Tuning Viral Capsid Nanoparticle Stability with Symmetrical

Aug 24, 2016 - The Dec protein naturally binds to the surface of the bacteriophage L capsid as trimers localized preferentially to the 60 quasi 3-fold...
0 downloads 10 Views 4MB Size
Tuning Viral Capsid Nanoparticle Stability with Symmetrical Morphogenesis Aida Llauró,†,‡ Benjamin Schwarz,§,‡ Ranjit Koliyatt,§ Pedro J. de Pablo,*,†,∥ and Trevor Douglas*,§ †

Departamento de Física de la Materia Condensada and ∥Condensed Matter Physics Center IFIMAC, Universidad Autónoma de Madrid, 28049 Madrid, Spain § Department of Chemistry, Indiana University, Bloomington, Indiana 47405, United States S Supporting Information *

ABSTRACT: Virus-like particles (VLPs) provide engineering platforms for the design and implementation of proteinbased nanostructures. These capsids are comprised of protein subunits whose precise arrangement and mutual interactions determine their stability, responsiveness to destabilizing environments, and ability to undergo morphological transitions. The precise interplay between subunit contacts and the overall stability of the bulk capsid population remains poorly resolved. Approaching this relationship requires a combination of techniques capable of accessing nanoscale properties, such as the mechanics of individual capsids, and bulk biochemical procedures capable of interrogating the stability of the VLP ensemble. To establish such connection, a VLP system is required where the subunit interactions can be manipulated in a controlled fashion. The P22 VLP is a promising platform for the design of nanomaterials and understanding how nanomanipulation of the particle affects bulk behavior. By contrasting single-particle atomic force microscopy and bulk chemical perturbations, we have related symmetry-specific anisotropic mechanical properties to the bulk ensemble behavior of the VLPs. Our results show that the expulsion of pentons at the vertices of the VLP induces a concomitant chemical and mechanical destabilization of the capsid and implicates the capsid edges as the points of mechanical fracture. Subsequent binding of a decoration protein at these critical edge regions restores both chemical and mechanical stability. The agreement between our single molecule and bulk techniques suggests that the same structural determinants govern both destabilizing and restorative mechanisms, unveiling a phenomenological coupling between the chemical and mechanical behavior of self-assembled cages and laying a framework for the analysis and manipulation of other VLPs and symmetric self-assembled structures. KEYWORDS: virus-like particles, atomic force microscopy, self-assembly, capsid stability, protein cages

V

mechanical effects of these structural changes in single particles and connecting those effects with population stability assays would provide insights into predictable mechanical design of VLPs, the use of VLPs as delivery systems, and better fundamental understanding of self-assembly and disassembly processes. To establish a relationship between precise modifications to a single particle and how the VLPs behave as a bulk population, a VLP system is needed in which precise repeatable structural changes can be introduced. The VLP derived from the bacteriophage P22 is a highly promising candidate for the design of nanomaterial systems with applications in catalysis, biomaterials, and biomedicine.3,8 In the P22 VLP, pores can be introduced selectively resulting in the highly regular interruption of subunit interactions that are anticipated to weaken

irus-like particles (VLPs) are noninfectious cage-like architectures derived from viral or nonviral sources that provide nanoscale platforms for the design and engineering of particles with functionalized exteriors and/or interiors.1,2 This potential for modification makes VLPs promising candidates for the design of a new class of nanomaterials with applications ranging from biomedicine to energy to electronics.3,4 Among other factors, the properties and related functionalities of these nanostructures originate from their symmetric assembly from a limited number of subunits.5 Selective interruption or reinforcement of the interactions between these subunits could lead to controllable changes in the overall stability of the capsid.6 In the assembly of infectious viral capsids, deviations from perfect symmetry, which result in missing subunit interactions, are anticipated to lead to changes in the performance of the virus.7 The disruption of symmetric subunit interactions and the subsequent effects these may have on VLP stability have not been examined because such an examination requires precise and repeatable control over subunit interactions. Examining the © 2016 American Chemical Society

Received: May 24, 2016 Accepted: August 24, 2016 Published: August 24, 2016 8465

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

www.acsnano.org

Article

ACS Nano the particle. P22 also permits the reinforcement of subunit contacts through the binding of a decoration (Dec) protein to the exterior capsid surface. P22 procapsid (PC) VLPs consist of 420 copies of a coat protein (CP) that assemble into a T = 7 icosahedral structure encapsulating 100−330 scaffolding proteins (SP). P22 VLPs can be produced heterologously through the co-expression of CP and SP, or they can be assembled in vitro through the combination of these two proteins.9 In the infectious phage, DNA packaging triggers the expansion of the immature PC form of P22 into the more angular mature phage.10−16 This structural transition from PC to the mature phage can be mimicked in the VLP system either by heating the sample to 65 °C or by briefly incubating the sample with low concentrations of sodium dodecyl sulfate (SDS), which results in an expanded (EX) VLP morphology (Figure 1A).17,18 Further heating to 75 °C leads to the release

VLP systems compared to the increased thermal tolerance evident in the infectious phage. In this system the heat-induced penton expulsion to form the WB morphology and the binding of Dec allow for the potential tuning of subunit interactions at symmetry specific sites on P22 VLPs. The stiffness and strength of nanoparticles are important characteristics that dictate how the particles behave within a biological environment.24 In addition, the P22-Dec system is a promising platform for the delivery and exterior presentation of whole proteins for biomedical applications and for the design of protein-based framework materials.8,25,26 The effects that both WB formation and Dec binding to P22 have on VLP stability are therefore of interest to better understand the potential of this system as a general purpose nanoplatform. Here we study the link between the mechanics of single capsids and the bulk chemical stability of four different forms of P22 VLPs (Figure 1D). Our results show that the selective interruption or reinforcement of intersubunit interactions results in a simultaneous modulation of both mechanical and chemical stability. Our single-molecule experiments reveal a symmetry-dependent variation of the mechanical properties and suggest that the capsid edges are responsible for breakage in the penton-less WB structures. Interestingly, the magnitude of this relationship would have been missed had the particles not been examined with the symmetry of the cage in mind. An examination of predicted intersubunit interaction energies based on buried surface area leads to a predictable explanation of this symmetry-specific stability modulation and allows for the extension of these concepts to other symmetric self-assembly structures. The coupling of symmetry-specific modulation of subunit interactions with bulk chemical and spatially resolved mechanical examination provides insights into the translation of single-particle mechanics into ensemble particle behavior and provides a promising framework for approaching the examination of symmetric nanostructures in general.

