Rescue of Maturation Off-Pathway Products in the Assembly of Pseudomonas Phage ϕ6

Abstract

Many complex viruses use an assembly pathway in which their genome is packaged into an empty procapsid which subsequently matures into its final expanded form. We utilized Pseudomonas phage ϕ6, a well-established virus assembly model, to probe the plasticity of the procapsid maturation pathway. The ϕ6 packaging nucleoside triphosphatase (NTPase), which powers sequential translocation of the three viral genomic single-stranded RNA molecules to the procapsid during capsid maturation, is part of the mature ϕ6 virion but may spontaneously be dissociated from the procapsid shell. We demonstrate that the dissociation of NTPase subunits results in premature capsid expansion, which is detected as a change in the sedimentation velocity and as defects in RNA packaging and transcription activity. However, this dead-end conformation of the procapsids was rescued by the addition of purified NTPase hexamers, which efficiently associated on the NTPase-deficient particles and subsequently drove their contraction to the compact naive conformation. The resulting particles regained their biological and enzymatic activities, directing them into a productive maturation pathway. These observations imply that the maturation pathways of complex viruses may contain reversible steps that allow the rescue of the off-pathway conformation in an overall unidirectional virion assembly pathway. Furthermore, we provide direct experimental evidence that particles which have different physical properties (distinct sedimentation velocities and conformations) display different stages of the genome packaging program and show that the transcriptional activity of the ϕ6 procapsids correlates with the number of associated NTPase subunits.

INTRODUCTION

The self-assembly of biological macromolecule complexes is a fundamental process involved in the formation of many central cellular operators, such as ribosomes, replicases, and viruses. Highly efficient viral in vitro assembly systems have both biomedical importance and nanotechnology potential. Consequently, viral capsid assembly pathways have become key models for macromolecule assembly. Cystoviruses, in particular Pseudomonas phage ϕ6, are among the leading virus assembly systems in which infectious virus particles have been self-assembled from purified protein and RNA constituents (1). The assembly pathway of these bacterial double-stranded RNA (dsRNA) viruses is particularly intriguing due to the complexity of their genome encapsidation pathway and the multiple enzymatic functions that are associated with assembly intermediates (2–4). Furthermore, the description of the assembly process down to an atomic level is facilitated by the high-resolution structures of key enzymatic and structural subunits (5–10).

Complex viruses, such as cystoviruses, tailed dsDNA bacteriophages, and herpesviruses, translocate their genomes to a preformed capsid (i.e., procapsid [PC]), using a genome packaging motor. In dsRNA bacteriophages (cystoviruses) a single viral gene product, P4, forms a hexameric ring-like conduit for single-stranded RNA (ssRNA) translocation into the empty PC (6, 11–13) and also acts as the packaging motor (NTPase). It requires NTP and a divalent metal ion (Mg²⁺or Ca²⁺) for hexamerization (14), is incorporated into the PC as a preformed hexamer during the assembly of the particle (1), and remains associated with the PC shell during its maturation (11, 15). Of the 12 potential P4 hexamer binding sites, approximately 11 are occupied in the virion (16). The PC framework is formed by 120 copies of protein P1 (17). Such capsid organization in which asymmetric dimers of the major capsid protein form a T=1 icosahedral lattice is common among dsRNA viruses but not observed in any other virus group (18). In addition to P4 and P1, the PC contains two minor subunits, the RNA-dependent RNA polymerase P2 (19, 20) and the assembly cofactor P7 (1, 21), both located at the interior of the PC shell close to the 3-fold axes of symmetry (22, 23).

The ϕ6 procapsid can be assembled through coexpression of protein components in an Escherichia coli recombinant expression system (24, 25) or by carrying out self-assembly of PCs using purified protein components (1). Empty PCs selectively recognize and package the viral ssRNA segments s, m, and l (26, 27), which are the precursors for the trisegmented dsRNA genome (composed of segments S, M, and L). Although each segment can be packaged in the absence of the other segments (3, 28), both in vitro and in vivo studies suggest a serial packaging mode where encapsidation of s is followed by m and then by l (3, 29, 30). The packaging of the l segment switches on RNA replication (minus-strand synthesis) that takes place inside the PC (26, 31). The resulting dsRNA segments, S, M, and L, serve as templates for subsequent RNA transcription, leading to the production of viral plus strands and their extrusion from the particle (26).

