ABSTRACT
During ϕX174 morphogenesis, 240 copies of the external scaffolding protein D organize 12 pentameric assembly intermediates into procapsids, a reaction reconstituted in vitro. In previous studies, ϕX174 strains resistant to exogenously expressed dominant lethal D genes were experimentally evolved. Resistance was achieved by the stepwise acquisition of coat protein mutations. Once resistance was established, a stimulatory D protein mutation that greatly increased strain fitness arose. In this study, in vitro biophysical and biochemical methods were utilized to elucidate the mechanistic details and evolutionary trade-offs created by the resistance mutations. The kinetics of procapsid formation was analyzed in vitro using wild-type, inhibitory, and experimentally evolved coat and scaffolding proteins. Our data suggest that viral fitness is correlated with in vitro assembly kinetics and demonstrate that in vivo experimental evolution can be analyzed within an in vitro biophysical context.
IMPORTANCE Experimental evolution is an extremely valuable tool. Comparisons between ancestral and evolved genotypes suggest hypotheses regarding adaptive mechanisms. However, it is not always possible to rigorously test these hypotheses in vivo. We applied in vitro biophysical and biochemical methods to elucidate the mechanistic details that allowed an experimentally evolved virus to become resistant to an antiviral protein and then evolve a productive use for that protein. Moreover, our results indicate that the respective roles of scaffolding and coat proteins may have been redistributed during the evolution of a two-scaffolding-protein system. In one-scaffolding-protein virus assembly systems, coat proteins promiscuously interact to form heterogeneous aberrant structures in the absence of scaffolding proteins. Thus, the scaffolding protein controls fidelity. During ϕX174 assembly, the external scaffolding protein acts like a coat protein, self-associating into large aberrant spherical structures in the absence of coat protein, whereas the coat protein appears to control fidelity.
INTRODUCTION
Due to their rapid replication cycle, detailed structural information, and well developed genetic and biochemical assays, the microviruses have become an ideal model system for experimental evolution (1–4). The morphogenetic pathway can be characterized under restrictive conditions and adaptive mutations can be mapped onto X-ray structures, providing insights into evolutionary mechanisms and protein structure-function relationships (2–8).
Proper virion assembly requires a series of precise protein-protein interactions proceeding along an ordered morphogenetic pathway. During the assembly of the T = 1 capsid, early intermediates are directed into large macromolecular structures by transiently associated scaffolding proteins. These proteins mediate conformational switches, lower thermodynamic barriers, and ensure morphogenetic fidelity (9, 10). Unlike most viruses, which use a single scaffolding protein, microviruses utilize a dual-scaffolding-protein system (Fig. 1A), which divides the pathways into two phases. During early assembly, five copies of the internal scaffolding protein B bind the underside of each viral coat protein F pentamer along with one DNA piloting protein H, yielding the 9S* particle (11, 12). These interactions allow the 9S* intermediates to associate with spike protein G pentamers (6S particles) to produce 12S* intermediates. During late assembly, 240 copies of the external scaffolding protein D, most likely arranged as asymmetric dimers, organize 12 12S* particles into procapsids (11, 13). As the procapsid is filled with single-stranded DNA (ssDNA), the internal scaffolding proteins exit the procapsid. Following genome packaging, the external scaffolding proteins dissociate to complete maturation.
ϕX174 procapsid morphogenesis and X-ray structures of the ϕX174 external scaffolding D protein. (A) The assembly pathway in cells with only wild-type D protein, only DG61X protein (X = valine, aspartic acid, lysine, or proline), and both wild-type and lethal dominant DG61X protein. (B) The assembly-naive DADB dimer (PDB accession number 1TX9 [13]). (C) The D protein tetramer (D1 to D4) as arranged in the atomic structure of the procapsid (PDB accession number 1CD3 [14]). Pink depict subunits in which α-helix 3 is straight, whereas this helix is kinked in the gray subunits. α-Helix 3 is highlighted in gold and DG61 in black with adjacent black dots to aid visualization.
