Swelling and Softening of the Cowpea Chlorotic Mottle Virus in Response to pH Shifts

Swelling and Softening of the Cowpea Chlorotic Mottle Virus in Response to pH Shifts

Biophysical Journal Volume 108 May 2015 2541–2549 2541 Article Swelling and Softening of the Cowpea Chlorotic Mottle Virus in Response to pH Shifts ...

900KB Sizes 0 Downloads 59 Views

Biophysical Journal Volume 108 May 2015 2541–2549

2541

Article Swelling and Softening of the Cowpea Chlorotic Mottle Virus in Response to pH Shifts Bodo D. Wilts,1 Iwan A. T. Schaap,1,2 and Christoph F. Schmidt1,* 1

Drittes Physikalisches Institut, Fakulta¨t fu¨r Physik, Georg-August Universita¨t, Go¨ttingen, Germany; and 2Center for Nanoscale Microscopy and Molecular Physiology of the Brain, Go¨ttingen, Germany

ABSTRACT Cowpea chlorotic mottle virus (CCMV) forms highly elastic icosahedral protein capsids that undergo a characteristic swelling transition when the pH is raised from 5 to 7. Here, we performed nano-indentation experiments using an atomic force microscope to track capsid swelling and measure the shells’ Young’s modulus at the same time. When we chelated Ca2þ ions and raised the pH, we observed a gradual swelling of the RNA-filled capsids accompanied by a softening of the shell. Control experiments with empty wild-type virus and a salt-stable mutant revealed that the softening was not strictly coupled to the swelling of the protein shells. Our data suggest that a pH increase and Ca2þ chelation lead primarily to a loosening of contacts within the protein shell, resulting in a softening of the capsid. This appears to render the shell metastable and make swelling possible when repulsive forces among the capsid proteins become large enough, which is known to be followed by capsid disassembly at even higher pH. Thus, softening and swelling are likely to play a role during inoculation.

INTRODUCTION Viruses replicate and self-assemble by hijacking their host cell’s resources (1). Viruses are typically composed of a capsid (a protein shell of simple, often icosahedral geometry) and an encapsulated viral genome, either DNA or RNA. The capsid is constructed from a well-defined number of copies of one or a few different types of proteins (2,3). Viral capsids are minimalistic nanocontainers that have to strike a balance between two seemingly contradictory requirements. They must be robust enough to protect the viral genome against physical and chemical damage, but still be able to release the genome into the target cell during infection. Viral genomes are often tightly packed in the capsids and, in the case of double-stranded DNA (dsDNA) bacteriophages, can exert considerable pressure on the shells (4–7). The mechanism of genome release varies among virus types and depends on the host organism. When the virus particle enters the host cell, the viral genome is commonly released via chemically triggered conformational transitions in the capsids (8). Examples of animal viruses that undergo structural changes in response to pH changes include polio virus (9) and influenza virus (10). Plant viruses of the bromovirus family appear to employ a structural transition that in some cases involves capsid swelling to aid genome delivery (11). Atomic force microscopy (AFM) has been established in recent years as a powerful tool for measuring virus capsid mechanics (12–18). AFM allows one to follow the structural

Submitted December 9, 2014, and accepted for publication April 15, 2015. *Correspondence: [email protected] Bodo D. Wilts’s present address is Adolphe Merkle Institute, University of Fribourg, Fribourg, Switzerland.

changes that occur during virus disassembly. In a study of influenza virus (17), a stepwise softening due to structural rearrangements was identified that correlated with the passage of the virus through the endocytic pathway during which the pH drops from 7 to 5. The importance of these structural changes was underlined by a >2-fold reduction of the infection rate when the endocytic acidification was skipped. Cowpea chlorotic mottle virus (CCMV) is an icosahedral plant virus of the bromovirus family that has been prominently explored as a nanocontainer for drug delivery since capsids readily assemble around cargoes other than the native RNA (1,19). CCMV capsid proteins self-assemble under various buffer conditions into different protein superstructures, ranging from rods to sheets and various polyhedral shapes (20,21). The native capsid is made out of 180 identical protein subunits arranged into an icosahedron with triangulation number T ¼ 3 (2). The capsid of CCMV thus has the symmetry properties of an icosahedron, with 2-, 3-, and 5-fold symmetry axes. The subunits are arranged into 12 pentamers and 20 hexamers. The 2-fold symmetry axes can be found between adjacent hexamers, the 3-fold (or quasi-6fold) symmetry axes are located within the hexamers, and the 5-fold symmetry axes are found within the pentamers (22). In addition, quasi-3-fold vertices are found at the junctions of two hexamers with a pentamer. CCMVs viral genome is multipartite, comprising three single-stranded RNAs. RNAs 1 and 2 (3171 and 2774 nt, respectively) are packed separately and encode for RNA replication. RNA 3 (2173 nt) encodes for viral movement and coat protein, and is packed together with the subgenomic RNA 4 (824 nt) (23–25). Thus, in the RNA-filled state, all CCMV capsids enclose an ~3-kb-long genome.

