Nucleic Acid Binding by Mason-Pfizer Monkey Virus CA Promotes Virus Assembly and Genome Packaging

All DNA manipulations were carried out according to standard subcloning techniques, and all plasmids were propagated in Escherichia coli DH5α. All newly created DNA constructs were verified by DNA sequencing. For in vitro assembly assays, the RKK mutations were introduced into pSIT ΔProCANC M-PMV plasmid, which carries the gene for an M-PMV capsid and nucleocapsid fusion protein lacking the N-terminal proline (34) under the T7 promoter. The ²⁰¹RKK²⁰³ to AAA or GPG mutations in pSIT ΔProCANC M-PMV were made by SLIM mutagenesis (35). To introduce mutations in the M-PMV proviral vector pSARM4 (36), we first used a helper vector (MHelppUC19) encoding M-PMV SacI-Eco72I fragment, which was prepared as described previously (37, 38). Mutations RKK/AAA and RKK/GPG were created by two-step PCR mutagenesis using primers carrying appropriate mutations and suitable NotI and XmaI restriction sites, respectively. The obtained PCR products were digested with SacI-NotI/NotI-Eco72I or SacI-XmaI/XmaI-Eco72I, and both fragments were ligated into the MHelppUC19. After sequence verification, the mutated SacI-Eco72I fragments of the MHelppUC19, carrying the appropriate mutation, were inserted into pSARM4 or pSARM-EGFP. For the single-round infectivity assay, the M-PMV Env expression vector pTMO (39) and the pSARM-EGFP vector in which EGFP (enhanced green fluorescent protein) replaces the env gene (40) were used. Further details of the cloning strategy and the full sequences of all PCR primers can be obtained from the authors upon request.

HEK 293T cells were grown in Dulbecco modified Eagle medium (DMEM; Sigma) supplemented with 10% fetal bovine serum (Gibco) and 1% l-glutamine (PAA Laboratories). Transfection of the HEK 293T cells was performed using FuGene HD transfection reagent (Roche Molecular Biochemicals) according to the manufacturer's instructions. At 24 or 48 h posttransfection, virions in the culture media were harvested, filtered through a 0.45-μm-pore-size filter, and centrifuged through a 20% sucrose cushion at 200,000 × g for 1 h in a Beckman SW41Ti rotor. M-PMV proteins were detected by Western blotting with a rabbit anti-M-PMV CA polyclonal antibody (41).

Luria-Bertani medium supplemented with ampicillin (100 μg/ml) was inoculated with a 0.1% (vol/vol) overnight culture of E. coli BL21(DE3) carrying pSIT ΔProCANC M-PMV or an AAA or a GPG mutant plasmid. The cells were cultivated at 37°C and 250 rpm until the culture reached the exponential growth phase (i.e., an optical density at 600 nm of 0.4 to 0.6). Protein expression was induced by the addition of 0.4 mM (final concentration) IPTG (isopropyl-β-d-thiogalactopyranoside). The cells were harvested at 4 h postinduction by centrifugation at 5,000 × g and stored at −20°C.

The purification of the proteins was performed according to previously published protocols (5, 42). Briefly, a high-salt containing buffer extraction was used to solubilize and release the protein from the cell lysate pellet. The extracted protein was subsequently purified by immobilized metal affinity chromatography on a Zn²⁺-charged column, followed by ion-exchange chromatography on a phosphocellulose column, to remove contaminating nucleic acids. Finally, the protein was dialyzed against storage buffer (50 mM phosphate, 500 mM NaCl, 1 μM ZnSO₄, 0.05% mercaptoethanol [pH 7.5]), concentrated using ultrafiltration on Amicon Ultra-15 Ultracel 10K (Millipore, Ireland) concentrators to ∼1 mg/ml, and stored at −20°C. The protein concentration was determined by using a Bradford protein assay.

To determine whether the viral protein assembled inside the bacterial cells during the induced expression process, 1 ml of the cell culture at 4 h postinduction was pelleted, and the cells were resuspended in 350 μl of lysis buffer (50 mM Tris, 100 mM NaCl, 1% [wt/vol] octylthioglucoside, 1 mg/ml lysozyme [pH 8.0]). The suspension was incubated on a rotation mixer for 10 min at room temperature. Using this mild lysis process, the intact virus-like particles (VLPs) were released into the lysate and were subsequently analyzed using transmission electron microscopy (TEM) after negative staining.

