0022-538X/80/12-0787/09$02.00/0 Vol.
Structural
Relationships Between the Surface Antigens of
Ground
Squirrel
Hepatitis Virus and Human Hepatitis B
Virus
WOLFRAM H.GERLICH,tMARK A. FEITELSON, PATRICIAL. MARION, AND WILLIAM S. ROBINSON*
Division of Infectious Diseases, Department of Medicine, Stanford University, Stanford, California 94305
Several physical, chemical, andserologicalproperties of surface antigen
parti-cles from ground squirrel hepatitis virus(GSHsAg) and human hepatitis B virus
(HBsAg) were compared. GSHsAg and HBsAg particles were purified from
positiveserabygel chromatographyandisopycnic centrifugation. Both antigens
consisted mainly of sphericalparticles with an average diameter ofapproximately
20 nm and a buoyant density in CsCl of approximately 1.19 g/ml. Their UV
absorptionspectraindicatedthe presence of more tryptophane than tyrosine and
theabsence of detectable nucleicacid.GSHsAgwas found tocontain two major
polypeptides of approximately 23,000 and 27,000 daltons, with electrophoretic
migration rates distinctly faster than those of the two major polypeptides of
HBsAg particles. After radiolabelingof purified antigen preparations with
Bolton-Hunterreagent,thetwomajorpolypeptidesofGSHsAgshowedalmostidentical
tryptic peptidemaps. The tryptic peptide map ofthe major polypeptide from
GSHsAgcontained 13of37spotsalsopresentinthemap ofthe majorHBsAg
polypeptide, and13of27spotsinthemapofthe major HBsAg polypeptide were
alsopresentin themapofthe majorGSHsAg polypeptide.This suggests
consid-erablesequencehomologybetween the major surface antigen polypeptidesof the
twoviruses.However,therewasonlyaweakserologicalcross-reactivity between
antigens of thetwoviruses.Usingananti-HBs-containingserumwitharelatively
strongcross-reactivity,GSHsAgwasfoundtoconsist ofatleasttwoantigenically
differentsubspecies.Themorestrongly cross-reactingformhadaslightlyhigher
buoyant densitythanthe otherantigenicform.
Whenhepatitis B virus(HBV)wasfirst
char-acterizeditwasfoundtohave severalinteresting
differences from viruses of therecognizedvirus
groups.Itsultrastructure andantigenicstructure
were considered unique (reviewed in 20). Its
small circular DNA hadan unusual structure,
includingthepresenceofalarge single-stranded regionof differentlengthindifferent molecules
whichwasrepairedbyavirion DNA
polymer-ase.Itcausedfrequent persistent infection with
completeviral forms (the42-nmDaneparticle) andincompleteviral forms(22-nmsphericaland filamentousparticles)continuouslyintheblood
inhighconcentrations,andpersistentinfection
wasassociated with chronichepatitisand
hepa-tocellular carcinoma. Both thecompleteandthe
incomplete viralforms inbloodappear to
con-tain on their surfaces a virus-specific antigen
now designated hepatitis B surface antigen
(HBsAg). HBsAg is a complex antigen. The
tPermanent address: Department of Medical Microbiol-ogy, HygieneInstitute, UniversityofG6ttingen,Gottingen,
FederalRepublicofGermany.
HBsAg'sfrom allseraappeartocontaina
group-specificdeterminanta aswellasdoryandw or r type-specific determinants (1, 15). Antigenic
variation of some of these determinants and
additional determinants suchas xand q have
also been described(4, 5).Severalpolypeptides
isolated frompurifiedpreparationsof22-nmand
filamentous HBsAg particles have been shown
tocontaintheHBsAgspecificities (7,10, 12,22).
Recently it has become apparent, however,
that HBV is not a unique virus. Viruses with
similarultrastructural, antigenic,molecular,and biologicalfeatures have been found in eastern
woodchucks (25), Beechey ground
squirrels
innorthem California(16),anddomestic ducksin
China (J. Summers, personal communication).
Thesevirusesprovideimportantanimal models
of HBVinfectioninhumans,and it is ofinterest
to determine how closely related they are.
Al-though the surfaceantigensof woodchuck
hep-atitis virus (29) and ground
squirrel
hepatitis
virus(GSHsAg) (16)clearlycross-reactin
sero-logicaltestswith HBsAg, itis not known how
787
on November 10, 2019 by guest
http://jvi.asm.org/
788 GERLICH ET AL.
similarthe surface antigens of these virusesare. Here we describephysical, chemical, and sero-logical properties of GSHsAg and compare it with HBsAg. The results clarify the degree of similarity of thesetwoantigens and indicate that the respective viruses differ serologically more thando HBV of differentHBsAg subtypes.
MATERLALS AND METHODS Isolation of 20-nm HBsAg and GSHsAg parti-cles.HBsAg (subtypeadw) concentrations inpatient
serahighin HBV-specific DNApolymerase activity (16) andGSHsAgingroundsquirrelserahigh in virion DNA polymerase activity (16) were determined by electroimmunodiffusion (11), using sheep antiserum (S-2/4) to HBsAg (anti-HBs). One-half to 2 ml of serumwithhighconcentrations ofHBsAgorGSHsAg
waschromatographedover aBio-Gel A-5M (Bio-Rad) column in TNEbuffer (0.01MTris-hydrochloride[pH 7.4],0.1 M NaCl, 5mMEDTA) as described in the legendtoFig.2.Fractionsstrongly positive for surface antigen in electroimmunodiffusion were pooled and concentratedapproximately fivefold with Ficoll (Phar-macia). Each 2.5 nil of concentrated material was made upto adensity of 1.30g/ml by addition of solid CsCl andlayeredover1.3mlofaCsClsolution with adensityof1.35g/mlinatube for theSpinco SW40
rotor. Two-milliliter volumes of CsCl solutions with densities of 1.27, 1.24, and1.18g/mlwereconsecutively layeredoverthesample, followedby0.8ml of 0.05 M phosphate buffer, pH 7.5. The surface antigen was centrifugedtoequilibriuminanSW40rotor at34,000 rpmfor40hat 10°C.Antigenwasagain located by electroimmunodiffusion, and positive fractions were pooledandstored inCsClat4°C.