Figure 1. Structure of various P22 morphologies. Cryo-EM reconstruction of (A) EX capsid (PDB: 3IYI)21 and (B) WB (PDB: 3IYH)21 (C) Dec protein bound to an EX capsid.23 An icosahedron is overlaid on the structure (red dashed lines). (D) Schematic of the transition of the VLPs in our study. EX VLPs were heated to expel pentons forming the WB morphology. Dec proteins were bound to either the EX capsids or the WB capsids at a ratio of 160 Dec trimers per capsid.

RESULTS AND DISCUSSION To explore the connection between symmetry-specific manipulation of the P22 VLP and bulk behavior of the VLP population, we examined the single-particle mechanical properties and bulk chemical stability of four different forms of P22 (Figure 1D). We first focused on the effects of penton release by comparing the mechanics and chemical stability of EX and WB. Second, a Dec protein was bound to both morphologies, and the changes due to Dec binding were compared (i.e., EX vs EX + Dec and WB vs WB + Dec). The Capsid Is Destabilized by the Loss of Pentons. P22 VLPs were produced in E. coli and purified as PCs with scaffold protein retained on the interior. Capsid samples were heated to form either EX or WB from a common PC sample, and expansion was monitored via nondenaturing agarose gel electrophoresis (Figure S1A). Samples were also assessed by size-exclusion chromatography (SEC) monitored by multiangle light scattering (MALS) and quasi-elastic light scattering (QELS). By SEC the EX and WB samples showed a shift in retention time, as compared to PC, suggesting larger capsids (Figure S1B). This was further supported by QELS, which reported a radius of hydration of 29 ± 1 nm for EX and 28 ± 1 nm for WB compared to 26.5 ± 0.8 nm for PC. By MALS, the number-average molecular weight of EX and WB was reduced compared to PC (Figure S1C,D). The measured molecular weight for EX of 20.3 ± 0.7 MDa was in good agreement with the expected weight of 19.7 MDa for a capsid completely

of the 12 pentons of the capsid resulting in the porous and less angular wiffle ball (WB) VLP morphology (Figure 1B).19−21 The mechanical stress of a icosahedron is concentrated at the vertices, and selective expulsion of the pentons in WB capsids likely relieves this stress.22 The Dec protein naturally binds to the surface of the bacteriophage L capsid as trimers localized preferentially to the 60 quasi 3-fold sites.17,20,23 Bacteriophage L is highly similar to P22 with only four conserved amino acid differences in the CP sequences. It has been shown that Dec binding to the mature P22 virus increases the heat tolerance of the infectious phage.17 Cryo-electron microscopy studies revealed that Dec also binds to the EX and WB P22 VLPs localizing to the quasi 3-fold axes with high affinity and with lower affinity to sites at the true 3fold axes (Figure 1C).20,23 While Dec binding is still anticipated to increase the strength of P22 VLP particles, the absence of a packaged genome may lead to significantly different behavior in 8466

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano devoid of SP. The molecular weight of WB of 18.3 ± 0.5 MDa indicated the loss of the majority of the pentons. However, the expected molecular weight for capsids with all 12 pentons removed was 16.9 MDa. This difference between the measured and expected mass likely indicates that the WB capsids retain a few pentons. Single-molecule atomic force microscopy (AFM) experiments confirmed that some of the WB capsids retain pentons at certain vertices (Figure S2). The primary interest of this study was to establish a connection between structural manipulation of single particles and the mechanics and stability of large ensembles of capsids in chemically aggressive environments. As such an analysis was needed that readily reported the integrity of the capsids in solution. The stability of a capsid sample can be measured in bulk by monitoring changes in static light scattering, which correlates with the assembly state of the VLPs in solution. Alternatively, the morphology and stability of P22 can be monitored by nondenaturing agarose gel electrophoresis. To assess the chemical stability of the P22 VLPs in bulk and resolve differences between the stability of EX and WB, a condition was needed that would disassemble the capsid slowly in a concentration-dependent manner. Guanidine hydrochloride (3M) disrupted the VLP assemblies but did not resolve any stability differences between morphologies (Figure S3A), while incubation in pH 4.5 at 65 °C (Figure S3B) or in 8 M urea (data not shown) had minimal impact on EX or EX with Dec bound over long time periods. Surprisingly both isothermal incubation at 75 °C and controlled temperature ramping from 40 to 95 °C revealed no resolvable difference between EX and EX with Dec bound (Figure S3C,D). This result is in contrast with previous reports that convincingly show that Dec binding to the infectious P22 virus leads to more retained infectivity after isothermal incubation at 50 °C in the presence of EDTA.17 These results suggest that the VLP system responds in a markedly different manner than the infectious virus when Dec is bound likely due to the absence of genetic cargo. Eventually, incubation of capsid samples in SDS was shown to degrade the capsid in both a time- and concentrationdependent manner and was selected as an appropriate stressor for comparing bulk stabilities of the VLP morphologies (Figure S4). While 0.5% SDS was chosen as the best concentration for resolving differences in stability between morphologies, similar trends were also observed at 0.2 and 0.1% SDS (Figure S4, data not shown). VLP samples were mixed rapidly with SDS, and the scattering at 320 nm was monitored. When the normalized loss in light scattering for EX and WB was compared, WB exhibited much lower stability with a 56% average reduction in scattering after 1 h incubation compared to only 31% for EX over the same time period (Figure 2A). The degradation of VLPs was also monitored with transmission electron microscopy (TEM) by incubating VLP samples with 0.5% SDS for 20 min and then depositing the sample on an EM grid. The percentage of intact capsids was counted across multiple fields of view. WB samples exhibited a decrease of 63% in the percentage of intact capsids, while EX samples remained largely intact with a decrease of 10% in the percentage of intact capsids (Figures 2B,C and S5). To examine the mechanics of the EX to WB transition at the single-particle level, we used AFM. VLPs were adsorbed on freshly cleaved HOPG (highly oriented pyrolytic graphite) surfaces, and their orientation was determined (Figure 2D). The 5-fold symmetry (S5) was most frequent oriented

