Enhanced Local Symmetry Interactions Globally Stab
http://www.100md.com
病菌学杂志 2006年第7期
Department of Molecular Biology, The Scripps Research Institute, La Jolla, California 92037
Department of Plant Sciences and Plant Pathology, Montana State University, Bozeman, Montana 59717
ABSTRACT
Structural transitions in viral capsids play a critical role in the virus life cycle, including assembly, disassembly, and release of the packaged nucleic acid. Cowpea chlorotic mottle virus (CCMV) undergoes a well-studied reversible structural expansion in vitro in which the capsid expands by 10%. The swollen form of the particle can be completely disassembled by increasing the salt concentration to 1 M. Remarkably, a single-residue mutant of the CCMV N-terminal arm, K42R, is not susceptible to dissociation in high salt (salt-stable CCMV [SS-CCMV]) and retains 70% of wild-type infectivity. We present the combined structural and biophysical basis for the chemical stability and viability of the SS-CCMV particles. A 2.7- resolution crystal structure of the SS-CCMV capsid shows an addition of 660 new intersubunit interactions per particle at the center of the 20 hexameric capsomeres, which are a direct result of the K42R mutation. Protease-based mapping experiments of intact particles demonstrate that both the swollen and closed forms of the wild-type and SS-CCMV particles have highly dynamic N-terminal regions, yet the SS-CCMV particles are more resistant to degradation. Thus, the increase in SS-CCMV particle stability is a result of concentrated tethering of subunits at a local symmetry interface (i.e., quasi-sixfold axes) that does not interfere with the function of other key symmetry interfaces (i.e., fivefold, twofold, quasi-threefold axes). The result is a particle that is still dynamic but insensitive to high salt due to a new series of bonds that are resistant to high ionic strength and preserve the overall particle structure.
INTRODUCTION
Assembly, stability, and disassembly of icosahedral virus particles are coordinated by protein-protein and protein-nucleic acid interactions. Cowpea chlorotic mottle virus (CCMV) provides a model system for examining these interactions. The advantages of this system include in vivo and in vitro assembly assays (4, 7, 12, 55), a reversible transition between distinctly different closed and expanded forms (6), a high-resolution crystal structure of the closed particle (42), an electron cryomicroscopy image reconstruction of the swollen particle (42), and a general understanding of the requirements for assembly and stabilization of the capsid (6, 53, 55, 56).
CCMV is a small RNA plant virus that belongs to the Bromovirus genus in the Bromoviridae family. The viral genome is composed of four (positive-sense) single-stranded RNA molecules that are encapsidated in three morphologically identical particles (8, 21, 27). RNA1 and RNA2 encode proteins involved in RNA-dependent RNA replication and are packaged in separate particles. RNA3, encoding the movement and capsid proteins, and RNA 4, a subgenomic RNA of RNA 3 encoding just the capsid protein, are copackaged in a third particle (2). The structure of the CCMV virion has been determined to a 3.2- resolution (42) revealing a T=3 truncated icosahedron having a diameter of 286 and composed of 180 identical protein subunits arranged as protruding pentamers and hexamers. The 190-amino-acid subunit is folded into an eight-stranded -barrel core with N and C termini that extend to make both numerous intercapsomere (facilitated by subunit dimers) and intracapsomere contacts (Fig. 1). Although not visible in the structure due to the low pH crystallization conditions in combination with the presence of chelating agents in the crystallization buffer, the particle is normally stabilized by intersubunit calcium or magnesium divalent cation binding at the quasi-threefold axes. Also, three to four bases of single-stranded RNA were visible stacking with the Trp47 side chain beneath the divalent cation binding sites. The N-terminal 26 amino acids were not visible in the crystal structure, correlating with previous studies that indicated that this highly basic region projects into the interior region of the virus particle and is required for viral RNA packaging (18, 50, 55).
A dramatic structural transition occurs in response to changes in pH, ionic strength, and divalent cation concentrations that results in swelling of the CCMV capsid volume by 10% (7, 23, 28, 49). In the absence of divalent cations, CCMV particles are stable at low pH (<6.0) and low ionic strength (ionic strength [I] of <0.2). In contrast, the particle dissociates into capsid protein dimers (the assembly unit) and protein-RNA complexes at high pH (>7.0) with increased ionic strength (I > 0.4) and the absence of divalent cations. If the pH is raised above 7.0 under low ionic strength (I < 0.2) conditions, particles rapidly transition to the swollen form. The driving force of the transition is believed to be charge repulsion within the cluster of acidic residues that form the putative divalent cation binding sites. This was partly confirmed with a 28- cryo-electron microscopy image reconstruction of the swollen particle that was fitted with the coordinates from the crystal structure (42). The swollen capsid state retains distinct pentamers and hexamers held together by largely intact dimer contacts, but large 2-nm pores open at each of the quasi-threefold axes with a clear split between the cation binding residues of opposing subunits. This is also observed using hydrogen exchange and mass spectrometry of the closely related brome mosaic bromovirus (BMV) (52). The largest exchange during BMV swelling occurs in regions around the quasi-threefold axes, while regions involved in dimer contacts actually showed decreased exchange, indicating a tightening of the structure around the extended C termini.
A point mutant of CCMV, isolated by nitrous acid treatment of CCMV RNA followed by serial single lesion transfers, does not dissociate at high pH and high ionic strength (1 M NaCl) while retaining about 70% of wild-type (WT) infectivity (9). Bancroft and colleagues determined that there was a single lysine (K)-to-arginine (R) change in the capsid protein based on amino acid analysis. Subsequent work confirmed the single amino acid substitution and identified residue 42 as the mutation site (22). Importantly, the salt-stable (SS) phenotype can be reconstituted by site-specific mutagenesis (K42R) of the WT capsid protein (22). Although modeling with the native CCMV crystal structure suggested that K42R could provide more protein-RNA interactions and create a new intersubunit salt bridge, the lack of a crystal structure of the SS particles prevented specific comparison of the structural characteristics of SS-CCMV relative to WT and rationalization of the new phenotype.
Protein dynamics are an important component of capsid function (for a recent review, see reference 24), and high resolution structural methods provide little data on these processes. Animal viruses have been shown to transiently expose protein domains to the capsid surface that are internal in high-resolution models based on X-ray diffraction data (10, 33). This solution phase motion facilitates the receptor binding and release of the viral genome that are critical to the life cycle of animal viruses (24, 38). Plant viruses are not known to require a dynamic capsid to assist in cell entry and are believed to lack this motion. A limited number of techniques have proven useful in the solution phase study of capsid protein dynamics, including the following: ultrasonic absorption (14), Raman spectroscopy (47), nuclear magnetic resonance (48, 49, 51), site-directed labeling (11), hydrogen-deuterium exchange (29, 30, 52), time-resolved fluorescence (17), and limited proteolysis (15, 37, 46). The last technique has proven to be the most generally applicable, and with the advent of matrix-assisted laser desorption ionization (MALDI) mass spectrometry, cleavage sites can be determined with a high degree of certainty. Thus, the combination of limited proteolysis and peptide mass mapping is a sensitive method for localizing dynamic regions of capsid proteins (10) and can also be used to measure relative differences in dynamic character (11, 32). The combination of structural and mass spectrometry techniques is likely to reveal a greater level of information about the CCMV particle dynamics between different particle states than either method alone.
We apply a multidisciplinary approach to decipher the relationship between the dynamics and structural mechanism of the stabilized SS-CCMV capsid and compare the results to WT-CCMV properties. Our results show that SS-CCMV remains dynamic and retains the ability to swell even though a series of new bonds created by the K42R mutation "cross-link" the six subunits that form hexamers at their N termini. It is surprising that a virus particle with such greatly enhanced stability maintains such a dramatic structural transition. Furthermore, this suggests that the structural transition may be required for CCMV infectivity.
MATERIALS AND METHODS
Virus purification and crystallization. WT and SS viruses were inoculated and propagated in Vigna unguiculata (L.) var. California blackeye as previously described (22). Briefly, a mixture of RNA 1, 2, and 3 (1:1:2) produced by runoff in vitro transcription from cDNA clones was used to inoculate 10-day-old plants by rubbing 5 μg of total RNA on a single leaf. Virus purification and crystallization were performed as previously described (41).
