The RNA Binding G-patch Domain in Retroviral Protease Is Important for Infectivity and D-type Morphogenesis of Mason-Pfizer Monkey Virus*

Retroviral proteases (PRs) cleave the viral polyprotein precursors into functional mature proteins late during particle release and are essential for viral replication. Unlike most retroviruses, β-retroviruses, including Mason-Pfizer monkey virus (M-PMV), assemble immature capsids within the cytoplasm of the cell. The activation of β-retroviral proteases must be highly regulated, because processing of the Gag-related polyprotein precursors occurs only after transport of immature capsids to the plasma membrane and budding. Several β-retroviral proteases have unique C-terminal extension sequences, containing a glycine-rich motif (G-patch), which specifically binds in vitro to single-stranded nucleic acids. In M-PMV PR the G-patch is removed in vitro as well as in vivo by autoproteolytic processing to yield truncated active forms of PR. To investigate the role of the G-patch domain on the virus life cycle, we introduced mutations within the C-terminal domain of protease. We found that the G-patch domain of M-PMV PR is not required for the processing of viral polyproteins, but it significantly influences the infectivity of M-PMV, the activity of reverse transcriptase, and assembly of immature capsid within the cells. These results demonstrate for the first time that the G-patch domain of M-PMV PR is critical for the life cycle of β-retroviruses, and its evolutionary conservation within members of this genus suggests its importance for retroviruses that display D-type morphology.

The maturation process of retroviruses is mediated by a virus-encoded protease (PR) 3 that functions to cleave the viral polyproteins into the mature structural proteins and enzymes found in infectious particles. The retroviral protease is translated as a domain of the Gag-related polyproteins. For most retroviruses, Gag and Gag-Pol polyproteins assemble into an immature particle at the plasma membrane during viral budding, in a process referred to as "C-type" viral assembly. Gagpolyprotein preassembly in the cytoplasm of the host cell, away from the plasma membrane, is termed "B/D-type" assembly and is unique to the ␤-retroviruses (B-and D-type retroviruses), and the spumaviruses.
Mason-Pfizer monkey virus (M-PMV) is the prototype of the type-D retroviruses. The genomic organization of M-PMV consists of four genes in the order 5Јgag-pro-pol-env-3Ј. The precursors to the viral enzymes, Gag-Pro and Gag-Pro-Pol, are produced by ribosomal frame shifts at the ends of the Gag and protease reading frames, respectively (1). Gag-polyproteins and genomic RNA assemble into immature viral capsids in the pericentriolar region of the infected cell (2). The dominant signal responsible for this targeting was identified within the N terminus of the Gag polyprotein in the matrix domain (MA) and has been termed the cytoplasmic targeting/retention signal (CTRS) (3,4). This motif of 18-amino-acid spanning residues 43-60 within MA sequence mediates intracytoplasmic targeting by interaction with the dynein/dynactin molecular motor complex (2). Mutation of a single amino acid (R55W) within M-PMV CTRS converts the D-type morphogenesis into a C-type assembly pathway (3), confirming that there are no inherent differences in the process of capsid assembly for different retroviruses. A bipartite membrane-targeting signal, consisting of the myristilated N-terminal 14 amino acids and a highly basic region of 17 residues of the matrix domain of human immunodeficiency virus, type-1 was shown to be responsible for assembly of viral polyproteins at the plasma membrane (5). Thus, structural signals located within Gag polyproteins direct polyprotein transport for capsid assembly.
The immature particle proceeds to a mature morphology during budding and virion release when the protease is activated. The first proteolytically active M-PMV PR excised from Gag-precursors is a 17-kDa protein (per monomer) (17PR), which undergoes further C-terminal self-processing yielding 13-kDa protease (13PR). These two forms of protease were detected in released virions. In vitro, autoprocessing of 17PR yields another truncated 12-kDa form of M-PMV PR (12PR) (6). All three forms display similar substrate specificities (7). The catalytic efficiencies of 13PR and 17PR are comparable; however, further truncation of 13PR into 12PR decreases the catalytic efficiencies more than 10-fold (8,9).
