INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Equine infectious anemia virus (EIAV) is a member of the lentivirus subfamily of retroviruses. EIAV is genetically the simplest lentivirus in that it contains only three accessory genes (rev, tat, and S2) in addition to the canonical gag, pol, and env genes. This genetic simplicity offers distinct advantages for using EIAV as a model system to study the basic aspects of lentiviral replication, including mechanisms involved in virion budding. As with all retroviruses, EIAV proviral expression in infected cells produces three predominant polyprotein precursors, Gag, Gag-Pol, and Env, that form the structural and enzymatic proteins of the viral particle (18). The Gag polyprotein of EIAV is about 55 kDa in size and comprises in order the matrix (MA p15), capsid (CA p26), nucleocapsid (NC p11), and p9 core proteins. Despite extensive studies into the molecular mechanisms of retrovirus budding over the past 20 years, the viral and cellular components and processes that mediate virion assembly and budding remain poorly defined.

To focus on the role of Gag proteins in retrovirus assembly and budding, numerous studies have employed assays of virion budding from cells transfected with plasmids expressing Gag polyproteins with or without the viral protease. The results of these studies have in general demonstrated several characteristic Gag functions, including the role of N-terminal myristylation of the MA protein in targeting the Gag polyprotein to the plasma membrane (M-domain) (2, 6, 22, 34, 35), the role of CA domains in multimerization of Gag proteins (I-domain) (7, 17), the role of NC motifs in genomic RNA incorporation, and the role of various Gag late assembly domains (L-domain) in releasing budding virions from the plasma membrane (2, 42, 50).

The retrovirus L-domain is of particular interest, since it is representative of highly specific virus-cell interactions that must occur to direct host cellular processes to achieve virus assembly and budding. Studies to date have demonstrated that different retroviruses may utilize different viral proteins and structural motifs to accomplish the same late budding function. For example, human immunodeficiency virus type 1 (HIV-1) (a lentivirus) (21, 27), Rous sarcoma virus (a type C oncornavirus) (39), human T-cell leukemia virus type 1 (a type C oncornavirus), and Mason-Pfizer monkey virus (a type D retrovirus) (51) all contain proline-rich L-domains consisting of PPXY and/or P(T/S)AP in various Gag protein locations. Interestingly, proline-rich motifs have also been identified in other enveloped viruses, such as vesicular stomatitis virus (a rhabdovirus) (14) and NS3 of bluetongue virus (47), suggesting a common role in viral budding. It has been demonstrated that proline-rich L-domains can in vitro bind specifically with WW-domains of cellular proteins, although there has been no direct demonstration of this putative interaction in virus-infected cells.

In contrast to the proline-rich L-domains frequently found in other viruses, we have used Gag budding assays to define a YPDL L-domain in the Gag p9 protein of EIAV (42) and demonstrated further that this L-domain specifically recruits the cellular adapter protein AP-2 to the sites of viral budding in infected equine cells (43). In addition, the involvement of YXXL domains has recently been demonstrated in tyrosine-based signal transduction and internalization (23, 33, 40) and in the budding and infectivity of other envelope viruses, including HIV-1 (16), bovine leukemia virus (28, 49), and Epstein-Barr virus (4). While the EIAV and HIV-1 L-domains are distinct, we have shown that the YPDL and PTAP motifs are functionally interchangeable and positionally independent in Gag budding assays (39), perhaps indicating different entry points into a common cellular process that facilitates viral budding. Further studies are required to elucidate the virus-cell interactions that so efficiently achieve the assembly and release of viral particles at the plasma membrane.

