Mutants of Sindbis virus. III. Host polypeptides present in purified HR and ts103 virus particles.

(1)

JOURNAL OF VIROLOGY, Nov.1978,P.466-474 0022-538X/78/0028-0466$02.00/0

Copyright X)1978 AmericanSociety for Microbiology

Vol.28, No. 2 Printed in U.S.A.

Mutants of

Sindbis

Virus

III. Host

Polypeptides

Present in

Purified

HR

and

tslO3

Virus

Particles

ELLEN G. STRAUSS

DivisionofBiology,CaliforniaInstitute of Technology, Pasadena, California 91125

Received for publication 5 June 1978

The amounts of host-encoded protein present in purified Sindbis virions of both the HR strain and of a mutant (ts103) whichmakes multicored particles wereexamined. Cells were labeled with [35S]methionine before infection and with [3H]methionine postinfection. Virions were purified by velocity sedimentation andisopycnic banding,and theirpolypeptideswere examinedbypolyacrylamide gels in a sodium dodecyl sulfate-containing discontinuous buffer system. Host prelabeled materialwasfoundprincipallyinasmallnumber of discrete polypep-tides in HRvirions,whichcontainedaslittleas0.2% host-encodedprotein. Virus-sized particles ofmutant tslO3 containedsignificantly more host material, and

multiploid particles from ts103 infection contained up to 12% host prelabeled

protein.

Sindbis virus and otheralphaviruses are the

simplest envelopedviruses(20).Theicosahedral nucleocapsid of the virus, which contains several hundred molecules of a single protein species surrounding the viralRNA,is assembled inthe cytoplasm of the infected cell. Upon diffusionto the cell surface, it interacts directly with the internalregions oftwovirus-encoded

glycopro-teins, El andPE2, present in theplasma mem-brane, leading to virus budding. During matu-rationaproteolytic cleavage of PE2to E2and E3occurs, and thematurevirionthus contains only three or four protein species, depending upon whetherE3 is retained in the virion. Other

envelopedviruses suchasrhabdoviruses,

ortho-myxoviruses, and paramyxoviruses mature by related mechanisms, but they contain a more complexnucleocapsidand alsocontain, in addi-tiontotheexternalglycoproteins,amatrix pro-tein. Thismatrixproteinisintimatelyassociated with the inner (cytoplasmic) surface of the

bi-layer and appears to interact with both the

capsid and the glycoproteins. In all cases the majorpolypeptide componentsof the virion are virusencoded (8, 14), whereas thelipid constit-uents arelargelyhostspecific, with little selec-tionby the virus (3, 9, 13, 15).

Early experiments on virus maturation sug-gested thathostcomponentswereexcludedfrom sitesof virusbudding (8, 14).Subsequent exper-imentsshowedthatthe virus-specific envelope polypeptidesarefree to diffuselaterally in the hostcell membranes and formpatches of modi-fied membranebyinteraction withintemal pro-teins ornucleocapsids (20). It wasthen tacitly assumed thatvirions contained only

virus-spe-cificpolypeptidesand that residual host material found in agiven virus preparation was the result ofinsufficientpurification(see review byLenard and Compans [11]). Recently, however, there hasbeen renewed interest in thepossibility that certaincellular polypeptides form integral mem-brane components of some enveloped viruses. Therefore, Sindbis virions were examined for the presence of host polypeptides which had been synthesizedbefore infection. The amount and type ofcellular material found in the HR strain andtslO3virionswerecompared. tslO3is amutantof Sindbis virus which forms aberrant

particles containing multiple nucleocapsids withinoneenvelope (19).

MATERIALS AND METHODS

Cell and virus. Preparation of monolayers of

chickenembryo fibroblasts has been described

previ-ously (16). Virus stocks usedwerethe HR strain of

Sindbis virus(2) and themutantts103 (19).

Labelingconditions.Rollerbottles (750 cm2) of

subconfluentchicken embryo fibroblasts were washed

for several hourswithEagle minimalessential medium

(6) containing Earlessalts, 2% dialyzed fetal calf

se-rum, and 1/3 to 1/5 the normal concentration of

essential amino acids. Then

[355]methionine

(460 to

500 Ci/mmol; Amersham/Searle) wasadded to 4 to

4.2mCi/roller bottle (depending ontheexperiment)

in 150 ml ofEagle medium containing 1/3x of all

essential amino acids except methionine (which was

at0.017mM or1/6X)and 3%dialyzed fetal calfserum.

In thismedium, themonolayers reached confluency in

48h,asopposedto 24hin mediumcontaining normal

aminoacid concentrations.

