Polypeptides of hepatitis B virus surface antigen produced by a hepatoma cell line.

0022-538X/79/12-0796/07$02.00/0

Polypeptides

of

Hepatitis

B Virus Surface

Antigen Produced

by

a

Hepatoma Cell Line

PATRICIA L.MARION,' FELIX H.SALAZAR,' JENNIFER J.ALEXANDER,2 ANDWILLIAM S.

ROBINSON'*

Departmentof Medicine,Stanford University, Stanford, California94305,1and National Instituteof

Virology, Johannesburg, SouthAfrica2

Received for publication25April1979

ThePLC/PRF/5cell linederived fromahumanhepatomaproduces hepatitis

Bsurface antigen (HBsAg) in 22-nm particles of the same buoyant density as

thosefound in the serum of infectedpatients.TheHBsAg particlesfrom this cell

linewere labeled with[3S]methionine andpurified,and the polypeptideswere

compared by sodium dodecyl sulfate-polyacrylamide gel electrophoresis with those of serum-derivedparticles. The two major polypeptides of serum-derived

HBsAg particles (p20and p23)werefound in the same relative amountsin the

particles from the cell line. The three smallest of the five minor components

observed inHBsAg particlesfrom serum were present inparticlesfromthe cell

line. These polypeptides (p31, p36, and p43), as well as p20 and p23, were

precipitated with anti-HBs-containing serum. The two largest polypeptides of serum particles (p49 and p66) were not detected in particles from these cells. When the PLC/PRF/5 HBsAg particles were radiolabeled with tritiated sugars, p23,andnotp20,wasfoundtocontainradioactivity, indicating that thepattem

ofpolypeptide glycosylation is similartothat ofserumHBsAg. None of the other possible gene products of hepatitis B virus was detected in the

PLC/PRF/5-derivedHBsAg particles,inthecells,orinthe cellsupernatants.

HepatitisBvirus(HBV)infection inhumans

isaccompanied bythe appearance of viral anti-gens in liverand blood. Several viral forms in

the blood contain hepatitis B surface antigen (HBsAg).Oneforn,the Daneparticle,contains

an intemal antigen designated the hepatitis B coreantigen (HBcAg), DNA, andDNA

polym-eraseactivity and is consideredtobe the

com-plete virion (18).Athird antigen, hepatitisB e

antigen (HBeAg), canbe detected in the blood ofsomeinfectedpatients(13). Recent evidence indicates thatHBeAg is foundin acryptic form in the interior of the Dane particle (22). The

mostabundantHBsAgformsare22-nm

spheri-calparticlesandfilamentous formsconsisting of

protein, carbohydrate, andlipid, and these are considered to be defective or incomplete viral forms withoutHBcAgorDNA.Several

polypep-tides isolated from these particles have been

shown tocontain both group- and type-specific

HBsAgdeterminants (4,5, 19). Ininfectedliver,

HBsAgisdetected incellcytoplasmand atcell surfaces, and HBcAg is detected in nuclei by immunofluorescence (1, 6).

Numerous studies have demonstrated an as-sociation betweenpersistentHBV infection and

primaryliver cancer in some parts of theworld

(2, 8, 21).Itis not known whether the

relation-shipisacausalone.In1976,atissue culturecell

line designated PLC/PRF/5 was isolated from

a primary liver carcinoma of an African man

withpersistentHBVinfection. The cell linewas

subsequentlyshowntoproduce smallamounts

of HBsAg, apparently in the form of 22-nm

particles (12). The antigenic subtype has been

reportedtobe ad (12). The other antigens

asso-ciated withHBV, HBcAg and HBeAg,were not

detected in the cells. These findings indicate that thecells containatleastsomeHBVgenes.

Since cellsin culture havenotbeen

success-fully infected with HBV,PLC/PRF/5cells offer the firstopportunity tostudyHBsAg formation in tissueculture.

HerewedescribeexperimentsinwhichPLC/ PRF/5 cells were incubated with radiolabeled methionine orsugars, the HBsAg particles were

purifiedfromculturemedium, andthe

radioac-tive polypeptideswereanalyzedby sodium

do-decylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) andcompared withHBsAg parti-clespurifiedfrompatientserum. Thecells and culture medium werealsoassayed for Dane

par-ticles, HBcAg,and HBeAg.