Since theconcentration ofHBsAg inthe human serawas,onthe average, 10-fold less than concentra-tionsofGSHsAgin theground squirrelsera,HBsAg inserum wasconcentrated and partiallypurified by polyethyleneglycol6000precipitation in some exper-ixnentsbefore Bio-Gel A-5M chromatography.Briefly, 200ml ofHBsAg-positive serum was made up to 6.8% (vol/vol) with polyethylene glycol (from a 30% [wt/ wt]stock) and centrifuged for5minat5,000 rpmand 4°C.Theresultingsupernatantwasmade up to 12.5% with additional polyethylene glycoland centrifuged again. The 12.5% polyethylene glycolpellet was dis-solvedin4mlof TNE.
Radiolabeling and chemical modification of HBsAgandGSHsAg.Forradioinmmunoassay, pun-fiedpreparations of HBsAg and GSHsAg were radio-labeledby oxidative iodination, using Iodogen (Pierce ChemicalCo.; 8). Free 125Iwasremoved by chroma-tographyin Sepharose6B-Clwith 1%bovineserum albumin.Specific activities up to 10,uCi/,ugofprotein wereobtained. Forpeptide mapping, both HBsAg and GSHsAg preparations were dialyzed against 0.01 M sodium boratebuffer (pH 8.5)-0.1% sodium dodecyl-sulfate (SDS), lyophilized, andlabeled with Bolton-Hunter reagent (3) as described by Feitelson et al. (submittedforpublication). Briefly,0.2to 0.5
Ag
of p-hydroxyphenylpropionic acid N-hydroxysuccinimideester (Calbiochem) dissolved in 1 to2
pl
ofdioxane wasadded to50pl
of 0.5Msodiumphosphate buffer,pH 7.5, andrapidlyiodinatedbyadditionof0.5mCi of
125I
and100,ugofchloramine-T. The labeledester wasimmediately extracted into 100 1p of benzene, the aqueousphasewas removed, and theorganic phase
wastransferredtothelyophilized protein sample. La-beling andextraction of theester werecarriedoutin less than 40 s to prevent excessive hydrolysis in aqueousmedia. The benzenewasdriven offbyagentle streamofair,the vialwascooledto0°C,and5
A1
of 0.1Msodium boratebuffer, pH 8.5,at0°Cwasadded toeachsample.Afterincubationat0°Cfor1 h,the sampleswerestoredat4°Covernight.Thenextday 200,ugofp-hydroxyphenylpropionic acid N-hydroxy-succinimideester(indioxane)wasaddedtoeachsam-ple,and the reaction mixtureswereincubatedat0°C for3h.ThepHwasthenadjustedto9in eachmixture, and the sampleswerereductively methylated as de-scribedpreviously(18), with minor modifications(M. A.Feitelson and F. 0.Wettstein, submitted for pub-lication).
Alternatively,somepreparationswereradiolabeled byoxidativeiodination,usingchloramine-T(13). All samplespreparedforanalysis bySDS-polyacrylamide gelelectrophoresis (PAGE) werereduced in 10%
2-mercaptoethanoland2Murea at100°Cfor3minand alkylated with approximately two chemical equiva-lents of iodoacetamide. The labeled proteins were
separated from the other reaction components by chromatographyonBio-GelP-10 in0.05MNH4HCO3 and0.1%SDS, lyophilized,andredissolved ina solu-tionwith final concentrations of2 M urea, 3%SDS, and 0.001%bromophenolblue forSDS-PAGE.
SDS-PAGE and peptide mapping. Discontin-uous SDS-PAGE was carried out with 1-mm-thick slabgelscontaining5%stacking gels (with N,N-meth-ylenebisacrylamide cross-linker) and 12.5% running gels (withN,N'-diallyltartardiamide cross-linker).The buffer system and other conditions usedwereidentical tothoseofLaemmli (14).Electrophoresiswasusually carriedoutovernight (16 h)at aconstant currentof 10mAandstoppedwhen thedyereached the bottom of therunninggel.Gels with non-radiolabeledproteins
were stained with 2% Coomassie blue in 10% acetic acid and 25%isopropanol overnight, destained in the samemixture without Coomassieblue,and dried un-dervacuum.Forradiolabeledproteins,thegelswere fixedovernightin10% acetic acid and 25%isopropanol, dried undervacuum for 2 h, and autoradiographed using an intensifying screen and Kodak XR-5 film. Individualradiolabeled bandswerecutfrom thegel, dried underhigh vacuumovernight, and then rehy-drated with0.5ml ofasolutioncontaining250Agof trypsin (Calbiochem) and0.05MNH4HCO3atroom temperatureovernight. The supematantswere lyoph-ilized,and eachwasredigestedwith 100,ug oftrypsin in20
pl
of0.01Msodiumboratebuffer, pH 8.5. Two-dimensionalpeptide mappingwascarriedout on cel-lulose thin-layer plates (20 by 20 cm; Brinkmann) under a layer of n-heptane as described elsewhere (Feitelsonetal., submitted for publication). Electro-phoresiswascarriedoutinthe first dimension for 2.25 hat14°C,using 8% acetic acidand2% formicacid, pH 2,asthe running buffer. After electrophoresis, thin-layer chromatographywascarriedoutin n-butanol-water-ethyl acetate-acidic acid-pyridine (120:60:20:4: J. VIROL.on November 10, 2019 by guest
http://jvi.asm.org/
1, vol/vol), using thetop phase of themixture only. Thin-layer plateswerethen exposedtoKodak XR-5 film, usinganintensifyingscreen(26).
