0022-538X/83/050538-09$02.00/0
Copyright © 1983, American Society for Microbiology
Deletion
Mutants
of Moloney Murine Leukemia Virus Which
Lack
Glycosylated
gag
Protein Are Replication
Competent
PAMELASCHWARTZBERG, JOHN COLICELLI,ANDSTEPHEN P.GOFF*
Departmentof Biochemistry, Columbia University, College of Physicians and Surgeons, Cancer Research
Center,
New York,New York10032
Received 28 December 1982/Accepted23February 1983
Aseries of deletion mutations localizednearthe 5' end of the Moloney murine
leukemia virusgenome was generated by site-specific mutagenesis of cloned viral
DNA. Themutants recoveredfrom such deleted DNAsfailed to synthesize the
normal glycosylatedgagprotein gPr80Wag.Twoof the mutants made no detectable
protein, and a third mutant, containing a66-base pair deletion, synthesized an
altered gag protein which was not glycosylated. All the mutants made normal
amounts of the internal PrO5gag protein. The viruses were XC positive and
replicated normally in NIH/3T3 cells as well as in lymphoid cell lines. These
results indicate that the additional peptides ofthe glycosylated gag protein are
encoded near the 5' end, that the glycosylated and internal gag proteins are
synthesized independently, andthatthe glycosylated gagprotein isnot required
during the normal replication cycle. In addition, the region deleted in these
mutants apparently encodesno cis-acting functionneeded for replication. Thus,
all essentialsequences, including those for packaging viralRNA, must lie outside
thisarea.
The structure of the Moloney murine
leuke-miavirus(M-MuLV)genome is known ingreat
detail; indeed, the complete nucleotidesequence
hasbeen determined (32). The detailed functions carriedoutby the
products
of thegag,pol,
andenv genes, however, are largely unknown (5).
For
example,
although it is known that thegagproteins are structural elements in the virion
particle,
the actual role of each matureprotein
(termed P15, P12, P30, and P10) is uncertain.We
have begunagenetic
analysis
of thegenomeof M-MuLVinanattempt todetermine the various functions of theproteins
and regulatoryse-quences
required
in the viral lifecycle.
The first 620 nucleotides of the viral RNA
(extending
from thecapped
5' end totheAUGcodon utilized for the synthesis of the Pr65gag
protein) are required for several functions
im-portant tothereplication ofthevirus. Contained
inthisregionare(i)thebinding site for
tRNAPro,
whichprimesDNAsynthesis (26);(ii)thesplicedonorsite for the formation ofenvmRNA; (32;
J.
Champoux, personal
communication); (iii) signals for packaging of viral RNA into virion particles (30);and(iv)aregion ofunknownsize encoding N-terminal peptides ofthe glycosylat-ed gag protein termed gPr80gag (7). The region fromthe splice donor site(atnucleotide205) tothe start of
PrO5gag
(at nucleotide 621)consti-tutes the largestregion ofthe virus with
uncer-tain function. Work on avian retrovirus mutants
(30)hasshown thatpartofthe
region
isessential for RNApackaging,
but theextent of theputa-tive recognition site is unknown. In the caseof M-MuLV, a large part of this RNA must also encode the
glycosylated
gagprotein
gPr80gag.
Thisproteincontains
essentially all of thepep-tides of Pr65gag and
approximately 4,000
to10,000 daltons of additional
protein
at the Nterminus (7); additional amino acids may be utilized as a signalpeptide
directing
theprotein
to the endoplasmic reticulum fortransport and glycosylation. Mannose-rich core
oligosaccha-ride chainsareaddedtotheprotein,
probably
inthe P15 and P30regions (7), andarethen
modi-fied furtherbytheaddition of
complex
oligosac-charides. The
protein
istransported
to the cell surface, and a portion may be shed from thecells (6). The AUG sequence at which
protein
translation starts is unknown. The function of this protein is equally unclear,
although
it is known to constitute the GCSAantigen
on thesurface of infectedcells;aroleinvirion
matura-tion has also beenproposed.
Many laboratorieshavegenerated mutants of leukemiaviruses which wereblockedatvarious
stages ofthe lifecycle (1, 9, 13, 20-22, 27, 31,
36, 40, 43); these mutantshavebeen
extremely
useful in definingthe variousfunctions carried
outby the virus. Theavailability of
fully
infec-tious DNA clones of the M-MuLV genome
makespossible theconstruction of defined
dele-538
on November 10, 2019 by guest
http://jvi.asm.org/
tion mutations at known sites within genes.
Thus, it should be possible to mutateany gene by modifying cloned DNA, to recover virus
fromcellstransfectedwith the altered DNA, and toobserve the effect ofthemutationon thelife
cycleofthe virus. To definemorepreciselythe
extentandpositionof thesequences neededfor
viral genomic packaging and for synthesis of
gPr809'9,
wehavecarriedoutsite-specificmuta-genesis ofacloned DNAcopy ofthe M-MuLV
RNAgenome. Surprisingly, several of the mu-tantDNAs gave riseto fully infectious replica-tion-competent virus,eventhough no
glycosyl-atedgag proteinwasmade.
MATERIALS AND METHODS
Cells and virus. NIH/3T3 fibroblastsweregrownin
Dulbecco modified Eagle medium (GIBCO
Labora-tories) supplementedwith10% calfserum(Flow
Labo-ratories,Inc.).M-MuLVreleased byNIH/3T3 clone4
wasthesourceofinfectious wild-type virus. This cell
linewasderivedas asingle-cell clone of NIH/3T3 cells
infected at lowmultiplicity (<0.05) with the clone 1 strain of M-MuLV(11). Viralinfectionsathigh
multi-plicity were carried out in the presence of 8 ,ug of
Polybrenepermlat37°Cfor 2h,and theinfectedcells
were then maintained in 1.6 jig ofPolybreneper ml
overnight.LymphocytecelllinesSWR/4 (34)and 70Z
(25) were grown in RPMI 1640 medium (GIBCO
Laboratories) supplementedwith10% fetal calfserum.
