Vol. 42, No. 3JOURNAL OF VIROLOGY, June 1982, p. 1046-10560022-538X/82/061046-11$02.00/0

Precursor Polypeptides to Structural Proteins of Visna VirusROBERT VIGNE,l* PIERRE FILIPPI,1 GILLES QUtRAT,1 NICOLE SAUZE,1 CHRISTIAN VITU,3

PIERRE RUSSO,3 AND PIERRE DELORI2Laboratoire de Virologie' and Laboratoire de Biochimie,2 Faculte de Medecine Nord, 13326 Marseille, Cedex

3, France, and Laboratoire National de Pathologie des Petits Ruminants, 06000 Nice, France3

Received 20 October 1981/Accepted 25 February 1982

Visna virus is a retrovirus which replicates in fibroblast-like cells of the sheepchoroid plexus through a lytic cycle. Visna virions contain three major low-molecular-weight proteins (p30, p16, and pl4) which, together with the genomicRNA and several molecules of reverse transcriptase, constitute the core structureof the virions. The core is surrounded by an envelope containing a majorglycoprotein (gp135). By analogy with the oncoviruses, these three groups ofstructural proteins (i.e., the internal proteins, the envelope glycoprotein, and thereverse transcriptase) are probably encoded by the gag, env, and pol genes,respectively. To elucidate the genetic organization of the visna virus genome andits expression, we studied the synthesis of viral proteins in infected sheep choroidplexus cells. Intracellular viral proteins were detected by immunoprecipitation ofpulse-labeled cell extracts with monospecific sera raised against p30, p16, andgp135 and resolution of the proteins by sodium dodecyl sulfate-polyacrylamide gelelectrophoresis. Immunoprecipitation with anti-p30 and anti-p16 sera allowed thecharacterization of the 55,000-dalton polypeptide precursor to internal virionproteins p30, p16, and p14 (Pr55Sag). Tryptic peptide mapping confirmed theprecursor-product relationship between Pr55Far and the three internal proteins. Inaddition, a gag-related polypeptide of 150,000 daltons was also detected. Thispolypeptide, which was less abundant than Pr55sag, is a likely precursor to theviral reverse transcriptase (Pr150gageP,I). Pr55gag and Pr1505agrP° are not glycosy-lated. The precursor related to major envelope protein gp135 is a glycosylatedpolypeptide with an average molecular weight of 150,000 (gPrlS0env). Pulse-chaseexperiments indicated that gPrlSOenv matures into glycoprotein gp135 intracellu-larly; however, gp135 was never preponderant in cell extracts. The non-glycosy-lated form of gPrlSoenv, which accumulated in the presence of 2-deoxy-D-glucose,appeared as a polypeptide of about 100,000 daltons. These results indicated thatvisna virus codes for the largest non-glycosylated env-related precursor among allof the retroviruses and therefore probably contains the largest env gene.

Visna virus, the prototype of the lentivirusgroup of the Retroviridae, contains a genome oftwo RNA molecules with nearly identical se-quences, each 10 kilobases long (4, 28, 29).Three low-molecular-weight non-glycosylatedproteins (p30, p16, and p14) constitute the core,and one major glycoprotein (gp135) is located inthe envelope of the virion (11, 12, 23). An RNA-dependent DNA polymerase, encoded by theviral genome and found in the viral particles, iscomposed of two subunits with an estimatedmolecular weight of 70,000 (16, 17). The geneswhich code for these three groups of proteins(i.e., the non-glycosylated core proteins, theenvelope glycoproteins, and the reverse tran-scriptase) and which occur in all of the retrovir-uses, have been designated gag, env, and pol,respectively (2).

Three classes of viral mRNAs have beenidentified previously in sheep choroid plexus(SCP) cells infected by visna virus. One is ofgenome size (36S), and the two smaller speciesare 27S and 21S (10). However, data obtainedrecently by hybridization blot techniques havesuggested that additional mRNAs are present ininfected cells (lOa).To determine which proteins are translated

from these RNAs, we started a study of viralprotein expression in infected SCP cells. Virus-specific proteins were detected by immunopreci-pitation of pulse-chase-labeled cell lysates withmonospecific sera raised against two core pro-teins (p30 and p16) and the major envelopeglycoprotein of the virus (gp135).

In this paper, we demonstrate that intracellu-lar production of the visna virus structural pro-

1046

VISNA VIRUS PROTEIN BIOSYNTHESIS 1047

teins involves the synthesis and processing ofseveral precursor polypeptides, as it does in theother retroviruses. A non-glycosylated precur-sor (Pr559aB) gives rise to the major internalstructural proteins. Another non-glycosylatedpolypeptide of 150,000 daltons is probably theprecursor common to the gag-related proteinsand the reverse transcriptase (PrlO5QagPpo). Fi-nally, a large, 150,000-dalton glycosylated poly-peptide (gPrl50env) is the precursor to the enve-lope glycoprotein (gp135).

MATERLALS AND METHODS

Virus and ceUls. Visna virus strain K1514 was usedthroughout this work. SCP cells were grown in Eagleminimal essential medium containing 2% newborn calfserum, 8% lamb serum, and antibiotics, as previouslydescribed (10). When the cells were going to beinfected, the growth medium was removed and re-placed by Eagle minimal essential medium containingonly 2% lamb serum since calf serum has been shownto inhibit visna virus infection (26).