Figure 2. The P22 VLP is destabilized by the loss of pentons. (A) Normalized loss in light scattering at 320 nm due to degradation of either EX (blue) or WB (red) in 0.5% SDS. Error bars represent the standard error on the mean of the normalized scattering at each time point for 3 runs. Electron micrographs of (B) WB (left) and WB after 20 min in 0.5% SDS (right) or (C) EX (left) and EX after 20 min in 0.5% SDS (right). Insets in the upper right corner show an expanded view of a representative particle from the field. Scale bars: 500 nm. (D) AFM images of single VLPs. Top row: S5 oriented EX (left) and WB (right). Bottom row: S3/S2 oriented EX (left) and WB (right). (E) AFM average height measurements of 30 EX and 29 WB capsids with either the 5- or 3/2-fold symmetry axes oriented upward. S3 and S2 axes were grouped because the orientations could not be reliably resolved. Error bars represent standard error on the mean.

perpendicular to the surface, representing 59% and 66% of EX and WB samples, respectively (Table S1). The other two orientations were classified together as 3- and 2-fold symmetries (S3/S2) because the precise position was difficult to resolve. The average height of the VLPs measured by AFM depended on the orientation. VLPs in the S5 orientation were higher than VLPs showing a S3/S2 orientation (Figure 2E). These differences in height were in agreement with the fact that in an icosahedron, the distance between two opposite vertices is larger than the distance between two opposite faces or edges. In both orientations WB capsids presented a lower height than the corresponding EX forms, which indicated that the loss of pentons had a significant effect on the deformability of the capsids after adsorption. Recent AFM experiments have shown that a lower height due to deformation of the capsid on the surface usually correlates with a softening of the overall structure.27 To further investigate this change in the mechanical properties with penton loss, we performed nanoindentation experiments on single VLPs. Nanoindentation assays have been extensively used to study the mechanical properties of viral and nonviral nanoparticles during the past decade, though only 8467

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano

Figure 3. Individual AFM nanoindentation assay. (A) AFM images of an EX capsid showing a S5 orientation before (left) and after (right) the nanoindentation. A pentagon with its five triangular faces is overlapped to the image to guide the eye (dashed lines). The red lines highlight the parts of the particle that were mostly affected by the nanoindentation. (B) Schematic at scale of the AFM tip (red sphere) indenting an EX capsid along the 5S axis. (C) AFM images of a WB capsid showing a S3 orientation before (left) and after (right) a nanoindentation. White arrows indicate penton vacancies. Dashed circles indicate the position of the missing pentons after breakage (the pentameric sites that become connected after the breakage are represented in red). (D) Schematic at scale of the AFM tip (red sphere) indenting an EX capsid30 along the 3S axis. Black dots are located at the pentons. The contact of the AFM tip can be seen to be localized to the capsid face. (E) Nanoindentation curves that induced the breakage in A (blue line) and C (red line). (F) Comparison of the elastic constant (K) and breaking force of 30 EX and 39 WB capsids. The error bar represents the standard error.

local rigidity in CCMV assemblies. 39 This mechanical sensitivity to the molecular interactions was also captured in this work when compared to previous studies examining bacteriophage T7, also a T = 7 virus of similar size.40 The expanded capsid form of P22 is isotropic with respect to stiffness (i.e., the same spring constant values were equivalent along the three different symmetry axes: KEX‑S5 ≈ KEX‑S3/S2 ≈ 0.20 N/m). In contrast, bacteriophage T7 capsids were found to be highly anisotropic (the 3S axis was two times stiffer than the 5S axis).38 In contrast to stiffness, our measurements on the breaking force in P22 show highly anisotropic variation (Figure 3F, top). The loss of pentons had an enormous impact on the mechanics of S3/S2 symmetry (green line) but did not affect the S5 direction (black line). This anisotropic decrease in breaking force of the selectively destabilized VLPs indicated that the removal of pentons had a very specific symmetry-dependent effect on capsid mechanics. The lack of pentons in WB reduced the strength of the capsids more dramatically along the S3/S2 axes, most likely because the hexamers (of the top face or edge) in contact with the AFM tip during deformation had lost their interaction with their neighboring pentons.21 Further loss of contacts is predicted from the intercapsomeric association energies provided by the VIPER database based on the buried surface area and estimated solvation energies present in available cryo-EM reconstructions of EX (3IYI) and WB (3IYH). In addition to the loss of penton−hexon connections, WB loses three predicted hexon−hexon interactions: two connections along the edge and one at the facet (see blue dots in Figure S6 and Table S3). Interestingly, most of WB deformations induced fractures along the edges connecting two pentameric vacancies (Figures 3C and S6), occasionally resulting in the collapse of an entire face. Effect of Dec Protein on P22 Stability. Dec protein from bacteriophage L has been previously shown to enhance the heat tolerance of infectious P22.17 Although not part of the P22 viral genome, Dec binds to both the EX and WB P22 VLPs preferentially at the quasi 3-fold sites, which straddle the 2-fold axes of the capsid.20 Interestingly, for the WB morphology fractures occurred along the 2-fold edges (Figures 3C and S6). Therefore, we predicted that binding of Dec directly adjacent to these points could lead to recovery of capsid stability through the introduction of additional intersubunit contacts. To