At 3 hbeforeinfection,the35S-containingmedium

wasreplacedwithEagle medium containing unlabeled

methionineat 0.2mM (2x)and2%dialyzed fetal calf

466

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

serum.Bottleswereinfected with10mlof

phosphate-buffered saline (5) containing 1% dialyzed fetal calf

serum, 1 ,igofactinomycin D per ml,and sufficient

HR or ts103 virus to give a multiplicity of5 to 10

PFU/cell. Mock-infected rollers received10ml of

dil-uentwithout virus.

At1hpostinfection the inoculumwasremoved and

replaced with 40 ml of low-salt medium (containing

NaCl at0.065M inplace of the normal concentration

of 0.115 M) per roller bottlewith 1/3X essential amino

acids exceptmethionine(whichwasat lxor2x)and

1 ,ug of actinomycin D per ml. After 1 h this was

replaced with mediumcontaining [3H]methionine (5

to10mCi/bottle),0.065MNaCl,1/3Xessential amino

acids exceptmethionine,3%dialyzedfetalcalf serum,

and atotalconcentrationof methionine of 1/2x to2x

depending upon theexperiment.

([Methyl-3H]methi-onine[5.5Ci/mmol]waspurchasedfromAmersham/

Searle, lyophilized to remove mercaptoethanol,

dis-solved in thelabeling medium, and sterilized by

filtra-tion.) Bottleswere incubated at30°C for 16 to 17 h

after HR virus infection and 20 to 21 h after tslO3

infection. The low-salt radioactive medium was

re-moved,and the virus was harvested by incubating the

monolayers with asmallvolume of medium containing

0.16M NaCl (16).

Purification of virus. The entire virus harvest

from oneroller bottle (-7ml)waslayered onto 32-ml

gradientsof15 to30%sucrosein 0.2MNaCl-0.05 M

Tris,pH 7.6-0.001 M EDTA-0.3% fetal calf serum and

sedimented inanSW27rotorfor90minat27,000 rpm

and4°C. Fractionswerecollectedby pumping,anda

smallportionof eachonewasassayed for 5Sand3H.

Fractions containing virion-sized tslO3, multiploid

tslO3 sedimenting 1.7x as rapidly, and 2.Ox

multi-ploidswerepooled separately.Comparable regions of thesucrosegradientscontaining HR and the harvest

fromamock-infection roller bottlewerealsopooled.

Pooledfractions were layered directly onto 11-ml

preparative isopycnic gradients containing 22 to44%

sucrose, 0.2M NaCl, 0.05 MTris (pH 7.6), 0.001 M

EDTA, and 0.3% fetal calfserumin 90%D20.

Centrif-ugation was for 12 h or more in an SW40 rotor at

32,000 rpm and4°C. Asmallamountof each of the

pools was also centrifuged in a parallel analytical isopycnic gradient with 32P-labeled HR virions as a

marker. In eithercasethegradientswerecollected by

pumping, and a small amount ofeach fraction was

counted inathree-channel Beckmanliquid

scintilla-tion spectrometer. Correcscintilla-tion of thecountsfor channel

overlapwasmadebycomputer.

Forquantitationof hostpolypeptides, pooled

frac-tions from the preparative isopycnic gradients were

diluted with5volumes of 0.2 M NaCl-0.05 MTris, pH

7.6-0.001M EDTA andfurtherpurified byasecond

cycle ofsucrose gradient velocity sedimentation

fol-lowedbyisopycnic separation.

Polyacrylamide gelelectrophoresis. Gels

con-taining 10%acrylamide, 0.26% bisacrylamide in 0.19M

Tris-chloride buffer(pH 8.6), and 0.1% sodium dodecyl

sulfate(SDS)werepreparedasdescribed byLaemmli

(10). Stacking gels contained 5%acrylamide, 0.067%

bisacrylamide,0.063MTris(pH6.8), and 0.1% SDS.

Sampleswerefirstdialyzed against buffer containing

0.05M Tris-chloride, pH 6.8, 1%SDS, 10% glycerol,

and 1to2%,8-mercaptoethanolandsubjectedto

elec-trophoresisuntil the bromophenol blue front was

ap-proximately0.5 cm from the bottom of the gel. Gels

werefrozen and sliced onaMickle gelslicer,and the

fractionswerecounted inascintillation cocktail

con-taining toluene,Liquifluor,andNCS.

Gelswerecalibrated for molecular weight by using

a series of standard proteins. Electrophoresis

condi-tionswerethesameasdescribed above, except that,

in the case of the unlabeledmarkers, the gels were

stained overnight with 0.05% Coomassie brilliant

blue-25% isopropyl alcohol-10% glacial acetic acid

and destained with several changes of

water-methanol-glacial acetic acid at 66:33:10. Gels were

scanned for absorbanceat580nm withagelscanning

attachment for an ISCO model UA-5 Absorbance

Monitor(Instrumentation SpecialtiesCo.).Standards

used in addition to the three virion proteins from

purified HR virus werephosphorylase A (molecular

weight, 94,000 [94K]), bovine serum albumin (66K),

pyruvatekinase(57K), heavy chain of gammaglobulin

(50K), actin (43K), D-amino acid oxidase (37K), and

soybeantrypsininhibitor(21K).