The in vitro assembly of purified wild-type and mutant ΔProCANC M-PMV proteins was performed as previously described (5). Briefly, 60 μg of purified protein in storage buffer was mixed with either MS2 phage genomic RNA or λ phage genomic DNA at a 10:1 or 5:1 (wt/wt) ratio in 100 μl of total reaction volume. This mixture was dialyzed against the assembly buffer (50 mM Tris, 100 mM NaCl, 1 μM ZnSO₄ [pH 8.0]) for 2 h at room temperature. When the effect of reducing conditions on the assembly process was studied, the assembly mixture contained 60 mM dithiothreitol (DTT), and the concentration of DTT in the dialysis buffer was 20 mM.

The in vitro particle assembly efficiency was determined by gradient centrifugation, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The dialyzed assembly mixture was loaded on top of 10 to 55% linear OptiPrep (Axis-Shield, Oslo, Norway) gradient and centrifuged for 40 min at 215,000 × g. The gradient was fractionated, and the individual fractions were analyzed by SDS-PAGE (12% gel). The gels were blue silver stained (43) and digitized using Uvitec Alliance 4.7 (Uvitec, United Kingdom), and the assembly efficiency was assessed by densitometric analysis of the lanes using the Fiji (ImageJ) software package (44). The protein content in the topmost fraction of the gradient represented the unassembled protein, whereas the protein content in fractions with an OptiPrep density of around 1.15 to 1.27 g/ml (wt/vol), i.e., fractions 6 to 9, represented the assembled VLPs (5, 6, 45).

For the electrophoretic mobility shift assay (EMSA), 5 or 1.7 μg of studied protein was mixed with 165 ng of 1-kb DNA ladder (Promega, USA) in 10 μl of total volume of buffered environment (25 mM Tris, 250 mM NaCl, 0.5 μM ZnSO₄ [pH 8.0]), corresponding to a protein/nucleic acid ratios of 30:1 (wt/wt) and 10:1 (wt/wt), respectively. The EMSAs were performed under reducing conditions, where the sample reaction mixture contained 60 mM DTT. The samples were incubated 45 min at room temperature. To prove that the nucleic acid shift was caused by the protein nucleic acid interaction, an equivalent reaction mixture was treated with proteinase K (5 μg/reaction) for 45 min at 37°C. All of the samples were analyzed by agarose gel electrophoresis (1% gel) at 8 V/cm. The gels were stained by ethidium bromide and digitized by UVIdoc HD2 (Uvitec).

The HEK 293T cells transfected with the appropriate DNA construct were grown for 48 h posttransfection, starved for 30 min in methionine- and cysteine-deficient DMEM (Sigma), and then pulse-labeled for 30 min with Tran³⁵S-label (M.G.P., Czech Republic) at 125 μCi/ml. The labeled cells were then chased in complete DMEM for 16 h. The cells from pulse and pulse-chase experiments were washed with phosphate-buffered saline (PBS), lysed in 1 ml of lysis buffer (1% Triton X-100, 1% sodium deoxycholate, 0.05 M NaCl, 25 mM Tris [pH 8.0]) on ice for 30 min, and clarified by centrifugation at 14,000 × g for 2 min. The culture medium of the chased cells was filtered through a 0.45-μm-pore-size filter, and SDS was added to a final concentration of 0.1%. Viral proteins were immunoprecipitated from the cell lysates and culture media with a polyclonal rabbit anti-M-PMV CA antibody (1:1,000 dilution) and separated by SDS-PAGE. Radiolabeled proteins were visualized using a Typhoon 9410 phosphorimager (Amersham Biosciences).

To quantify the particle release, the radiolabeled protein bands of ³⁵S-pulse-labeled Gag (Pr78) and pulse-chase-labeled virion-associated CA (p27) were quantified using ImageQuant TL (Amersham Biosciences). The released viral proteins are shown as relative concentrations of CA correlated to the levels of intracellular Gag in individual samples.