MATERIALS AND METHODS Ceilculture. ThePLC/PRF/5cellswere grown in Dulbecco-modifiedEagle medium supplementedwith 10% fetalbovine serum.Thecellshavebeen passaged 796

on November 10, 2019 by guest

http://jvi.asm.org/

HBV POLYPEPTIDES PRODUCED BY HEPATOMA CELLS 797 weeklyand are nowatpassage 44 in our laboratory.

Production of HBsAg has remained at a constant level. Thecells are infected with mycoplasma.

Labelingof cells. At2 to 3days after reaching confluency in 10-cm culture dishes, PLC/PRF/5cells

were incubated with 50

lACi

of L-[3S]methionine (Amersham Corp., 800 to900 Ci/mmol) in 10 ml of methionine-free medium supplemented with 1% me-thionine-containing medium and 10% fetal bovine se-rum.After3days, the medium was replaced with fresh medium containing L-[3S]methionine and incubated foranother2days. Thepooledradioactivecell super-natants werecentrifuged at 12,000xgand4°C for 30 min to removecellular debris.

Similar cultureswereincubated withL-[6-3H]fucose (AmershamCorp., 17Ci/mmol) and

D-[6-3H]glucosa-mine(AmershamCorp.,20Ci/mmol) (20uCieach)in 10 ml ofDulbecco-modified Eagle medium for the sametimesdescribed for[3S]methionine.

DetectionofHBsAg. Solid-phase radioimmunoas-say(AusriaII,AbbottLaboratories)wasusedtodetect HBsAg. Samples from CsCl gradient fractions were diluted in TNE buffer(0.01MTris-hydrochloride,pH 7.5; 0.15 MNaCl;0.001MEDTA)atleasteightfold to avoid the effects of concentrated CsCl onthe assay.

Purification of HBsAg. Clarified cell superna-tants from PLC/PRF/5 cells were centrifuged at 25,000 rpm and4°Cfor24h inaSpinco SW27 rotor to pellet the HBsAgparticles. All but0.5 ml of the supernatantswereremovedbysuction; the remaining 0.5mlcontained concentrated fetal calfserumproteins andlooselypelleted HBsAg particles.Thepellets were suspended in the remaining supematant and were made up to 7 ml with TNE buffer. Solid CsCl was addedtogiveadensity of1.2g/ml, and the material wascentrifuged at 45,000 rpm and 4°C for 68 h in a

Spinco50Tirotor.Fractionswerecollectedfrom the bottomof the tube andassayedforradioactivity and HBsAg. Peak fractions ofHBsAgwerepooled,diluted to 7 ml with CsCl at 1.2 g/ml in TNE buffer, and

centrifugedasecondtime. TheHBsAg peakfractions were again pooled and dialyzed at 4°C overnight

against TNE buffer to remove CsCl. The material

(approximately2ml)wasthenlayeredontop ofa 33-mlpreformed 5 to20%sucrosedensity gradient con-tainingTNE buffer and1mgof bovineserumalbumin per mlandcentrifugedat25,000rpm and4°Cfor5h inaSpinco SW27rotor. HBsAgpeak fractionswere

dialyzedovernightagainstTNEbuffertoremovethe sucrose.TheHBsAgwasthen banded twicemorein CsCldensity gradientsasdescribed above. The frac-tions containing purified HBsAgparticles from the finalgradientwerethenpooledandkeptat4°Cuntil used.

HBsAgfromserum waspurifiedin thesame man-ner,but aLowryproteinassay(10)wasusedtodetect total protein in gradient fractions, and HBsAg was

detectedbycomplementfixation.

Antisera. Aguinea pig anti-HBs/adw-containing

serumwas agiftfromJohnGerinof the OakRidge

NationalLaboratory. Theserumwasabsorbed with rabbit liver powder and has a complement fixation titerof1:32.

The immunoglobulin G fraction of serum from HBsAg carrierswithhightitersofanti-HBcwas pre-pared by ammonium sulfate precipitation and

em-ployed in asolid-phase radioimmunoassayforHBcAg (16).

Sera from HBsAg carriers with high titers of anti-HBe were used to detect anti-HBeAg by double immuno-diffusion (detection to a serum dilution of 1:128) and by anenzyme-linked immunosorbent assay (14) used in thislaboratory (detection to adilutionof1:64,000). PAGE. Purified HBsAg particles in CsCl were di-luted 1.5-fold with TNE buffer, and 20

jig

each of bovineserumalbumin and cytochrome c were added. Atotal of5to 10volumes of ethyl alcohol (EtOH) was added, and the precipitated protein was pelleted by

centrifugation for 10min at 10,000 x g. The pellets

weresuspended insample buffer (0.07 M Tris-hydro-chloride, pH6.8; 11% glycerol; 0.0015% bromophenol blue; 3%SDS; 10% ,B-mercaptoethanol) and heated at 100°C for 3 min.

