JOURNAL OF VIROLOGY,
Copyright ? 1997, American Society for Microbiology
Mar. 1997, p. 2083–2091 Vol. 71, No. 3
Altered Rous Sarcoma Virus Gag Polyprotein Processing and
Its Effects on Particle Formation
YAN XIANG,1TODD W. RIDKY,1NEEL K. KRISHNA,2AND JONATHAN LEIS1*
Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio 44106,1and
Department of Microbiology and Immunology, Pennsylvania State University College of Medicine,
Hershey, Pennsylvania 170332
Received 10 July 1996/Accepted 26 November 1996
Proteolytic processing of the Rous sarcoma virus (RSV) Gag precursor was altered in vivo through the
introduction of amino acid substitutions into either the polyprotein cleavage junctions or the PR coding
sequence. Single amino acid substitutions (VP2S and PP4G), which are predicted from in vitro peptide substrate
cleavage data to decrease the rate of release of PR from the Gag polyprotein, were placed in the NC portion
of the NC-PR junction. These substitutions do not affect the efficiency of release of virus-like particles from
COS cells even though recovered particles contain significant amounts of uncleaved Pr76gagin addition to
mature viral proteins. Single amino acid substitutions (AP3F and SP1Y), which increase the rate of PR release
from Gag, also do not affect budding of virus-like particles from cells. Substitution of the inefficiently cleaved
MA-p2 junction sequence in Gag by eight amino acids from the rapidly cleaved NC-PR sequence resulted in
a significant increase in cleavage at the new MA-p2 junction, but again without an effect on budding. However,
decreased budding was observed when the AP3F or SP1Y substitution was included in the NC-PR junction
sequence between the MA and p2 proteins. A budding defect was also caused by substitution into Gag of a PR
subunit containing three amino acid substitutions (R105P, G106V, and S107N) in the substrate binding pocket
that increase the catalytic activity of PR. The defect appears to be the result of premature proteolytic
processing that could be rescued by inactivating PR through substitution of a serine for the catalytic aspartic
acid residue. This budding defect was also rescued by single amino acid substitutions in the NC-PR cleavage
site which decrease the rate of release of PR from Gag. A similar budding defect was caused by replacing the
Gag PR with two PR subunits covalently linked by four glycine residues. In contrast to the defect caused by the
triply substituted PR, the budding defect observed with the linked PR dimer could not be rescued by NC-PR
cleavage site mutations, suggesting that PR dimerization is a limiting step in the maturation process. Overall,
these results are consistent with a model in which viral protein maturation occurs after PR subunits are
released from the Gag polyprotein.
Translation of retroviral genomic RNA on free ribosomes in
the cell cytoplasm leads to synthesis of two polyproteins, Gag
and Gag-Pol, which migrate to the plasma membrane. This
transport is directed by a signal located at the amino terminus
of Gag referred to as the membrane binding (M) domain. In
mammalian viruses, this signal includes an N-terminal myristic
acid (4). In avian viruses, the signal is contained within MA
(32). The viral env gene products are translated concomitantly
on the rough endoplasmic reticulum-associated ribosomes.
They are modified posttranslationally during transport through
the reticulum to the plasma membrane, where particles bud
from the cell surface (18). Aggregation of Gag polyproteins
(Gag and Gag-Pol) is necessary to form particles. This process
is driven in part by the M domain concentrating Gag on the
membrane but is assisted by an internal Gag interaction (I)
domain that in Rous sarcoma virus (RSV) maps to the NC
protein (28, 32). Efficient budding is also dependent on a
domain required late in the budding process, referred to as the
L domain (33). In RSV, the L domain maps to a proline-rich
sequence, PPPPYV, in p2 of Gag. This sequence probably
interacts with a cell protein containing a WW domain (35), a
conserved motif that has two tryptophan residues separated by
20 to 22 amino acids. The formation and release of immature
virus particles from cells do not require proteolytic processing,
incorporation of viral RNA, or pol or env gene products. How-
ever, the presence of the latter normal components alters the
characteristics of particles released.
During or after the budding process, polyproteins are
cleaved by a virus-encoded protease (PR) to release mature
proteins (18). Concomitant with processing, there are morpho-
logical changes in the appearance of particles released from
cells, and virions become infectious (11, 15, 19, 26). This is the
rationale for the development and use of antiprotease drugs in
the treatment of AIDS (12, 17, 23). An active PR is composed
of a dimer of two identical subunits, and Gag and Gag-Pol
polyprotein processing requires PR to recognize at least nine
unique cleavage sites of eight amino acids each (7, 9, 16, 23, 24,
27, 34). The molecular basis for this remarkable specificity is
dependent primarily on optimization of van der Waals inter-
actions between amino acid side chains in the polyprotein
cleavage site and key amino acid residues in the enzyme’s
substrate binding pocket (7, 23, 24).
In avian retroviruses, the PR subunit is found at the carboxyl
terminus of Gag (Fig. 1). Cleavage at the NC-PR boundary is
believed to be catalyzed by PR itself (3, 5) and seems to be
required before processing of other viral protein junctions in
Gag occurs (6). The timing of this activation is tightly regulated
and appears to coincide with the process of virus budding.
When budding is prevented by disruption of the Gag M do-
main, the level of polyprotein processing is greatly diminished
(2, 28). Particle formation is also aborted if PR is activated
* Corresponding author. Mailing address: Department of Biochem-
istry, Case Western Reserve University, 2119 Abington Rd., Cleveland,
OH 44106. Phone: (216) 368-3360. Fax: (216) 368-4544. E-mail: jx18
prematurely, before Gag reaches the plasma membrane (2, 5).
This is most likely because premature cleavage separates the
Gag proteins from the M-domain transport signal. Similar re-
sults were obtained with processing of human immunodefi-
ciency virus type 1 (HIV-1) Gag polyproteins (20).
