depends on by

Vol. 88, pp. 8332-8336,October 1991 MedicalSciences

Expression of proteins encoded by Epstein-Barr virus trans-activator genes depends on the differentiation of epithelial cells in oral hairy leukoplakia

(Epstein-Barrviralantigens/immunotainlng/human Immunodeficiency virusinfection)

J. BECKER*, U. LESERt, M. MARSCHALLt, A. LANGFORD*, W. JILGt, H.

GELDERBLOMt,

AND H. WOLFt§

*Abteilung furzahnarztlicheChirurgie/Oralchirurgieder Poliklinikenffr Zahn-, Mund- und Kieferheilkunde- Nord-, FreieUniversitat,Berlin, Federal Republic of Germany;tMaxvonPettenkoferInstitutder Ludwig Maximilians Universitat, Munich, Federal Republic ofGermany;andtRobertKoch InstitutdesBundesgesundheitsamtes, Berlin,FederalRepublicofGermany

Communicated by Maurice R.Hilleman, June5, 1991 (received for review March 12, 1991) ABSTRACT The Epstein-Barr virus (EBV) immediate

early gene product BZLF1 waslocalizedbyindirectimmuno- fluorescencetothecytoplasmof the basal epithelial layerat the lateral border and dorsum of tongue in human immunodefi- ciency virus-infectedand-seronegative patients. Twobiopsies of oral hairy leukoplakia revealed a sporadic cytoplasmic staining ofthe BHRF1 and BRLF1 geneproducts in the basal epitheliallayer. ThewidespreadpresenceofBZLF1 in the basal epithelial layer indicated that this cell layer contained EBV DNA and was probably directly infected by EBV. Nuclear localization of the immediate early and early gene products BZLF1, BHRF1, BRLF1,and BMLF1 waslimitedtooralhairy leukoplakia in human immunodeficiency virus-seropositive patients and revealed a codistribution with the virus capsid antigen. Ourresultsindicatethat theepithelium ofthe tongue is apotentialreservoirforEBVand that inheavily immuno- compromised patientsEBVmay movefrom thecytoplasmto the nucleus withincreasing differentiationand becoactivated thereduringtheterminaldifferentiationofepithelialcells at the lateral border and dorsum of tongue.

Epstein-Barr virus (EBV) isaubiquitousherpesviruswitha linear double-stranded DNA genome,which has beencom- pletely sequenced (1, 2).Itcausespolyclonal immortalization of human B cells bothin vivoandin vitro(for review,seeref.

3) and persists within latently infected cells as multiple, covalently closed circular episomes (4). EBV infection of lymphoid cells isprimarilylatentasindicatedbytheabsence of antibodiestoearlyEBV-encodedantigens,suchasp54and p138inhealthyEBV-seropositive individuals(5).Inlatently infectedcells viralgeneexpressionis restrictedtoafewviral products: the EBVnuclearantigens (forreview, seeref. 6), the BNLF1 membraneprotein (latentmembraneprotein;ref.

7), EBV-encoded small RNAs(8), andterminalprotein (9).

Thecellular and viral factors that determine whether EBV infection is latent or productive are mainly unknown and dependonthe hostcell (10).

The concept of strict B-lymphocyte tropism has been centraltodiscussion ofthebiologyofEBVsince the discov- eryof this herpesvirus. The observation thatepithelialcells innasopharyngealcarcinomaarelatentlyinfected with EBV and thatlyticexpression ofEBVoccursinepithelialcells of theparotid gland(11) and other sites of the humanbody (12) clearly points to acomplexbiologyof thisherpesvirus.

Virus binding toB cells occurs viaa specific interaction betweenthemajorenvelope glycoprotein gp340and the C3d receptor molecule (CD21; refs. 13and14) and it seemsthat

thespecificity of gp340-CD21 binding plays a major role in determining the viral Blymphotropism(15).

The presence of EBV in the upper two-thirds of the epithelium in oral hairy leukoplakia (OHL) was originally described by Greenspanetal.(16) and waslater confirmed by ultrastructural studies (17), immunohistochemistry, and in situhybridization(18-20).