Structure determination and analysis. Diffraction data extending to 2.7 resolution were collected at the Advanced Photon Source beamline 14-BM-C from a single flash-cooled (30% 2-methyl-2,4-pentanediol) crystal. The data were indexed, integrated, merged, scaled, and reduced to a unique reflection set (Table 1) with DENZO and SCALEPACK (35). No reflections were rejected in data processing. The crystals have P212121 space group symmetry (a = 365.5 , b = 374.9 , and c = 402.5 ) and four particles in the unit cell; thus, there is one particle per asymmetric unit with 60-fold noncrystallographic symmetry. This is similar to, but not isomorphous with, crystals of the native virus. The virus particle orientation was determined with locked self- and cross-rotation functions (using native coordinates) in the program GLRF (45), and the particle position (0.251, 0.242, and 0.250) was determined with manual translation of the WT-CCMV coordinates based on particle packing constraints. Initial phases were computed to 3.8- resolution using oriented and translated WT-CCMV coordinates and then refined and extended to 2.7- resolution with CCP4 (16) programs FFT (44), RSTATS, SFALL (1), SFTOOLS, and SIGMAA (39) and the RAVE package (26) using the 60-fold noncrystallographic symmetry for real-space averaging. Electron density quality in the averaged electron density map is excellent, and coordinates for the SS mutant, including R42, were readily built using the WT-CCMV coordinates as a template with the program O (25). The structure was refined against all data using several rounds of torsion molecular dynamics, conjugate gradient minimization, and B-value refinement as implemented in the CNS (crystallography and NMR [nuclear magnetic resonance] system) program (13), followed by further phase refinement. The final rounds of positional and B-value refinement with 3,615 protein atoms and 126 water molecules gave an Rcryst of 24.5% for all data (Table 1).
PROCHECK (31) was used to examine polypeptide geometry. Potential hydrogen bonds and salt bridges were identified with the hydrogen bond tool in Chimera for OS X (36); this program was also used to create figures (Fig. 1A and C; see also Fig. 2A and B). Default length and angle constraints had to be relaxed to identify some of the interactions of R42 and R26. Other distance measurements and root mean square deviation (RMSD) values were calculated using the program O (25).
Swelling assay. Purified virus was dialyzed against swelling buffer (100 mM Tris-HCl, 100 mM NaCl, 1 mM EDTA, 1 mM sodium azide, pH 7.5) or closed buffer (100 mM Tris-HCl, 100 mM NaCl, 5 mM CaCl2, 1 mM sodium azide, pH 7.5) overnight at 4°C. After dialysis, 0.5-mg samples were loaded onto 5 to 25% sucrose gradients having the same final buffer concentrations and centrifuged in an SW41 rotor (Beckman, Fullerton, CA) at 38,000 rpm for 2 h at 4°C. After centrifugation the gradients were fractionated and the absorbance at 260 nm was monitored.
MALDI-TOF. MALDI-time of flight (MALDI-TOF) mass analysis was conducted on a Perceptive Biosystems Voyager Elite equipped with delayed extraction. Reflectron mode was used for ions of <5,000 m/z. Intact capsid protein was cocrystallized with a saturated solution of 1,5-dimethoxy-4-hydroxycinnamic acid in water-acetonitrile-trifluoroascetic acid (50:50:0.05). Protease-released peptides were cocrystallized with a saturated solution of 1-cyano-4-hydroxycinnamic acid.
Protease digestions. WT and SS virus at 10 mg/ml were digested with trypsin (Promega, Madison, WI) or Glu-C (Sigma) in either closed or swelling buffer, as described above. An enzyme-to-virus ratio of 1:1,000 (wt/wt) was used in all reactions. Samples were diluted 10-fold with 10 mM acetic acid to inhibit proteolysis at the specified time intervals and used directly for MALDI-TOF, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, or for visualization by transmission electron microscopy analysis. Protease activity in closed and swelling buffer was tested by monitoring the cleavage rate of the insulin -chain peptide, and no difference was detected. Trypsin and Glu-C activities were equalized using the same peptide, which has three sites for each protease.
Protein structure accession number. The coordinates and structure factors for the SS-CCMV structure determination have been deposited in the Protein Data Bank (http://www.rcsb.org/pdb/) under accession code 1ZA7.
RESULTS
Structure of the SS virion. The chemical basis for the SS phenotype was suggested to be the result of an enhanced protein-RNA interaction and/or the presence of an additional intersubunit salt bridge (22). We determined the crystal structure of SS-CCMV at a 2.7- resolution in an attempt to explain the K42R capsid stabilization. As in the WT structure, the virion is made up of 180 copies of the coat protein arranged with T=3 quasi-symmetry and organized into 20 hexameric and 12 pentameric capsomeres. Each subunit has the characteristic extended N- and C-terminal arms that form an intricate network of ropes that appear to tie the subunits together (Fig. 1). Residues 40 to 190 of the A subunits (which comprise the pentamers) and residues 26 to 190 of the B and C subunits (which comprise the hexamers) were built into well-defined electron density maps averaged over 60-fold noncrystallographic symmetry and refined to a final R factor of 24.5% (Table 1). There are two additional ordered residues on the A subunit N terminus and a single additional ordered residue on the B and C subunit N termini compared to the WT structure. The quality of the electron density was good for R42 in each subunit environment, and the side chain was readily fitted (Fig. 2A). In contrast, electron density for the K42 side chain in the WT structure was disordered past the gamma carbon. Potential RNA electron density beneath W47 in all three subunit environments is present but poorly defined relative to the WT structure. Attempts to build and refine RNA coordinates were abandoned as they resulted in a higher refinement R factor, thermal B factors several times higher than those of nearby protein coordinates, and a poor fit to the density. Other than the lack of RNA coordinates and the point mutation, the mutant subunit structures are closely similar to those of WT, having RMSDs of only 0.4 to 0.6 for the C-atom positions between residues 42 to 190.
The only structural differences between WT- and SS-CCMV involve new interactions created by the K42R mutation. Arginine 42 on the A subunits of SS-CCMV had no new protein interactions compared to WT-CCMV but may have new RNA interactions. The R42 side chain points toward the RNA coordinates fitted in the WT structure, bringing the guanidinium group within 3 of the phosphate backbone and demonstrating that a new salt bridge could be present. The RNA density under the A subunit is the least ordered of the three subunit environments in SS-CCMV, suggesting a weak interaction.
Both B and C subunit R42 residues form new salt bridges and hydrogen bonds around the -hexamer, a small, parallel -cylinder formed at the quasi-sixfold (icosahedral threefold) axes by alternating B and C subunit N termini that run through the centers of the hexamers (42). The salt bridges form between R42(i) and quasi-sixfold related (i + 5) E34 residues (Fig. 2B) (i.e., an ArgB42i-GluC34i + 5 intersubunit interaction). Additionally, both nitrogens at the tip of the R42 guanidinium group hydrogen bond the carbonyl oxygen of R26 from the subunit directly across the quasi-sixfold axis (i + 3) (i.e., an ArgB42i-GluC26i + 3 intersubunit interaction). R26 was not ordered in the B and C subunits of the WT-CCMV crystal structure, which only showed ordered structure up to Val27. In SS-CCMV electron density for R26 is not well defined but is contiguous, and the side chain can be unambiguously oriented and fitted up to the delta carbon in electron density maps contoured at 1.0. Due to the high fidelity of the quasi-sixfold symmetry, the reciprocal interaction also forms between R42 and R26 (i.e., C42i + 3-B26i). The extended N termini of the opposing pairs of B and C subunits in each -hexamer reach completely across the structure, allowing the last ordered residue (R26) to interact with R42 near the beginning of the opposing, quasi-equivalent N-terminal arm.
R26 also interacts with neighboring N termini in the -hexamer (Fig. 2B). On both the B and C subunit N termini, the R26 side chain is oriented inward toward the quasi-sixfold (icosahedral threefold) axis. Hydrogen bonds form with the side chain oxygens of Gln29 on both the icosahedral threefold and quasi-sixfold related subunits (i.e., ArgC26i + 3-GlnC29i + 5 and ArgC26i + 3-GlnB29i + 4) and the main chain carbonyl oxygen of Val27 also of the quasi-sixfold related subunit (i.e., ArgC26i + 3-ValB27i + 4) (Fig. 2B). In the only break of quasi-sixfold symmetry in the new SS-CCMV interactions, R26 of the B subunit does not hydrogen bond the Gln29 side chain oxygen of the quasi-sixfold related C subunit (not shown). This hydrogen bond is not formed due to slightly larger distances and angles between these side chains. As with R42, the R26 interactions are only observed in the SS-CCMV structure although this is the WT residue. Interestingly, the network of new bonds associated with each R42 residue (including R26 and its interactions) involves all but two of the other subunits comprising the -hexamer and effectively circles the quasi-sixfold axis. In total, there are two new salt bridges and nine new hydrogen bonds formed by a B and C subunit pair in the SS-CCMV structure. This amounts to 33 new bonds per -hexamer and 660 new bonds across the entire mutant capsid.