Mechanisms which delay PR activation until budding are not
yet understood. To investigate this process of virus maturation,
we have introduced a series of amino acid changes into Gag
polyprotein cleavage sites that are predicted from cleavage
data derived from peptide substrates to either increase or
decrease their cleavage rates. Despite the fact that cleavage of
the NC-PR site in vivo might be influenced by accessibility
and/or conformation of the site in the polyprotein, we find that
there is a correlation between the steady-state kinetic cleavage
data of the NC-PR peptides in vitro and cleavage of the altered
sites in vivo. Analysis of these sites, coupled with mutations
that change the intrinsic enzymatic properties of PR, has re-
vealed that there is also a correlation between the rate of viral
polyprotein processing and particle budding. Moreover, these
data further support a model in which PR must be released
from the Gag polyprotein before processing of other Gag
cleavage sites occurs.
MATERIALS AND METHODS
Reagents. All reagents were as previously described (33). Oligodeoxynucleo-
tides were purchased from Midland Certified Reagent Company (Midland, Tex.)
and used directly for mutagenesis. The wild-type RSV gag gene is from pATV-8,
an infectious molecular clone of the RSV Prague C strain. The plasmid
pSV.Myr0 is a simian virus 40-based mammalian expression vector carrying a
wild-type copy of the RSV gag allele, the product of which efficiently directs
production of virus-like particles from COS-1 cells (2, 28, 30, 32, 33, 35).
Oligodeoxynucleotide-directed mutagenesis. Site-directed mutagenesis of the
RSV gag gene was carried out by overlap extension mutagenesis as described by
FIG. 1. Substitutions in the RSV Gag polyprotein. The RSV Gag polyproteins are represented by rectangular boxes. The vertical lines inside the boxes represent
cleavage sites between the different proteins. A series of substitutions that change PR76gagare listed below the boxes and were expressed from mutant plasmids
constructed by overlap extension mutagenesis (1) using oligodeoxynucleotides listed in Table 1. (A) Single amino acid substitutions in the NC-PR cleavage site. The
eight-amino-acid sequence representing the NC-PR cleavage site is expanded, and the positions relative to the scissile bond are indicated by P4 to P4?. The full-length
black line represents the wild-type (WT) NC-PR cleavage site in Gag. Amino acid substitutions introduced into the P4 to P1 positions are as indicated and are aligned
with the expanded NC-PR cleavage site sequence. (B) Substitutions that change the sequence of the MA-p2 cleavage site to that of the NC-PR cleavage site. The amino
acid sequence of the MA-p2 cleavage site is expanded, and the amino acid substitutions are indicated below. The one-amino-acid difference between MA-p23NC-PR
and MA-p23NC(PP4G)-PR, MA-p23NC (AP3F)-PR, or MA-p23NC(SP1Y)-PR is indicated in boldface. (C) Substitutions in PR which alter its activity. Part of the
amino acid sequence of RSV PR is expanded, and the numbers below indicate positions relative to the amino terminus of mature PR. Amino acid substitutions are
as indicated. (D) Constructs that combine the RSV PR(S105-7) and NC-PR cleavage site substitutions. The altered PR is shown as PR* in Gag. All the other
representations are as in panel A. (E) Deletions in Gag. The shaded areas represent deleted sequence. In the ?NC construct, most of NC has been removed with the
exception of seven amino acids from the amino and carboxyl termini. The positions of two copies of the I domain are also indicated. In the ?(GSGL) construct, the
p2-p10 cleavage junction is expanded. The shaded amino acids have been deleted; this region includes the junction between p2 and p10 (reverse arrow). Other notations
are as in panel A. (F) Constructs that combine a linked PR dimer with NC-PR cleavage site substitutions. The four-glycine linker is shown between two PR subunits.
All the other representations are as in panel A.
2084 XIANG ET AL.J. VIROL.
Aiyar et al. (1). A list of mutations made is presented in Fig. 1. In most cases, the
mutations create a new restriction enzyme site to facilitate identification of
clones containing the desired mutation (Table 1). The mutant ?NC was created
by oligodeoxynucleotide-directed mutagenesis using single-stranded MGAG
template DNA, containing the RSV gag gene (32), and a synthetic oligonucleo-
tide having the sequence 5?-GCAGTAGTCAATAGAGAGAGGCCTGAGCC
ACCTGCCGTC-3?. The deletion removes all of the NC coding sequence except
that for the first and last seven residues. The mutation was inserted into the
mammalian expression plasmid by transferring the BglII-BssHII fragment of gag.
The presence of the mutation was confirmed by DNA sequencing, and multiple
clones of the mutant were characterized in transfection experiments.
Transfection of mammalian cells. COS-1 cells grown in Dulbecco’s modified
Eagle’s medium supplemented with 3% fetal bovine serum and 7% calf serum
(HyClone Inc.) were transfected by the DEAE-dextran-chloroquine method as
described previously (30). Plasmid DNAs, at a concentration of 25 ?g/ml, were
digested with XbaI and incubated with T4 DNA ligase before transfection. This
removes the bacterial plasmid sequence and joins the 3? end of the gag gene with
the simian virus 40 late polyadenylation signal (30).
Metabolic labeling and immunoprecipitations. In most experiments, cells in
35-mm-diameter dishes were labeled 48 h after transfection with L-[35S]methi-
onine (1,000 Ci/mmol, 75 ?Ci/ml of tissue culture medium) for 2.5 h at 37?C as
described previously (30). The cells or growth medium from each labeled culture
was mixed with lysis buffer containing protease inhibitors. Rabbit antiserum
directed against whole RSV (reactive with MA, CA, NC, and PR Gag proteins)
was added, and immunoprecipitation was carried out for 2 h at 4?C. Precipitates
were collected by using protein A-agarose (Gibco-BRL), and the proteins were
subjected to electrophoresis (35). Pulse-chase experiments were carried out with
sets of identically transfected cells incubated in methionine-free medium for 30
min, pulse-labeled with [35S]methionine for 15 min, and chased with a 1,000-fold
excess of cold methionine in serum-free medium for the indicated times. Den-
sitometric quantitation of the fluorograms was carried out by using the Bioanaly-
sis program on a Sci-Scan 5,000 (U.S. Biochemicals).