Editor: Matthias Rief. Ó 2015 by the Biophysical Society 0006-3495/15/05/2541/9 $2.00

http://dx.doi.org/10.1016/j.bpj.2015.04.019

2542

The capsid of CCMV assumes two distinct structural states, and one can switch it between these states by changing the pH and controlling the Ca2þ ions (26–28). The unswollen state of CCMV capsids (RNA filled or empty) is stable at pH 5 (15,21), has an outer diameter of ~28 nm and an inner diameter of ~21 nm, and is referred to as the native state (28). The expanded, or swollen, state can be observed when the pH is raised to ~7 at low ionic strength and in the absence of divalent ions (29). The swelling is suggested to be driven by electrostatic repulsion due to deprotonation at specific sites. Addition of divalent metal ions eliminates this repulsion and prevents swelling (30–32). The shell radius is increased by ~10% in the swollen state compared with the native state, as measured by electron microscopy (28). A further pH increase leads to complete disassembly into coat proteins and RNA. In the expanded state, pores ~2 nm in diameter open along the quasi-threefold axes of the capsid. Such swelling with increasing pH has also been found for other members of the bromovirus family (11,33) and may be a universal part of the physiological inoculation mechanism. Intriguingly, CCMV is stable at low pH and loses integrity at neutral pH, which is exactly the opposite of the behavior shown by the aforementioned animal viruses. This particular pH dependence is likely to be optimized for the infection pathway of bromoviruses. The stability at low pH may be an adaptation to the low pH in the gastrointestinal tract of the (insect) vector that is responsible for transmission (34). Wild-type (WT) virus can be studied in the expanded state, whereas empty capsids typically dissociate before the swelling transition is reached. A salt-stable mutant, which resists ionic strengths up to 1 M and pH values up to pH 8 without dissociation, was reported by Bancroft et al. (35) in the 1970s and later characterized biochemically by Fox et al. (36). The salt-stable version differs by only one amino acid in the N-terminal arm of the capsid protein (K42R). This mutation has been speculated to enhance a ring of salt bridges around the quasi-6-fold vertices and possibly to also stabilize the pentameric units (36). In our earlier AFM nano-indentation studies, we probed CCMV capsids in their native state (15), and later we investigated the softening of the empty CCMV at pH 6.0 (37), before any swelling could be seen. The softening of the shells was found to be accompanied by a transition from brittle to elastic behavior, which has also been studied in numerical simulations (38,39). Recently, Vaughan et al. (40) reported that brome mosaic virus (BMV) softened when the pH was changed from 4.5 to 6.2, but they did not observe swelling. Cieplak and Robbins (41) modeled the low-pH state of CCMV capsids by coarse-grained molecular dynamics with a focus on the nonlinear response regime and shell rupture, and showed symmetry breaking that was not captured by continuum modeling. Kononova et al. (42) performed AFM indentation experiments on empty CCMV capsids at pH 5.0 with high force, accompanied by modeling to describe nonlinear response and capsid collapse, and Biophysical Journal 108(10) 2541–2549

Wilts et al.

found experimental stiffness values consistent with the earlier experiments (15). To date, a mechanical characterization of the actual swelling transition of RNA-filled CCMV up to pH 7 is still lacking. Here, we measured the elastic response parameters of CCMV virus shells while they expanded with an increase in pH. In all cases, we imaged immobilized viral particles in liquid using AFM in tapping mode. We were able to characterize the swelling process over a pH range between 4.8 and 7.5. We then compared the elastic response and swelling behavior of WT RNA-filled CCMV with those of empty shells and shells made by the salt-stable mutant at different pH values. We found that shell softening and swelling are two separate processes. MATERIALS AND METHODS Virus particles All virus samples used in this work were kind gifts from Charles Knobler (UCLA) and Mark Young and Chris Broomell (University of Montana). Wild-type RNA-filled viruses and salt-stable mutants were prepared by ultrafiltration from infected Cowpea California Black Eye No. 5 plants as described previously (15,36). Empty virus capsids were prepared by in vitro reassembly of purified viral capsid (21) and salt-stable virus mutants (K42R) were prepared by infecting plants with in vitro transcribed mutated RNA (36). Virus solutions were stored at 80 C and imaged at pH 4.8 in storage buffer (100 mM sodium acetate, 1 mM EDTA, pH 4.8). To image the viruses in different pH conditions, we prepared buffers using 20 mM sodium phosphate, adjusted to pH values of 5.6, 6.0, 6.3, 6.9, and 7.5, respectively, by adding PBS. EDTA (200 mM) was added to the buffers to keep the buffers free of Ca2þ and Mg2þ and to chelate the 180 bound Ca2þ ions from the quasi-3-fold vertices of each virus capsid.

AFM For imaging and elastic-response measurements by AFM, virus particles were adsorbed to microscope coverslips. The glass coverslips were first cleaned in a saturated solution of KOH in ethanol and then either rendered hydrophobic by silanization with hexamethyldisilazane (HMDS, C6H19NSi2, 99.9% pure; Sigma-Aldrich) vapor or positively charged by silanization with N1-[3-(Trimethoxysilyl)-propyl] diethylenetriamine (DETA, C10H27N3O3Si; SigmaAldrich) (15,43). Viral particles were immobilized using HMDS-coated surfaces at low pH in the native state and DETA-coated surfaces in higherpH environments (pH > 5.6). The exterior of CCMV is negatively charged (44) and thus is expected to bind to DETA surfaces, but we found that binding to the HDMS hydrophobic surface was stronger at low pH. Stock solutions (~1 mg/ml for the RNA-filled WT capsids, ~0.1 mg/ml for the empty WT capsids, and ~2 mg/ml for the salt-stable mutant virus) of each type of CCMV capsid were diluted to ~10 mg/ml in phosphate buffer at the desired pH. A 60 ml droplet of viral solution was deposited on the silanized substrate and incubated for 20 min to allow the capsids to adsorb. Then the sample was washed twice in the respective buffer to remove unadsorbed particles. The AFM tip was prewetted with a 30 ml drop of the same buffer. Imaging and elastic-response measurements were performed in buffer at room temperature with an atomic force microscope (MFP3D-BIO; Asylum Research, Santa Barbara, CA) using tapping mode (45) for imaging and force-volume mode (46) to measure the elastic response. Silicon-nitride, gold-coated cantilevers (BL-RC150VB Bio-Lever; Olympus, Tokyo, Japan) with nominal spring constants of 0.03 N/m and a nominal tip radius of 30 nm were calibrated by the thermal-noise method (47,48) before every

