Activation of the Mason–Pfizer monkey virus protease within immature capsids in vitro

Abstract

For all retroviruses, the completion of the viral budding process correlates with the activation of the viral protease by an unknown mechanism, and, as the structural (Gag) polyproteins are cleaved by the viral protease, maturation of the immature virus-like particle into an infectious virion. Unlike most retroviruses, the Mason–Pfizer monkey virus Gag polyproteins assemble into immature capsids within the cytoplasm of the cell before the viral budding event. The results reported here describe a unique experimental system in which Mason–Pfizer monkey virus immature capsids are removed from the cell, and the protease is activated in vitro by the addition of a reducing agent. The cleavage of the protease from the precursor form is a primary event, which proceeds with a half time of 14 min, and is followed by authentic processing of the Gag polyproteins. Activity of the viral protease in vitro depends on pH, with an increase in catalytic rates at acidic and neutral pH. The initiation of protease activity within immature capsids in vitro demonstrates that viral protease activity is sensitive to oxidation-reduction conditions, and that the viral protease can be activated in the absence of viral budding.

The formation of an infectious retrovirus begins as the structural (Gag) polyproteins assemble into an immature capsid in the cytoplasm of the cell. An infrequent ribosomal frame-shifting event near the C terminus of the gag gene results in the translation of a Gag-Pol polyprotein, which is incorporated into the immature capsid and includes the viral aspartyl protease. As the immature capsid is released from the host cell in the viral budding process, the Gag polyproteins are cleaved by the viral protease. Gag processing by the viral protease is associated with a morphologic maturation of the viral particle: the spherical immature capsid condenses to form the asymmetric viral capsid found in the mature virion.

Retroviral proteases, with the exception of HIV type-1 in selected expression systems, are generally inactive in the form of Gag-Pol precursor polyproteins during the assembly of immature capsids within the cell (1–15). When imaged by electron microscopy, particles composed of Gag and Gag-Pol polyproteins remain in an immature form throughout the entire viral budding process, whereas most of the particles that have been released have proceeded to a mature morphology (6, 11, 16). Although the cleavage of Gag polyproteins seems to proceed very quickly at a late stage of viral budding, it is difficult to measure the kinetics of the Gag cleavage events in vivo where the particle-maturation events are not synchronized (4, 17). The mechanism by which the initiation of viral protease activity is coupled with the viral budding process is unclear. However, the oxidation of a highly conserved cysteine residue (Cys-95) at the dimer interface in the HIV type-1 protease, or a highly conserved methionine residue in the same location (Met-95) in the HIV type-2 protease, has been proposed as a possible mechanism for the suppression of protease activity before the completion of viral budding (18, 19).

For most retroviruses, Gag and Gag-Pol polyproteins assemble into an immature particle at the plasma membrane during membrane extrusion and viral budding, in a process referred to as “C-type” viral assembly. Particle assembly in the cytoplasm of the host cell, away from the plasma membrane, is a feature unique to the betaretroviruses (B- and D-type retroviruses), and the spumaviruses. Mason–Pfizer monkey virus (M-PMV) is the prototype D-type retrovirus, and because the fundamental assembly steps (Gag particle formation, transport to the membrane, membrane interaction/viral budding, and viral release) are separate sequential processes, this virus has served as an important model of retroviral assembly. During maturation of an M-PMV virion, the Gag polyproteins (Pr78^Gag) are processed by the protease into six mature products. These are, in N- to C-terminal order, matrix (MA-p10); the phosphoprotein (pp24/pp16–18); p12 (internal scaffold domain); capsid (CA-p27); nucleocapsid (NC-p14); and p4 (unnamed) (20, 21). As with most retroviruses, the M-PMV protease is expressed by ribosomal frame-shifting. In this virus a separate Gag-Pro precursor is synthesized with a Gag to Gag-Pro molar ratio of ≈6:1. The M-PMV reverse transcriptase and integrase are expressed in a separate reading frame (pol) that is accessed by infrequent ribosomal frame-shifting near the C terminus of the protease reading frame. Thus, three polyproteins are made as a product of M-PMV Gag translation and they are, in decreasing order of amounts found within the immature particle, Pr78^Gag, Pr95^Gag-Pro, and Pr180^Gag-Pro-Pol. The active form of the M-PMV protease is released as a 17-kDa product (p17) from the Pr95 precursor by proteolytic cleavage at the N terminus of the protease domain (22–24). Several M-PMV Gag domain mutants have been characterized in which viral assembly is arrested at different stages, including transport of assembled particles to the plasma membrane, initiation of viral budding, and the release of tethered particles from the host cell (9, 10, 25). With each of these mutants, immature capsids accumulate at the stage of the assembly block, but the Gag precursor polyproteins are not processed into mature products. Thus, M-PMV particle assembly and the completion of each of these assembly steps are not sufficient to initiate viral protease activity. Like most C-type retroviruses, M-PMV protease activity is strictly limited to the completion of the viral budding process.