In the empty PC, the 5-fold vertices are pressed toward the interior of the capsid to form inward-oriented “cups” which are occupied by the P4 hexamers, while the virion core, which has exactly the same protein composition as the empty PC, is slightly convex at the vertices, giving it a spherical appearance (15, 17). In addition, two intermediate conformations, intermediates 1 and 2, have been described for the ϕ6 PC (17, 32). These can be produced by chemical or physical stress, e.g., low pH, high salt, or high temperature, from empty PCs or dsRNA-filled virion cores (17, 32).

A model has been put forward (32, 33) that makes a connection between the sequential packaging of the three ssRNA segments and the stepwise expansion of the PC: the empty compact PC can package only the s segment; packaging of s results in PC expansion (intermediate 1), exposing m binding sites; and packaging of m induces further expansion of the particle (intermediate 2), resulting in the packaging of l ssRNA. A point mutation in ϕ6 gene 1, which encodes the shell forming protein P1, results in the formation of PCs that package the m segment without the packaging of the s segment, and such particles were proposed to be in the intermediate 1 conformation displaying the “m segment packaging mode” (32, 34). A similar phenotype has also been described for particles that have the amino acid change of Ser₂₅₀ to Gln in P4 (12).

P4 hexamers nucleate PC assembly and are efficiently incorporated at the 12 potential binding sites on the PC shell during the course of the in vitro assembly reaction (1, 16). However, the interaction of the P4 hexamer with the PC is not stable (35): a majority (∼90%) of P4 hexamers dissociate from the PC surface if the particles are exposed to detergents (12). Similar instability in P4 association with the PC is also observed in recombinant particles that contain the amino acid change of Ser₂₅₀ to Gln (25). In both cases, an equivalent of approximately one P4 hexamer remains associated with a particle. P4-deficient particles (containing the Ser₂₅₀-to-Gln change in P4) display ssRNA packaging kinetics similar to those of the wild-type recombinantly produced PCs (12), which indicates that only one vertex is active in packaging at any given time. These particles also actively replicate the packaged RNA segments but are deficient in transcription (12).

During the life cycle of phage ϕ6, P4 hexamers drive the assembly of empty PCs (1), act as a molecular motor for the translocation of the genomic ssRNA precursors into the PC (12, 14), and apparently also have an essential role in the production of viral mRNA molecules (12, 36). In this investigation, we show that it is possible to control the assembly and packaging processes so that a reversible system is achieved where biological functions—packaging order of the ssRNA genome segments, replication, and transcription—are controlled by conformational changes in the viral particle caused by the assembly and disassembly of the hexameric packaging NTPase.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The wild-type PC and S250Q particle (which harbor the amino acid change of Ser₂₅₀ to Gln in P4) were produced in E. coli strain JM109 (37). E. coli strain HMS174(DE3) (38) was used for the expression of recombinant P4. Pseudomonas syringae pv. phaseolicola HB10Y was used as a host for Pseudomonas phage ϕ6 (39).

Plasmids pLM687 (40) and pLM1224 (25) were used to produce the wild-type PC and the S250Q particle, respectively. Protein P4 was expressed from plasmid pJTJ7 (41). Plasmids pLM659 (42), pLM656 (43), and pLM687 (40), containing the cDNA copies of the ϕ6 L, M, and S segments, respectively, were used to produce corresponding plus-sense ssRNAs.

Preparation of PC particles and purification of P4.