Isopycnic sucrose gradient analysis. COS-1 cells expressing the ?NC mutant
were radiolabeled, and the particles released into the medium were subjected to
isopycnic sucrose gradient analysis as previously described (2). Unlabeled Molo-
ney murine leukemia virus (MoMLV) was added to the gradient as an internal
control and detected by reverse transcriptase activity.
SDS-polyacrylamide gel electrophoresis and fluorography. Immunoprecipi-
tated proteins were separated by electrophoresis in sodium dodecyl sulfate
(SDS)-polyacrylamide gels containing acrylamide monomer and N,N?-methyl-
enebisacrylamide at a ratio of 30:1.2. Resolving gels contained 12% acrylamide,
0.1% SDS, and 400 mM Tris-HCl (pH 8.8), while the stacking gels contained 3%
acrylamide, 0.1% SDS, and 60 mM Tris-HCl (pH 6.8). After electrophoresis, gels
were fixed in a solution of 7.5% methanol and 7% acetic acid. Radiolabeled
proteins were detected by fluorography using Kodak BIO MAX MR film at room
temperature. Overnight exposures were typically required.
In vitro protease assay. As described previously (23), RSV protease activity
was assayed in a volume of 25 ?l of 100 mM sodium phosphate (pH 5.9), 2.4 M
sodium chloride, 100 ?M peptide substrate, and 50 ng of PR. Reactions were
initiated by the addition of protease, incubated at 37?C for 3 to 15 min, and
stopped by the addition of 300 ?l of 0.5 M sodium borate (pH 8.5). Twenty
microliters of 0.05% (wt/vol) fluorescamine was then added. Relative fluores-
cence intensity was measured on a Perkin-Elmer LS-50B luminescence spectro-
photometer, using an excitation wavelength of 386 nm and an emission wave-
length of 477 nm. Each activity measurement represents the mean of at least
three independent experiments. In each case, the standard error for all experi-
ments did not exceed 20% of the value reported. Kinetic constants were deter-
mined by using the assay described above. Initial rate data from substrate satu-
ration curves were fit to the Michaelis-Menten equation by using the NFIT
program. Correlation coefficients of the fit were greater than 0.98, and the
standard deviation of the constants reported was less than 20%.
Peptides. Peptide substrates, 9 to 12 amino acids in length and representing
various RSV Gag polyprotein cleavage sites, were prepared as described previ-
ously (24). These peptides include the eight amino acids, P4 to P4?, required for
efficient and specific cleavage by the retroviral protease. Peptides were solubi-
lized in 1 mM dithioerythritol, and their concentrations were determined by
quantitative amino acid composition analysis. The presence of Pro on the amino
terminus renders the substrate nonreactive to fluorescamine; the inclusion of Arg
residues on the carboxyl terminus improves solubility of the peptides (24).
Purification of soluble bacterially expressed mutant RSV proteases. Polyhis-
tidine-tagged RSV protease was expressed in Escherichia coli M15 pDM1.1 as
previously described (24). The active RSV protease is greater than 95% pure as
judged by SDS-polyacrylamide gel electrophoresis and was obtained in a final
yield of about 2 mg. The RSV PR(S105-107) and the covalently linked dimer
protease were prepared as described for wild-type PR.
Single amino acid substitutions in the NC-PR cleavage site
alter overall Gag processing. It was reported previously that a
small deletion introduced into the RSV NC-PR junction of
Gag blocked processing of all Gag cleavage sites (6). While this
result suggests that processing of Gag cleavage sites is depen-
dent on the release of PR from the polyprotein, this experi-
ment could not exclude the possibility that the deletion pre-
vented dimerization and activation of the PR subunits (27, 36).
To examine more closely the relationship between cleavage of
the NC-PR junction and initiation of viral polyprotein process-
ing, we introduced single amino acid substitutions into the NC
side of the NC-PR cleavage junction. These mutations are
predicted from in vitro cleavage data (Table 2) to either in-
crease or decrease the ability of the sequence to be cleaved by
PR. The effects of these substitutions were analyzed in vivo by
expressing the Gag alleles in COS-1 cells in the presence of
[35S]methionine, immunoprecipitating the viral proteins from
both the cell lysate and cell media, and fractionating the pre-
cipitated protein by SDS-polyacrylamide gel electrophoresis
(see Materials and Methods). The viral proteins detected in
the media are contained in virus-like particles released from
the cell surface.
Substitution of Gly for Pro in the P4 or Ser for Val in the P2
position of the NC-PR cleavage site results in an order of
magnitude decrease in the ability of wild-type PR to cleave this
TABLE 1. Mutagenic oligodeoxynucleotides
Sequence of mutagenic oligodeoxynucleotideb
MA-p2 3 NC-PR(P4)5?ACACCTAAAACCGTTGGCGGTGCTGTAAGCTTAGCGATGACAGCTATTGGCTGTAAT3?
MA-p2 3 NC-PR(P3)5?ACACCTAAAACCGTTGGCCCTTTCGTAAGCTTAGCGATGACAGCTATTGGCTGTAAT3?
MA-p2 3 NC-PR(P1)5?ACACCTAAAACCGTTGGCCCTGCCGTCTACCTAGCGATGACAGCTATTGGCTGTAAT3?
aSubstitutions in the Gag NC-PR and MA-p2 cleavage junctions are defined as in Fig. 1.
bOligodeoxynucleotides used to introduce substitutions into the RSV Gag NC-PR and MA-p2 junctions.
cNew restriction sites were introduced to facilitate identification of mutated clones.