RESULTS
Antigenic heterogeneity of GSHsAg. GSHsAg was detected in the sera of infected
ground squirrels by its cross-reaction with anti-bodiesagainst HBsAg. For its qualitative deter-mination, a commercially available test kit for HBsAg (Ausria II, Abbott Laboratories) was
used,asdescribedearlier(16). Positivesera were
assayed quantitatively by electroimmunodiffu-sion, using asheep antiserun (S-2/4) directed
moststrongly against the group-specifica
deter-minant of HBsAg. Most other antisera against HBsAg tested didnotshow sufficient
cross-reac-tivitytopermitquantitation of GSHsAg. In
con-trast toHBsAg, which producedonestrongline with this antiserum (Fig. la, well 1), GSHsAg always producedtwoweakerprecipitationlines
of different intensities (Fig. la,wells3-8). This
result suggests antigenic heterogeneity in both GSHsAg and the cross-reacting antibodies. Al-though the relative amounts of both GSHsAg components were similar in different positive
ground squirrelsera, thefinding thatone
com-ponentformedastrongerprecipitationarccloser
totheoriginthanthe othersuggests that there
was more cross-reacting antibody directed
against the former subpopulation than against the latter.Heterogeneity in buoyant densitywas
also demonstratedbyCsCldensity gradient
cen-trifugation (Fig. lb).
Size of GSHsAg. GSHsAg-containing ground squirrelserumwaschromatographedon a columnof 4% agarose beads (Sepharose 4B). Totalproteinwasdetectedbyabsorbanceat 280
nmandGSHsAgwasdetectedby
electroimmu-nodiffusion. GSHsAgwaseluted inasingle 280-nmpeak characteristic ofserummacroglobulins
(Fig. 2). Incontrast, HBsAg preparations have been shown to elute with immunoglobulin M andbeta-lipoprotein, but slightly before the
a-macroglobulin peak (11). If the Stokes radius of humana-macroglobulin is 8.9 nm, and the
di-ameterandcomposition of this moleculearethe same in both species, GSHsAg would have an
averagediameter of 18to20nm.Larger
filamen-tousandspherical particles, detectedearlierby
electronmicroscopy (16), were apparently
bro-ken up oradsorbed tothe column, because all
fractionsbefore theantigen peakwerenegative even by the sensitive enzyme-linked
immuno-sorbents assay for GSHsAg (data not shown). Further, there was no detectable DNA
polym-coN ._E
E
sm
3
nc
30
20
10 '
FIG. 2. Sepharose 4B chromatography of ground squirrelserumcontaining GSHsAg. Serum (0.8 ml) wasloadedon an80-cm columncontaining135mlof Sepharose4B in 0.02 MTris-hydrochloride(pH 7.4)-0.13 M NaCI andeluted at4 ml/h with the same buffer. The solid line is a tracing of the 280-nm absorbanceof the eluate passing through a
3-mm-light path cuvette. Fractions were assayed for GSHsAg by electroimmunodiffusion asdescribedin thelegendtoFig. 1, and the lengths ofthe long (0) and short (0) precipitin arcs were measured. Bar shows thefractions pooled for further purification of GSHsAg.
a
11 2 3 4 5 6 7 8 15 16 17 18 19 20
FIG. 1. Electroimmunodiffusion of GSHsAgandHBsAg with sheep anti-HBs. Ten microliters ofeach
HBsAg-containing sampleor a1:10 dilutionofeachGSHsAg-containing samplewaselectrophoresedat20 mAfor8 h inaslab(25 by75by1.5mm) of1%agarose(Bio-Rad) containing0.034 M sodiumbarbital(pH 8.4) and 20%(vol/vol)antiserum. Thegelwasthenwashed, dried,andstained with 0.2% Coomassie blue in 7.5% acetic acid. (a) Well 1, humanserum with 35pgof HBsAg (adw)perml; well2, normalground squirrel serum;wells3 to8,AusriaII-positive ground squirrelsera.(b) Wells 15through 20, fractions15through 20, respectively, from the CsCl density gradient described in the legend to Fig. 3a after centrifugation to
equilibrium forpurification of GSHsAg.
20F 30F
40k
50h
70-80 Vt
40 50 60 70 80 90 100 110 120 30 140
36,
IF
on November 10, 2019 by guest
http://jvi.asm.org/
[image:3.496.245.439.256.343.2] [image:3.496.96.387.473.580.2]790 GERLICH ET AL.
erase activity in any of the fractions, even
though the starting materialwasvery strongly
positive.
Density ofpurifiedGSHsAg. GSHsAg
par-tiallypurifiedfrom0.8ml ofserumbygel
chro-matographywas adjustedto asolution density
of1.3g/mlwithCsCl andcentrifugedto
equilib-rium inaCsCldensity gradient. Antigenic
activ-itywasdetectedbyelectroimmunodiffusionina
prominent280-nmabsorption peakat adensity
of 1.183 g/ml (Fig. 3a). The low base line of
absorption around thispeak indicated that the
a
1.0 I 1.3
0.5 1.
o 0~~~~~~~~~~~~~~~~~~~~~~~e
0sL0
o
11.0
A
0.5 1.