Thepolypeptides were separated by electrophoresis invertical slabgelsasdescribed by Laemmli (9). Gels of13or 8 to20%polyacrylamide were runusuallyat 19mA/13.5 cmfor 5 h. Gelswerefixed in 20% meth-anol-7% acetic acid, stained in 0.2% Coomassie bril-liant blue in 50% methanol-7% acetic acid, and de-stained in the fixative. To enhance detection of radio-activity, we performedfluorography as described by Bonner andLaskey(3)onallgels before drying with a slab gel dryerfrom Hoefer Scientific Instruments. The driedgels were exposed toXR-5X-rayfilm (East-man Kodak Co.) at -70°C and developed. Protein standards of known molecular weightswere fB-galac-tosidase, bovineserumalbumin, ovalbumin, aldolase,

chymotrypsinogenA, and cytochrome c.

Immunoprecpitation.Antigenswere

immunopre-cipitatedwithstaphylococcal protein A prepared by the method of Kessler (7). One volume of HBsAg

particleswas diluted with either 3 volumes of TNE bufferplus0.1%,8-mercaptoethanoland1mgof bovine serumalbumin per ml(TNEMEBSA)or2volumes of 1XRIPA buffer(0.01 M Tris, pH 7.2; 0.15 M NaCl; 1%

deoxycholate; 1% Triton X-100; 0.1% SDS) plus 1 volume of2x RIPA buffer. Serum containing

anti-HBs(251d/ml)wasadded,and the mixturewas

incu-bated for1hat37°C.Asuspensionofprotein A (10%,

vol/vol)wasthen added(50

p1/5

p1

ofantiserum), and the mixturewasincubated for30minat room temper-ature.Aftercentrifugation (10,000rpm inaBrinkman 3200 centrifuge for 10 min), the pellet was washed oncewith0.5ml ofTNEMEBSAor lx RIPAbuffer,

repelleted,andresuspendedin50pl ofsample buffer

containing6%SDS.Sampleswereheatedat100°C for 3 to 5min, theproteinAwasremovedby centrifuga-tion, andsampleswereanalyzed bySDS-PAGE.

RESULTS

Purification of HBsAg particles from PLC/PRF/5 cells. As culture fluids of PLC/

PRF/5 cells contained greater amounts of

HBsAg than did disrupted cells, these fluids

wereusedasthe sourceofHBsAgparticles for

purification.After removalof cellular debris

by

low-speed centrifugation, the HBsAg particles were pelletedin theultracentrifuge. The

parti-cles were then purified as described above by

fourequilibriumcentrifugationsinCsCl andone

VOL. 32,1979

on November 10, 2019 by guest

http://jvi.asm.org/

798 MARION ET AL.

rate sedimentation in a sucrose gradient. The

finalCsCl gradientservedmainlyto concentrate

the particles without the losses incurred when other concentratingmethods, such aspelleting orultrafiltration,wereused.

The firstand lastCsCl centrifugationsteps of

[35S]methionine-labeled HBsAg particle purifi-cation are shown in Fig. 1. Figure 1A shows

trichloroacetic acid-precipitable 35S andHBsAg

measured by solid-phase radioimmunoassay in the fractions of the firstgradient. The HBsAg particles accounted forsucha smallfraction of theacid-precipitable[35S]methioninethatthese

particles couldnotbedistinguishedas aseparate 35S component in this early purification step.

35S-labeled HBsAg particlesbecame detectable

onlyatthe third or,rarely, secondpurification

step. In the final gradient (Fig. 1B), the major 35Speakandthat ofHBsAgcoincided.Thepeak

fractions ofthe final gradientwere pooled and

stored at 4°C. Several purification steps for HBsAg were needed because of thelargeamount

ofnon-HBsAg 35S-labeled material released by these cells. 35S-labeled HBsAg represented about1% of thetotaltrichloroacetic acid-precip-itableradioactivityinthe clarified cell

superna-tants. SDS-PAGE analysis of the non-HBsAg 35S-labeled material revealed a continuous

dis-tribution ofpolypeptideswithmolecularweights

of 5,000(5K) to100K,and nopredominant com-ponents were visible.

HBsAg particles from serum were found at

the same buoyant density (1.23 g/ml) in CsCl

densitygradients.

