Cooperative role of MHR and the CA dimerization helix in the maturation of the functional retrovirus capsid

Abstract

The second helix in the C-terminal domain of retroviral capsid (CA) protein functions as the site of dimerization between subunits in capsid assembly and is believed to participate in a unique interface between Gag molecules in immature particles. This study reports isolation of two substitutions in the dimerization helix of Rous sarcoma virus CA protein that have the ability to suppress lethal defects in core maturation imposed by alterations to the major homology region (MHR) motif just upstream. Together with two previously published suppressors, these define an extended region of the dimerization helix that is unlikely to contribute directly to CA-CA contacts but whose assembly-competence may be strongly affected by conformation. The broad-spectrum suppression and temperature-sensitivity exhibited by some mutants argues that they act through modulation of protein conformation. These findings provide important biological evidence in support of a significant conformational change involving the dimerization helix and MHR during maturation.

Introduction

A highly conserved motif named the major homology region (MHR) lies within the CA region of the Gag structural proteins of orthoretroviruses and a diverse collection of other reverse transcribing elements like hepadnaviruses, the yeast retrotransposon Ty3, LINE elements, and telomeric retrotransposons of Drosophila (Craven et al., 1995; McClure, 1991; Orlinsky et al., 1996; Pardue et al., 2005; Wills and Craven, 1991; Zlotnick et al., 1998). Outside of the MHR the amino acid sequences of CA proteins are quite diverse, yet all share a common structural organization with two independently-folded α-helical domains, the N-terminal domain (NTD) and C-terminal domain (CTD) (Fig. 1A, 1B) (Berthet-Colominas et al., 1999; Campos-Olivas et al., 2000; Cornilescu et al., 2001; Gamble et al., 1996; Gamble et al., 1997; Gitti et al., 1996; Jin et al., 1999; Khorasanizadeh et al., 1999; Kingston et al., 2000; Momany et al., 1996; Mortuza et al., 2004; Tang et al., 2002; Worthylake et al., 1999). The MHR spans approximately 20 amino acids of the CTD, specifically the strand immediately following the inter-domain linker through the turn and first α-helix of the domain (Fig 1B). Conserved hydrophobic residues of the MHR contribute to the core of the folded domain, while polar ones form a hydrogen-bond network that links together the strand and first two α-helices (Kingston et al., 2000; Worthylake et al., 1999).

Figure 1.

(A) Alignment of the RSV and HIV CA CTD sequences. Position of α-helices is marked by string of H’s. MHR is underlined, with the identical and similar residues bolded. Position of suppressors in the dimerization helix is bolded. (B) Position of the MHR residues and their second-site suppressors in the RSV CA structures. Ribbon diagram was created using the PDB files 1EM9 and 1EOQ for the NTD and the CTD respectively and connecting them by a dotted line. The two domains are not drawn to scale. Start and end of the MHR is marked by arrow-heads. The side-chains of the MHR residues examined in this study (in blue) and the residues where suppressor mutations map (in red) are shown. The rescue of the L171V mutant virus by the A38V suppressor mutation was previously reported by our lab. (C) Position of the second-site suppressor mutations on the presumptive model of RSV CTD dimer. The model was created by superimposing the RSV CTD structure (PDB file 1EOQ) onto the HIV-1 CTD dimer (PDB file 1A43) (Kingston et al., 2000; Worthylake et al., 1999). Structural alignment was done using the program SARF2 (Alexandrov and Fischer, 1996).

The evolutionary conservation of the MHR predicts a critical role of the motif in one or more stages of the viral life cycle and the retrotransposition process. Extensive mutagenesis of the MHR in the Rous sarcoma virus (RSV, a member of the Alpharetrovirus genus), other retroviruses and the Ty3 retrotransposon has demonstrated a range of phenotypes caused by alteration to the conserved residues. The observed effects on Gag function include interference with normal Gag protein membrane targeting or binding, distortion of the morphology of released particles or a complete loss of immature particle assembly (Alin and Goff, 1996; Cairns and Craven, 2001; Craven et al., 1995; Mammano et al., 1994; Orlinsky et al., 1996; Strambio-de-Castillia and Hunter, 1992; von Schwedler et al., 2003; Willems et al., 1997). Other substitutions allow the normal formation and release of particles but cause a subsequent failure of the structures to mature normally to form functional core structures (Alin and Goff, 1996; Cairns and Craven, 2001; Craven et al., 1995; Mammano et al., 1994; Orlinsky et al., 1996; Strambio-de-Castillia and Hunter, 1992; von Schwedler et al., 2003; Willems et al., 1997). This suggests overlapping roles of the motif in both the Gag functions needed to build an immature particle and in the creation of the capsid shell from the CA subunits that are subsequently released upon protease activation at or near the time of provirion release. A total of twenty-five substitutions introduced at the most conserved positions of the RSV MHR motif were unable to identify particular sub-domains of the region that are critical for one part of the assembly pathway but not the other.