sometimes is this analysis categorized by the orientation of the particle as was done here.28,29 Briefly, the experiment consists of deforming individual capsids with an AFM tip until mechanical failure is observed. The force vs indentation curve (FIC) obtained during this irreversible deformation provides information about the rigidity K (elastic constant) and the strength of the particle (breaking force and critical deformation).28 A representative set of AFM images and model schematics is shown in Figure 3. The example displays an EX VLP indented along the S5 axis (Figure 3A,B) and a WB VLP indented along the S3 axis (Figure 3C,D) before and after breakage. The corresponding FICs for these capsids are plotted together in Figure 3E. To quantify the mechanical parameters, 30 EX and 39 WB capsids were examined, and the results classified by symmetry (Figure 3F and Table S2). Measurements of the spring constant suggested that the loss of pentons was accompanied by a softening of the structure. A decrease in particle stiffness of 25% was observed between EX and WB along both symmetry axes (Figure 3F, bottom). These results were in good agreement with continuum elastic theory and Monte Carlo coarse-grained simulations predicting that pentons are the most stressed regions of the capsid and that their removal should make the capsid softer.22,31 Penton-less capsid forms exist across a wide size range of icosahedral viruses including adenovirus (T = 25), HSV1 (T = 16), HK97, and P22 bacteriophages (both T = 7).19,32−34 It is worth noting that known penton-less forms seem to be localized to T ≥ 7 VLPs, which agrees well with predictions that as the T number increases, more stress is localized to the 5fold sites.22 This general occurrence of penton-less capsids suggests that this state is a consequence of the overall mechanics of the icosahedral cage architecture. However, the symmetry-specific mechanical properties of a virus particle originate at molecular level interactions.35−38 For example, single-point mutations in MVM change the elastic response of the capsid more than 2-fold in a symmetry-dependent manner.6 In recent work by Cieplak et al., 35 virus capsids, with T numbers from 1 to 7, were tested in a molecular model showing that the change in mechanical properties did not correlate with the virus size or the T number but with the mean number of interactions between neighboring subunits.5 Hespenheide et al. also utilized intra and intersubunit interactions to predict the 8468

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano examine this possibility, the SDS-mediated degradations of EX and WB were compared to Dec bound samples of both VLP morphologies. While Dec reinforcement had little effect on the degradation of EX in 0.5% SDS, the degradation of WB was significantly reduced. WB + Dec showed a 20% reduction in the normalized scattering signal in 0.5% SDS over 1 h compared to a 56% reduction for WB alone (Figure 4A).

Figure 5. Mechanical changes due to Dec binding and AFM elastic energy required to break each configuration. (A) Evolution of the breaking force and elastic constant (K) for the transition EX → EX + Dec along the S5 axes (black) and S3/S2 axes (green). (B) Evolution of the breaking force and elastic constant (K) for the transition WB → WB + Dec along the S5 axes (black) and S3/S2 axes (green). (C) Example of a nanoindentaion experiment performed on a WB + Dec along the S3 axis. The AFM images before (top) and after (bottom) the rupture show that the top face was completely removed (white arrows indicate the penton-less vacancies). The area below the elastic FIC (gray area) represents the energy provided to the system to produce the breakage. (D) Statistical analysis of the different energies of rupture: along the S5 axes (left) and S3/S2 axes (right). The most significant decrease in energy was observed for the S3/S2 orientation after penton removal (one asterisk). Likewise, the most significant increase in energy was observed for the S3/S2 orientation after Dec binding (two asterisks).

Figure 4. Dec selectively reinforces WB VLPs. (A) Normalized loss in light scattering at 320 nm due to degradation of (top) WB (red) or WB + Dec (black) and (bottom) EX (red) or EX + Dec (black) in 0.5% SDS. Error bars represent the standard error on the mean of the normalized scattering at each time point for three runs. (B) Electron micrographs of WB (top) or WB + Dec (bottom) after 20 minutes in 0.5% SDS. Insets in the upper right corner show an expanded view of a representative particle from the field. Scale bars are 500 nm. (C) AFM height of capsids with either the S5 axis or the S3/S2 axes oriented upward. Error bars represent standard error on the mean. (D) Single particle AFM images of either the S5 (top row) or S3 (bottom row) orientations of EX + Dec (left) or WB + Dec (right).

in the elastic constant (K) were observed in WB populations after Dec binding, but a moderate increase of 13% was observed for EX + Dec along the S5 axes. Distinct differences were observed in the breaking force of all samples. The binding of Dec to EX VLPs was accompanied by a similar increase in the breaking force for the S5 (15%) and S3/S2 (12%) symmetry axes (Figure 5A). In contrast, WB VLPs showed a recovery of stability that depended significantly on the particle orientation. While the breaking force only increased by 6% along the S5 symmetry axes, an increase of 31% was observed for the S3/S2 symmetry axes (Figure 5B). Remarkably, this anisotropic reinforcement compensated for the observed decrease in the breaking force after the expulsion of pentons (Figure 3F, top). To collectively compare the multiple contributing mechanical parameters, we estimated the elastic energy transferred in order to fracture the capsids (gray area in Figure 5C). The energy required to break EX VLPs was similar to that of WB VLPs along the S5 axes, but this value dropped by 50% along the S3/S2 axes (Figure 5D, red bars). As expected, after Dec binding, the required energy increased for both morphologies, but the effect was not proportional. Whereas along the S5 axes, the effect of Dec was similar (25% vs 30%), and along the S3/

This recovery of stability was supported by TEM. As mentioned above, micrographs of WB treated with 0.5% SDS for 20 min displayed a 62% average reduction in the ratio of intact to disrupted capsids compared to untreated WB. With Dec bound, only an 11% reduction in the ratio of intact to damaged capsids was observed (Figures 4B and S5). AFM images showed that Dec bound WB capsids increased in average height (3% along the S5 axes and 5% along the S3/ S2) to a greater extent than EX capsids (0.2% and 2%, respectively) (Figure 4C), which is in agreement with TEM results. The greater increase in the height of WB with Dec binding was attributed to increased resistance of the capsid to deformation in addition to the height gained from Dec protruding from the capsid surface. The presence of Dec attached to the capsid surface was also evident in the less geometric appearance of the VLP structures and decreased resolution of individual subunits (Figure 4D). The effect of Dec binding was also explored by single-particle nanoindentation assays (Figure 5A,B). No significant changes 8469

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano

pansion and degradation will likely be isotropic about the capsid, these processes may be more pronounced at sites of greater mechanical stress. Our AFM results showed that for the WB morphology, fractures occurred along the edges, the second most stressed regions of the capsid after the pentons (Figure 3C and S7).22 An inspection of the intercapsomeric energies between subunits revealed that along those edges, interactions were lost after the transition, which explained the symmetry-dependent destabilization observed in our single-molecule experiments. Subsequent Dec binding at those critical points was able to restore the stability of the capsid in a symmetry-dependent way. We hypothesize that the significant recovery of chemical stability with Dec binding to WB VLPs may indicate that capsid rupture under these conditions also occurs at the intercapsomeric contacts along the edges, exposed during the transition from EX to WB, which would allow for more ready intercalation of SDS and faster degradation of the capsids.