2 6 10 14 18 22 26
[image:4.496.66.255.208.509.2]Fraction Number
FIG. 3.
Isopycnic
centrifugation
ofGSHsAg
(a)orHBsAg (b)in CsCl.After gelchromatography of0.8 mlofserumwith GSHsAgand2mlofserum with HBsAgasdescribed in thelegendtoFig. 2,thepooled antigen-containing fractions ofeach were
concen-tratedbyosmoticdehydrationwithdryFicollto4ml andadjustedto adensity of1.30g/mlwithCsClfor centrifugation in a CsCl density gradient as de-scribed in thetext.After centrifugation inaSpinco SW40rotorfor40 hat34,0X)rpmand10°C, fractions
were collectedfromthe bottomofthe tube. Absorb-anceat280nm(0)ina cuvettewitha1-cmlightpath
andsolutiondensity(Ax)byrefractometrywere deter-minedforeachfraction. Bars representthelengths oftheprecipitation arcsin the
electroimmunodiffu-sion assayforGSHsAgorHBsAg.
antigen was already highly purified after only
two purification steps. HBsAg from 2 ml of.
serum, purified in parallel by the same
proce-dure(Fig.3b),wasfoundina280-nmpeakat a
density of 1.190 g/ml. Although the starting
volume of humanserum was 2.5times greater
than the volume ofground squirrelserumused,
thedensitygradient profiles showed muchmore
GSHsAg than HBsAg. In another experiment
with serum from a secondpositive animal, an
evenlargeramountofGSHsAgwas extracted,
suggesting that the serum concentration of
GSHsAg in infected ground squirrels was
ap-proximately10times that ofHBsAginthe
high-titer human sera used.
Electroimmuno-diffusionanalysis of the three fractions making
uptheGSHsAg density gradient peak(fractions
17, 18,and19,Fig. 3a) revealedtwoprecipitation
lines in the two fractions of highest density
(fractions17and18, Fig. lb) andonly the faint
j.recipitation line in the antigen-containing
frac-tionof lowestdensity (fraction19,Fig. lb). This
resultsuggested thatGSHsAgwas notonly
an-tigenically heterogeneous but also
heteroge-neousinbuoyant density. Electron microscopy
of the GSHsAg peak fractions revealed both
sphericaland short filamentous forms withmean diameters of 18 to 20 nm. Although the
mor-phologyof theseparticleswassimilartothat of
HBsAgparticles, theapparentaveragediameter
was slightly less than that reported (2, 6) for
sphericalHBsAg particles(20to 22nm).
UV absorption spectrum of purified GSHsAg. The UV absorption spectrum of GSHsAg,like that ofHBsAg (11), had peaksat
280and 290.5nm as well as ashoulder at 275 nm (Fig. 4), which were indicative of a high tryptophane/tyrosineratio. Sincemostproteins have alow
tryptophane/tyrosine
ratio,and theabsorption spectrum of tyrosine obscures the
290.5-nm peak, the UV absorptionspectrum of
thispreparationsuggests arelativelyhigh degree
ofHBsAgpurity. Further, thespectrumshowed
no evidence of thepresence ofnucleic acid. If
the specificabsorbance ofGSHsAg isassumed tobethe same asthatofHBsAg, which is 4.2 at 280 nmfora 1-mg/ml solution, then the serum
from which the GSHsAg was purified in this
experiment would have contained
approxi-mately 300,ug ofantigen protein per ml. Sera
fromotheranimals had between40 and 300% of
that concentration when tested in
electroim-munodiffusion.
Serologicalcross-reactivityof antibodies
against HBsAg and GSHsAg in
radioim-munoassay. Antibodies directed against
HBsAgorGSHsAgdeterminants were detected
in sera bysolid-phase radioimmunoassay (21).
Briefly, dilutionsof test serawere added to U-J. VIROL.
on November 10, 2019 by guest
http://jvi.asm.org/
SURFACE ANTIGENS FROM HEPATITIS VIRUSES 791
0.3-,03
0.2-240 250 260 270 280 290 300 310 320 Wavelength nm
FIG. 4. UV absorption spectrum of purified GSHsAg. Fraction 19 from the CsCI density gradient shown in Fig. 3a was measured in a Zeiss PM2 spectrophotometer at 0.1-mm slit width.
shaped wells of microtiter plates coated with
GSHsAg or HBsAg, and antibody bindingwas
detected by subsequent binding of
"2'I-labeled
antigenof thesametypeusedto coatthewells.
Mouse and ground squirrel antisera against
GSHsAg and guinea pig and sheep antisera
against HBsAgweretitratedby thisassay,using
purified GSHsAgorHBsAgasthereagent(Fig.
5). All antiserafrom four mice immunizedwith
GSHsAg bound labeled GSHsAg tosaturation
(75% of theradioactivityadded)atlowdilutions,
and 50% was boundwhen antiserum dilutions
between 1:2,000 and 1:10,000 were used. With
HBsAgasreagent, twoof the four antiserawere
negative(<0.3%binding), and the othertwo sera
bound 0.5 and 2% of the added
"2'I-labeled
HBsAg at a dilution of 1:10 (Fig. 5a and b).
Using HBsAg,antibodycouldbe detected in the
two anti-HBs-containing sera diluted up to 1:
10,000. The sheep antiserum (S-2/4) boundup
to48%ofGSHsAgwhenused undiluted and 28%
at a 1:10 dilution (Fig. 5c), and the guinea pig
antiserum boundonly5.6% ofGSHsAgata1:10
dilution (Fig. 5d). It should be noted that the
sheep antiserum S-2/4 was directed mainly
against thea determinant ofHBsAgbecauseit
wasprepared by boostingwithan HBsAg
sub-type different from that used for the primary
inoculation. Six monoclonal antibodies to
HBsAg (agiftof LizJones)werealsotested with
GSHsAg. Onlyonebound 0.5% of theGSHsAg
atdilutions upto 1:100, whereasthe otherfive
antibodies showednodetectablebindinginthis
sensitiveassay.