A structural study of the dimerized CTD of the CA of human immunodeficiency virus type 1 (HIV-1) has provided a model that may help explain how the MHR contributes to Gag assembly. Relaxation of the bend between the MHR and the adjacent dimerization helix by deletion of a single residue allowed the CTD to dimerize in a “domain-swapped” arrangement creating an extensive anti-parallel interface (Ivanov et al., 2007). Several of the conserved MHR residues, instead of forming only intra-domain interactions, establish intermolecular interactions across the dimer interface that would appear to explain their conservation. Because of the structural homology of this domain-swapped HIV CTD dimer to the dimer of SCAN interaction domains of cellular proteins that are evolutionarily related to Gag (Ivanov et al., 2005; Kingston and Vogt, 2005), it was predicted that this arrangement exists between Gag proteins in immature virions. This proposed mode of dimerization is consistent with the loss of normal Gag assembly observed in many MHR mutants. However, a recently published cryo-electron tomographic study of immature HIV could neither confirm nor refute its presence (Wright et al., 2007). Thus, additional biological evidence beyond the loss-of-function studies available to date will be necessary in validating its importance.

The means by which the MHR residues contribute to formation of the mature core structure is less clear. The accepted model for the mature capsid shell formation posits that a hexameric lattice is formed via interactions between the α-1, α-2 and α-3 helices of neighboring NTDs and that inter-hexamer associations are accomplished via a dimeric interface involving the α-2 helix (dimerization helix) of the CTD in parallel alignment (Briggs et al., 2003; Ganser et al., 1999; Ganser-Pornillos et al., 2007; Li et al., 2000; Mayo et al., 2003; Mortuza et al., 2004; Yeager et al., 1998). Analysis of 2-D crystals of the murine leukemia virus, HIV and RSV proteins dictate that the same packing mode likely applies to all orthoretroviruses (Barklis et al., 1998; Ganser et al., 2003; Mayo et al., 2002). Additional inter-domain contacts in which the NTD docks with the CTD of a neighbor appear to form only during maturation and are a hallmark of the final assembled capsid (Ganser-Pornillos et al., 2007; Lanman et al., 2003; Lanman et al., 2004). The MHR residues do not contribute directly to any of these known CA-CA interfaces.

The isolation of genetic suppressors of crippling CA mutations by our laboratory has proven to be a novel means of identifying interactions between distant residues of CA that are important for normal core assembly and function. Beginning with RSV viruses bearing MHR mutations that prevent formation of a functional core structure without obvious alterations of Gag functions, four second-site suppressors were isolated that could restore part or all of the infectious potential (Bowzard et al., 2001). Two suppressors fell in the NTD, predicting the involvement of an NTD-CTD interaction at some point in assembly or maturation of the particle (Bowzard et al., 2001). Such an interface in mature HIV virions, but not immature particles, has since been confirmed by biochemical and structural studies (Ganser-Pornillos et al., 2007; Lanman et al., 2003; Lanman et al., 2004). An additional suppressor, I190V, mapped to the second CTD helix; a fourth fell in the cleavage site separating CA and the downstream peptide.

This present study describes the isolation of two additional MHR suppressors that lie on the dimerization helix. These, together with the previously described I190V MHR suppressor, define an extended region that overlaps the residues that are involved in subunit CTD-CTD dimerization. The position of the MHR suppressors in CA as well as the broad-spectrum nature of suppression argues that the mutations likely work by controlling conformation of the protein. This is further supported by the temperature-sensitive growth of some of the mutants. These findings provide important biological evidence in support of a major structural change in the CA CTD involving the MHR and the α-2 helix such as that described by Ivanov et al. during the maturation and assembly of the functional core.