SS capsid swelling. The characteristic swelling of WT-CCMV is driven by repulsion between negatively charged side chains in the absence of divalent cations at high pH. Analysis of the mutant structure revealed that the acidic residues proposed to be important for swelling (42) were present in the same relative positions. Therefore, based on the structure presented here, SS virions would be expected to swell similar to WT given proper conditions. To test whether SS-CCMV does swell, particles in different solution conditions were analyzed by velocity centrifugation through sucrose. Solution conditions that induce swelling in WT-CCMV (absence of divalent cations at high pH) were compared with conditions (presence of divalent cations) in which WT-CCMV particles are closed. Samples were dialyzed against either 100 mM Tris-HCl, 100 mM NaCl, and 1 mM EDTA at pH 7.5 or 100 mM Tris-HCl, 100 mM NaCl, and 5 mM CaCl2 at pH 7.5. Samples were loaded on to 5 to 25% sucrose gradients and centrifuged for 2 h. Gradients were fractionated, and the absorption profile at 260 nm was plotted. Under these conditions, WT-CCMV was present as swollen and closed particles, respectively. WT-CCMV reproducibly migrated at two different rates dependent on the particle form (Fig. 3A), which is consistent with previous reports (6, 7, 42). SS-CCMV also had a distinct form in each buffer condition, both migrating closely similar to comparably treated WT-CCMV (Fig. 3A). In order to confirm that the change in particle form was nearly identical for WT- and SS-CCMV particles, the experiment was repeated after mixing particles from both phenotypes. WT- and SS-CCMV comigrate in both closed and swollen buffer conditions, confirming that the K42R mutant can swell, and produce a swollen particle form with the same sedimentation rate as the swollen form of WT-CCMV (Fig. 3B).
WT and SS capsid protein dynamics. To elucidate the extent of the capsid motion of particles in solution and the effect of the added chemical stability of the SS mutant on this behavior, we initiated a series of protease-based experiments. Both WT-CCMV and SS-CCMV particles were prepared in the closed and swollen forms and incubated with the trypsin and Glu-C proteases. Virus samples were analyzed after 3 and 8 h to determine if the different forms and phenotypes of the particles had an effect on the rate of digestion. All forms and phenotypes of CCMV treated with trypsin, which cleaves at the C-terminal side of the lysine and arginine residues, were readily digested, producing three relatively stable forms of the capsid protein (Fig. 4). Transmission electron micrographs confirmed that the virions remain intact up to 8 h during the protease digestion (data not shown). Interestingly, incubation of all forms and phenotypes of CCMV with the acidic protease Glu-C, which cleaves at the C-terminal side of glutamic and aspartic acid residues, did not visibly degrade the capsid protein. The mutation responsible for the SS phenotype, K42R, does not change the number of potential cleavage sites for either trypsin or Glu-C. Enzymatic activity was tested by monitoring the cleavage rate of a control peptide using trypsin or Glu-C and was not affected by the buffer conditions used for the closed or swollen forms (data not shown). The N terminus of the protein contains 12 basic residues, whereas the C terminus contains 8 acidic residues. Therefore, proteolysis by trypsin and the lack of proteolysis by Glu-C suggested that the extended terminal regions of the capsid protein behave very differently from each other in solution.
The rate at which capsid protein was digested by trypsin was different depending on the form and phenotype of the particles. After 8 h, there was clearly less intact protein present in the WT lanes compared with the SS lanes (Fig. 4). The relative stability of the degradation products was also different between WT and SS particles. Quantitation of the intact capsid protein after 8 h of digestion with trypsin was used to compare the rates of digestion. The intensity of the Coomassie-stained intact capsid protein bands (Fig. 5) demonstrated that the closed and swollen SS-CCMV capsids were 1.7 and 2.4 times more resistant to enzyme digestion than WT-CCMV, respectively. Interestingly, comparing the closed versus swollen forms of the WT particles shows that the closed form is more resistant to digestion (1.4 times), but this difference was much more subtle between the two forms of the SS-CCMV capsids (1.1 times).
Tightly folded protein domains or regions involved in protein-protein interaction are less susceptible to proteolysis than extended or dynamic regions (20). For this reason, we were interested in identifying which parts of the capsid protein were represented in the intact bands visible in the SDS-PAGE analysis (Fig. 4). MALDI-TOF mass spectrometry was used to identify the cleavage products present in the trypsin-treated samples. Spectra representing the mass region containing intact capsid protein for both phenotypes (WT and SS) and forms (closed and swollen) of the virus were consistent with the SDS-PAGE analysis. Intact capsid protein and at least three major cleavage products were present in each sample (Fig. 6). These were identified by mass mapping as residues 15 to 190, 27 to 190, 43 to 190 and 2 to 190 (intact capsid subunit) (Table 2). The closed form of the particles had an additional minor cleavage product present in the mass spectra corresponding to residues 24 to 190 (ion 3).
The SDS-PAGE and protein mass mapping experiments provided comparative data on capsid stability and folded subunit domains, but neither method provided a clear view of the first sites accessed by the protease. To address this question, we applied the mass spectrometry-based technique to the low mass region (<5,000 m/z). Low concentrations of peptides can easily be detected in this region because it is devoid of ions in the undigested samples. This allowed us to map the very first sites of protease cleavage by monitoring the initial minutes of the reactions. Analysis of the released peptides after 1, 5, 10, and 20 min revealed that the pattern was similar for both forms and phenotypes of the virions. MALDI-TOF mass mapping of the released tryptic peptides localized the cleavages to the N terminus of the capsid protein (Fig. 7). The mass mapping experiment was also conducted using Glu-C. As before, no significant degradation of capsid protein was detected. These data confirm that only the N-terminal region of the CCMV capsid protein is accessible to protease in the intact capsid and that the dynamics are retained at a lower level in SS-CCMV.
DISCUSSION
Additional local symmetry contacts globally stabilize the icosahedral particle. Speculation on the physical basis for the increased stability of the SS mutant of CCMV has covered nearly 30 years. We present structural and biochemical data elucidating the basis, and further characteristics, of the SS phenotype. Of particular interest is that the dramatic increase in salt stability is accompanied by increased protease resistance and only has a small effect on infectivity, particle swelling, and virion dynamics. The crystal structure of SS-CCMV establishes that the structural basis for the increased particle stability is the addition of 660 new intersubunit interactions within the 20 -hexamers, which are a direct result of the K42R mutation. The slight additional length and multivalent bonding potential of the R42 side chain allows bivalent hydrogen bonding of the R26 main-chain oxygen, which appears to order and orient R26 to contribute half of the new interactions. The appearance of R26 in SS-CCMV can also be attributed, in part, to experimental improvements such as the higher resolution data collected from a single frozen crystal. For residues R42 and R26 in each hexamer (B or C) subunit, one in every six new interactions is a salt bridge, and the remaining five are a mixture of side chain-to-main chain and side chain-to-side chain hydrogen bonds. Together, the new bonds form an intricate network that "cross-links" the N termini of all six subunits that form hexamers. Moreover, each hexamer subunit is part of a strong dimer interaction that "cross-links" pentamers to hexamers (A-B dimer at quasi-twofold axis) or hexamers to hexamers (C-C dimer at icosahedral twofold axis) (Fig. 1B) via interchange and "clamping" of the subunit C termini (Fig. 2A). The gain in stability from the increased contacts local to the hexamer subunit N termini may be more widely distributed throughout the particle by the intercapsomer interactions of the C termini, which would partly compensate for the lack of any new bonds involving pentamer subunits. The generally accepted strength of hydrogen bonds in water is 0.5 to 1.5 kcal/mol (19, 40). Thus, using a greatly simplified view of this complex bonding structure, where each of the 660 new interactions are all hydrogen bonds and contribute equally to particle stabilization independent of other factors in the surrounding environment, gives a potential net free energy gain of between 330 to 990 kcal/mol for each particle even without contributions from the pentamers. No new interactions are made at any other subunit interface except for the possibility of new protein-RNA interactions by R42 of the pentamer subunits. However, the SS phenotype is also demonstrated by empty particles (22), suggesting that any new RNA interactions made by the pentamers do not contribute significantly to the increased stability. The increase in stability is due to new bonds created only at the quasi-sixfold axes just as retention of WT functions by the stabilized capsid are due to the lack of changes at the fivefold, twofold, quasi-threefold, and quasi-twofold axes.
Stabilization does not prevent particle expansion. Since the structural changes in SS-CCMV are local to the quasi-sixfold axes, they do not preclude the capsid from swelling via expansion of the openings at the quasi-threefold axes. Indeed, sucrose sedimentation velocity experiments reproducibly detected swelling of SS-CCMV particles at pH 7.5 by chelating divalent cations with EDTA. These data demonstrate that the SS-CCMV particles do expand under the same conditions, and to the same extent, as WT particles. The use of EDTA was not cited in two previous studies of SS-CCMV swelling; one concluded that the mutant particles did not expand (3), and the other concluded that the mutant particles did not expand to the same degree as WT when divalent cations were absent (9). It is possible divalent cations remained in reduced concentrations in the previous studies as both used only dialysis to reach the higher pH condition, and this low concentration was sufficient to inhibit swelling.