Although Gag budding assays are useful for determining the function of these proteins in virus assembly and budding, it has been well established that Gag proteins also perform critical functions during virus infection of target cells, indicating the multifunctional nature of these relatively small proteins. Thus, it is likely that the late assembly functions identified with HIV-1 p6 and EIAV p9 reflect only a single aspect of a multifunctional Gag protein. In support of this hypothesis, recent studies have indicated that HIV-1 p6 is also involved in the incorporation of Vpr (3, 29, 31), envelope (36), and polymerase (52) proteins into viral particles, in ubiquitination of Gag polyproteins (41, 45, 47, 48), and in the determination of particle size (19, 20, 27). Thus, it is important that studies of Gag functional properties be performed in the context of virus replication to reveal interactions with other viral and cellular proteins necessary for viral infectivity.

To examine further the function of EIAV p9 in viral replication we have now performed a comprehensive characterization of the replication properties of a series of truncation and site-directed mutants constructed in a reference pathogenic proviral clone, EIAV_uk (12). The results of the studies described herein reveal new important insights into the fundamental mechanisms of EIAV assembly and budding that were not evident in previously reported Gag polyprotein budding assays and also indicate a critical role of p9 sequences distinct from the L-domain in viral infectivity. Thus, these data demonstrate the multifunctional roles of EIAV p9 in replication and suggest complex interactions with other viral proteins and with cellular proteins in viral exit and entry.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

DNA mutagenesis. Two versions, EIAV_uk and cmvEIAV_uk, of proviral DNA plasmids were used for this report. EIAV_uk is an in vivo pathogenic proviral DNA cloned in a low-copy-number plasmid, pLG338-30 (12, 15). When transfected into fetal equine kidney (FEK) and equine dermal (ED) cells, EIAV_uk is replication competent and produces viruses that are also infectious to FEK and ED cells. The p9 mutants in the context of the EIAV_uk backbone were used to test their replication competencies in both FEK and ED cells. The cmvEIAV_uk version is fundamentally the same as EIAV_uk except that a 650-bp-long DNA fragment containing the cytomegalovirus promoter was inserted between the U3 and R regions of the 5' long terminal repeat (LTR) to obtain a higher level of protein expression in transfected Cos-1 cells. Therefore, the cmvEIAV_uk constructs were used when large amounts of viruses were needed for biochemical analysis.

View larger version (29K): [in this window] [in a new window]   FIG. 1.   Summary of p9 truncations engineered into the reference EIAVuk proviral clone. (A) Diagram of the Gag and Pol genome and amino acid sequences of EIAV p9. The boxed YPDL is the L-domain for EIAV budding and release. (B) Schematic summary of p9 truncations. Each p9 mutant is denoted by a single-letter amino acid followed by a number to represent the residue of p9 that is replaced with a stop codon. For example, the K31 mutant designation indicates that the lysine residue at the 31st position of p9 is replaced with a stop codon (asterisk).

Cell transfection. FEK, ED, and African green monkey kidney Cos-1 cells were cultured in minimal essential medium. All media were supplemented with 2 mM L-glutamine, 10% (vol/vol) fetal bovine serum, 100 U of penicillin, and 100 µg of streptomycin per ml. The cell culture products were obtained from Life Technologies Inc. (Gibco BRL). All plasmid DNA preparations used for transfection were isolated using the Midiprep kit (Qiagen). Transfections were carried out using GenePorter 2 reagents (Gene Therapy Systems) according to the manufacturer's recommendations. Typically, 60 mm petri dishes were used in this study unless otherwise noted. Briefly, a total of 6 µg of purified plasmid DNA was mixed with DNA diluent and GenePorter 2 lipid film and then added to cells that were split the day before transfection and at 50 to 70% confluence. One day after transfection, media were replaced with fresh minimal essential medium.