Results and Discussion

MHR mutants are rescued by second-site mutations in the α-2 helix of the CTD

In the first MHR suppressor hunt, three substitutions in the α-helical portion of the MHR yielded four suppressors located in distant regions of the CA protein structure (Bowzard et al., 2001). To assess whether or not MHR mutations that lie in structurally distinct locations can be corrected by a similar array of secondary changes, we have since sought second-site suppressors of a D155Y mutation (which lies on the strand at the upstream boundary of the MHR) and an E162Q substitution (immediately preceding the α-helix, contributing to the H-bonding network in the domain) (Fig. 1B). Both the D155Y and E162Q mutations allow release of particles that exhibit defects in core integrity (Cairns and Craven, 2001). D155Y and E162Q revertant viruses were isolated by introducing proviral plasmids bearing the MHR mutations into turkey embryo fibroblasts by transfection to allow synthesis of the crippled viruses, and these were allowed to evolve during several weeks of serial passages (Bowzard et al., 2001). Periodic sampling of the culture medium detected a rise in reverse transcriptase activity indicative of a reversion event. Sequences containing the original MHR mutations were detected in the genomic DNA of each culture and, in addition, a novel mutation was detected in each (R185W in the D155Y culture and F193L in the E162Q culture).

Each of the two suspect suppressors was recreated in a GFP-coding proviral vector in combination with the MHR mutation with which it had evolved and also in an otherwise wild-type (WT) plasmid. The potential of each to suppress the MHR mutation was assessed in DF1 (chicken) cells at 37°C, monitoring the spread of virus through the cell population by flow cytometry to detect GFP expression (Figs. 2A, 2B). The growth rate of each single- and double-mutant virus was calculated from the exponential phase of GFP spread (Table 1)(Wang and Bushman, 2006). Under these assay conditions, growth of virus bearing the single D155Y MHR mutation was undetectable, i.e. indistinguishable from mock-infected cultures. However, the infectivity was restored to half that of the WT virus by addition of the R185W mutation (Fig. 2A). Similarly, the very sluggish growth of E162Q virus (about one-tenth that of the WT virus) was reversed to within half of the WT growth rate by the second-site F193L substitution (Fig. 2B). Thus, both newly identified CA mutations have suppressor activity. Not surprisingly, neither F193L nor R185W was lethal when present alone, as was documented for the previously published MHR suppressors.

Figure 2.

Confirmation of suppression activity of the R185W and the F193L mutations. (A) R185W suppressor mutation rescues the infectivity of the D155Y MHR mutant. (B) F193L suppressor mutation rescues the infectivity of the E162Q MHR mutant.

Table 1.

Summary of the viral growth kinetics of the single- and double-mutant viruses. Mean growth rate and standard deviation for each of the viruses was calculated from at least three independent experiments as described by Wang and Bushman (Wang and Bushman, 2006). Viral infectivity of each mutant was then compared to the wild-type virus, and expressed by the following symbols: “++++” >90 % of the WT virus, “+++” 65–90% of the WT virus, “++” 40–65% of the WT virus, “+” 15-40% of the WT virus, “-” <15% of the WT virus.

	Substitutions in the MHR
	WT	D155Y	E162Q	F167Y	L171V
	+ + + +	−	−	−	−
	1.03±0.05	<0.05	0.08±0.09	0.08±0.02	<0.05
	+ + +	+ +	+ + +	+ + + +	−
	0.83±0.08	0.56±0.05	0.70±0.15	1.01±0.03	<0.05
	+ + +	−	+ +	−	−
	0.75±0.03	<0.05	0.60±0.07	<0.05	<0.05
	+ + +	−	+	+ + +	−
	0.76±0.05	0.05±0.04	0.21±0.10	0.92±0.02	<0.05
	+ + +	−	−	+ + +	+ +
	0.68±0.01	<0.05	<0.05	0.71±0.04	0.54±0.05

The R185W and the F193L suppressors lie on the α-2 helix of the CTD, one on either side of the I190V suppressor that was originally described as a suppressor of the crippling F167Y MHR mutation. This helix is well documented to serve a critical role in capsid assembly as the principal site of CTD-CTD dimerization in the mature capsid shell. It has also been implicated in Gag protein dimerization (Datta et al., 2007; Ivanov et al., 2005; Ivanov et al., 2007). A fourth mutation, C192R, described by Spidel et al. as a suppressor of a β-hairpin mutation (K5R) in the NTD of RSV CA, falls within the same region of the α-2 helix. Thus, we have identified a cluster (“hot-spot”) of four suppressors along this helix.