VOL. 71, 1997ACTIVATION OF RSV POLYPROTEIN PROCESSING 2085
sequence in vitro (Table 2). Introduction of either of these
substitutions into the gag allele resulted in a decrease in the
processing of all Gag polyprotein cleavage sites. This is de-
tected most readily in the medium fraction (Fig. 2B), as seen by
the accumulation of Pr76gagand the concomitant decrease in
mature CA and PR bands (Fig. 2B; compare lanes 2, 3, and 6
with lane 1). This is also observed in the lysate fractions. While
Gag processing was decreased, the overall rate of particle
release from cells into the medium fraction was not affected.
This is expected, because an active PR is not required for virus
budding (31; see also Fig. 4B, lane 7). In a separate set of
experiments, the ability to cleave the NC-PR junction was
increased by substitution of Phe for Ala in P3 or Tyr for Ser in
P1 (Table 2). The AP3F substitution resulted in slightly faster
processing of Pr76gagin the lysate but again had little effect on
the overall particle release. In the case of the SP1Y substitu-
tion, there was a significant increase in processing of Pr76gagin
the lysate. This is seen by the dark band migrating with a
molecular mass of 63 kDa (Pr63). This band corresponds to
that expected for a Gag polyprotein lacking the PR subunit
(Fig. 2A, lanes 7 and 8). Despite the fact that this substitution
increased the ability to cleave the NC-PR cleavage site as a
peptide substrate, the overall amount of particles released
from cells was again similar to that of the wild type. Particles
released from cells under these conditions, however, contained
Pr63 in addition to fully processed proteins. These results
support a model in which PR that dimerizes within the struc-
ture of the polyprotein cleaves initially at the NC-PR junction.
PR subunits must be released from the viral polyprotein before
processing occurs at other polyprotein sites.
Replacing the MA-p2 cleavage junction in Gag with the
amino acid sequence from the NC-PR cleavage junction. To
determine whether cleavage of the NC-PR site and subsequent
PR subunit release from Gag play a unique role in processing,
we substituted the same set of NC-PR cleavage site sequences
for the natural site between the MA and p2 proteins (Fig. 1B).
The MA-p2 cleavage site is one of the slowest-cleaved sites in
Gag and Pol analyzed both in vitro and in vivo (8). As a
consequence, the MA protein is usually detected in virus-like
particles released from cells as an MA-p2 fusion of approxi-
mately 23 kDa (Fig. 3B, lane 1). When the more efficiently
cleaved NC-PR is substituted for the MA-p2 cleavage junction,
processing at this site is significantly increased in vivo. This is
seen by the appearance of mature MA and the concomitant
FIG. 2. Single amino acid substitutions in the NC side of the NC-PR cleavage site alter the overall processing of Gag. All labeling and protein detection were done
as described in Materials and Methods. The positions of migration of Pr76gagand its cleavage products are indicated on the left. Substitutions in the NC-PR cleavage
site of Gag are indicated above the lanes and are defined in Fig. 1A. In almost every case, two independent clones for each mutation were constructed and analyzed.
WT, wild type.
TABLE 2. Steady-state kinetic activities of various RSV PRs on
Wild type PAVS-LAMT
aSteady-state kinetic parameters were determined with the wild-type, RSV
PR(S105-107), and PR-GGGG-PR covalently linked proteases, using peptide
substrates as described in Materials and Methods.
bAll peptides contain two Arg residues not shown on the carboxyl terminus.
PAVS-LAMT and TSCY-HCGT represent the NC-PR and MA-p2 cleavage
sites, respectively. Lines under residues indicate changes from wild-type se-
cThese peptides contain an additional Pro residue on the amino terminus (not
2086XIANG ET AL. J. VIROL.
loss of the MA-p2 fusion band in the virus-like particles (Fig.
3B, lane 5). The doublets observed with MA-containing bands
are due to partial phosphorylation (14, 21) of the protein in
COS cells. While this site is processed more efficiently, there is
no detectable loss of particles released from cells. Moreover,
when the mutation was cloned into the virus genome and
analyzed in vivo, infectious virus particles, which spread
throughout the cells in culture, were obtained (29).
The same single amino acid substitutions in NC that in-
creased or decreased the rate of cleavage of NC-PR were then
introduced into the substituted site at the MA-p2 junction. In
contrast to the data shown in Fig. 2B (lanes 2 and 3), substi-
tution of Gly in P4 did not prevent the processing of Gag (Fig.
3B, lane 4), nor did it affect the release of virus-like particles
from cells. Thus, it is the substitution of Gly in P4 of NC-PR
and the release of the PR subunits from the polyprotein that
were responsible for the change in processing observed in Fig.
2. The combination of replacing MA-p2 with the NC-PR se-
quence and modifying its P1 or P3 position to increase further
PR cleavage resulted in a significant decrease in virus-like
particles released from cells (Fig. 3B, lanes 2 and 3) compared
to the wild type (lane 1). The loss of budding with the substi-
tutions in P1 and P3 may be the result of the premature
cleavage of the Gag polyprotein. Alternatively, it is possible
that the P1 or P3 substitution directly influenced the budding
process even though neither did so in the NC-PR junction at
the wild-type position. Also, the wild-type MA-p2 junction
sequence is not required for budding since it can be replaced
by the NC-PR sequence (Fig. 3B, lane 5).