The X-ray crystal structure of the ϕX174 assembly-naive external scaffolding protein D is an asymmetric dimer. Each 16,806-Da subunit consists of seven α-helices separated by six loop regions (13). Each dimer subunit assumes a different conformation, DA and DB (Fig. 1B). In the procapsid atomic structure (14, 15), each coat protein is decorated with four structurally distinct D proteins, arranged as dimers of dimers, D1D2 and D3D4 (Fig. 1C). This arrangement allows D protein interactions across the procapsid's 2-fold axes of symmetry. To achieve this unique, non-quasi-equivalent organization, D proteins exist in two major conformations: the D1 and D3 subunits are bent 30° in α-helix 3, while the corresponding helices in the D2 and D4 proteins are straight (Fig. 1C). This critical kink occurs at glycine residue 61 (DG61) and is also present in the DA subunit of the assembly-naive dimer (Fig. 1B). Based on Ramachandran plot predictions, only a glycine residue can kink the helix to this extent.
Mutant D genes that encode large side chain substitutions (DG61X, where X is valine, aspartic acid, lysine, or proline) do not complement nullD strains and inhibit wild-type (WT) morphogenesis (12). During nullD infections, DG61X proteins do not interact with assembly intermediates. However, in wild-type infections (Fig. 1A), these mutant proteins inhibit morphogenesis by removing 12S* intermediates into the insoluble fraction, thus displaying a dominant lethal phenotype (12). As the wild-type D protein forms dimers and higher-order oligomers in solution (11), these observations suggested that lethal dominant G61X proteins cannot self-oligomerize, a hypothesis directly tested in this study.
We previously showed that mutations conferring resistance to the presence of DG61X proteins map to the coat (F) and internal scaffolding (B) genes. Resistance was not specific to the DG61X mutant allele to which they were selected (5, 6). A strain with a robust resistance phenotype was experimentally evolved. Wild-type ϕX174 was cultured continually by serial passages in exponential-phase cells while the induction of the lethal dominant gene was increased incrementally. The first resistant strain isolated contained three mutations in the coat protein F (FD44H, FD205N, and FS227P), which were identical or similar to resistance mutations isolated in single-step selections. In general, all of these mutations cluster underneath the D3 subunit in the X-ray structure (12, 16). There was also a mutation in the DNA pilot protein H (HD136G). However, it is unclear whether this mutation contributes to the resistance phenotype. Mutations in this region of the DNA pilot protein are commonly observed during experimental evolution under various selective pressures (17–20). Thus, it may have arisen due to the propagation method. In this study, this strain is referred to as the “intermediary strain.” However, the “most evolved” strain achieved optimal fitness after acquiring an additional Asp→Gly mutation within the genome-encoded D protein (DD34G).
In the absence of the DG61D protein, the intermediary and evolved resistant strains demonstrated a decrease in fitness relative to the wild-type virus (Table 1 [data were adapted from reference 5), suggesting that the coat protein mutations may produce prudent less-reactive12S* intermediates that compensate for promiscuous, off-pathway reactions mediated by inhibitory D proteins. The coat protein mutations may allow 12S* particles to entirely exclude mutant dimers from the assembly pathway or from certain locations within the procapsid lattice. Intermediary-strain fitness was not significantly affected by the presence of the DG61D protein (Table 1); however, its presence greatly increased the most evolved strain's fitness (Table 1). The DD34G mutation does not confer a resistance phenotype in an otherwise wild-type background. Thus, the most evolved strain's elevated fitness appeared to be a function of the intragenic DD34G mutation in combination with the other mutations found in the intermediary strain.
Fitness values for the wild-type, intermediary, and evolved strains in the presence of each indicated D gene expressed in transa
Experimental evolution in vivo is a powerful tool, providing many genetic insights into possible adaptive mechanisms. However, details needed to test specific mechanistic hypotheses often cannot be determined without more detailed in vitro biochemical analyses. In this study, in vitro biophysical and biochemical methods were utilized to elucidate the mechanistic details and evolutionary trade-offs of resistance and also determine the extent to which in vivo fitness and in vitro assembly can be correlated.
RESULTS
The DG61D protein remains monomeric in solution.Earlier results suggested that the inhibitory DG61D protein alone cannot properly dimerize but will form inhibitory oligomers with wild-type subunits (11, 12). To investigate this further, purified mutant and wild-type D proteins (0.5 mg/ml) were examined by analytical ultracentrifugation (AUC). The sedimentation velocity profiles in Fig. 2 indicate that the DG61D and wild-type D proteins sediment at 2.1 S and 3.3 S, respectively. Each single peak was used to estimate the molecular mass for the corresponding sedimenting species. The estimated molecular mass of the DG61D protein was 17.6 kDa, whereas the value obtained for the wild-type D protein was 37.2 kDa. The theoretical molecular masses of a protein D monomer and dimer are 16.8 kDa and 33.6 kDa, respectively. These calculated weights are within 4.6% (monomer) and 9.7% (dimer) of the theoretical values, confirming that DG61D subunits cannot form homodimers.