S2 axes, the increase was more significant for the WB morphology (10% vs 50%) (Figure 5D). In summary, our energetic study demonstrated two things: (1) the formation of pores, and thus the removal of subunit contacts, had a symmetry-dependent effect, being more important for the S3/ S2 orientation (see one asterisk in Figure 5D), and (2) Dec binding has a similar symmetry-dependent effect in the recovery of stability (see two asterisks in Figure 5D). An analysis of the position of Dec protein in the structures sheds some light on these results (Figure S6). Dec proteins bind nearest the icosahedral 2-fold axes, reinforcing the edges of the capsid (see trimers in Figure S6). This binding site coincides with the positions where the intercapsomeric interactions were estimated to be lost after the transition from EX to WB (blue dots, Figure S6). It seems likely that in the same way that those interactions were responsible for the symmetry-dependent loss of stability, the re-establishment of connections at these sites through Dec binding results in a restoration of the mechanical strength. Auxiliary proteins, such as Dec, are a common strategy used by different viruses to stabilize their structures against the remarkable mechanical forces involved in the viral lifecycle.21 Recent publications have shown that binding of a Dec protein to a viral capsid enhances the mechanical stability of the capsid.30,41 While Dec protein binding is usually symmetric, the locations of the binding sites are not ubiquitous between viruses. For instance, the Dec protein binding sites for HSV1 are directly adjacent to the pentons, while for λ phage, the binding sites are similar to Dec at the quasi and true 3-fold sites.21,30,41 This difference likely results from the Dec protein being localized to the points of greatest stress, the distribution of which changes with capsid size.22 If this is the case, it is expected that the capsid reinforcement, mediated by auxiliary proteins, would have a symmetry-dependent effect to compensate for the anisotropic stress on the capsid. Our results support this argument by showing an anisotropic reinforcement of the P22 shell after Dec binding. Chemical and Mechanical Determinants of Stability. Despite the stark differences in the means of perturbation, our bulk chemical and single-particle mechanical stability experiments reflected the same trends between capsid morphologies. Interestingly, many of the mechanical trends between morphologies are only evident when single-particle measurements are grouped according to symmetry. The agreement in the single particle and bulk results likely originates from similar structural elements of the capsid being interrupted. The degradation of the P22 VLPs in bulk does not provide direct insights into the weakening or recovery of the capsid because the technique lacks the spatial resolution to be able to pinpoint which sites of the capsid are being affected by SDS. However, single-particle, symmetry-specific mechanical perturbation results provide some insights into the similar trends seen in the chemical experiments. Chemical destabilization and degradation of the particle with SDS likely functions by interrupting both intra- and intersubunit hydrophobic contacts. SDS has been previously shown to facilitate the expansion of the PC form of the capsid to the EX form presumably by providing more intrasubunit flexibility allowing for transition of subunits to the EX structure without the loss of intersubunit interactions.17 Our results demonstrate that at higher concentrations, SDS facilitates the degradation of P22 VLPs likely through the interruption of intersubunit interactions. Although both SDS-mediated ex-

CONCLUSIONS The symmetric interruption of subunit contacts in P22 VLPs modifies both the chemical and mechanical properties of the capsid in a striking way. This characterization utilized symmetric examination of single particles revealing anisotropic destabilization and stabilization, which manifest as bulk destabilization and stabilization under chemical stress. Under chemical stress, WB VLPs exhibited much lower stability than EX VLPs. AFM analysis suggested that the stability loss due to morphology change was dependent on the orientation of the capsid. This indicates that symmetry-specific mechanical changes may directly translate into changes in bulk stability against chemical stress. An inspection of the patterns of breakage and interaction energies between capsid subunits allowed us to suggest that loss of stability was concentrated at the edges of the cage. With the addition of a Dec protein that bound directly over those edges, the decrease in mechanical strength was restored. We found that the binding of Dec enhanced the overall mechanical properties of the VLPs but, notably, had a more significant effect on the mechanical variables most weakened after the removal of pentons. In agreement with these AFM results, our light-scattering and TEM experiments showed that Dec binding had a greater stabilizing effect toward WB VLPs than toward EX VLPs. Our results unveil a phenomenological coupling between the chemical and mechanical behavior of self-assembled protein cage systems. While the P22 system provides an ideal model for the validation of this type of examination through the selective expulsion of pentons and the binding of the Dec protein, this system also presents a broadly applicable platform for whole protein delivery and display.3,26,42 The selective control of mechanics demonstrated here may lead to a better understanding in the future of how these mechanics might be translated into subtle changes in the in vivo behavior of these capsids. In a more general sense, we show how the introduction of symmetric structural changes tunes the mechanical and chemical stability of a self-assembly nanostructure in a coupled fashion. This relationship between spatially discrete destabilization and reinforcement may be broadly applicable to a wide range of self-assembly systems. 8470

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano

sample was mixed at 160 Dec trimers per capsid (i.e., 2× excess Dec per binding site) and allowed to bind for 1 h at room temperature prior to the adsorption. Nanoindentations were done at a loading rate of 60 nm/s with forward elongation of 100 nm. FICs were obtained from force vs Z piezo (FZ) curves as previously explained.44 In brief, to obtain the FIC, one has to subtract from the FZ (which includes the deformation of the tip cantilever and the particle) the compliance of the tip cantilever. The elastic constant was obtained by fitting the initial linear part of each FIC; the breaking force and critical indentation were measured using the WsxM software.47 To calculate the energy, we assumed a linear deformation and integrated the area below the FIC, considering the elastic constant and the critical indentation.