Cloned DNAs. Plasmid p8.2 contains a full-length permutedcopyof M-MuLVcircular viralDNA, joined
attheuniqueHindIll site in thecenterof thegenome
and containing only one copy of the long terminal
repeat(LTR).Theclonewaspreparedby subcloning
theHindIll viral DNAfragmentof lambda8.2(33)into
theHindIlI site ofplasmid pBR322 bystandard meth-ods.
PlasmidpZAPcontainsafull-lengthproviralcopyof M-MuLV plus approximately 8 kilobases offlanking rat sequencessubcloned from phage lambdainto the EcoRI site ofplasmid pBR322(15).PlasmidpSV2GPT
(gift from R. Mulligan) contains theEscherichia coli
gptgenelinkedtoasimian virus 40promoter(24).
DNAenzymology.Partial restrictionenzymedigests
ofplasmidDNAswerecarriedout at20°Cover arange
of time periods (5 min to 1 h). Full-length linear molecules were isolated by electrophoresis in 0.6% agarose (SigmaChemicalCo.) gelcastinTEA buffer
(50mMTris-hydrochloride, 1 mMEDTA,aceticacid
topH 8.0) containing1 ,ugof ethidiumbromideperml
and were purified bythe glass powder method (41).
This linear DNAwasdigestedwith Bal 31exonuclease
at25°Cfor 20sinbuffer(0.3MNaCl,6mMCaC12,6
mM MgCl2, 10 mMTris-hydrochloride [pH 8.0], 0.5
mM EDTA). The reaction was quenched with the
additionof EDTA toafinalconcentrationof20mM,
and the DNAwaspurified bysuccessive extractions
withphenoland ether andprecipitationwithethanol.
The DNAwascircularized with T4ligase (New
En-glandBioLabs)ataDNA concentration of1 Fg/mlin
buffer (50 mM Tris-hydrochloride [pH 7.5], 10 mM
MgCl2,10 mMdithiothreitol,1 mM ATP).
Bacterial cultures. E. coliHB101(4)wasgrownin L broth(10goftryptone[DifcoLaboratories]perliter,5
g of yeast extract [Difco] per liter, 5 g of NaCl per liter,
1 mM NaOH) and was transformed to ampicillin
resistance by the CaCl2 method (4). Selective plates and media contained 50 1tg of ampicillin per ml.
Plasmid DNA wasisolated from small cultures (2 ml)
by the rapid preparation of Holms and Quigley (16)
andfrom large cultures (200 ml) by equilibrium
centrif-ugation incesiumchloride-ethidium bromide (17).
DNAsequence analysis. Mutant DNAs (1 to 5 ,ug)
were cleaved with BstEII (New England BioLabs),
and the terminal 5' phosphate was removed by
treat-ment with calf intestinal phosphatase (Boehringer
Manheim Corp.) in buffer (10 mM Tris [pH 8.3], 10
mMMgCl2) for 1 hat37°C. The DNAwaslabeled with
polynucleotide kinase (Bethesda Research
Labora-tories) and [-y-, 32P]ATP (Amersham Corp.) and was
recleaved with Sacl (New England BioLabs). The
end-labeled DNA fragment containing the mutation
was isolated by electrophoresis on a10o
polyacryl-amide gel andsubjected tothechemical degradation
protocol of Maxam and Gilbert (23). Cleavage
prod-ucts weredisplayedon8%polyacrylamide-8 M urea
gels(34 by 14cm) andexposedtoX-ray film (Kodak
XR5, Eastman KodakCo.)withanintensifyingscreen
(Lighting Plus, Du Pont Co.)at -70°C.
Mammalian cellular transfections and
transforma-tions.Cloned mutated viral DNAswereexcised from
the pBR322vectorplasmid by cleavage with HindlIl
(Bethesda Research Laboratories) andwerereligated
at a concentration of 100 to 500 ,ug/ml to ensure
extensive concatemerization. Plasmid DNAs pZAP
andpSV2GPTwereusedascircular DNAs.NIH/3T3
cells were transfected with the DNAbythe calcium
phosphate precipitation method (14, 35) with salmon
sperm DNA as the carrier. Transfections for XC
plaque assayswereperformed by precipitating0.5 ,ug
of cloned DNA and6,ugofcarrier inavolumeof 0.5
ml andapplyingthemto6-cm dishescontaining
10'
to1.5 x 10- cells per dish for4 to 6 h. At3 days after
transfection, cells were split 1:10, replated in 10-cm
dishes, and allowedtogrowtoconfluency.The
pres-ence of viable virus was determined by overlaying
withXC cells (29) after UV irradiation of the
mono-layer.
Cotransfections with plasmidpSV2GPTDNAwere
carriedoutessentially in the same fashion.Amixture
of1 ,ugof viral DNA,1 ,ugofpSV2GPT, and 12 ,ug of
carrierwasappliedto3 x 105to5 x 105cells ona
10-cmplate. At2 days after transfection, cells were fed
with XAT medium (15 mM hypoxanthine, 0.2 mM
aminopterin,5 mMthymidine, 250 mM xanthine, 150
mMglutamine, 5 mM glycine,25 mMmycophenolic
acid).Cellswererefed every 5days, and colonies were
cloned 3 weeks later. The presence ofreverse
tran-scriptase activity in the media was assayed by the
methodofGoffet al. (13).
