Identification of a Conserved Residue of Foamy Virus Gag Required for Intracellular Capsid Assembly

Abstract

In contrast to all retroviruses but similar to the hepatitis B virus, foamy viruses (FV) require expression of the envelope protein for budding of intracellular capsids from the cell, suggesting a specific interaction between the Gag and Env proteins. Capsid assembly occurs in the cytoplasm of infected cells in a manner similar to that for the B- and D-type viruses; however, in contrast to these retroviruses, FV Gag lacks an N-terminal myristylation signal and capsids are not targeted to the plasma membrane (PM). We have found that mutation of an absolutely conserved arginine (Arg) residue at position 50 to alanine (R50A) of the simian foamy virus SFV cpz(hu) inhibits proper capsid assembly and abolishes viral budding even in the presence of the envelope (Env) glycoproteins. Particle assembly and extracellular release of virus can be restored to this mutant with the addition of an N-terminal Src myristylation signal (Myr-R50A), presumably by providing an alternate site for assembly to occur at the PM. In addition, the strict requirement of Env expression for capsid budding can be bypassed by addition of a PM-targeting signal to Gag. These results suggest that intracellular capsid assembly may be mediated by a signal akin to the cytoplasmic targeting and retention signal CTRS found in Mason-Pfizer monkey virus and that FV Gag has the inherent ability to assemble capsids at multiple sites like conventional retroviruses. The necessity of Env expression for particle egress is most probably due to the lack of a membrane-targeting signal within FV Gag to direct capsids to the PM for release and indicates that Gag-Env interactions are essential to drive particle budding.

Foamy viruses (FV), classified in the Spumavirus genus of the Retroviridae family, are unique viruses sharing morphogenic features found among many diverse types of enveloped viruses, including the human hepatitis B virus. Although FVs cause substantial cytopathic effects in tissue culture (hence the name “foamy”, referring to the highly vacuolated appearance of infected cells), an asymptomatic and persistent infection is seen in nature in a wide variety of organisms including nonhuman primates, cats, cattle, and horses (22, 27, 42, 46). The genomic organization of the FVs, including the prototype molecular cloned simian foamy virus SFVcpz(hu), is similar to that of other complex retroviruses, with several additional open reading frames located 3′ of the canonical gag, pol, and env genes, including the transcriptional transactivator gene, tas (23, 29, 36). Unlike in retroviruses, however, the Pol protein is expressed from a unique spliced mRNA, not as a Gag-Pol fusion, and therefore must be specifically incorporated into newly forming capsids (14, 28, 47). In addition, reverse transcription of the genome is initiated early, during budding of capsids, viral egress, or prior to infection of new cells, suggesting a novel coordination of morphogenesis (50).

The main FV structural protein, Gag, is also different from that of other retroviruses because of the lack of proteolytic processing into the MA, CA, and NC domains, and, correspondingly, extracellular viral particles possess an immature morphology (9). The majority of viral protease-specific Gag cleavage is limited to a single event near the C terminus of the protein, releasing an approximately 3-kDa protein from the 71-kDa precursor peptide (13, 16, 35), although it has been proposed that additional cleavage occurs on infection of new cells during viral uncoating (16, 34, 39). In addition, FV Gag lacks the major homology region found in the capsid proteins of all other retroviruses, as well as the signature Cys-His box motifs found in all retroviral Gag NC proteins (43a). Instead, FV Gag possesses three glycine-arginine rich domains, termed GR boxes I to III, situated at the C terminus of the protein, which are involved in nucleic acid binding and nuclear localization (41, 48) and possibly particle density (4). By 24 h postinfection, a strong nuclear Gag signal is seen in all cell types infected with FVs (the serological hallmark of FV infection), although transport of Gag to the nucleus is not essential for the production of infectious virus and the role of Gag nuclear localization is not known (22, 41, 48).

FV particle assembly occurs in the cytosolic compartment (8, 10). Similar to the intracisternal A-type particles but distinct from all other retroviruses, FV capsids bud through the endoplasmic reticulum (ER) membrane. The FV envelope (Env) protein is also retained in the ER by means of a trilysine motif located at the C-terminal cytoplasmic tail of Env (18, 19). FV morphogenesis requires the presence of the Env protein to allow release of virus from the cell, a mechanism also employed by hepatitis B virus, such that capsid budding is completely inhibited in the absence of Env expression (2, 5, 15). In contrast, all other retroviruses have the ability to assemble and bud capsids from a variety of cell types on the sole expression of the gag gene (3; Wills and Craven, Editorial).