MATERIALS AND METHODS Materials. Chemically competent BL21 (DE3) E. coli cells were purchased from Lucigen. TEM grids were purchased from Electron Microscopy Sciences. All other chemical reagents were purchased from Fisher Scientific. Protein Production. E. coli strains transformed with a pET Duet plasmid containing either the CP and SP genes or an N-terminal 6x his-tag Dec gene were grown in LB medium at 37 °C in the presence of ampicillin. Expression was induced at mid log phase (OD600 = 0.6) with isopropyl β-D-thiogalactopyranoside added to a final concentration of 0.5 mM. Four h postinduction, cells were harvested by centrifugation, and cell pellets stored at −80 °C overnight. Cell pellets were resuspended in PBS (100 mM sodium phosphate, 50 mM sodium chloride, pH 7) and lysozyme and RNase were added. Protease inhibitor, cOmplete minitabs (Roche), were added to Dec cultures. Cell suspensions were incubated for 30 min at room temperature. Suspensions were lysed by sonication, and cell debris was removed by centrifugation at 12,000 g for 45 min at 4 °C. P22 VLPs were purified from the supernatant by ultracentrifugation through a 35% (w/v) sucrose cushion, and the resulting pellet was resuspended in PBS. VLPs were purified using a S-500 Sephadex size exclusion column and a Biorad Biologic Duoflow FLPC. Flow rate for SEC purification was 1 mL/min. Fractions containing P22 were concentrated by ultracentrifugation. Dec was purified via his-tag affinity chromatography using a 5 mL cOmplete his-tag column (Roche) and a Biorad Biologic Duoflow FLPC. Samples were initially eluted using a 60 mL gradient from 20 to 250 mM imidazole. Fractions were dialzyed against PBS and reloaded onto the column. Samples were concentrated and further purified using a short 10 mL 20−250 mM gradient and subsequently redialyzed against PBS. Expansion. P22 PC VLPs were diluted to ∼1.5 mg/mL total protein in PBS and heated at either 65 or 75 °C for 25 min. Samples were at 16,000× g for 5 min to remove any aggregates, and the supernatant analyzed for expansion. Samples were analyzed by nondenaturing gel electrophoresis as previously described.43 SDS Destabilization. For TEM and initial concentration screening experiments, samples were incubated at room temperature in varying concentration of SDS in PBS and quenched at the point of analysis either by loading on a nondenaturing agarose gel or by deposition on a TEM grid. For light-scattering experiments, samples were diluted to ∼1 mg/ mL CP. Dec-bound samples were made by mixing Dec with P22 VLP samples at a ratio of 160 Dec trimers per capsid prior to dilution. The intensity of light scattering was monitored through the optical density (OD) at 320 nm with a background subtraction of the OD at 800 nm to account for any baseline fluctuations. Data were normalized to the starting OD320 value, and all runs were performed n = 3. Runs were plotted using IGOR Pro 6.3. TEM. Samples (5 μL, 0.1 mg/mL protein) were incubated for 30 s on carbon-coated copper grids. Grids were then washed with 5 μL of distilled water and stained with 5 μL 2% uranyl acetate. Images were captured at an accelerating voltage of 80 kV on a JEOL 1010 transmission electron microscope. AFM. AFM experiments were performed as described previously in Llauró et al.44 Experiments were done with a Nanotec Electrónica microscope (Madrid, Spain) operating in jumping mode plus.45 Imaging forces were kept between 60 and 150 pN. Rectangular siliconnitride cantilevers (RC800PSA, Olympus, Center Valley, PA) with a nominal spring constant of 0.05 N/m were used. Before every measurement cantilevers were calibrated by Sader’s method.46 Experiments were carried out in buffer conditions (100 mM phospathe, 50 mM NaCl, pH = 7) at a controlled temperature of 17 °C. A 20 μL drop of diluted stock solution was incubated on a freshly cleaved highly ordered pyrolytic graphite surface (ZYA quality; NT-MDT, Tempe, AZ). After 30 min, sample was washed with buffer solution until a volume of 90 μL was reached. AFM images were processed with the WSxM software (http://www.wsxmsolutions. com).47 For the binding of Dec (50 mM HEPES, 100 mM NaCl)

ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.6b03441. Tables with the mechanical characterization, AFM images, SEC chromatograms, nondenaturing agarose gels and TEM analysis are presented in the supplemental. An analysis of the change in subunit interaction based on the structure of the capsid is also shown (PDF)

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions ‡

These authors contributed equally.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS This work was supported by grants FIS2014-59562-R, FIS201571108-REDT, Fundación BBVA, and the “Mariá de Maeztu” Programme for Units of Excellence in R&D (MDM-20140377) to P.J.P.; NSF-BMAT DMR-1507282 to T.D. B.S. was supported by the Department of Defense (DoD) Air Force Office of Sponsored Research through the National Defense Science & Engineering Graduate Fellowship (NDSEG) Program. REFERENCES (1) Uchida, M.; Klem, M. T.; Allen, M.; Suci, P.; Flenniken, M.; Gillitzer, E.; Varpness, Z.; Liepold, L. O.; Young, M.; Douglas, T. Biological Containers: Protein Cages as Multifunctional Nanoplatforms. Adv. Mater. 2007, 19, 1025−1042. (2) Douglas, T.; Young, M. Viruses: Making Friends with Old Foes. Science 2006, 312, 873−875. (3) Schwarz, B.; Douglas, T. Development of Virus-Like Particles for Diagnostic and Prophylactic Biomedical Applications. Wiley Interdiscip Rev. Nanomed Nanobiotechnol 2015, 7, 722−735. (4) Lee, Y. J.; Yi, H.; Kim, W. J.; Kang, K.; Yun, D. S.; Strano, M. S.; Ceder, G.; Belcher, A. M. Fabricating Genetically Engineered HighPower Lithium-Ion Batteries Using Multiple Virus Genes. Science 2009, 324, 1051−1055. (5) Cieplak, M.; Robbins, M. O. Nanoindentation of 35 Virus Capsids in a Molecular Model: Relating Mechanical Properties to Structure. PLoS One 2013, 8, e63640. (6) Castellanos, M.; Perez, R.; Carrasco, C.; Hernando-Perez, M.; Gomez-Herrero, J.; de Pablo, P. J.; Mateu, M. G. Mechanical Elasticity as a Physical Signature of Conformational Dynamics in a Virus Particle. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 12028−12033. 8471