While the exact steps in CCMV capsid expansion remain unclear, correct dimer formation is required for particle assembly and swelling. Mutational studies have shown that removing the C-terminal arm of the subunit completely abrogates particle assembly (55), and hydrogen exchange studies of the closely similar BMV capsid showed a tightening of the structure in the vicinity of the C-termini as the BMV particle expands (52). Outside the large reorganization at the quasi-threefold axes, both structural and computation studies concur that the transition between the WT closed and swollen states is accommodated by the flexible dimer interface, which remains largely intact and binds the pentamers and hexamers together (34, 42, 43). Portions of the N-terminal arm are also involved in dimer contacts (Fig. 2A), suggesting that alterations to the native, extended dimer interactions can have a measurable impact on particle transitions without altering the dimer interface itself. The subunit contacts within the hexamers and pentamers are also largely unchanged although these capsomers have rotated slightly compared to the native state. In SS-CCMV, the N terminus of each hexamer subunit is involved in new contacts. These additional interactions would seem to favor the closed capsid state under swelling conditions. The N termini are likely to be more strongly anchored in the -hexamer, preventing distortion of the dimer interface, reducing the sampling of favorable subunit orientations, and averting dissolution of the -hexamer to the point necessary to achieve WT capsomer rotation and particle expansion. However, the swollen SS-CCMV particles are indistinguishable from swollen WT particles in sucrose sedimentation experiments when both are exposed to the swelling buffer overnight.
Stabilized particles retain dynamic pentamers. Past structural and biochemical studies indicate that the N-terminal region of the capsid protein is located inside the closed and swollen forms of WT-CCMV and inside the closed form of SS-CCMV (5, 42, 50, 54, 55). However, protease mass mapping experiments reveal that in solution the N terminus of the capsid subunit is readily accessible to trypsin. This agrees well with proteolysis studies of the BMV capsid, which produced primarily N-terminal peptides (52). For the protease to access these N-terminal residues, they must have intermittent exposure to the particle surface, implying a breathing motion for the capsid. The new interactions in SS-CCMV at the hexamer N termini could reduce their exposure at the particle surface and/or restrict the capsid breathing motion. Reductions in N termini exposure or capsid breathing would reduce the level of protease digestion leading to the greater protease resistance of the SS-CCMV particles.
The rate and sites of protease cleavage are a function of protein dynamics, allowing us to compare the solution phase properties of the particles (10). A highly biased distribution of amino acids creates chemically distinct regions and allows the differential probing of protein regions based on protease specificity. Proteases with basic or acidic specificity will, therefore, report primarily on the accessibility of the CCMV coat protein N- or C-terminal regions, respectively. Although SS-CCMV is dramatically stabilized, the sites and apparent early rate of protease cleavage in WT- and SS-CCMV are very similar, indicating that identical regions of the overall capsid and protein subunit are equally dynamic in the two particles (Fig. 6 and 7). All of the primary cleavages are within the first 42 residues of the protein, with one of the primary cleavage sites in both WT- and SS-CCMV being residue 42. Involvement of Arg42 in the new bonding network of the SS-CCMV particle apparently does not significantly reduce the preference for this site by trypsin. Indeed, protease cleavage sites map to position 42 at very early time points in both forms of the capsid. This suggests that the early appearance of this site in SS-CCMV is contributed by subunits within the pentamer. With no changes to the bonding of pentamer N termini, changes in dynamics of this region could indicate that the increased stability of hexamers propagates throughout the particle. This set of N termini can cycle back and forth to the particle surface at the WT rate without affecting the stabilizing interactions centered at the hexamers and be cleaved in the protease mapping experiments at nearly the same rate as WT-CCMV. This defines two sets of dynamic N termini in SS-CCMV. The set located at the center of the pentamers is unaffected by the K42R mutation and makes it appear that the stabilized particles have unaltered dynamics. The set located in the hexamers is contributing to a web of new interactions responsible for the stability and slightly abnormal swelling of the mutant capsids, and this set is probably cleaved more slowly than the equivalent set in WT particles. Alternatively, it may be that only subunits within pentamers are cleaved at the early time points in either type of particle. The difference in the stability of WT- and SS-CCMV would then be due mainly to an overall decrease in SS-CCMV dynamics connected to the increase in hexamer stability. The detailed hydrogen bonding network created by the K42R mutation supports the idea that initially only subunits in pentamers are cleaved.
Although SS-CCMV has an isometric particle structure, its dynamics may be quite asymmetric. The K42R mutation can lead to a clear distinction between capsomer dynamics, where pentamer N termini have a faster surface exchange rate than those of hexamers, or at least more so than those of WT-CCMV. The hexamer N termini in the WT particle have more visibly ordered structure and interactions that could slow their surface exposure compared to those of the pentamers, hinting at partly asymmetric dynamics even in the WT particle. With the additional interactions created by the K42R mutation, the difference in exchange rates for pentamer versus hexamer N termini in SS particles can become even greater, resulting in increased salt stability and protease resistance. Increasingly asymmetric, dynamic movements that act to stabilize a highly symmetric virus particle seem to contradict the structural foundations of the particle itself; however, this is caused by a reduction of the dynamics that make the WT particle more vulnerable to proteases and high-salt conditions. Concurrently, the same reduction in dynamics does not seem to affect overall particle swelling and has little effect on infectivity of the SS-CCMV particle. Dynamic hexamers, at least of the level found in WT particles, appear not to be required for these large structural transitions associated with disassembly.
Thus, the viable increase in particle stability is a result of concentrated tethering of subunits at a highly symmetrical interface (i.e., quasi-sixfold axes) that does not interfere with the function of other key symmetry interfaces (i.e., fivefold, twofold, and quasi-threefold axes). Indeed, the relatively large quaternary rearrangements that lead to the reversible swelling of the capsid appear to have less of an effect on particle dynamics than the K42R point mutation. The effect is a particle that is still dynamic and insensitive to the high-salt conditions but is not disrupted due to a new, interleaving series of bonds that preserve the overall particle structure.
ACKNOWLEDGMENTS
The authors thank Jason Lanman and Lu Gan for insightful discussions, Tianwei Lin for assisting in X-ray data collection, and Gary Siuzdak for providing access to the TSRI Center for Mass Spectrometry.
This study was supported by Public Health Service grant GM54076 from the National Institutes of Health (J.E.J).
J.A.S. and B.B. contributed equally to this work.
REFERENCES
Agarwal, R. C. 1978. A new least-squares refinement technique based on the fast Fourier transform algorithm. Acta Crystallogr. A 34:791-809.
Ahlquist, P. 1992. Bromovirus RNA replication and transcription. Curr. Opin. Genet. Dev. 2:71-76.
Albert, F. G., J. M. Fox, and M. J. Young. 1997. Virion swelling is not required for cotranslational disassembly of cowpea chlorotic mottle virus in vitro. J. Virol. 71:4296-4299.
Allison, R. F., M. Janda, and P. Ahlquist. 1988. Infectious in vitro transcripts from cowpea chlorotic mottle virus cDNA clones and exchange of individual RNA components with brome mosaic virus. J. Virol. 62:3581-3588.
Annamalai, P., S. Apte, S. Wilkens, and A. L. Rao. 2005. Deletion of highly conserved arginine-rich RNA binding motif in cowpea chlorotic mottle virus capsid protein results in virion structural alterations and RNA packaging constraints. J. Virol. 79:3277-3288.
Bancroft, J. B. 1970. The self-assembly of spherical plant viruses. Adv. Virus Res. 16:99-134.
Bancroft, J. B., G. J. Hills, and R. Markham. 1967. A study of the self-assembly process in a small spherical virus: Formation of organized structures from protein subunits in vitro. Virology 31:354-379.
Bancroft, J. B., and R. W. Horne. 1977. Bromovirus (brome mosaic viurs) group, p. 287-302. In K. Maramorosch (ed.), The atlas of insect and plant viruses, vol. 8. Academic Press, New York, N.Y.
Bancroft, J. B., M. W. Rees, M. W. Johnson, and J. R. O. Dawson. 1973. A salt stable mutant of cowpea chlorotic mottle virus. J. Gen. Virol. 21:507-513.
Bothner, B., X. F. Dong, L. Bibbs, J. E. Johnson, and G. Siuzdak. 1998. Evidence of viral capsid dynamics using limited proteolysis and mass spectrometry. J. Biol. Chem. 273:673-676.
Bothner, B., A. Schneemann, D. Marshall, V. Reddy, J. E. Johnson, and G. Siuzdak. 1999. Crystallographically identical virus capsids display different properties in solution. Nat. Struct. Biol. 6:114-116.
Brumfield, S., D. Willits, L. Tang, J. E. Johnson, T. Douglas, and M. Young. 2004. Heterologous expression of the modified coat protein of cowpea chlorotic mottle bromovirus results in the assembly of protein cages with altered architectures and function. J. Gen. Virol. 85:1049-1053.
Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and G. L. Warren. 1998. Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54:905-921.
Cerf, R., B. Michels, J. A. Schulz, J. Witz, P. Pfeiffer, and L. Hirth. 1979. Ultrasonic absorption evidence of structural fluctuations in viral capsids. Proc. Natl. Acad. Sci. USA 76:1780-1782.
Chidlow, J., and J. H. Tremaine. 1971. Limited hydrolysis of cowpea chlorotic mottle virus by trypsin and chymotrypsin. Virology 43:267-278.
Collaborative Computational Project, Number 4. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50:760-763.
Da Poian, A. T., J. E. Johnson, and J. L. Silva. 2002. Protein-RNA interactions and virus stability as probed by the dynamics of tryptophan side chains. J. Biol. Chem. 277:47596-47602.
Douglas, T., E. Strable, D. Willits, A. Aitouchen, M. Libera, and M. Young. 2002. Protein engineering of a viral cage for constrained nanomaterials synthesis. Adv. Mater. 14:415-418.
Fersht, A. R., J. P. Shi, J. Knill-Jones, D. M. Lowe, A. J. Wilkinson, D. M. Blow, P. Brick, P. Carter, M. M. Waye, and G. Winter. 1985. Hydrogen bonding and biological specificity analysed by protein engineering. Nature 314:235-238.
Fontana, A., M. Zambonin, P. Polverino de Laureto, V. De Filippis, A. Clementi, and E. Scaramella. 1997. Probing the conformational state of apomyoglobin by limited proteolysis. J. Mol. Biol. 266:223-230.
Fox, J. M., G. Wang, J. A. Speir, N. H. Olson, J. E. Johnson, T. S. Baker, and M. J. Young. 1998. Comparison of the native CCMV virion with in vitro assembled CCMV virions by cryoelectron microscopy and image reconstruction. Virology 244:212-218.
Fox, J. M., X. Zhao, J. A. Speir, and M. J. Young. 1996. Analysis of a salt stable mutant of cowpea chlorotic mottle virus. Virology 222:115-122.
Jacrot, B. 1975. Studies on the assembly of a spherical virus. II. The mechanism of protein aggregation and virus swelling. J. Mol. Biol. 95:433-446.
Johnson, J. E. 2003. Virus particle dynamics. Adv. Protein Chem. 64:197-218.
Jones, T. A., J. Y. Zou, S. W. Cowan, and M. Kjeldgaard. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A. 47:110-119.
Kleywegt, G. J., and T. A. Jones. 1999. Software for handling macromolecular envelopes. Acta Crystallogr. D 55:941-944.
Lane, L. C. 1981. Bromovirus, p. 333-336. In E. Kurstak (ed.), Handbook of plant virus infection and comparative diagnosis. Elsevier/North-Holland, Amsterdam, The Netherlands.
Lane, L. C. 1974. The bromoviruses. Adv. Virus Res. 19:151-220.
Lanman, J., T. T. Lam, M. R. Emmett, A. G. Marshall, M. Sakalian, and P. E. Prevelige, Jr. 2004. Key interactions in HIV-1 maturation identified by hydrogen-deuterium exchange. Nat. Struct. Mol. Biol. 11:676-677.
Lanman, J., and P. E. Prevelige, Jr. 2004. High-sensitivity mass spectrometry for imaging subunit interactions: hydrogen/deuterium exchange. Curr. Opin. Struct. Biol. 14:181-188.
Laskowski, R. A., M. W. MacArthur, D. S. Moss, and J. M. Thornton. 1993. PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26:283-291.
Lewis, J. K., B. Bothner, T. J. Smith, and G. Siuzdak. 1998. Antiviral agent blocks breathing of the common cold virus. Proc. Natl. Acad. Sci. USA 95:6774-6778.
Li, Q., A. G. Yafal, Y. M. Lee, J. Hogle, and M. Chow. 1994. Poliovirus neutralization by antibodies to internal epitopes of VP4 and VP1 results from reversible exposure of these sequences at physiological temperature. J. Virol. 68:3965-3970.
Liu, H., C. Qu, J. E. Johnson, and D. A. Case. 2003. Pseudo-atomic models of swollen CCMV from cryo-electron microscopy data. J. Struct. Biol. 142:356-363.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276:307-326.
Pettersen, E. F., T. D. Goddard, C. C. Huang, G. S. Couch, D. M. Greenblatt, E. C. Meng, and T. E. Ferrin. 2004. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25:1605-1612.
Pfeiffer, P. 1980. Changes in the organization of bromegrass mosaic virus in response to cation binding as probed by changes in susceptibility to degradative enzymes. Virology 102:54-61.
Phelps, D. K., B. Speelman, and C. B. Post. 2000. Theoretical studies of viral capsid proteins. Curr. Opin. Struct. Biol. 10:170-173.
Read, R. J. 1986. Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallogr. A 42:140-149.
Sheu, S. Y., D. Y. Yang, H. L. Selzle, and E. W. Schlag. 2003. Energetics of hydrogen bonds in peptides. Proc. Natl. Acad. Sci. USA 100:12683-12687.
Speir, J. A., S. Munshi, T. S. Baker, and J. E. Johnson. 1993. Preliminary X-ray data analysis of crystalline cowpea chlorotic mottle virus. Virology 193:234-241.
Speir, J. A., S. Munshi, G. Wang, T. S. Baker, and 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.
Tama, F., and C. L. Brooks III. 2002. The mechanism and pathway of pH-induced swelling in cowpea chlorotic mottle virus. J. Mol. Biol. 318:733-747.
TenEyck, L. F. 1973. Crystallographic fast Fourier transforms. Acta Crystallogr. A 29:183-191.
Tong, L., and M. G. Rossmann. 1997. Rotation function calculations with GLRF program. Methods Enzymol. 276:594-611.
Tremaine, J. H., and H. O. Agrawal. 1972. Limited proteolysis of potato virus X by trypsin and plant proteases. Virology 49:735-744.
Verduin, B. J., B. Prescott, and G. J. Thomas, Jr. 1984. RNA-protein interactions and secondary structures of cowpea chlorotic mottle virus for in vitro assembly. Biochemistry 23:4301-4308.
Vriend, G., M. A. Hemminga, B. J. M. Verduin, J. L. De Wit, and T. J. Schaafsma. 1981. Segmental mobility involved in protein-RNA interaction in cowpea chlorotic mottle virus. FEBS Lett. 134:167-171.
Vriend, G., M. A. Hemminga, B. J. M. Verduin, and T. J. Schaafsma. 1982. Swelling of cowpea chlorotic mottle virus studied by proton nuclear magnetic resonance. FEBS Lett. 146:319-321.
Vriend, G., B. J. M. Verduin, and M. A. Hemminga. 1986. Role of the N-terminal part of the coat protein in the assembly of cowpea chlorotic mottle virus. A 500 MHz proton nuclear magnetic resonance study and structural calculations. J. Mol. Biol. 191:453-460.
Vriend, G., B. J. M. Verduin, M. A. Hemminga, and T. J. Schaafsma. 1982. Mobility involved in protein-RNA interaction in spherical plant viruses, studied by nuclear magnetic resonance spectroscopy. FEBS Lett. 145:49-52.
Wang, L., L. C. Lane, and D. L. Smith. 2001. Detecting structural changes in viral capsids by hydrogen exchange and mass spectrometry. Protein Sci. 10:1234-1243.
Willits, D., X. Zhao, N. Olson, T. S. Baker, A. Zlotnick, J. E. Johnson, T. Douglas, and M. J. Young. 2003. Effects of the cowpea chlorotic mottle bromovirus beta-hexamer structure on virion assembly. Virology 306:280-288.
Wolfs, C. J., M. van der Graaf, and M. A. Hemminga. 1993. Conformational changes in oligo-nucleotides upon binding to a peptide representing the N-terminal region of a viral coat protein. An optical spectroscopy study. Biochim. Biophys. Acta 1172:1-4.
Zhao, X., J. M. Fox, N. H. Olson, T. S. Baker, and M. J. Young. 1995. In vitro assembly of cowpea chlorotic mottle virus from coat protein expressed in Escherichia coli and in vitro transcribed viral cDNA. Virology 207:486-494.