SDS-PAGE analysis of GSHsAg- and
HBsAg-associated proteins. Purified
prepa-rationsofHBsAgandGSHsAgeachcontaining
50 ,ug ofprotein were reduced, alkylated, and
analyzed bySDS-PAGE,andtheproteinbands
wererevealedbyCoomassie bluestaining. The
HBsAg preparation yieldedtwomajor
polypep-tides with apparent sizes of 25,000 and 29,000
daltons (Fig. 6a, lane 2), similar to the findings
inprevious studies (7, 10, 12, 17, 22). Two major
polypeptides were also found in the GSHsAg
preparation (Fig. 6a, lane 1), and although their
apparent sizes were similar, they consistently
migrated fasterduringelectrophoresis(apparent
sizesof23,000and 27,000 daltons) than the
cor-responding majorHBsAgpolypeptides. Surface
antigen preparations radiolabeled by
chlora-mine-T (Fig. 6b)orBolton-Hunterreagent (data
notshown) revealedthe same two major
com-ponentsby autoradiography.
Several minor components migrating more
slowly in SDS-PAGE than the major
polypep-tides werealso foundin surface antigen
prepa-rationsfrom bothviruses when Coomassie blue
staining or either method of radiolabeling was
usedtodetectpolypeptides.Minor polypeptide
components have been repeatedly found in
HBsAg preparations previously (7, 10, 12, 17,
22).
Table1summarzesthe findings fromseveral
SDS-PAGE
experiments
in whichpolypeptidesweredetected by Coomassie blue stainingorby
autoradiography ofpreparations radioiodinated
by the iodogen procedure. None of the minor
Coomassie blue staining polypeptides with ap-parentmolecularweightsgreaterthan30,000in
GSHsAg preparations comigrated with any of
those from HBsAg. In contrast, the major and
-, -2 -3 -4 -, -2 -3 -4 0 -, -2 -3 -4 -5 -, -2 -3 -4 '0Log Dilution
FIG. 5. Cross-reactivity of antisera against GSHsAg andHBsAginsolid-phase radioimmunoas-say. A50-,lI portion ofmouseantiseraagainst puri-fied GSHsAg (a and b), sheep antiserum against HBsAgdeterminant a (c),orguinea pigantiserum againstHBsAgsubtypeadw(d), eachdiluted in 1% bovineserumalbumin (BSA), wasincubatedin mi-crotiterwells precoated with 10 ngofpurifiedHBsAg (0)orGSHsAg(0)for15hat4°C.Aftertwo wash-ings withBSA, 40,000 cpmof'2I-labeledHBsAg(2.5
,uCi/pg)or8,000cpmof"LI-labeledGSHsAg(0.5liCi/
pg) in50
j1
ofBSAwereincubatedfor24hat4°C. After threewashingsthe wellswerecounted. Approx-imately 75%ofeitherradiolabeledantigen wasthe maximumamountbound. Thebackgroundwithnor-malhumanseruminsteadofantiserumwasbetween 0.2and 0.4%(datanotshown).
VOL. 36, 1980
on November 10, 2019 by guest
http://jvi.asm.org/
[image:5.496.43.237.52.225.2]792 GERLICH ET AL.
a
FIG. 6. Polypeptides ofGSHsAgan ticle preparations.Purified preparatic (a, lane1) andHBsAg (a, lane2) we) SDS-PAGE and Coomassie blue $tai GSHsAg preparations from ground s
lane1)and 147(b, lane 2)wereradioio iodogen method andanalyzed by SI autoradiography.
TABLE 1. Apparent molecularweig)
polypeptidesseparatedby SDS-PAG.
GSHsAg and HBsAg preparc GSHsAg
Animalno. 83
Animalno. Coomassie 125i 147(1251)
blue
75 75
70 70
43 42 42
39 39 39
36 37
34 34 35
27a 27a 27a
23a 23a 23a
20 20
aMajorcomponents.
most of the minor '25I-labeled polypeptides of
) GSHsAg preparationsfromtheseraoftwo
dif-ferent ground squirrels (no. 83 and 147)
radio-2 labeled
by
theiodogen procedure appeared
tobe identical(Table 1). Therewas,however,
var-iabilityin the intensity andposition oftwo of
theminorcomponentsmigrating in the
34,000-to37,000-molecular-weight range(Table 1, Fig.
6b) in the two GSHsAg preparations.
Radiola-_ 5 beling of bothHBsAg and GSHsAg by the
chlor-7e anmine-T, iodogen, and Bolton-Hunter
proce-dures revealed several autoradiographic bands
42. inthe 14,000- to20,000-molecular-weight range
39 which were not detected on gels stained with
37 Coomassie blue (Fig. 6, Table 1).
35 Typtic peptide mapping of the major
polypeptides ofGSHsAg and HBsAg. The
relationship between the two major
polypep-tides ofGSHsAgwas examinedbytryptic
'25I-labeled peptide mappingof each afterlabeling
2 antigen preparations with Bolton-Hunter
re-agentandrecovering the appropriate
"25I-labeled
2 C polypeptide from a polyacrylamide gel after
SDS-PAGE. Figure 7a shows the map of
'25I-labeled peptides -detected by autoradiography
for the 23,000-dalton polypeptide (P-23) of
4
HBsAg
par-GSHsAg.