Themeansby which EBV gainsaccess toepithelial cells in general and the questions of which epithelial cell layer in OHLcould beinfectedby EBV and of how the virusreaches thereplicating cells in OHL have beenthe subjectofnumer- ous studies (15, 21-24). Different modes of infection have beensuggested: Receptor molecules similarbut notidentical toCD21 have been reported from undifferentiated epithelial cells of the oropharynx (22, 25). It is possible that basal epithelial cells in OHL could be infected through such receptors orvia cellfusion(21)byEBV-infectedB lympho- cytes from peripheral blood. Production of EBV in OHL could be initiated later as a consequence of cellular differ- entiation (26). It has been shown that human epithelial cells of normalnasopharynx and tonsil (15),aswellasof theupper andmiddle spinous cell layer of OHL (23),express a200-kDa surface molecule that is antigenetically related to, but not identical with, the CD21 antigen of B cells (15). It is possible that thiscell layer, wherelyticexpression of EBV is observed inOHL, could be infected by free virus in the saliva from the parotidgland. But this receptormolecule,which is detected by monoclonal antibody HB5, is not limited to the lateral border oftongueandis alsopresentin otherparakeratinized areas-i.e.,gingiva,softpalate (23)-wherelytic expression ofEBVhas sofarnotbeen observed.

To getfurtherinsightinto the mode and time of infection of epithelial cells we used antibodies to regulatory and structural gene products of EBV on biopsy material of HIV-seronegative and HIV-infectedpatients.

METHODS

Biopsies.Thirty-threebiopsies (Table1)weretaken under localanesthesia fordiagnosticreasons orduringsurgeryfor otherreasons.Biopsiesof uninvolvedmucosa weretaken in asystematicmannerfrom volunteers after information about theproposed study.Immediatelyafter removalbiopsieswere divided. One portion was fixed with 10%o formalin and embedded in paraffin for routine histology (hematoxylin/

eosin and periodic acid/Schiff staining). The otherportion wasfrozen inliquid nitrogenand storedat -750C.

Abbreviations: EBV, Epstein-Barr virus; OHL, oral hairy leuko-

plakia; HIV, human immunodeficiency virus; VCA, viral capsid antigen;ORF,openreadingframe.

§Towhomreprint requests should be addressed.

8332 Thepublicationcostsof this articleweredefrayedin partbypagecharge payment.This article must therefore beherebymarked"advertisement"

in accordance with 18 U.S.C. §1734 solelytoindicate this fact.

Proc. Natl.Acad. Sci. USA 88 (1991) 8333 Table 1. Clinical dataofbiopsies examined

Histologicaldiagnosis Localization n Sex/age, yr HIVstatus CDC stage

Hairy leukoplakia Lateral border of tongue 12 M/30(23-59) + CIV

Normaloral mucosa Lateral borderof tongue 3 M/41 (34-52) + CIII

Normaloral mucosa Orthokeratinizedgingiva 5 M/32(24-37) + CIV

Smokers' leukoplakia Lateral borderof tongue 4 M/42 (30-57) -

Normaloral mucosa Tongue 6 M/32 (5-53) -

Normal oral mucosa Tongue 1 F/40

Squamous cell carcinoma Tongue 1 M/72

Fibroma Orthokeratinizedgingiva 1 M/67

n, No.individuals; M, male; F, female; ages are given as mean and range; CDC, Centers for Disease Control.

Allpatientswereexamined for humanimmunodeficiency virus (HIV) infection by ELISA and Western blotting. The clinical status was diagnosed according tothe Centers for Disease Control definition (27). All patients wereexamined for EBVinfection and demonstratedIgG andIgM anti-viral capsid antigen (VCA) antibodies except the youngest, a

5-year-old patient, who didnotshowanti-EBVIgG andIgM antibodies.