While the mechanism of intracellular capsid assembly is not fully understood for the B- and D-type retroviruses, recent experiments with the D-type retrovirus Mason-Pfizer monkey virus (MPMV) have show that an 18-amino-acid domain near the N terminus of Gag, termed the cytoplasmic targeting and retention signal (CTRS) and centered on a highly conserved arginine (Arg) residue at position 55, is required to direct the cytoplasmic assembly of capsids (7, 32, 38, 40). Mutation of Arg55 of Gag to tryptophan (R55W) results in a switch of morphology to the default type C capsid assembly at the plasma membrane (PM), a process dependent on the N-terminal myristylation signal (37, 38). Remarkably, fusion of the CTRS to green fluorescent protein (GFP) resulted in discrete staining at cytoplasmic sites, and, furthermore, addition of the CTRS to the C-type retrovirus Moloney murine leukemia virus Gag protein conferred intracellular assembly (7).

All sequenced FV Gag proteins have a high level of conservation near the N terminus, including an absolutely conserved arginine at amino acid 50. Comparison of this region of FV Gag with the CTRS of MPMV reveals a number of identical residues, comprising a domain of FV Gag [GXWGX₃RX₇L(Q/V)D], centered on the conserved Arg (7, 42). We predicted that if cytoplasmic assembly of FV capsids is mediated by sequences in this region, mutation of conserved residues might block assembly altogether, since the FV Gag protein is not myristylated and is not known to possess a membrane-targeting signal. We found that alanine or tryptophan substitution at position 50 (R50A/W) of SFVcpz(hu) severely disrupts capsid assembly in transfected cells and completely abolishes the release of virus from the cell even in the presence of Env. Capsid assembly and budding of the R50A mutant was restored if Gag was given a targeting signal via the addition of an N-terminal Src myristylation signal (33, 43, 44). Extracellular particles were produced at wild-type levels but were not infectious, and Gag cleavage was completely blocked. Surprisingly, addition of a myristylation signal bypassed the strict requirement for Env in particle budding, suggesting that Gag does not possess an inherent targeting signal and therefore relies on interaction with Env for particle egress.

MATERIALS AND METHODS

Recombinant plasmid DNAs.

The infectious molecular clone of SFVcpz(hu), designated pHSRV13, was used for all experiments described below (30). Proviral mutants R50A and R50W were created by PCR mutagenesis using the Morph kit (5 Prime→3 Prime) with a modified version of FV subclone 1 (2) renamed G3 (AvrII and EcoRV sites deleted from the polylinker, consisting of a 2,885-bp fragment of pHSRV13 from EagI to SwaI) as a template and using the oligonucleotides primers MAR1A (5′-GGACAAATT GAGGCATTTCAGATGG-3′) and MAR1W (5′-GTGGGGACAAATTGAGTGGTTTC AGATGGTACG-3′) to change the Arg residue (AGA) to Ala (GCA) and Trp (TGG), respectively. The Src myristylation signal (GSSKSKPKD) was introduced into the G3 subclone by inverse PCR with the oligonucleotides SRC+ (5′-GGCTCATCGAAGAGCAAGCCTAAGGACGAACTTGATGTTGAAGC-3′) and SRC− (5′-GTCCTTAGGCTTGCTCTTCGATGAGCCCATTGTCTATTGGCTTT-3′) to create the Myr and Myr-R50A proviruses. The Env deletion mutants were cloned with a 2,536-bp fragment (PacI [position 4644] to BlpI [position 9193]) from the provirus ΔMN (deleted from MroI [BspEI] [position 6957] to NdeI [position 8970] provided by Martin Loechelt) (2).