Thepeptide
mapof the27,000-dalton
)nsof GSHsAg polypeptide (P-27) of GSHsAgwasverysimilarreanalyzed by but not identical.Figure 7b shows acomposite
ning. Purified diagramof thepeptidemapsofthetwoGSHsAg
quirrels 83(b, polypeptides. The spot at position 1 was
de-dinatedby the tected in the mapof 27 and not in that of
P-)S-PAGE and 23. The spot at position 2 was very heavily labeled in P-23 and barely detectable in P-27.
itS _Xl-3)
ofThus,
the two major polypeptides ofGSHsAgts (xlO.).of are closely related in their primarysequences.
tonfprified
Figure 7c shows the12II-labeled
peptide map ofthe 25,000-dalton polypeptide (P-25) derived
fromHBsAg, andadiagram of thismapis shown
HBsAg in Fig. 7d. The tryptic
"2'1-labeled
peptides in(Coomassie themaps of the major polypeptides of the two
blue)
viruses
werecomparedinautoradiograms
of thetwo mapped separately and by digesting and
72 mappingamixture ofthetwo onasingle
thin-layer plate. Thirteen
125I-labeled
peptidesinthe63 map ofGSHsAg P-23 were found to be identical 56 to spots in the map of HBsAg P-25, and the
54 remainder ofthe spots weredifferentin the two
maps. Specifically, the shaded spots (13 of 37 spots) in the diagram of the peptide map of
GSHsAg P-23 (Fig. 7b) are identical to spots
present in the map of HBsAg P-25. The
un-31 shaded spots areunique toGSHsAgP-23. The
29a shadedspots (13of 27 spots) in thediagram of
the peptide map ofHBsAg P-25 (Fig. 7d) are
25" those also found in the peptide map, of GSHsAg
P-23. The unshaded spots were unique to HBsAg P-25. It can be concluded that, although
themajorpolypeptides of GSHsAg and HBsAg
J. VIROL.
t
on November 10, 2019 by guest
http://jvi.asm.org/
[image:6.496.74.241.76.333.2]a
b
.a
a
0.
c
d
I
2:
_ v
-rA
-FIG. 7. Tryptic'LI-labeledpeptidemaps of the majorpolypeptidesofHBsAgandGSHsAgradioiodinated by theBolton-Hunter method. (a) "LI-labeledpeptidemap of the23,000-daltonpolypeptide(P-23)ofGSHsAg. (b)Diagram of the compositemapof the samepolypeptideand the27,000-dalton polypeptide ofGSHsAg. The darkened spots are identical to spots in the map of the 25,000-daltonpolypeptide (P-25) ofHBsAg. (c) 1251. labeledpeptide map ofHBsAgP-25. (d)Diagram of the map of the same polypeptide, with the darkened spots correspondingtothose identical to spots in the map ofGSHsAgP-23.
possessdifferent mobilitiesonSDS-PAGE,they
share a considerable amount of structural
ho-mology.
DISCUSSION
By physical, chemical, and serological
tech-niques the similaritiesanddifferences between
thespherical surface antigen particles fromHBV
andground squirrel hepatitis viruswere
inves-tigated. The most numerous particles bearing
the surfaceantigens of both viruses hadsimilar
sizes andmorphologies, thesamebuoyant
den-sity,unusually high tryptophane/tyrosine ratios,
serologicalcross-reactivity, and a pair of major
structural polypeptides with considerable
ho-mology in primary structure. These common
featuressuggest that the molecular architecture
ofintact GSHsAg particles is probably similar
tothatofHBsAgparticles. Withoneantiserum
(S-2/4) from asheep immunized withHBsAg,
sufficient cross-reactivity was found to detect
andquantitate GSHsAgin
electroimmunodiffu-sionandbyenzymeimmunoassay.Surprisingly,
this antiserumdistinguished between two
anti-genicforms in GSHsAg, although no such het-erogeneityhasbeendescribed forHBsAg.Since
this antiserum was raised against HBsAg, it is
probable that HBsAg itself hasasimilar
heter-ogeneity, which maybe obscured by the
pres-enceofmanycommondeterminantsonpartially
heterogeneousparticles.Theunevendensity
dis-tribution of thestrongercross-reactive
compo-nents suggests that itsspecificity is related to
glycosylationormasking by lipid. That thereare
commondeterminantson HBsAgandGSHsAg
suggestsasimilarconfigurationofthe
polypep-tide chainsatthesurface of theparticles.It will
be of interest to determine whether chemical
reduction ofGSHsAgreduces itsantigenicityas
muchashas beenreportedforHBsAg (27).
Itis of interest thatGSHsAgparticlescontain
two major structural polypeptides similar but
notidenticalin sizetothemajorcomponents of
HBsAg,providingfurtherevidencefora
similar-ityin structure of the surfaceantigen particles
of thetwoviruses.Based upon amino acid
com-position (22, 28), limited
sequencing
data (19),andsimilarantigenicproperties (12, 22),thetwo
36, 1980
1- 5--" "t
3=- -..
-Iqmk
".Al&
Mts..