Reverse transcriptase (RT) assays. Approximately 10 to 15 µl of cell culture supernatant was used for each RT activity assay as described previously (30). In certain cases, supernatant virus was first pelleted by ultracentrifugation, and the viral pellet was resuspended for the RT assay. Briefly, the supernatant was mixed with 50 µl of reaction buffer containing 100 mM Tris-HCl (pH 8.0), 12 mM KCl, 10 mM MgCl₂, 10 mM dithiothreitol, 0.1% NP-40, 10 U of poly(rA) · poly(dT)_12-13 (Amersham Pharmacia Biotech), and 0.25 µl of [³H]TTP (TRK424; Amersham). After 1.0 to 1.5 h of incubation at 37°C, the mixtures were transferred to a DEAE membrane (catalog no. 1205-405; Wallac Inc.) and dried. The membrane was then washed three times in 1× SSC (0.15 M NaCl plus 0.015 M sodium citrate) for 10 min each time. The membrane was quickly rinsed with 95% ethanol to facilitate drying. Finally, the membrane was assayed for radioactivity with a 96-well reader in a Wallac Betaplate liquid scintillation counter.

Protein analysis. Clarified transfection supernatants were centrifuged to pellet viral particles through a 20% glycerol-phosphate-buffered saline cushion in an SW55 rotor at 34,800 rpm at 4°C for 50 min. Then the viral pellets were resuspended in phosphate-buffered saline solution for protein analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting. To examine the intracellular expression of viral proteins, transfected Cos-1 cells grown on 60-mm-diameter dishes were lysed with 500 µl of lysis buffer (25 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% deoxycholic acid, 1% Triton 100, protease inhibitor cocktail, 0.1% SDS). Immunoblot analysis was performed as previously described (11). SuperSignal West Pico chemiluminescent substrate (Pierce, Rockford, Ill.) was used to visualize the bands after blotting with a monoclonal antibody against p26 (10) or a reference horse (Lady) immune serum (32). Horse peroxidase-conjugated goat anti-horse immunoglobulin G (IgG) F(ab')₂ was purchased from Jackson ImmunoResearch, and anti-mouse immunoglobulin G peroxidase conjugate was purchased from Sigma (catalog no. A4416).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of EIAV p9 sequences required for replication competence. We have previously reported that the deletion of EIAV p9 or site-directed mutagenesis of the p9 Y₂₃P₂₄D₂₅L₂₆ L-domain produces a defect in the release of budding virus particles from Cos-1 cells transfected with Gag expression plasmids (42). To extend the p9 function observations made in the Gag budding assay, we sought to examine the role of EIAV p9 in viral replication in equine cells. Towards this goal, we generated a series of p9 truncation mutants in our reference infectious proviral clone (12), EIAV_uk, by introducing termination codons at various positions in the p9 sequence, as summarized in Fig. 1. Although these mutations also altered the preprotease sequences in the overlapping pol reading frame, the coding sequences of mature viral protease were not affected and the protease activity was retained (see below). EIAV p9 truncation mutations were designated according to the site of the engineered termination codon. For example, K31 substitutes a termination codon at the position of lysine-31, such that the truncated protein product should contain only the N-terminal 30 amino acids of p9. A total of 12 different C-terminal truncations were constructed to express protein products containing the N-terminal segments of the parental (51 amino acids) p9 protein ranging in length from 2 to 44 amino acids.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Replication properties of EIAV proviral p9 mutants in transfected ED cells as monitored by extracellular RT activity. ED cells were transfected in parallel with the various proviral mutant DNA samples, and supernatants from transfected cells were collected at weekly intervals posttransfection for measurements of RT activity, as described in Materials and Methods. Proviral p9 mutants are under the EIAVuk backbone as described in Fig. 1.

Effects of p9 truncations on extracellular virion production While the preceding experiments define the minimum p9 sequences required for EIAV replication, they do not elucidate the specific defect in p9 truncations that fail to support viral replication. In the light of previous studies indicating the role of EIAV p9 in virus budding, one possible explanation is that replication-defective p9 truncation mutants failed to produce virus particles from transfected cells. To examine this hypothesis, we next assayed the level of extracellular virion production by both replication-competent and -defective p9 truncation mutants compared to the parental EIAV_uk provirus. For these assays we utilized a cmvEIAV_uk proviral plasmid in which viral gene expression is driven by a cytomegalovirus promoter inserted into the viral LTR to increase levels of viral gene expression in transfected Cos-1 cells. The p9 truncations shown in Fig. 1 were subcloned into cmvEIAV_uk expression plasmids, and the resulting constructs were then used to transfect Cos-1 cells. Cos-1 cells are not susceptible to infection by EIAV_uk and thus were chosen as transfection targets to prevent subsequent rounds of infection by progeny virions produced from the transfected proviral clones.