For indirect immunofluorescence,cryostatsections (5 ,um)

wereair dried atroom temperature for 2 h and fixed with chloroform/acetone for 5 min at 5°C. Sections were incu- bated first with primary antibody (dilution, 1:20) and then with tetramethyl-rhodamine isothiocyanate (TRITC)- conjugated goat anti-rabbit or fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse second antibody (dilu- tion, 1:20; Jackson ImmunoResearch). After each stepsec-

tions were washed three times with phosphate-buffered sa-

line, pH 7.2. The localization of BZLF1 and BcLF1 (p150) included double-staining experiments. These sections were

incubated first with antiserum to BZLF1 followed by the TRITC-conjugated second antibody and then with monoclo- nalantibodytoBcLF1followedby theFITC-conjugatedgoat anti-mouseIgG(dilution, 1:20; Jackson ImmunoResearch).

Controls using second antibodies only wereincluded for

every biopsy. Immunostaining was visualized with a Leitz Orthoplan microscope andphotographedonIlford HP5 films.

Antibodies. Amonoclonal antibodyspecific foropenread- ing frame (ORF) BcLF1wasproduced against authentic viral protein by inoculation of EBV-VCApositive P3HR-1 cells (W.J., Helga Mairhofer, and H.W.,unpublished results) and

wasdemonstratedtobe highly specificonsemithincryostat sectionsof OHL(17). Seraagainsttheproteins expressedby ORFs BZLF1 (amino acids 90-166 of the ORF; exon 1),

FIG. 1. Normal oralmucosa of the dorsumof thetongue of^a HIV-seronegative patient. Strong cytoplasmic fluorescence for BZLF1isshown in the basalepithelial layer. (x600.)

BMLF1 (aminoacids1-165), BRLF1(amino acids 400-600), BHRF1 (amino acids24-138)have beenderived using,8-ga- lactosidase fusion proteins as described (28, 29). A second polyclonal antiserum specific for amino acids 175-189 of BZLF1 (exon 2) was raised by inoculation of synthetic peptides in rabbits.Ac-fos antiserumwaskindly providedby R.Muller (Marburg, F.R.G.).

RESULTS

Biopsies fromHIV-SeronegativePatients.Allbiopsieswere completely negative for BRLF1, BMLF1, and BcLF1(p150) and for the antiserum against c-fos. Four biopsies from normal tongue (three males andone female) andtwo from smokers'leukoplakiarevealedanintensecytoplasmatic flu- orescence for BZLF1 in the basal epithelial layer(Fig.1). The presence ofBZLF1 in three of these biopsies from normal tongue was notlimitedtothe lateralborder oftongue(Fig.2).

The fluorescence for BZLF1 was not present in all basal epithelial cells, extensiveareas werenegative(Fig. 3). Two biopsiesfrom smokers' leukoplakia and threebiopsiesfrom normal tongue, including multiple sections from an EBV- seronegative five-year-old patient, didnotshowanyfluores- cence forBZLF1 in the basalepithelial layerat the lateral border oftongue. The antibodies against exons 1 and 2 of BZLF1 revealed similar results.

Biopsies from HlIV-Infected Patients. Immunolabeling for antibodies against the ORFs BcLF1 (VCA), BRLF1, BMLF1,and BHRF1 waslimitedtoOHLbiopsies(Fig. 3).

Theantiserum against c-fos didnotshowanyimmunoreac- tion as was alsonoted in HIV-seronegative patients.

BcLFJ(p50). TheVCAwasobservedonly in the nuclei of epithelial cells and in the extracellular space in the upper stratum spinosum(Fig. 4A).

$

~~~~~BZLF,

FIG. 2. Schematicdrawingof the dorsumand lateral borderof tongue. Areas that were found to be positive for BZLF1 in the cytoplasmof the basal epithelial layerin our biopsy material are indicated.

Medical Sciences: Beckeretal.

8334 Medical Sciences: Becker et al.

Normal oral mucosa:

Str.corneum Str.granulosum

Str.spinosum -

Str. basale -

BZLF., .

)~~~~~~~~~~~

Connectivetissue

Oral hairy leukoplakia:

2 ~ VCA(BcLF.)

BMLFj.BRLF

<_X -BZLF1,. BHRF.

BZLF.