To characterize the processing of Gag polyproteins and the mechanism by which viral protease activity is regulated within an assembled immature capsid, we have developed an experimental system in which M-PMV immature capsids are harvested from cell lysates, and the viral protease is activated by the addition of a reducing agent. This system demonstrates the importance of reduction potential in regulating the activity of the viral protease and, because protease activation in all immature capsids is synchronized, the kinetics of Gag and Gag-Pro precursor processing into mature viral proteins can be measured over time. A kinetic analysis demonstrates that the viral protease (p17) is cleaved from the Pr95^Gag-Pro precursor with a half time of 14 (±4) min, and the Pr78^Gag polyproteins are processed with a half time of 20–25 min.

Materials and Methods

Expression of M-PMV Proviral Clones and Labeling of M-PMV Gag Polyproteins in Cultured Mammalian Cells.

COS-1 cells were grown in DMEM, (GIBCO/BRL) supplemented with 10% FBS (GIBCO), at 37°C in 5% CO₂. Cells were transfected by M-PMV proviral plasmid DNA [pSHRM15 (26)] with a liposomal transfection reagent following the manufacturer's protocol (FuGENE 6, Roche Molecular Biochemicals). Thirty-six hours after transfection, cells were incubated in methionine-free medium (Sigma-Aldrich) for 30 min followed by pulse labeling for 60 min with ³⁵S protein labeling mix (EXPRE³⁵S³⁵S, NEN) in methionine-free media at a concentration of 0.5 mCi/ml. The labeling mix was removed, and the cells were incubated in medium without labeling mix for 60 min.

Purification of M-PMV Immature Capsids from Cell Lysates.

The cells were lysed in 1% Triton X-100 (Sigma) buffer at 4°C, and the immature capsids were purified by centrifugation at 77,000 × g through a series of sucrose gradients as described (27).

Activation of the M-PMV Protease within Immature Capsids in Vitro.

Equal volumes of ³⁵S-labeled immature capsids were diluted into buffers with a range of pH from 5.0 to 8.5 (pH 5.0 and 5.5, 50 mM sodium acetate; pH 6.0–8.5, 50 mM Tris⋅HCl) and with a final concentration of 5 mM DTT (Roche). The reactions were incubated at 37°C, and aliquots were removed at measured intervals and mixed with an equal volume of 2X protein gel loading buffer before analysis by SDS/PAGE (28).

Quantitative Analysis of Reaction Products.

Dried SDS/PAGE gels were used to expose phosphor storage screens, and the screens were read by a Cyclone storage phosphor system and analyzed by optiquant image analysis software (Packard).

Electron Microscopy of M-PMV Immature Capsids.

M-PMV immature capsids were incubated in 5 mM DTT at pH 7.5 for 10–15 min and placed on Formvar carbon-coated grids (Electron Microscopy Sciences, Fort Washington, PA) before staining with phosphotungstic acid and imaging by electron microscopy (29).

Matrix Assisted Laser Desorption Ionization–Time of Flight Mass Spectroscopy of in Vitro Protease Reaction Products.

M-PMV immature capsids were purified from Spodoptera frugiperda insect cells after infection with a recombinant baculovirus expressing the M-PMV gag, pro, and pol genes (30), using the methods described above. The particles were incubated in 50 mM DTT at pH 7.5 for 1 h and analyzed by mass spectroscopy as described (27).

Results

Purification of M-PMV Immature Capsids from Cell Lysates.