Viralproteinanalysis.Cells (2 x 105per6-cm dish)
wereinfected with filtered virus harvested from clones
of transfectedmutants.After1day, cellswerepassed
1:10 into 10-cmdishes. Whenmonolayers were half
confluent,thecells werestarved for0.5 hin Dulbecco
modified Eagle medium minus methionine, and
[35S]methionine (200 ,uCi;AmershamCorp.)was
add-edfor15min.The cells werelysed,and theproteins
wereimmunoprecipitated with goat anti-M-MuLV
se-rumorrabbitanti-gagserum507(giftofD.Baltimore)
by standard procedures. Immunoprecipitated proteins
on November 10, 2019 by guest
http://jvi.asm.org/
were analyzed by sodium dodecyl
sulfate-polyacryl-amide gel electrophoresis and fluorography as de-scribedpreviously (18, 39).Theproteinsweretreated
withendo-p-N-acetylglucosaminidase(endoH;gift of
P.Robbins)asdescribed(38).
RESULTS
Construction of deletion mutants incloned
M-MuLV DNA. The DNA used as the parent
molecule for all the mutagenesis described in
this work was plasmid p8.2, an infectious
full-lengthcopyof the M-MuLVgenomeinplasmid
pBR322. This clonewasprepared by cleavage of
proviral circular DNAatthesingle HindIII site
before cloning; thus, the M-MuLV insert was
circularly permuted relative to the usual linear
form of the genome. A map of the insert is
shown in Fig. 1.
Mutantswereconstructedby deleting
particu-lar restriction endonuclease recognition sites
and part of the flanking DNA. To direct these
deletionstotheregion encoding the5'end of the
viral RNA, recognition sites for the enzymes
PstI and PvuIwereselectedastheinitialtargets.
These enzymes made cleavages in the area of
interest,although they also cleave elsewhere in
the M-MuLV genome and in the plasmid
pBR322 vector DNA. Plasmid p8.2 DNA was
partially digested with either PstIor PvuI, and
full-length linears (cleaved only once)were
iso-lated fromanagarosegel.This DNAwastreated
with Bal31 exonuclease to remove a fewbase
pairs from each end. It was next purified and
recircularizedbyligationatalowconcentration.
The modified DNA was used to transform E.
coliHB101toampicillin resistance, and 12
resis-tantcolonies werepicked and grown into small
culturesfor eachstartingenzyme.Plasmid DNA
isolated from each culture was analyzed by
agarose gel electrophoresis after cleavage with
PvuII and Sacl to determine the position and
approximate size of the deletion.
Approximately one-third to one-half of the
plasmids examined contained deletions in M-MuLV sequences. Two ofthe DNAs
mutagen-izedbycleavage with PstI contained deletions of
the entire 176-base pair(bp) PstI fragment
be-tween 1.0 and 1.15 map units and were not
examined further. A fewmutants had deletions
atthePstI and PvuI sites inplasmidpBR322,but
neverthelesstheymusthaveencoded functional
ampicillinaseenzymes.Thoseplasmids
contain-ing small deletions mappcontain-ingateither the PvuI or
PstIsitenearthe5' end of theM-MuLVgenome
wereamplified in E. coli HB101 and purifiedon
CsClgradients for further analysis.
Mapping and sequencing of mutants. Three
mutantscontaining small deletions in the sites of
interestwere selected forfurther study. Two of
these mutants, termed d1902 and dl904, were
prepared by cleavage with PvuI; mapping
showed that they had lost the PvuI site at
nucleotide position421andapproximately 75bp of flankingDNA. The third mutant,d11003,was
prepared from PstIlinears and had lost the PstI siteat position 565as wellas approximately 50 bpnearby.Inallthree cases,theremaining PvuI and PstI sites wereintact.
The nucleotide sequences ofthe mutated
re-gions were determined as described above. Briefly, each mutant DNA was cleaved with BstEII, the 5' ends were labeled with 32p, and
the DNAs were recutwith Sacl. Maxam-Gilbert
cleavage reactions (23) were performed on the
fragment containingthesiteof the deletion, and
thesequence wasdetermined byelectrophoresis
ofthe DNA fragments. The numbers and
posi-tions of the bases missing in each mutant are
shown in Fig. 1. Mutants d1902 and
d1904
con-tained deletions of 56 and 83 bp, respectively,
and mutant d11003 contained a 66-bp deletion
veryclose to the start ofPr65gag. The sequences
obtained in this region agree with the data of Shinnicket al.(32),exceptforonebase
substitu-tion(aG to Achangeatnucleotide 617)near the
AUGfor Pr659'9 (Fig. 1).
Biological activity. Plasmid p8.2is apermuted
cloneofM-MuLV containing onlyone LTR. To
study theinfectivity ofthe mutagenized clones,
eachDNA wascleaved withHindIlI and ligated
intopolymerstoreconstruct acompleteproviral
structure.DNAs wereappliedtosensitive NIH/
3T3 fibroblasts in acalcium phosphate
precipi-tate, and the cells were tested for release of
viable virusby overlay with XC cells. All three
deletion mutants, dl902,
d1904,
and dl1003, formed XCsyncytia
similar in appearance and number to those formed by the wild-type viral clone p8.2. Media from these plates containedmeasurable reverse transcriptase activity (data
not shown).