Virion and cell labeling. Visna virus strain K1514was metabolically labeled with 100 ,uCi of [35S]meth-ionine (600 Ci/mmol; Radiochemical Centre, Amer-sham, England) per ml in methionine-free Eagle mini-mal essential medium containing 2% lamb serum andantibiotics for 16 h at 37°C. The virus was purifiedfrom the tissue culture fluid by rate and densitysucrose gradient centrifugation, as described previous-ly (28).

[35S]methionine labeling of the infected SCP cellsused to isolate intracellular virus-specific proteins wassimilar to the labeling described above, except that thecells were labeled for shorter periods of time. For thepulse experiments, the labeling time was 40 min. Forthe chase periods, complete Eagle minimal essentialmedium containing 2% lamb serum was added to thepulse-labeled cells for varying times (see below). Tolabel the carbohydrate moieties of the glycoproteins,20 ,uCi of D-[14C]glucosamine (22.6 mCi/mmol; Radio-chemical Centre) per ml was added for 1 or 16 h to aculture grown in glucose-free medium containing 2%lamb serum. Cells pulse-labeled for 1 h were chased byadding complete medium. To inhibit the glycosylationof the glycoproteins, infected cells grown in glucose-free and amino acid-free medium containing 2% lambserum were treated with 15 mM 2-deoxy-D-glucose for2 h. After the medium was removed, the cells werepulse-labeled with [3H]leucine (20 Ci/mmol; Radio-chemical Centre) in the same medium containing 15mM 2 deoxy-D-glucose; 2-h chases were performed incomplete medium with or without the glycosylationinhibitor.

Antisera. To raise antiserum against whole virions,visna virus strain K1514 or K796 was purified asdescribed previously (28). A 100-,ug amount of purifiedvisna virus was suspended in 1 ml of saline (0.13 MNaCl, 7.5 mM MgCl2, 5 mM KCl), heat denaturated,mixed with an equal volume of Freund completeadjuvant, and injected subcutaneously into the back ofa female albino white rabbit. Two additional subcuta-neous booster injections of 200 ,ug of heat-denaturedvirus were administered into the back of the rabbit at

2-week intervals, and serum was collected by cardiacpuncture 7 to 10 days after each injection.To raise monospecific sera against p30, p16, and

gpl35 (strains K796 and K1514), viral proteins werepurified on 6 to 18% polyacrylamide slab gel as de-scribed below. A 2-mg amount of purified virus wasdisrupted in 1 ml of the lysis buffer used for immuno-precipitation and diluted in 1 ml of sample buffer (seebelow). After heat denaturation viral proteins wereseparated on a preparative electrophoresis slab gel byusing a 10-cm-wide well. Viral proteins were usuallylocated by staining a 0.5-cm vertical band at bothedges of the preparative gel with Coomassie blue. Thebands of acrylamide containing the separated proteinswere excised from the gel. Then these bands werecrushed with a Dounce piston and, after suspension ofthe gel particles in 1 to 2 ml of saline, mixed with anequal volume of Freund complete adjuvant. Rabbitswere injected as described above for the completevirus.

Sheep anti-gp135 serum obtained by collecting se-rum of sheep experimentally infected by visna virusK1514 was kindly provided by M. Brahic and A.Haase.An examination of the specificities of the antisera by

immunoprecipitation of [35S]methionine-labeled pro-teins from visna virions revealed that the anti-wholevirus sera were mainly reactive with gp135, p30, andp16, but were not reactive with p14. The rabbit seraraised against p30, p16, and gpl35 (strains K796 andK1514) and the anti-gp135 sheep serum (strain K1514)were monospecific and reacted with equal efficiencieswith the corresponding antigens of both visna virusstrains (K796 and K1514).

Immunoprecipitation and protein gel electrophoresis.After labeling, the cells were washed once with phos-phate-buffered saline and then lysed with lysis buffer(0.01 M Trishydrochloride, pH 7.5, 0.1 M NaCl, 0.001M EDTA, 1% Triton X-100, 0.5% sodium deoxycho-late, 0.1% sodium dodecyl sulfate [SDS]). This bufferwas removed from the cells after 20 min at 4°C, and thelysate was clarified by centrifugation at 100,000 x gfor 30 min. Portions of the lysate were incubated with2 to 5 ,ul of antisera for 1 to 4 h at 0°C. Immunecomplexes were collected by adsorption to protein A-bearing Staphylococcus aureus (13), which was thenwashed four times to remove unbound material. Afterwashing, the adsorbed proteins were eluted from thebacteria in sample buffer (2% SDS, 5% P-mercaptoeth-anol, 10% glycerol, 0.125 M Tris-hydrochloride, pH6.8, 0.001% bromophenol blue), and the bacteria wereremoved by centrifugation. The same procedure wasused for labeled virus. SDS-polyacrylamide gel elec-trophoresis was performed in 10- or 20-cm gels whichcontained a 6 to 18% acrylamide gradient in the buffersystem of Laemmli (14). Radioactive polypeptideswere detected by fluorography (15).

Tryptic peptide mapping. Tryptic peptide mappingwas performed as described previously (30). Briefly,polypeptides to be mapped were located in the gel byautoradiography. The bands were excised, and theproteins were eluted, precipitated with carrier protein(150 ,ug of ovalbumin), oxidized with performic acid,and digested with tolylsulfonyl phenylalanyl chloro-methyl ketone-treated trypsin. [35S]methionine-la-beled peptides were separated in two dimensions oncellulose thin-layer plates (20 by 20 cm; E. M. Labora-

VOL. 42, 1982

1048 VIGNE ET AL.

tories). The first dimension was an electrophoresis at600 V for 100 min in an aqueous buffer containing 3%acetic acid and 2.5% pyridine (pH 4.5), and the seconddimension was chromatography with buffer containingpyridine, water, acetic acid, and n-butanol (60:60:15:75). Labeled peptides were detected by autoradiog-raphy with Kodak NS 54T film at room temperature.