Sedimentation velocity analysis of wild-type D (WT), DG61D, and a mixture of WT and DG61D proteins (1:1).
To determine whether DG61D and wild-type D proteins can form heterodimers, the purified components were mixed at a 1:1 ratio before AUC. This resulted in a single peak consistent with a dimer (Fig. 2), suggesting that homodimeric, wild-type D proteins disassociate and then reassociate with DG61D monomers after mixing. The absence of a monomer peak suggests that once WT/DG61D heterodimers form, they remain stable and do not easily disassociate (see Discussion). Wild-type homodimers and WT/DG61D heterodimers were analyzed by circular dichroism. In agreement with the X-ray structure (13–15), the wild-type D protein homodimer was largely helical. The heterodimers displayed a slight loss of helicity (data not shown). The DD34G protein, which was previously shown to elevate the most evolved strain's fitness, was also analyzed for the ability to form homodimers and heterodimers with the DG61D protein (11, 12). It behaved like the wild-type species in these assays (data not shown). Thus, the DD34G mutation did not appear to alter oligomerization dynamics.
The mutant DG61D protein does not interact with wild-type 12S* early assembly intermediates in vitro.During in vivo morphogenesis (Fig. 1A), the DG61D protein does not complement a nullD strain but inhibits wild-type morphogenesis, thus displaying a dominant lethal phenotype (12). To further characterize this phenomenon, the previously developed in vitro assembly system (11) was adapted to analyze assembly kinetics. In these experiments, reagents were mixed in the presence of polyethylene glycol, and procapsid assembly was monitored by light scatter at 340 nm as a function of time.
A mixture of wild-type 12S* particles and wild-type D protein produced a rapid increase in detectable light scatter (Fig. 3A), a result consistent with the nucleation and elongation phases of procapsid assembly (21). However, the wild-type D protein alone also produced significant scatter, indicating the formation of larger macromolecular structures. Therefore, samples were examined by electron microscopy. Reaction mixtures containing both wild-type 12S* particles and wild-type D protein produced procapsids with uniform dimensions similar to those of native procapsids (Fig. 3B), whereas the wild-type D protein alone formed a heterogeneous array of protein lattices (Fig. 3C). Similar heterogeneous structures have been reported at wild-type D protein concentrations higher than 4.0 mg/ml (22), conditions likely mimicked by the crowding agent in our assays. Unlike the reactions performed with the wild-type D protein, samples containing only 12S* particles, only the DG61D protein, or a mixture of these two components produced little or no detectable light scatter, indicating the absence of significant macromolecular assembly. Thus, the in vitro results are consistent with previous in vivo findings.
(A) In vitro assembly kinetics of purified protein components. The components of each reaction are color coded as indicated in the figure. (B) Electron micrograph of the in vitro reaction with WT 12S* plus WT D protein, producing procapsids. (C) Electron micrograph of the WT D protein alone, forming heterogeneous D protein lattice structures lacking internal density.
In vitro analysis of assembly kinetics defines the nature of an evolutionary trade-off.As previously described, serial passages of wild-type ϕX174 in cells expressing the dominant lethal DG61DD gene first produced an intermediary strain with a resistant phenotype. That strain contained three substitutions within the viral coat protein F (6). After resistance was established, a mutation in gene D arose in the most evolved strain that stimulated fitness (virion doublings per hour) to values as high or higher than that of the uninhibited wild-type strain (Table 1). Although resistant to the inhibitory scaffolding protein, the intermediary strain displayed significantly reduced fitness in the absence of DG61D expression compared to the wild type, indicating that an evolutionary trade-off had occurred (11, 12). To elucidate the nature of the trade-off, the rate of in vitro assembly with purified wild-type and evolved-strain 12S* particles in the presence of the wild-type D protein was examined (Fig. 4A). Compared to the wild-type 12S* reaction, the mixture of evolved-strain 12S* particles and wild-type D protein displayed a delayed start and reduced light scattering rate. The delay may indicate that the selection of a resistant coat protein coselected for a less reactive 12S* intermediate in the presence of only wild-type D protein.