All retroviral proteases belong to the family of aspartic proteases. The active retroviral protease is a homodimer with a single active site. A key structural feature of PR homodimer is a four-stranded antiparallel ␤ sheet that results from an interaction of N-terminal sequences of one subunit and C-terminal sequences of the second subunit (for review see Ref. 10). NMR structural study of the 12-kDa form of M-PMV protease revealed that its overall structure follows the conservative structural motifs of retroviral proteases despite the lack of C-terminal sequences. 12PR can exist in solution in the form of a fully folded monomer, which dimerizes and becomes active in the presence of substrate or inhibitor. The C-terminal sequence of M-PMV 17PR likely forms a separate structural motif independent of the catalytic core of the molecule (8).
A bioinformatic analysis of proteins containing RNA binding sequences revealed a conserved glycine-rich module, termed the G-patch, in M-PMV Gag-polyproteins within the C-terminal sequence of protease (11). Recently, the C-terminal domains in M-PMV protease and the protease of mouse intracisternal A-type particles endogenous retrovirus have been demonstrated to specifically bind in vitro to singlestranded DNA and RNA oligonucleotides (9). A similar domain was identified in over 100 eukaryotic proteins many of which are involved in RNA processing (11).
Here, we report an analysis of the influence of the C-terminal domain of the M-PMV protease on the life cycle of the virus. We show that although this C-terminal G-patch-containing sequence is not necessary for virus maturation, it is important for its infectivity and an activity of the reverse transcriptase. Surprisingly, this domain affects the morphogenesis of viral particles.
Bacterial Expression and Purification of M-PMV Protease-Expression of PR and its mutants in Escherichia coli BL21 (DE3) from pBP phagemids were carried out in LB (Luria-Bertani) medium containing ampicillin (at a final concentration of 100 g/ml). When the cells reached A 600 ϳ 0.9 -1.1 the expression was induced by adding isopropyl 1-thio-␤-D-galactopyranoside to a final concentration of 1 mM. The cells were harvested 2 h postinduction by centrifugation. The PR and all of its mutants were isolated from the inclusion bodies by solubilization in 50 mM Tris-HCl buffer, pH 7.5, with 8 M urea, followed by dialysis against 50 mM Tris-HCl buffer, pH 7.5, with 0.1% mercaptoethanol and by batch chromatography on QAE-Sephadex A25 under the same conditions (6).
Autocatalytic Cleavage of Protease in Vitro-The protease at a concentration of 0.2 mg/ml was dialyzed into 50 mM acetate buffer, pH 5.3, containing 0.3 M NaCl, 0.1% mercaptoethanol and then incubated at 37°C for different time (0 -10 days). The aliquots were then analyzed by SDS electrophoresis in 15% polyacrylamide gel.
Cell Cultures and Transfection-COS-1 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. Transfection of the DNA constructs into COS-1 cells was performed by using the FuGENE 6 reagent according to the instruction manual.
Radiolabeling and Immunoprecipitation-At 20 h posttransfection, cells were starved for 30 min in Cys/Met-free Dulbecco's modified Eagle's medium and then labeled for 30 min in Cys/Met-free Dulbecco's modified Eagle's medium containing 35 S-labeled methionine-cysteine (100 Ci/ml) (ICN). Pulse-labeled cells were lysed in lysis buffer A (0.025 M Tris-HCl, 1% Triton X-100, 1% sodium deoxycholate, 0.05 M NaCl, pH 8). For the chase, the cells were incubated in complete Dulbecco's modified Eagle's medium for an additional 4 h and then lysed in lysis buffer A. Nuclei were then pelleted from both the pulse and chase cell lysates at 13,000 rpm in a microcentrifuge. After addition of SDS to the final concentration of 0.1% the viral proteins were immunoprecipitated with goat anti-M-PMV antiserum or with rabbit anti-CA antiserum. The chase-plate culture medium containing radiolabeled viral particles was filtered through a 0.45-m pore size filter. It was then loaded onto a 20% (w/v) sucrose cushion in phosphate-buffered saline and centrifuged for 60 min at 55,000 rpm in a 90Ti rotor (Beckman). The pelleted virus particles were then lysed in lysis buffer B (0.025 M Tris-HCl, 1% Triton X-100, 1% sodium deoxycholate, 0.05 M NaCl, 0.1% SDS, pH 8), and the proteins were immunoprecipitated with anti-M-PMV or anti-CA antisera. Radiolabeled proteins were separated on SDS electrophoresis in 15% polyacrylamide gel and detected by autoradiography.