RESULTSVisna virion major proteins: size and composi-

tion. Before we started to study the intracellularviral proteins, we analyzed the [35S]methionine-labeled proteins of purified visna virions (strainK1514) on a 14% polyacrylamide slab gel todetermine the protein compositions of the viri-ons and the sizes of the major proteins (Fig. 1A).As previously described (11, 12, 18), about 20polypeptides are packaged into the viral parti-cles. However, among these 20 proteins, onlythe following 4 are considered major: the largeenvelope glycoprotein gpl35 and the three smallinternal proteins p30, p16, and p14. By electro-phoresing these proteins in a continuous 14%polyacrylamide slab gel (Fig. 1A), we obtainedexactly the same sizes as other workers havefound for these polypeptides (25).Among the 16 minor polypeptides, two large

polypeptides with molecular weights of 97,000(p97) and 80,000 (p80) were particularly visiblein many analyses of the [35S]methionine-labeledproteins from purified visna virions. Figure 1Bshows an example of such an analysis on a 6 to18% polyacrylamide gradient slab gel. As previ-ously suggested, these additional proteins couldbe either cellular proteins specifically packagedin the free viral particles or precursor polypep-tides of viral structural proteins (18).To approach this question, we compared p97

and p80 with the four major viral proteins byusing tryptic peptide mapping of 35S-labeledproteins. Figure 2 shows autoradiograms of thetryptic peptide maps. p30 contained five majorpeptides (peptides 1 to 5), p16 had two majorpeptides (peptides 6 and 7), and p14 had onepeptide (peptide 8). Although in this experimentthe p14-specific peptide was slightly present inthe p16 map, we concluded that the three majorinternal proteins are unique entities. The sameconclusion could be reached for the three largepolypeptides gpl35, p97, and p80. The smallnumber of peptides in the map of gpl35 com-pared with the number in the map of p30 wasartificially emphasized by the fact that the mapof p30 was overexposed and revealed manyweakly labeled peptides. However, another pos-sible explanation is that gpl3S contains a smallnumber of methionine residues and thus a smallnumber of labeled peptides. Relative to p97 andp80, the peptide maps suggest that these poly-peptides are not high-molecular-weight precur-sors to the envelope or internal proteins; rather,

.B)

gp 1 3 5 - :,:.*

-gp135- p97

.-_ p80.w..*>; ..%i;i..... ::..

p30-4o

p 16-p 14 - _

_omo - p30

__ - pa6- p14

FIG. 1. Major proteins of visna virions. [35S]methi-onine-labeled proteins of purified visna virus weredissociated with lysis buffer, diluted with an equalvolume of sample buffer, and heat denatured. Thedissociated proteins were electrophoresed in an SDS-14% polyacrylamide gel (A) or in an SDS-6 to 18%polyacrylamide gel (B). After electrophoresis the gelswere fixed and treated for fluorography as described inthe text.

they are either precursors to other viral proteinsor cellular proteins packaged into the virions.

Intracellular viral polypeptides related to themajor structural proteins of visna virus. To iden-tify the virus-specific polypeptides related to themajor structural proteins of visna virus, a seriesof pulse-chase experiments were performed withinfected SCP cells. Extracts of SCP cells labeledwith [35S]methionine for 5 h were immunopreci-pitated with appropriate sera and analyzed onpolyacrylamide slab gels. When these extractswere precipitated with monospecific anti-p30sera, several polypeptides with molecularweights of more than 30,000 were observed, inaddition to mature p30. These gag-related poly-peptides had apparent molecular weights of35,000 (P35), 50,000 (P50), 55,000 (P55), and150,000 (P150). Polypeptide P150 migrated as adoublet containing a major form and a minorform (Fig. 3, lanes B and C). The proteinspresent in the molecular weight range of 80,000were probably cellular since they were alsopresent with control normal sheep serum (Fig. 3,lane F). Immunoprecipitation of the extractswith anti-p16 serum showed the presence of P50,P55, and P150 (only the major form), in additionto p16 (Fig. 3, lane D); P35 was not present.Figure 3, lane E, shows that the serum raisedagainst glycoprotein gpl35 recognized a majorpolypeptide of 150,000 daltons and minor poly-

J. VIROL.

VISNA VIRUS PROTEIN BIOSYNTHESIS 1049

41

0

.A

FIG. 2. Tryptic peptide maps of major proteins of visna virions. [35S]methionine-labeled protein bands wereexcised from gels as described in the legend to Fig. 1, and the proteins were eluted, precipitated with carrierprotein, oxidized with performic acid, and digested with tolylsulfonyl phenylalanyl chloromethyl ketone-treatedtrypsin. The resulting peptides were separated on thin-layer cellulose plates by electrophoresis at pH 4.5 in thefirst dimension and chromatography in n-butanol-water-pyridine-acetic acid in the second dimension. The[35S]methionine labeled peptides were visualized by autoradiography. The resulting viral protein fingerprints areshown for p30 (A), p16 (B), p14 (C), gp135 (D), p97 (E), and p80 (F). The major peptides of p30, p16, and p14 arenumbered.

peptides that appeared on the fluorogram as asmear in the molecular weight range of maturegp135. The relationship of these minor polypep-tides to gp135 and the major 150,000-daltonpolypeptide is discussed below.