Clonal NIH/3T3 cell lines producing virus wereisolated bycotransfection withtwoDNAs, namely plasmid pSV2GPT DNA carrying the
selectablemarkerof resistancetomycophenolic
acid, and each of the polymerized mutant
DNAs. Cells expressing the gpt gene were
se-lected in XAT medium(seeabove), and several
mycophenolic acid-resistant colonies were
cul-tured and tested forreversetranscriptase
activi-ty. Between one-third and all of the colonies
exposedto DNAsofthe mutants werepositive
inthis assay(Fig. 2). To test thetransmissibility
of theseapparentlyviableviruses, medium from
the reverse transcriptase-positive clones was
filtered and used to infect new NIH/3T3 cells.
Cellsinfected in thismanneralsoproducedvirus
asmeasuredbythe XC assay, with titers in the
range of
104
to105
PFU/ml(data notshown).Insummary, these viruses renderedNIH/3T3
on November 10, 2019 by guest
http://jvi.asm.org/
r J
VOL. 46, 1983
t0(I 00
_ 0
O 0 N N
go%3 N' = 0N
co 3Q
O-0 O
r Co r'° N
0
<N
_. Coo Co O
a. Z
3°-U) o'
N 0
-z 3* n
ft CFO;
o
n-0
OcX o r_
0Q
CD >
B
0 0o
CA 0 w
o 0.eN,
000 0
C)
r--*0.
N CA
x N
om
C 'O
ON oN 0.
<,o3_<
on November 10, 2019 by guest
http://jvi.asm.org/
1 2 3 4 5 6 7
pZAP * @ * *
di902*
209*
di904 *
d11003 Control
FIG. 2. Reverse transcriptase assays of
supema-tant fluids from clones of cells transfected by various DNAs. Each viral DNA was oligomerized, mixed with
plasmid pSV2GPT DNA, and applied to NIH/3T3
cells. Individual clones resistant to mycophenolic acid
were isolated, and the supernatantfluid was assayed
for reverse transcriptase on an exogenous template (13). Results for several independent clones derived
fromeach transfection are shown.
cells XC positive upon transfection, caused
themtoreleasereverse transcriptase-containing virions, and weretransmissible tofibroblasts.
Growth of mutant viruses in lymphocytes. A
number of strains of murine leukemia
viruses,
including the Grossvirusand a BALB/c
endoge-nousstrain, replicate poorly inlymphocyte cell lines and alsoactas poorhelpersfor the stable
transmission of Abelson MuLV into such cells (28). To examine whether these deletions could
cause such a
phenotype,
twolymphocyte
lineswereinfected with wild-typeor mutantvirusat a
multiplicity ofapproximatelyone, and the virus
released by these cells was monitored over
several days.In both70Z,aB-cell leukemia cell
line, and SWR/4, an Abelson MuLV-induced
(pre-B-cell) leukemiacellline,the mutant virus-es replicated normally and were secreted into
the medium at least as wellas wild-type virus
(Fig. 3). Thus,theregiondeleted isnotessential
for replication in theselymphocytes.
Analysis
ofgag proteins. Cells infected withthemutantviruseswere
pulse-labeled
for 15 minwith [35S]methionine, and extracts of protein
wereprepared foranalysis. Viralproteinswere
immunoprecipitated with avariety ofsera, and
theimmune
complexes
were collectedby
bind-ing to Formalin-fixed Staphlococcus aureus.
Proteinswereexaminedbysodiumdodecyl
sul-fate-polyacrylamide gel electrophoresis
fol-lowedby fluorography (Fig. 4). Wild-typevirus
(in NIH/3T3 clone 4 cells) clearly directed the
synthesisoftwo
gag-specific
proteins,
theinter-nalPr65Rar andtheglycosylated
gPr8g9a9
(which migratedslightly
faster than theenvelope
pro-tein gPr8Oe"v). Cells containing mutants d1902
andd1904 showed
only
theinternal gagprotein
andnodetectablelarger
species
evenuponlong
overexposure ofthe fluorograph. The cells
in-fected with mutant
d11003, however,
synthe-sized PrO5gag and a
large
amount of a novelprotein migrating at
approximately 71,000
dal-tons,detectedby
gag-specific
sera(Fig. 4).
Thedifference inapparent molecular massbetween the wild-type gPr80gag and the
dl1003
mutantIL E
IL
0
ccI3
0
a
SWR/4
70
Z2 4 6 8 10 2 4 6 8 10
Days Post
Infection
FIG. 3. Growth curves of various M-MuLV isolates inlymphocyte cell lines SWR/4(A) and 70Z(B). The
indicated cell lines were infected with virus on day zero, and supernatantfluidswere collected on successive
days. PFU in thesesupernatantswere titrated by XC assay in NIH/3T3 cells.Mock-infected cultures released no
PFU. Virus isolates used were clone 4 virus(0), d1902 virus (0), d1904 virus (O), and d11003 virus (A).
J. VIROL.
on November 10, 2019 by guest
http://jvi.asm.org/
[image:5.490.90.209.77.178.2] [image:5.490.104.391.430.641.2]VIABLE DELETIONS OF M-MuLV 543
N N
O
0
° Oa7) CY 0) _
_ _ _ _
tl-' rl--
ll-0 0 0
Z U Z 0 Z )
g
Pr
80e9-o
Pr6
5Q9_-
Pr 71904
Pr
65
9g9
4
1 2 3 4 5 6 7 8 9 10 11 12
FIG. 4. Analysis of viralgagproteins synthesized in infected NIH/3T3 cells. Cell populationsinfected with
various M-MuLVpreparations werelabeled with [35S]methionine, and thelabeled proteinswereanalyzed by
immunoprecipitation,sodium dodecyl sulfate gel electrophoresis, and fluorography. Lanes1through4:Proteins
fromcells infected withwild-typeM-MuLV(clone4cells)wereprecipitatedwithnormalgoatserum(lane 1),
rabbitanti-gagserum507 (lane 2),goatanti-Moloneyserum(lane3),andgoatanti-RauscherPr60ga serum(lane
4).Thepositionsof thewild-type proteins Pr65gag andgPr80ga areindicated. Lanes5through12:Cells infected
withmutantsd/902(twoisolates), dl904, and d11003,asindicated,wereanalyzedwithnormalgoatserum(N)or rabbitanti-gagserum507 (507). The positions ofproteins Pr65gag and Pr7lgagareindicated.