pH Dependence of CCMV Mechanics

2543

measurement to ensure proper force calibration. Viral capsids were first imaged in a large field of view (1  1 mm or 1.5  1.5 mm, divided into 128  128 pixels) to determine their density and position on the surface. Imaging forces were minimized (<100 pN) (49) to avoid dislodging or deforming the capsids. Subsequently, single capsids were scanned with higher resolution using a field of view between 200  200 nm and 300  300 nm (128  128 pixels). The height of the particle was determined from these high-resolution images (Fig. 1). The elastic response of the capsids was mapped by recording a rapid succession of force-distance (FD) curves using an approach speed of 0.5 mm/s, taking ~1 s per curve, in a raster-scan fashion covering an area of ~200 nm  200 nm (Fig. 2). In this mode, lateral displacements occur only when the tip is out of contact with the sample, which minimizes lateral shear forces. We used a maximum axial force of 200 pN. As a control to show that response mapping did not destroy the viruses, we imaged the particles at the end again in tapping mode. We found by trial and error that a setpoint of 200 pN maximal tip-sample force left the virus capsids unharmed.

Evaluation of viral stiffness The force-mapping routine used in this work is described in detail elsewhere (50). The recorded elastic-response maps contained one FD curve for each pixel of the designated scan area (Fig. 2). Force maps were evaluated using a custom-written LabVIEW routine (National Instruments, Austin, TX). The FD curves from an area of ~10 nm radius around the viral center were automatically selected, aligned, and subsequently linearly fitted between 60 pN and 150 pN. The capsid’s effective stiffness was then determined by means of a simple two-spring model (12,51), which assumes that the combination of cantilever and virus capsid acts as two undamped Hookean springs in series on the relevant timescales. With this model, the shells’ spring constant can be calculated as

ksh

kcl km ¼ ; kcl  km

Finite-element modeling We used the program Comsol 3.4 (COMSOL Multiphysics, Go¨ttingen, Germany) for finite-element modeling (FEM). We created a simple thick-shell model similar to the one described by Gibbons and Klug (52). The native state of CCMV was approximated as a 2.8-nm-thick shell with an outer diameter of 26.4 nm. The Young’s modulus was set to 250 MPa to match the stiffness of the model to the experimentally found values. The Poisson ratio n was set to 0.4. We also tested other values for n from 0.3 to 0.5, which changed the stiffness by not more than 510% (see Fig. 6). Changing E or v affects the absolute stiffness, but not the relative softening of the capsid at larger diameters. We neglected the effect of the encapsulated RNA in the model because the experiment showed only a small difference between empty and filled capsids, and because adding RNA to the model would involve assumptions about the geometry and the coupling to the capsid for which we have no independent evidence. We placed the model on a solid support and hindered it from rotating by setting appropriate boundary conditions. To simulate the indentation by the AFM probe, we modeled the tip apex as a parabolic cone with a tip radius of 20 nm. Inertia and drag are not expected to play a role on the timescales of the experiments and were not included in the simulations. We used the axial symmetry module of the program to reduce the model to a 2D slice of the sphere, which was meshed with ~1000 elements. The stiffness of the model capsid was obtained by linear fitting of the calculated indentation curve between 60 pN and 150 pN. To model the expanded state, we assumed conservation of material and rescaled the native-state model in radial increments of 0.5 nm to a final maximum outer diameter of 29.9 nm, corresponding to a swelling of 13%. The material parameters were not changed to estimate the change in the effective shell stiffness due to capsid expansion and shell thinning only.

RESULTS AND DISCUSSION (1)

where kcl is the cantilever’s spring constant and km is the spring constant measured from the slope of the (approach) FD curve. Only data in which the complete virus could be mapped were added to the histograms. To check for a possible effect of the substrate, the same procedure was performed on background curves recorded from spots next to virus particles (Fig. 2).

Imaging CCMV capsids The capsid diameter could be determined most accurately from the measured height of the particles (Fig. 1) because the measured height was not affected by the AFM tip radius. The measured particle width, in contrast, was substantially larger than the actual particle diameter due to tip broadening

FIGURE 1 Imaging of virus particles. (A) Height contours of surface-adsorbed CCMV particles in the native state. Three exemplary images of RNA-filled WT viruses at pH 4.8 (scale bar: 20 nm). Possible different attachment orientations of the icosahedral shells on the surface could not be distinguished from the topography or height due to tip dilation effects (Fig. S1). (B) Line profiles across the centers of the viruses from the top row. (C) Histogram of measured heights from all WT CCMV viruses imaged at this pH. To see this figure in color, go online.

Biophysical Journal 108(10) 2541–2549

2544

FIGURE 2 Elastic response of virus particles. (A) Height image of a WT virus particle at pH 4.8, which is reconstructed from a 2D array of FD curves (scale bar: 60 nm). (B) Representative FD curves taken from the background around the particle (left, gray dotted lines: taken at locations marked by orange squares on the substrate in A) and from the center of the viral capsid (right, pink dotted lines: taken at locations marked by red squares on the virus in A), showing the reproducibility of the FD curves. The maximum applied force was always kept at <200 pN to prevent damage to the capsids. The black (left) and red (right) lines in the FD plots are averaged curves. The distance is the motion of the cantilever base by the z-piezo; application of the two-spring model (Eq. 1) to the linear part of the response curve (marked by arrows) yields a stiffness of 0.21 N/m for this virus. To see this figure in color, go online.