gag-related precursor polypeptides: processingand composition. To determine whether precur-sor-product relationships exist among the gag-related polypeptides, pulse-chase experimentswere performed. Visna virus-infected SCP cellslabeled with [35S]methionine for 40 min werechased for 1, 3, and 5h, and their cell lysateswere immunoprecipitated with anti-p30 serum.Figure 4, lane E, shows the presence of the p30-related polypeptides P150, P55, P50, and P35 inaddition to mature p30 after 40 min of pulse-labeling. During the chase (Fig. 4, lanes F to H),the concentrations of the higher-molecular-weight polypeptides decreased, particularly theconcentration of P150. After a 1-h chase, onlythe upper form of the P150 doublet was visible(lane F). In contrast, p30 occurred at nearly thesame level throughout the chase period. Anexplanation for this constant level of p30 during

the chase period may be that the rate of exportof p30 outside the cells is equal to the rate ofproduction from the precursor P55 through theintermediates P50 and P35. The same resultswere obtained when anti-p16 serum was used,except that P35 was not detected, implying thatP35 does not contain p16 (data not shown).These pulse-chase experiments demonstratedthat mature internal viral proteins p30 and p16are found inside cells as precursor polypeptides,the largest of which has a molecular weight of150,000. This polypeptide, which is large enoughto contain the three major internal proteins plusthe reverse transcriptase (molecular weight,70,000 [20]), probably corresponds to the largeprecursor of the RNA-dependent DNA polymer-ase found in all of the retroviruses.The major polypeptide precursor to core pro-

teins p30 and p16 has a molecular weight of55,000. To determine whether P55 contains p14as well as p30 and p16, we compared the[35S]methionine-labeled peptides obtained froma tryptic digest of P55 with the peptides of amixture of p30, p16, and p14. P55 and intracellu-

VOL. 42, 1982

a

0

I.-

II

II

A*A

I

1050 VIGNE ET AL.

V A 8 C D E F

Ulllinmmmo

gpl35-4

p97_

p80 -

AwY

*!:::

p3

.4

p16 - _,

p4 -

cated at one of its ends, a related larger peptidecontaining amino acid sequences adjacent tothose of p30 should be present in the precursor,P55. Despite this imprecision, we concluded thatP55 was the major precursor of the internalvirion proteins and designated it Pr55ga'. Byanalogy with other retroviruses, P150 was pre-

en sumably the precursor of the reverse transcrip-- Pl 50 tase and was designated Pr1509a"Po.

env-related precursor polypeptides: processingand glycosylation. Pulse-chase experiments werealso performed to detect env-related polypep-tides. SCP fibroblasts were labeled with[35S]methionine for 40 min and chased for 1, 3,and 5 h. env-related polypeptides were immun-

- P 5 ag oprecipitated with anti-gpl35 serum and separat-- P50a ed by acrylamide gel electrophoresis (Fig. 4).

After a short labeling period (40 min), the im-- 3 5g ag munoprecipitated polypeptides migrated as a

doublet in the molecular weight range of 150,000(Fig. 4, lane I). In contrast, only the upper bandof the doublet was detected after a long labelingperiod (Fig. 3, lane E). The amount of precur-

4UI

A 3 C D E F G H s 1<. _

FIG. 3. Polypeptide precursors to structural pro-teins of visna virus. Extracts of SCP cells pulse-labeled with [35S]methionine for 5 h were immunopre-cipitated with appropriate sera and analyzed on anSDS-6 to 18% polyacrylamide gel. Lane V, [35S]methi-onine-labeled virion proteins used as standards; laneA, anti-whole virus serum (strain K796); lane B, anti-p30 serum, bleed 1; lane C, anti-p30 serum, bleed 2;lane D, anti-p16 serum; lane E, anti-gp135 serum; laneF, normal sheep serum. P150¶ag appears as a doubletcontaining a major form and a minor form. After a longlabeling period, P150env appeared as a single band.

lar p30 were metabolically labeled with[35S]methionine and isolated by immunoprecipi-tation with anti-p30 serum from SCP cells. Viri-on polypeptides p30, p16, and p14 were isolatedfrom [ S]methionine-labeled virus purified fromthe tissue culture fluid of infected SCP cells. Acomparison of the P55 tryptic peptide map (Fig.SB) with the map of the mixed virion p30-p16-p14 preparation (Fig. SD) showed that P55 con-tains four peptides of p30 (peptides 1 to 4), bothpeptides 6 and 7 of p16, and unique peptide 8 ofp14. The presence of an additional peptide in thepeptide maps of the p30 found in both infectedcells and virions (Fig. SA and C) compared withP55 is difficult to interpret. Ifwe assume that p30contains a new methionine-labeled peptide lo-

U"_of

I.

-. Pi60 Y

jp 5 o

- P35*4;*3_ p3l

FIG. 4. Pulse-chase labeling of env- and gag-relat-ed precursors to products. Cultures of SCP cellsinfected by visna virus were pulse-labeled for 40 minwith [35S]methionine and chased for 1, 3, or 5 h incomplete medium, as described in the text. As de-scribed in the legend to Fig. 3, intracellular viralproteins were analyzed by immunoprecipitation withnormal serum (lanes A through D), anti-p30 serum(lanes E through H), and anti-gp135 (lanes I throughL). Lanes A, E, and I, 40-min pulse; lanes B, F, and J,1-h chase; lanes C, G, and K, 3-h chase; lanes D, H,and L, 5-h chase. [35S]methionine-labeled visna virus(lane V) was included as a standard.

J. VIROL.

v

VISNA VIRUS PROTEIN BIOSYNTHESIS 1051

14> ..'I

a

44_wrs

22 3

._,..3

2

07/

©F11 3

a

4

Ihr:I.I.

