conformation forIFNj-induction was also suggested by our pre-vious observalion that a mutation on the glyco-protein M, which should cause important confor-

tnational changes, greatly impaired IFNot-induetion (Laude et at., submitted for publica-tion). The existence of a membrane receptorinvolved in IFNa-induction was suggested by theresults from Lebon [18] that IFNa-induetion wassensitive to lysosomotropic drugs.

Alternatively, carbohydrate moieties of viralpolypeptides may influence IFN:(-induction byincreasing the stability of complexes formedbetween virus panicles or virus-infeeted cells andPBMC. With that respect, adhesion receptorstermed "seleetins" were recently demonstrated onlymphocytes [25]: these molecules contain an N-terminal lectin domain potentially able to interactwith carbohydrates. Cell-surface molecules equi-valent to such 'seleetins' could bind virus-asso-ciated carbohydrates, which in turn would facili-tate stable interactions between PBMC and theactual "IFN-inducing domain' of the viral poly-peptide, leading lo IFNa synlhesis. Obviously itwould be interesting to know whether carbo-hydrates are important for induction of IFNa byviruses other than TGFV. and which of themechanisms we suggested (i.e.. adequate confor-mation or adhesion requirement) is involved inthe IFNa-induction process. A consequence ofour findings is that synthetic peptides corres-ponding to the M protein or Esherichia coli-produeed unglycosytated reeombinant M pro-teins by themselves will be unlikely to efficientlyinduce IFNa, unless used at high concentrations.In addition, carbohydrates such as "complex-type" oligosaccharides may be useful tools toprobe the infrequent leucocyte subpopulationinvolved in the production of IFNa.

ACKNOWLEDGMENTS

We thank Drs R. Hedriek (Davis. Calif, USA)and H. Laude (Jouy en Josas. France) for criticalreading of the manuscript.

REFERENCES

1 Baenziger. J. L), & Fictc. D. Structural determinantsof Concanavalin A specilicity for oligosaccharides.J. Biol. Ctiem. 254. 2400. iy79.

2 Capobianchi. M.R., Facchini, J.. Di Marco. P.,Anlonelli. G. & Dian7iini. F. Induction of alphainlerferon by membrane iutcractioti between viralsurface and peripheral blood tnononuciear cells.Prac. Soc. E.xp. Biot. Med. 178. 551. 1985.

440 B. Charli'v, L. Latcnant & B. Detmas

3 Cederblad. B. & Aim. G.V. Infrequent btit efficientinterferon-a-producing human mononuclcar leuko-cytes induced by Herpes Simplex Virus in titrostudies hy immuno-pluqtte and limiting dilutionassays. J. Interferon Res. 10, 65. 1990,

4 Charley, B. & Laude. H. Induction of alpha inter-fcron by transmissible gastroenteritis coronavirus:Role of tran.smembrane glycoprotein El. 7. I'irot.62, S, [988.

5 Charley. B. & Lavenant. L. Characterization ofblood mononuciear cells producing IF'Ny followinginduction by corona virus-infected cells (PorcineTransmissible Gastroenteritis Virus). Res. Immunot.141. 141, 1990.

6 Delmas. B. & Laudc, H. Assembly of coronavirusspike protein into trimers iind its role in epitopeexpression. J. Virot. 64, 5367, 1990.

7 Dianzani. P. & Capobianchi. M.R. Mechanism ofinduction ot\ilpha interleron. Pp. 21 .10 in Baron.S.. Dianzani. P., Stanton. J. & Fleischmann. W.R.(eds) Ttw Interferon Sy.stem. A Current Rerien tot9H7. University oTTexas Medical Branch Scries inBiomedieal Science, I9S7.

8 Elbein. A.D. Inhibitors of the hiosyrthesis andprocessing of N-linked oligosaecharide chains. Ann.Rev. Biocttem. 56. A-il. !9X7.

9 Fitzgerald-Bocarslv, P., Feldman, M.. Mendelsohn,M.. Curl. S. & Lope?, C. Human mononuciear cellswhich produee interferon-alpha during NK (HSV-FS) assays are HLA-DR positive cells distinct iromcytolylic Natural Killer eflectors. J. Leuk. Biot. 43,323. 1988.

10 Gaillard, C , Chilmonezyk. S. & de Kinkelin. P.Interferon synthesis indueed by Viral Hemorrhagic.Septiecmia Virus (VHSV) in rainbow trout leuko-cytes in vitro. Proceedings ofthe Xlilth Eish HeatlhSection!American Fisheries Society Meeting p. 52.19i*9.