Zlotnick, A., R. Aldrich, J. M. Johnson, P. Ceres, and M. J. Young. 2000. Mechanism of capsid assembly for an icosahedral plant virus. Virology 277:450-456.(Jeffrey A. Speir, Brian B)
Department of Plant Sciences and Plant Pathology, Montana State University, Bozeman, Montana 59717
ABSTRACT
Structural transitions in viral capsids play a critical role in the virus life cycle, including assembly, disassembly, and release of the packaged nucleic acid. Cowpea chlorotic mottle virus (CCMV) undergoes a well-studied reversible structural expansion in vitro in which the capsid expands by 10%. The swollen form of the particle can be completely disassembled by increasing the salt concentration to 1 M. Remarkably, a single-residue mutant of the CCMV N-terminal arm, K42R, is not susceptible to dissociation in high salt (salt-stable CCMV [SS-CCMV]) and retains 70% of wild-type infectivity. We present the combined structural and biophysical basis for the chemical stability and viability of the SS-CCMV particles. A 2.7- resolution crystal structure of the SS-CCMV capsid shows an addition of 660 new intersubunit interactions per particle at the center of the 20 hexameric capsomeres, which are a direct result of the K42R mutation. Protease-based mapping experiments of intact particles demonstrate that both the swollen and closed forms of the wild-type and SS-CCMV particles have highly dynamic N-terminal regions, yet the SS-CCMV particles are more resistant to degradation. Thus, the increase in SS-CCMV particle stability is a result of concentrated tethering of subunits at a local symmetry interface (i.e., quasi-sixfold axes) that does not interfere with the function of other key symmetry interfaces (i.e., fivefold, twofold, quasi-threefold axes). The result is a particle that is still dynamic but insensitive to high salt due to a new series of bonds that are resistant to high ionic strength and preserve the overall particle structure.
INTRODUCTION
Assembly, stability, and disassembly of icosahedral virus particles are coordinated by protein-protein and protein-nucleic acid interactions. Cowpea chlorotic mottle virus (CCMV) provides a model system for examining these interactions. The advantages of this system include in vivo and in vitro assembly assays (4, 7, 12, 55), a reversible transition between distinctly different closed and expanded forms (6), a high-resolution crystal structure of the closed particle (42), an electron cryomicroscopy image reconstruction of the swollen particle (42), and a general understanding of the requirements for assembly and stabilization of the capsid (6, 53, 55, 56).
CCMV is a small RNA plant virus that belongs to the Bromovirus genus in the Bromoviridae family. The viral genome is composed of four (positive-sense) single-stranded RNA molecules that are encapsidated in three morphologically identical particles (8, 21, 27). RNA1 and RNA2 encode proteins involved in RNA-dependent RNA replication and are packaged in separate particles. RNA3, encoding the movement and capsid proteins, and RNA 4, a subgenomic RNA of RNA 3 encoding just the capsid protein, are copackaged in a third particle (2). The structure of the CCMV virion has been determined to a 3.2- resolution (42) revealing a T=3 truncated icosahedron having a diameter of 286 and composed of 180 identical protein subunits arranged as protruding pentamers and hexamers. The 190-amino-acid subunit is folded into an eight-stranded -barrel core with N and C termini that extend to make both numerous intercapsomere (facilitated by subunit dimers) and intracapsomere contacts (Fig. 1). Although not visible in the structure due to the low pH crystallization conditions in combination with the presence of chelating agents in the crystallization buffer, the particle is normally stabilized by intersubunit calcium or magnesium divalent cation binding at the quasi-threefold axes. Also, three to four bases of single-stranded RNA were visible stacking with the Trp47 side chain beneath the divalent cation binding sites. The N-terminal 26 amino acids were not visible in the crystal structure, correlating with previous studies that indicated that this highly basic region projects into the interior region of the virus particle and is required for viral RNA packaging (18, 50, 55).
A dramatic structural transition occurs in response to changes in pH, ionic strength, and divalent cation concentrations that results in swelling of the CCMV capsid volume by 10% (7, 23, 28, 49). In the absence of divalent cations, CCMV particles are stable at low pH (<6.0) and low ionic strength (ionic strength [I] of <0.2). In contrast, the particle dissociates into capsid protein dimers (the assembly unit) and protein-RNA complexes at high pH (>7.0) with increased ionic strength (I > 0.4) and the absence of divalent cations. If the pH is raised above 7.0 under low ionic strength (I < 0.2) conditions, particles rapidly transition to the swollen form. The driving force of the transition is believed to be charge repulsion within the cluster of acidic residues that form the putative divalent cation binding sites. This was partly confirmed with a 28- cryo-electron microscopy image reconstruction of the swollen particle that was fitted with the coordinates from the crystal structure (42). The swollen capsid state retains distinct pentamers and hexamers held together by largely intact dimer contacts, but large 2-nm pores open at each of the quasi-threefold axes with a clear split between the cation binding residues of opposing subunits. This is also observed using hydrogen exchange and mass spectrometry of the closely related brome mosaic bromovirus (BMV) (52). The largest exchange during BMV swelling occurs in regions around the quasi-threefold axes, while regions involved in dimer contacts actually showed decreased exchange, indicating a tightening of the structure around the extended C termini.
A point mutant of CCMV, isolated by nitrous acid treatment of CCMV RNA followed by serial single lesion transfers, does not dissociate at high pH and high ionic strength (1 M NaCl) while retaining about 70% of wild-type (WT) infectivity (9). Bancroft and colleagues determined that there was a single lysine (K)-to-arginine (R) change in the capsid protein based on amino acid analysis. Subsequent work confirmed the single amino acid substitution and identified residue 42 as the mutation site (22). Importantly, the salt-stable (SS) phenotype can be reconstituted by site-specific mutagenesis (K42R) of the WT capsid protein (22). Although modeling with the native CCMV crystal structure suggested that K42R could provide more protein-RNA interactions and create a new intersubunit salt bridge, the lack of a crystal structure of the SS particles prevented specific comparison of the structural characteristics of SS-CCMV relative to WT and rationalization of the new phenotype.
Protein dynamics are an important component of capsid function (for a recent review, see reference 24), and high resolution structural methods provide little data on these processes. Animal viruses have been shown to transiently expose protein domains to the capsid surface that are internal in high-resolution models based on X-ray diffraction data (10, 33). This solution phase motion facilitates the receptor binding and release of the viral genome that are critical to the life cycle of animal viruses (24, 38). Plant viruses are not known to require a dynamic capsid to assist in cell entry and are believed to lack this motion. A limited number of techniques have proven useful in the solution phase study of capsid protein dynamics, including the following: ultrasonic absorption (14), Raman spectroscopy (47), nuclear magnetic resonance (48, 49, 51), site-directed labeling (11), hydrogen-deuterium exchange (29, 30, 52), time-resolved fluorescence (17), and limited proteolysis (15, 37, 46). The last technique has proven to be the most generally applicable, and with the advent of matrix-assisted laser desorption ionization (MALDI) mass spectrometry, cleavage sites can be determined with a high degree of certainty. Thus, the combination of limited proteolysis and peptide mass mapping is a sensitive method for localizing dynamic regions of capsid proteins (10) and can also be used to measure relative differences in dynamic character (11, 32). The combination of structural and mass spectrometry techniques is likely to reveal a greater level of information about the CCMV particle dynamics between different particle states than either method alone.
We apply a multidisciplinary approach to decipher the relationship between the dynamics and structural mechanism of the stabilized SS-CCMV capsid and compare the results to WT-CCMV properties. Our results show that SS-CCMV remains dynamic and retains the ability to swell even though a series of new bonds created by the K42R mutation "cross-link" the six subunits that form hexamers at their N termini. It is surprising that a virus particle with such greatly enhanced stability maintains such a dramatic structural transition. Furthermore, this suggests that the structural transition may be required for CCMV infectivity.
MATERIALS AND METHODS
Virus purification and crystallization. WT and SS viruses were inoculated and propagated in Vigna unguiculata (L.) var. California blackeye as previously described (22). Briefly, a mixture of RNA 1, 2, and 3 (1:1:2) produced by runoff in vitro transcription from cDNA clones was used to inoculate 10-day-old plants by rubbing 5 μg of total RNA on a single leaf. Virus purification and crystallization were performed as previously described (41).
Structure determination and analysis. Diffraction data extending to 2.7 resolution were collected at the Advanced Photon Source beamline 14-BM-C from a single flash-cooled (30% 2-methyl-2,4-pentanediol) crystal. The data were indexed, integrated, merged, scaled, and reduced to a unique reflection set (Table 1) with DENZO and SCALEPACK (35). No reflections were rejected in data processing. The crystals have P212121 space group symmetry (a = 365.5 , b = 374.9 , and c = 402.5 ) and four particles in the unit cell; thus, there is one particle per asymmetric unit with 60-fold noncrystallographic symmetry. This is similar to, but not isomorphous with, crystals of the native virus. The virus particle orientation was determined with locked self- and cross-rotation functions (using native coordinates) in the program GLRF (45), and the particle position (0.251, 0.242, and 0.250) was determined with manual translation of the WT-CCMV coordinates based on particle packing constraints. Initial phases were computed to 3.8- resolution using oriented and translated WT-CCMV coordinates and then refined and extended to 2.7- resolution with CCP4 (16) programs FFT (44), RSTATS, SFALL (1), SFTOOLS, and SIGMAA (39) and the RAVE package (26) using the 60-fold noncrystallographic symmetry for real-space averaging. Electron density quality in the averaged electron density map is excellent, and coordinates for the SS mutant, including R42, were readily built using the WT-CCMV coordinates as a template with the program O (25). The structure was refined against all data using several rounds of torsion molecular dynamics, conjugate gradient minimization, and B-value refinement as implemented in the CNS (crystallography and NMR [nuclear magnetic resonance] system) program (13), followed by further phase refinement. The final rounds of positional and B-value refinement with 3,615 protein atoms and 126 water molecules gave an Rcryst of 24.5% for all data (Table 1).