(53). The image resolution was likely limited by the softness of the capsid, and not by the sharpness of the AFM tip we used. Studies using identical AFM tips have obtained higher resolutions on stiffer capsids (13). The measured particle heights were very reproducible, and the observed small variation of ~1 nm can be largely explained by unevenness of the reference background and possibly by particle attachment in different orientations with respect to the shell symmetries. Fig. 3 shows the dependence of the capsid diameter on buffer pH for RNA-filled WT virus, empty WT shells, and RNA-filled salt-stable mutant virus (K42R). Additional images are shown in Fig. S2 in the Supporting Material. Height histograms obtained at each pH value, measured in different samples of the three species of virus, are shown in Fig. S3. Table 1 summarizes the measured values for both particle diameter and mechanical response. The average diameter we measured for CCMV viruses in the native state (at pH 4.8) of 27.6 5 0.1 nm is roughly consistent with the value of 27.94 nm along the 3-fold axis obtained from cryo-electron microscopy (cryo-EM) reconstructions (28). The slightly smaller diameter measured by AFM might be due to a residual indentation effect, but the fact that the values are very close proves that the residual force we exerted with the AFM tip did not substantially reduce the particle diameter. The diameters of the WT empty and filled capsids were experimentally indistinguishable. The salt-stable mutant RNA-filled viruses had a slightly larger radius (28.2 5 0.2 nm), which is even closer to the cryo-EM results (28). This is likely due to the increased stiffness we also measured with the mutant (Fig. 4), which would lead to slightly less residual indentation by the imaging forces. We did not observe multimodal height distributions for a given capsid type at one pH, which might have been expected due to the different orientations in Biophysical Journal 108(10) 2541–2549

Wilts et al.

FIGURE 3 Swelling transition of three types of CCMV particles at increasing pH. (A) Schematic of the three types of particles studied (left to right): empty reconstituted capsid, WT RNA-filled virus particle, and salt-stable mutant RNA-filled virus particles. (B) Height versus pH plot for the CCMV capsids (mean 5 SE). Under native-state conditions (low pH), all viruses have approximately the same mean height. Increasing pH led to dissociation of the empty capsids (at pH 6). RNA-containing WT capsids increased in diameter, but the salt-stable mutant virus particles did not. To see this figure in color, go online.

which the capsids can attach to the surface. All imaged particles also appeared to have close-to-spherical circumferences and equal heights within 51 nm. To determine what differences we could reasonably expect in AFM images from capsids adsorbed to the substrate in different orientations, we constructed simulated images of CCMV virus particles on a solid surface, based on the atomic structure of the shells assuming an AFM tip diameter of 15 nm (Fig. S1) (53–55). Although the atomic structure of CCMV capsids has surface corrugations that lead to different diameters along the different symmetry axes, the 15 nm tip effectively filters out such localized features. We found that the different orientations of the capsids (along the primary symmetry directions) should result in only very minor differences in outer shape and height (Fig. S1), consistent with our experiments. In addition, modeling the mechanical response of the different orientations of CCMV (39) showed that the stiffness values for the small deformations that were used in our measurements should be indistinguishable at the level of accuracy we could attain. An increase of pH combined with chelation of divalent ions led, as expected, to a clearly measurable expansion of the RNA-filled WT viruses (Fig. 3). We could not resolve any pores in our images, probably due to the compliance of the shells. The transition progressed gradually with increases in pH, beginning at pH values of ~6.3 and progressing in the pH range of 6.3–7.5. The lack of a pronounced sigmoidal shape of the transition across the relatively large pH range indicates that the expansion is not a strongly

pH Dependence of CCMV Mechanics TABLE 1

2545

Mean 5 SE for capsid heights, spring constants, and measurement days for the RNA-filled WT CCMV

pH Days Height (nm) Stiffness (N/m)

4.8 7 27.56 5 0.13 N ¼ 43 0.195 5 0.006 N ¼ 35

5.6 4 27.85 5 0.12 N ¼ 22 0.194 5 0.012 N ¼ 17

6.0 3 28.16 5 0.08 N ¼ 28 0.161 5 0.006 N ¼ 18

cooperative phenomenon within a given shell, i.e., it is not equivalent to a phase transition. We also did not find any evidence for coexistence between expanded and native shells at intermediate pH values, as the widths of the size distributions did not noticeably broaden during the transition (Fig. S3). The maximal average diameter measured at pH 7.5 was 30.3 5 0.11 nm, corresponding to an expansion of 9.7%. This increase is consistent with the reported increase of ~10% from cryo-EM (28) and an all-atom multiscale simulation (32). The pH value at which the expansion of the shells started is slightly lower than that reported from proton NMR studies (29). Proton NMR suggested a pH of 7–7.2 for the beginning of the expansion, but proton NMR is not capable of providing direct structural information. Therefore, it remains unclear whether the increased proton movement shown by NMR correlates exactly with the structural expansion of the shell, which we measured directly by AFM. We could not probe the reversibility of the swelling transition because we had to use different substrate surfaces to attach virus particles to the surface in the different pH regimes, i.e., when we changed the pH on a given sample (either from low to high or from high to low), particles would come off the surface. Empty WT virus could not be observed at pH > 6 because the capsids disassembled. This is consistent with previous reports of disassembly at pH 6 without expansion (37), although the stiffness we observed with empty capsids at pH 6 (Fig. 3; Table 2) was somewhat lower than what was reported earlier (37).