M-PMV Gag polyproteins were labeled within COS-1 cells for 1 h with [³⁵S]methionine, and assembled immature capsids were harvested from cell lysates after a 1-h chase period in media without label. Previous kinetic studies of M-PMV assembly and release in vivo demonstrated that immature particles assemble in this time frame but have not completed the viral budding process (11, 26). An SDS/PAGE analysis of velocity gradient fractions containing M-PMV immature capsids showed the Pr78^Gag and Pr95^Gag-Pro polyproteins are the major constituents of the immature capsids, with smaller amounts of the Pr180^Gag-Pro-Pol polyprotein (Fig. 1). The products of ribosomal initiation at an internal Gag AUG (in-frame) codon, corresponding to the methionine residue at the end of the matrix domain, can also be seen as bands migrating ahead of the Pr78 and Pr95 polyproteins (identified as Pr68 and Pr85, respectively, in Fig. 1). Mature Gag products and the viral protease (p17) are not seen in purified immature capsids, demonstrating that the viral protease is not activated during purification and manipulation of the immature capsid outside of the cell. However, the ≈80-kDa N-terminal fragment of Pr95 (termed Pr95ΔPro) remaining after the cleavage of p17 (protease) can be detected within the immature capsids, suggesting that the release of p17 from the Pr95 precursor may proceed as a relatively inefficient process before or, less likely, after the particle assembly event.

Figure 1.

Separation of M-PMV immature capsids from cell lysates. As a final purification step, M-PMV immature capsids were centrifuged through a 5–20% (wt/vol) sucrose velocity gradient. The gradient fractions are labeled from Top (lane 1) to Bottom (lane 11, including the pellet). Of each gradient fraction, 1% is analyzed directly by SDS/PAGE. A region of the gel corresponding to lane 4, in which the Gag and Gag-Pro precursors are found, is expanded on the right of the figure for identification of the individual precursor polyproteins. Most of the M-PMV immature capsids sediment into fractions 4–6 and are composed of Pr78^Gag, Pr95^Gag-Pro, and Pr180^Gag-Pro-Pol polyproteins, as well as the Pr95ΔPro cleavage product. Other Gag products include Pr85 and Pr68, which are a consequence of internal initiation during the translation of Pr95 and Pr78, respectively.

Activation of the M-PMV Protease within Immature Capsids in Vitro.

The extended incubation of M-PMV immature capsids under a variety of conditions, including pH values from 5.0 to 9.0, NaCl concentrations from 0 to 1.0 M, and the addition of fractionated cell lysates, including cell membranes, all at 37°C, does not result in the cleavage of the Pr78^Gag and/or Pr95^Gag-Pro precursor polyproteins into mature viral proteins (data not shown). However, when immature capsids with an active protease were treated with a reducing agent (5 mM DTT), the Pr78^Gag and Pr95^Gag-Pro precursors were processed over time into smaller products that, with one exception, appear identical by SDS/PAGE to the proteins found within mature M-PMV virions (Fig. 2, lane 4). Although a small fraction of p17 was also released from Pr95 when capsids were incubated in 10 mM reduced glutathione, DTT was much more effective in activating proteolytic processing (data not shown). Wild-type (wt) M-PMV immature capsids were compared with immature capsids made from a proviral construct with a point mutation within the protease active site (D26N) as a negative control with no proteolytic activity (16). The Gag polyproteins within D26N immature capsids (Fig. 2, lane 2) remained intact after treatment with DTT. Thus, Pr78^Gag processing induced by exposure to DTT must be a consequence of viral protease activity. The mature viral proteins identified in Fig. 2 include the major capsid protein (p27CA), phosphoprotein products (pp24 and pp16–18), the matrix protein (p10MA), and the viral protease, (p17^Pro). Other Gag cleavage products (p12, p14NC, and p4) do not contain methionine residues and were not seen in this assay. During M-PMV maturation, the phosphoprotein (pp24) is cleaved at an internal site, and the C-terminal half (pp16–18) is found within the virion (20). However, when the viral protease is activated in vitro, the phosphoprotein remains intact. The internal cleavage of pp24 may be very inefficient or depend on an intravirion environment that is not reproduced within immature capsids in vitro.

Figure 2.