--Aim
on November 10, 2019 by guest
http://jvi.asm.org/
[image:7.496.100.385.74.358.2]794 GERLICH ET AL.
major structuralpolypeptides ofHBsAgappear
tohaveverysimilarprimarysequences.Periodic
acid-Schiffstaining (23) andradiolabeling (17)
experiments indicate that the major HBsAg
polypeptide with higher apparent molecular
weight, and not the smaller one, contains
car-bohydrate and this maycompletelyaccountfor
the difference in the mobilities of the two in
SDS-PAGE. Because of the small amounts of
antigenavailable, thepresence ofcarbohydrate
in the GSHsAg polypeptides has notyet been
studied, but the largerof the twomaycontain
carbohydratebyanalogywith theHBsAg
poly-peptides. The peptide maps ofthe two major
polypeptides ofGSHsAgwere almostidentical,
differing in only two spots (Fig. 7b). Although
differences in the amino acid sequence of the
two proteins could account for these small
changes in the peptide map, glycosylation (or
otherpost-translationalmodifications) ata
sin-gle site could also account for the changes. At
present,studiesarebeing directed toward
defin-ing thenatureofthesedifferences.
The molecular weight estimate by
SDS-PAGE for themajornon-glycosylated
polypep-tide ofHBsAg (25,000) isveryclosetothe
the-oretical molecular weight calculated from the
DNA sequence that appears to code for this
polypeptide (9). If the estimates of therelative
sizes of themajorpolypeptides ofGSHsAgand
HBsAg are accurate, an approximately 10%
smaller DNAcodingsequencewould berequired
to specify the 23,000-dalton GSHsAg
polypep-tide thanto specify the 25,000-dalton
non-gly-cosylatedcomponentofHBsAg.
Forpeptide mapping, the majorpolypeptides
from surface antigen particles of both viruses
wereradiolabeled in vitrobytheBolton-Hunter
procedure, which has been showntolabel
pro-teins more uniformlythan oxidative iodination
(Feitelson et al., submitted for publication).
Wheninterpretingstructuralrelationshipsfrom
peptide maps ofproteins labeled with
Bolton-Hunter reagentanddigested with trypsin,it is
importanttonotethat each-spotdoes not
nec-essarily correspond to a single tryptic peptide.
Dependinguponthe number oflysine residues
withfreeepsilonaminogroups in a givenpeptide
andtheextent to which these groups are
acyl-atedbytheBolton-Hunterreagent, thisreaction
introduces somechemicalheterogeneity into a
populationofotherwise identicalprotein
mole-cules. The relatively large number of spots,
therefore, isareflection ofthischemical
heter-ogeneity introduced by the acylationprocedure.
This is in contrast to the relatively few spots generated from direct iodination with chlora-mine-T oriodogenwhere thischemical
hetero-geneity isnotpresent. This isespecially
impor-tant when the differences involve spots which areminorbyautoradiography. Underidentical
reaction conditions the heterogeneity intro-ducedby this method isquitereproduciblesince
different radiolabeledpreparationsof aprotein
giveidenticalpeptide maps.
Thecomposite peptide mapsinFig. 7band d
define therelationship between the major struc-tural polypeptides of GSHsAg (Fig. 7a) and
HBsAg (Fig. 7c). InFig. 7b, about one-third of
thespotsinthemapof theGSHsAgpolypeptide
(P-23) werealso in thepeptide mapof the
ap-propriate HBsAg polypeptide (P-25). Nearly
half of the spotsin themapofHBsAg P-25were
also in themapofGSHsAg P-23 (Fig. 7d). That
common tryptic peptides signify considerable
sequence homology has recently been
demon-strated inthecomparison of simian virus40and BK virus T antigens (30). These different
pa-povavirus antigens share 7 of 20 or 21 tryptic
peptides, andyetby DNA basesequence
anal-ysis, they share 70% homology. This suggests
that the percent homology derived from the
peptide maps described here probably
repre-sents aminimumestimate. These resultsarein apparent contrast to the minimal serological
cross-reactivity of HBsAg andGSHsAg, which
was undetectable or less than 10-4 with the
antisera used here. Itappearspossible that the
selectivepressureof the hostimmune response
results in greater variability in the partof the
proteinmanifestingsurfaceantigenic reactivity
than in the overallsequence.Thecoreantigens
from ground squirrel hepatitis virus and HBV
havemore serological cross-reactivitythan the
surfaceantigens (W. Gerlichetal., unpublished
data). This observation confirms for thisgroup
ofviruses thegreatervariabilityof virionsurface
than internalantigens found formanyenveloped
viruses. Asthe molecularbiologyof theground
squirrel virus is further elucidated and compared
with that ofHBV, theextent towhich this
virus-host system can be used as a model for HBV
infection in humans will becomemoreapparent.
ACKNOWLEDGMENTS
This studywas supported by Public Health Service
re-search grant AI13526from theNational Institutes of Health. M.A.F. was supported by postdoctoral fellowship PF-1792
from theAmerican CancerSociety.W.H.G. received grant Ge 345/5from theDeutscheForschungsgemeinschaft.
LITERATURE CITED
1. Bancroft, W. H., F. K. Mundo, and P. K.Russel.1977.
Detection ofadditional determinants ofhepatitis B antigen.J.Immunol.109:842-848.
2. Bayer, M. E.,B.S.Blumberg, and B. Werner. 1968. Particles associatedwithAustralia antigen in the sera
ofpatients with leukemia, Down's syndrome and hep-atitis.Nature(London)218:1057-1059.
J. VIROL.
on November 10, 2019 by guest
http://jvi.asm.org/
3. Bolton, A. E., and W. M. Hunter. 1973. The labelling of proteinsto high specificradioactivities by conjugation to a l25I-containingacylating agent. Biochem. J. 133: 529-539.
4. Courouce, A. M., P. V.Holiand,J.Y. Muller, and J. P.Soulier. 1976. HB. antigensubtypes.Proceeding of theinternationalworkshop on HB. antigensubtypes. Bibl. Haematol. (Basel) 42:1-158.
5. Courouce-Pauty,A. M., and J. P.Soulier.1974. Fur-therdata onHB.antigen subtypes-geographical dis-tribution. Vox Sang. 27:533-549.