6.3 4 28.88 5 0.93 N ¼ 26 0.124 5 0.01 N ¼ 15

6.9 4 29.62 5 0.13 N ¼ 36 0.076 5 0.006 N ¼ 25

7.5 2 30.3 5 0.11 N ¼ 24 0.046 5 0.003 N ¼ 15

The salt-stable mutant of CCMV, in contrast, could be observed at high pH as expected. Interestingly, it did not expand when the pH was increased to 6.9 (Fig. 3; Table 3). This is consistent with earlier experiments (35,36) in which expansion was found to be absent or much less pronounced than in WT virus, but inconsistent with other experiments that showed the same expansion as in WT virus (56). The reason for this behavior may be a change in the interaction between the N-terminal arms of the coat proteins that stabilize the quasi-6-fold vertices and the pentameric units (30–32). The point mutation from lysine to arginine at position 42 (K42R) in the N-terminal arms may thus stabilize against the expansion of the virus by enhancing protein-protein and/or protein-RNA interactions elsewhere in the capsid. The apparently inconsistent results for swelling that have been reported may point to an important fact: the mutant shells may become only marginally unstable around pH 7.5, such that small differences in solvent conditions, divalent ions, temperature, etc. can lead to large conformational changes. Elastic-response measurements Average elastic-response curves measured in the central region of RNA-filled WT capsids are shown in Fig. 4. Compared with the indentation curve recorded on the plain substrate surface, a soft (i.e., gradual) transition to contact is apparent in all cases, and is increasingly soft at higher pH. The transition zone corresponds to an indentation of 1–4 nm and likely results from a combination of thermal tip fluctuations, roughness of the tip and the virus shells, and the local compression of the capsid wall under the tip, causing a Hertzian-type soft onset of the indentation curves (39,51). It is difficult to fit the response in this transition zone because many parameters, including shell radius and thickness and AFM tip size, influence the relative importance of the respective contributions. At contact forces > 50 pN, the response was linear, as expected for a weakly indented TABLE 2 Mean 5 SE for capsid heights, spring constants, and measurement days for the empty WT CCMV

FIGURE 4 Force-indentation measurements on individual CCMV particles at increasing pH. Reference curve on substrate (small circles, grey). Average force-indentation curves of individual RNA-filled WT viruses (from left to right) at pH 4.8 (red, N ¼ 9), 6.3 (green, N ¼ 11), and 7.5 (blue, N ¼ 8), showing a decrease in slope with increasing pH. The inset defines measured parameters. To see this figure in color, go online.

pH Days Height (nm) Stiffness (N/m)

4.8 4 27.2 5 0.34 N ¼ 18 0.165 5 0.006 N ¼ 18

6.0 4 27.4 5 0.31 N ¼ 15 0.13 5 0.01 N ¼ 13

Biophysical Journal 108(10) 2541–2549

2546

Wilts et al.

TABLE 3 Mean 5 SE for capsid heights, spring constants, and measurement days for the RNA-filled salt-stable mutant (K42R) pH Days Height (nm) Stiffness (N/m)

4.8 3 28.2 5 0.23 N ¼ 30 0.275 5 0.009 N ¼ 30

6.9 4 28.1 5 0.15 N ¼ 17 0.107 5 0.0067 N ¼ 17

elastic shell (12), and we focused on this part of the response curves to obtain elasticity values. Full distributions of elastic response for each pH are shown in Fig. S4. The elastic response of RNA-filled WT virus particles dramatically changed with increasing pH. The change of shell stiffness with pH is clearly visible in a decrease in slope of the indentation curves at higher forces (Fig. 4). A distinct softening of the RNA-filled WT capsids started at approximately pH 6 (Fig. 5). This correlated with the onset of swelling of the shells (Fig. 3). The softening continued with increasing pH to ~25% of the initial stiffness at the highest pH tested (pH 7.5). Measured height and stiffness are roughly linearly correlated (Fig. 6) for the RNA-filled WT virus capsids. This prompted us to ask whether the softening of the elastic response of the capsid might simply be due to the increase in radius that would necessarily go along with a decrease in effective shell thickness if the density of the shell were to remain constant. To test for this possibility, we estimated the change in capsid elastic response by FEM. We used a simple approach, similar to what has been described previously (15,51), to model the virus as a homogeneous elastic thick shell with a radius R and wall thickness t, indented by a parabolic tip of 20 nm radius. We

FIGURE 5 Softening mechanical response of CCMV particles to indentation with increasing pH. Plot of the averaged stiffness values (error bars represent mean 5 SE) measured for surface-adsorbed CCMV capsids versus pH. Under native-state conditions (pH 4.8), the three different virus particles we probed (WT RNA-filled (black squares), WT empty (red circles), and salt-stable mutant (blue triangles)) had different stiffnesses. Virus particles at pH 4.8 were adsorbed on HMDS-coated and at pH 5.6– 7.5 on DETA-coated surfaces. Inset: pH-dependent stiffnesses normalized by the respective pH 4.8 values. To see this figure in color, go online. Biophysical Journal 108(10) 2541–2549

FIGURE 6 Correlation of particle stiffness and height for the RNA-filled WT capsids. Data were obtained from measurements between pH 4.8 and 7.5. The correlation was close to linear (linear regression: red dotted line). The three blue and green lines give the dependency of stiffness on height according to the thick-shell simulations (see text). To see this figure in color, go online.

calculated the shell thickness t for the swollen viruses assuming constant density. The Young’s modulus of the shell material was set to 250 MPa to reproduce the measured response of WT capsids. The calculated spring constants for the swollen capsids are plotted in Fig. 6 together with the measured stiffnesses. The modeling results show a clearly weaker decrease in stiffness with shell expansion (only reaching 69% of the initial stiffness at a diameter increase of 10%) than experimentally measured. Recent more detailed modeling, based on the atomic structure of CCMV capsids (39), predicted a somewhat greater softening to ~50% of the initial value from a simple shell expansion, which still does not explain the measured ~4-fold decrease in stiffness of the swollen RNA-filled WT capsids at pH 7.5. These results show that the shell softening is caused not merely by the expansion of the shell but also by a substantial weakening of the shell structure due to changed bond structures and/or effects of the packed RNA. Recent advances in the mechanical modeling of virus shells that take the molecular structure into account may provide more detailed insight into how the structural changes are responsible for the observed shell softening (32,41,42,57). Under native-state conditions (pH 4.8), the three different virus particles we probed (WT RNA-filled, WT empty, and salt-stable mutant) had different stiffnesses. Increasing pH led to a softening of all three types of particles. The softening behavior was similar for all species and could be described by a common curve after normalization to the initial values at pH 4.8 (Fig. 5, inset). Further, surface treatment had no obvious influence on measured stiffnesses (pH 4.8: HMDS-coated surface; pH 5.6: DETA-coated surface). Unfortunately, the use of different surfaces precluded us from measuring the reversibility of the changes in stiffness by reversing the pH from high to low again on the same sample (17). The stiffnesses we measured for the three different types of virus capsids at pH 4.8 (Table 1) agree well with our