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano

Protective Immune Response to Influenza. ACS Nano 2013, 7, 3036− 3044. (26) Uchida, M.; LaFrance, B.; Broomell, C. C.; Prevelige, P. E.; Douglas, T. Higher Order Assembly of Virus-Like Particles (Vlps) Mediated by Multi-Valent Protein Linkers. Small 2015, 11, 1562− 1570. (27) Llauro, A.; Coppari, E.; Imperatori, F.; Bizzarri, A. R.; Caston, J. R.; Santi, L.; Cannistraro, S.; de Pablo, P. J. Calcium Ions Modulate the Mechanics of Tomato Bushy Stunt Virus. Biophys. J. 2015, 109, 390− 397. (28) Roos, W. H.; Bruinsma, R.; Wuite, G. J. L. Physical Virology. Nat. Phys. 2010, 6, 733−743. (29) Mateu, M. G. Mechanical Properties of Viruses Analyzed by Atomic Force Microscopy: A Virological Perspective. Virus Res. 2012, 168, 1−22. (30) Hernando-Perez, M.; Lambert, S.; Nakatani-Webster, E.; Catalano, C. E.; de Pablo, P. J. Cementing Proteins Provide Extra Mechanical Stabilization to Viral Cages. Nat. Commun. 2014, 5, 4520. (31) Klug, W. S.; Roos, W. H.; Wuite, G. J. L. Unlocking Internal Prestress from Protein Nanoshells. Phys. Rev. Lett. 2012, 109, 168104. (32) Perez-Berna, A. J.; Ortega-Esteban, A.; Menendez-Conejero, R.; Winkler, D. C.; Menendez, M.; Steven, A. C.; Flint, S. J.; de Pablo, P. J.; San Martin, C. The Role of Capsid Maturation on Adenovirus Priming for Sequential Uncoating. J. Biol. Chem. 2012, 287, 31582− 31595. (33) Li, Y.; Conway, J.; Cheng, N.; Steven, A.; Hendriz, R.; Duda, R. Control of Assembly: Hk97 ’Whiffleball’ Mutuant Capsids without Pentons. J. Mol. Biol. 2005, 348, 167−82. (34) Newcomb, W.; Trus, B.; Booy, F.; Steven, A.; Wall, J.; Brown, J. Structure of the Herpes-Simplex Virus Capsid - Molecular Composition of the Pentons and the Triplexes. J. Mol. Biol. 1993, 232, 499−511. (35) Carrasco, C.; Carreira, A.; Schaap, I. A. T.; Serena, P. A.; Gomez-Herrero, J.; Mateu, M. G.; de Pablo, P. J. DNA-Mediated Anisotropic Mechanical Reinforcement of a Virus. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 13706−13711. (36) Carrasco, C.; Castellanos, M.; de Pablo, P. J.; Mateu, M. G. Manipulation of the Mechanical Properties of a Virus by Protein Engineering. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 4150−4155. (37) Snijder, J.; Uetrecht, C.; Rose, R. J.; Sanchez-Eugenia, R.; Marti, G. A.; Agirre, J.; Guerin, D. M.; Wuite, G. J.; Heck, A. J.; Roos, W. H. Probing the Biophysical Interplay between a Viral Genome and Its Capsid. Nat. Chem. 2013, 5, 502−509. (38) Hernando-Perez, M.; Pascual, E.; Aznar, M.; Ionel, A.; Caston, J. R.; Luque, A.; Carrascosa, J. L.; Reguera, D.; de Pablo, P. J. The Interplay between Mechanics and Stability of Viral Cages. Nanoscale 2014, 6, 2702−2709. (39) Hespenheide, B.; Jacobs, D. J.; Thorpe, M. Structural Rigidity in the Capsid Assembly of Cowpea Chlorotic Mottle Virus. J. Phys.: Condens. Matter 2004, 16, S5055. (40) Ionel, A.; Velazquez-Muriel, J. A.; Luque, D.; Cuervo, A.; Caston, J. R.; Valpuesta, J. M.; Martin-Benito, J.; Carrascosa, J. L. Molecular Rearrangements Involved in the Capsid Shell Maturation of Bacteriophage. J. Biol. Chem. 2011, 286, 234−242. (41) Sae-Ueng, U.; Liu, T.; Catalano, C. E.; Huffman, J. B.; Homa, F. L.; Evilevitch, A. Major Capsid Reinforcement by a Minor Protein in Herpesviruses and Phage. Nucleic Acids Res. 2014, 42, 9096−9107. (42) Patterson, D. P.; LaFrance, B.; Douglas, T. Rescuing Recombinant Proteins by Sequestration into the P22 Vlp. Chem. Commun. 2013, 49, 10412−10414. (43) Kang, S.; Hawkridge, A. M.; Johnson, K. L.; Muddiman, D. C.; Prevelige, P. E., Jr. Identification of Subunit-Subunit Interactions in Bacteriophage P22 Procapsids by Chemical Cross-Linking and Mass Spectrometry. J. Proteome Res. 2006, 5, 370−377. (44) Llauro, A.; Guerra, P.; Irigoyen, N.; Rodriguez, J. F.; Verdaguer, N.; de Pablo, P. J. Mechanical Stability and Reversible Fracture of Vault Particles. Biophys. J. 2014, 106, 687−695. (45) Ortega-Esteban, A.; Horcas, I.; Hernando-Perez, M.; Ares, P.; Perez-Berna, A. J.; San Martin, C.; Carrascosa, J. L.; de Pablo, P. J.;