Recombinant wild-type PCs and S250Q particles were isolated as previously described using Triton X-114 extraction and centrifugation in a linear gradient of 5 to 20% (wt/vol) sucrose in 20 mM Tris (pH 8.0) and 150 mM NaCl (Sorvall TH641 rotor at 110,100 × g and 10°C for 110 min) and stored at −80°C (3, 12, 25). The detergent-treated particles (OG-PCs) were prepared by adding 1.5% (wt/vol) n-octyl-β-d-glucopyranoside (OG) to freshly isolated wild-type PCs as described by Pirttimaa et al. (12). The following sedimentation was carried out in a linear gradient of 15 to 40% (wt/vol) sucrose, and isolated OG-PCs were stored at −80°C. Recombinant P4 hexamers were purified as previously described (14) and stored at −80°C in 20 mM Tris (pH 7.5), 450 mM NaCl, 5 mM MgCl₂, 2 mM CaCl₂, 5 mM ATP, and 0.1 mM EDTA.

Protein concentrations were determined by the Coomassie brilliant blue method using bovine serum albumin as a standard (44). Bacteriophage ϕ6 was propagated and isolated as described previously (45) and used as a protein size marker in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (46).

In vitro assembly of P4 on preformed particles.

Purified P4 hexamers were mixed with the isolated P4-deficient (OG-PC and S250Q) particles in 60 mM Tris (pH 8.0), 1 mM ATP, 75 to 90 mM NaCl, 0.005 to 0.15 mM MgCl₂, and 0.002 to 0.06 mM CaCl₂, in the presence or absence of 6% (wt/vol) polyethylene glycol 4000 (PEG 4000). The reactions were carried out at room temperature for 30 to 60 min with a particle concentration of approximately 0.15 mg/ml. Subsequently, PCs were separated from unassembled P4 subunits by rate-zonal centrifugation in a linear gradient of 15 to 40% (wt/vol) sucrose in 20 mM Tris (pH 8.0) and 150 mM NaCl (Sorvall SW50.1 rotor at 114,000 × g and 10°C for 60 to 80 min). The light-scattering zones of the gradients were collected using a BioComp gradient fractionator, and the protein compositions of the collected fractions were analyzed by SDS-PAGE. The estimation of relative protein amounts was based on the known relationship between the protein band intensity in a Coomassie brilliant blue-stained SDS-polyacrylamide gel and relative protein quantity; relative protein quantities were converted to copy numbers by use of the known molecular weights of the proteins and the copy number 120 for P1 (16).

In vitro synthesis of ϕ6 plus-sense ssRNA genomic precursors.

Synthetic ϕ6-specific plus-sense ssRNAs, s, m, and l, were produced by in vitro transcription with T7 RNA polymerase. Templates for T7 transcription were PCR amplified from plasmids pLM659, pLM656, and pLM687 using primers complementary to the terminal sequences of the ϕ6 segments and forward primers with an upstream T7 RNA polymerase promoter (47).

In vitro ssRNA packaging, replication, and transcription assay.

Plus-strand synthesis activity assays (i.e., combined in vitro ssRNA packaging, replication, and transcription assays [4]) were carried out in 25-μl reaction mixtures containing 50 mM Tris (pH 8.9), 2 mM dithiothreitol (DTT), 0.1 mM EDTA, 5 mM MgCl₂, 6% PEG 4000, 80 mM NH₄Ac, 1 mM each ATP, GTP, CTP, and UTP, 1 U/μl of RNase inhibitor (Fermentas), 2 μCi of [α-³²P]UTP (PerkinElmer Inc.), 0.5 to 1 μg of complete PC or P4-deficient particles, and 0.5 to 1.1 μg of s, m, and l ssRNA (in equimolar ratios). The reaction mixtures were supplemented with purified P4 as indicated below. After 90 min of incubation at 30°C, the reactions were stopped by adding 2×U loading buffer (48), containing 6.7% (wt/vol) ethidium bromide, and incubating the mixtures at 50°C for 5 min. Subsequently, the reaction products were analyzed by electrophoresis using a 1.0% (wt/vol) agarose gel. The autoradiograph was recorded by using a Fuji BAS-1500 phosphorimager and BAS1500 image plate (Fujifilm). The quantitative estimation of band signal intensity was performed using AIDA image analyzer software (Raytest, Isotopenmeßgeräte GmbH).