Comparison of the polypeptides of [36S]-methionine-labeled HBsAg with HBsAg from human serum. The polypeptides of pu-rified [35S]methionine-labeled HBsAg particles

(subtype ad) fromPLC/PRF/5cellsand of

pu-rifiedHBsAg(subtype adw) from patient serum that was HBeAgpositivewereanalyzed by

elec-trophoresisin an 8 to20%SDS-polyacrylamide gradient gelasdescribed above.

The polypeptides of the serumHBsAg prep-aration, shownwithcytochrome c in Fig. 2, track b, appear to be similar to previous findings of

others (18).Twopolypeptidescorresponding to

molecular weights of 20 and 23K (p20 and p23)

constitute the majority ofthe Coomassie

bril-liantblue-staining material. p20was presentin

greater amounts thanp23. Several minor

Coo-massiebrilliant blue-stainingcomponents in

po-sitionscorrespondingtomolecularweights of 31, 36, 43,49,and66K(p31, p36, p43, p49, and p66) were also regularly observed. p66 comigrates with human serum albumin. An eighth

poly-peptide migrating at 97K was frequently

ob-served, althoughnotin thegelshown.

5 l0 15 20

m z U,

3

0 0

u 0

o

0

E

CL)

C)

5 10 15 20

FRACTION NUMBER

FIG. 1. Isopycnic centrifugation of 35S-labeled HBsAg particles in CsCl gradients. (A) The first equilibrium CsCldensitygradient centrifugation for 35S-labeled HBsAg purificationwascarriedoutasdescribedin the text,andfractionswerecollectedfromthe bottomof the tube. The trichloroaceticacid-precipitable

radioac-tivitywasdeterminedon1,il of each fraction, and theamountofHBsAg activitywasmeasuredineveryother

fraction by radioimmunoassay after 100-fold dilution of2.5gil.(B) The final equilibrium CsCl density gradient

centrifugationstep in 5S-labeled HBsAgparticle purificationwascarriedoutasdescribed in thetext, and fraction collection,radioactivity, andHBsAgassayswerecarriedoutasdescribed in (A), using 5 ,ul for the 35S

determinationand 5,uldiluted50-fold fortheantigenassay.

0

o

0x

E

C')

C/)

U-ur

i m z *I

--3

0 0

u

on November 10, 2019 by guest

http://jvi.asm.org/

[image:3.504.70.457.376.575.2]

HBV POLYPEPTIDES PRODUCED BY HEPATOMA CELLS 799

In Fig.2, tracka, thepositions of

[3S]methi-onine-labeledpolypeptidesfrom the PLC/PRF/

5 HBsAg preparation electrophoresed in the samegelcanbe observedafterautoradiography

tobesimilartothepositions ofthe polypeptides

of serum-derived HBsAg, with certain

excep-tions.p20 andp23are presentin thesame

gen-eral proportions.p31,p36, and p43 are also pres-ent, although the high background makes it

difficult todistinguish them. The two prepara-tionsdifferedbythepresence of a 27K (p27)

3S-polypeptide less heavily labeled than p20 and p23, butmoreinevidence than the otherminor componentsandnotdetected by Coomassie

bril-liant bluestaining of the serum HBsAg

polypep-tides. p27 was present in the same relative amountsintwootherpreparations of 3S-labeled HBsAgparticles.

CarbohydrateinHBsAgfromPLC/PRF/

5cells. One of themajor polypeptides of puri-fiedserumHBsAg (our p23) has been observed by others (18)tostain withperiodic acid-Schiff,

c

p66 _I 65.4 K

p49

p43

u.,.A.i ...~~~~~1.

p31

.*..w

_p~~~23

_p20

45 K 0

I

0

--- -39.5K _Ci

0

_~25 K x

E

-t)

[image:4.504.53.240.303.501.2]

_,1 .413.4K

FIG. 2. Comparison ofthepolypeptides of HBsAg particles from PLC/PRF/5 cells and from human

serum.HBsAg particles fromserumand

[uS]methi-onine-labeledparticles fromPLC/PRF/5 cellswere

purifiedandprecipitatedwith EtOHasdescribed in the text, and thepolypeptides ofeachwereseparated on an 8 to 20% SDS-polyacrylamide gradient gel.