~~~~~ r..~~~~~~~Blo

) 0 C-e

(

(A.z

.\-4' ^{La~~c~e .l-}

FIG.3. InHIV-seronegativepatients,BZLF1 waslocalizedtothecytoplasmof the basalepitheliallayeratthe lateral border or dorsumof tongue,withtheimmunolabelingshowingas apatchlike distributionandextensiveareasbetween labeledcells remaining unstained.InOHL patients,BZLF1 fluorescence was observed in thecytoplasmof nearly theentire basalepithelial layer.Nuclearstainingforimmediate early andearlyantigensaswell asvirus structuralproteins (BcLF1, VCA) indicatesthe limitation of thelyticcycleof EBV to the upperstratum spinosumand theoverlyingballoonedkeratinocytes. All proteinsexamined revealedanuclearcodistribution in this location.

BZLFL.AllOHLbiopsiesshowedanintensefluorescence forBZLF1throughoutalmost the entire basalepitheliallayer (Fig. 4B). The adjacent stratum spinosum was completely negative.Inthe upperstratumspinosumandin the ballooned keratinocytesanuclearfluorescencewasnoted,anddouble immunofluorescence experiments revealed in this area a nuclearcodistribution ofBZLF1 and BcLF1(Fig. 4).

Onlyone biopsy from normal lateral tongue (Centers for Disease Control stage CIII) showed intense intracytoplas- matic fluorescenceinthebasalepithelial layer.Theothertwo biopsies from lateral tongue were negative as was the or- thokeratinized gingiva. The antibodies against the exons 1 and 2 gavesimilar results.

BRLFL. OHLbiopsiesshowedanuclear fluorescencefor BRLF1 in the upperstratum spinosumand in theoverlying ballooned keratinocytes (Fig. 3). Two patients revealed a sporadicbut strongcytoplasmatic fluorescence inepithelial cellsof the basalepithelial layer.

BMLFL. BMLF1 fluorescence was observed only in the upper stratum spinosum of OHL biopsies and showed a nuclearcodistribution withBcLF1fluorescence(Figs.3 and 5).

BHRFL.BHRF1 showed as a nuclearfluorescencein the upper stratum spinosum and in the overlying ballooned keratinocytes ofOHLbiopsies as wasobserved forBcLF1 (Fig. 3). In twopatients afurther cytoplasmic stainingwas notedin the basalepithelial layer.

DISCUSSION

In vitro models have shown that latent EBV infection is controlledbycomplex interactions betweenBandTcellsas well as their secreted products, immunoglobulins, interfer- ons, andinterleukins. Animportantcomponent of the inhi- bition oflytic expressionin vivo seems tobecooperationof

HLA-restricted cytotoxicT cells and theCD4' subset(30, 31). This EBV-directed immune eliminationdoes not reach all sitesof the humanbodyandit has been shown thatcells of the ductalepithelium oftheparotid glandcanundergoa complete lytic cycle in nonimmunocompromised patients (11).

The replication of EBV inepithelial cells ofthe stratum spinosumatthe lateral borderoftongue inimmunocompro- misedpatientsisanobvioussignthat theimmunesystem has lost theabilitytoeliminatecellsexpressing otherthan latent geneproductsof EBV. Recent individual reports haveshown thatthelyticcycle ofEBVreplicationinOHLisnotlimited toimmunocompromised patients underHIV infection(32).

OHL is therefore not aspecific sign ofHIVinfection,but is related toimmunosuppression in general (33).

Wehave shown the presenceinHIV-seropositivepatients of the EBV immediate early gene product BZLF1 and occasionally ofBHRF1and BRLF1 inthecytoplasm ofthe basal epithelial layerof the tongue. Thesefindingsindicate thatthis celllayer contains EBV genomes and isprobably directly infected byEBV. Inparticular thedemonstrationof BZLF1 in immunocompromised HIV-infected as well as HIV-seronegativepatients indicatesfurther that the tongue is apotentialreservoir for EBV.

Thepatchlikedistribution ofBZLF1-specificcytoplasmic fluorescence in the basalepithelial layerinHIV-seronegative patients suggests that this epithelial layer is probably not continuously infected. Thepresenceof EBV in basalepithe- lialcells could resultfromahorizontaltransmissionof viral genomes during the proliferation of epithelial progenitor cells.