All 293T cell transfections were conducted with proviruses containing the immediate-early promoter from the cytomegalovirus (CMV) in place of the U5 region of the 5′ long terminal repeat (LTR) sequence of pHSRV13, such that expression of the viral RNA is initiated from the same nucleotide as for the wild-type RNA transcribed from U3 (D. Baldwin and M. Linial, unpublished data). Briefly, the CMV immediate-early promoter was PCR amplified from pCR3.0 (Invitrogen) with oligonucleotide primers containing EagI and XhoI sites (5′ and 3′, respectively) to clone into a modified pHSRV13 subclone (sub1) described previously (2) containing a linker with an additional XhoI site between EagI and XhoI (Baldwin and Linial, unpublished). The R50A and Myr-R50A mutants were introduced into CMVsub1 with restriction sites AvrII and PflMI.

Cells and transfections.

FAB indicator cells, BHK cells containing an integrated copy of the β-galactosidase gene under the control of the SFVcpz(hu) LTR (49), were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal bovine serum. Human embryonic lung cells (HEL), thymocytes (CF3TH), and 293 T cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. FAB cell transient transfections were conducted using the Lipofectamine reagent (Life Technologies, Inc.) as previously described (2) with 5 μg of proviral DNA, as well as the Fugene reagent (Roche Molecular Biochemicals) in which 10 μg of proviral DNA was added to 250 μl of DMEM containing 25 μl of Fugene, mixed and incubated at room temperature for 15 min, added to cells in 10 ml of medium, and rinsed after 24 h. 293 T cell transient transfections were conducted using a modified calcium phosphate method (6) in which 8 μg of proviral DNA plus 2 μg of LTR-GFP reporter plasmid were combined with 0.5 ml of 0.25 M CaCl₂, then added to 0.5 ml of 2× BES buffered solution mixed and incubated at room temperature for 20 min, and added for 18 to 20 h to 10-cm-diameter dishes containing cells at approximately 75 to 85% confluency in 10 ml of medium.

Western blotting.

Proteins were analyzed by Western blotting of cell lysates and viral supernatants with the anti-Gag polyclonal antiserum as previously described (2), with the following modifications. FAB and 293 T cells were scraped in phosphate-buffered saline (PBS) 40 to 42 h posttransfection, pelleted, and rinsed three times with PBS. Cell pellets were lysed in 1 ml of antibody buffer (20 mM Tris [pH 7.5], 50 mM NaCl, 0.5% NP-40, 0.5% sodium dodecyl sulfate [SDS], 0.5% deoxycholate, 0.5% aprotinin, supplemented with 100 μg of phenylmethylsulfonyl fluoride per ml and 1 μg of leupeptin per ml), passed through a 23-gauge syringe to shear chromosomal DNA, cleared of cell debris by centrifugation in the microcentrifuge, mixed with SDS protein sample buffer, boiled, and loaded on SDS-polyacrylamide gel electrophoresis (PAGE) minigels (10% polyacrylamide). Culture supernatants were passed through 0.45-μm-pore-size syringe filters and pelleted through a 20% sucrose cushion (2 ml) at 24,000 rpm and 4°C for 2 h in a total volume of 17.5 ml (SW28 rotor; Beckman). Viral pellets were resuspended with protein sample buffer and loaded onto SDS-PAGE minigels (10% polyacrylamide). After separation, proteins were transferred to Immobilon-P membranes (Millipore), blocked in 5% nonfat milk in PBS, and incubated with anti-Gag antiserum at 1:2,000 overnight. The membranes were washed three times in PBS containing 0.1% Tween 20 (PBS-T) and incubated with horseradish peroxidase-conjugated anti-rabbit immunoglobulin (Amersham) antibody at 1:7,500 dilution for 1 h. They were then washed four times for 20 min in PBS-T and visualized by enhanced chemiluminescence (Amersham).

Linear-velocity sedimentation gradients.