We have found that small changes inpH

signifi-cantly influence the rate ofmigration of the virion

polypeptidesrelativetothebromophenolblue marker

but havelittle effectonthemigrationof theSindbis

polypeptidesrelativeto oneanother.For thisreason allmeasurementsofmigrationrates weremade

rela-tivetothemigrationrateof Sindbiscapsidprotein.

Laemmli slab gels containing anexponential

gra-dient ofacrylamidefrom8 to20%wereusedtobetter

displaythelower-molecular-weightcomponents. Stan-dard proteinsused to calibrate this gelsystemwere

insulin (two chains of molecular weights 2.3K and

3.4K), a-bungarotoxin (8K), cytochromec(11.7K), ,8-lactoglobin (18.4K), and soybean trypsin inhibitor

(21K).

Immuneprecipitation.Asafinal method of

pu-rification of certain virusfractions, the materialwas

precipitated by atwo-stageimmuneprecipitation, in

which rabbit anti-Sindbis immunoglobulin G (IgG)

and goat anti-rabbitIgGwereused. The anti-Sindbis

IgGfractionwas agiftofC.Birdwellandwasabsorbed

three timesonmonolayersof uninfected chickencells

before use. Goat anti-rabbit IgG was obtained as a

lyophilized powderfrom Calbiochem. Fractions from

theisopycnicgradientswereprecipitatedat4°Cwith

anti-Sindbis serum for 4 h, followed by incubation

overnightat4°Cwith goatanti-rabbitserum.

Precip-itateswere collectedby centrifugation at27,000x g

for20minand dissolved in thesamebufferasused for

dialysisofgel samples.Bromophenolbluewasadded,

and thesampleswereboiled for 2 min.Electrophoresis,

slicing,andcountingwere asdescribed above.

RESULTS

Amountofprelabeled proteininHR and ts103. Several prelabeling experiments were

performed,withsomechangesinprocedural de-tail,

especially

inthespecificactivities of methi-onine used; with each experiment an attempt wasmade to

improve

theefficiency ofthe chase. Several preliminary experiments were done in

http://jvi.asm.org/

(3)

468 STRAUSS

which thepurification consisted of onlyasingle

cycle of velocityandisopycnic centrifugation.In theseexperiments,theamountof hostmaterial in lx tslO3 virions varied between 0.5 and 2.8%, dependingonthepreparation. Afinal

ex-perimentwitha two-cycle purificationreduced

the amount of host material to an apparent

minimum and yielded the quantitative data shown in Table 1.

Chicken embryo fibroblasts were labeled for

2dayswith [35S]methionineataspecific activity

of 1,700mCi/mmol,infectedwith eithertsl03or

HRormockinfected,andlabeledpostinfection

in low-salt medium containing [methyl-3H]me-thionine at 1,150 mCi/mmol; virus was

har-vested in normal salt. The entire virus harvest from each infection wassedimented in sucrose

velocitygradients,andmaterialssedimentingat

lx(the majorviruspeak-),1.7x(containing mul-tiploidparticles sedinenting1.7xasfastasHR

virions), and 2,OX werepooledfor further anal-ysis.

Eachof thepooledfractionswaslayeredonto both analytical and preparative isopycnic

gra-dients. Figure 1 shows profiles obtained with preparative isopycnic gradients of lx (virion-sized) and1.7x material forHR,tslO3,andmock

infections. The gradients from virus-infected material contained well-definedpeaksof3H la-bel, which was coincident with the32p marker HR virus on the parallel analytical gradients.

Themock-infected monolayer showed no such

peak and illustrated thatmost of the

35S

label associatedwith the virus from the infectedcells

was specific to virus infection. The location of the virus region shown on this gradient was

determined byreference tothe32p markerHR

on theanalytical gradient. In all three samples

J. VIROL.

from the 1.7x region there was a secondpeak

(band 2; usually split into twopoorly resolved peaks) near the density of 1.19 g/ml (at 5°C),

midway between the sample overlay and the Sindbispeakat1.21g/ml. From the densitywe

postulate that this fraction contained membrane fragments which sedimented at greater than 400S.

The viruspeaks from such isopycnic gradients

were diluted and resedimented in velocity

su-crose gradients, and the appropriate fractions wereagain selected and rebanded on isopycnic

gradients. Selected isopycnic gradients from suchatwo-cycle purification areshownin Fig.