pH Dependence of CCMV Mechanics

earlier results (15). We found small deviations from our earlier results, close to the error margins, in the stiffnesses of both the salt-stable mutant virus (0.275 5 0.009 N/m compared with 0.31 5 0.02 N/m) and the empty WT virus (0.165 5 0.006 N/m compared with 0.15 5 0.01 N/m). RNA is likely to play a role in reinforcing the capsid shell as it binds to the interior of the capsid. This explains the overall higher stiffness values for the RNA-containing viruses at pH 4.8. The RNA strengthened the WT shells by ~20% and the salt-stable mutant shells by ~100% compared with the empty WT virus. The RNA-shell protein interaction also is likely to be the reason for the stability of the WT capsid at pH values higher than 6.0. Recently, it was found for BMV that the particular RNA strand of the tripartite genome that is packed can also affect the measured stiffness (40). Since in our experiments we could not differentiate among the three different genomes, such a genome-specific reinforcement may be hidden in the spread of the data. Interestingly, the pH-dependent elastic responses of the three different types of virus capsids we examined very nearly collapse onto a single master curve when normalized by the respective pH 4.8 value (Fig. 5, inset). This suggests that the mechanism of pH-induced shell softening is very similar for all examined types of CCMV shells, even though the initial shell stiffness may vary. Although swelling was not observed for the salt-stable capsids at pH 6.9, the capsids nevertheless showed a distinct softening by a factor of ~3, comparable to the softening of the WT RNA-filled capsids. The fact that salt-stable mutant capsids softened in exactly the same way as the other shells, but failed to swell, indicates that the change in stiffness is uncoupled from the change in diameter of the capsids. We saw indications of a corresponding effect in earlier work with WT empty capsids at pH 6 (37), which were also discussed in the context of modeling results (39). This effect points to a structural transition in the connectivity in the shell. WT empty capsids softened to ~80% of their initial stiffness at pH 6, which is comparable to the softening of the RNA-filled WT capsids at the same pH and slightly less drastic than our previous result (37). We also tested the mechanical failure of the capsids at high forces and mostly found catastrophic breakthrough even for the softer shells at high pH (Fig. S5), somewhat in contrast to our earlier study (37) in which we saw a completely reversible strong indentation. This may be due to the use of predamaged shells in those earlier measurements and the increased tip-positioning accuracy of the current force-map-based methods. Structural basis of capsid softening Multiple studies have shown that capsid swelling involves large structural changes around the quasi-3-fold vertices that lead to an early loss of interactions around those points (30–32,58). This picture is entirely consistent with the notion

2547

that deprotonation at the calcium-binding sites with increasing pH and loss of divalent metal ions at the same locations serve as triggers for the transition. Since within a continuum elasticity framework, overall shell stiffness scales linearly with the shell material’s Young’s modulus, the close-to-linear decrease in the stiffness with pH points to an approximately linear change in the effective Young’s modulus of the shell with pH. The loss of interactions around the pseudo-3-fold vertices together with the mobility of the building blocks of the capsid that allows swelling could explain such a decreased effective Young’s modulus (37), as was also theoretically described in a continuum theory (59). Interestingly, the all-atom simulations of Miao et al. (32) showed an increase of fluctuations at the onset of swelling, which is consistent with shell softening. It is important to note that the loss of connectivity alone only changes the response of the shell, and a gradient in the overall energy is necessary in addition to drive swelling. With a soft response, however, only a small energy difference is needed to produce a large conformational change. Our results can thus be interpreted as primarily showing the loss of interactions around the quasi-3-fold vertices resulting in a distinct softening of all capsids tested, whereas subsequent spontaneous swelling depends on small details in protein structure and differs between WT and salt-stable mutant shells. CONCLUSIONS We directly observed the swelling and softening of single CCMV viruses in buffers of elevated pH while chelating divalent cations. These results provide a direct measurement of the physical consequences of the structural transitions that occur in the virus shell. Interestingly, we see that the softening of the shell can occur independently of swelling. Other plant viruses are believed to share the appearance of soft modes, which are not always connected to swelling. This suggests that the physiological role of shell softening may be more important than swelling. Although the precise pathways of CCMV virus transmission from host to host are still unknown, it appears likely that a change of pH could serve as a signal for genome release. One could speculate that this genome release is regulated through the different degrees of capsid breakdown and softening, followed by swelling and then disintegration. It will be particularly interesting in the future to also track the dynamics of the transition and shell fluctuations at intermediate pH conditions and to test whether such fluctuations may already allow RNA release. However, the high stability of CCMV at low pH indicates a need to persist in acidic environments, such as the intestines of insects that are believed to act as transmission vectors (60). From a nano-engineering perspective, investigators could try to modify the change of stiffness and its control parameters by targeted mutations. This might open up promising avenues for the design of drug delivery vehicles. Biophysical Journal 108(10) 2541–2549

2548

SUPPORTING MATERIAL Five figures are available at http://www.biophysj.org/biophysj/supplemental/ S0006-3495(15)00399-9.

AUTHOR CONTRIBUTIONS B.D.W. performed the experiments, evaluated data, and co-wrote the manuscript. I.A.T.S. supported the experiments and data evaluation, performed the simulations, and co-wrote the manuscript. C.F.S. conceived the study, directed the project, and co-wrote the manuscript.

ACKNOWLEDGMENTS We thank Chuck Knobler, Mark Young, and Chris Broomell for providing CCMV virus samples; Stefan Vinzelberg for providing the force mapping software; and Chuck Knobler and Bill Gelbart for helpful discussions. This work was supported by the Cluster of Excellence ‘‘DFG Research Center Nanoscale Microscopy and Molecular Physiology of the Brain’’ (CNMPB).