Infectivity Assay-COS-1 cells in 100-mm diameter plates were transfected with individual DNA constructs. At 48 h posttransfection the medium was collected, filtered and the relative levels of CA protein were determined for each sample by Western blot analysis with chemiluminescent detection using a CCD camera. The levels of CA protein concentration in the samples were normalized by dilution with complete medium, and the normalized supernatants were used to infect COS-1 cells on a 60-mm diameter plate in the presence of Polybrene (10 g/ml). After overnight incubation the medium was replaced with fresh Dulbecco's modified Eagle's medium, and the cells were incubated for another 48 h. The medium was collected, and the viral particles were pelleted through a sucrose cushion as described above and were resuspended in 50 l of PLB (25 mM Tris-HCl, 0.5% 2-mercaptoethanol, 1% SDS, 0.05% bromphenol blue, 5% glycerol, pH 6.8). The level of CA protein released into the medium was determined by Western blot analysis using a polyclonal rabbit anti-CA antiserum and detected on a CCD camera using chemiluminescent substrate West Femto (Pierce).
Activity of the Reverse Transcriptase-COS-1 cell cultures cultivated in 100-mm diameter plates were transfected with individual DNA constructs. 48 h posttransfection the medium was collected and filtered, and the pelleted viral particles were lysed in lysis buffer C (50 mM Tris-HCl, pH 7.8, 100 mM KCl, 0.05% Triton X-100, 2 mM dithiothreitol). The relative levels of CA protein were determined for each sample by Western blot analysis using West Femto (Pierce) chemiluminescent substrate. The reaction mixture for the RT assay contained 40 l of reaction buffer (50 mM Tris-HCl, pH 7. Electron Microscopy-At 20 h posttransfection, COS-1 cells were washed with phosphate-buffered saline and fixed with 2.5% (v/v) glutaraldehyde in cacodylate buffer (0.1 M, pH 7.4) for 1 h at 4°C. Fixed cells were scraped, washed with cacodylate buffer, postfixed with 1% osmium tetroxide, dehydrated through an increasing ethanol series (including a 30-min incubation in 1.5% uranyl acetate in 70% ethanol), and embedded in Agar 100 resin (Gröpl, Tulln, Austria). Sections (70 nm) were contrasted with a saturated uranyl acetate solution in water (10 min, room temperature) and Reynold's lead citrate (7 min, room temperature). The electron microscopy images were obtained with a JEOL JEM-1010 electron microscope operating at 80 kV.

Mutations within the C-terminal Domain of M-PMV Protease Do Not Affect Specificity of Protease and Rate of Virion-associated Proteolysis-
Previous studies showed that proteolytic processing of the 17-kDa form of M-PMV PR into the truncated form both in vitro and in vivo is a time-dependent process. The freshly released mature particles contain only 17PR and the truncated 13PR appears in virions within 2 h after their release (6). In vitro experiments confirmed that the first autocatalytic processing of 17PR occurs at the C terminus in position Ala 114 -Gln 115 and it is followed by cleavage in position Ser 107 -Pro 108 yielding the 13 and 12 kDa forms of M-PMV PR, respectively (see Fig. 2). To determine an influence of the G-patch and the rate of its processing on the virus life cycle, we introduced several mutations into the C-terminal domain of PR (see Fig. 2). Asn 109 , which forms the P2Ј position of the 12PR/13PR processing site, was mutated into Ile, Asp, Lys, Ser, Glu, Leu, Thr, Met, and Val. Furthermore, Ala 114 in the P1 position of the 13PR/ 17PR processing site was replaced by Arg. Gln 115 from the PЈ1 site of this cleavage site was mutated into Lys, Arg, Glu, and Ile. These protease mutants were expressed in bacteria and purified, and their autocatalytic cleavage was monitored by SDS-PAGE at different time intervals (Fig.  3). The purified sample of the wt M-PMV PR (17PR) contained 17PR and 13PR and traces of 12PR. Three of the analyzed mutations significantly changed the rate of the PR processing. Arg introduced into the P1 position of the 13/17 cleavage site (mutant A114R) slowed down the autoprocessing and prolonged the existence of 17PR in the sample. The replacement of Gln 115 by Ile in the P1Ј position of this cleavage site (mutant Q115I) significantly accelerated the C-terminal processing of PR. The 17-kDa form was processed into truncated PRs during the renaturation, and the purified sample contained mostly the 12-kDa form of PR. After two additional days of incubation at 37°C the sample contained only the shortest form of PR. In contrast, Ile introduced into the P2Ј position of the 12/13 cleavage site decreased the autoprocessing of 13PR into 12PR (data not shown). To obtain a mutant of PR that yielded a more stable 13PR, the double mutations (N109I and Q115I) were cloned into the protease processing sites. Mutations A114R, Q115I, and N109I/Q115I influencing the C-terminal autoprocessing of M-PMV PR were then cloned into the infectious proviral clone pSARM4. In addition, proviral clones containing truncated PR mutants (PR⌬118-145) and (PR⌬111-145) were constructed. Finally, a mutation Y121S in the PR, which was shown to significantly decrease binding of nucleic acids to the G-patch in vitro (9), was introduced in pSARM4.