(A) In vitro kinetics of procapsid assembly with wild-type or evolved strain 12S* particles. (B) In vitro kinetics of procapsid assembly with wild-type 12S* particles supplemented with wild-type D and DG61D proteins. (C) In vitro kinetics of procapsid assembly with evolved strain 12S* particles supplemented with wild-type D and DG61D proteins.
In vitro analysis of assembly kinetics and liquid chromatography-mass spectrometry distinguishes between two resistance mechanisms.Two simple contrasting models can explain the resistance mechanism. The coat protein mutations may (i) exclude inhibitory D protein dimers from the assembly pathway or (ii) allow the inhibitory dimers to be productively utilized. To distinguish between these hypotheses, in vitro kinetic reactions were carried out with either the wild-type or evolved-strain 12S* particles in the presence of various wild-type and mutant DG61D protein ratios. As expected, the presence of the DG61D protein lowered procapsid assembly efficiency in the wild-type 12S* reaction (Fig. 4B). Overall, lower WT/DG61D ratios correlated with less efficient assembly reactions. In contrast, the evolved-strain 12S* particles assembled into procapsids more rapidly in the presence of the DG61D protein (Fig. 4C). Moreover, procapsid assembly appeared to be most efficient when the WT and DG61D proteins were present in a 3:1 ratio. The reaction products were examined by electron microscopy, which verified the presence of procapsids (data not shown). The evolved 12S* particle's increase in assembly efficiency in the presence of the DG61D protein favors the model in which the virus evolved to productively utilize the inhibitory protein.
To further verify this conclusion, liquid chromatography-mass spectrometry was used to determine whether the DG61D protein was incorporated into procapsids. Procapsids were generated in lysis-resistant cells infected with the ϕX174 intermediary and evolved strains. The DG61D protein was supplied by a cloned IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible gene. To facilitate isolation and increase yields, an amber mutation was placed in gene C to block DNA packaging and the procapsid-to-virion transition. Two different induction conditions were employed: the higher-induction conditions yields a WT/DG61D D protein ratio of approximately 1:1, whereas the lower-induction conditions produce a 3:1 ratio (12).
Figure 5A and B display the calculated masses for the wild-type and DG61D protein peaks of the D proteins found in the procapsids. These values are consistent with their respective theoretical masses: 16,806 (WT) and 16,864 Da (DG61D). There were two D protein peaks found in intermediary-strain procapsids (Fig. 5A and B), demonstrating that both D proteins were incorporated into procapsids generated in vivo under low- or high-induction conditions.
Mass spectrometry verification and calculated masses of the D proteins that were incorporated into in vivo-generated procapsids for the intermediary (A and B) and evolved (C and D) strains. During infections, the DG61D cloned gene was induced to produce either low (A and C) or high (B and D) intracellular concentrations of the mutant D protein.
DD34G and DG61D together increase the in vitro production rate and overall yield of procapsids using the evolved 12S* particle.While the mutations in the viral coat protein contribute to the resistance phenotype, they do not dramatically elevate in vivo fitness to uninhibited wild-type levels (5, 6). Instead, the genome-encoded DD34G appears to stimulate fitness, but only in the presence of the dominant lethal DG61D protein and the resistant coat protein mutations. To determine if the DD34G mutation affected assembly kinetics, in vitro reactions were conducted with wild-type and resistant 12S* intermediates and various D protein combinations.
Figure 6A shows the light scatter profiles for the in vitro assembly reactions over a 20-min period. As expected, the wild-type 12S* reaction mixture supplemented with wild-type D protein produced the highest rate and highest yield of scatter. Reaction mixtures supplemented with the DD34G protein alone or in 1:1 and 3:1 combinations of DD34G to DG61D proteins displayed lower rates of increased light scatter and lower yields. The lowest yield correlated to the highest ratio of DD34G to DG61D D proteins.
In vitro kinetics of procapsid assembly with wild-type, DD34G, and DG61D proteins. (A) WT 12S* particles; (B) Evolved strain 12S* particles. The external scaffolding proteins used in these experiments are color coded as indicated in panel A.