A structure of the mature, dimerized CTD of RSV has not been published as yet. However, it was possible to examine the relationship of the residues involved in suppression to the presumed interface by utilizing a model of the RSV CTD dimer based upon the well-established HIV-1 CTD dimer structure (R. Kingston, pers. comm.) (Fig. 1C). The RSV CTD dimer model was generated by aligning the structure of the RSV CTD (PDB file E1OQ) and the HIV-1 CTD dimer (PDB file 1A43) using SARF2 program, which only generated one solution due to similarity in the length and orientation of the helices in the two structures (Alexandrov and Fischer, 1996; Kingston et al., 2000; Worthylake et al., 1999). In the model of the RSV CTD dimer, it is clear that the F193, I190 and R185 residues span the region that is likely to be involved directly in the dimer interface but are unlikely to participate in the interface. Rather, the side-chains of each point away from the predicted interface with F193 buried in the hydrophobic core of the CTD, while I190 and R185 are more exposed. Furthermore, examination of the published model for the CTD monomer shows that none of the MHR residues are situated in a position that would place their side chains in the proximity of the residues responsible for suppression. Preliminary attempts to model the RSV CTD protein upon the proposed immature HIV-1 CTD dimer (Ivanov et al., 2007) were unsuccessful due to limited amino acid similarity between the two proteins outside of the MHR, but a comprehensive computational study has not been conducted. However, in the immature HIV-1 CTD dimer model, the side-chains of the MHR residues and the side-chains of the residues corresponding to the RSV suppressors were not in close vicinity (Ivanov et al., 2007). This argues that the second-site suppressors do not act by making direct contact with the MHR residues or by directly affecting the CTD-CTD dimerization interface. The R185W residue, in contrast to the other nearby suppressors, is located near the top of the dimerization helix. The analogous position in the HIV protein is involved directly in the NTD-CTD interface, raising the possibility that suppression of the D155Y defect by R185W in RSV involves strengthening of the inter-domain interface (Ganser-Pornillos et al., 2007; Lanman et al., 2003). No models of this interface in assembled HIV CA are as yet of sufficient resolution to allow the modeling of the RSV interface to examine this interpretation more carefully.

MHR mutations and suppressor mutations are not allele-specific

All the MHR suppressor mutations isolated lie close to or in regions that have been implicated in the process of maturation of the virus capsid, suggesting that compensation could be accomplished by modulating protein conformation and/or the kinetics of the maturation process. If so, then some degree of cross-suppression between different alleles might be predicted. To test this, the infectivity of mutant viruses bearing double mutations in various combinations was measured. In addition to the three suppressors on the dimerization helix, the L171V MHR mutation and its A38V suppressor in the NTD were included to evaluate whether or not inter-domain suppression is allele-specific. Table 1 compares the average growth rate for each virus at 37°C. The results for the mutants carrying a single MHR mutation or either of the previously described suppressors (I190V and A38V) are consistent with published infectivity data obtained with turkey embryo fibroblast cells using less quantitative methodologies (Bowzard et al., 2001).

The pattern of suppression was found to vary depending upon the MHR allele examined. The E162Q and F167Y mutations were each effectively suppressed by more than one secondary mutation (Fig. 3A, 3B). The growth defect of E162Q was corrected as readily by R185W as it was by the original suppressor F193L (Fig. 3A). Suppression of the E162Q phenotype by I190V was less effective but still readily detectable. In contrast, the NTD substitution A38V not only failed to suppress but actually ablated the weak infectivity of E162Q. The growth-crippling F167Y MHR allele exhibited a unique suppression pattern—it was strongly suppressed by R185W and the inter-domain suppressor A38V but not at all by F193L (Fig. 3B). In fact, the F193L secondary mutation eliminated the slight residual replication that is characteristic of the F167Y single mutant. In addition to A38V and R185W, the replication defect in F167Y was also shown previously to be suppressed by a single substitution S241L in the cleavage site at the end of CA (Bowzard et al., 2001). Notably, the R185W substitution (originally isolated as a suppressor of D155Y) was even more effective in correcting the growth defects of E162Q and F167Y making it a powerful broad-spectrum suppressor.