(7) Wang, J. C.-Y.; Chen, C.; Rayaprolu, V.; Mukhopadhyay, S.; Zlotnick, A. Self-Assembly of an Alphavirus Core-Like Particle Is Distinguished by Strong Intersubunit Association Energy and Structural Defects. ACS Nano 2015, 9, 8898−8906. (8) Schwarz, B.; Patterson, D.; Douglas, T., Virus-Like Particle Enzyme Encapsulation: Confined Catalysis and Metabolic Materials. In Viral Nanotechnology; CRC Press: Boca Raton, FL, 2015; pp 371− 382. (9) Prevelige, P. E.; Thomas, D.; King, J. Scaffolding Protein Regulates the Polymerization of P22 Coat Subunits into Icosahedral Shells Invitro. J. Mol. Biol. 1988, 202, 743−757. (10) Earnshaw, W.; Casjens, S.; Harrison, S. C. Assembly of Head of Bacteriophage P22 - X-Ray-Diffraction from Heads, Proheads and Related Structures. J. Mol. Biol. 1976, 104, 387−410. (11) Parker, M. H.; Casjens, S.; Prevelige, P. E. Functional Domains of Bacteriophage P22 Scaffolding Protein. J. Mol. Biol. 1998, 281, 69− 79. (12) Jiang, W.; Li, Z.; Zhang, Z.; Baker, M. L.; Prevelige, P. E.; Chiu, W. Coat Protein Fold and Maturation Transition of Bacteriophage P22 Seen at Subnanometer Resolutions. Nat. Struct. Biol. 2003, 10, 131−135. (13) Weigele, P. R.; Sampson, L.; Winn-Stapley, D.; Casjens, S. R. Molecular Genetics of Bacteriophage P22 Scaffolding Protein’s Functional Domains. J. Mol. Biol. 2005, 348, 831−844. (14) Chen, D. H.; Baker, M. L.; Hryc, C. F.; DiMaio, F.; Jakana, J.; Wu, W.; Dougherty, M.; Haase-Pettingell, C.; Schmid, M. F.; Jiang, W.; Baker, D.; King, J. A.; Chiu, W. Structural Basis for ScaffoldingMediated Assembly and Maturation of a Dsdna Virus. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 1355−1360. (15) Casjens, S.; Adams, M. B.; Hall, C.; King, J. AssemblyControlled Autogenous Modulation of Bacteriophage-P22 Scaffolding Protein Gene-Expression. J. Virol. 1985, 53, 174−179. (16) Zhang, Z.; Greene, B.; Thuman-Commike, P. A.; Jakana, J.; Prevelige, P. E., Jr; King, J.; Chiu, W. Visualization of the Maturation Transition in Bacteriophage P22 by Electron Cryomicroscopy1. J. Mol. Biol. 2000, 297, 615−626. (17) Gilcrease, E. B.; Winn-Stapley, D. A.; Hewitt, F. C.; Joss, L.; Casjens, S. R. Nucleotide Sequence of the Head Assembly Gene Cluster of Bacteriophage L and Decoration Protein Characterization. J. Bacteriol. 2005, 187, 2050−2057. (18) O’Neil, A.; Prevelige, P. E.; Basu, G.; Douglas, T. Coconfinement of Fluorescent Proteins: Spatially Enforced Communication of Gfp and Mcherry Encapsulated within the P22 Capsid. Biomacromolecules 2012, 13, 3902−3907. (19) Teschke, C. M.; McGough, A.; Thuman-Commike, P. A. Penton Release from P22 Heat-Expanded Capsids Suggests Importance of Stabilizing Penton-Hexon Interactions During Capsid Maturation. Biophys. J. 2003, 84, 2585−2592. (20) Parent, K. N.; Deedas, C. T.; Egelman, E. H.; Casjens, S. R.; Baker, T. S.; Teschke, C. M. Stepwise Molecular Display Utilizing Icosahedral and Helical Complexes of Phage Coat and Decoration Proteins in the Development of Robust Nanoscale Display Vehicles. Biomaterials 2012, 33, 5628−5637. (21) Parent, K. N.; Khayat, R.; Tu, L. H.; Suhanovsky, M. M.; Cortines, J. R.; Teschke, C. M.; Johnson, J. E.; Baker, T. S. P22 Coat Protein Structures Reveal a Novel Mechanism for Capsid Maturation: Stability without Auxiliary Proteins or Chemical Crosslinks. Structure 2010, 18, 390−401. (22) Zandi, R.; Reguera, D. Mechanical Properties of Viral Capsids. Phys. Rev. E 2005, 72, 021917. (23) Tang, L.; Gilcrease, E. B.; Casjens, S. R.; Johnson, J. E. Highly Discriminatory Binding of Capsid-Cementing Proteins in Bacteriophage L. Structure 2006, 14, 837−845. (24) Huang, C. J.; Butler, P. J.; Tong, S.; Muddana, H. S.; Bao, G.; Zhang, S. L. Substrate Stiffness Regulates Cellular Uptake of Nanoparticles. Nano Lett. 2013, 13, 1611−1615. (25) Patterson, D. P.; Rynda-Apple, A.; Harmsen, A. L.; Harmsen, A. G.; Douglas, T. Biomimetic Antigenic Nanoparticles Elicit Controlled 8472

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473

Article

ACS Nano Gomez-Herrero, J. Minimizing Tip-Sample Forces in Jumping Mode Atomic Force Microscopy in Liquid. Ultramicroscopy 2012, 114, 56− 61. (46) Sader, J. E.; Chon, J. W. M.; Mulvaney, P. Calibration of Rectangular Atomic Force Microscope Cantilevers. Rev. Sci. Instrum. 1999, 70, 3967−3969. (47) Horcas, I.; Fernandez, R.; Gomez-Rodriguez, J. M.; Colchero, J.; Gomez-Herrero, J.; Baro, A. M. Wsxm: A Software for Scanning Probe Microscopy and a Tool for Nanotechnology. Rev. Sci. Instrum. 2007, 78, 013705.

8473

DOI: 10.1021/acsnano.6b03441 ACS Nano 2016, 10, 8465−8473