Transmission electron microscopy.

For negative staining, a 5- to 10-μl aliquot of an ∼0.12-mg/ml particle preparation was deposited on a grid and stained with 1% (wt/vol) phosphotungstic acid (pH 7.0) for 10 s. The micrographs were taken with a Jeol JEM-1400 transmission electron microscope operating at 80 kV with a magnification of ×5,000 (Electron Microscopy Unit, Institute of Biotechnology, University of Helsinki). The ratio of expanded particles to compact ones was determined by counting several micrograph grids. For the definition of the particle conformations, see the work of Butcher et al. (17) and Nemecek et al. (32).

RESULTS

P4-deficient ϕ6 PCs accept external P4 hexamers.

Three derivatives of recombinant ϕ6 PC particles were used in this study (Fig. 1): wild-type PCs that have a protein composition similar to that of the viral cores (16, 24) (Fig. 1, lane 2) and two types of P4-deficient particles, OG-PC and S250Q, which have equivalent amounts of P4 to form one P4 hexamer per particle (12, 25) and thus 11 potential empty P4 hexamer binding sites (Fig. 1, lanes 3 and 5, respectively).

Fig 1.

Reconstitution of P4 on P4-deficient particles. Purified particles with reduced amounts of P4 (OG-PC and S250Q) were incubated with or without purified P4 hexamers at room temperature for 60 min. The particle-associated P4 hexamers were separated from the unassembled protein subunits by rate-zonal centrifugation. The light-scattering zone of the gradients was collected and analyzed by SDS-PAGE. Recombinant PCs (PC wt) from E. coli and purified ϕ6 virions were added as protein size markers. The four PC proteins are indicated on the left. The ϕ6 virion contains seven additional virus-specific proteins which are associated on the envelope or form the nucleocapsid shell structure. Six of these are visible in the gel.

P4 hexamers are essential for the in vitro assembly of the ϕ6 PCs (1, 16), but a substantial amount of P4 can be disassociated from the PC surface without apparent effects on the integrity of the P1 shell (12, 35).We applied purified P4-deficient PCs (OG-PC and S250Q) and purified P4 hexamers to investigate the reversibility of this process. P4 was initially mixed with OG-PC or S250Q particles at a molar ratio of 11:1, corresponding to a mass ratio of 15:85. After a 60-min incubation, the particles were separated from the nonassociated P4 hexamers by rate-zonal centrifugation. Incubation with P4 significantly increased the amount of P4 associated with the OG-PC (Fig. 1, lane 4) and S250Q (Fig. 1, lane 6) particles, and the average copy number for P4 hexamers in the particles increased from an equivalent of 1 to 10 (Fig. 1, compare lanes 3 and 5 with lanes 4 and 6, respectively; for the estimation of the average copy numbers, see Materials and Methods). Similar results were also obtained if the P4 reassociation reactions were carried out in the absence of PEG 4000 (ϕ6 PC self-assembly conditions [1]; see Fig. 3A).

Fig 3.

PC subpopulations displaying different sedimentation velocities have different protein compositions, morphologies, and ssRNA packaging modes. Subpopulations of particles originating from the P4 reassociation reaction mixture containing purified P4 hexamers and OG-PC particles in a 5.5:1 molar ratio were separated by rate-zonal centrifugation, and the two light-scattering zones (see Fig. 2A, lane 3, OG-PC+1/2×P4) were collected using a BioComp gradient fractionator. (A) SDS-PAGE analysis of the protein composition in the slower [PC(s)]- and faster [PC(f)]-sedimenting PC particles. Recombinant PCs (PC wt), OG-PC particles, and purified ϕ6 virions were added as protein size markers and controls for P4 quantity. The PC proteins are indicated on the left. (B) Agarose gel analysis of the reaction products of combined in vitro ssRNA packaging and replication assay with the indicated PC particles and ϕ6-specific ssRNA molecules s, m, and l. The positions of the double-stranded ϕ6 genomic segments S, M, and L are indicated on the left. Micrographs of negative-stain transmission electron microscopy of OG-PCs (C) and OG-PCs preincubated with purified P4 hexamers (in a molar ratio of 1:22) (D). The expanded particles are indicated with white arrowheads and compact particles with white arrows. The scale bar represents 100 nm and applies to panels C and D.