PROCHECK (31) was used to examine polypeptide geometry. Potential hydrogen bonds and salt bridges were identified with the hydrogen bond tool in Chimera for OS X (36); this program was also used to create figures (Fig. 1A and C; see also Fig. 2A and B). Default length and angle constraints had to be relaxed to identify some of the interactions of R42 and R26. Other distance measurements and root mean square deviation (RMSD) values were calculated using the program O (25).
Swelling assay. Purified virus was dialyzed against swelling buffer (100 mM Tris-HCl, 100 mM NaCl, 1 mM EDTA, 1 mM sodium azide, pH 7.5) or closed buffer (100 mM Tris-HCl, 100 mM NaCl, 5 mM CaCl2, 1 mM sodium azide, pH 7.5) overnight at 4°C. After dialysis, 0.5-mg samples were loaded onto 5 to 25% sucrose gradients having the same final buffer concentrations and centrifuged in an SW41 rotor (Beckman, Fullerton, CA) at 38,000 rpm for 2 h at 4°C. After centrifugation the gradients were fractionated and the absorbance at 260 nm was monitored.
MALDI-TOF. MALDI-time of flight (MALDI-TOF) mass analysis was conducted on a Perceptive Biosystems Voyager Elite equipped with delayed extraction. Reflectron mode was used for ions of <5,000 m/z. Intact capsid protein was cocrystallized with a saturated solution of 1,5-dimethoxy-4-hydroxycinnamic acid in water-acetonitrile-trifluoroascetic acid (50:50:0.05). Protease-released peptides were cocrystallized with a saturated solution of 1-cyano-4-hydroxycinnamic acid.
Protease digestions. WT and SS virus at 10 mg/ml were digested with trypsin (Promega, Madison, WI) or Glu-C (Sigma) in either closed or swelling buffer, as described above. An enzyme-to-virus ratio of 1:1,000 (wt/wt) was used in all reactions. Samples were diluted 10-fold with 10 mM acetic acid to inhibit proteolysis at the specified time intervals and used directly for MALDI-TOF, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, or for visualization by transmission electron microscopy analysis. Protease activity in closed and swelling buffer was tested by monitoring the cleavage rate of the insulin -chain peptide, and no difference was detected. Trypsin and Glu-C activities were equalized using the same peptide, which has three sites for each protease.
Protein structure accession number. The coordinates and structure factors for the SS-CCMV structure determination have been deposited in the Protein Data Bank (http://www.rcsb.org/pdb/) under accession code 1ZA7.
RESULTS
Structure of the SS virion. The chemical basis for the SS phenotype was suggested to be the result of an enhanced protein-RNA interaction and/or the presence of an additional intersubunit salt bridge (22). We determined the crystal structure of SS-CCMV at a 2.7- resolution in an attempt to explain the K42R capsid stabilization. As in the WT structure, the virion is made up of 180 copies of the coat protein arranged with T=3 quasi-symmetry and organized into 20 hexameric and 12 pentameric capsomeres. Each subunit has the characteristic extended N- and C-terminal arms that form an intricate network of ropes that appear to tie the subunits together (Fig. 1). Residues 40 to 190 of the A subunits (which comprise the pentamers) and residues 26 to 190 of the B and C subunits (which comprise the hexamers) were built into well-defined electron density maps averaged over 60-fold noncrystallographic symmetry and refined to a final R factor of 24.5% (Table 1). There are two additional ordered residues on the A subunit N terminus and a single additional ordered residue on the B and C subunit N termini compared to the WT structure. The quality of the electron density was good for R42 in each subunit environment, and the side chain was readily fitted (Fig. 2A). In contrast, electron density for the K42 side chain in the WT structure was disordered past the gamma carbon. Potential RNA electron density beneath W47 in all three subunit environments is present but poorly defined relative to the WT structure. Attempts to build and refine RNA coordinates were abandoned as they resulted in a higher refinement R factor, thermal B factors several times higher than those of nearby protein coordinates, and a poor fit to the density. Other than the lack of RNA coordinates and the point mutation, the mutant subunit structures are closely similar to those of WT, having RMSDs of only 0.4 to 0.6 for the C-atom positions between residues 42 to 190.
The only structural differences between WT- and SS-CCMV involve new interactions created by the K42R mutation. Arginine 42 on the A subunits of SS-CCMV had no new protein interactions compared to WT-CCMV but may have new RNA interactions. The R42 side chain points toward the RNA coordinates fitted in the WT structure, bringing the guanidinium group within 3 of the phosphate backbone and demonstrating that a new salt bridge could be present. The RNA density under the A subunit is the least ordered of the three subunit environments in SS-CCMV, suggesting a weak interaction.
Both B and C subunit R42 residues form new salt bridges and hydrogen bonds around the -hexamer, a small, parallel -cylinder formed at the quasi-sixfold (icosahedral threefold) axes by alternating B and C subunit N termini that run through the centers of the hexamers (42). The salt bridges form between R42(i) and quasi-sixfold related (i + 5) E34 residues (Fig. 2B) (i.e., an ArgB42i-GluC34i + 5 intersubunit interaction). Additionally, both nitrogens at the tip of the R42 guanidinium group hydrogen bond the carbonyl oxygen of R26 from the subunit directly across the quasi-sixfold axis (i + 3) (i.e., an ArgB42i-GluC26i + 3 intersubunit interaction). R26 was not ordered in the B and C subunits of the WT-CCMV crystal structure, which only showed ordered structure up to Val27. In SS-CCMV electron density for R26 is not well defined but is contiguous, and the side chain can be unambiguously oriented and fitted up to the delta carbon in electron density maps contoured at 1.0. Due to the high fidelity of the quasi-sixfold symmetry, the reciprocal interaction also forms between R42 and R26 (i.e., C42i + 3-B26i). The extended N termini of the opposing pairs of B and C subunits in each -hexamer reach completely across the structure, allowing the last ordered residue (R26) to interact with R42 near the beginning of the opposing, quasi-equivalent N-terminal arm.
R26 also interacts with neighboring N termini in the -hexamer (Fig. 2B). On both the B and C subunit N termini, the R26 side chain is oriented inward toward the quasi-sixfold (icosahedral threefold) axis. Hydrogen bonds form with the side chain oxygens of Gln29 on both the icosahedral threefold and quasi-sixfold related subunits (i.e., ArgC26i + 3-GlnC29i + 5 and ArgC26i + 3-GlnB29i + 4) and the main chain carbonyl oxygen of Val27 also of the quasi-sixfold related subunit (i.e., ArgC26i + 3-ValB27i + 4) (Fig. 2B). In the only break of quasi-sixfold symmetry in the new SS-CCMV interactions, R26 of the B subunit does not hydrogen bond the Gln29 side chain oxygen of the quasi-sixfold related C subunit (not shown). This hydrogen bond is not formed due to slightly larger distances and angles between these side chains. As with R42, the R26 interactions are only observed in the SS-CCMV structure although this is the WT residue. Interestingly, the network of new bonds associated with each R42 residue (including R26 and its interactions) involves all but two of the other subunits comprising the -hexamer and effectively circles the quasi-sixfold axis. In total, there are two new salt bridges and nine new hydrogen bonds formed by a B and C subunit pair in the SS-CCMV structure. This amounts to 33 new bonds per -hexamer and 660 new bonds across the entire mutant capsid.
SS capsid swelling. The characteristic swelling of WT-CCMV is driven by repulsion between negatively charged side chains in the absence of divalent cations at high pH. Analysis of the mutant structure revealed that the acidic residues proposed to be important for swelling (42) were present in the same relative positions. Therefore, based on the structure presented here, SS virions would be expected to swell similar to WT given proper conditions. To test whether SS-CCMV does swell, particles in different solution conditions were analyzed by velocity centrifugation through sucrose. Solution conditions that induce swelling in WT-CCMV (absence of divalent cations at high pH) were compared with conditions (presence of divalent cations) in which WT-CCMV particles are closed. Samples were dialyzed against either 100 mM Tris-HCl, 100 mM NaCl, and 1 mM EDTA at pH 7.5 or 100 mM Tris-HCl, 100 mM NaCl, and 5 mM CaCl2 at pH 7.5. Samples were loaded on to 5 to 25% sucrose gradients and centrifuged for 2 h. Gradients were fractionated, and the absorption profile at 260 nm was plotted. Under these conditions, WT-CCMV was present as swollen and closed particles, respectively. WT-CCMV reproducibly migrated at two different rates dependent on the particle form (Fig. 3A), which is consistent with previous reports (6, 7, 42). SS-CCMV also had a distinct form in each buffer condition, both migrating closely similar to comparably treated WT-CCMV (Fig. 3A). In order to confirm that the change in particle form was nearly identical for WT- and SS-CCMV particles, the experiment was repeated after mixing particles from both phenotypes. WT- and SS-CCMV comigrate in both closed and swollen buffer conditions, confirming that the K42R mutant can swell, and produce a swollen particle form with the same sedimentation rate as the swollen form of WT-CCMV (Fig. 3B).