Activation of the M-PMV protease within immature capsids in vitro. The contents of M-PMV immature capsids with an inactive protease (D26N, lanes 1 and 2) were analyzed by SDS/PAGE with autofluorography and compared with immature capsids with an active protease (wt, lanes 3 and 4) before (lanes 1 and 3) and after (lanes 2 and 4) treatment with DTT. The cleavage products found in wt immature capsids are compared with the mature viral proteins found within M-PMV virions (lane 5) and are labeled as major capsid protein (p27 CA), uncleaved phosphoprotein (pp24), the cleaved phosphoprotein (pp16–18), the viral protease (p17^Pro), and the matrix protein (p10 MA).

To confirm the identity of each Gag cleavage product, M-PMV immature capsids with an active protease were harvested from a recombinant baculovirus-insect cell expression system by the same methods described above. The immature capsids were incubated in 50 mM DTT at pH 7.0 and 37°C for 2 h, followed by SDS/PAGE analysis. Each of the in vitro Gag and Gag-Pro cleavage products was removed from a Coomassie blue-stained gel, digested with trypsin, and subjected to analysis by mass spectroscopy. The identity of each cleavage product was confirmed by matching the trypsin digestion products with predicted trypsin digestion products within the Pr78^Gag and Pr95^Gag-Pro polyproteins (data not shown). The entire in vitro reaction was also analyzed by mass spectroscopy, and the mass for each reaction product was compared with the predicted mass for the corresponding viral protein, based on amino acid content and known posttranslational modifications (Table 1, and Fig. 7, which is published as supporting information on the PNAS web site, www.pnas.org). With the exception of p14NC, all of the mature viral proteins (p27, pp24, p12, p10, and p4) were identified in the in vitro reactions and confirmed as authentic cleavage products.

Table 1.

Mass of in vitro reaction products

	p10	pp24	p12	p27	p4	p17Pro
In Vitro	12,050	12,667*	9,349	24,684	3,970	16,979
		12,748†
		12,828‡
Predicted	12,049	12,666	9,349	24,686	3,968	16,982

The mass of each in vitro terminal reaction product was measured by mass spectroscopy (labeled as in vitro) and compared with the predicted mass for each of the corresponding mature M-PMV viral proteins based on amino acid content and known post-translational modifications (labeled as predicted). The phosphoprotein (pp24) is present in unphosphorylated form ( 

^*), and with phosphorylation at one ( 

^†) or two ( 

^‡) sites. 

Electron Microscopy of M-PMV Immature Capsids.

To determine whether activation of protease was accompanied by morphological changes that mimic maturation, wt and D26N immature capsids were imaged by electron microscopy. The structure of the immature capsid itself is unaffected by the presence of DTT because the D26N particles show no major morphological differences after treatment with DTT (compare Fig. 3 C and D). Thus, the initiation of protease activity is not based on disruption of immature capsid structure. In contrast, there is a dramatic change in the morphology of wt particles after treatment with DTT for 10 min, with evidence of a condensation of Gag proteins within the particle (compare Fig. 3 A and B).

Figure 3.

Electron microscopy of M-PMV immature capsids. M-PMV immature capsids with an inactive protease (C and D) were imaged by negative stain electron microscopy and compared with immature capsids with an active protease (A and B) before (A and C) and after (B and D) the addition of DTT. The particles within the high-magnification Insets are representative of a minimum of 50 particles observed in multiple images for each sample.

Processing of M-PMV Gag and Gag-Pro Precursor Polyproteins by the Viral Protease in Vitro Over Time.

The ability to initiate synchronous processing of the capsid precursors provided a unique opportunity to study the kinetics and order of this process. To establish the time course of the Gag processing events in vitro, aliquots were removed at measured intervals from a sample of [³⁵S]methionine-labeled wt immature capsids incubated in 5 mM DTT at pH 7.0 and 37°C. The samples removed at each time point were compared by SDS/PAGE with an equivalent amount of immature capsids that were not treated with DTT (Fig. 4, lane C). The viral protease (p17) appears at the earliest time points (2.5–5 min), and increasing amounts of p17 over time are associated with decreasing amounts of the Pr95^Gag-Pro precursor and increasing amounts of a large N-terminal Gag-Pro cleavage product (Pr95ΔPro). The appearance of p17 is also associated with Pr78^Gag processing into the p10MA and pp24 cleavage products that appear between 15 and 30 min after addition of DTT. The p27CA protein is not released efficiently as a terminal cleavage product and appears later in the reaction. The release of the p17 protease from the Pr95^Gag-Pro precursor is nearly complete at the 60-min time point. Processing of the Pr78^Gag precursor into mature viral proteins does not proceed to completion in this in vitro system, as it does during maturation of an enveloped virion. Although Pr78 is completely cleaved, several partial Gag cleavage products remain in the reaction at 120 min. The inefficiency of Pr78^Gag processing in vitro may be a consequence of protease diffusion out of these immature capsids that are not bound by viral membranes. Efficient Gag processing may also depend on molecular rearrangements that occur during the viral budding process, which do not occur in these in vitro reactions.