P4 incorporation on P4-deficient ϕ6 PCs is dose dependent.

During particle assembly, a P4 hexamer is incorporated at each, or at a majority, of the 5-fold symmetry positions of the P1 shell (16). To probe the pathway for P4 association on the potential 11 empty P4 hexamer binding sites in the P4-deficient particles, we titrated the amount of P4 in the reassociation reaction mixtures from 1/4× to 4×, where “1×” refers to a molar ratio of 11:1 between P4 hexamers and P4-deficient particles. Analyses of the recovered particles revealed that the intensity for P4 increased with an increasing amount of P4 in the reassociation reaction mixture (for OG-PC and S250Q particles, see Fig. 2C and D, respectively). The average copy numbers were 3, 7, and 10 P4 hexamers per particle in the particle populations originating from reaction mixtures with 1/4×, 1/2×, and 1× P4, respectively. At higher P4/particle ratios (22:1 and 44:1 for 2× P4 and 4× P4 reaction mixtures, respectively), potential P4 binding sites in the PC were saturated and approximately 12 P4 hexamers were detected in the resulting particles (Fig. 2C and D). These results imply that P4 hexamers can efficiently assemble on preformed ϕ6 PC particles in vitro and that the binding affinities for P4 hexamers are similar in OG-PC and S250Q particles.

Fig 2.

Dose-dependent incorporation of P4 hexamers on P4-deficient PCs is reflected in the sedimentation profile. In vitro P4 reassociation assays were carried out with increasing amounts of purified P4 and OG-PC (A) or S250Q (B) particles, where 1× P4 refers to a molar ratio of 11:1 between P4 hexamers and P4-deficient particles. The reaction products were analyzed using a linear 15 to 40% (wt/vol) sucrose gradient. The light-scattering zones of the gradients (A and B) were collected, and their protein compositions were analyzed by SDS-PAGE (C and D; particles originating from reassociation reaction mixtures containing OG-PC or S250Q particles, respectively). If two closely sedimenting light-scattering bands were observed in a gradient, they were mixed and analyzed together. Recombinant PCs (PC wt) from E. coli and purified ϕ6 virions were added as protein size markers. The PC proteins are indicated on the left.

Reassociation of P4 on P4-deficient PC changes the sedimentation velocity of the particle.

Close inspection of the sedimentation profiles of ϕ6 PCs originating from P4 reassociation reactions revealed a clearly detectable and reproducible difference in the sedimentation of P4-deficient particles and particles with high occupancy of P4 hexamers (Fig. 2A and B, compare the leftmost and rightmost sedimentation profiles). This type of difference in sedimentation can arise from a change in the frictional coefficient or volume due to particle collapse and/or a change in mass or density of the PC due to P4 binding. Interestingly, two distinct subpopulations of particles which differed in their sedimentation velocities (Fig. 2A and B) were detected under conditions in which only a limited amount of P4 was applied in the reassociation reaction (1/4× or 1/2× P4 reaction mixtures). Those particles that sedimented faster were designated PC(f), and those having a slower velocity were designated PC(s). The position of PC(f) corresponded to that observed for particles containing a high number of P4 hexamers (originating from the reaction mixtures containing 2× and 4× P4), while the PC(s) particles had a mobility in the sucrose gradient similar to that of the P4-deficient particles (OG-PCs). The formation of two subpopulations of PCs with clearly distinct sedimentation velocities (two distinct light-scattering zones in the sucrose gradient [Fig. 2A and B]) suggests a global conformational change triggered or stabilized at a defined number of P4 hexamers associated with the particle. A continuous light-scattering zone from the position of the PC(f) particles to the position of the PC(s) particles would be expected if the change was gradual.