11 Gobi. A.E.. Funa, K. & Aim. G.V, DilTerenlinduction patterns of iiiRNA lor IFN-x and -p inhuman mononuciear leukocytes after in vitro stimu-lation with Herpes Simplex virus-infected fibro-biastsand Sendai virus../. Immunot. 140,3605. 198X.

12 Ito. Y,, Nishiyama. Y.. Shimokala, K.. Nagata, J..Takeyama. H. & Kunii. A. The mechanism ofjnterferon induction in mouse .spleen cells stimu-lated with HV.I. r(W./(.^i-88, 128. 1978.

13 Kurane. I., Meager. A. & Ennis. F,A. Induction ofinterfcron alpha and gatiima from human lympho-cytes by Dengue virus-infeeted eells. 7, Gen. Virot.67, 1653. 19S6.

14 La Bonnardiere. C. & Laude. H. High interlerontiter in newborn pig intestine during experimentallyinduced vira! enteritis, tnf. tmmun. 32, 28, 1981.

15 Laude, H.. Chapsal, .I.M., Gelti. J.. Labiau, S. &Grosclaude, J. Antigenie structure of transmissiblegastroenteritis virus. 1. Properties of monoclonal

antibodies directed against virion proteins. 7. Gen.[•(>W. 65, 119, [986.

16 Laudc. H..Ra.sshaert.D.& Huet. J.C. Sequence andN-termina! processmg of the trans-membrane pro-lein E| ofthecoronavirustransmissiblegastroenter-itis virus. J. Gen. Viriot. 68, 1687. 1987.

[7 Lebon. P.. Commoy-Chevalier. M.J., Robert-Cal-liot & Chany. C. Different mechanisms for a and fiinterferon indtietion. I'irotogy 119, 504, 1982.

18 Lebon, P., Bernard, A. & Boumsell. L. Humanlymphocytes involved in a-intcrreron productioncan be identilied by monoelonal antibodies directedagainst cell surface antigens. CR. .4cad. Sei. Pari.s295,79. 1982.

19 Lebon, P,. Inhibition of Herpes Simplex virus typet-induced interfcron synthesis by monoclonal anti-bodies against vira! glycoprotein D and by lysoso-motropic drugs. J. Gen. Virot. 66, 2781. 1985.

20 Machamer. C.E.. Florkiewicz. R.Z. & Rose. J.K. Asingle N-linked oligosaceharide at eilher ofthe twonormal sites is sufficient for transport of vesicularstomatitis virus G protein to the cell surfaee. Mol.Cett. Biot. 5. 3(t74. 1985.

21 Oh, S,H.. Bandyopadhyay, S.. Miller. D,S. & Starr,S.E. Cooperation between CD16(Lcu-l Ib)-^ NKcells and HLA-DR* cells in Natural Killing ofHerpesvirus-infected fibroblasts. / . Immunot. L39,2799. 1987.

22 Peter, H.J., Dalligge. H.. Zawatzky, R.. Euler. S..Leibold. W. & Kirchner, H. Human peripheral nulllymphocytes. IL Producers of type 1 interferonupon stimulation with tumor cells, herpes simplexvirus and Corynelmcterium parvutn. Eur. J. Immu-not. 10,547, 1980.

23 Ronnblom.L.. Cederblad. B.,Sandberg.K.& Aim,G.V. Determination of Herpes Simplex Virus-induced alpha interfcron-seereting human bloodleucocytes by a filter immuno-plaijue assay. Scand.J. Immunot. 11, 165. 1988.

24 Sandberg, K., Matsson, P. & Aim. G.V. A distinctpopulation of nonphagocytic and low level CD4*null lymphocytes produce IFN-a after stitnulationby Herpes'Simplex Virus-infected cells.,/. Immwtot.145,1015,1990,

25 Springer, T.A. Adhesion reeeptors of the immunesystem. Nature 346, 425. 1990.

26 Sutton, C. Lectin affinity ehromatography. Pp. 268-282 in Harris. E.L-V. & Angal, S- (eds) ProtetnPurification Methods, a Practicat .Approach. IRLPress. Oxford. 1989.

27 Tarentino. A.L.. Trimble. R.B. & Plummer. T.H.Enzymatic approaches for studying the structure,synthesis, and processing of glycoproteins. Methodsin Celt Biol. 32, 111. 1989.

Received 16 October 1990Accepted 15 November 1990