Infectivity of M-PMV wt and CA-CTD mutants was determined as described earlier (37, 46). Briefly, HEK 293T cells were cotransfected with either wt or RKK/AAA mutant pSARM-EGFP expression vector, together with the glycoprotein expression vector pTMO. At 48 h posttransfection, the culture supernatants were collected and filtered through a 0.45-μm-pore-size filter, and each sample was normalized for capsid protein content by quantitative Western blotting (37). The volume of culture supernatant used to infect HEK 293T cells was adjusted to 4 ml with complete DMEM, and the cells were incubated for an additional 48 h. The cells were fixed with 4% formaldehyde, and the number of GFP-positive cells was determined using flow cytometry (BD FACSAria).

Reverse transcription-PCR (RT-PCR) was performed as described previously (37). Briefly, the HEK 293T cells were transfected with wt or RKK/AAA mutant proviral construct. At 48 h posttransfection, the virus-containing medium was filtered and centrifuged, and the M-PMV CA content in the viral pellets was normalized by semiquantitative Western blot analysis. Encapsidated RNA was isolated using the QIAamp viral RNA minikit (Qiagen) according to the manufacturer's instructions. The amount of isolated RNA was quantified by measuring the A₂₆₀. To specifically quantitate the genomic RNA, RT using RevertAid H Minus M-MuLV reverse transcriptase (Fermentas) of isolated total RNA (of the CA-normalized amount) was performed. Subsequently, 2 μl of each RT product was used for real-time PCR using a Light Cycler 480 II Real-Time PCR system (Roche) and DyNAmo Hot Start SYBR green mix (Finnzymes). Three pairs of MPMV-CA derived primers—CA1ss (GTGGAATCTGTAGCGGACAA) and CA1as (ATTACCGGCTTGTTGGTTTC), CA2ss (GAAACCAACAAGCCGGTAAT) and CA2as (GAGCAAACAATCCTGGATCA), and CA3ss (TATTGGGCCCTCTTATCAGC) and CA3as GCAACACCCTCCTTTCTCTT—were used for each wt and RKK/AAA sample. Three control samples containing (i) total RNA isolated from wt and RKK/AAA mutant, (ii) cDNA prepared from mock-infected cells, and (iii) water were used for real-time quantitative PCR. The viral genomic RNA content in individual samples was determined in three independent experiments.

HEK 293T cells transiently expressing the wild-type or RKK/AAA mutant M-PMV proviral constructs grown in 100-mm plastic dishes were incubated with leptomycin B (LMB; final concentration, 20 nM) at 24 h posttransfection for 30, 60, and 120 min. The cells were then washed with PBS, and nuclear and cytosolic fractions were isolated using a Nuclei EZ Prep Nuclei Isolation kit (Sigma-Aldrich) according to the manufacturer's protocol. Individual samples of total cell lysate, and nuclear and cytosolic fractions were then analyzed by Western blotting with the following antibodies: rabbit anti-M-PMV CA (the present study), rabbit anti-lamin A (L1293; Sigma-Aldrich), rabbit anti-cyclophilin A (sc-20360; Santa Cruz), and monoclonal anti-β-actin (clone AC-74; Sigma-Aldrich).

HEK 293T cells transiently expressing the wild-type or mutant M-PMV proviral constructs grown in 100-mm plastic dishes were washed with PBS, scraped into a microtube, and prefixed with freshly prepared 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.5). After a washing step with 0.1 M cacodylate buffer (pH 7.5), the cells were postfixed in 1% osmium tetroxide, dehydrated in an ethanol series (30, 50, 70, 80, 90, and 100%), and embedded in fresh Agar 100 epoxy resin. Ultrathin sections (70 nm) of cells were cut with a diamond knife on a Leica UC6 ultramicrotome (Leica Microsystems, Wetzlar, Germany). The thin sections were collected on Parlodion-coated microscopy grids and contrasted using saturated uranyl acetate and lead citrate. For each sample, we analyzed approximately 30 infected cells.

In vitro-assembled VLPs were deposited on carbon-coated copper grids for 3 to 6 min. The grid was washed twice on a drop of deionized water for 20 s and stained with sodium silicotungstic acid (4%, pH 7.4) for 30 s. The excess stain was wicked off using a filtration paper, and the samples were dried in air and analyzed using a JEM-1010 transmission electron microscope (Jeol, Japan) operated at 80 kV and equipped with a Megaview III CCD (charge-coupled device) camera. The images were processed using the AnalySIS software suit (Olympus, Japan).