Subpopulations of particles displaying different velocities in rate-zonal centrifugation have different ssRNA packaging modes.

To further analyze the biochemical properties of the different ϕ6 PC particles, the two particle subpopulations (the two light-scattering zones) originating from the reassociation reaction mixture containing a limited amount of P4 (1/2× P4) were collected using a BioComp gradient fractionator. Analysis of the protein composition confirmed that the faster-sedimenting subpopulation, PC(f), contained an equivalent of eight hexamers per particle, while the PC(s) subpopulation contained an equivalent of one hexamer per particle (Fig. 3A). Furthermore, the two particle subpopulations displayed different specificities in RNA packaging and minus-strand synthesis reactions (Fig. 3B). The PC(s) particles (Fig. 3B, lane 3) and OG-PC particles (12) (Fig. 3B, lane 2) preferentially packaged and replicated m and l segments. The PC(f) particles had the phenotype of the wild-type PC (Fig. 3B, lane 1), and they carried out the sequential encapsidation of the three ssRNA molecules (s, m, and l) and their replication to dsRNA (Fig. 3B, lane 4).

The current hypothesis regarding the control of sequential packaging of the three viral genomic segments s, m, and l into the ϕ6 PC proposes that the transition from s to m ssRNA packaging mode involves conformational changes in the PC (33, 49). The observed changes in sedimentation (Fig. 2A) and ssRNA replication (Fig. 3B) induced by P4 reassociation on OG-PCs encouraged us to explore particle conformation by negative-stain transmission electron microscopy (Fig. 3C and D). Both expanded particles (∼80%) and wild-type-like compact particles (∼20%) were observed in the preparation of OG-PCs (Fig. 3C). However, if the OG-PCs were incubated with purified P4 for 30 min before staining of the sample, the amount of expanded particles was only ∼10% (Fig. 3D). This indicates that the reassociation of P4 on the expanded P4-deficient particle changes the particle conformation.

The reconstitution of P4 on P4-deficient PCs rescues transcription activity.

OG-PC and S250Q particles can perform packaging and minus-strand synthesis reactions but do not display transcriptional activity (12). We carried out transcription reactions with the P4-deficient particles and supplemented the reaction mixtures with different amounts of purified P4 (1/4× to 4× P4 [Fig. 4]). Under the reaction conditions applied, empty wild-type PC packages the three plus-strand genomic precursor molecules s, m, and l (3). Inside the particle the ssRNAs are replicated to genomic dsRNA molecules (S, M, and L) which are subsequently applied as templates for transcription, resulting in the production of predominantly s and m transcripts (4) (Fig. 4A and B, lane 1). No production of plus strands was detected in reaction mixtures with OG-PC (Fig. 4A, lane 2) or S250Q (Fig. 4B, lane 2) particles. However, if purified P4 was applied in the reaction mixtures, the transcription activity of P4-deficient particles was rescued (Fig. 4A and B). The increase in transcription activity was proportional to the amount of P4 added into the reaction mixtures until the P4 hexamer-to-particle ratio was 22:1 (reaction with 2× P4 addition [Fig. 4C and D]). Under these conditions the OG-PCs approached the transcription activity of the wild-type PC (Fig. 4C). At higher P4 concentration (4×P4 reaction) the transcription activity of the OG-PC and S250Q particles decreased substantially (Fig. 4C and D). Apparently the large excess of free P4 hexamers results in an efficient turnover of NTP substrates, reducing their availability for the polymerase subunit.

Fig 4.