Gag polyproteins with modified PR subunits cause defects
in budding. The relationship between Gag processing and for-
mation of virus-like particles was further studied by introduc-
ing an altered PR subunit containing substitutions R105P,
G106V, and S107N into the Gag allele (Fig. 1C). This modified
PR, termed RSV PR(S105-107), cleaves the RSV NC-PR and
MA-p2 peptide substrates 9- and 50-fold faster, respectively,
than the wild-type PR (Table 2). The PR(S105-107), like the
wild type, must dimerize to form an active enzyme. When the
PR(S105-107) substitution was introduced into Gag, a de-
FIG. 3. Moving the NC-PR cleavage site sequence to the junction between
the MA and p2 proteins in Gag. The NC-PR wild-type (WT) and altered cleav-
age site sequences have been moved to the junction between the MA and p2
proteins in Pr76gagas shown in Fig. 1B. Specific changes are indicated above the
lanes. Viral proteins from the cell lysate and media were analyzed as described
in the legend to Fig. 2. Migration positions of viral proteins are indicated on the
FIG. 4. Gag polyproteins with modified PR subunits have budding defects resulting from increased Gag processing inside the cell. Substitutions that alter the
sequence of the PR subunit in Pr76gagare shown in Fig. 1C and are indicated above the lanes. A four-amino-acid deletion (GSGL), which disrupts the p2-p10 junction
in Gag, is as in Fig. 1E. Viral proteins from the cell lysate and media were analyzed as described in the legend to Fig. 2. Migration positions of viral proteins are indicated
on the left. WT, wild type.
VOL. 71, 1997ACTIVATION OF RSV POLYPROTEIN PROCESSING2087
crease in the level of virus-like particles released from cells was
observed (Fig. 4B; compare lane 5 to lane 4). The defect can be
rescued by inactivating the PR(S105-107) by substitution of a
serine for the catalytic aspartic acid residue (Fig. 4B, lane 6). In
addition to the defect in particle release, maturation of the
carboxyl terminus of CA was incomplete. This is seen by the
predominant CA1 band (22), representing CA plus a nine-
amino-acid spacer sequence (Fig. 4A, lane 4).
The budding defect caused by the PR(S105-107) substitu-
tion can be rescued by slowing the rate of cleavage at the
NC-PR junction. If cleavage of the NC-PR site is a prerequisite
for PR-dependent processing of Gag as indicated in Fig. 2,
then introduction of amino acid substitutions at the NC-PR
cleavage site which slow down the release of the PR subunit
are predicted to rescue defects observed with the RSV
PR(S105-107) (Fig. 1D). When the RSV PR(S105-107) substi-
tution was combined with either the NC (VP2S)-PR or the
NC(PP4G)-PR substitution, wild-type levels of virus-like parti-
cles were released from transfected cells (Fig. 5B; compare
lanes 1, 2, 7, and 8 to lane 9). Both of these substitutions are
FIG. 5. The defect in budding observed with PR(S105-7) can be rescued by slowing down the rate of cleavage of the NC-PR junction in Gag. Constructs that
combine the PR and NC-PR cleavage site substitutions are shown in Fig. 1D and are indicated above the lanes. ?(p2b) deletes the 11 amino acids defined as p2b in
Fig. 1A. This deletion disrupts the L assembly domain as previously described (34). Viral proteins from the cell lysate and media were analyzed as described in the
legend to Fig. 2. Migration positions of viral proteins are indicated on the left. WT, wild type.
FIG. 6. Defects in particle release observed with a covalently linked PR dimer in Gag cannot be rescued by substitutions to the NC-PR cleavage site. Constructs
that combine the linked PR dimer with the NC-PR cleavage site substitutions are as shown in Fig. 1F and are indicated above the lanes. Viral proteins from the cell
lysate and media were analyzed as described in the legend to Fig. 2. Migration positions of viral proteins are indicated to the left of the panels. WT, wild type.
2088XIANG ET AL. J. VIROL.
predicted to lower the rate of cleavage of the NC-PR junction
(Table 2). Interestingly, while release of virus-like particles
from cells was normal, removal of the spacer peptide from the
end of CA was still defective (Fig. 5B; compare lanes 1, 2, 7,
and 8 to lane 9). When the rate of cleavage of the NC-PR
junction was increased by the introduction of either the
NC(SP1Y)-PR or the NC(AP3F)-PR substitution, a budding
defect that was more severe than that observed with RSV
PR(S105-107) by itself was seen (Fig. 5B, lanes 3 to 6). As a
control, we analyzed the ability of a small deletion mutation
that inactivates the p2-p10 cleavage junction to rescue the
budding defect observed with the RSV PR(S105-107) enzyme
(Fig. 1E). The ?(GSGL) deletion interferes with cleavage at
the p2-p10 junction (Fig. 4B, lane 3), resulting in a p33 band
representing the MA-p2-p10 proteins detected in released par-
ticles from Gag alleles with wild-type PR. The ?(GSGL) de-
letion does not disturb the biological function of the adjacent
L domain, a conserved amino acid sequence (PPPPYV) re-
quired for budding of virus particles from cells (33, 35). Thus,
wild-type levels of virus-like particles are released from cells
(Fig. 4B, lane 3). When the ?(GSGL) deletion was combined
with the RSV PR(S105-107) substitution, there was no rescue
of the particle defect (Fig. 4B, lane 1). Thus, altering a Gag
polyprotein junction other than NC-PR has little effect on PR
activation. Taken together, these results indicate that the bud-
ding defect caused by PR(S105-107) is dependent on release of
modified PR subunits from the Gag polyprotein.
Defective particle release observed with a covalently linked
PR dimer cannot be rescued by substitutions in the NC-PR
cleavage site. A budding defect is observed when a PR dimer
linked by four glycine residues is introduced into the Gag allele
(5). This enzyme, purified from the soluble fraction of bacterial
cell lysates, has a specific activity with the reference NC-PR
peptide substrate similar to that seen with wild-type PR (Table
2). However, in contrast to the results shown with RSV
PR(S105-107) in Fig. 5, decreasing the cleavage rate of the
NC-PR site by introduction of either the NC(VP2S)-PR or the
NC(PP4G)-PR substitution (Fig. 1F) does not rescue the as-
sembly defect observed with the linked dimer (Fig. 6B, lanes 1
to 4). Increasing the cleavage rate of the NC-PR junction also
had little detectable effect (Fig. 6B, lanes 5 to 8). These results
suggest strongly that initial cleavage at the NC-PR junction,
which is a requisite for PR dimerization, is the rate-limiting
step in PR activation and subsequent polyprotein processing.