2. On these gradients, little or noradioactivity,

either 35S prelabel or 3H postlabel, was found

outside the virus peak. In the case ofthe 1x

ts103the twohalves of thefinal isopycnic

gra-dientwerepooled separately (lxtslO3"dense"

andlxtslO3 "light").In thecaseof 1.7x tslO3,

the virus band was split according to density

after thefirstisopycnic gradient (see below). Figure 3showstheresult ofthesecond-cycle velocitygradientofthe 1.7x HR. Themajority of the 3H label (>60%) sedimented at lx, with only25% ofthe materialresedimentingat1.7x. This is characteristic ofHRpreparations; mul-tiploid particles are produced in very small

amounts and are often obscured by reversible

aggregates of lx particles sedimenting in the

sameregion. (The1.7xparticles from the second

cycle all sedimentat1.7x,however.) It is readily apparent from thisfigure thattherewas much more35Sassociatedwith themultiploid particles

than with the lx virions (also seebelow). The

1.7x particles and the lx particles from this velocity gradient were pooled separately and

sedimented to equilibrium on isopycnic

gra-TABLE 1. Quantitation of host polypeptides in Sindbis virus

%Host Total[3H]methi-

%s

El+ Total[3S]methi- %asEl+

Sample polypep- onine radioactiv- E2+C onineradioactiv-

Ey

+C

tidesb ity (cpm) o ity (cpm)o E2 + C

lxHR 0.24 385,600 94.3 9,690 68.0

lxHR from 1.7x HR pool 0.56 28,700 89.4 1,010 49.7

1.7xHR 4.01 5,300 77.0 770 24.1

lxts103dense 0.58 110,100 88.8 3,950 49.1

lxtslO3light 1.14 30,170 84.6 1,560 33.5

1.7x tslO3 intermediate den- 3.80 9,050 81.5 1,320 24.5

sity

1.7xts103light 12.31 2,650 75.5 1,070 18.2

aPreparationandpurification of these samplesisdescribed in thetext.

bHostpolypeptidesarecomputed as

follows:

% host polypeptides=(counts of 35S not in

El

+ E2 +C/counts

of3HinEl+E2+C)x(specific activityof3Hin counts per minute per mole/specificactivity of 3S in counts

per minute permole). This formulaexpresses the 35S present corrected for the amountreutilizedin virion

proteins but uncorrected foranyother turnover andnormalizedtothe total amount of virion polypeptide on

thegel. Itgivesamassratioifone assumesthatthe average methionine content of the host polypeptides is

similartothatof the virion polypeptides.

c35Scountshave been corrected fordecay.

http://jvi.asm.org/

[image:3.499.61.456.481.588.2]

(4)

I0

x 0

N

I V

re

12 1

N

0 x 8

-0 N N CI) 4 )

CL u

20 40 60 20 40 60

FRACTION NUMBER

FIG. 1. Preparative isopycnic gradients of Ix and 1.7x regions from HR, ts103, and mock infections.

Samples (20 ,u) ofeach fraction were counted in Aquasol to determine the distribution of prelabel and

postlabel. (A)Ix HR; (B) Ix tsl03; (C) Ix mock;(D) 1.7xHR; (E) 1.7x tsl03; (F) 1.7xmock. ( )

[35S]-methionineprelabel; (---)[3H]methioninepostlabel. Sedimentationwasfromrighttoleft.

dients. They are referred to below as lx HR from 1.7x pool and1.7x HR,respectively.

The 1.7x tslO3 particles show considerable density heterogeneity on isopycnic gradients

(19).Therefore, after the firstisopycnic

centrif-ugation, three regions of the virus band were pooled:the denseside, the central(intermediate) region, and the light side. Each of these was diluted and resedimented separately on 15 to

30% sucrose gradients, where virtually all the

radioactivity resedimented at 1.7x. The 1.7x material from the three second-cycle velocity gradientswere pooledseparatelyandrebanded onthreeisopycnicgradients,where each formed asingleband. Thesepreparations are referred to

as 1.7x tslO3 dense, 1.7x ts103 intermediate,

and1.7x ts103light,respectively.

Selected fractions from these gradients were analyzed by acrylamide gel electrophoresis for radiolabelinvirion andhostpolypeptides. Rep-resentative gels are shown in Fig. 4, and the resultsof one suchexperiment are summarized in Table 1. Several conclusions can be drawn from these data. The first is that the amount of hostpolypeptide foundin HR virions was strik-ingly small, as little as 0.24%. The amount of prelabelin lx ts103 was severalfoldhigher, on theotherhand,anddependedonthedensityof the particle. Multiploid particles, whether HR or

tslO3,

contained more host material; for

http://jvi.asm.org/

[image:4.499.112.394.68.460.2]

(5)

A.

density.

These results

suggest

thatabnormalities

30-xHR in

budding

lead to the inclusion of more host

material in the virionsproduced.