Effect of P4 on the plus-strand synthesis activity of P4-deficient particles. Plus-strand synthesis reactions (i.e., combined in vitro ssRNA packaging, replication, and transcription reactions) were carried out with purified OG-PC (A) or S250Q (B) particles and particles supplemented with different numbers of purified P4 hexamers. “1× P4” corresponds to a molar ratio of 1:11 between particles and P4 hexamers. Positions of double-stranded (uppercase letters) and single-stranded (lowercase letters) RNA segments are indicated on the left. The reaction with an equal amount of recombinant PC (PC wt) was applied as a positive control. The relative transcription activities, based on phosphorimager quantitation of the labeled ssRNA molecules, are shown with bar graphs for OG-PC (C) and S250Q (D). Error bars represent standard deviations of the means for three replicates.

The addition of P4 hexamers to the transcription reaction also increased label incorporation to the dsRNAs (Fig. 4A and B). This reflects the semiconservative strand displacement transcription mechanism (50). The newly synthesized strand displaces the parental plus strand, resulting in the presence of radiolabeled NTPs, to label incorporation into the dsRNA during the synthesis of the first transcripts. Both OG-PC and S250Q particles transcribed m segment more efficiently than s segment, and l segment-specific plus strands were not detected (Fig. 4A and B). The inefficient transcription of the L segment reflects the transcriptional control observed late during ϕ6 infection (51, 52) and arises from a single nucleotide difference between L and the other two segments (4, 53).

DISCUSSION

The assembly of macromolecule complexes, especially large virus capsids, displays sequentiality (54). The assembly proceeds in distinct steps in which molecular interactions formed in one step are a prerequisite for those of the following one. Further adjustment of molecular interactions takes place in complex viruses during the maturation of the PC, which, as demonstrated for several dsDNA bacteriophages, may involve exothermic processes that drive the refolding of the subunit proteins to form a more stable entity (55–59). Unidirectionality of the maturation pathway may also be further promoted through stabilization of the end product by covalent modifications, such as cross-linking of the capsid proteins (60, 61), incorporation of cementing proteins that abrogate contraction (62), or proteolytic processing (63–65).

However, truly unidirectional, irreversible macromolecule assembly pathways do not allow error correction, and such processes may lead to the accumulation of misassembled structures and off-pathways, dead-end products, or conformations. In the cystovirus system, we established that the maturation pathway of the PC displays substantial flexibility (for a summary, see Fig. 5). The premature expansion of the empty PC that is accompanied by reduction in the P4 content and defects in RNA packaging and transcription is fully reversible; free P4 hexamers efficiently assemble on expanded P4-deficient particles (Fig. 2), driving the particle to its naive compact conformation (Fig. 3C and D) and leading to the rescue of the packaging capacity as well as transcription activity (Fig. 3B and 4). The reversibility of the process enables correction of dead-end maturation pathway products, which implies that energy and materials are not wasted; rather, the production of infectious virions is maximized. Such property might have provided ϕ6 a selection advantage during its evolution.

Fig 5.

Conformational flexibility in the maturation expansion pathway of ϕ6 PC allows reprogramming of biological activities. The ϕ6 PC shell (composed mainly of protein P1) is presented in light blue. The star shape represents the compact PC shell, while the pentagon represents the expanded particle (intermediate 1). P4 hexamers are located at the icosahedral 5-fold symmetry positions of the PC shell. Wild-type P4 is in yellow and the mutant P4 (S250Q) in orange. Wild-type ϕ6 PCs are multifunctional molecular machines that carry out sequential packaging of the three single-stranded genomic RNA molecules s, m, and l, replication of the ssRNA to dsRNA, and transcription of new plus strands from the dsRNA template inside the particle. P4 hexamers can spontaneously dissociate from the PC shell. P4-deficient PCs can also be produced by detergent treatment or through expression of a cDNA clone of the ϕ6 L segment that harbors mutations in gene 4. These particles contain an equivalent amount of P4 to form one P4 hexamer per particle, thus leaving 11 P4 hexamer binding sites unoccupied. Regardless of the origin, the P4-deficient particles display a similar phenotype: they package and replicate only m- and l-specific RNAs and do not express transcription activity. Assembly of the P4 hexamers on the PC shell results in the stabilization of the compact naive conformation of the PC and also rescues its natural, biologically relevant activities.