Electrophoresis was carried out for 16 h at 40 V. (Track a) Polypeptides of 3S-labeledHBsAg parti-cles(8,000 cpm) after fluorography for3weeks; (track b) polypeptides ofserum HBsAg particles stained with Coomassiebrilliantblue; (track c)marker

pro-teinsrunin thesamegelandstained with Coomassie brilliant blue were bovine serum albumin (65.4K), ovalbumin(45K),aldolase(39.5K), chymotrypsinogen A(25K),andcytochromec(13.4K).

suggesting thatit was glycosylated. Minor com-ponents migrating at 27, 32 or 35, and 53.3K

have also appeared to contain carbohydrate in

previous studies (18).

Todetermine whether any of the polypeptides in the HBsAg preparation from PLC/PRF/5

cellswere glycosylated, we incubated the cells with both

L-[3H]fucose

and

D-[3H]glucosamine

andpurified the HBsAgparticles in the culture

supernatants as described above. The 3H and

HBsAgprofilesinthe finalCsCl gradient

coin-cidedasshown inFig.3.The3H-labeledHBsAg

particleswere EtOH precipitated andanalyzed

by SDS-PAGE alongside EtOH-precipitated

[35S]methionine-labeled HBsAg.Amajor

prob-lem with this analysiswastheverylowlevel of

radioactivity incorporatedintoHBsAgparticles

in these experiments. Although 2.5 mCi of

[3S]methionine

incubated with 108 cellsas de-scribed yielded a purified particle preparation containingapproximately150,000 cpm,thesame amountof tritium in thesugarsusedatthesame

specific activity yielded no -more than 37,000

cpminpurified particlesinanyexperimentand

required fluorography for detection by

autora-diography. Thepattem of3H-labeled

polypep-tidescanbeseenin track b ofFig.4and

com-pared with the 3S-labeled polypeptides shown intracka.Amajoramountof3Happearsinp23,

i1.4

1.3

a

1.2 zn

z

en

1.1

0

3 I

o

5 10 15 20 25

FRACTION NUMBER

FIG. 3. Final isopycnic banding ofHBsAg parti-cles labeledwithL-[6-3H]fucose andD

-[6-3H]gluco-samine. Theparticles were radiolabeled and

par-tially purified as described in the text. Fractions collectedfromthefinalCsCl density gradient were

analyzed for radioactivityandHBsAgasdescribed in thelegendtoFig.IB,exceptthat the3Hcontentof 10,ul ofeachfractionwasmeasured.

a

b

VOL. 32,1979

A

on November 10, 2019 by guest

http://jvi.asm.org/

[image:4.504.253.448.367.579.2]

800 MARION ET AL.

andnoneis detectedinp20.Althoughnotvisible

inFig.4,radioactivityis alsopresent inp3l and

p36,indicating that thepattemofglycosylation

inthePLC/PRF/5 HBsAgpolypeptidesis

sim-ilar tothat ofHBsAg from patientserum.

Immunospecificity of labeled

polypep-tides ofPLC/PRF/5 HBsAg. To determine which35S-labeledpolypeptides haveHBsAg

de-terminants, wedissociated 35S-labeled

polypep-tides ofpurifiedparticleswith RIPA bufferand

immunoprecipitatedthem withaguinea pig

se-rum containing anti-HBs/adw and protein A.

Forty to fifty percent of the radioactivity in various particle preparations was precipitated

and recovered bythis method, comparedwith

50 to70%recoverybyEtOHprecipitation.The

SDS-polyacrylamidegelpattern ofpolypeptides

precipitated by HBsAg-specific antiserum is

shown inFig. 5, trackb,andcomparedwith the

same amountofEtOH-precipitatedHBsAgfrom

thesamepreparationin tracka.Thehigh

back-ground along the track of EtOH-precipitated

polypeptides is not present with the

immuno-precipitated

material, and only p20, p23, p31,

p36, and p43aredetected in similarproportions

tothepolypeptidesofserumHBsAg.Thethird

major polypeptide found in EtOH-precipitated

HBsAg,p27, was notdetected, and thereare no

[image:5.504.289.434.62.206.2]

visible bands corresponding to the serum-de-rived HBsAg components at 49, 66, and 97K.

FIG. 4. Patternofglycosylationofthepolypeptides

of HBsAgparticles produced byPLC/PRF/5cells. HBsAgparticles labeled withL-[3H]fucose andD

-[3Hlglucosamine (track b) and [35S]methionine-la-beledparticles(track a)werepurified,EtOH precip-itated, andanalyzed bySDS-PAGE(13% gel,19mA for5h) asdescribed in thetext. Bothsamples con-tained12,000 cpm, and thefluorogramswereexposed for3weeks.

a

ww

..:.

[image:5.504.92.231.387.566.2]

.-b

..d:05::.