protein was approximately 9,000 daltons and
wastoo greattobeaccounted for by the 66-bp
deletion. One possible explanation for this
dis-crepancy is that the deleted protein may also
lack thecarbohydrate foundonnormalgPr8099a.
To determine whether thenovel protein
con-tained carbohydrate, the immunoprecipitated
proteins were digested with endo H, which
cleaves mannose-richcarbohydrate sidechains,
and the treated proteinwas analyzed as before
(Fig. 5). Wild-typegPr8099a wasreduced in size
by endo H treatment, migrating as a protein of
approximately 73,000 daltons in these gels. The
mutant d11003 protein was completely
un-changed by endo Htreatment(Fig. 5). Thus,the
mutant protein appeared to lack the normal
glycosylation foundonthewild-type large
mem-brane-bound gagprotein.
Alltheseprotein phenotypeswerecompletely
stable for many cell generations (data not
shown). Thus, reversionorrecombination with
endogenous viruses did notinterfere with these
assays.
DISCUSSION
The site-specific mutations described in this
work were constructed by Bal 31 exonuclease
treatment of cloned viral DNA linearized by
restriction enzymes. This method provided a
rapid and reliable means for introducing
muta-tions at selected sites throughout the viral
genome: approximately 80%oofplasmids
isolat-ed after the mutagenesis contained deletions.
From suchtreatmentofclonep8.2 DNA,
linear-ized with PstI and PvuI, we obtained three
mutantscontaining small deletionsin theregion
between the5' LTR and theinitiation AUG for
the major gag and gag-polprecursor proteins.
Several larger deletions which were not viable
werealso obtained (datanotshown).
When these threeDNAswereappliedtoNIH/
3T3 cells in a calcium phosphate precipitate,
viable virus wasproduced, asdemonstrated by
XCplaqueassayandbythepresenceofreverse
transcriptase activity.When NIH/3T3 cellswere
cotransfected with viral DNAs and plasmid
0
:
.Lto-01
O-
2 (rZo
M Z 0OVOL.46, 1983
4 .
A
on November 10, 2019 by guest
http://jvi.asm.org/
[image:6.490.60.435.70.351.2]544 AND GOFF
Wild Type di1003
Endo H M
g PrB Pr4O"n
j- kP Pr659og
FIG. 5. Analysis of viralgagproteinssynthesized
bymutantdl1003 virus. Immunoprecipitates from cells
infected with wild-type M-MuLV or mutant d11003
were prepared with rabbit anti-gag serum 507. The
immunoprecipitates were solubilized, and portions
were mock treated or treated with endo H before
analysis by sodium dodecyl sulfategelelectrophoresis
and fluorography. Lane 1, wild-type M-MuLV
pro-teins without endo H; lane 2, wild-type M-MuLV
proteins with endoH treatment; lane 3, dl1003
pro-teins without endo H; lane 4, dl1003 proteins with
endoH; lane 5, marker proteinsproduced by
immuno-precipitation ofproteinsinclone 4 cellswithgoat
anti-M-MuLV serum.
pSV2GPT,amajority of the clones surviving in
XAT medium produced reverse transcriptase
activity in the supernatant medium, indicating
the presence of released viral particles. In this
way,producer cell lineswerereadily isolated. It
should be noted, however, that the percentage
of clones cotransfectedmaybeanoverestimate,
since we cannotrule outthepossibility of virus
spreadontheoriginal transfection plate. Each of
these clones produced fully transmissible virus,
which could be filtered and passedtonewNIH/
3T3 cellsto render them XCpositive.
Murine leukemia virusescontainanunusually
long leader sequence(620 bpin thecase of
M-MuLV) between the 5'end of the viral RNA and
the initiation codon for thePr658agand
Pr2009'9-PO/
precursors (32). We have shown thatdele-tions within an approximately 200-bp region
apparently do notaffect viral functions inNIH/
3T3 fibroblasts. Thus, we conclude that the
deleted regions containnocis-acting sequences
required for viral replication. R. Mann has
re-cently isolatedoneM-MuLV mutantcontaining
a 350-bp deletion in this region (personal
com-munication); thismutantisdefective and unable
to transmit stably its own genome to infected
cells. Because our mutantsare not impaired in
replicationfunctions,wecanfurthernarrowthe
window in which these control sequences are
located toa region close to the 5' end;
accord-ingly,we caneliminatemost ofthe region close
togag.
Itisparticularly strikingthat mutantd11003is
viable. This deletion eliminates sequences upto
19bp away from the initiation codon forPrO5gag
without affecting protein translation ofthis gene. Apparently, elimination of these sequences
causes no adverse effect on ribosomal binding
for Pr65gag and
Pr200gag-po
protein translation;the signals for recognition of this AUG by the
ribosome, if they exist at all, must lie very close
tothe gene.
Analysis of gPr8OIga
proteip.