Although tailed dsDNA and dsRNA bacteriophages represent distinct viral lineages (66), they display surprising similarities in the mechanisms and pathways of capsid assembly and genome packaging, implying mechanistic or functional molecular convergence. The presence of reversible steps in the overall unidirectional maturation pathway of the PC appears to be an additional feature shared by these viruses, as shown here for the dsRNA bacteriophage ϕ6 (Fig. 3) and previously for several tailed dsDNA phages (67–69). The interaction between the PC and the packaging machinery displays substantial flexibility, especially in the ϕ6 and T4 systems (Fig. 2 and reference 68, respectively). Nevertheless, there are details that are specific for different viral lineages and families: e.g., the PCs of dsDNA bacteriophages (T4 and λ) may package the viral genome regardless of their maturation stage (67, 68), whereas the sequential packaging program of the segmented ϕ6 genome is affected by the PC expansion (Fig. 3B). Consequently, in order to direct the expanded ϕ6 PC to its productive maturation pathway, the P4-induced conformational changes should occur prior to the onset of ssRNA packaging. Apparently, the conformational flexibility that allows correction of off-pathway maturation complexes is crucial for the efficient reproduction of these phages, and it is likely that similar flexibility has also been integrated into other macromolecule assembly pathways during their evolution.

Our results indicate that the ϕ6 PC transits between two conformers [compact and intermediate 1, i.e., PC(f) and PC(s), (Fig. 2 and 3)] without populated intermediates. Interestingly, the conformational change from the intermediate 1 to the naive compact conformation was accompanied by efficient capture of free P4 hexamers present in the environment, whereas the P4 content of the expanded particles [PC(s) particles] remained low (Fig. 3A). These results suggest that the expanded P4-deficient particles (intermediate 1, represented by OG-PC particles) may transiently adopt a conformation that exposes high-affinity binding sites for P4 hexamers, leading to P4 association with these particles and fast stabilization of the compact conformation. In other words, the initiation of this process must be much slower than the addition of the P4 subunits into the capsid. The overall procedure displays characteristics of a nucleation-limited reaction resembling the progression of PC assembly in dsRNA and dsDNA bacteriophages (70). Furthermore, the reversibility of the PC expansion implies that there are no major energy barriers between the two conformers.

A critical yet unsolved process in the assembly of complex dsRNA viruses is the selective encapsidation of all the genomic segments. Previous studies on ϕ6 have implied that mutations in genes 1 and 4 (resulting in amino acid changes of Glu₃₉₀ to Ala in P1 and Ser₂₅₀ to Gln in P4, respectively) may lead to defects in the packaging program (12, 34) similar to those observed in this study (Fig. 3B). Independent structural studies showed that the particles that carry the amino acid change of Glu₃₉₀ to Ala in P1 have a conformation similar to that of the wild-type PCs (22), except that they were expanded to the intermediate 1 stage if exposed to low pH (32). Interestingly, the authors also reported that some preparations of expanded PC particles (intermediate 1) contain very little P4-related density (32). Here we provide direct experimental evidence that particles which display different physical properties (distinct sedimentation velocities and conformations) represent different stages of the genome packaging program (Fig. 2 and 3). Our results support the model that the sequential packaging of the three ϕ6 genomic ssRNA molecules s, m, and l is controlled through conformational changes in the particle (49). Our results further imply that the expanded particles that are in the m segment packaging mode have significantly reduced affinity for P4 hexamers (Fig. 2 and 3). Nevertheless, stable association of P4 on the particle is crucial in the following steps of the viral life cycle in which the presence of P4 is needed for viral mRNA production (Fig. 4).

ϕ6 P4 provides the conduit for nucleic acid translocation and powers the actual RNA translocation (6, 12, 13). Despite its apparent simplicity (71), P4 is a multifunctional protein that is involved in many essential steps in the viral life cycle (nucleation of capsid assembly, genome packaging, maturation control, and transcription). These features make P4 hexamers an attractive model and potential tool for nanotechnological applications.