View larger version (50K): [in this window] [in a new window]   FIG. 3.   Effects of p9 truncations on production of extracellular virions. Cos-1 cells were transfected in parallel with equal amounts of cmvEIAVuk plasmid DNA preparations from the indicated p9 truncation mutants. After 3 days, equal volumes of supernatant medium from each culture were collected and subjected to ultracentrifugation to pellet virus particles. The quantities of virions from different p9 truncations were then analyzed by SDS-PAGE and immunoblotting with a high-titer reference equine immune serum to blot virion Gag proteins. Full-length Gag precursor p55, the intermediates p47 (MA-CA-NC) and p40 (MA-CA), and the mature capsid protein p26 are indicated by arrows. The data presented in this figure are representative of at least three independent transfections. Estimations of relative levels of extracellular virion Gag protein production by each proviral mutant were made using densitometer scans of Gag protein bands in the immunoblot.

Effects of p9 truncations on virion RT incorporation and activity. Another possible explanation for the replication-defective phenotype of the p9 truncation mutants is that the p9 protein may perform a role in Pol incorporation into viral particles, as previously described for HIV-1 p6 (52). Alternatively, truncations in p9 may cause defects in Gag-Pol processing, preventing the activation of protease and the maturation of RT in viral particles. To test directly the influence of p9 truncations on virion RT incorporation and activity, we analyzed the levels of RT activity in viral particles produced by replication-competent and -defective p9 truncations, normalized to total Gag protein content. After 3 days, virus particles produced from the Cos-1 cells transfected with each p9 truncation mutant were pelleted from the culture medium by ultracentrifugation and resuspended to achieve a standard concentration of EIAV Gag proteins. Then equal amounts of virus were assayed for virion RT activity (Fig. 4). The results of these assays revealed similar amounts of RT activity per unit of Gag protein, regardless of the portion of p9 contained in the Gag polyprotein. For example, virions produced from a truncation mutant (Q3) containing only the N-terminal two amino acids of p9 displayed about 90% of the RT activity observed in the same amount of viral particles produced by the parental EIAV_uk provirus containing a full-length p9 protein. Similarly, the replication-competent E32 and replication-defective K31 viral particles displayed virtually identical normalized RT values. These data indicated that the EIAV p9 protein does not influence incorporation of Pol into viral particles or the cleavage processing of Gag-Pol precursor proteins to active RT in the virion. Thus, the replication-competent and replication-defective phenotypes of p9 truncation mutants evidently cannot be correlated with RT incorporation or processing. The cleavage and activation of the RT enzyme from the Gag-Pol polyprotein also support the conclusion that the various p9 mutations did not affect viral protease activity, despite modifications in the preprotease sequences in certain mutants.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Effect of p9 truncations on incorporation of RT activity into progeny virions. Cos-1 cells were transfected in parallel with equal quantities of the indicated cmvEIAVuk proviral plasmid DNAs using standard transfection conditions. After 3 days, virus particles were pelleted by centrifugation of equal volumes of cell supernatant medium, and the quantity of EIAV Gag p26 was determined by Western blotting with a reference murine monoclonal antibody, as described in Materials and Methods. Based on p26 content, equal amounts of EIAV virions were then assayed for RT activity. The presented data represent the averages of results from at least three experiments.