Theincrease ofBZLF1-positivecellsin the basalepithelial layerof OHL may beaconsequenceof thegeneral loss of the immune systemtoeliminate cellsexpressingother thanlatent

Proc. Natl. Acad. Sci. USA 88

(1991)

i;

Proc. Nati.Acad. Sci. USA 88 (1991) 8335

FIG. 4. OHLbiopsies. (A)BcLF1.(VCA). Strongfluorescenceis shown in the nuclei of epithelial cells and in theextracellular space in the upper stratumspinosum.(B) BZLF1.Strongcytoplasmicfluorescenceis shown in the basal epithelial layer and weaker nuclear staining is shown in the upper stratum spinosum. (x600.)

geneproducts asindicated bythelytic cycleof EBV expres- sioninthestratum spinosum. Whenthis isinterpretedas a defectin theeliminationofcellsenteringalytic cycle ofviral replication onecanalsoassume thatfurther epithelialcells could be infected through EBV-expressing B lymphocytes from peripheralblood.

Wehave shownthatthe presenceofthe EBVimmediate earlygeneproductBZLF1 in thecytoplasm ofbasalepithelial cellsisnotlimitedtothe lateralborder ofthe tongue. Recent

FIG. 5. In OHLbiopsiesBMLF1 fluorescencewaslimitedtothe nuclei ofepithelial cells in the stratum spinosum. It was first observedattheperipheryof the nucleiin the lower stratumspinosum (arrow). (x750.)

clinicalfindings underline these results with the description oflytic expressionof EBVatthedorsum oftongue(34,35).

To exclude cross-reactivity between BZLF1 and c-fos in basalepithelialcells(36) weusedantibodiesagainst

dihferent

exonsofBZLF1and anantibody to c-fos. Our results exclude across-reactivitywith c-fos in ourbiopsy material. A further hint for thespecificityof theimmunohistochemicaldetection of BZLF1 in the basallayerwasthedetection ofBHRF1and BRLF1in this location. Thefactthat BHRF1 wasobserved in twosamplesis in agreement withobservationsby Austin etal. (37) of transientexpression of this protein inlympho- cytes.However, ourobservationdoes notnecessarilysup- portdirect involvementof this gene ininduction ofthelytic cycle. The sporadic observation of BRLF1 may reflect a tissue-specificregulatoryeventand lowlevels ofexpression.

Inthe stratumspinosum,where BZLF1wasdetectedin the nuclei, late viral genes areexpressedinregionspositivefor BZLF1. It has been shown that expression of BZLF1 is sufficient todisrupt latency (38, 39). Our findings strongly suggest thatanuclearlocalizationfor BZLF1 (29)isneces- saryfor its trans-activating function on other EBV genes (39-42). BZLF1 maymovefromcytoplasmtonucleus with increasing differentiation in immunocompromised patients and the mechanism may be similartothat ofnuclear factor KB inHIV-infectedTcells(43), which is present in aninactive formduetoformation ofacomplexwithacellularproteinin nonactivated cells and changes its location in response to external stimuli. Alternatively, a cellular protein may be requiredfor activation ofBZLF1, suchasthe cellularprotein E2F,which activates the adenovirus-encoded EMAproteinin adenovirus-infectedcells(44).

Various studiesemployingin situhybridizationhave failed todetect lowcopy numbersofEBV-specific nucleic acidsin the basal epithelial layer of OHL because of the limited sensitivity of this technique (18, 45). Independentproofof EBVinfection of the basallayer,inadditiontothe demon- Medical Sciences: Beckeretal.

8336 MedicalSciences: Becker etal.

stration of BZLF1, is desirable. But there art technical constraints at this time to block otherwise widely used technicalapproaches. Thepolymerasechainreaction,which isahighly sensitive techniquetodetect EBVinHIV-infected patients (45), can in our opinion not be applied for this question,becauseof thedifficultiestoextractonly the basal layerof the stratified epitheliumandtoexclude contamina- tion withfree virus shedding from oropharyngealmucosa.