REFERENCES 1. Douglas, T., and M. Young. 2006. Viruses: making friends with old foes. Science. 312:873–875. 2. Caspar, D. L., and A. Klug. 1962. Physical principles in the construction of regular viruses. Cold Spring Harb. Symp. Quant. Biol. 27:1–24. 3. Fields, B. N., D. M. Knipe, and P. M. Howley. 2007. Fields Virology. Wolter Kluwer Health, Philadelphia. 4. Evilevitch, A., L. Lavelle, ., W. M. Gelbart. 2003. Osmotic pressure inhibition of DNA ejection from phage. Proc. Natl. Acad. Sci. USA. 100:9292–9295. 5. Tzlil, S., J. T. Kindt, ., A. Ben-Shaul. 2003. Forces and pressures in DNA packaging and release from viral capsids. Biophys. J. 84:1616– 1627. 6. Smith, D. E., S. J. Tans, ., C. Bustamante. 2001. The bacteriophage straight phi29 portal motor can package DNA against a large internal force. Nature. 413:748–752. 7. Hernando-Pe´rez, M., R. Miranda, ., P. J. de Pablo. 2012. Direct measurement of phage phi29 stiffness provides evidence of internal pressure. Small. 8:2366–2370. 8. Suomalainen, M., and U. F. Greber. 2013. Uncoating of non-enveloped viruses. Curr Opin Virol. 3:27–33. 9. Wetz, K., and T. Kucinski. 1991. Influence of different ionic and pH environments on structural alterations of poliovirus and their possible relation to virus uncoating. J. Gen. Virol. 72:2541–2544. 10. Fontana, J., G. Cardone, ., A. C. Steven. 2012. Structural changes in Influenza virus at low pH characterized by cryo-electron tomography. J. Virol. 86:2919–2929. 11. Incardona, N. L., and P. Kaesberg. 1964. pH-induced structural change in bromegrass mosaic virus. Biophys. J. 4:11–21. 12. Ivanovska, I. L., P. J. de Pablo, ., G. J. L. Wuite. 2004. Bacteriophage capsids: tough nanoshells with complex elastic properties. Proc. Natl. Acad. Sci. USA. 101:7600–7605. 13. Carrasco, C., A. Carreira, ., P. J. de Pablo. 2006. DNA-mediated anisotropic mechanical reinforcement of a virus. Proc. Natl. Acad. Sci. USA. 103:13706–13711. 14. Carrasco, C., M. Castellanos, ., M. G. Mateu. 2008. Manipulation of the mechanical properties of a virus by protein engineering. Proc. Natl. Acad. Sci. USA. 105:4150–4155. 15. Michel, J. P., I. L. Ivanovska, ., C. F. Schmidt. 2006. Nanoindentation studies of full and empty viral capsids and the effects of capsid protein Biophysical Journal 108(10) 2541–2549

Wilts et al. mutations on elasticity and strength. Proc. Natl. Acad. Sci. USA. 103:6184–6189. 16. Roos, W. H., R. Bruinsma, and G. J. L. Wuite. 2010. Physical virology. Nat. Phys. 6:733–743. 17. Li, S., C. Sieben, ., I. A. T. Schaap. 2014. pH-Controlled two-step uncoating of influenza virus. Biophys. J. 106:1447–1456. 18. Kasas, S., G. Longo, and G. Dietler. 2013. Mechanical properties of biological specimens explored by atomic force microscopy. J. Phys. D Appl. Phys. 46:133001. 19. Douglas, T., and M. Young. 1998. Host-guest encapsulation of materials by assembled virus protein cages. Nature. 393:152–155. 20. Lavelle, L., M. Gingery, ., J. Ruiz-Garcia. 2009. Phase diagram of self-assembled viral capsid protein polymorphs. J. Phys. Chem. B. 113:3813–3819. 21. Lavelle, L., J.-P. Michel, and M. Gingery. 2007. The disassembly, reassembly and stability of CCMV protein capsids. J. Virol. Methods. 146:311–316. 22. Johnson, J. E., and J. A. Speir. 1997. Quasi-equivalent viruses: a paradigm for protein assemblies. J. Mol. Biol. 269:665–675. 23. Johnson, J. M., D. A. Willits, ., A. Zlotnick. 2004. Interaction with capsid protein alters RNA structure and the pathway for in vitro assembly of cowpea chlorotic mottle virus. J. Mol. Biol. 335:455–464. 24. Allison, R. F., M. Janda, and P. Ahlquist. 1989. Sequence of cowpea chlorotic mottle virus RNA-2 and RNA-3 and evidence of a recombination event during bromovirus evolution. Virology. 172:321–330. 25. Dzianott, A. M., and J. J. Bujarski. 1991. The nucleotide sequence and genome organization of the RNA-1 segment in two bromoviruses: broad bean mottle virus and cowpea chlorotic mottle virus. Virology. 185:553–562. 26. Bancroft, J. B., and E. Hiebert. 1967. Formation of an infectious nucleoprotein from protein and nucleic acid isolated from a small spherical virus. Virology. 32:354–356. 27. Bancroft, J. B., E. Hiebert, ., R. Markham. 1968. Properties of cowpea chlorotic mottle virus, its protein and nucleic acid. Virology. 34:224–239. 28. Speir, J. A., S. Munshi, ., J. E. Johnson. 1995. Structures of the native and swollen forms of cowpea chlorotic mottle virus determined by X-ray crystallography and cryo-electron microscopy. Structure. 3:63–78. 29. Vriend, G., M. A. Hemminga, ., T. J. Schaafsma. 1982. Swelling of cowpea chlorotic mottle virus studied by proton nuclear magnetic resonance. FEBS Lett. 146:319–321. 30. Tama, F., and C. L. Brooks, 3rd. 2002. The mechanism and pathway of pH induced swelling in cowpea chlorotic mottle virus. J. Mol. Biol. 318:733–747. 31. Konecny, R., J. Trylska, ., J. A. McCammon. 2006. Electrostatic properties of cowpea chlorotic mottle virus and cucumber mosaic virus capsids. Biopolymers. 82:106–120. 32. Miao, Y., J. E. Johnson, and P. J. Ortoleva. 2010. All-atom multiscale simulation of cowpea chlorotic mottle virus capsid swelling. J. Phys. Chem. B. 114:11181–11195. 33. Robinson, I. K., and S. C. Harrison. 1982. Structure of the expanded state of tomato bushy stunt virus. Nature. 297:563–568. 34. Gergerich, R. C., H. A. Scott, and J. P. Fulton. 1986. Evidence that ribonuclease in beetle regurgitant determines the transmission of plant viruses. J. Gen. Virol. 67:367–370. 35. Bancroft, J. B., M. W. Rees, ., J. R. O. Dawson. 1973. Salt-stable mutant of cowpea chlorotic mottle virus. J. Gen. Virol. 21:507–513. 36. Fox, J. M., X. Zhao, ., M. J. Young. 1996. Analysis of a salt stable mutant of cowpea chlorotic mottle virus. Virology. 222:115–122. 37. Klug, W. S., R. F. Bruinsma, ., G. J. L. Wuite. 2006. Failure of viral shells. Phys. Rev. Lett. 97:228101. 38. Vliegenthart, G. A., and G. Gompper. 2006. Mechanical deformation of spherical viruses with icosahedral symmetry. Biophys. J. 91:834–841.