Figure 3.

Rescue of the MHR mutants by the suppressor mutations is not allele-specific. Viral growth kinetics of allele-swapped double-mutant viruses was measured as in Fig 2. (A) The E162Q MHR mutation in combination with various suppressor mutations. (B) The F167Y MHR mutation in combination with various suppressor mutations.

In sharp contrast to the above, the D155Y and L171V MHR substitutions, both of which cause a complete loss of detectable replication at 37°C, were resistant to correction by any secondary change other than the original suppressor (R185W and A38V, respectively, Table 1). This is likely a reflection of the profound effects of these MHR mutations on protein structure and virus replication. Since these same two suppressors each have some broad-spectrum activity, it is clear that none of the pairs of MHR mutations and suppressors that were originally associated with them are truly specific for one another.

Several mutants exhibit temperature-sensitivity

As might be expected of mutations that affect either folding pathways or protein conformation, some of the mutants exhibit temperature-sensitive growth. Those single- and double-mutants that showed detectable growth at 37°C were also screened for growth at 42°C. Three substitutions proved to have a temperature-dependent effect on growth (the A38V suppressor in the NTD and the F167Y and E162Q MHR mutations). The slow growth of the F167Y and E162Q viruses was completely inhibited by increasing the growth temperature at 42°C, whereas the elevated temperature had no deterrent effect on the growth rate of WT virus (Fig. 4A, 4B). The E162Q/F193L, E162Q/R185W and E162Q/I190V double-mutants were also temperature-sensitive (Fig. 4A, data not shown), although viruses bearing the F193L, R185W, or I190V suppressor without an MHR mutation were not. Thus, no suppressor was able to counteract the temperature-sensitivity due to E162Q. The NTD A38V substitution also caused a very strong temperature-sensitive effect that was observed both in the context of the single-mutant or in combination with either L171V or F167Y (Fig. 4B, 4C).

Figure 4.

Temperature sensitivity of mutants. Viruses were tested for their growth kinetics at 37°C and 42°C. (A) Temperature sensitive phenotype of the E162Q/F193L and the E162Q mutant. (B) Temperature sensitive phenotype of the L171V/A38V and the A38V mutant. (C) Temperature sensitive phenotype of the F167Y/A38V and the F167Y mutant.

When the E162Q/F193L, L171V/A38V or WT viruses produced at 37°C (permissive temperature) were used to infect DF1 cells growing at 37°C and at 42°C and the percent infected cells were measured 24 hrs post-infection, it was noted that a reproducibly higher percentage of cells were infected when the target cells were growing at 42°C than at 37°C. Similar result was also obtained when the viruses were produced at 42°C. This argues that the temperature-sensitive growth of these viruses is not due to the effect of temperature on the early steps of the virus life-cycle (i.e. virus attachment, entry, reverse-transcription, nuclear entry, integration or expression).

Finally, of the four MHR mutations examined above only E162Q causes a detectable effect on Gag function—the efficiency of virus assembly and release at 37°C is diminished to ~60% that of the WT in QT6 cells (Cairns and Craven, 2001)(Fig. 5). This reduced particle release was restored to normal by the F193L suppressor indicating that at least some of the replication deficit in E162Q is due to effects on Gag assembly function (Fig. 5). When the virus release in two independent experiments was normalized for the amount of expression, the budding efficiency at 37°C was calculated to be 54% for E162Q, 87% for F193L and 89% for E162Q/F193L as compared to the WT. Although F193L suppressor could improve the budding efficiency of the E162Q virus, it could not rescue the infectivity of the virus at 42°C, suggesting the presence of an additional defect presumably during core maturation.

Figure 5.

Budding of the viruses from QT6 cells at 37°C. The immuno-precipitated viral proteins harvested from the lysates of the cells labeled for 5 min (L) and from the media of the cells labeled for 2.5 hrs (M) were detected by radiography.