In vitro-assembled mutant ΔProCANC M-PMV RKK/AAA was diluted in PBS containing 10-nm colloidal gold and transferred to glow-discharged C-Flat 2/2 Holey carbon grids in a high-humidity chamber. Cryogrid preparation was performed using a manual plunging device (EMBL, Heidelberg, Germany). Grids were blotted from the back, frozen in liquid ethane, and then stored under liquid nitrogen conditions until imaging.

Data acquisition and image processing were performed as previously described (32). In brief, tilt series were imaged on a FEI Titan Krios electron microscope operated at 200 keV, with a GIF2002 post-column energy filter (using a slit width of 20 eV) and a 2k×2k Gatan Multiscan 795 CCD camera. Low-magnification montages for search and navigation were acquired using SerialEM (47). Tilt series were then acquired at appropriate positions using FEI tomography software version 4 in automated batch mode. The nominal magnification was 33,000, giving a calibrated pixel size of 2.87 Å. The tilt range was from −45° to +60° in 3° steps, collecting first from 0° to −45° and then from 3° to 60°, with a total dose of ∼40 e Å⁻² being applied to each tilt series. Defoci ranged between −1.5 and −3.5 μm. Tomograms were reconstructed using the IMOD software suite (48).

Subtomogram averaging was performed as described previously (32) using MATLAB scripts derived from the TOM (49) and AV3 (50) packages. The Dynamo software package was used for generation of masks and for FSC (Fourier shell correlation) calculations (51). Initial alignment was performed on 3× binned data. Processing was carried out entirely independent for two half data sets. Each half data set contained roughly the same number of tubes and an equal distribution of defoci. To obtain an initial structure, one tomogram with a defocus of −3.5 μm was chosen for each half-data set. Extracted subtomograms from this tomogram were assigned initial angles based only upon the geometry of the tubes and were averaged to generate a smooth starting reference. The subtomograms from this tomogram were then iteratively aligned and averaged in six dimensions against the reference as described previously (50). After the structure stabilized it was used as starting reference for its respective half-data set. Subsequently, all subtomograms within each half-data set were aligned and averaged against their respective independent starting reference for six iterations. After the first two iterations, a cross-correlation-based cleaning was performed to remove subtomograms that contained no density corresponding to the M-PMV ΔProCANC protein layer. No symmetry was applied in the alignments performed with binned, non-CTF-corrected data.

The defocus of each tomogram was measured by fitting theoretical contrast transfer function (CTF) curves to averaged power spectra from 512 square pixel tiles generated from all images in a tilt series using MATLAB scripts. CTF correction was performed using the program “CTF phase flip” implemented in IMOD (52).

Subvolumes with a size of 310 ų were extracted from unbinned, CTF-corrected tomograms at the positions determined in the 3× binned alignments. The subtomograms were subjected to two further alignment iterations in which an additional 2-fold symmetry was applied. The two final references were aligned, averaged, and multiplied with a Gaussian filtered mask. Subsequent comparison of the two final references (averaged from 52,566 and 53,038 asymmetric units in each of the half-data sets, respectively) by FSC indicated a resolution of 10.9 Å. The final structure was sharpened applying a negative B-factor of −1,500 Ų, while filtering to the resolution determined at the 0.143 FSC threshold. The wt M-PMV ΔProCANC tube structure used for comparison was EMD-2089 (31). Visualization of tomograms or electron microscopy density maps was performed using either IMOD (48), UCSF chimera (53), or Amira4 (FEI Visualization Sciences Group) with the electron microscopy toolbox (54).

We first used a bacterial system to study the role of the RKK basic region in the assembly of immature particles. The RKK/AAA and RKK/GPG mutations were introduced into truncated M-PMV Gag (ΔProCANC), and its ability to assemble in bacterial cells was assessed using TEM. In contrast to wt ΔProCANC protein that forms tubular and spherical VLPs in E. coli cells, none of the RKK mutant proteins formed any particles, indicating that the RKK sequence is important for the assembly of VLPs in bacterial cells.