The expression of M-PMV-specific proteins was analyzed 20 h after transfection of the respective plasmids into COS-1 cells. Both cell-associated and released virion-associated proteins were analyzed by immunoprecipitation with rabbit anti-M-PMV antiserum (Fig. 4). Similar levels of Gag, Gag-Pro, and Gag-Pro-Pol precursor proteins were synthesized in cells transfected with truncated mutants as well as with wild-type genomes (Fig. 4a), confirming that deletion of sequences encoding a substantial part of the C-terminal domain of protease did not disturb the ribosomal frameshifts. Following virus release, Gag precursors were processed by wt PR as well as by the protease mutants (Fig. 4b).
The truncation of the most C-terminal domain (mutants PR⌬118-145 and PR⌬111-145) did not affect the proteolytic processing of protein precursors in released viral particles. Similar processing of Gag polyproteins and a comparable amount of the capsid protein in the released virions was also detected for the G-patch mutant, Y121S. Viral particles of constructs Q115I and N109I/Q115I contained, besides the mature proteins, a low level of unprocessed Gag polyprotein even after a 24 h incubation of virions at 37°C (Fig. 4b). In some cases we also observed the other viral precursors, Gag-Pro (Pr95) and Gag-Pro-Pol (Pr180) in these particles (data not shown). The presence of unprocessed polyproteins suggests that the cleavage of Gag-polyproteins by these PR mutants was impaired compared with the wild type. In contrast, the in vitro autoproteolytic cleavage of the isolated 17-kDa protease containing Q115I and N109I/Q115I mutations was significantly faster than the wild type, confirming that these mutations affect in vivo substrate specificity of PR and not the activity per se.
Mutations within C Terminus of Protease Decrease Virus Infectivity-To analyze the effect of the mutations introduced into the C terminus of the protease region on the virus infectivity, virus particles released from the COS-1 cells transiently transfected by mutant and wild-type viral constructs were normalized to the content of the capsid protein, and equivalent amounts of viral particles were used for infection of the fresh COS-1 cells. The level of CA protein released into the medium was determined by Western blot analysis using polyclonal rabbit anti-CA antiserum. Mutant A114R retained ϳ70% of the infectivity of wt virus (Fig. 5), but a significant decrease in infectivity was observed for mutants displaying accelerated C-terminal autoprocessing (Q115I, N109I/ Q115I) as well as for the C-terminally truncated protease mutants (PR⌬111-145, PR⌬118-145). Importantly, the G-patch mutant, Y121S, with defective nucleic acid binding retained only ϳ20% infectivity of the wt virus, confirming that this region plays an important role in virus infectivity.  The C-terminal Domain of Protease Is Important for the Activity of Reverse Transcriptase-As the decrease of virus infectivity caused by deletion or substitution mutations within the C-terminal protease domain was not because of impaired protease activity or virion maturation, we analyzed the virion-associated RT activity of all protease mutants described here (Fig. 6). Mutant and wild-type viral particles were isolated from media of transiently transfected COS-1 cells, and the rate of incorporation of [ 35 S]dATP in an RT reaction was measured. As expected, replacement A114R had no significant influence on the RT activity, which is in agreement with the results obtained from the infectivity assay. However, for constructs Q115I, N109I/Q115I, and Y121S, RT activity was reduced by 50% and deletion the G-patch region (PR⌬111-145, PR⌬118-145) decreased RT activity by 70% in comparison with wt. These results demonstrate that the G-patch domain of PR is necessary for the activity of M-PMV reverse transcriptase. Both the virus infectivity and the activity of reverse transcriptase of all the constructs presented in Figs. 5 and 6 are average values of three independent experiments.