Implications

The retroviral CA protein is a dynamic protein that possesses a dramatic flexibility in the range of intermolecular interaction it forms and the ability of the individual domains to change conformations upon cleavage, mutations or binding of various molecules (Alcaraz et al., 2007; Barrera et al., 2006; Berthet-Colominas et al., 1999; Gamble et al., 1997; Gross et al., 2000; Jin et al., 1999; Nandhagopal et al., 2004; Ternois et al., 2005; von Schwedler et al., 1998; Worthylake et al., 1999). Both the inherent intra-domain flexibility and flexibility in intermolecular contacts are likely to be critical to the ability to form a capsid shell (Ganser-Pornillos et al., 2007; Li et al., 2000). The ability of the CA to successfully undergo the dynamic events needed to support different functions at different points in the replication cycle almost certainly requires a balancing of various structural constraints that has been finely tuned through evolution. Our data suggest that the MHR and the dimerization helix modulate conformational changes during maturation.

Genetic screens for second-site suppressor mutations have proven a useful tool to identify distant regions of CA that cooperate to control the conformation changes and/or binding events that are needed to build a functional capsid. The clustering of four suppressor mutations along the dimerization helix indicates that, in addition to serving as a primary point of contact between CTDs in the mature capsid, this helix as a whole is probably a site of significant conformational change during maturation. Furthermore, it appears to work in cooperation with the MHR to control the ability of the CTD to achieve an assembly-competent conformation during maturation. This is supported by the fact that some of the MHR mutations disrupt the ability of the RSV CA protein to assemble in vitro into capsid-like structures, while the corresponding suppressors rescue this phenotype (Purdy et al., manuscript submitted). Further, it suggests that MHR mutations are lethal due to effects on the conformation of the following helix, a prediction that is currently being tested by our laboratory. Control of the α-2 conformation by the MHR residues and their suppressors is supported by both the broad-spectrum suppression exhibited by three secondary substitutions (R185W, I190V and A38V) and by temperature-sensitivity in case of E162Q and F167Y. Importantly, these mutations provide biological evidence in support of the suggestion that the formation of a slight kink in the helix (exactly overlapping I190 and F193) during maturation may be crucial to the ability of the protein to achieve assembly-competence (Campos-Olivas et al., 2000; Ivanov et al., 2007; Khorasanizadeh et al., 1999; Kingston et al., 2000).

The R185W suppressor lies in a position where, in addition to potentially affecting α-2 conformation, it may act by directly modulating the formation of NTD-CTD interface (Ganser-Pornillos et al., 2007; Lanman et al., 2003). Two MHR suppressors were previously described in the NTD. One of these (P65Q at the base of α-4 NTD helix) is predicted to participate directly in the inter-domain interface. A38V, however, is not expected to participate directly in interfaces but does lie within a region (α-2 helix in the NTD) where it could induce conformational effects that could affect NTD multimerization and the ability of the NTD to form inter-domain interactions (Ganser-Pornillos et al., 2007; Mortuza et al., 2004). Again, such a mode of action is supported by both its ability to suppress two different MHR alleles and by its temperature-sensitivity.

The promiscuous suppression of F167Y by four different secondary mutations that alter the CTD, the NTD, or the rate of proteolytic cleavage of the downstream spacer-peptide from the C-terminus of the CA is particularly intriguing. In the proposed immature interface, this residue participates in a hydrophobic interaction across the interface, but upon maturation it must move into an internal position within the CTD core. We suggest that the F167Y substitution, which is tolerated by Gag during immature assembly, cripples normal core formation by affecting the propensity of the domain to proceed normally through the conformational and folding changes that are triggered by proteolytic processing. The same interpretation could apply to the L171V substitution that lies one turn of the helix past F167Y at the immature dimer interface. The suppression of F167Y by secondary changes in two different domains suggests that compensation for a limited representation of the assembly-competent CTD conformation in a population of folding intermediates might be accomplished by either a secondary change in the α-2 helix or by contact between the α-2 helix and the NTD that could stabilize the proper conformation. A role for binding-induced conformational changes (induced fit) in forming the NTD-CTD and CTD-CTD interfaces has been suggested (Ganser-Pornillos et al., 2007; Mateu, 2002).