Host polypeptides in purified virus par-ticles and in other fractions. Figure 4shows

20_ three of the profiles from which the data in

Table 1 were obtained andillustratesthat

puri-__5 fiedvirions from HR andts103 differ in

compo-sition of prelabeled material. In these

prepara-0_ tions, a

small

number of discrete species of

pre-labeled host polypeptides was identified. The same host polypeptides appeared reproducibly

O

0 / in Oall

prelabeled preparations

examined and

0 l have been numbered in order of decreasing

mo-B. lecular

weight.

lx HRvirions

(Fig.

4A) contain

i-

*t

x1

1s

103 _2

polpeptides

called 1 (molecular weight, 68K), 3 (41K), 7 (22K), 8 (21K), and F (polypeptides of

4 - <20K which

migrated

with the solvent

front).

o

'O

(For

reference,

the

apparent

molecular

weights

l x of El, E2, and C in this system are 54K, 48K,

IN _ I _ and

32K,

respectively.)

Prelabel

35S

was also

o 0 found in E1, E2, and C, presumably due to

(M_2 inefficiencies in the chase or from breakdown

roI andreutilization of the label afterinfection, but

E'

IK there was very

little 35S

in other

regions

of the

L\Egel. lx tslO3 (Fig. 4B) contained the host

poly-0 o peptides found in lx HR as well as peakscalled

C.L

1.7 xts

103

LIGHTj

4 1~~~~.7xHR

l

12

.10~~~~~~~~~~~~~~~~~~~2

20-~~~~~~

021 ,05

A: 15 8 X

x

-0

[image:5.499.64.246.49.516.2]

FRACTION NUMBER5

FIG. 2. Second-cycle preparative isopycnic gra-L X /

dients of(A) Ix HR, (B) Ix ts103, and (C)1.7x ts103_

light. The origin of the samples isdescribed in the 20 40

text. ( ) [35S]methionineprelabel; (- --)[3H]me- FAT NNME

thionineFpostlabel.

NU

FIG. 3. Second cycle ofpreparative velocity

sedi-ample,thelight 1.7x tslO3 containedup to 12% mentationof 1. 7x HR. 1. 7x HR material was purified host protein, 10 to 20 times as much as was byvelocity sedimentation followed by isopycnic

cen-found in lx tslO3. Despite these differences trifugation,and thevirusbandwasdiluted and

cen-earlierexperimentshave shown that thesemul- trifuged on a15 to30%sucrosegradient. Portions of

tiploids are fully infectious. The 1.7x HR are eachsamplewerecountedfor

'S

and3H. Samples14

also

nlot beenasr weyocharacterized

naturaly

occurring multiploids,

but have through 23 were

pooled,

banded on an

isopycnic

mu 9 Ne

thnave

gradient, and designated 1.7x HR. Samples 30

not been as well characterized (19). Note that through 36 were treated similarly and designatedlx

1.7x HRweremade inverysmallamounts and HRfrom 1.7x HR pool. (-)

[3SJmethionine

pre-contained levels ofprelabeled material compa- label; (---)

[3H]methionine

postlabel. Sedimenta-rabletothatfound in1.7xts103 ofintermediate tion wasfrom right to left.

470 STRAUSS J. VIROL.

--- ---I - - ---1

http://jvi.asm.org/

[image:5.499.249.455.84.549.2] [image:5.499.264.455.355.549.2]

(6)

polypeptides and in thesmall material migrating atthe front.

Selected fractions from first-cycle isopycnic gradients (such asthose shown in Fig. 1) were

also examined as acheck onvirus purification

and to determine the distribution of polypep-tidesthroughout thegradients. Figure 5shows gel profiles ofthe band 2 material from mock-infected and tslO3-mock-infected preparations. One predominant peak (6, 25Kto26K)wasfound in

addition toF. Polypeptide 6 (aswell asF) was

labeled with both 35S and 3H in both cases,

showing that it is a species which turns over

rapidly, is synthesized after actinomycin treat-ment of the cells, and is independent of virus infection. Band 2 fromtslO3 infection contained the Sindbispolypeptidesinadditionto6and F. The absence ofpolypeptide 6from the various purified virion preparations in Fig. 4

demon-To

stratesthatthe viruspreparationswerenot

de-x tectably contaminated with band 2 material.

The

origin

of band 2 is

unclear,

although

its isopycnicdensity of1.18to1.19indicates that it

a.

o contains membrane

fragments.

Whatever its

or-igin, it is curious that it contains suchalimited

distributionofpolypeptides.

It shouldalso be pointed outthat host poly-peptides 1,5,and 7 couldnotbedetected inany

fraction examined from uninfected cells (whether the virus region, band 2,orthe soluble

material at the top of the isopycnic gradient). Thesepolypeptidesappeartobespecifically

as-sociated with virusparticlesinsomeway.