FIG. 5. Comparison of tryptic peptide maps of Pr55y'S, p30, p16, and p14. [35S]methionine-labeled trypticpeptides were prepared and separated on thin-layer cellulose plates as described in the legend to Fig. 2. (A)Intracellular p30. (B) Pr559ag. (C) Virion-associated p30. (D) Mixture of p30, p16, and p14.

sors decreased with the chase periods, withoutgiving raise to a large amount of gp135. Theabsence of a distinct gp135 band after chaseperiods suggested that the gp135 molecules areexported to the extracellular fluid at a higherrate than they are produced. Because the ex-tracts of pulse-chase-labeled cells did not con-tain high concentrations of eventual mature gly-coproteins as gp135, we were not able todetermine whether a small protein related to envand linked to mature gp135 by disulfide bonds asthe piSE protein of murine leukemia viruses (27)was derived from the large precursor. There-fore, we labeled cells for 24 h with [35S]methio-nine to increase the concentrations of labeledproteins in the cell extracts 10 to 20 times.Figure 6 shows that highly labeled cell extracts

contained polypeptides in the molecular weightranges of 150,000 and 135,000 but not smallproteins in the molecular weight range of 15,000(Fig. 6, lanes A through C; lanes A and B showthe proteins precipitated by normal serum andanti-p30 serum, respectively). However, twopolypeptides with molecular weights 85,000 and50,000 were detected with anti-gp135 serum;these molecules were probably degradationproducts of the large env-related polypeptides.Because these proteins were present at lowconcentrations during viral replication in cellextracts and in mature virions (Fig. 1), wesuggest that they represent degradation productsof mature proteins which are not necessary tothe replication cycle.These experiments were not conclusive with

VOL . 42, 1982

6 .*el

:,a

1052 VIGNE ET AL.

A a

9p135p97p80

Pr55 g

p3O

FIG. 6. gp135 is the majcvisna virus. Cultures of S(virus were labeled for 16 hdescribed in the legend toproteins were analyzed bynormal serum (lane A), ananti-gp135 serum (lane C). ]thionine-labeled visna virus

C v respect to the absence of a pl5E equivalent invisna virus. For example, the failure to detect apl5E equivalent in visna virus by direct immu-noprecipitation experiments could reflect theabsence of the correct antibody in the sera if

_Prl 50 plSE is not linked to gp135 by disulfide bonds ora methionine deficiency in the small protein.To determine whether P150"fV is glycosylated,

SCP cells infected by visna virus strain K1514were labeled with [14C]glucosamine for 16 or 1 h:::.2 2)"0 and chased for 1 or 4 h. In contrast to [35S]meth-ionine labeling (Fig. 4) or [3H]leucine labeling(data not shown), P150"" was almost undetect-able after the 40-min pulse (Fig. 7, lane A), butappeared to be more labeled after the chaseperiods of 1 and 4 h (Fig. 7, lanes B and C). Thisslow incorporation of [14C]glucosamine into theenv-related polypeptides was expected since theprocessing that this compound requires to beincorporated is more complex and less efficientthan the processing of the amino acid isotopes.After 16 h of labeling, P150"ev was strongly

r envelope glycoprotein of labeled, and polypeptides in the moleculariP ctlls in5fmete.dbyAvisna weight range of 135,000 were also visible (Fig. 6,Fig. 3, intracellular viral lanes D and D'). A tissue culture fluid of infectedimmunoprecipitation with SCP cells was also examined; throughout theiti-p30 serum (lane B), or labeling, no P150enV was exported from the cellsLane V contained [35S]me- (Fig. 7, lanes E to G). The proteins larger thanproteins. gp135 present in Fig. 7, lane H, were not env-

. a 3, C D E G H ,D/

a,emwoo - qpl 35

-- p16

FIG. 7. env-related precursor polypeptide is glycosylated. Cultures of SCP cells infected by visna virus werepulse-labeled for 40 min with [14C]glucosamine and chased for 1 or 4 h in complete medium as described in thetext. Intracellular env-related proteins were immunoprecipitated by anti-gp135 serum and analyzed in a 6 to 18%polyacrylamide gel. In parallel, the culture fluid was collected during the pulse-chase experiment to performimmunoprecipitation experiments, as for cell extracts. Lane V contained standard virion proteins. Lanes Athrough D contained intracellular env-related proteins (lane A, 40-min pulse; lane B, 1-h chase; lane C, 4-h chase;lane D, 16 h of labeling; lane D', same as lane D, but a shorter exposure). Lanes E through H containedextracellular env-related proteins (lane E, 40-min pulse; lane F, 1-h chase; lane G, 4-h chase; lane H, 16 h oflabeling).

J. VIROL.

q"P!W,.

VISNA VIRUS PROTEIN BIOSYNTHESIS 1053

related precursor molecules since they were alsoprecipitated by normal sheep serum (data notshown). By contrast, gp135 was clearly presentafter a long labeling period (Fig. 7, lane H).Glycosylated P150 was the precursor to themajor envelope glycoprotein gp135 and was des-ignated gPrl5Oenv. A similar experiment withanti-p30 serum revealed neither Pr55gag norPr15V'9-P'1. Moreover, no new gag-related pre-cursors were detected (data not shown). Conse-quently, in visna virus infections, only the env-related proteins appear to be glycosylated.