:

p43--p36' p31

p27

-p23 p20

FIG. 5. Immunoprecipitation of35S-labeled HBs-Agpolypeptides with guinea pig anti-HBs/adw. Pu-rified HBsAg particles labeled with [3S]methionine weredivided intotwoparts. Thefirstpart(track a) wasconcentrated by precipitation withEtOH, and thesecond(trackb)wasprecipitated in RIPAbuffer with the anti-HBs andproteinAasdescribedin the text.The polypeptideswereanalyzed bySDS-PAGE asdescribed in thelegendtoFig.4.

Although p23 from the immunoprecipitate

seems to bepresent in agreater concentration

thanp20 in thisexperiment, subsequentstudies have shownp20 tobe thepredominant immu-noprecipitatedpolypeptide.

To determine whetherp27 was a component

ofHBsAgparticleslackingHBsAgspecificityor acontaminantwhichcopurified with the HBsAg particles,wecarriedoutimmunoprecipitationin abuffer withoutdetergents. Comparisonof35S-. labeled HBsAg afterimmunoprecipitationin the

non-dissociating buffer and analysis on SDS-PAGE withEtOH-precipitated and

RIPA-pro-tein A-immunoprecipitated

[35S]methionine-la-beledantigen is shown in Fig.6. The polypep-tides immunoprecipitated

under.non-dissociat-ingconditions (trackb) appear tobe thesame

(i.e.,nop27)aswhenthey were

immunoprecip-itated afterdissociation with RIPA buffer(track

c). Therefore,p27isprobablynot a component

ofHBsAg particles, but a contaminant of the

purifiedHBsAgpreparations.

Todeterminewhether p27 or anyof the other

polypeptidescontainedHBcAgorHBeAg

spec-ificities,weincubatedequal amounts of purified

t[S]methionine-labeled

HBsAg particles with

purified human immunoglobulin G containing

anti-HBc and with human serum containing

anti-HBe. The latterserum alsocontained

anti-HBs. After precipitation with protein A, the

polypeptides were compared by SDS-PAGE

with polypeptides from equal amounts of the sameparticle preparation precipitated,

respec-tively,withEtOH, anti-HBs(Fig. 6, tracks a, b, andc), and normal human serum (track f). It

J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

HBV POLYPEPTIDES PRODUCED BY HEPATOMA CELLS 801

P43

p36

P31

p27

- p23-p20 -4

FIG. 6. Im Agwithanti humanserun

ticles were d

wereprecipit

(track b) ant non-dissocial proteinA in

anti-HBc an

anti-HBe pit buffer, and (ti A in RIPA I thepellet wa

inthetextan

canbeseen

precipitated human seru

tides precipi by theserun

e). Thus, th HBcAg-reac HBsAg par

ments.

Wedidn(

ucts of HB cells. NoHI

noassay ofs

fold by ultr by freeze-tl cellsperml

materials A

bent assay.

measurable fluids.

We havei

particlespri

derived fron

the polypep

from human labeledpoly the minorp

imrlmunopre

distinguisha

a b c d e f from

polypeptides

inserum-derived

particles

de-tected by Coomassie

brilliant

blue staining after SDS-PAGE. However, the

higher-molecular-weight minor polypeptides of serum HBsAg,

whichmigrate at49, 66,and 97K in oursystem, were not detectedinthePLC/PRF/5-produced HBsAg. One of the HBsAg polypeptides, p66, has been found

by

others to comigrate with humanalbumin, a protein foundinpreparations of

HBsAg

particles

purified

fromserum(20). p66

__m

has been the component

varying

the most in

we __ Z_ amount in different preparations (20). Its

ab-senceinthe hepatoma HBsAg raises the possi-bility that p66 has

actually

represented only human albumin in HBsAg preparations from

munoprecipitation of

35S-labeled

HBs-

serum in previous studies and that the HBsAg

-HBs, anti-HBc, anti-HBe, and normal reactivity of p66 found in some experiments was

n.Purifiedand'S-labeled

HBsAg

par- only due to a trailing of lower-molecular-weight

iivided into six equal fractions which antigen along the gel track. p27, a component

ated with the following: (track a)EtOH, fnd

in

ged

trag p27,aa

from

ti-HBs/adw serum and protein A in a found in

purified

HBsAg

preparations

from

ting buffer, (track c)

anti-HBs/adw

with PLC/PRF/5 cellsandnotinHBsAg from serum, a dissociating buffer

(RIPA),

(track d)

appeared

tobe a

contaminant,

since it failedto

d protein A in RIPA buffer, (track e) precipitate when intact HBsAg particles were

us anti-HBs with protein A in RIPA immunoprecipitated with anti-HBs. Thus, it did rack f) normal human serum andprotein not react with anti-HBs and was not a

compo-5uffer. Each was then centrifuged, and nent of HBsAg particles.