Combining RSV PR(S105-107) and covalently linked PRs
with deletions of the Gag I domain. The I domain in Gag
promotes aggregation of polyproteins during particle assembly
(2, 10, 28) and is therefore predicted to influence the process
of PR dimerization in the polyprotein. To test this hypothesis,
we made use of mutant ?NC (Fig. 1E), which lacks Gag resi-
dues 496 to 570, including both copies of the I domain. Cells
FIG. 7. The I domains of Gag can facilitate dimerization of viral PR. (A)
Expression of ?NC. Cells transfected with the indicated DNAs were metaboli-
cally labeled as described in Materials and Methods. WT, wild type. (B) Particle
density analysis of ?NC. Forty-eight hours posttransfection, COS-1 cells were
labeled for 8 h in 1 ml with leucine-free, serum-free Dulbecco’s medium sup-
plemented with 500 ?Ci of L-[4,5-3H(N)]leucine. The sample was spiked with
unlabeled MoMLV (as an internal control [2, 28] for density of wild-type retro-
virus particles) and loaded onto a 10 to 50% sucrose gradient, which was spun in
an ultracentrifuge at 70,000 ? g for 16 h. Fractions containing MoMLV (å) were
identified by assaying for reverse transcriptase (RT) activity. ?NC proteins (F)
from each fraction were immunoprecipitated by using anti-RSV serum, subjected
to electrophoresis on an SDS–12% polyacrylamide gel, and visualized by fluo-
rography. The optical density (O.D.) of each band was determined by densitom-
etry. The arrow indicates the direction of sedimentation. (C) Pulse-chase exper-
iments with [35S]methionine were carried out as described in Materials and
Methods in cells expressing Gag polyproteins containing PR(S105-7) (?),
PR-PR (E), PR(S105-7) plus the ?NC mutation (I), or PR-PR plus the ?NC
mutation (F). The ?NC mutation is shown in Fig. 1E. The percentage of
detectable labeled Gag was graphed as a function of chase time.
VOL. 71, 1997ACTIVATION OF RSV POLYPROTEIN PROCESSING 2089
transfected with this mutant released particles less efficiently
than the wild type (Fig. 7A). Unlike previously described I-do-
main mutants of RSV (2, 28), ?NC contains complete CA and
PR sequences. Since these proteins participate in interactions
within the virion, we determined whether the ?NC clone pro-
duced particles of low rather than high density, the criterion
used to define the effects of I-domain mutations (2, 28). Anal-
ysis of ?NC particles in sucrose density gradients revealed that
the particles were of low density and thus behaved as expected
for I-domain mutants (Fig. 7B).
We then introduced the ?NC deletion into Gag constructs
containing either RSV PR(S105-107) or covalently linked PR.
Pulse-chase experiments were conducted to characterize the
stability of Gag in whole-cell extracts, including released par-
ticles (Fig. 7C). Under these conditions, Gag containing RSV
PR(S105-107) or any of the covalently linked enzymes had an
estimated half-life of 0.75 h. Introduction of the I-domain
deletion into Gag containing the RSV PR(S105-107) enzyme
increased the half-life of Pr76gagabout twofold, to 1.5 h,
whereas there was little effect when the deletion was intro-
duced in combination with the covalently linked PR. While the
differences observed are small, the same magnitude of effect
was observed in two separate pulse-chase experiments. These
results are consistent with the postulated role of the I domain
in promoting Gag polyprotein interactions which facilitate viral
Processing of viral polyproteins by retroviral PR is a late
event in replication which is required for production of infec-
tious virus. If processing occurs too early, Gag proteins are
separated prematurely from the amino-terminal M-domain
transport signal. This prevents them from reaching the site of
virus assembly and results in failure to release particles. The
exact mechanism by which PR activation is delayed is not fully
understood. A model presented in Fig. 8, which depicts this
process, is based on previous as well as present studies. Pre-
cursor processing appears to be initiated by release of PR
subunits from the polyprotein. While a RSV PR dimer, as part
of the Gag polyprotein, is capable of cleaving the NC-PR
junction, processing of all other Gag cleavage sites requires
release of PR subunits from the polyprotein. This is shown by
the results of Burstein et al. (6) and findings reported here
where inefficient cleavage of mutant NC-PR junctions delays
subsequent processing of all other Gag cleavage sites. These
mutations were placed in the NC side of the NC-PR junction
and do not alter the catalytic properties of PR. Furthermore,
introduction of a catalytically enhanced PR [RSV PR(S105-
107)], which has to dimerize, or a covalently linked PR dimer,
which does not, results in premature processing of the Gag
polyprotein. However, the budding defect created by the RSV
PR(S105-107) is rescued by slowing release of the altered PR
from Gag, whereas the defect caused by a covalently linked PR
is not rescued. Similarly, it is reported that in HIV, precursor
forms of PR are less efficient in Gag processing (37). Precursor
processing is dependent on release of the PR subunit, probably
because of the requirement for PR dimerization. This process
occurs slowly in the polyprotein but rapidly with free PR sub-
units. This observation is supported by recent studies of Sellos-
Moura and Vogt (25) which demonstrate that recombinant
RSV NC-PR fusion and PR proteins migrate on gels as mono-
mer and dimer, respectively.
In the processing model depicted in Fig. 8, a small amount of
PR as part of the polyprotein has to dimerize to initiate the
autoproteolytic processing. Whether the initial PR cleavage
occurs in cis, with the PR dimer cleaving itself out of its own
polyproteins, or in trans, with the PR dimer cleaving PR sub-
units from different Gag polyproteins, is not known. However,
structural considerations make the cis cleavage seem unlikely.