Figure 4.

Processing of M-PMV Gag and Gag-Pro precursor polyproteins by the viral protease over time. (A) M-PMV immature capsids with an active viral protease were treated with DTT, and the precursor polyproteins, as well as the cleavage products, were analyzed by SDS/PAGE with autofluorography at measured intervals (2.5, 5, 10, 15, 30, 60, and 120 min) and compared with untreated immature capsids (control, labeled as “C”). The Gag, Gag-Pro, and Gag-Pro-Pol precursor polyproteins are identified as well as the final cleavage products—major capsid protein (p27 CA), phosphoprotein (pp24), protease (p17^Pro), and matrix protein (p10 MA). (B) The relationship of the individual Gag and Gag-Pro cleavage products within the precursor polyproteins is diagramed, with arrows delineating the protease cleavage sites, and “fs” delineating the site of the frame shift.

Quantitative Kinetics of M-PMV Gag and Gag-Pro Processing by the Viral Protease in Vitro.

The intensity of each of the bands at each time, from experiments similar to that described above, was quantitatively measured by a phosphor storage screen. The intensity of each band was adjusted for methionine content, and the average amount of each precursor (Pr78^Gag, Pr95^Gag-Pro) from three independent time course experiments was plotted over time relative to the amount measured in the control sample, which was not treated with DTT (Fig. 5). The average amounts for each of the mature cleavage products (p17, p27CA, pp24, and p10MA) were also plotted, relative to the sum of the amounts of the appropriate precursor products in the control sample. When protease activity is initiated with the addition of DTT, the Pr95^Gag-Pro precursors are processed at early time points primarily by cleavage into p17 and Pr95ΔPro with an apparent half time of 10–15 min. The amount of the N-terminal Pr95 cleavage product (Pr95ΔPro) increases with the release of p17 from the Pr95^Gag-Pro precursor over the first 10 min of the in vitro reaction and then disappears from the reaction as this Gag precursor is processed into Gag cleavage products by the viral protease.

Figure 5.

Kinetics of M-PMV Gag and Gag-Pro cleavage into mature viral proteins. The amounts of the Gag (Pr78) and Gag-Pro (Pr95) precursor polyproteins, the Pr95ΔPro cleavage product, were quantitatively measured at precise intervals (2.5, 5, 10, 15, 30, 60, and 120 min) after the addition of DTT and plotted as “% remaining” relative to the amount of precursor measured in an equal quantity of an untreated control sample. The major capsid (p27 CA), phosphoprotein (pp24), protease (p17^Pro), and matrix protein (p10 MA) cleavage products were also plotted (as “% product”) relative to the amount of the appropriate precursor(s) in the control sample.

There is a delay in the processing of Pr78^Gag relative to the release of p17 from the Pr95^Gag-Pro precursor, suggesting that the release of p17 is a primary event (in cis or in trans) and that cleavage within the Gag domain begins only after significant amounts of p17 are present within the immature capsid. Although Pr78^Gag processing is delayed by 5–10 min, the parallel slope of the lines representing the percent Pr78 and Pr95 remaining (beyond 5–10 min) indicates that both Pr78 and Pr95 precursors are processed at similar rates after the generation of p17 within the immature capsid. Approximately half of the Pr78 Gag remains in the reaction at 20–25 min. At the end of the in vitro reactions, the mature Gag cleavage products are present at 40 (p10MA), 35 (pp24), and 15% (p27CA) of the amount of precursor polyproteins present in the untreated control sample.