Transiently transfected 293 T cells were washed with PBS and lysed in NP-40 lysis buffer (1% NP-40, 50 mM NaCl, 10 mM Tris-Cl [pH 7.4], 5 mM EDTA) for 30 min on ice. Lysates were cleared with an initial low-speed centrifugation at 4,000 × g for 10 minutes followed by centrifugation at 15,000 × g for 5 min in a microcentrifuge, and the resulting supernatants were pelleted through 750 μl of 20% sucrose in TNE (20 mM Tris [pH 7.5], 150 mM NaCl, 1 mM EDTA) at 25,000 rpm for 2 h and 4°C in a total volume of 5 ml (Ti55 rotor; Beckman). The pellets were resuspended in 300 μl of TNE, placed onto 5 ml of NP-40 lysis buffer containing sucrose step gradients consisting of 1.2 ml of 20%, 2 ml of 40%, and 1.5 ml of 66% sucrose, and ultracentrifuged at 30,000 rpm for 1 h (Ti55 rotor; Beckman) at 4°C, and eight 635-μl fractions were collected from the top, as well as the pellet fraction (resuspended in NP-40 lysis buffer) (26). Fractions were precipitated for 1 h at −20°C with trichloroacetic acid (TCA) (consisting of a final concentration of 25%) and 10 μg of yeast tRNA, centrifuged at 16,000 × g for 20 min in a microcentrifuge, washed with 10% TCA and then with 100% acetone, air dried, and resuspended in 1× protein sample buffer. Fractions were subjected to SDS-PAGE (10% polyacrylamide) and Western blotting. Extracellular virus was recovered from culture supernatants by pelleting through 20% sucrose and analyzed on these gradients after removal of viral envelopes by treatment with 1% NP-40 in TNE.

Indirect immunofluorescence (IFA).

Transiently transfected FAB cells on glass coverslips were rinsed with PBS and fixed for 5 min at room temperature with 4% paraformaldehyde at 36 to 40 h posttransfection. The cells were permeabilized with 1% Triton-X in PBS for 5 min, washed with PBS, and blocked in 5% heat-inactivated bovine serum albumin for at least 30 min at 4°C. The coverslips were then rinsed and incubated with anti-Gag serum (1:2,000) for 1 h at 37°C, rinsed with PBS for 15 min three times, incubated with anti-rabbit fluorescein isothiocyanate (FITC)-conjugated antibody (1:1,000) for 45 min at 25°C, and rinsed with PBS as before. They were then stained with 4′,6-diamidino-2-phenylindole (DAPI; 0.2, μg/ml) for 5 min in double-distilled H₂O washed in double-distilled H₂O, and mounted in Vectashield (Vector Laboratories). Imaging was performed using a Nikon TE300 microscope and Metamorph software.

RESULTS

Point mutation of the arginine at position 50 (R50) of FV Gag inhibits particle assembly and blocks viral budding.

Analysis of the Gag amino acid sequence of all FV isolates shows an absolute conservation of the arginine at position 50 (R50) from the N terminus (Fig. 1A). The sequence of SFVcpz(hu) in the vicinity of this conserved R50 is reminiscent of the CTRS of the D-type retrovirus MPMV (Fig. 1B) (7, 38), including a number of identical residues (7, 38), comprising a domain of FV Gag [GXWGX₃RX₇L(Q/V)D] (42). To determine whether R50 of SFVcpz(hu) is involved in intracytoplasmic assembly, we made substitutions to alanine (R50A) and tryptophan (R50W) (Fig. 1C). Proviral mutants under the control of the CMV immediate-early promoter were transfected into 293T cells along with the wild-type (wt) molecular clone of SFVcpz(hu), HSRV. Intracellular Gag proteins and extracellular virus that was purified from culture supernatants were then analyzed by Western blotting. The cellular Gag expression levels (Fig. 2A) of R50A and R50W were comparable to those of wt HSRV. Extracellular release of virus (Fig. 2B), however, was completely abolished with these mutant proviruses compared to the budding competent virus. wt HSRV cannot spread in this cell type and therefore provides a proper control for a single-cycle infection. We performed the same transient-transfection experiments with FAB cells, which are permissive for replication and viral spread, to determine whether the block to assembly and viral release was cell type specific, and we observed identical results (data not shown).

FIG. 1.