In conclusion, ourfindings show that the basal epithelial celllayer of the lateralborder andthedorsum oftongueis probably apotentialreservoirforEBVinHIV-seropositive and in nonimmunocompromised HIV-seronegative individ- uals. In heavily immunosuppressed patients latently EBV- infected epithelial cellsmaybecoactivated during terminal differentiation oftheepithelium. Due totheregularstratifi- cationoftheepithelium OHLisanexcellentmodeltostudy the hostcell-dependent activation ofEBV genesusinganti- bodiestodifferentgeneproductsuntil more-sensitivetech- niques forinsituhybridization are available.

Wethank SusanneModrow (Max vonPettenkofer Institute)for BZLFl-derived synthetic peptides and helpful- discussions. This work was supported in patt by the Deutsche Forschungsgemein- schaft (GrantFA-138)andthe Bundesministeriumffr Forschungund Technologie (GrantII-071-88).

1. Skare, J. & Strominger, J.L. (1980) Proc. Natl.Acad. Sci.

USA77, 3860-3864.

2. Baer, R., Bankier, A. T., Biggin, M.D., Deininger, P.L., Farrell,P.J.,Gibson,T.J.,Hatful,G., Hudson, G. S., Satch- well,S. C., Seguin, C.,Tuffnell,P. S. &Barrel,B. G. (1984) Nature(London) 310,207-211.

3. Miller,G. (1985) in Virology, ed. Fields,B. N. (Raven,New York), pp. 563-589.

4. Adamns,A.&Lindahl,T.(1975)Proc. Nadl.Acad. Sci. USA72, 1477-1481.

5. Wolf, H.,Motz,M.,Kilhbeck,R.,Jilg, W., Fan, J., Pi, G.H.

&Zeng,Y.(1987)inEBVandItsDiseases, eds. Levine, P. H., Ablashi,D; V., Nonoyama,M.,Pearson, G.R.&Glaser,R.

(Humana, Clifton,NJ), pp. 211-215.

6. Dinner,J.&Kallin,B.(1988)in Adv. Cancer Res.50,95-157.

7. Modrow, S. &Wolf,H.(1986)Proc.Ndal.Acad.Sci. USA83, 5703-5707.

8. Howe,J.G. &Steitz,J. A.(1986)Proc.NatI.Acad.Sci. USA 83,9006-9010.

9. Laux,G., Perricaudet, M. & Farrell, P. (1988)EMBOJ. 7, 769-774.

10. Rowe, M.,Rowe, D.T., Gregory, C. D., Young,L. S.,Far- rell, P.J.,Rupani,H.&Rickinson, A.B.(1987) EMBOJ.6, 2743-2751.

11. Wolf,H.,Haus,M.&Wilmes,E.(1984)J. Virol.51, 795-798.

12. Sixbey,J.W., Lemon, S.M.&Pagano,J. S.(1986)Lancet2, 1122-1124.

13. Nemerow,G. R., Mold,C.,Kievens, Schwerd, V.,Tollefson, V. &Cooper,N.R. (1987)J. Virol. 61,1416.1420.

14. Tanner, J.,Weis, J., Fearon,D.,Whang,Y.&Kieff,E.(1987) Cell 50, 203-213.

15. Young, L. S., Dawson, C.W., Brown, K. W. & Rickinson, A. B. (1989)Int. J. Cancer43,786-794.

16. Greenspan, J.S., Greenspan, D., Lennette, E.T., Abrams,

D. I.,Conant,M. A.,Petersen,V. &Freese,U. K.(1985)N.

Engl.J. Med.313, 1564-1571.

17. Zhang, X., Langford, A., Becker, J., Rabanus, J.P.,Pohle, H.D., Reichart,P.&Gelderblom,H.(1988)VirchowsArch.A 412, 533-542.

18. Loning,Th., Henke, .R.-P.,Reichart, P. &Becker, J. (1987) Virchows Arch. A412, 127-133.

19. Sydinen,S.,Laine,P. &Valle,S.-L.(1988)Proc. Finn. Dent.

Soc.84,127-132.