sanalyzedbySDS-PAGEasdescribed The presenceof five separate

"S-labeled

pol-z4in thelegendtoFig.4. ypeptides in an

immunoprecipitate

of purified

that nodiscerniblepolypeptide was particles, using anti-HBs-containing serumand Iwith anti-HBc (track d) or normal a

dissociating

buffer,

indicates that

multiple

iin (track f), and only the polypep-

HBsAg-reactive

polypeptides

are found in

itated by anti-HBs were precipitated HBsAgparticlesfromhepatomacells, aswellas nwith anti-HBe and anti-HBs (track in serum

particles.

This

multiplicity

is

probably

Lere was no evidence for HBeAg- or notdueto

degradation

by

serumproteases

(11),

ctive polypeptides in the purified unless such proteases are also present in these ticle preparations in these experi-

PLC/PRF/5 cells.

Peterson et al. (15) have proposed that the ot find any other possible gene prod- two main

polypeptides

in the serum

antigen

V associated with the PLC/PRF/5 particles

(our

p20

and

p23)

differ

only by

the BcAg was detectable by radioimmu-

carbohydrate

moiety

found in p23. This

post-supernatant fluids concentrated150- translational modification ofa

polypeptide

with acentrifugation or of cells disrupted HBsAg

reactivity

canaccount for the presence iawing at a concentration of 2 x107 of at least these two components of different 1.No HBeAg was found in the same apparent molecular weight and similar

antige-vith an enzyme-linked immunosor- nicity.Consistent with this conceptare our find-No DNA polymerase activity was ings that the hepatoma cells incorporate

3H-in the concentrated

supernatant

labeled sugars into p23 and not

p20

and that both are precipitated by anti-HBs. As the

rela-DISCUSSION tive

proportions

of

p20

andp23appeartobethe

same in HBsAg particles from both PLC/PRF/ radiolabeled and purified the HBsAg 5 cells and serum, these two components in this oducedby the PLC/PRF/5 cell line ratio may be a necessary feature of the particle n a human hepatoma and compared structure. How the remaining threeminor

com-tides with those found in particles ponents

(p31,

p36, and p43) may be related to iserum. The major

[3S]methionine-

p20,

to p23, and to each other remains to be

!peptides

(p20and p23) and threeof

determined.

Since others have found similar an-olypeptides(p31,p36, and

p43)

were tigenic specificity in several polypeptides iso-cipitated with anti-HBs and were in- lated from HBsAg particles (4, 19), different Lble in position and relative amount polypeptides must share some amino acid

se-VOL. 32,1979

on November 10, 2019 by guest

http://jvi.asm.org/

[image:6.504.61.252.58.213.2]

802 MARION

quences,andone ormoreof thelarger

polypep-tides maybe uncleaved precursors of the smaller ones.

Although a large proportion of HBV (Dane

particle) DNA base sequences appears to be present in PLC/PRF/5 cells (unpublished

ob-servations),only HBsAg,andnototherprobable

HBV gene products, e.g., HBcAg, HBeAg, or

Daneparticles, has been found in the cells.

ACKNOWLEDGMENTS

We thank Aleem Siddiqui for helpful discussions and George Scullard and Janet Kahle for assistance with the

antigen assays.

Thiswork wassupported by Public Health Service grant AI 13526 fromtheNational InstituteofAllergy and Infectious Diseases. P.L.M.istherecipient of postdoctoral fellowship5

F2CA06083fromthe NationalInstitutesofHealth. LITERATURE CITED

1. Barker, L. F., F. V.Chisari,P. P. McGrath,D.W. Dalgard, R. L. Kirschstein, J. D.Almeida,T. S. Edgington, D. G. Sharp,and M. R. Peterson. 1973.

Transmission of type B viralhepatitistochimpanzees. J. Infect. Dis.127:648-662.

2. Blumberg, B.S.,B.Larouze,W.T.London,B. Wer-ner,J. E.Hesser, I. Millman, G. Saimot, and M. Payet. 1975. The relationship ofinfection with the hepatitisBagent to primaryhepatic carcinoma. Am. J. Pathol.81:669-682.