Schematic representation of FV Gag sequence alignments, genomic structure map, and location of proviral mutants. (A) Sequence alignment of the N-terminal portion of the Gag protein from all characterized FV molecular clones including the amino acid position number. ∗, absolute conservation; /, conservative change. (B) Sequence alignment (Clustal W) of the N-terminal portion of SFVcpz(hu) Gag with the CTRS domain from the D-type retrovirus MPMV, with the absolutely conserved Arg residue highlighted in bold. (C) Genomic structure of SFVcpz(hu) gag including the 5′ region of CMV-driven HSRV proviral constructs containing the CMV immediate-early promoter in place of the 5′ LTR for initiation of viral transcription. The location of all proviral mutant sequences is shown in relation to wt HSRV. The Myr mutant contains the 10-amino-acid N-terminal Src myristylation signal (MGSSKSKPKD) in place of the first 10 amino acids of Gag. R50A and R50W have alanine and tryptophan substitutions, respectively, of the conserved arginine residue at position 50. ΔEnv proviruses have a 2-kb deletion in env (2).

FIG. 2.

R50A/W mutation blocks the release of virus from FV-transfected cells. Western blot analysis of transiently transfected 293T cells expressing proteins from wt and mutant CMV-driven proviruses is shown. (A) Cellular lysates prepared 40 h posttransfection from cells transfected with HSRV or the conserved Arg substitution mutants R50A and R50W, as well as a mock-transfected negative control. (B) Extracellular virus isolated from the same culture supernatants by 20% sucrose cushion sedimentation, including wt HSRV, the conserved Arg substitution mutants R50A and R50W, and the mock-transfected negative control. Viral proteins were visualized using the anti-Gag sera. MW, molecular weight in thousands.

Next, we tested the hypothesis that the lack of particle budding observed with the R50A/W mutants is caused by the inhibition of intracellular capsid assembly. We first used a standard equilibrium density centrifugation method to detect assembled retroviral particles in the cell. Transiently transfected FAB cell lysates were placed on gradients by previously described methods (1), and Gag protein was found to band at a density of approximately 1.14 g/ml for both the wt and R50A mutant proviruses on Western blots (data not shown). This finding was both inconsistent with our prediction of FV intracellular assembly occurring via a CTRS-type domain and puzzling since we observed a complete lack of extracellular virus produced from the R50A mutant provirus. However, previous experiments performed in the laboratory using proviral mutants lacking gag (ΔGag) indicated that viral polymerase proteins exhibited banding characteristics that are indistinguishable from banding seen with wt virus on Western blots of fractions collected from identical density equilibrium gradients (Baldwin and Linial, unpublished). Thus, the equilibrium centrifugation does not appear to distinguish between unassembled viral protein complexes and assembled viral capsids under these conditions. Therefore, we used the linear-velocity sedimentation techniques that were previously used to examine the human immunodeficiency virus (HIV) capsid assembly pathway (26). We used a modified lysis procedure with 1% NP-40 in an attempt to discriminate bona fide intracellular capsids from unassembled Gag monomers and protein aggregates. Cell lysates were centrifuged through 20% sucrose, and the pelleted material was placed on linear-velocity sedimentation gradients made from 20 to 66% sucrose containing 1% NP-40 and centrifuged under conditions found to band HIV capsids (approximately 750S) in the middle of the gradient (fraction 5) (26). We found that whereas cellular expression levels of HSRV and R50A mutant Gag were identical (Fig. 3A), the R50A mutant gradient was devoid of Gag protein compared with the wt gradient containing Gag capsids in fraction 6 (Fig. 3B). This indicated that the initial 20% sucrose spin separated actual capsids (HSRV) from protein aggregates (R50A).

FIG. 3.

R50A mutation inhibits capsid assembly. Western blot analysis of transiently transfected 293T cells expressing proteins from wt and mutant CMV-driven proviruses is shown. (A) Total-cell lysates. (B) Cell lysates from HSRV and the R50A mutant. These lysates were pelleted through 20% sucrose, resuspended in TNE, and layered onto 20, 40, and 66% sucrose linear velocity sedimentation gradients. Fractions were collected from the top of the gradients (lane 1) and TCA precipitated. Viral proteins were visualized using the anti-Gag sera. The arrow indicates the location of HIV ∼750S particles analyzed on parallel gradients.

Plasma membrane targeting of R50A restores capsid assembly and extracellular release of virus.