The gPr80gagproteincouldbe readilydetected in cells
infect-edwith wild-type
virus,
but none of thisproteincouldbe detected in cells infected withmutants
dl902 and d1904. Longexposures of the
fluoro-grams showed that the protein was present at
less than 5% of wild-type levels. The gels also
showed that the serum wasveryspecificfor gag
proteins and did not precipitate the closely
mi-grating gPr8oe"v protein. These results suggest
thatgPr80gag is notnecessary for replication or
transmission of M-MuLV in NIH/3T3
fibro-blasts. Furthermore, because these mutants
eliminate
gPr80gag,
they confirm that this region contains sequences necessary for the synthesis of thisprotein.Immunoprecipitation of gag proteins from
mutant dllO03 revealed a protein
migrating
atapproximately 71,000 daltons; we interpreted this protein to be a shortened
gPr809'9.
Theobserved shift inmobilitywastoo great,
howev-er, to be accounted for by a 22-amino-acid
deletion. Since
gPr80gag
is a glycosylatedpro-tein, the large shift in mobility of the mutant
protein might result fromthe combination ofa
lackof carbohydrate side chains andashortened polypeptide. When the mutant dl1003 protein
was treated with the enzyme endo H, which selectively removes the mannose-rich side chains, nochange in mobility wasdetected. In
contrast, wild-typegPr809a9was reduced in
ap-parentsizebyendoHtreatment to
approximate-ly 73,000 daltons, onapproximate-ly slightapproximate-ly larger than the
mutant protein. The difference in the apparent
size ofthedeglycosylated wild-typeproteinand thedl1003 mutantproteinis small: about 2,000
daltons,
very close to the difference predicted from a 66-bp DNA deletion. Thisgoodagree-ment suggests that the polypeptide moiety of
both proteins begins and ends at identical
se-quences and that the only difference is the
internally
deleted region of dl1003.Previous studies (7)
tentatively
map the sitesof
glycosylation
inwild-type
gPr80g'g
to the region shared with themajor
gagproteins
P30andP15,far downstreamfromthe
position
of thedeletion in d11003. This mutation must, there-fore,eliminate
glycosylation
of sitesataconsid-erable distance. We consider two
possibilities
J. VIROL.
on November 10, 2019 by guest
http://jvi.asm.org/
[image:7.490.80.219.74.221.2]+
f
10) to leave a larger translation product. TheA
gP_
_gr8o0°
d11003
protein, despite
itsdeletion,
would beR US 1_ Pr65°° largerthanwild-type
gPr80g'g
ifaleaderpeptide
TY" Ii -- of 30 amino acids were not removed. Thus,
B Pr71 g
gPr809ag
may resemble the case of ovalbumin,_P15
P12P6
p9P0
5 whichcontainsaninternalsignal
sequenceneed-dl U ed for membrane transport and which does not
undergo
cleavage
ofaleaderpeptide
(19).
C P P P0_
Pr659°
Structure of the gene forgPr8O'g.
Thedele-dIO2SU5 v tion mutantsdescribed inthispaper
help
definedl
W4the
region
which encodes theglycosylated
gag,base
protein of M-MuLV. The extra amino acids [image:8.490.50.242.58.202.2]° 500 1000 1500 found in
gPr80gag
and notinPr65gag
havebeenFIG. 6. Proposed structure of RNA genomes localized by peptide analysis (7) to the amino
(lines) and protein products (hatched boxes) of various terminal segment of the longer protein. The
M-MuLVisolates. (A) Two gag proteins,gPrgosasand
simplest interpretation
of the structure ofPr65gag,
are synthesized by wild-type M-MuLV. The gPr80gag isthat translation is initiatedatasite instructure of the RNA template for the synthesis of the leader RNA andcontinues in phase into the
gPr8.g2g
is unknown. Carbohydrates are added to sequence of the regular Pr65gag(Fig. 6). ThreegPr80gag at portions corresponding to P15 and P30 in AUG initiator methionine codons are found in
the maturecleavage products of
Pr659a8.
(B) Proteins the published sequence (32) upstream of thesynthsizeby mtantdllO0 arePr7ga adP68. gag
synthesized bymutantd11003are
Pr7.
a andp5 . AUG for the Pr65 ,butnone of these codonsThe66-bp deletion in thisgenome causes ashortening is in thesametranslational
reading
frame astheofthelargergagprotein withoutaffectingPr658a8 The gag
deletion also abolishes the addition of the carbohy- Pr65 . It is possible that proper translational
drate chainsusually found at the downstream sites.(C) initiation or RNA
splicing
sites have not beenOnly Pr65gag is synthesized by mutants
d1902
and identified. A lesslikely possibility
is that thisdl904; production of the larger protein is completely sequence may be translated starting from a
abolished by the deletions. valine GTG codon like that found in procaryotic
proteins
orfromanentirely
different codon.Possible protein function. Our work
provides
which could accountfor theeffect of this muta- thefirst
example
ofmutantsin theglycosylated
tion. Onepossibility
isthat thedeletion indl1003
membrane-bound gagprotein.
Previous efforts alters the structure of thisprotein
in a manner to isolate such mutantsby
immunoselection ofwhich decreases the
accessibility
oftheglycosy-
infected cells withanti-gag
sera and comple-lation sites. On the average,only
one-third of mentlysis
haveonly
resulted in isolation of the classictripeptide
sites(-Asn-X-Thr/Ser)
are cellularmutants defective in membraneprotein
found
glycosylated
in membrane-bound and se-processing
(8, 12).