3.Bonner, W.M.,andR.A.Laskey.1974.Afilmdetection method for tritium-labeledproteins andnucleicacids in polyacrylamidegels. Eur. J. Biochem.46:83-88. 4. Dreesman,G. R., R.Chairez, M. Suarez, F. B.

Hollin-ger,R. J.Courtney, and J. L. Melnick. 1975. Pro-duction ofantibody to individual polypeptidesderived

frompurified hepatitis B surface antigen. J. Virol. 16:

508-515.

5. Gold, J. W. N., J. W.Shih, R. H. Purcell, and J. L. Gerin.1976.Characterization ofantibodiestothe struc-turalpolypeptides of HBsAg: evidence for subtype-spe-cific determinants. J. Immunol.117:1404-1406. 6. Gudat, F., L.Bianchi,W.Sonnabend, G. Thiel, W.

Aenishaenslin, and G. A. Stalder. 1975.Pattern of

coreand surfaceexpressioninliver tissuereflects state

ofspecificimmuneresponse inhepatitis B. Lab. Invest. 32:1-9.

7. Kessler, S. W.1975.Rapid isolation of antigensfromcells with a Staphylococcal protein A-antibody adsorbent: parametersof the interaction ofantibody-antigen com-plexes with proteinA. J.Immunol.115:1617-1624.

8. Kew, M., R. Geddes, G. Macnab, andI. Bersohn. 1974. HepatitisBantigen and cirrhosis in Bantu patients with primary liver cancer. Cancer(Philadelphia) 34:539-541.

9. Laemmli, U. K. 1970. Cleavage of structural proteins duringtheassembly of the head of bacteriophage T4. Nature(London) 227:680-685.

10.Lowry,0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement withthe Folin

phenol reagent. J. Biol. Chem. 193:265-275.

11. MacKay, P.,and C. J. Burrell. 1976. Examination of the polypeptides of hepatitis B surface antigen. J. Gen. Virol.33:181-191.

12. Macnab, G. M., J. J.Alexander, G. Lecatsas, E. M. Bey,and J. M. Urbanowicz.1976.Hepatitis B surface antigenproduced by a human hepatoma cell line. Br. J. Cancer 34:509-515.

13. Magnius, L. O., A. Lindholm, P. Lundin, and S. Iwar-son. 1975. A new antigen-antibody system. Clinical significanceinlong-termcarriers of hepatitis B surface antigen. J.Am.Med. Assoc. 231:356-359.

14. Mathiesen, L. R., S. M. Feinstone, D. C. Wong, P. Skinhoej, and R. H. Purcell. 1978. Enzyme-linked immunosorbent assay for detection of hepatitis A anti-gen instool andantibody to hepatitis A antigen in sera: comparison with solid-phase radioimmunoassay,

im-muneelectronmicroscopy, and immune adherence he-magglutination assay. J. Clin. Microbiol. 7:184-193. 15. Peterson,D.L., I.M.Roberts, and G. N. Vyas. 1977.

Partial amino acidsequence of two major component polypeptides ofhepatitisBsurfaceantigen. Proc. Natl. Acad.Sci. U.S.A.74:1530-1534.

16. Purcell, R. H., J. L. Gerin, and J. B. Almeida. 1973/

74.Radioimmunoassay forthedetection of the core of

the Daneparticleandantibodyto it.Intervirology 2: 231-243.

17. Rao, K. R., andG. N. Vyas.1973.HepatitisBantigen activity in protein subunits produced by sonication. Nature(London)NewBiol.241:240-241.

18. Robinson, W.S. 1977. The genome of hepatitis virus.

Annu.Rev.Microbiol.31:357-377.

19. Shih, J. W., and J. L. Gerin. 1975. Immunochemistry of hepatitisBsurfaceantigen (HBsAg): preparation and characterization of antibodies totheconstituent poly-peptides. J. Immunol.115:634-639.

20. Shih, J. W., and J. L.Gerin.1977.Proteins ofhepatitis

Bsurfaceantigen:aminoacidcompositions of the major polypeptides. J. Virol.21:1219-1222.

21. Szmuness, W. 1978.Hepatocellularcarcinoma and the hepatitis B virus: evidence for a causal association. Prog. Med.Virol.24:40-69.

22. Takahashi, K., Y. Akahane,T.Gotanda, T. Mishiro, M.Imai, Y.Miyakawa, andM.Mayumi. 1979. Dem-onstration ofhepatitisBantigenin the coreof Dane particles. J. Immunol.122:275-279.

on November 10, 2019 by guest

http://jvi.asm.org/