We next tested whether the R50A mutant, lacking a putative intracellular signaling domain, could synthesize particles if supplied with an alternate targeting signal. There is precedent for this type of experiment, since studies with the mouse intracisternal A-type particle retrovirus, which assembles and buds capsids into the ER that remain within the ER lumen, have shown that redirection of Gag to the PM by myristylation allows extracellular release of virus (43). We substituted the Src targeting signal, a 10-amino-acid N-terminal peptide previously reported to act as a dominant plasma membrane-targeting domain dependent on myristylation and the presence of 3 lysine residues, for the first 10 amino acids at the N terminus of the R50A mutant Gag protein to produce the Myr-R50A provirus (Fig. 1B) (44, 45). CMV-driven proviral mutants or HSRV were transiently transfected into 293 T cells, and cellular lysates and viral supernatants were analyzed. We first checked the ability of the myristylated form of the R50A mutant (Myr-R50A) to form intracellular capsids by using the intracellular assembly assay described above. Cellular lysates from HSRV- or Myr-R50A-transfected cells (Fig. 4A) were pelleted through 20% sucrose, placed on linear-velocity gradients, and fractionated. Remarkably, these results indicated that the Myr-R50A mutant Gag was able to pellet through 20% sucrose and was found to possess Gag sedimentation characteristics indistinguishable from those of the wt on the linear-velocity gradients (Fig. 4B).

FIG. 4.

Plasma membrane targeting of R50A mutant with the Src- myristylation signal restores intracellular capsid assembly. Western blot analysis of transiently transfected 293T cells expressing proteins from wt and mutant CMV-driven proviruses is shown. (A) Total-cell lysates. (B) Linear-velocity sedimentation gradient analysis of cell lysates from HSRV (top) and Myr-R50A mutants (bottom). Viral proteins were visualized using the anti-Gag sera.

Next, we determined whether myristylation of R50A Gag could restore viral budding. Again, 293 T cells were transfected with the proviral constructs and Western blot analyses were performed from whole-cell lysates. The results show similar expression levels of cellular Gag (Fig. 5A) for HSRV, R50A, and Myr-R50A, although proteolytic processing of Gag was inhibited. Analysis of purified culture supernatants (Fig. 5B), however, revealed a rescue of viral release on addition of the PM-targeting signal (Myr-R50A), at levels comparable to those for HSRV, whereas the R50A mutant did not release virus. We consistently observed the absence of Gag cleavage with the extracellular Myr-R50A viruses. Identical results were observed using FAB cells (data not shown). In addition, extracellular virus produced from cells transfected with HSRV or the Myr-R50A mutant proviruses (Fig. 5C, cell lysate) was analyzed on linear-velocity sedimentation gradients after removal of the viral envelope with 1% NP-40. The results show that the Myr-R50A mutant exhibited sedimentation characteristics identical to those of wt HSRV, with Gag banding in fraction 6 (Fig. 5C). The infectivity of the Myr-R50A was assessed using the FAB assay (49), and the virus was found to be noninfectious, containing less than 1 infectious unit (IU) per ml of culture supernatant, compared with 10⁵ to 10⁶ IU/ml for wt virus (data not shown).

FIG. 5.

Plasma membrane targeting of R50A mutant restores viral budding. Western blot analysis of transiently transfected 293T cells expressing proteins from wt and mutant CMV-driven proviruses is shown. (A) Cell lysates prepared 41 h posttransfection including HSRV, the R50A mutant, and the myristylated R50A mutant, Myr-R50A, as well as a mock-transfected negative control. (B) Extracellular virus isolated from culture supernatants of the same cell transfections by 20% sucrose cushion centrifugation. (C) Linear-velocity sedimentation gradients of extracellular virus (treated with 1% NP-40 to remove viral envelope) produced from cells transfected with the HSRV (top) and Myr-R50A (bottom) proviruses (whole-cell lysates shown at left). Viral proteins were visualized using anti-Gag sera. MW, molecular weight in thousands.

Subcellular localization of mutant Gag proteins and assembled capsid structures.

Subcellular localization of the mutant viral proteins was analyzed by indirect immunofluorescence (IFA) of transiently transfected FAB cells using the anti-Gag serum. Expected distinctive nuclear fluorescence was seen in cells transfected with HSRV (Fig. 6B), as well as the R50A mutant (Fig. 6C) (41). Addition of the Src membrane-binding domain to wt and R50A mutant Gag significantly altered protein localization. Cell nuclei appeared dark with a corresponding punctate staining of cytoplasmic regions including PM sites (Fig. 6D and E). Comparison of the Myr mutants with a mutant lacking the NLS in GR box II, ΔNLS (50) (Fig. 6F), revealed that whereas nuclei were not stained in either case, the Myr-tagged proteins appeared to accumulate in brightly staining regions within the cytoplasm and at the PM, as opposed to the diffuse cytoplasmic staining seen in cells transfected with the ΔNLS mutant. These staining patterns are reminiscent of HIV Gag localization with and without N-terminal myristylation, respectively (31).