Unglycosylated form of gPrl5Oe"v. Becauseglycosylation of polypeptide chains raises theirmolecular weights, we determined the size of theunglycosylated polypeptide chain of gPrl5Oenv.Visna virus-infected SCP cells were preincubat-ed with 15 mM 2 deoxy-D-glucose to blockglycosylation of the envelope-related polypep-tides (8, 24) and labeled with [3H]leucine for 1 h.env-related proteins were immunoprecipitatedfrom lysates with anti-gp135 serum. As a con-trol, cells were also labeled without 2-deoxyglu-cose treatment (Fig. 8). Figure 8, lane V, showsthe [35S]methionine-labeled visna virion proteins(strain K1514). In lane A the doublet ofgPrl5oenv described above (Fig. 4 and 5) ap-peared in the absence of the inhibitor. When thecells were treated with inhibitor, the doublet ofgPrl5oenv was absent. Instead, a new polypep-

A It G HE I K L V

we

c qPrlO1 5 0: =

Pr 55

-.

p30 -

FIG. 9. Lack of glycosylation of gag-related pre-cursors. Immunoprecipitations of p30-related proteinsfrom the extracts described in the legend to Fig. 8 wereperformed with normal sheep serum (lanes A throughD). Lane A, Untreated cells, 40-min pulse; lane B,treated cells, 40-min pulse; lane C, treated cells, 2-hchase with inhibitor; lane D, treated cells, 2-h chase inthe absence of inhibitor. The same series of analyseswere performed with anti-p30 serum (lanes E throughH) and with anti-gp135 serum as the reference (lanes Ithrough L). Lane V contained [35S]methionine-labeledvisna virus proteins. The proteins were electropho-resed in a 6 to 18% polyacrylamide gel.

V ABC DE

gPr 150 =

P105

_ ^~~~~~~~~~~~~~~~~~~~~~~.

~~-gpl35p97p80

_,som - p30

_amft- p16

FIG. 8. Effect of 2-deoxy-D-glucose on the glyco-sylation of the env-related precursor. Cultures of SCPcells infected by visna virus were treated with 15 mM2-deoxy-D-glucose for 2 h, pulse-labeled for 1 h with[3H]leucine, and chased for 2 h with or without theglycosylation inhibitor. Immunoprecipitation and gelanalysis of gp135-related proteins were performed asdescribed in the legend to Fig. 7. Lane V, Standardvirion proteins; lane A, untreated cells, 1-h pulse; laneB, treated cells, 1-h pulse; lane C, treated cells, 2-hchase with inhibitor; lane D, treated cells, 2-h chasewithout inhibitor; lane E, control containing the un-treated cell extract as in lane A, immunoprecipitatedby normal sheep serum.

tide which migrated faster was visible (Fig. 8,lane B). In fact, this 105,000-dalton protein washeterogeneous and contained two bands. Duringa 2-h chase in the presence of the drug (Fig. 8,lane C) and more clearly in the absence of theinhibitor (lane D), the modified polypeptides hadslightly lower electrophoretic mobilities. Thissuggests that glycosylation was still occurring ata low rate. Thus, we cannot exclude the possibil-ity that the 105,000-dalton polypeptides werestill partially glycosylated. This conclusionseems to be strengthened by the fact that a veryfaint band was visible on the X-ray film (but noton the reproduction shown in Fig. 8) in the areaof the gel bounded by the virion markers p97 andp80. The polypeptide with a molecular weight ofabout 50,000 present in Fig. 8, lanes A and B,was probably a nonviral protein since it wasprecipitated as well with control normal sheepserum (Fig. 8, lane E).

Similar experiments were performed to deter-mine the glycosylation of the gag-related poly-peptides. Figure 9, lane F, shows that the gag-related polypeptides synthesized in presence of2-deoxyglucose were not modified in their elec-trophoretic mobilities or in their levels of syn-thesis compared with those produced in absence

VOL. 42, 1982

1054 VIGNE ET AL.

of inhibitor (lane E); this was particularly truefor PrlS5garPol and Pr55rag. Figure 9, lane J,shows the modified env-related polypeptides de-scribed above. Moreover, the absence of150,000-dalton polypeptides among these modi-fied env polypeptides demonstrates clearly thatthe two kinds of precursors (Pr15O9'9-P'1 andPrlSOenv) are unrelated and that our sera weremonospecific.The 2-deoxyglucose experiment indicated that

many env-related polypeptide intermediatescould be present during the pulse experiment,from the unglycosylated precursor of about100,000 daltons to the glycosylated precursor of150,000 daltons. The diffuse bands of polypep-tides observed during the pulse-chase experi-ments (Fig. 3, 4, 6, 7, and 9) could be interpretat-ed either as partially glycosylated intermediatesto the gPrlSOenv or as intermediates processedfrom the precursor stage represented bygPrlSoenv to the mature stage represented byenvelope glycoprotein gp135 of the virion.Moreover, it is likely that a small number ofmature gp135 molecules are also present in thecytoplasm of the cells, since the env-related pre-cursors are probably processed intracellularly.