In summary, the phenotypes of the mutations presented provide support for a model of core maturation in which the MHR and α-2 helix undergo significant rearrangement during maturation, such as that proposed by Ivanov et al (Ivanov et al., 2007). Because of the documented relevance of retrovirus maturation inhibitors as a viable approach for antiviral therapy, understanding the molecular details of maturation is critical. The MHR mutations and suppressors provide a useful tool to be utilized in defining the important transient intermediates of maturation, as well as providing an important means for evaluating the validity of developing models of core assembly.

Materials and Methods

Proviral genomes

Proviral plasmid pRC.V8, which was used for revertant virus isolation, was derived from an RCAN plasmid by replacement of the Schmidt-Ruppin A gag gene with that of Prague C. Identified mutations were then engineered into pRS.V8.eGFP, an analogous splicing-competent plasmid encoding eGFP (Callahan and Wills, 2003). To create the R185W substitution, codons 185 and 186 were changed from CGGGCT to TGGGCA which also created an additional silent mutation destroying the adjacent BanII site. The F193L was produced by changing codon 193 from TTT to TTA, which also created a novel AflII site. The D155Y, E162Q, F167Y, L171V, I190V and A38V mutations have previously been described (Bowzard et al., 2000; Cairns and Craven, 2001). The various single and double mutations were engineered into the pRS.V8.eGFP plasmid by either oligonucleotide-directed mutagenesis and/or by restriction fragment exchange to create various combinations of mutations, and confirmed by sequencing.

Identification of second-site suppressor mutations

Primary turkey embryo fibroblast cultures were transfected with either the E162Q or D155Y mutant proviral genome and serially passaged for 24 days. Genomic DNA was extracted from the cultures that showed a marked increase in extracellular reverse transcriptase over this time-frame and the gag region was PCR-amplified and sequenced to identify the novel mutations.

Infectivity experiments and growth rates

QT6 (transformed quail) cells were transfected with pRS.V8.eGFP genomes to produce suspensions of WT or mutant viruses, which were then normalized by exogenous reverse transcriptase assay and used to infect DF1 (chicken fibroblast) cells. The fraction of cells that were expressing GFP were measured by flow cytometry at various times post-infection (Spidel et al., 2004). Each mutant was analyzed in at least three independent experiments. Two independently derived clones were tested for each of the mutants, except for the E162Q/F193L and the previously published F167Y, I190V and F167Y/I190V viruses, where a single clone was stested several times. Growth rates were computed for each virus as described previously (Wang and Bushman, 2006). Briefly, data from each culture was plotted as ln(% green cells) versus time, and the exponential phase for each growth curve was identified by visual examination of the data. The slope of the line, fit by linear regression to the exponential phase for each curve, was calculated using GraphPad Prism 4 software. This slope was expressed as the growth rate in units of ln(% infected cells)/time. Mean and standard deviation for each mutant was calculated by combining data from several independent experiments and presented in Table 1. For certain non-infectious mutants, the values for percent GFP-expressing cells were usually below 0.1%, which we have set operationally as our background detection limit. This prevented us from accurately fitting a linear-exponential model. The growth rates of such non-infectious mutants were assigned a value of < 0.05.

Testing the temperature-sensitivity of the viruses

To test the effect of temperature on the virus infectivity, WT and mutant viruses were produced from transfected QT6 cells at 37°C and at 42°C. All virus preps produced at 37°C were normalized to each other and used to infect DF1 cells at 37°C. Similarly viruses produced at 42°C were normalized and used to infect DF1 cells at 42°C. Spread of virus in the cultures was detected as mentioned above.

Testing the budding-efficiency of viruses

The budding-efficiency of the viruses was measured as described previously (Cairns and Craven, 2001). Briefly, duplicate plates of QT6 cells growing at 37°C were transfected with WT or mutant pRS.V8.eGFP genomes, and at 24 hrs post-transfection the cells were radiolabelled with [³⁵S]-methionine (50 µCi, 1,000 Ci/mmol). Cells from one set of plates were lysed after 10 minutes to examine Gag expression, while the extracellular viruses were harvested from a second set of cells that were labeled for 2.5 hrs. The viral proteins from the cell lysates and the media samples were immuno-precipitated using a polyclonal anti-RSV serum and subjected to SDS-polyacrylamide gel electrophoresis, followed by autoradiography. Budding efficiency for each virus was measured by calculating the ratio of the CA protein found in media samples to the level of Gag protein in the cell-lysates and expressed as a percentage of the WT.