To calculate the rate of the enzymatic reaction in which p17 is cleaved from the Pr95 precursor, the reaction was assumed to be of the first order (autocatalytic). The amount of p17 remaining in each of three reactions at each time point was subtracted from the initial amount of the protease precursors in the untreated control, and the natural logarithm of this value was plotted vs. time. The plots generated a straight line (correlation coefficient = 0.991–0.999), and the half time for the reaction was calculated to be 14 min, with as SD of 4 min, when averaged over 3 time course experiments. Thus, this synchronized in vitro system allows an unprecedented opportunity to measure the initial rate for release of the protease from the Pr95 precursor.

Processing of M-PMV Gag and Gag-Pro Precursor Polyproteins by the Viral Protease as a Function of Reaction pH.

Previous work on bacterially expressed protease had defined an optimal pH of 4.5 (23). To determine whether this optimal pH value changed in the context of the immature capsid, the processing of both Pr78^Gag and Pr95^Gag-Pro precursor polyproteins was measured over time during in vitro reactions at pH 7.0 and compared with in vitro reactions at pH values ranging from 5.5 to 8.5 (Fig. 6). Although the efficiency of p17 release from the precursor was similar at acidic to neutral pH (pH 5.5 and 7.0 in Fig. 6B), processing of the Gag polyprotein was somewhat more efficient at acidic pH when compared with neutral pH conditions (compare pH 5.5 and 7.0 in Fig. 6A). For both sets of reactions, there was a progressive decrease in catalytic rates at higher pH (pH 7.5–8.5).

Figure 6.

Processing of M-PMV Gag and Gag-Pro precursor polyproteins by the viral protease as a function of reaction pH. The amounts of the Gag [Pr78 (A)] and Gag-Pro [Pr95 (B)] precursor polyproteins were quantitatively measured at precise intervals (2.5, 5, 10, 15, 30, 60, and 120 min) after the addition of DTT, at pH values ranging from 5.5 to 8.5, and plotted relative to the amount of precursor measured in an equal quantity of an untreated control sample. Values for pH 6.0 and 6.5 are omitted for clarity because they overlaid the values for pH 5.5.

Discussion

When M-PMV immature capsids are harvested from cell lysates and manipulated in vitro, the viral protease remains in the form of a stable inactive Pr95^Gag-Pro precursor. In contrast, when M-PMV immature capsids are treated with DTT in vitro, the Gag and Gag-Pro precursor polyproteins are processed into mature viral proteins. This specific activation of viral protease activity in vitro demonstrates that the viral budding process is not necessary for protease activity and Gag polyprotein processing. The treatment of immature capsids with DTT may reproduce the mechanism that initiates protease activity in vivo. However, DTT could reduce cysteine disulfide bonds formed as particles are oxidized during purification or activate the protease in other ways that are unrelated to mechanism(s) that regulate protease activity in vivo. The activity of the HIV type-1 and -2 protease within the enveloped immature virus-like particle is enhanced in the presence of DTT, and this effect depends on the presence of a cysteine (HIV type-1) or methionine (HIV type-2) residue at the dimer interface near the C terminus of the protease (18, 19). It has been proposed that the activity of the HIV type-1 protease may be inhibited within the cell when cysteine residues form mixed disulfides with glutathione, and the effect may be reversed at the completion of viral budding by a thioltransferase at the plasma membrane (18, 31). This mechanism of regulation may be conserved among different retrovirus genera because a group of cysteine and/or methionine residues is conserved in a similar context in the C-terminal half of the protease domain of the B- and D-type retroviral proteases (see Fig. 8, which is published as supporting information on the PNAS web site), and as we show here proteolysis can be activated in the presence of DTT. The cytosol is a reducing environment, which contains 1–10 mM reduced glutathione (32, 33). That 10 mM glutathione did not efficiently activate the protease in vitro is consistent with the concept that it is necessary to reduce a mixed disulfide bond in order for activation to occur and that the reduction potential found within the cell is insufficient for this function.