FIG. 6.

Subcellular Gag localization using IFA. Transiently transfected FAB cells (phase-contrast images [left panel]) were fixed 36 to 40 h posttransfection, incubated with anti-Gag sera, and visualized with FITC (center panel). Nuclei are stained with DAPI (right panel). The proviruses are mock (A) HSRV (B), the R50A mutant (C), Myr (D), and Myr-R50A (E), as well as the control mutant lacking the NLS in GR box II, ΔNLS (F) (48).

Redirection of Gag to the plasma membrane allows FV particle budding in the absence of Env.

We next examined whether the intracellular assembly mutant virus containing the PM-targeting signal, Myr-R50A, required Env glycoprotein synthesis for extracellular release. Our previous finding that particle budding is dependent on the presence of Env suggests that FV capsids do not possess an inherent membrane-targeting signal and that specific Gag-Env interactions are required for capsids to be released from the cell. We hypothesized that if Gag was given a specific mechanism for localization, such as the Src plasma membrane-targeting signal, the Env requirement could be bypassed and capsids would bud from the PM in the absence of Env expression. We deleted most of the env gene from each provirus to create mutants (ΔEnv) that do not express Env but do express all other proteins at wt levels and conducted transfection experiments as before (2). Western blot analysis of cellular lysates from 293 T cells transiently transfected with the mutant proviral clones showed comparable levels of intracellular Gag (Fig. 7A). Analysis of extracellular virus purified from these culture supernatants (Fig. 7B), however, showed that in the absence of Env expression, the Myr/ΔEnv (lane 8) and Myr-R50A/ΔEnv (lane 9) mutants released virus at levels at or surpassing those of budding-competent noninfectious control viruses D936I (integrase mutant, lane 2), and HSRV-D/A (protease active site mutant, lane 3). In contrast, HSRV/ΔEnv (lane 7) was completely inhibited in particle release, similar to the R50A mutant (lane 4) and mock transfected cells (lane 1).

FIG. 7.

Myr-R50A mutant virus budding in the absence of Env. Western blot analysis of CMV-driven mutant proviruses transfected into 293 T cells is shown (A) Cell lysates prepared 41 h posttransfection, including the integrase mutant, D936I (lane 2), and the protease active-site mutant, HSRV-D/A (lane 3), R50A (lane 4), Myr (lane 5), Myr-R50A (lane 6), the HSRV envelope deletion, HSRV/ΔEnv (lane 7), Myr/ΔEnv (lane 8), and Myr-R50A/ΔEnv (lane 9), as well as mock-transfected cells (lane 1). (B) Extracellular virus purified from the same culture supernatants by 20% sucrose cushion centrifugation, with the same lane designations. Viral proteins were visualized using anti-Gag sera.

DISCUSSION

In this study, we have shown that a single substitution of an absolutely conserved residue in the N-terminal region of FV Gag severely inhibits intracellular capsid assembly and ablates viral particle release. This residue (R50) is situated in a region of high conservation among all characterized FVs and has moderate homology to the recently defined CTRS signal from the D-type retrovirus MPMV (7). Nuclear magnetic resonance analysis of the MA domain of MPMV Gag suggests that the 18-amino-acid CTRS region exists on an exposed loop of the protein, not found in the matrix domain of Gag from C-type viruses (11, 12). The CTRS is proposed to target and retain Gag molecules at a specific site within the cell, allowing for localized increases in protein concentrations such that Gag-Gag interactions and assembly may proceed (7). It is also likely that a cellular factor(s) is involved in the targeting process, a hypothesis supported by the recent report of the discovery of an insect cell line defective in the transport of assembled capsids to the PM (32). We propose that FV assembly occurs free of membranes in a fashion similar to the B- and D-type retroviruses and that capsid formation is mediated by a signal akin to the CTRS domain.