Mutations within the C-terminal Domain of PR Change the Type of Assembly-To analyze the effect of mutations within the C terminus of protease on the morphogenesis and morphology of the viral particles, ultrathin sections were prepared and analyzed by transmission electron microscopy (Fig. 7). The wt M-PMV particles assembled within the cytoplasm and were subsequently transported to the plasma membrane, representative of D-type morphogenesis as described by Chopra and Mason (15). Unexpectedly, all constructs with a mutated C-terminal domain assembled particles with both D-type (assembly of spherical particles within the cytoplasm) and C-type (assembly at the plasma membrane) morphogenesis. The average number of C-type assembled particles for the A114R mutant was ϳ20% (TABLE ONE), which is in good agreement with lower virus infectivity and exogenous RT activity. Mutants with accelerated C-terminal autoprocessing (Q115I, N109I/ Q115I) and truncated protease mutants (PR⌬111-145, PR⌬118-145) displayed ϳ40% of the C-type assembly. The G-patch mutant assembled ϳ60% of the particles at the plasma membrane. Importantly, the release of viral particles was not impaired for any of the mutants because the amount of virus-associated proteins was comparable to that of the wt virus, indicating that the C-type assembled particles were not retained in COS cells.

DISCUSSION
Proteases from several endogenous retroviruses and ␤-retroviruses (MIA-14, HERV K10, rabbit endogenous retrovirus H, M-PMV) undergo autoprocessing from the C terminus both in vitro (16 -19) and in vivo (6) yielding proteolytically active proteases shorter by ϳ4 kDa. However, the significance of the C-terminally extended domain of PR and its processing for their life cycle has not been elucidated. The first active form of the M-PMV protease is released as a dimer of two 152 amino acid subunits (17PR), which undergoes further autoprocessing from its C terminus yielding proteolytically active homodimers that contain 114 (13PR) or 107 (12PR) amino acid monomers (6). Importantly, the RNA-binding module, whose function is in most eukaryotic proteins associated with RNA processing and also with specific proteinprotein interactions (11,20), is present at the C terminus (amino acid residues 108 -152) of M-PMV 17PR as well as in closely related simian retroviruses I and II (SRV) and in squirrel monkey retrovirus (SMRV) PRs. The G-patch is a region of ϳ45 amino acids, which contains six highly conservative glycines and a number of hydrophobic and basic residues (see Fig. 2). The position following the first Gly (Gly 120 in M-PMV PR) is occupied almost invariably by an aromatic residue (Tyr 121 in M-PMV PR) that has been shown to be critical for interaction with nucleic acid bases (21). Recently, we demonstrated that mutation Y121S in M-PMV PR prevented binding of single-stranded DNA and RNA oligonucleotides to the protease (9). Interestingly, a phylogenetic analysis showed that the G-patch is present in most endogenous retroviruses of class II that are clustered with ␤-retroviruses. This domain was probably acquired early in evolution (22), suggesting that it could fulfill a specific role in the life cycle of ␤-retroviruses.
M-PMV protease is expressed by ribosomal frame-shifting as a domain of the Gag-Pro and Gag-Pro-Pol polyproteins. Thus, three polyproteins Pr78 Gag , Pr95 Gag-Pro , and Pr180 Gag-Pro-Pol are targeted to the pericentriolar region of the cell where the immature particle is formed. The processing of polyproteins is initiated only after transport of the capsids to the plasma membrane, budding, and release of the particle when the protease is activated. M-PMV protease might utilize an additional control mechanism to delay the process of the activation in assembled immature capsids. It is tempting to speculate that the conservative G-patch domain of M-PMV PR and/or its time-dependent  release might regulate the proteolytic activity of PR and might prevent the premature processing of the Gag-polyproteins within the immature particle. However, the experiments with mutated M-PMV proviral genomes presented here demonstrate that truncation of M-PMV PR does not affect the maturation of viral particles and that the C-terminal domain of PR is not necessary for proteolysis. Even the activity of the PR mutant ⌬111-145, corresponding to the shortest variant, 12 PR, which has been never detected in released particles (6), was sufficient for maturation of M-PMV when subcloned into the proviral genome. Unexpectedly, virion-associated proteins of Q115I and N109I/Q115I mutants contained a low level of unprocessed Gag-polyproteins (Fig.  4b), and a fraction of the released particles was immature as documented by electron micrographs (Fig. 7), even though the in vitro autoprocessing of proteases containing these mutations was accelerated. We can hypothesize that the mutated 13/17PR processing site in PR is preferentially cleaved within the embedded polyprotein and yields a conformationally altered Gag-product with a sterically less accessible N-ter-minal cleavage site of the protease domain. Because the initial proteolytic cleavage within the Gag-Pro precursor is primarily directed at the N terminus of the protease, 4 release of the protease could be altered in the Q115I and N109I/Q115I mutants and processing of the Gag-polyproteins might be decreased. The results presented here clearly confirm that the G-patch domain does not affect maturation of released viral particles.