Thelow-molecular-weight material migrating with thesolventfrontonthesegels (F) has been

showntobecomposedofpolypeptides. Itis

non-dialyzable, precipitable in hot trichloroacetic acid, and insoluble in chloroforn-methanol. When Fwasdisplayedon8to20% Laemmli slab

SLICE NUMBER

FIG. 4. Acrylamide gel electrophoresis of virus

particles purified throughtwocycles ofsedimentation

followed by isopycnic density banding.

Electropho-resis wasfrom left toright; Fcomigrated with the

bromophenol blue marker. (A) Ix HR.Same

prepa-rationasinTable1,line 1.(B)Ix tsl03, dense side

ofviruspeak.SamepreparationasinTable1,line 4. (C) 1.7x tsl03, lightsideofviruspeak.Same

prepa-ration as in Table1, line 7. (-) [3S]methionine

prelabel;(---)[3H]methioninepostlabeL

4 (33K) and5 (29K).Allof the host peakswere more prominent in the tslO3 material, and in

additiontherewasmoreofabackgroundsmear

of35Sthroughout the gel. Thelight 1.7x ts103 particles (Fig. 4C) contained 35S primarily in peptides1andF, plusageneralsmear.InFig.4,

the 3Hpostlabel wasfoundonlyin theSindbis

C~4

I

24

10

0

20 '0

x

U)

10E

0

40 80 120 40 80 120

[image:6.499.58.225.51.490.2]

SLICE NUMBER

FIG. 5. Polyacrylamidegel electrophoresis ofband

2fractions from isopycnic gradients. Fcomigrated

with the bromophenol blue marker. (A) 1.7x mock, band 2(pool of fractions41,42, 43fromgradientin

Fig. IF). (B)1.7x tsW03,band 2(pool offractions 37,

38, 39fromgradientinFig.IE). (-)

[3S]methio-nineprelabel; (---)[3H]methionine postlabel.

I0

x

I

A. B. F

1.7xMock Bond2 IxtsIO3 Bond2 F

4 ~~~6

6 E

'E2

http://jvi.asm.org/

[image:6.499.254.448.482.596.2]

(7)

472 STRAUSS

gels with appropriate low-molecular-weight markers, both theprelabelandpostlabelshowed a small number of discretebands, betweentwo andfourdependingonthesample, with molec-ular weights between10Kand 20K.F was pres-ent in the most highly purified virus prepara-tions examined (Fig. 4A and B) and was not removed by immunoprecipitation with anti-Sindbis antibody, implying that some orall of these smallpolypeptides may be integral

com-ponents of the virion. On the other hand, it is possible that these low-molecular-weight

com-ponents are in part generated by proteolysis

occurringduring purification. DISCUSSION

Hostpolypeptides in purified Sindbis virions have been examinedtodetermine the absolute level of host proteinpresent invirusparticles,to ascertain thenatureanddegree of heterogeneity of hostproteinfound invirions,andtodetermine whether the amount or type of host material dependsupontheefficiencyofbudding.First of all, the host materialfound in HR virionswas a verysmall percentage of theprotein of the Sind-bisvirion,althoughit isalwaysdifficulttoknow whether residual host label invirions is dueto truevirioncomponentsor toresidual nonvirion contaminants. Pfefferkorn and Clifford (14) placed an upper limit of 5% host material in Sindbisvirus,butimproved techniquesandthe

availabilityofhigherspecificactivitylabels have reduced this limitconsiderably. Thepresent re-sults indicate that during a standard infection withthevirus, theamountof host protein found inpurified virionswas aslow as0.2 to 0.5% of the totalprotein,whichtranslatestoonlytwo to four host polypeptides per particle. What pro-portion of theprotein ofthe infected cell

plas-malemma isrepresented byEl and PE2 is un-known, but it isprobablyless than50%,

consid-eringthelengthof theinfection and the relative ratesofproteinsynthesis in infected and unin-fectedcells. Thus, anarea of membrane which contains morethan 400 host proteins and less than 400virus-specificmoleculesatthe onset of budding produces a mature virion containing 800viruspolypeptidesand not more than2to 4 host polypeptides in its envelope. This rather remarkableselectivityin anessentiallyfluid sys-temisprobably provided bytheinteraction be-tweenthenucleocapsidand the transmembran-ousportionsoftheglycoproteins (20).

In comparing HR virus with tslO3, it was found that lx ts103 contained more prelabel materialthan didthecomparable HR fraction. In addition to the discrete species of polypep-tides seen in HR particles, tslO3 particles also

containedmore of a smearofmanypolypeptides, suggestinganonspecific inclusionofhost mate-rial.Larger tslO3 virionparticles, the multiploids which contain four to eightnucleocapsids in a single envelope, contained significantly more host polypeptides than did the lx virionswith asinglecore,and thesmearofbackground poly-peptides was quite pronounced. Furthermore, thetslO3multiploids of lowestdensity contained up to 12% 35S prelabel; these particles have a higherproportion ofenvelope tocore, in other words, loosely fitting membranes. It is also of notethatmultiploids of HR containmore host protein than does the standard HRvirion; this isreadilyapparentinFig. 3and Table1.