Theproduction
of viablecreted
proteins,
and thus it has beenproposed
virusby
someof these cell linesdespite
the lack(37) that the addition of
carbohydrate
side chains ofaglycosylated
gagprotein
on theircell sur-isdependentonthetertiary
structure surround- facesis in agreement with ourfindings.
ing the
glycosylation
site.Adeletion of22amino Theconstruction of viable deletionmutantsof acids may wellchange
the conformation of M-MuLV which lackgPr80gag provides
conclu-gPr80gag
so as toeliminateglycosylation.
Alter- sive evidencethatthisprotein
isnotrequired
fornatively,
the deletion indl1003
may interferereplication
and transmission in NIH/3T3 fibro-with asignal
sequence whichnormally
directs blasts or inlymphocytes.
No known function thisprotein
to themembrane. The mutantpro- has yet been associated with theglycosylated
tein may not be translated on the
rough
endo- gagprotein,
but theprotein
may beinvolvedinplasmic reticulum,
andaccording
to current somelittle-understood function of leukemogene-modelsfor thesynthesis
of membraneproteins,
sis. Theavailability
ofmutants ingPr80gag
willit would not be glycosylated or
transported
to facilitate a critical evaluation ofpossible
rolesthe cellmembrane. Weare
currently testing
the for thisprotein
in the transformation of lym-localization of theprotein by
surfaceprotein
phoid cells in vivo.iodination with
lactoperoxidase
andby
surface immunofluorescent stains.If the
dl1003 protein
haslost part ofasignal
sequence
normally
tagging
it as a membraneprotein, it must have one unusual property.
Normally, short leader
peptides
of about 30amino acids are cleaved from membrane
pro-teins
during secretion,
and mutations in thisleader block
glycosylation
andcleavage (2, 3,
ACKNOWLEDGMENTS
This work was supported by a grant from the National CancerInstitute(5R01CA-30488)andbyanawardfromthe Hirschl Foundation.
LITERATURECITED
1. Alevy,M. C.,and P. K.Vogt. 1978. Tspolmutantsof avian sarcoma viruses: mapping and demonstration of
single-cyclerecombinants.Virology87:21-23.
46,
on November 10, 2019 by guest
http://jvi.asm.org/
2. Bedouelle, H.,P. J.Bassford, Jr., A. V. Fowler, I. Zabin, J. Beckwith,and M. Hofnung. 1980. Mutations which alter the function of the signalsequence of the maltosebinding protein ofEscherichia coli. Nature (London) 285:78-81. 3. Blobel, G., and B. Dobberstein. 1975.Transferof proteins
across membranes. I. Presence ofproteolytically proc-essed and unprocproc-essed nascent immunoglobulin light chains onmembrane-bound ribosomes ofmurine myelo-ma. J. CellBiol. 67:835-851.
4. Boyer, H. W., and D. Roulland-Dussoix.1969. A comple-mentationanalysis of the restriction andmodification of DNA in E.coli. J. Mol. Biol. 41:459-472.
5. Dickson, C., R. Eisenman, H. Fan, E. Hunter, and N. Teich. 1982. Proteinbiosynthesis and assembly, p. 513-648. In R. Weiss, N. Teich, H. Varmus, and J. Coffin (ed.), RNA tumor viruses, 2nd ed. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.
6. Edwards, S.A., and H. Fan. 1979.gag-Related
polypro-teins of Moloney murine leukemia virus: evidence for independentsynthesis ofglycosylated andunglycosylated forms. J.Virol. 30:551-563.
7. Edwards, S.A., and H. Fan. 1980.Sequencerelationship
ofglycosylated andunglycosylated gagpolyproteinsof Moloneymurine leukemia virus. J.Virol. 35:41-51. 8. Edwards, S.A., and H. Fan. 1981.Immunoselection and
characterization of Moloney murine leukemia virus-in-fected celllines deficientin surfacegagantigen expres-sion. Virology 113:95-108.
9. Elsenman, R. N., W. E. Mason, and M. Linial. 1980. Synthesis andprocessing ofpolymeraseproteins of wild-type andmutant avianretroviruses. J. Virol. 36:62-78. 10. Emr, S. D., J.Hedgpeth, J. M. Clement, T. J.Silhavy,and
M. Hofnung. 1980. Sequenceanalysisof mutations that prevent export oflambda-receptor, an Escherichia coli outermembraneprotein. Nature(London) 285:82-85. 11. Fan, H., and M. Paskind. 1974. Measurement of the
sequencecomplexity of clonedMoloney murine leukemia virus 60 to70S RNA: evidence for a haploid genome. J. Virol.14:421-429.
12. Fittig,T., M.Ruta, and D. Kabat. 1981. Mutant cells that abnormallyprocess plasma membraneglycoproteins en-coded bymurineleukemiavirus. Cell 24:847-858. 13. Goff,S., P. Traktman, and D. Baltimore. 1981. Isolation
and properties of Moloney murine leukemia virus mu-tants: use of a rapid assay for release of virionreverse transcriptase.J. Virol. 38:239-248.
14. Graham, F. L., and A. J. van der Eb. 1973. A new technique forthe assayofinfectivityof human adenovirus 5 DNA.Virology 52:456-471.
15. Hoffmann, J.W., D. Steffen, J.Gusella,C.Tabin, S. Bird, D. Cowing,and R. A.Weinberg.1982. DNAmethylation
affecting theexpressionof murine leukemiaproviruses.J. Virol. 44:144-157.
16. Holmes, D. S., and M. Quigley. 1981. A rapid boiling
method for thepreparation of bacterial plasmids. Anal. Biochem. 114:193-197.
17. Katz, L., D. T. Kingsbury, and D. R. Helinski. 1973. Stimulation bycyclic adenosinemonophosphate of
plas-middeoxyribonucleic acid replicationandcatabolite re-pression of the plasmid deoxyribonucleic acid-protein
relaxationcomplex. J. Bacteriol. 114:557-591.