Effects of p9 truncations on Gag protein expression and processing. The preceding studies indicated similar levels of virion production and RT activity in viral particles produced by replication-competent and -defective p9 truncation proviral mutants but did not address the issue of Gag polyprotein expression in transfected cells and maturation cleavages in budding virions. Thus, we next compared the levels of Gag polyprotein expression and processing in Cos cells transfected with selected replication-competent and replication-defective p9 mutant proviruses. As described above, we once again utilized our cmvEIAV_uk expression plasmids to achieve high levels of proviral gene expression in transfected Cos-1 cells. After 3 days, the supernatants were harvested from the transfected cells and analyzed for virus production, while the transfected cells were lysed, and the cellular extracts were analyzed for EIAV Gag protein expression and processing.

View larger version (41K): [in this window] [in a new window]   FIG. 5.   Effects of p9 truncations on Gag protein expression and proteolytic processing. Equal amounts of plasmid DNA samples containing the respective p9 truncation mutations were used to transfect Cos-1 cells. After 3 days, the transfected cells and supernatant medium were collected for analysis of intracellular and virion-associated EIAV Gag proteins, respectively. (A) Transfected cells were lysed, and intracellular EIAV Gag protein expression was analyzed by SDS-PAGE and Western blotting using the reference (Lady) equine immune serum. (B) Supernatant media from the transfected cells were subjected to ultracentifugation to pellet virus particles that were resuspended and quantified for p26 content using a murine monoclonal antibody. Based on virion p26 content, equal amounts of pelleted virus were then analyzed by SDS-PAGE and immunoblotting with Lady serum for Gag protein profiles. The various p9 mutants are listed on the top of the gel. The full-length Gag precursor p55 is indicated with an arrow. The p9 truncation mutations generated a series of Gag precursors with decreasing molecular weights from right to left, as expected. The major processing intermediates, p47 (MA-CA-NC) and p40 (MA-CA), and mature capsid protein p26 are also indicated by arrows.

Role of the K₃₀K₃₁ sequences of p9 in viral replication. The preceding studies together focused attention on the K₃₀K₃₁ segment of EIAV p9 as being critical for the maintenance of replication competence. In this regard, recent studies have revealed that 2 to 5% of HIV-1 p6 found in virions is monoubiquitinated at lysine residues, perhaps suggesting an involvement of ubiquitination in HIV-1 assembly or infectivity (38, 45). We next examined the potential role of this p9 segment in viral replication by mutating the K₃₀K₃₁ doublet to M₃₀M₃₁ in the context of the EIAV_uk proviral clone. Methionine substitutions were selected to maintain the general size properties of the lysine residues, while changing the charge and abolishing the potential for ubiquitination at this site. The replication properties of the KK/MM p9 mutant were then compared in parallel to those of the parental EIAV_uk in transfected ED cells, as described for the p9 truncation mutants (cf. Fig. 2). The data in Fig. 6 demonstrated similar replication kinetics and levels for the KK/MM p9 mutant and the parental EIAV_uk provirus in transfected ED cells. Thus, these data indicated that the K₃₀K₃₁ sequence of p9 could be functionally replaced by an MM sequence, evidently suggesting that the charged residues and their potential ubiquitination were not required for viral budding and infectivity of EIAV. These observations are consistent with recent studies indicating that the two lysine residues in HIV-1 p6 that are targets for ubiquitination can be mutated with any effect on virus replication (37).

View larger version (14K): [in this window] [in a new window]   FIG. 6.   Role of the K30K31 sequences of EIAV p9 in viral replication. The EIAVuk proviral clone was specifically mutated to substitute methionine residues for both the K31 and K32 residues. The replication properties of the KK/MM mutant were then analyzed by transfection of ED cells and compared to a parallel transfection with an identical amount of the parental EIAVuk proviral plasmid. Viral replication was monitored by extracellular RT activity, as described in the legend for Fig. 2.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have previously reported experiments using Gag polyprotein budding assays that indicated a critical role for the EIAV p9 protein in the release of budding virions from transfected Cos-1 cells, similar to the assembly role defined for HIV-1 p6 (8, 42). Based on selected mutagenesis of the EIAV p9 protein, we further defined a YPDL motif in p9 that provided the late budding function mediated by PTAP-based L-domains in HIV-1 Gag polyprotein budding, and we demonstrated a specific interaction of EIAV p9 with the adapter protein AP-2 in virus-infected equine cells (43). The results of these studies suggest that the EIAV p9 protein, and the YPDL L-domain in particular, may specifically recruit components of the cellular endocytosis machinery for viral budding from plasma membranes of infected cells.