The initial dimerization process would be facilitated both by
aggregation of the precursor polyproteins at the site of budding
and through protein-protein interactions involving the I do-
main (2, 28). It has been shown that in HIV, NC and p6 can
facilitate PR dimerization (38). In this report, we find that an
I-domain deletion, which interferes with Gag aggregation, can
decrease the processing rate of Gag containing RSV PR(S105-
107) but not that of Gag containing the covalently linked
dimer. This suggests that I-domain-mediated interactions
among Gag molecules facilitate the formation of PR dimers at
the level of the polyprotein. The budding defect caused by
premature PR activation is experimentally similar to the bud-
ding defect observed by deletion of the Gag L domain in p2
(Fig. 5B, lane 10). It differs, however, in that the budding
defect caused by premature PR activation can be rescued by
inactivating PR (Fig. 4B, lane 6) whereas that caused by the
L-domain mutation cannot (33).
Is there a preferred cleavage order for the processing sites in
Gag? The NC-PR site must be cleaved first, because cleavage
of all other sites in the Gag polyprotein is dependent on re-
lease of PR and assembly of active homodimers. In contrast,
removal of the spacer peptide between the CA and NC pro-
teins and cleavage at the MA-p2 junction appear to be late
events, as polyprotein intermediates of MA-p2 and CA-spacer
can be detected in newly budded particles from cells. Removal
of the spacer peptide between the CA and NC proteins is
FIG. 8. Model for proteolytic processing of viral polyproteins. Gag polyproteins are represented by gray boxes. The PR cleavage sites in Gag are represented by
thick black lines. The M domain and Gag-Gag I domain are indicated as M and I, respectively. The PR subunit is shown as a half scissors, while a dimerized PR is
shown as a full scissors.
2090 XIANG ET AL. J. VIROL.
correlated with the rate of particle budding. When the budding
process is delayed, removal of the spacer is slowed concomi-
tantly (22, 35), suggesting that the two processes are linked
Substitutions introduced in the Gag p2-p10 cleavage site,
which decreased its predicted cleavage rate, resulted in MA-
p2-p10 fusion proteins being accumulated in particles (33).
This occurred without an observable effect on overall Gag
processing or particle release (Fig. 4, lanes 3). Changing the
cleavage rate of the MA-p2 junction by more than 10-fold also
results in little effect on particle budding except that mature
MA is released more rapidly. That cleavage rates of Gag
polyprotein sites can be altered by more than 10-fold without
causing a biological defect in assembly underscores the re-
markable flexibility built into the viral replication process. This
wide tolerance allows for production of infectious virus even
when polyprotein processing is altered by protease mutations
conferring resistance to PR-directed drugs used to treat pa-
tients with AIDS. This replication allows further mutations to
develop at the cleavage sites that can partially compensate for
mutated PR. One such case has recently been reported (13).
We thank Terry Copeland, NCI-Frederick, for the preparation of
peptide substrates, Beth Morrison for construction of the MA-
p23NC-PR clones depicted in Fig. 1B, and A. M. Skalka for critical
reading of the manuscript.
N.K.K. is supported by Public Health Service grant CA-47482 and
American Cancer Society grant FRA-427 awarded to John W. Wills at
Pennsylvania State University, in whose laboratory the ?NC clone was
constructed and analyzed. This work was supported in part by Public
Health Service grants CA38046 and CA52047 from the National Can-
1. Aiyar, A., Y. Xiang, and J. Leis. 1995. Site-directed mutagenesis using over-
lap extension PCR. Methods Mol. Biol. 57:177–191.
2. Bennett, R. P., T. D. Nelle, and J. W. Wills. 1993. Functional chimeras of
Rous sarcoma virus and human immunodeficiency virus Gag proteins. J. Vi-
3. Bizub, D., I. T. Weber, C. E. Cameron, J. P. Leis, and A. M. Skalka. 1991. A
range of catalytic efficiencies with avian retroviral protease subunits geneti-
cally linked to form single polypeptide chains. J. Biol. Chem. 266:4951–4958.
4. Bryant, M., and L. Ratner. 1990. Myristoylation-dependent replication and
assembly of human immunodeficiency virus 1. Proc. Natl. Acad. Sci. USA
5. Burstein, H., D. Bizub, and A. M. Skalka. 1991. Assembly and processing of
avian retroviral Gag polyproteins containing linked protease dimers. J. Virol.
6. Burstein, H., D. Bizub, M. Kotler, G. Schatz, V. M. Vogt, and A. M. Skalka.
1992. Processing of avian retroviral Gag polyprotein precursors is blocked by
a mutation at the NC-PR cleavage site. J. Virol. 66:1781–1785.
7. Cameron, C., B. Grinde, P. Jacques, J. Jentoft, J. Leis, A. Wlodawer, and I.
Weber. 1993. Comparison of the substrate-binding pockets of the Rous
sarcoma virus and human immunodeficiency virus type 1 proteases. J. Biol.
8. Cameron, C. E., B. Grinde, J. Jentoft, and J. Leis. 1992. Mechanism of
inhibition of retroviral protease by a Rous sarcoma virus peptide substrate
representing the cleavage site between Gag p2 and p10 proteins. J. Biol.
9. Cameron, C. E., T. W. Ridky, S. Shulenin, J. Leis, I. T. Weber, T. Copeland,
A. Wlodawer, H. Burstein, D. Bizub-Bender, and A. M. Skalka. 1994. Mu-
tational analysis of the substrate binding pockets of the Rous sarcoma virus
and human immunodeficiency virus-1 proteases. J. Biol. Chem. 269:11170–
10. Campbell, S., and V. M. Vogt. 1995. Self-assembly in vitro of purified CA-NC
proteins from Rous sarcoma virus and human immunodeficiency virus type
1. J. Virol. 69:6487–6497.
11. Crawford, S., and S. P. Goff. 1985. A deletion mutation in the 5? part of the
pol gene of Moloney murine leukemia virus blocks proteolytic processing of
the gag and pol polyproteins. J. Virol. 53:899–907.
12. Debouck, C. 1992. The HIV-1 protease as a therapeutic target for AIDS.
AIDS Res. Hum. Retroviruses 8:153–164.