Targeting domains for all characterized retroviral Gag proteins reside at the amino-terminal MA domain (24). There is no obvious membrane-targeting signal on FV Gag, and, correspondingly, we found that a disruption of the cytoplasmic assembly signal blocked capsid formation altogether. If a PM targeting signal was provided, however, in the form of the Src myristylation signal, as in Myr or Myr-R50A, capsid assembly proceeded and particle release was restored to wt levels. A switch to C-type assembly at the PM was not detected with these mutants when transiently transfected cells were analyzed by electron microscopy, but this mechanism of capsid formation cannot be ruled out. Indeed, we found that the total amounts of intracellular Myr-R50A mutant Gag were always significantly reduced compared to HSRV (Fig. 5A) in cells transfected with similar levels of provirus as detected by cotransfection of an LTR-GFP plasmid. Extracellular Myr-R50A virus, however, was found to be present at wt levels (Fig. 5B), indicating that PM targeting of Gag may be increasing the efficiency of capsid assembly and budding.

Of considerable interest is the fact that myristylated Gag proteins created capsids that budded from cells in the absence of Env expression. FV capsids are capable of budding from the PM rather than the ER. SFVcpz(hu) mutant viruses, which lack the Env ER retention signal, will bud at wt levels from the PM, where Env is localized (17, 18). Interestingly, the equine and bovine FV possess sequences homologous to the intracellular assembly domain (Fig. 1A) but lack the ER retention signal on the Env protein and, correspondingly, bud virus solely from the PM (21, 42). Recently, it has been reported that the N terminus of the leader peptide of SFVcpz(hu) Env, termed the budding domain, is required for efficient particle release (25). We propose that intracellular budding of FV capsids is a process normally driven by a Gag-Env interaction but that if capsids are given an alternate localization signal, such as the N-terminal Src membrane-targeting signal on Gag, budding may occur from the PM in the absence of Env.

Also of interest is the observation that extracellular virus produced from the Myr-R50A mutant exhibited reduced Gag cleavage and was noninfectious. While many FV mutations cause slight alterations in intracellular Gag cleavage, virus released from cells containing the Myr-R50A Gag mutant provirus appeared to be devoid of any proteolytic processing (Fig. 5B; compare lanes 6 and 8; Fig. 7B, compare lanes 2 and 6), whereas the provirus with only the Src-targeting signal (Myr) produced virus with an intermediate cleavage phenotype (Fig. 7B, compare lanes 2, 5, and 6). Several possibilities could account for this deficiency in Gag cleavage. Addition of the Src signal to Gag may alter the proper folding and conformation of the protein such that processing is prevented. Also, it is conceivable that these mutant viral particles lack a cellular factor required for protease activation or that these mutant virus particles have reduced levels of Pol. If capsid formation is highly site specific, perhaps a specific mechanism exists to target Pol into assembling particles; thus, any perturbation of the site of assembly will affect Pol incorporation. Alternatively, these particles may be devoid of viral nucleic acid, which may also influence the levels of Pol in virus. A recent study analyzing the region upstream of the primer-binding site of SFVcpz(hu) indicates that deletion of a 29-nucleotide domain in the U5 region of the 5′ LTR, while having no effect on packaging of nucleic acid, inhibits Gag cleavage in extracellular virus (20). Perhaps assembly of RNA and Pol into capsids is a highly coordinated event occurring at a discrete location and subtle mutations in Gag have a substantial impact on this process.

The block to infection by the Myr-R50A virus could be due to the lack of cleavage, Pol and/or genome packaging, or failure to incorporate Env. Previous experiments by others have shown that C-terminal Gag cleavage is essential for infectivity (13, 51). However, Myr-R50A viruses, which appear to bud from the plasma membrane, are also likely to be missing Env. Lack of Env would block the Myr-R50A viruses from entering new cells. The presence of Pol, Env, and RNA in these viruses is currently under investigation. Our data indicate that the site of assembly of FV capsids is critical for the formation of an infectious viral particle. Although FV capsids have the inherent ability to assemble at different sites within the cell including the PM, assembly at the CTRS-mediated intracellular site appears to be essential for proper incorporation of all viral components.