In addition to providing insight into the possible mechanism for regulating viral protease activity, the experimental system described here allows an opportunity to quantitatively study the proteolysis of Gag and Gag-Pro polyproteins within the context of an assembled immature capsid in the absence of protease inhibitors. By activating the viral protease within the immature capsid in vitro, the initiation of proteolysis is synchronized, and the kinetics of the Gag and Gag-Pro processing events can be precisely measured. In this way we have shown that in vitro, the M-PMV protease is released from the Gag-Pro precursor with a half time of 10–15 min and the Gag polyproteins are processed with half times of 20–25 min. There is no experimental measurement for the rate at which Gag polyproteins are processed in vivo that would allow us to compare the kinetics of viral maturation in vivo with the quantitative measurements from the in vitro assay described here. Although Gag processing seems to be almost simultaneous with the release of a viral particle from the cell, it is difficult to determine when proteolysis begins relative to the time at which viral particles can be harvested and assayed for the presence of a mature vs. immature state. Although the M-PMV Gag polyproteins are incompletely processed by the viral proteinase in this in vitro system, it seems likely that the initial rates we have observed are valid and precede the potential diffusion of substrate and enzyme from the nonenveloped particle. Indeed, previous work with rapidly harvested Rous sarcoma virus (RSV) virions showed that particles harvested every 5 min contained a significant amount of unprocessed Gag precursor polyproteins, arguing that the rates we describe here are relevant to the in vivo maturation process (17). The M-PMV immature capsids used in vitro are different from budding virions in that they are not bound by a viral envelope and have not been altered by conformational changes that may occur during the interaction with the plasma membrane and the viral budding event. Both of these differences may contribute to the incomplete Gag processing observed in vitro and could also affect the rate of Gag processing.

When the protease is activated in vitro, the mature form of the viral protease (p17) is the first product to appear in the reaction, and Gag precursor processing is a relatively slow process until significant amounts of p17 have been released from the Gag-Pro precursor. Thus, the initial activity of the protease within the Gag-Pro precursor is primarily directed at the N-terminal p17 cleavage site. This is consistent with the observation that mutations in this cleavage site dramatically interfere with Gag polyprotein processing (refs. 34–38 and M. Andreansky and E.H., unpublished data). The mature Gag products generated by activation of the protease in vitro are the same as those observed in mature virions, with one notable exception. Although the M-PMV phosphoprotein is cleaved into N- and C-terminal halves during viral maturation in vivo, protease activity in vitro yields pp24 as the exclusive terminal phosphoprotein cleavage product. This lack of internal cleavage is similar to what is observed in simian retrovirus type-2, where the pp21 phosphoprotein does not undergo internal cleavage (39).

The activity of the mature M-PMV protease (p17) has been well characterized when incubated in solution with small peptides containing M-PMV Gag cleavage sites. Under these conditions, the protease is most efficient at acidic pH (5.0–5.5) and has an affinity for the p10-pp24 Gag cleavage site that is 10 times greater than that for other Gag cleavage sites (23). However, the activity of the protease described here in the context of an assembled immature capsid is equally efficient at acidic pH and at neutral pH. In addition, the p10 matrix protein and the pp24 phosphoprotein are released from assembled immature capsids in vitro with nearly equal efficiency, suggesting that the viral protease is equally active at both cleavage sites. Thus, there are significant differences in the optimal conditions for protease cleavage of peptides in solution and cleavage of the same sites within the assembled immature capsid. Moreover, it makes it less likely that a change in pH value serves as a mechanism in the control of viral protease activity in vivo.

In conclusion, the proteolytic activity within the M-PMV Pr95^Gag-Pro precursor is effectively suppressed within an assembled immature capsid, but with addition of DTT, there is a dramatic increase in protease activity. The activation of the protease results in the cleavage of p17 from the Pr95 precursor as a primary event, followed by processing of the Gag polyproteins into other mature viral proteins. The experimental system described here will provide a unique way to define the control mechanisms that regulate viral protease activity in vivo. Furthermore, this system provides a more authentic environment in which to characterize the structural constraints and substrate requirements of Gag cleavage events that occur within a viral particle during viral maturation.

Abbreviations

M-PMV

Mason–Pfizer monkey virus

wt

wild type

Note Added in Proof.

M-PMV immature capsids have been prepared, and incubated at 37°C, in the presence of 0.1 mM reduced glutathione without evidence of proteolytic processing (see Fig. 9, which is published as supporting information on the PNAS web site). Thus, the activation of the viral protease with the addition of DTT is not due to the reversal of an artificial oxidation of the immature capsids during purification.