13. Doyon, L., G. Croteau, D. Thibeault, F. Polin, L. Pilote, and D. Lamarre.
1996. Second locus involved in human immunodeficiency virus type 1 resis-
tance to protease inhibitors. J. Virol. 70:3763–3769.
14. Erikson, E., J. Brugge, and R. Erikson. 1977. Phosphorylated and nonphos-
phorylated forms of avian sarcoma virus polypeptide p19. Virology 80:177–
15. Gelderblom, H. 1990. Morphogenesis, maturation, and fine structure of
lentiviruses, p. 159–180. In L. H. Pearl (ed.), Retroviral proteases: control of
maturation and morphogenesis. Stockton Press, New York, N.Y.
16. Grinde, B., C. Cameron, J. Leis, I. Weber, A. Wlodawer, H. Burstein, D.
Bizub, and A. M. Skalka. 1992. Mutations that alter the activity of the Rous
sarcoma virus protease. J. Biol. Chem. 267:9481–9490.
17. Ho, D. D., A. U. Neumann, A. S. Perelson, W. Chen, J. M. Leonard, and M.
Markowitz. 1995. Rapid turnover of plasma virions and CD4 lymphocytes in
HIV-1 infection. Nature 373:123–126.
18. Katz, R. A., and A. M. Skalka. 1994. The retroviral enzymes. Annu. Rev.
19. Kohl, N. E., E. A. Emini, W. A. Schleif, L. J. Davis, J. C. Heimbach, R. A.
Dixon, E. M. Scolnick, and I. S. Sigal. 1988. Active human immunodeficiency
virus protease is required for viral infectivity. Proc. Natl. Acad. Sci. USA
20. Krausslich, H. G. 1991. Human immunodeficiency virus proteinase dimer as
component of the viral polyprotein prevents particle assembly and viral
infectivity. Proc. Natl. Acad. Sci. USA 88:3213–3217.
21. Leis, J., N. Phillips, X.-D. Fu, P. T. Tuazon, and J. A. Traugh. 1989. Phos-
phorylation of avian retrovirus matrix protein by Ca??/phospholipid-depen-
dent protein kinase. Eur. J. Biochem. 179:415–422.
22. Pepinsky, R., J. A. Papayannopoulos, E. P. Chow, N. K. Krishna, R. C.
Craven, and V. M. Vogt. 1995. Differential proteolytic processing leads to
multiple forms of the CA protein in avian sarcoma and leukemia viruses.
J. Virol. 69:6430–6438.
23. Ridky, T., and J. Leis. 1995. Development of drug resistance to HIV-1
protease inhibitors. J. Biol. Chem. 270:29621–29623.
24. Ridky, T. W., D. Bizub-Bender, C. E. Cameron, I. T. Weber, A. Wlodawer, T.
Copeland, A. M. Skalka, and J. Leis. 1996. Programming the Rous sarcoma
virus protease to cleave new substrate sequences. J. Biol. Chem. 271:10538–
25. Sellos-Moura, M., and V. M. Vogt. 1996. Proteolytic activity of purified avian
sarcoma and leukemia virus NC-PR protein expressed in Escherichia coli.
26. Stewart, L., G. Schatz, and V. M. Vogt. 1990. Properties of avian retrovirus
particles defective in viral protease. J. Virol. 64:5076–5092.
27. Weber, I. T. 1990. Comparison of the crystal structures and inter subunit
interactions of human immunodeficiency and Rous sarcoma virus proteases.
J. Biol. Chem. 265:10492–10496.
28. Weldon, R. A., Jr., and J. W. Wills. 1993. Characterization of a small (25-
kilodalton) derivative of the Rous sarcoma virus Gag protein competent for
particle release. J. Virol. 67:5550–5561.
29. Wills, J. W. Personal communication.
30. Wills, J. W., R. C. Craven, and J. A. Achacoso. 1989. Creation and expression
of myristylated forms of Rous sarcoma virus gag protein in mammalian cells.
J. Virol. 63:4331–4343.
31. Wills, J. W., and R. C. Craven. 1991. Form, function, and use of retroviral gag
proteins. AIDS 5:639–654.
32. Wills, J. W., R. C. Craven, R. A. Weldon, Jr., T. D. Nelle, and C. R. Erdie.
1991. Suppression of retroviral MA deletions by the amino-terminal mem-
brane-binding domain of p60src. J. Virol. 65:3804–3812.
33. Wills, J. W., C. E. Cameron, C. B. Wilson, Y. Xiang, R. P. Bennett, and J.
Leis. 1994. An assembly domain of the Rous sarcoma virus Gag protein
required late in budding. J. Virol. 68:6605–6618.
34. Wlodawer, A., and J. W. Erickson. 1993. Structure-based inhibitors of HIV-1
protease. Annu. Rev. Biochem. 62:543–585.
35. Xiang, Y., C. E. Cameron, J. W. Wills, and J. Leis. 1996. Fine mapping and
characterization of the Rous sarcoma virus PR76gaglate assembly domain.
J. Virol. 70:5695–5700.
36. Zhang, Z. Y., R. A. Poorman, L. L. Maggiora, R. L. Heinrikson, and F. J.
Kezdy. 1991. Dissociative inhibition of dimeric enzymes. Kinetic character-
ization of the inhibition of HIV-1 protease by its COOH-terminal tetrapep-
tide. J. Biol. Chem. 266:15591–15594.
37. Zybarth, G., H. G. Krausslich, K. Partin, and C. Carter. 1994. Proteolytic
activity of novel human immunodeficiency virus type 1 proteinase proteins
from a precursor with a blocking mutation at the N terminus of the PR
domain. J. Virol. 68:240–250.
38. Zybarth, G., and C. Carter. 1995. Domains upstream of the protease (PR) in
human immunodeficiency virus type 1 Gag-Pol influence PR auto process-
ing. J. Virol. 69:3878–3884.
VOL. 71, 1997 ACTIVATION OF RSV POLYPROTEIN PROCESSING2091