In contrast, the resistant 12S* particles exhibited higher yields in assembly reaction mixtures that were supplemented with a combination of the DD34G and DG61D proteins (Fig. 6B). The rate of light scatter was detectably higher when these proteins were present in a respective 3:1 ratio. The wild-type D protein no longer promoted the most efficient reaction as was seen with wild-type 12S* particles (Fig. 6A). These results suggest that the evolved strain adapted a mechanism to utilize the DD34G/DG61D protein heterodimers to increase the efficiency of procapsid assembly. In vivo characterization of the evolved strain indicates that optimal fitness was achieved at a relative DD34G-to-DG61D protein ratio of 3:1. The in vitro results are in agreement with this observation.
The combination of DD34G with the DG61D proteins has a kinetic effect on assembly. However, it may also increase the amount of DG61D incorporated into evolved-strain procapsids. To investigate this hypothesis, procapsids containing the DD34G and DG61D proteins were generated in vivo as described above and examined by liquid chromatography-mass spectrometry. Figure 5C and D represent the results obtained with the evolved-strain procapsids under low and high DG61D gene induction conditions. Based on these spectra, both of the mutant D proteins, DD34G and DG61D, are also incorporated into procapsids, and the calculated masses agree with each protein's theoretical mass (16,748 and 16,864 Da, respectively). However, the relative amount of DG61D protein incorporated into in vivo-generated procapsids appears to be independent of the intracellular concentration of the DG61D protein.
The inhibitory DG61D protein appears to be incorporated into procapsids in relatively similar amounts regardless of the presence of the stimulatory DD34G protein. These results indicate that the DD34G protein does not dramatically increase the number of incorporated DG61D proteins and suggest that the stimulatory effects of the DD34G protein are exclusively kinetic.
DISCUSSION
As described in the introduction, lethal dominant ϕX174 external scaffolding D proteins have been characterized, and a resistant strain was experimentally evolved in vivo (5, 6, 12, 16). The results indicate that the inhibitory and compensatory mutations target a very specific morphogenetic step: the 12S*-to-procapsid transition. The ability to reconstitute this reaction in vitro (11) allows mechanistic hypotheses to be directly tested and affords the unique opportunity to qualitatively relate in vivo fitness to in vitro assembly efficiency.
External scaffolding protein oligomerization dynamics and the mechanism of inhibition.As depicted in Fig. 1C, there are four structurally unique D proteins, D1 to D4, per coat protein. In subunits D1 and D3, a 30° kink in α-helix 3 occurs at glycine residue 61 (G61), whereas this helix is straight in the other two subunits. The X-ray structure predicts that proteins unable to switch between kinked and straight conformations should form long helical oligomers (13), which would not interact with other viral proteins. Although this prediction explains the lack of in vivo DG61D protein activity in the absence of the wild-type subunit (12), the results of AUC analysis indicate that the prediction is incorrect. The DG61D protein cannot self-dimerize, a requirement for morphogenesis. The DG61D mutation would inhibit the kinking of α-helix 3. Amino acid residues at the end of the helix participate in dimer formation (13–15). The altered orientation of the helix may inhibit those interactions. The inability to form homodimers is also consistent with the mutant protein's lack of activity in the absence of wild-type subunits.
The DG61D protein did dimerize with the wild-type protein (Fig. 2), which likely allows it to enter the assembly pathway. When mixed at a 1:1 ratio with the wild-type protein, only a single peak was observed after AUC. This suggests that the Kd (dissociation constant) of the wild-type homodimer is greater than the Kd of the heterodimer at equilibrium. If the heterodimers displayed a Kd greater than or equal to that of wild-type dimers, a dimer and monomer peak would have been detected, which was observed when the DG61D protein was in excess (data not shown).
The mechanism of resistance and its inherent evolutionary trade-off.Procapsid assembly of wild-type and evolved strain 12S* particles was analyzed in vitro with wild-type D protein. Although procapsids were observed in each sample by electron microscopy, the coat protein resistance mutations created a less reactive 12S* particle. The mutant particles exhibited a lower in vitro assembly rate and a decreased yield, which may partially explain the fitness loss in the absence of the inhibitory protein (Table 1).
Assembly was more efficient in the presence of the DG61D protein than in its absence, indicating that the resistant 12S* particles were becoming specialized to assemble with heterodimers. This phenomenon was not previously detected by in vivo fitness assays (Table 1). The single-targeted focus of the in vitro analysis may have elucidated a detail obscured by more complex and encompassing in vivo assays (see below).