20. Reichart, P., Langford, A., Gelderblom, H., Pohle, H.-D., Becker,J. &Wolf,H.(1989)J.Oral Pathol. Med. 18 410-415.

21. Bayliss,G. J. &Wolf,H.(1980)Nature(London)287,164-165.

22. Sixbey,J. W.,Davis,D. S., Young,L.S., Hutt-Fletscher, L., Tedder, T. F. & Rickinson, A. B. (1987) J. Gen. Virol. 68, 805-811.

23. Corso, B., Eversole, L. R. &Hutt-Fletscher, L. (1989)Oral Surg. 67,416-421.

24. Sixbey,J. W. (1989)Advances ViralOncol.8, 187-202.

25. Sixbey,J.W., Nedrud,J.G.,Raab-Traub, N., Hanes, R.A.&

Pagano,J. S.(1984) N. Engl. J. Med. 310,1225-1230.

26. Becker,J., Leser, U., Marschall, M., Langford, A., Jilg,W., Wilmes, E., Reichart,P.,Gelderblom,H. & Wolf, H.(1988) J.

Exp.Clin.Cancer Res.7,80.

27. Centersfor DiseaseControl(1986)Morb. Mortal. Wkly.Rep.

35, 757-766.

28. Seibl, R.,Motz,M. &Wolf,H.(1986)J. Virol. 60,902-909.

29. Marschall, M., Leser, U., Seibl,R. &Wolf,H.(1989)J. Virol.

63, 938-942.

30. Konttinen, Y. T.,Bluestein,H.G. &Zvaifler,N.J.(1985)J.

Immunol. 134,2287-2293.

31. Blumberg,R.S., Paradis, T., Byington, R., Henle, W.,Hirsch, M.S. & Schooley,R. T.(1987)J.Infect. Dis. 155,877-890.

32. Greenspan,D.&Greenspan,J. S.(1989)J.Oral Pathol. Med.

18, 32-34.

33. Sydjinen,S.,Laine, P., Happonen,R.P. &Niemela,M.(1989) J. Oral Pathol. Med.18,28-31.

34. Eversole,L.R., Jacobsen,P.,Stone, Ca E. & Frackleton, V.

(1986)OralSurg. 61,249-255.

35. Kabani, S., Greenspan, D., de Souza, Y.,Greenspan, J. &

Cataldo,E. (1989)Oral Surg. 67, 411-415.

36. Farrell, P., Rowe, D. T., Rooley, C. M. & Kouzarides, T.

(1989)EMBO J. 8, 127-132.

37. Austin, P.J., Flemington E., Yandava, C. N., Stromiinger, J.L. &Speck,S. H. (1988)Biochemistry85,3678-3682.

38. Takada, K.,ShimizujN., Sakuma, S. & Ono, Y. (1986)J. Virol.

57,1016-1022.

39. Countryman, J., Jenson, H.,Seibl, R., Wolf, H. & Miller, G.

(1987)J.Virol. 12, 3672-3679.

40. Chevallier-Greco, A., Manet, E., Chavrier,P., Mosnier, C., Daillie, J. &Sergeant,A; (1986) EMBO J.5,3243-3249.

41. Hardwick,J. M.,Lieberman,P. M. &Hayward,S. D.(1988) J.Virol.62, 2274-2284.

42. Kenney, S. J., Kamine, J., Markovitz, D., Fenrick, R. &

Pagano,J.(1988) Proc. Matd. Acad. Sci. USA85, 1652-1656.

43. Baeuerle, P. A. &Baltimore,D. (1988)Cell53,211-217.

44. Thalmeier, K., Synovzik, H., Mertz, R.,Winnacker,E. L.&

Lipp, M. (1989) Genes Dev. 3,527-536.

45. Snajders, P. J. J.,Schulten,E. A. J. M., Mullink, H.,tenKate, R.W., Jiwa, M., van der Waal, I., Meyer, C. J. L. M. &

Walboomers,J. M. M. (1990)Am. J. Pathol.137,659-666.

Proc. NaM Acad Sci. USA 88

(1991)