18. Laemmli, U. K. 1970. Cleavage of structural proteins
during the assembly of the head of bacteriophage T4. Nature(London)227:680-685.
19. Lingappa, V. R., J. R. Lingappa, and G. Blobel. 1979. Chicken ovalbumin gene contains an internal signal se-quence. Nature (London) 281:117-121.
20. Linial, M., and W. S. Mason. 1973. Characterization of two conditional early mutants of Rous sarcoma virus. Virology53:258-273.
21. Mason, W., and C. Yeater. 1977. A mutant of Rous sarcoma virus with aconditional defect in the determi-nants of viralhost range. Virology77:443-456. 22. Mason, W. S., C. Yeater,T. V. Bosch, J.A.Wyke, and
R. R.Fris. 1979. Fourteentemperature-sensitive
replica-tion mutants ofkoussacoma virus.Virology99:226-240.
23. Maxam, A., and W.Gilbert. 1979. Sequencing end-labeled DNA with base specific chemical cleavages. Methods Enzymol. 65:499-599.
24. Mulligan, R. C., and P. Berg. 1981. Selection for animal cells that express the Escherichia coli genecoding for xanthine-guanine phosphoribosyltransferase. Proc. Natl. Acad.Sci.U.S.A. 78:2072-2076.
25. Paige, C. J., P. W.Kincade, and P. Ralph. 1978. Murine B cellleukemia line withinduciblesurfaceimmunoglobulin expression. J. Immunol. 121:641-647.
26. Peters, G.,andJ. E.Dahlberg. 1979.RNA-directed DNA synthesisin Moloney murine leukemia virus: interaction between theprimer tRNA and the genome RNA. J. Virol. 31:398-407.
27. Rohrschneider, J. M., H. Diggelman, H. Ogura, R. R. Friis, and H. Bauer. 1976. Defectivecleavageof a precur-sorpolypeptide in a temperature-sensitive mutant of avian sarcomavirus.Virology 75:177-187.
28. Rosenberg, N., and D. Baltimore. 1978. The effect of helper virus onAbelsonvirus-induced transformation of lymphoid cells. J. Exp. Med. 147:1126-1141.
29. Rowe, W.P., W. E. Pugh, and J. W.Hartley.1970. Plaque assaytechniques for murine leukemiaviruses. Virology 42:1136-1139.
30. Shank, P. R., and M. Linial. 1980. Avian oncornavirus mutant(SE21Q1b) deficient in genomic RNA: character-ization of a deletion in the provirus. J. Virol. 36:450-456. 31. Shields, A.,0. N.Witte, E. Rothenberg, and D. Baltimore. 1978. Highfrequency ofaberrant expression of Moloney
murine leukemia virus in clonal infections. Cell 14:601-609.
32. Shinnick, T. M.,R. A.Lerner, and J. G.Sutcliffe. 1981. Nucleotide sequenceof Moloney murine leukemia virus. Nature(London) 293:543-548.
33. Shoemaker, C., S. Goff,E. Gilboa, M. Paskind, S.W.
Mitra, and D. Baltimore. 1980. Structure of a cloned circularMoloney murine leukemia virus DNA molecule containing an inverted segment: implications for retro-virus integration. Proc. Natl. Acad. Sci.U.S.A. 77:3932-3936.
34. Silverstone, A. E., N. Rosenberg, V. L. Sato. M. P.
Schneid,E. A.Boyse,and D.Baltimore. 1978. Control of cellproliferation. vol. 5. ColdSpring Harbor Laboratory, ColdSpringHarbor, N.Y.
35. Smotkin, D., A. M. Gianni, S. Rozenblatt, and R. A. Weinberg.1975.Infectious viral DNA of murine leukemia virus. Proc.Natl. Acad. Sci. U.S.A. 72:4910-4913. 36. Stephenson, J. R.,andS.A.Aaronson.1973.
Character-ization oftemperature-sensitive mutants of murine leuke-miavirus.Virology 54:53-59.
37. Struck,D.K.,and W.J. Lennarz. 1980. Thefunction of
saccharide-lipidsinsynthesisofglycoproteins,p.35-80. InW. Lennarz (ed.), Thebiochemistryofglycoproteins
andproteoglycans.PlenumPublishingCorp.,NewYork. 38. Tarentino, A. L., and F. Maley. 1974. Purification and properties ofan endoB-N-acetyl-glucosaminidase from Streptomyces griseus.J.Biol. Chem.249:811-816. 39. Traktman, P.,and D. Baltimore.1982.Proteasebypassof
temperature-sensitive murine leukemia virus maturation mutants.J.Virol. 44:1039-1046.
40. Tronikc,S.R., J.R.Stephenson,I. M.Verma,andS.A. Aaronson.1975. Thermolabilereversetranscriptase ofa mammalian leukemia virusmutanttemperaturesensitive in itsreplication and sarcomavirus helperfunctions. J. Virol. 16:1476-1482.
41. Vogelstein, B., and D. Gillespie. 1979. Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad.Sci. U.S.A. 76:615-619.
42. Wensink,P.C.,D.J.Finnegan, J.E.Donelson,andD. S. Hogness. 1974.AsystemformappingDNAsequencesin the chromosomes of Drosophila melanogaster. Cell 3:315-325.
43. Wong,P. K. Y., L.J.Russ,andJ.A.McCarter. 1973. Rapid selective procedure for isolation of spontaneous temperature-sensitive mutants ofMoloney leukemia vi-rus. Virology51:424-431.
J. VIROL.