We next purified the expressed proteins and studied their ability to assemble into VLPs in vitro using TEM. Our previous results showed that the wt ΔProCANC of M-PMV could form in vitro either spherical (5) or tubular (31) particles, depending on the assembly conditions. The RKK/GPG ΔProCANC mutant protein did not assemble into any particles in vitro under any assembly conditions tried (see Materials and Methods). In the presence of λ phage DNA at reducing conditions, the RKK/AAA mutant of ΔProCANC protein formed tubular VLPs that were similar to those formed by the wt protein (Fig. 1) but were typically longer (up to several microns) than those of the wt protein. This mutant assembled also in the presence of MS2 phage RNA, although at both reducing and nonreducing conditions it formed only a few regular spherical particles (as seen in Fig. 2C), alongside fragments of particles and aggregated protein. Particle assembly was dependent on the presence of nucleic acid, an observation consistent with our previous finding that the efficiency of wt ΔProCANC VLP assembly is significantly facilitated by the presence of nucleic acids (5).

To assess the efficiency of in vitro VLP assembly, we performed OptiPrep gradient ultracentrifugation of particles assembled under reducing conditions from both wt and mutant ΔProCANC proteins in the presence of either MS2 RNA or λ phage DNA, followed by SDS-PAGE analysis of individual fractions and by densitometric analysis of the bands (Fig. 2). All VLPs assembly reactions, together with ultracentrifugations and protein quantitation analyses, were performed in triplicate from independent protein isolations. Independent of the type of added nucleic acid, the in vitro assembly of wt ΔProCANC protein proceeded efficiently. The proportion of the assembled protein ranged between ca. 60% and ca. 50% in the presence MS2 RNA and λ DNA, respectively (Fig. 2A). For the RKK/GPG mutant, for which no particles were observed by TEM, the assembly efficiency was negligible, and no visible bands corresponding to assembled particles were observed in the SDS-PAGE gel. The yield of the RKK/AAA mutant particles was significantly lower than that of wild type under both conditions (i.e., with MS2 RNA or λ DNA) and, unlike the wild-type particles, it was more dependent on the amount of nucleic acid added. Increasing the amount of MS2 RNA in the assembly mixture from 6 to 12 μg enhanced the VLP assembly efficiency of RKK/AAA mutant from 6.4% to 21% while the wt protein assembled most efficiently when 6 μg of MS2 RNA was added (Fig. 2A). On the other hand, increased concentration of the λ DNA slightly reduced the assembly efficiency of the RKK/AAA mutant from ca. 21% ± 1% to ca. 16% ± 2% (Fig. 2A), which presumably reflects the redistribution of particles in the gradient as a result of aggregation. We inspected the gradient fractions using TEM to confirm the presence of assembled particles in the relevant fractions (Fig. 2B to E). As expected, spherical or tubular assembled particles were observed in fractions 6 to 9.

We next assessed whether the RKK/AAA mutation affects the structure of the assembled tubular particles. ΔProCANC M-PMV RKK/AAA tubes assembled in vitro under reducing conditions in the presence of λ DNA were subjected to cryoelectron tomography. Consistent with the negative-stain electron microscopy results, long tubular arrays were observed (Fig. 3A). Along the surface of the tubes hexagonal patches were visible, similar to those observed in cryoelectron tomograms of wt ΔProCANC tubes (32). We performed subtomogram averaging and obtained an ∼11-Å resolution structure (Fig. 3B). We compared the mutant structure with the available wt ΔProCANC tube structure (31). No differences were seen in the capsid region (Fig. 3B), indicating that the RKK/AAA mutation does not influence the tertiary or quaternary structure of the capsid domains. Differences were observed in the region underlying the C-terminal part of capsid (Fig. 3B and C), where the nucleocapsid and nucleic acids are located. Here, the RKK/AAA mutant lacks a filamentous density that was present in the wt tube (Fig. 3B, yellow). The dimensions of this density are approximately that of a nucleic acid double helix. We previously suggested that this density was nucleic acid recruited by the RKK motif in M-PMV that is not present in the HIV CA sequence nor in the cryoelectron microscopy structure (30, 31). The absence of this structure in the RKK/AAA mutant strongly supports this hypothesis.