Despite the informative nature of the Gag polyprotein budding assays for assessing p9 function, we remained concerned about two limitations of these functional assays. First, the functions of p9 are defined in the absence of the expression of other non-Gag viral proteins that may also contribute to viral assembly and budding. For example, interactions between Gag and Env proteins influence polarity of virus budding and release (13, 26, 36). In addition, certain accessory gene proteins, such as Vif, appear to influence virus assembly in HIV-1 (5) and caprine arthritis encephalitis virus (24), and Vpu can enhance virus particle release through an L-domain-independent mechanism (46). Second, the Gag polyprotein budding system monitors p9 functions related to virus assembly but does not monitor the potential functions of the p9 protein in virus infection of target cells. The latter limitation has become increasingly relevant as there has been accumulating evidence that retroviral Gag proteins perform critical functions during both viral infection and budding. Therefore, we initiated the present studies to examine the functional roles of EIAV p9 in the context of an infectious proviral clone and a full complement of viral protein expression in infected or transfected cells.

The results of the present studies indeed reveal for the first time novel aspects of p9 function in viral replication that were not evident from Gag polyprotein budding assays. The analysis of a series of p9 truncation mutants in the context of a reference proviral construct clearly demonstrated that only the first 31 amino acids of the p9 protein are required for replication competence, suggesting that the C-terminal 20 amino acids are not essential for viral budding or infection. However, it should be noted that replication-competent p9 proviral mutations did display delayed replication kinetics and lower replication levels than the parental EIAV_uk proviral clone, indicating an optimization of p9 functional interactions by the complete protein. It is noteworthy that all of the replication-competent p9 truncation mutants retained the YPDL L-domain of p9 and that replication-defective truncation mutants lacking the YPDL sequences displayed threefold-lower levels of virus budding than parental EIAV_uk. These observations are in general in agreement with the previously defined role of p9 in virus budding from the plasma membrane. However, several of the replication-defective p9 truncation mutants also contained the YPDL L-domain, indicating the presence in p9 of other functional domains necessary for viral replication.

In this regard, a remarkable observation was the absolute demarcation between replication-competent and replication-defective phenotypes in the p9 C-terminal truncation series with the progression from the E₃₂ to K₃₁ mutants that differed by only one amino acid. These p9 variants displayed similar levels of expression and processing of Gag polyproteins, production of extracellular virions, and incorporation of virion-associated RT activity. Despite these similarities in viral gene expression, processing, and viral budding, the virions produced by the E₃₂ mutant were replication competent and the K₃₁ virions were completely replication defective. These observations indicate for the first time a critical function of the p9 protein in viral infectivity, in addition to its previously defined role in viral budding. Based on the hypothesis that the EIAV p9 protein recruits cellular endocytosis machinery to facilitate viral budding, it is possible that the p9 protein may interact with cellular endocytic molecules during viral penetration and uncoating in target cells. This model then implies that the N-terminal 30-amino-acid sequence of p9 is functionally budding competent and infectivity defective, while the addition of a single amino acid restores the infectivity function.