Mechanism of stimulation.During experimental evolution, the coat protein mutations conferring resistance appeared first. The last mutation to sweep through the population was DD34G in the genome-carried D gene. This single mutation almost doubled fitness in the presence of the DG61D protein (Table 1) and likely specializes assembly with heterodimers. In its absence, fitness was not affected but remained lower than wild type. Resistant 12S* particle assembly kinetics were followed in vitro using various combinations of DD34G, DG61D, and wild-type D proteins. Assembly was most efficient and gave the highest yields when reactions mixtures contained a 3:1 DD34G-to-DG61D protein ratio, but it was not possible to attribute this to higher nucleation or elongation rates. Faster nucleation is most reliably detected with a decreased lag phase, which was not observed. The curve's initial slope was steeper right after scatter was detected (data not shown). While this may reflect faster nucleation, faster elongation and higher yields can also affect this region of the graph. The latter may indicate that lower critical concentrations can nucleate assembly, but the presented experiments do not rigorously address this point.
To determine whether the DD34G mutation also increased the amount of DG61D protein allowed into the assembly pathway, procapsids were characterized by liquid chromatography-mass spectrometry. The DD34G mutation did not appear to increase the amount of DG61D protein found in the procapsid. Thus, the effect of the DD34G mutation appears to be exclusively kinetic. This observation may indicate that the system has evolved an upper limit on the amount of DG61D protein that can be incorporated. Regardless of the induction conditions used to generate procapsids, the amount of incorporated DG61D protein appeared to be constant, approximating one-fourth of the total D protein. This may suggest that the heterodimer is occupying a unique position in the D protein lattice.
Correlating in vivo fitness with in vitro analyses.In general, in vitro assembly efficiency qualitatively correlated with in vivo fitness. However, there were some disparities, and the two phenomena could not be quantitatively related. The in vitro assay examines a very specific morphogenetic step, whereas in vivo fitness measures the entire infection process. In vitro assembly reactions begin with substrate levels well above the effective concentrations to nucleate assembly. In vivo, the reactants must accumulate to their critical concentrations to nucleate assembly before programmed cell lysis, which occurs independently of progeny formation (23, 24). Several factors could influence when these critical concentrations are achieved in an infected cell. Mutant proteins may require higher critical concentrations and/or be more prone to interact with host cell proteins not present in vitro. In either instance, the timing of in vivo nucleation could be delayed, which would result in lower fitness. Moreover, procapsids must be filled after formation, which requires the external scaffolding lattice to interact with ssDNA synthesizing and packaging machinery. Inefficient genome synthesis and packaging may account for some disparities. For example, the DD34G mutant in an otherwise wild-type background exhibits fitness values significantly less than wild type (5), yet the protein efficiently assembles procapsids in vitro (data not shown).
The redistributed biophysical roles in a two-scaffolding-protein system.The microvirus internal scaffolding protein shares many properties and functions with the scaffolding proteins found in other viruses (7, 9, 10, 14, 15, 25–41). However, it does not physically construct the procapsid or control morphogenetic fidelity. These are functions of the external scaffolding protein. Both in vitro and in vivo, the external scaffolding protein behaves more like the viral coat protein in other systems, whereas the 12S*-containing coat protein behaves more like a scaffolding protein. For example, coat proteins from other systems form aberrant assemblies in the absence of scaffolding protein, whereas scaffolding proteins alone are relatively inert. When mixed, the scaffolding proteins control fidelity, suppressing the formation of heterogeneous aberrant structures containing only the coat protein (21, 33, 42–50). In contrast, the ϕX174 external scaffolding protein D self-associates to form large heterogeneous spherical complexes (Fig. 2C), whereas coat proteins in the 12S* particles do not self-associate in vitro, in vivo, or in the assembled procapsid (11, 12, 51). Similarly, the presence of the 12S* particles in reaction mixtures prevents the formation of complexes containing only the external scaffolding coat protein.
MATERIALS AND METHODS
Phage plating, media, buffers, and stock preparation.Media, buffers, plating, and stock preparation have been previously described (52).
Escherichia coli cell lines, ϕX174 strains, and plasmids.The ϕX174 strains and plasmids used in these study have been previously described (5, 6, 11, 12, 53). The am(C)S10 mutation, which facilitated procapsid isolation, was introduced by site-directed mutagenesis (54). Strains were propagated in BAF7 supD (52). The ϕX174 Δ9D fs440 strain (53) contains a deletion of the first nine D gene codons, nucleotides 396 to 419 in the genome sequence, and a frameshift mutation, a deletion of nucleotide 440. It was used to generate wild-type 12S* particles. The Δ9 mutation was engineered into the evolved ϕX174 strain as previously described (53). the Δ9D fs440 and Δ9D strains were grown in BAF30 pDNco (11). The C900 E. coli strain used to generate 12S* particles and procapsids contains the slyD mutation, which confers resistance to E protein-mediated lysis (23).