The observations presented so far suggest a role for the RKK basic region in binding of nucleic acids. To investigate whether the RKK/GPG and RKK/AAA mutations alter nucleic acid binding, we mixed the studied proteins with a 1-kb DNA ladder in two different ratios (protein/DNA ratios of 30:1 [wt/wt] and 10:1 [wt/wt]) and performed an EMSA. The ratios of protein to nucleic acid were based on the optimization of reaction conditions, where the DNA was bound efficiently to wt and mutant protein. None of the tested proteins (wt, RKK/AAA, or RKK/GPG) showed any traces of contaminating nucleic acids (Fig. 4, lanes PC), and no shifts were observed when the equivalent reaction mixtures were also treated with proteinase K (Fig. 4, lane Pr), indicating that any observed shifts are caused exclusively by the interaction of the protein with added nucleic acid. At a protein/DNA ratio of 30:1 the wt protein quantitatively bound DNA content from the reaction mixture (Fig. 4, left panel, wt lane P), whereas at a protein/DNA ratio of 10:1 some DNA was not bound by the protein and remained free in the solution (Fig. 4, right panel, wt, lane P). The RKK/AAA and RKK/GPG mutants showed a much lower ability to bind DNA at both ratios (Fig. 4, AAA and GPG, lanes P), the reduction was particularly severe for the RKK/GPG mutant. The higher-molecular-weight fragments from the 1-kb ladder showed greater retention than those of the lower-molecular-weight fragments (compare Fig. 4, lanes P). The EMSA indicates that mutations in the RKK region indeed lead to reduced DNA binding that correlates with a reduced efficiency of particle assembly.

To investigate whether the effect of mutations in the RKK region on nucleic acid binding and particle assembly observed in vitro are reflected in changes in the viral life cycle, HEK 293T cells were transfected with proviral wt pSARM4 plasmid and the corresponding RKK/AAA and RKK/GPG mutants. Transfection with all of the constructs led to the expression of similar amounts of Gag, Gag-Pro, and Gag-Pro-Pol polyproteins, indicating that the level of expression was not affected by the mutations (Fig. 5A). At 48 h posttransfection, the expression and processing of the wt and mutant M-PMVs were assessed by pulse-chase assay and Western blot analysis of transfected cells.

We measured the amount of protein in the supernatant (after a 16-h chase) relative to the amount of cell-associated proteins (before chase) in order to estimate the amount of virus released (Fig. 5C). The efficiency of release for the RKK/AAA mutant was only ca. 45% that of the wt. Both the wt and the RKK/AAA virions contained processed CA protein, indicating that proteolytic processing of capsid was not affected by the RKK/AAA mutation (Fig. 5B). In the RKK/GPG mutant the ratio of CA protein detected in the medium was <20% that of the wt.

To analyze the assembly pathway of the virions, we performed TEM analysis of thin sections of transfected HEK 293T cells (Fig. 6). We inspected about 30 infected cells of each wt or mutant viruses and semiquantitatively assessed the virus particle morphogenetic types. Transfection with wt M-PMV gives rise to large numbers of intracellularly assembled D-type particles in the pericentriolar region. We counted approximately 100 intracellular particles which were exclusively D-type. In the case of the RKK/AAA mutant, we observed not only D-type pericentriolar assembly but also a considerable amount of particles assembling at the plasma membrane (C-type assembly) (Fig. 6B and C, respectively). In 30 cells we counted 58 intracellular D-type particles and 29 partially assembled C-type particles budding from the plasma membrane. The intracellular RKK/AAA M-PMV particles had a spherical shape similar to that of the wt M-PMV (compare Fig. 6A and B). Released virions were also observed. The GPG mutant failed to form immature particles inside the cells; however, a large amount of electron-dense material (presumably Gag) accumulated underneath the cell plasma membrane (Fig. 6D) (55). This compact layer on the membrane did not form any released virus particles or budding structures. Based on this analysis (see further discussion below), we concluded that the CA protein detected in the medium of the RKK/GPG mutant most likely represents free protein and not virus particles.

The infectivity of the released particles was tested by single-round infectivity assay (Fig. 7A). The infectivity of the RKK mutants was below the detection limit (<1%) compared to that of the wt M-PMV, indicating that the mutation affected some key step of the infectivity process.

The above-mentioned results of particle release and TEM analysis (Fig. 5C and 6, respectively) indicate that mutation of the RKK motif leads to a reduction in particle release efficiency and to a partial relocalization of the assembly site from the pericentriolar region to the plasma membrane. The mutant particles are released from the cells but show severely abolished infectivity.