DISCUSSIONVisna virions contain the following four major

structural proteins: one large 135,000-dalton gly-coprotein and three internal proteins (p30, p16,and p14). By analogy with the genetic organiza-tions of other retroviruses (2), the only glycopro-tein of the viral envelope (gp135) is encoded bythe env gene, and the three major internal struc-tural proteins are encoded by the gag gene.Immunoprecipitation and peptide mapping tech-niques showed that all four proteins are unique,without any overlap in amino acid sequence.pol, the third replicating gene of visna virus,codes for the reverse transcriptase. Previously,visna virus reverse transcriptase monomerswere shown to have an average molecularweight of 70,000 (17). Besides these four struc-tural proteins and the reverse transcriptase, weand others (11, 12, 18) have shown that addition-al polypeptides are present in the virions. Sever-al of these molecules could be polypeptide inter-mediates related to the structural proteins. Forexample, a virion 35,000-daltons polypeptideseems to correspond exactly in electrophoreticmobility to the intracellular gag-related interme-diate P35gag. By contrast, a comparison of thepeptides maps of p97 and p80 with the maps ofthe major structural virion proteins indicatedthat these two large polypeptides are not gag orenv related. They either are related to the re-verse transcriptase gene, are related to anotherviral gene still unknown, or originated from thecell genome and are packaged specifically in

viral particles. Further analyses will be neces-sary to distinguish among these possibilities.

gag-related precursor polypeptides were char-acterized by pulse-chase labeling and immuno-precipitation techniques. A large 150,000-daltonpolypeptide was detected. Based on analogywith other retroviruses, such a polypeptide islikely to represent the gag-pol precursor. Itappeared after a pulse period of few hours as adoublet. Since the visna virus was plaque puri-fied, this heterogeneity of size was not related toa heterogeneity of variants present in our stocksof viruses but rather to distinct precursor poly-peptides.The major precursor of the internal virion

protein has a molecular weight of 55,000, whichis relatively low compared with other mammali-an retroviruses (1, 5, 20, 22, 27). Since in manyretroviruses the sum of the molecular weights ofthe gag-related mature proteins is very close tothe sum of the molecular weights of the precur-sor polypeptides (3, 6, 19, 21), we suggest thatvisna virus Pr55gag is a precursor to only thethree mature proteins p30, p16, and p14. Pulse-chase experiments showed that the mature inter-nal virion proteins are present inside the cells.This suggests that cleavage and maturation ofPr55gag occur mainly intracellularly, in contrastto mouse mammary tumor virus and murineleukemia virus, in which cleavage is largelyvirion associated (5, 27). This intracellular matu-ration of the internal virion proteins was con-firmed by a polypeptide analysis of the intracel-lular ribonucleoprotein complexes (lOa). Theseribonucleoprotein complexes, which are notcontaminated by extracellular or budding viri-ons, contain as a major gag-related species themature p30 polypeptide.We detected a glycosylated polypeptide pre-

cursor to gp135 which had a molecular weight ofapproximately 150,000. This molecule appearedafter 1 h of pulse-labeling as a doublet. Thisheterogeneity most likely derived from distinctprecursor polypeptides rather than from variantspresent in our stocks of viruses. Similarly, adoublet nature for env-related precursors hasbeen described for Friend murine leukemia virus(9).Among retroviruses, the env-related precur-

sor of visna virus is the largest env-relatedprecursor which has been isolated. As weshowed with 2-deoxyglucose inhibition of glyco-sylation, the molecular weight of the env-related150,000-dalton precursor was overestimated byat least 45,000 since the unglycosylated form hada molecular weight of approximately 105,000.Although we found that glycosylation was notcompletely stopped by the inhibitor, we believethat the env-related unglycosylated precursorshould have a molecular weight of approximate-

J. VIROL.

VISNA VIRUS PROTEIN BIOSYNTHESIS 1055

ly 90,000 for two reasons. We observed a faintband at this molecular weight as the smallestprotein made in our inhibition studies. Recentdata on the molecular weights of the unglycosy-lated env-related polypeptides of retroviruseshave shown that they have molecular weightsnot below 60% of their glycosylated sizes (7, 8)(i.e., 90,000 for visna virus).The env-related precursor of visna virus ma-

tures into glycoprotein gp135. We obtained noevidence for a small protein analogous to pl5Eof murine leukemia viruses. However, the fail-ure to detect such a protein by direct immuno-precipitation experiments could reflect the ab-sence of the correct antibody in the sera or amethionine deficiency in the protein.The env-related precursor is not exported

from the cells, which suggests that maturation isintracellular. However, gp135 was poorly repre-sented in the cell extracts during the pulse-chaseexperiments (Fig. 3, 4, 6, 7, and 9). The low levelof gp135 could be explained by a rate of synthe-sis lower than the rate of export of the extracel-lular fluid. Additional experiments will be neces-

sary to elucidate viral envelope synthesis,particularly at the level of intracellular ribonu-cleoprotein particles, where the env-related pre-cursors are still preponderant (1Oa).The present data on protein synthesis indicate

that the visna virus genome codes for two gag-related precursors and one env-related precursorof the major structural proteins of virions (i.e.,(Pr55a, pPr15gag-Po, and gPrl5oenv). We arenow investigating the expression of the viralgenome and its organization by analyzing the invitro translation products of genomic and viralmRNAs.

ACKNOWLEDGMENTS

We thank J. Tamalet for support during this work, J.Delamarter for critically reading the manuscript, and M.Brahic and A. T. Haase for providing serum samples.

This work was supported by grants from Association duDeveloppement de la Recherche sur le Cancer (Villejuif),Fondation de la Recherche Medicale (Paris et Marseille),Institut de la Sante et de la Recherche Medicale (CRL 802012),and Mission de la Recherche.

LITERATURE CITED

1. Arcement, L. J., W. L. Karshin, R. B. Naso, G. JamJoom,and R. B. Arlinghaus. 1976. Biosynthesis of Rauscherleukemia viral proteins: presence of p30 and envelope p15sequences in precursor polypeptides. Virology 69:763-774.

2. Baltimore, D. 1974. Tumor viruses. Cold Spring HarborSymp. Quant. Biol. 39:1187-1200.

3. Barbacid, M., J. R. Stephenson, and S. A. Aaronson. 1976.gag gene of mammalian type-C RNA tumor viruses.Nature (London) 262:554-559.