There are three possible origins for the host material whichwasfound associated with virus. The first is that the 35S material consists of residualcontaminants whichareadsorbedonto the surface of the virions. Such a mechanism wouldnotexplain the differencesseenbetween lxHR and1.7xHRorbetween lxHR andlx tslO3, which should have very similar surface properties foradsorption. The second possibility is that the35S-labeledprotein is duetointrinsic contamination, i.e.tomaterial whichcopurifies withparticular virus classes. Thisseemshighly unlikely because of the absence ofcomparable prelabeled material in mock-infected prepara-tions and because of thedisparate levels of host polypeptides present in the various virus sub-classes. Ipersonallyfavor thethirdalternative, i.e., thatsomeof theprelabeled material (above the0.2 to0.5%level whichmaybeameasureof thepurity of the preparations) represents inte-gral host membrane componentswhich are in-serted intheenvelope of the virion.

All of the results support the general model forSindbis maturationpreviouslypresented (20) as well as our explanation of the defect inthe mutanttslO3 (19). The tslO3 defect isprobably inthenucleocapsidprotein;wehave shown that themajorityofnucleocapsidssynthesized early ininfectionare aberrant in their sedimentation behavior, but thatmostofthese defectivecores do not bud and are not recovered in released virions.Wehave postulatedthatallof the man-ifestations of the tslO3 mutation can be ex-plained if the ts103 nucleocapsidcanonly inter-act weakly with the glycoproteins in the cell surface. Whereas tslO3 veryseldombudssingly and usually buds as multiploid particles, it is easytoseehowhost materialcouldbeincluded adventitiously.

Themosthighly purifiedpreparationsofHR virionscontain theequivalentofonly two to five molecules of hostpolypeptides pervirion, and thesearepresentasseveralpolypeptide species.

J. VIROL.

http://jvi.asm.org/

(8)

Incontrast tothis, actin has been foundtobea

regular component of both paramyxoviruses, suchasSendaivirus(17),andtheavian, monkey,

and murine RNAtumorviruses(21).Comprising

upto8% of the proteinmass ofthe

Rous-asso-ciated virus-2 virion, this large amount of a

cellularprotein incorporatedisthoughttoresult fromabudding mechanism which requires the

juxtapositionofmicrofilamentstotheemerging virion. Actin,whichshould migrateatthe posi-tionof hostpolypeptide 3, could be present in only insignificant quantities in Sindbis virions. Inaddition, there have been numerousreports whichsuggestan associationof polypeptides of

the major histocompatibility complexwith

en-veloped virus particles. On the one hand, H-2

antigenshavebeenreportedtobeintegral

com-ponentsofvesicular stomatitis virionsgrownin

mouse L-cells (7). On the other hand, studies

with several viruses, including lymphocytic cho-riomeningitis virus and Rauscherleukemia

vi-rus,haveshown thatsensitizedcytotoxic

T-lym-phocytes are most efficient in killing

virus-in-fectedcellswhen the target and thekillercells

sharethe samehaplotypeofthe H-2

histocom-patibility locus, suggesting that the cell surface antigen recognized might be a hybrid of viral

andcellularcomponents (4, 18). Sindbis virions havenotbeen testeddirectly for suchantigens, but they could be present only in very small

quantitiesinthesepreparations.

There have also been numerous reports of

host cell enzymatic activities associated with purified viruses, but whether these represent polypeptides necessaryfor production of

infec-tiousvirions, adventitiouscomponentswhichare

engulfedorincluded in theenvelopeatthe time of budding, or stubborn contaminants which

copurifywith virions is unclear(1). It has been suggested that the host-encoded ribosomes which are found in the arenaviruses belong to

the second category and are not required for virusreplication (12).

Finally, there is supporting qualitative evi-dence indicating that alphaviruses are much more selective in their envelope components

thanother virusgroups.Arecentseries of phe-notypic mixingexperimentshas shown that it is possible to produce pseudotypes containing either vesicular stomatitis virusoraviantumor

virusgenomeswithSindbis envelope

glycopro-teins. However, noparticles could be detected

containing Sindbisnucleocapsids and

heterolo-gous viral antigens (22). This implies that the interactionof thealphavirus nucleocapsidwith its envelope is highly specific, much more so

thanthe interaction between glycoproteinsand the matrixproteins ofrhabdoviruses or

oncor-naviruses. Thismay

explain

the almost

complete

exclusion ofhost polypeptides from

alphavirus

virions which has beenobserved, in contrastto the situationwith other enveloped viruses dis-cussed above.