To assess whether mutation of the RKK motif influences nucleic acid incorporation in infected cells, we determined the total RNA and the genomic RNA content of the released particles. We found that the total amounts of RNA (measured spectrophotometrically at A₂₆₀) incorporated into the mutant and wt virions were comparable (wt [100% ± 5%] and RKK/AAA mutant [118% ± 18%]). To quantify genomic RNA incorporated into released M-PMV particles, the isolated total RNA from the wt and RKK/AAA mutant was reverse transcribed and analyzed by real-time qPCR using three various CA-specific primer pairs. No positive signals (i.e., threshold cycle [C_T] > 38) were determined in the control samples in which cDNA isolated from mock-infected HEK 293T cells, isolated total RNA, or water were used as a template for real-time quantitative PCR. Surprisingly, significant differences between the RKK/AAA mutant and the wt particles were observed by specific quantification (quantitative RT-PCR) of the genomic RNA (Fig. 7B). The RKK/AAA mutant particles contained only about 9% (±1.5%) of genomic RNA in comparison to the wt. The data represent three independent transfections, where the wt values were arbitrarily assigned as 100%.

It was previously observed that mutation of basic amino acid patches in the pp24 domain of M-PMV Gag affects its localization to nuclear pores (56). We therefore next assessed whether the genome packaging defect observed in the RKK/AAA mutant might also relate to altered nuclear trafficking of Gag. To delay fast export of proteins from the nucleus, we treated the cells at 24 h posttransfection with LMB. This inhibitor of CRM1-dependent nuclear export was selected because it was suggested to affect M-PMV Gag exit from the nucleus (22). Analysis of LMB-treated cells thus should reveal even transient presence of Gag inside the nuclei. Nuclear and cytosolic fractions of cells transfected with wt or RKK/AAA mutant M-PMV vector were analyzed for the presence of Gag (Fig. 8).

The level of wt M-PMV Gag in the nuclear fractions gradually increased with increasing length of LMB treatment over a 120-min time frame. For the RKK/AAA mutant, the initial amount of Gag in the nucleus was notably higher than for the wt but did not increase during longer incubations. These results are consistent with a reduced rate of nuclear export for the RKK/AAA mutant of Gag. A reduced export rate would lead to increased accumulation of RKK/AAA Gag during the 24-h period of posttransfection before LMB treatment. Upon reaching saturation, no increase in nuclear RKK/AAA was observed when export was inhibited by LMB.

Another remarkable observation is the difference in the presence of the wt and RKK/AAA mutant CA proteins inside the cell nucleus (Fig. 8). Although nuclear localization is not surprising for the wt CA protein since it was also reported for the HIV-1 CA protein in primary human macrophages and HeLa cells at the early stages of infection (57, 58), the AAA mutation of the RKK motif in the M-PMV CA protein almost completely abolished the import of the CA protein into the nucleus (Fig. 8). The detectable amounts of M-PMV CA in the nucleus may originate from CA present in a preintegration complex in cells newly infected by the released particles. The absence of CA in the nuclei of cells transfected with the RKK mutant is consistent with the fact that the RKK/AAA mutant virus lacking the vRNA is noninfectious and thus unable to introduce mature CA into the cells.

The mutation of the RKK region to either AAA or GPG had wide-ranging effects on Gag assembly, nucleic acid incorporation, and virus production. In the case of mutation to GPG, assembly, nucleic acid incorporation and virus production were essentially abolished. We speculate that the insertion of a proline residue in this region may cause structural defects in CA which interfere with its proper function.

Mutation of RKK to AAA was still permissive for in vitro assembly. This allowed us to test our previous hypothesis—that the extended filament of density underlying CA in in vitro-assembled ΔProCANC tubes represents an ordered nucleic acid structure bound to the basic RKK region (30). Indeed, structural analysis showed that this filament is lost in the RKK/AAA mutant, indicating that the nucleic acid is either absent in the particles or is no longer bound to the RKK motif. Consistent with this observation, the efficiency of nucleic acid binding by the mutant protein is reduced. We observed that the mutation also leads to significantly reduced efficiency of Gag assembly, suggesting that the interactions between RKK and nucleic acid may facilitate the assembly. This could be achieved either by promoting CA dimerization or further oligomerization into a hexameric lattice. Since the wt ΔProCANC assembles into particles of similar morphology both in E. coli and in vitro, the failure of the ΔProCANC RKK/AAA mutant to assemble in E. coli suggests that RKK-mediated CA oligomerization may be particularly important in crowded cellular environments. Together, these observations are consistent with a model in which nucleic acid binding by the RKK motif creates a scaffold promoting Gag assembly.