pH Dependence of CCMV Mechanics

2549

39. Gibbons, M. M., and W. S. Klug. 2008. Influence of nonuniform geometry on nanoindentation of viral capsids. Biophys. J. 95:3640–3649.

50. Li, S., F. Eghiaian, ., I. A. T. Schaap. 2011. Bending and puncturing the influenza lipid envelope. Biophys. J. 100:637–645.

40. Vaughan, R., B. Tragesser, ., C. C. Kao. 2014. The tripartite virions of the brome mosaic virus have distinct physical properties that affect the timing of the infection process. J. Virol. 88:6483–6491.

51. Schaap, I. A. T., C. Carrasco, ., C. F. Schmidt. 2006. Elastic response, buckling, and instability of microtubules under radial indentation. Biophys. J. 91:1521–1531.

41. Cieplak, M., and M. O. Robbins. 2010. Nanoindentation of virus capsids in a molecular model. J. Chem. Phys. 132:015101.

52. Gibbons, M. M., and W. S. Klug. 2007. Nonlinear finite-element analysis of nanoindentation of viral capsids. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 75:031901. 53. Villarrubia, J. S. 1997. Algorithms for scanned probe microscope image simulation, surface reconstruction, and tip estimation. J. Res. Natl. Inst. Stand. Technol. 102:425–454.

42. Kononova, O., J. Snijder, ., V. Barsegov. 2013. Structural transitions and energy landscape for Cowpea Chlorotic Mottle Virus capsid mechanics from nanomanipulation in vitro and in silico. Biophys. J. 105:1893–1903. 43. Schaap, I. A. T., P. J. de Pablo, and C. F. Schmidt. 2004. Resolving the molecular structure of microtubules under physiological conditions with scanning force microscopy. Eur. Biophys. J. 33:462–467. 44. Suci, P. A., M. T. Klem, ., M. Young. 2005. Influence of electrostatic interactions on the surface adsorption of a viral protein cage. Langmuir. 21:8686–8693.

54. Heymann, J. B., D. J. Mu¨ller, ., A. Engel. 1999. Charting the surfaces of the purple membrane. J. Struct. Biol. 128:243–249. 55. Schmitz, S., I. A. Schaap, ., C. Veigel. 2010. Malaria parasite actin polymerization and filament structure. J. Biol. Chem. 285:36577– 36585.

45. Hansma, P. K., J. P. Cleveland, ., V. Elings. 1994. Tapping mode atomic-force microscopy in liquids. Appl. Phys. Lett. 64:1738–1740.

56. Speir, J. A., B. Bothner, ., J. E. Johnson. 2006. Enhanced local symmetry interactions globally stabilize a mutant virus capsid that maintains infectivity and capsid dynamics. J. Virol. 80:3582–3591.

46. Vanderwerf, K. O., C. A. J. Putman, ., J. Greve. 1994. Adhesion force imaging in air and liquid by adhesion mode atomic-force microscopy. Appl. Phys. Lett. 65:1195–1197.

57. Zink, M., and H. Grubmu¨ller. 2010. Primary changes of the mechanical properties of Southern Bean Mosaic Virus upon calcium removal. Biophys. J. 98:687–695.

47. Burnham, N. A., X. Chen, ., S. J. B. Tendler. 2003. Comparison of calibration methods for atomic-force microscopy cantilevers. Nanotechnology. 14:1–6. 48. Hutter, J. L., and J. Bechhoefer. 1993. Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64:1868–1873. 49. Schaap, I. A. T., C. Carrasco, ., C. F. Schmidt. 2011. Kinesin walks the line: single motors observed by atomic force microscopy. Biophys. J. 100:2450–2456.

58. Liu, H., C. Qu, ., D. A. Case. 2003. Pseudo-atomic models of swollen CCMV from cryo-electron microscopy data. J. Struct. Biol. 142:356–363. 59. Gue´rin, T., and R. Bruinsma. 2007. Theory of conformational transitions of viral shells. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 76:061911. 60. Hobbs, H. A., and J. P. Fulton. 1979. Beetle transmission of cowpea chlorotic mottle virus. Phytopathology. 69:255–256.

Biophysical Journal 108(10) 2541–2549