4. Beemon, K. L., A. J. Faras, A. T. Haase, P. H. Duesberg,and J. E. Maisel. 1977. Genomic complexities of murineleukemia and sarcoma, reticuloendotheliosis, and visnaviruses. J. Virol. 17:525-537.

5. Dickson, C., and M. Atterwill. 1978. Polyproteins related

to the major core protein of mouse mammary tumor virus.J. Virol. 26:660-672.

6. Dickson, C., and M. Atterwill. 1979. Composition, ar-rangement and cleavage of the mouse mammary tumorvirus polyprotein precursor Pr77gag and p110oa9. Cell17:1003-1012.

7. Diggelman, H. 1979. Biosynthesis of an unglycosylatedenvelope glycoprotein of Rous sarcoma virus in the pres-ence of tunicamycin. J. Virol. 30:799-804.

8. Dresler, S., M. Ruta, M. J. Murray, and D. Kabat. 1979.Glycoprotein encoded by the Friend spleen focus-formingvirus. J. Virol. 30:564-575.

9. Evans, L. H., S. Dresler, and D. Kabat. 1977. Synthesisand glycosylation of glycoprotein precursors to the inter-nal core proteins of Friend murine leukemia virus. J.Virol. 24:865-874.

10. Filippi, P., M. Brahic, R. Vigne, and J. Tamalet. 1979.Characterization of visna virus mRNA. J. Virol. 31:25-30.

10a.Filippi, P., R. Vigne, G. Querat, C. Jouanny, and N.Sauze. 1982. Intracellular ribonucleoprotein complexes ofvisna virus are infectious. J. Virol. 42:1057-1066.

11. Haase, A. T. 1975. The slow infection caused by visnavirus. Curr. Top. Microbiol. Immunol. 72:101-156.

12. Haase, A. T., and Baringer. 1974. The structural polypep-tides of RNA slow viruses. Virology 57:238-250.

13. Kessler, S. W. 1975. Rapid isolation of antigens from cellswith staphylococcal protein A-antibody adsorbant: pa-rameter of the interaction of antibody-antigen complexeswith protein. A. J. Immunol. 115:1617-1624.

14. Laemmli, U. K. 1970. Cleavage of structural proteinsduring the assembly of the head of bacteriophage T4.Nature (London) 227:680-685.

15. Laskey, R. A., and A. D. Mills. 1975. Quantitative filmdetection of 3H and "4C in polyacrylamide gels by fluorog-raphy. Eur. J. Biochem. 56:335-341.

16. Lin, F. H., M. Genovese, and H. Thormar. 1973. Multipleactivities of DNA polymerase from visna virus. Prep.Biochem. 3:525-539.

17. Lin, F. H., and M. Papini. 1979. Evidence for two formsof RNA-dependent DNA polymerase in visna virus. Bio-chim. Biophys. Acta 561:383-395.

18. Lin, F. H., and H. Thormar. 1979. Precipitation of visnaviral proteins by immune sera of rabbits and sheep. J.Virol. 29:536-539.

19. Murphy, E. C., and R. B. Arlinghaus. 1978. Trypticpeptide analysis of polypeptides generated by prematuretermination of cell-free protein synthesis allows a determi-nation of the Rauscher leukemia virus gag gene order. J.Virol. 28:929-935.

20. Racevskis, J., and N. H. Sarkar. 1978. Synthesis andprocessing of precursor polypeptides to murine mammarytumor virus structural proteins. J. Virol. 25:374-383.

21. Reynolds, R. K., and J. R. Stephenson. 1977. Intracis-tronic mapping of the murine type C viral gag gene by useof conditional lethal replication mutants. Virology 81:328-340.

22. Schochetman, G., C. W. Long, S. Oroszlan, L. Arthur, andD. L. Fine. 1978. Isolation of separate precursor polypep-tides to mouse mammary tumor virus glycoproteins andnonglycoproteins. Virology 85:168-174.

23. Scott, J. V., L. Stowring, A. T. Haase, 0. Narayan, and R.Vigne. 1979. Antigenic variation in visna virus. Cell18:321-327.

24. Shapiro, S. Z., M. Strand, and J. R. August. 1976. Highmolecular weight precursor polypeptides to structuralproteins of Rauscher murine leukemia virus. J. Mol. Biol.107:459-477.

25. Stowring, L., A. T. Haase, and H. P. Charman. 1979.Serological definition of the lentivirus group of retrovi-ruses. J. Virol. 29:523-528.

26. Thormar, H., and B. Sigurdardottir. 1962. Growth ofvisna virus in primary tissue cultures from various animalspecies. Acta Pathol. Microbiol. Scand. 55:180-186.

27. Van Zaane, D., M. J. Dekker-Michielsen, and H. P.Bloemers. 1976. Virus-specific precursor polypeptides in

VOL. 42, 1982

1056 VIGNE ET AL.

cells infected with Rauscher leukemia virus: synthesis,identification and processing. Virology 75:113-129.

28. Vigne, R., M. Brahic, P. Filippi, and J. Tamalet. 1977.Complexity and polyadenylic acid content of visna virus60-70S RNA. J. Virol. 21:386-395.

29. Vigne, R., P. Filippi, M. Brahic, and J. Tamalet. 1978.

Absence of circularly permuted and largely redundantsequences in the genome of visna virus. J. Virol. 28:543-550.

30. Vigne, R., J. C. Neil, M. L. Breitman, and P. K. Vogt.1980. Recovered src genes are polymorphic and containhost markers. Virology 105:71-85.

J. VIROL.