6. Dane, D. S., C. H. Cameron, and M. Briggs. 1970. Virus-likeparticles in serum of patients with Australia antigenassociatedhepatitis. Lancetii:695-698. 7. Dressman, G. R., R.Chairez,ILSuarez, F. B.
Hollin-ger, R. J. Courtney,and J. LMelnick. 1975. Pro-duction of antibody to individual polypeptidesderived frompurffiedhepatitisBantigen. J. Virol. 16:508-515. 8. Fraker,P.J., and J. C.Speck. 1978.Protein andcell membrane iodinations with asparinglysoluble chlor-amide, 1,3,4,6-tetrachloro-3a,6a-diphenyl-glycoluril. Biochem.Biophys.Res.Commun. 80:849-857. 9. Galibert,F., E.Mardart,F.Fitoussi,P.Tiollais,and
P.Charnay. 1979. Nucleotide sequence of hepatitis B virusgenome (subtype ayw) cloned in E. coli. Nature (London)281:646-650.
10.Gerin, J.L.1974.Structure ofhepatitis B antigen (HB Ag), p.215-224.In W. S. Robinson and C. F. Fox (ed.), Mechanisms of virusdisease. W. A. Benjamin, Menlo Park,Calif.
11.Gerlich,W., andR.Thomssen. 1975. Standardized de-tectionofhepatitisBsurfaceantigen: determination of its serum concentration inweight units per volume. Dev.Biol. Stand.30:78-86.
12.Gold,J. W.M.,J.W.Shih,R. H.PurcelLand J. L GeriD.1976.Characterizationof antibodiestothe
struc-turalpolypeptides ofHB.Agevidence for subtype-spe-cificdeterminants. J.Immunol. 117:1404-1406. 13.Greenwood,F.C.,W. M.Hunter,and J. S. Glover.
1963. Preparationof'3'I-labelledhuman growth hor-moneofhigh specificradioactivity.Biochem. J. 89:114-123.
14.Laemmli, U. K. 1970. Cleavageof structural proteins during theassemblyof the headofbacteriophageT4. Nature(London)227:680-685.
15.LeBouvier,G.L 1971. TheheterogeneityofAustralia antigen. J. Infect. Dis. 123:671-675.
16.Marion,P.L,LOshiro,D.C. Regnery,G. H. Scul-lard,and W.S. Robinson. 1980.Avirus inBeechey groundsquirrelswhich is relatedtohepatitisBvirus of man.Proc.Natl. Acad. Sci. U.S.A. 77:2941-2945.
17. Marion, P.L, F. H. Salazar, J. J. Alexander,andW. S. Robinson. 1979. Polypeptides of hepatitis B surface antigenproduced by a hepatoma cell line. J. Virol. 32: 796-802.
18. Means, G. E*, and R. E. Freeney. 1968. Reductive alkylation of amino groups in proteins.Biochemistry 7: 2192-2201.
19. Peterson, D. L, L. M.Roberts,and G. N.Vyas. 1977. Partial amino acid sequence of two major component polypeptides ofhepatitis B surface antigen. Proc. Natl. Acad. Sci. U.S.A. 74:1530-1534.
20. Robinson,W. S. 1977. The genome ofhepatitis B virus. Annu.Rev. Microbiol. 31:357-377.
21. Schober, A., R. M. Bigawar,and R.ThomRsen. 1975. Direct solid phase radioixmunoassay (dsp-RIA) for detection ofantibody to hepatitis B surface(HB,) an-tigen. Dev. Biol.Stand. 30:93-102.
22. Shih,J.W.,and J. LGerim. 1975.Immunochemistryof hepatitisBsurfaceantigen(HB.Ag):preparation and characterization of antibodies to theconstituent poly-peptides. J. Immunol. 115:634-639.
23. Shih,J.W., and J. L Gern. 1976. Proteins of hepatitis B surfaceantigen. J. Virol. 21:347-357.
24. Shih, J.W.,andJ. LGern. 1977. Proteins ofhepatitis B surfaceantigen: amino acid composition of themajor polypeptides.J.Virol. 21:1219-1222.
25. Summers, J.,J. M.Smolec, and R.Snyder.1978. A
virussimilar to human hepatitis B virus associated with hepatitis and hepatoma in woodchucks. Proc. Natl. Acad. Sci. U.S.A.75:4533-4537.
26. Swanstrom, R., andP.R.Shank.1978.X-ray intensi-fyingscreensgreatlyenhance thedetectionby
autora-diographyof theradioactiveisotopes'Pand "1I.Anal. Biochem.86:184-192.
27. Vyas, G. N., K. R. Rao, and A. B. Ibrahim. 1972.
Australia antigen (hepatitis B antigen): a conforma-tional antigendependentondisulfidebonds. Science 178:1300-1301.
28.Vyas,G.N.,E. W.Williams,G. G. B.Klaus,and H. E. Bond. 1972. Hepatitis-associated Australia antigen: protein,peptidesand aminoacidcompositionofpurified antigenwith tissue in determiningsensitivity ofthe hemagglutinationtest.J.Immunol. 108:1114-1118. 29. Werner,B.G.,J.M.Smolec, RI Snyder,and J.
Sum-mers.1979.Serologicrelationahipofwoodchuck hepa-titis virus and humanhepatitisBvirus. J.Virol.32: 314-322.
30.Yang,R.C.A.,A.Young,and R. Wu.1980.BKvirus DNA sequence codingfor the tand T antigens and evaluation of methods fordeterminingsequence ho-mology.J.Virol.34:416-430.
on November 10, 2019 by guest
http://jvi.asm.org/