12S* particle and D protein purification.The protocols for 12S* particle and D protein purification have been previously described (11).
AUC and circular dichroism.Sedimentation velocity experiments were performed at 22°C with a D protein concentration of 0.5 mg/ml in 20 mM Tris-HCl, pH 8. The sedimentation at 50,000 rpm in a An-60 Ti rotor was followed by the absorbance at 280 nm over time with a Beckman Optima XL-A analytical ultracentrifuge (AUC) (International Mi-Ss, Inc., Corona, CA). The partial specific volume was calculated by amino acid sequence, and the solvent density was calculated according to the solvent composition using the Sednterp program found at http://bitcwiki.sr.unh.edu/index.php/Main_Page (55). The sedimentation coefficient distributions were calculated with a continuous c(s) distribution model of Lamm equation solution implemented in the SEDFIT program found at http://www.analyticalultracentrifugation.com/default.htm (56).
Circular dichroism spectra of wild-type and mutant scaffolding proteins were obtained on an AVIV model 62DS spectropolarimeter (Lakewood, NJ) equipped with a single-cell thermoelectric device. Data were collected on 1.0-mg/ml protein solutions at 13°C, from 250 nm to 190 nm in 1.0-nm increments, with a 16-s averaging time per point. The spectrum for baseline correction was identical to the sample solution less protein. Raw data, in millidegrees, were baseline corrected, smoothed, and multiplied by a scaling factor to obtain spectra in units of mean residue ellipticity.
Generation and purification of procapsids synthesized in vivo.To isolate in vivo-generated procapsids, 1.0 liters of E. coli C900 (slyD) was grown to a concentration of 1.0 × 108 cell/ml at 37°C. Cells were infected, as described above, with either the intermediary or evolved am(C)S10 strain at a multiplicity of infection (MOI) of 3.5. Protein C is essential for DNA packaging. Thus, procapsids accumulate in infected cells (57). Particles were isolated and purified as previously described (58).
Analysis of in vivo-generated procapsids by mass spectrometry. In vivo-generated procapsids were purified as previously described and were injected onto a Phenomenex Aeris 3.6-μm SB-C18 reverse-phase column in 5% acetonitrile and 0.1% formic acid. Species were eluted from the column with a linear gradient from 5% to 95% acetonitrile into a Q-ToF Premier mass spectrometer. Species were identified by mass.
In vitro assembly reactions and subsequent analyses.The 12S* particles and D proteins were mixed with buffer (100 mM Tris-HCl [pH 7.5], 5 mM NaCl, 20 mM MgCl2) at a 1:12 respective molar ratio in 400-μl reaction mixtures at room temperature. The final concentrations of the buffering components were kept constant at 50 mM Tris-HCl, 2.5 mM NaCl, 9.0 mM MgCl2, 20 mM NaCl, and 10% polyethylene glycol (PEG) 3350 as previously described (11). Reactions were followed every 20 s for at least 30 min at a 340-nm wavelength through the use of a Beckman 640 spectrophotometer.
Electron microscopy.The protein samples were diluted into dialysis buffer containing 20 mM Tris-HCl (pH 7.8), 50 mM NaCl, 1 mM MgSO4, 1 mM CaCl2, applied to glow-discharged 400-mesh carbon-only grids (Electron Microscopy Sciences), washed with 2 drops of dialysis buffer, and negatively stained with 1% uranyl acetate. The grids were observed in an FEI Tecnai F20 electron microscope operated at 200 kV and imaged with a Gatan Ultrascan 4000 charge-coupled device (CCD) camera at a magnification of ×65,500 or ×81,200.
ACKNOWLEDGMENTS
This research was supported by National Science Foundation grant MCB 1408217 to B.A.F. and an R.O.A. supplement to J.E.C.
We thank Sarah M. Doore and Aaron P. Roznowski for discussion.
FOOTNOTES
- Received 16 September 2016.
- Accepted 18 October 2016.
- Accepted manuscript posted online 26 October 2016.
- Copyright © 2016 American Society for Microbiology.