When placed in the context of the infectious virus, the RKK/AAA mutation leads to a ca. 50% drop in virus particle release, which is consistent with the assembly defect observed in vitro. Strikingly, the mutation also influences the site of assembly: whereas wild-type M-PMV assembles D-type particles in the pericentriolar region (1, 59), in the AAA mutant, approximately one-third of the observed particles appeared as C-type particles assembling at the plasma membrane. In the GPG mutant, dense layers of accumulated protein are seen underlying the plasma membrane, which are not seen in noninfected cells. Based on the thickness of these layers, and because similar structures have previously been shown to be formed by assembly-incompetent Gag polyproteins (55, 60, 61), we presume that these extra layers observed under the plasma membrane are formed by accumulated mutant Gag protein.

Trafficking of M-PMV Gag to the pericentriolar assembly site is dependent on the interaction between a CTRS signal in MA with the Tctex-1 component of dynein (4). We consider it unlikely that mutation of RKK influences this interaction because the MA site is distal to CA. We prefer the following alternative explanation for the relocation of assembly. The assembly of C-type retroviruses at the plasma membrane is promoted by CA-CA interactions, NC-RNA-NC interactions, and MA-membrane-MA interactions. In the case of M-PMV, a D-type retrovirus, assembly in the cytoplasm is promoted by CA-CA, NC-RNA-NC, and CA(RKK)-nucleic acid-CA(RKK) interactions. In the case where the CA(RKK)-nucleic acid-CA(RKK) interaction is abolished by mutation of the RKK, the cytosolic assembly efficiency is reduced, and part of Gag is transported to the PM, where MA-membrane-MA interactions promote C-type assembly.

Surprisingly, while the RKK/AAA mutation does not affect the total amount of RNA being incorporated into virus particles, it dramatically reduces the packaging of genomic RNA. This is most likely the cause of the observed reduced infectivity. Specific packaging of genomic RNA in retroviruses is typically mediated by NC (62–64). While some studies have indicated that interactions between MA and nucleic acid may modulate this effect (18, 19), a role for CA in genome packaging has not been described. We cannot rule out that the RKK motif influences packaging by direct interaction with sequences in the genomic RNA, but this seems unlikely. A number of other hypotheses seem more likely. The RKK motif may help to stabilize the Gag-vRNA complex, or it may contribute to a structural arrangement of the NC region which is required for packaging. The reduction in genome packaging may be a secondary effect resulting from the relocalization of viral assembly from the pericentriolar region to the plasma membrane. Alternatively, packaging may be hindered by a change in the nuclear transport of Gag.

Nuclear trafficking of Gag has previously been shown to be important for genomic RNA incorporation for RSV (24, 25). In the case of M-PMV, Gag associates with nuclear pores (56) and may enter the nucleus (22). We found both RKK/AAA and wild-type Gag in the nucleus. The phenotype we observed for the RKK/AAA Gag is consistent with reduced efficiency of nuclear export. Because the RSV nuclear export signal consists mainly of nonpolar amino acids in the p10 domain (65), we expect that the export of Gag from the nucleus is not directly inhibited by the mutation of the polar RKK region, but it is rather mediated by the Gag-vRNA interactions and their simultaneous export from the nucleus. This could lead to a reduction in genomic RNA export and packaging into virus particles. Alternatively, if nuclear export is promoted by a Gag-vRNA complex, then reduced export might result indirectly from the RKK/AAA mutation interfering with genomic RNA binding. In either case, the reduction in genomic RNA incorporation and the altered nuclear trafficking would be related phenotypes.

In summary, our data suggest that the RKK region of CA modulates interactions between Gag and the viral genomic RNA in a manner that is also accompanied with changes in nuclear trafficking. The nucleic acid recruited by RKK serves a scaffolding function that promotes Gag assembly. We speculate that D-type retroviruses such as M-PMV may require an additional assembly scaffold to replace the function of the plasma membrane in the assembly of C-type retroviruses.