The replication phenotypes associated with the series of C-terminal p9 proviral truncation mutants clearly highlighted the highly charged K₃₁E₃₂ sequences as a critical functional borderline in terms of viral infectivity and thus replication competence. In light of recent reports indicating the importance of the ubiquitination of HIV-1 p6 lysine residues for viral assembly, it seemed plausible that the functional properties of EIAV p9 may depend on ubiquitination of the K₃₀K₃₁ residues at the functional border. To test this hypothesis directly, we engineered and evaluated the replication properties in transfected ED cells of a proviral mutant in which the K₃₀K₃₁ residues of p9 were mutated to M₃₀M₃₁. The results of these transfection experiments revealed that the KK/MM p9 proviral mutant displayed replication kinetics and levels similar to those of the parental EIAV_uk provirus in transfected cells. These observations demonstrated conclusively that the K₃₀K₃₁ residues are not required for EIAV budding or infectivity, and thus ubiquitin modification of these particular lysines cannot be an important factor in viral replication, although it does not exclude the possibility of the ubiquitination of other lysine residues in p9. Thus, these EIAV studies suggest that additional experiments are required to assess further the importance of ubiquitination as a common factor in the budding and infectivity of retroviruses.

In addition to revealing a functional role of EIAV p9 in viral infectivity, the comparative study of the series of p9 truncation proviral mutations also elucidated new aspects of the role of p9 in viral assembly and budding that were not evident from earlier Gag budding assays (42). For example, Gag budding assays indicated an absolute requirement for the YPDL L-domain for virion budding from transfected Cos-1 cells; Gag mutants lacking the p9 protein or with a YPDL deletion or specific L-domain point mutations in the Gag polyprotein failed to produce detectable extracellular particles from transfected cells (cf. Fig. 5 and reference 42). In contrast, the present experiments using p9 mutations in the context of a complete provirus genome clearly demonstrated that virion budding was reduced by only about threefold in truncation variants lacking the YPDL L-domain. These marked differences in the influence of the p9 mutation and its L-domain on virion budding efficiency in the Gag polyprotein and proviral budding assays cannot be attributed to the differences in experimental conditions. Both assays utilized transfection of Cos-1 cells with expression plasmids that produced similar levels of viral protein. Thus, we interpret these observations to imply that other non-Gag proteins may also contribute to viral budding, perhaps by providing a redundant and somewhat less efficient L-domain budding function. These results also emphasize the importance of assessing viral protein functions in the context of replicating virus and natural target cells.

The present studies also demonstrated for the first time new information about the role of EIAV p9 and other viral proteins during assembly. Several published reports have indicated the importance of HIV-1 p6 in the incorporation of viral RT in virions (52), but this function does not appear to be associated with the EIAV p9 protein. Both replication-competent and -defective p9 truncation proviral mutants contained similar levels of RT activity when normalized to virion Gag protein content. Thus, the infectivity defect observed in proviruses containing less than the N-terminal 31 amino acids of p9 could not be attributed to a lack of Pol protein incorporation or processing to active RT. We interpret the EIAV data to indicate that the primary function of EIAV Gag p9 is interaction with cellular proteins rather than other viral proteins. In this regard, it is of interest that localization experiments indicate a lack of association of p9 with other viral core proteins (44). Also, the fact that Gag p6 is absent from HIV-1 core preparations (1) is consistent with our interpretation.

Taken together, our data are consistent with a general model in which EIAV p9 performs at least two functions during viral replication, one related to viral budding and the other related to viral infectivity. The viral budding function has previously been correlated at least in part with a specific recruitment by p9 of the cellular adapter protein AP-2 at the site of viral budding. It is possible that the infectivity function of p9 may be mediated by a similar interaction, since the cellular adapter proteins are part of the endocytic machinery of the cell and localized to the plasma membrane where virus penetration occurs. Alternatively, p9 may interact with different components of the endocytosis and exocytosis machinery of the cell during viral budding and infection. These possible mechanisms are currently under investigation. Regardless of the exact functional mechanism of EIAV p9, the present studies support the concept that the "structural" proteins of retroviruses likely perform multiple functions in adapting cellular processes to efficiently accomplish viral assembly and budding from cells and effective penetration and uncoating during infection of target cells. A better understanding of these Gag-related functions can lead to the identification of new virus-specific functions that can be targeted by novel antiviral drugs.