Importantly, we demonstrate that deletion of the C-terminal domain in PR⌬118-145 and PR⌬111-145 mutants, acceleration of the C-terminal processing of PR in Q115I, N109I, and Q115I mutants, and/or mutation of the RNA binding sequence (Y121S mutant) decreased the activity of RT in released virions by 50 -70% and the infectivity of the virus by 80 -95%. The M-PMV reverse transcriptase domain is expressed by the Ϫ1 ribosomal frameshift at the 3Ј-end of the protease reading frame within the Gag-Pro-Pol polyprotein. In this long precursor the protease region is followed by the reverse transcriptase sequence. It is likely that the proteolytic digestion of this polyprotein yields the reverse transcriptase covalently linked at its N terminus with the G-patch sequence of PR, since Entin-Meer et al. (23) reported that mature mouse mammary tumor virus RT contained C-terminal flanking residues of the protease within virions. In vitro experiments with this extended RT confirmed that this transframe form of RT is active. However, mouse mammary tumor virus is the only virus from the ␤-retroviruses that does not contain the C-terminal extension and the G-patch at the C terminus of the protease domain (see Fig. 2).
A surprising observation of this study was that the G-patch of the protease influences the type of virus assembly. Electron micrographs of all the mutants studied here show that morphogenesis is shifted, to different degrees, from the D-type to the C-type morphogenetic pathway. M-PMV as the prototypical representative of D-type retroviruses assembles very few particles (Ͻ1%) via the C-type pathway (24). In contrast, ϳ40% of PR⌬111-145, PR⌬118-145, Q115I, and N109I/Q115I mutant particles assembled at the plasma membrane and as many as 60% of Y121S particles displayed C-type morphogenesis. M-PMV Gag precursors together with genomic RNA assemble into capsids at a pericentriolar region of the cytoplasm in a microtubule-dependent process through interactions between a short peptide signal, known as the cytoplasmic targeting-retention signal, and the dynein/dynacting motor complex (2,25). The C-and D-type morphogenetic pathways appear distinct; however, a point mutation of Arg 55 within the CTRS signal in the matrix domain, abrogates the pericentriolar targeting of Gagpolyproteins, and this mutant follows the C-type assembly pathway (3). Insertion of M-PMV CTRS into the same region of murine leukemia virus (MoMuLV), which exhibits C-type morphogenesis, results in intracytoplasmic assembly of the Gag-polyproteins. However, only 20 -30% of MoMuLV Gag-polyproteins containing M-PMV CTRS were incorporated into stable, pelletable particles within the cytoplasm (4), suggesting that efficient assembly of M-PMV immature capsids in the cytoplasm requires other domains that are present in the Gag-related polyproteins. The G-patch of the protease, which contains the basic as well as the hydrophobic amino acids might represent a signal that promotes the efficient intracytoplasmic assembly of immature capsids by both the binding of genomic RNA during the assembly process and/or by the protein-protein interaction with cellular proteins involved in intracellular trafficking. The influence of the G-patch on the ratio of incorporated Gag-Pro and Gag-Pro-Pol polyproteins into the immature particles might be also considered.
In summary, the observations presented here describe the so far uncharacterized function of the RNA binding domain in ␤-retroviruses. We showed that this domain located within the C terminus of the protease does not affect the maturation; however, more surprisingly it influences the morphogenesis and the virus infectivity. The mutations that impair the function of the G-patch, which is evolutionarily conserved among endogenous retroviruses and ␤-retroviruses, allow the polyprotein precursors to escape from the intracytoplasmic site of assembly and redirect them to the cytoplasmic membrane.

The frequency of particles assembled with the C-type morphogenesis in COS-1 cells transfected with wt M-PMV and protease mutants
A total of 300 -400 viral capsid structures for each mutant were analyzed by transmission electron microscopy to yield the percentage for C-and D-type morphogenesis.

Construct
Frequency