ACKNOWLEDGMENTS

IthankEdithLenchesandMaryStammreich-Martinfor excellenttechnical assistance, Charles Rice for his participa-tion inthepreliminaryexperiments, and James H.Straussfor stimulatingdiscussions and invaluable assistance in the prep-aration of thismanuscript.

This research was supported by Public Health Service grant AI-10793 from the National Institute of Allergy and Infectious Diseases.

LITERATURE CITED

1. Bishop, D.H. L., and M. S. Smith. 1977. Rhabdoviruses, p.167-280. In D. B. Nayak (ed.), The molecular biology ofanimal viruses. Marcel Dekker, Inc., New York. 2. Burge, B.W.,andE.R.Pfefferkorn. 1966. Isolation

and characterization of conditional lethal mutants of Sindbis virus. Virology 30:204-213.

3. David, A. E. 1971. The lipid composition of Sindbis virus. Virology 46:711-720.

4. Doherty, P. C., and R. M. Zinkernagel. 1975. H-2

compatibility isrequired for T-cell-mediated lysis of

targetcellsinfected withlymphocytic choriomeningitis virus. J.Exp. Med. 141:502-507.

5. Dulbecco, R., andM.Vogt.1954.One-step growth curve ofwesternencephalomyelitisvirusgrown in vitro and analysisof the virus yieldsfromsingle cells. J. Exp. Med.99:183-199.

6. Eagle, H. 1959. Amino acidmetabolism inmammalian cellcultures. Science 130:432-437.

7. Hecht,T.T.,and D. F. Summers. 1976. Interactions of vesicular stomatitis virus with murinecellsurface an-tigens. J. Virol. 19:833-845.

8. Holland,J.J., and E. D.Kiehn.1970.Influenza virus effects on cell membraneproteins. Science 167:202-205. 9. Klenk, H.-D.,and P. W.Choppin.1969.Lipids of plasma membranes of monkey and hamster kidneycellsand of parainfluenza virions grown in these cells. Virology 38:255-268.

10. Laemmli, U. K. 1970. Cleavage of structural proteins during theassemblyof thehead ofbacteriophage T4. Nature(London)227:680-685.

11. Lenard, J.,and R. W.Compans.1974.The membrane structureoflipid-containingviruses. Biochim.Biophys. Acta 344:51-94.

12. Leung,W.-C.,and W. E. Rawls. 1977. Virion-associated ribosomes are notrequired for thereplication of Pi-chindevirus.Virology81:174-176.

13. McSharry,J.J.,and R. R.Wagner. 1971.Lipid com-positionofpurifiedvesicular stomatitis virus. J. Virol. 7:59-70.

14. Pfefferkorn,E.R., and R. L.Clifford.1964.The origin oftheproteinofSindbis virus.Virology23:217-223. 15.Pfefferkorn,E.R.,and R.L. Clifford.1964.The origin

oftheproteinof Sindbis virus.Virology 20:446-456. 16.Pierce, J.S.,E.G.Strauss, and J.H.Strauss. 1974.

Effectof ionicstrengthonthebinding of Sindbis virus tochickcells.J. Virol. 13:1030-1036.

17.Scheid,A., and P. W.Choppin.1974.Identification of biologicalactitities ofparamyxovirus glycoproteins. Ac-tivationofcellfusion,hemolysisandinfectivity of pro-teolytic cleavage of an inactive precursor protein of Sendai virus.Virology 57:475-490.

18. Schrader, J.C.,B.A.Cunningham,and G. M. Edel-man. 1975.Functional interactions of viral and

http://jvi.asm.org/

(9)

474 STRAUSS

compatibility antigens at tumor cell surfaces. Proc. Natl. Acad. Sci. U.S.A. 72:5066-5070.

19. Strauss, E. G., C. R. Birdwell, E. M.Lenches, S. E. Staples, and J. H.Strauss.1977. Mutantsof Sindbis virus. II. Characterization ofa maturation defective

mutant,tslO3.Virology 82:122-149.

20. Strauss,J. H., and E. G. Strauss.1977.Togaviruses,p.

111-166. InD.Nayak (ed.),The molecularbiologyof animalviruses, vol,1.MarcelDekker,Inc., New York. 21. Wang, E., B. A. Wolf, R. A. Lamb, P. W. Choppin,

and A. R. Goldberg. 1976.Thepresenceof actinin enveloped viruses,p.589-600.InR. Goldman, T.

Pol-lard, and J. Rosenbaum (ed.), Cold Spring Harbor Con-ferencesonCell Proliferation, vol.3.Cold Spring Har-bor, N.Y.

22. Zavadova, Z., J. Zavada, and R.Weiss. 1977. Unilat-eral phenotypic mixing of envelope antigens between togaviruses and vesicular stomatitis virusoravian RNA

tumorviruses. J. Gen. Virol. 37:557-567.

J. VIROL.

http://jvi.asm.org/