WT and SS capsid protein dynamics. To elucidate the extent of the capsid motion of particles in solution and the effect of the added chemical stability of the SS mutant on this behavior, we initiated a series of protease-based experiments. Both WT-CCMV and SS-CCMV particles were prepared in the closed and swollen forms and incubated with the trypsin and Glu-C proteases. Virus samples were analyzed after 3 and 8 h to determine if the different forms and phenotypes of the particles had an effect on the rate of digestion. All forms and phenotypes of CCMV treated with trypsin, which cleaves at the C-terminal side of the lysine and arginine residues, were readily digested, producing three relatively stable forms of the capsid protein (Fig. 4). Transmission electron micrographs confirmed that the virions remain intact up to 8 h during the protease digestion (data not shown). Interestingly, incubation of all forms and phenotypes of CCMV with the acidic protease Glu-C, which cleaves at the C-terminal side of glutamic and aspartic acid residues, did not visibly degrade the capsid protein. The mutation responsible for the SS phenotype, K42R, does not change the number of potential cleavage sites for either trypsin or Glu-C. Enzymatic activity was tested by monitoring the cleavage rate of a control peptide using trypsin or Glu-C and was not affected by the buffer conditions used for the closed or swollen forms (data not shown). The N terminus of the protein contains 12 basic residues, whereas the C terminus contains 8 acidic residues. Therefore, proteolysis by trypsin and the lack of proteolysis by Glu-C suggested that the extended terminal regions of the capsid protein behave very differently from each other in solution.
The rate at which capsid protein was digested by trypsin was different depending on the form and phenotype of the particles. After 8 h, there was clearly less intact protein present in the WT lanes compared with the SS lanes (Fig. 4). The relative stability of the degradation products was also different between WT and SS particles. Quantitation of the intact capsid protein after 8 h of digestion with trypsin was used to compare the rates of digestion. The intensity of the Coomassie-stained intact capsid protein bands (Fig. 5) demonstrated that the closed and swollen SS-CCMV capsids were 1.7 and 2.4 times more resistant to enzyme digestion than WT-CCMV, respectively. Interestingly, comparing the closed versus swollen forms of the WT particles shows that the closed form is more resistant to digestion (1.4 times), but this difference was much more subtle between the two forms of the SS-CCMV capsids (1.1 times).
Tightly folded protein domains or regions involved in protein-protein interaction are less susceptible to proteolysis than extended or dynamic regions (20). For this reason, we were interested in identifying which parts of the capsid protein were represented in the intact bands visible in the SDS-PAGE analysis (Fig. 4). MALDI-TOF mass spectrometry was used to identify the cleavage products present in the trypsin-treated samples. Spectra representing the mass region containing intact capsid protein for both phenotypes (WT and SS) and forms (closed and swollen) of the virus were consistent with the SDS-PAGE analysis. Intact capsid protein and at least three major cleavage products were present in each sample (Fig. 6). These were identified by mass mapping as residues 15 to 190, 27 to 190, 43 to 190 and 2 to 190 (intact capsid subunit) (Table 2). The closed form of the particles had an additional minor cleavage product present in the mass spectra corresponding to residues 24 to 190 (ion 3).
The SDS-PAGE and protein mass mapping experiments provided comparative data on capsid stability and folded subunit domains, but neither method provided a clear view of the first sites accessed by the protease. To address this question, we applied the mass spectrometry-based technique to the low mass region (<5,000 m/z). Low concentrations of peptides can easily be detected in this region because it is devoid of ions in the undigested samples. This allowed us to map the very first sites of protease cleavage by monitoring the initial minutes of the reactions. Analysis of the released peptides after 1, 5, 10, and 20 min revealed that the pattern was similar for both forms and phenotypes of the virions. MALDI-TOF mass mapping of the released tryptic peptides localized the cleavages to the N terminus of the capsid protein (Fig. 7). The mass mapping experiment was also conducted using Glu-C. As before, no significant degradation of capsid protein was detected. These data confirm that only the N-terminal region of the CCMV capsid protein is accessible to protease in the intact capsid and that the dynamics are retained at a lower level in SS-CCMV.
DISCUSSION
Additional local symmetry contacts globally stabilize the icosahedral particle. Speculation on the physical basis for the increased stability of the SS mutant of CCMV has covered nearly 30 years. We present structural and biochemical data elucidating the basis, and further characteristics, of the SS phenotype. Of particular interest is that the dramatic increase in salt stability is accompanied by increased protease resistance and only has a small effect on infectivity, particle swelling, and virion dynamics. The crystal structure of SS-CCMV establishes that the structural basis for the increased particle stability is the addition of 660 new intersubunit interactions within the 20 -hexamers, which are a direct result of the K42R mutation. The slight additional length and multivalent bonding potential of the R42 side chain allows bivalent hydrogen bonding of the R26 main-chain oxygen, which appears to order and orient R26 to contribute half of the new interactions. The appearance of R26 in SS-CCMV can also be attributed, in part, to experimental improvements such as the higher resolution data collected from a single frozen crystal. For residues R42 and R26 in each hexamer (B or C) subunit, one in every six new interactions is a salt bridge, and the remaining five are a mixture of side chain-to-main chain and side chain-to-side chain hydrogen bonds. Together, the new bonds form an intricate network that "cross-links" the N termini of all six subunits that form hexamers. Moreover, each hexamer subunit is part of a strong dimer interaction that "cross-links" pentamers to hexamers (A-B dimer at quasi-twofold axis) or hexamers to hexamers (C-C dimer at icosahedral twofold axis) (Fig. 1B) via interchange and "clamping" of the subunit C termini (Fig. 2A). The gain in stability from the increased contacts local to the hexamer subunit N termini may be more widely distributed throughout the particle by the intercapsomer interactions of the C termini, which would partly compensate for the lack of any new bonds involving pentamer subunits. The generally accepted strength of hydrogen bonds in water is 0.5 to 1.5 kcal/mol (19, 40). Thus, using a greatly simplified view of this complex bonding structure, where each of the 660 new interactions are all hydrogen bonds and contribute equally to particle stabilization independent of other factors in the surrounding environment, gives a potential net free energy gain of between 330 to 990 kcal/mol for each particle even without contributions from the pentamers. No new interactions are made at any other subunit interface except for the possibility of new protein-RNA interactions by R42 of the pentamer subunits. However, the SS phenotype is also demonstrated by empty particles (22), suggesting that any new RNA interactions made by the pentamers do not contribute significantly to the increased stability. The increase in stability is due to new bonds created only at the quasi-sixfold axes just as retention of WT functions by the stabilized capsid are due to the lack of changes at the fivefold, twofold, quasi-threefold, and quasi-twofold axes.
Stabilization does not prevent particle expansion. Since the structural changes in SS-CCMV are local to the quasi-sixfold axes, they do not preclude the capsid from swelling via expansion of the openings at the quasi-threefold axes. Indeed, sucrose sedimentation velocity experiments reproducibly detected swelling of SS-CCMV particles at pH 7.5 by chelating divalent cations with EDTA. These data demonstrate that the SS-CCMV particles do expand under the same conditions, and to the same extent, as WT particles. The use of EDTA was not cited in two previous studies of SS-CCMV swelling; one concluded that the mutant particles did not expand (3), and the other concluded that the mutant particles did not expand to the same degree as WT when divalent cations were absent (9). It is possible divalent cations remained in reduced concentrations in the previous studies as both used only dialysis to reach the higher pH condition, and this low concentration was sufficient to inhibit swelling.
While the exact steps in CCMV capsid expansion remain unclear, correct dimer formation is required for particle assembly and swelling. Mutational studies have shown that removing the C-terminal arm of the subunit completely abrogates particle assembly (55), and hydrogen exchange studies of the closely similar BMV capsid showed a tightening of the structure in the vicinity of the C-termini as the BMV particle expands (52). Outside the large reorganization at the quasi-threefold axes, both structural and computation studies concur that the transition between the WT closed and swollen states is accommodated by the flexible dimer interface, which remains largely intact and binds the pentamers and hexamers together (34, 42, 43). Portions of the N-terminal arm are also involved in dimer contacts (Fig. 2A), suggesting that alterations to the native, extended dimer interactions can have a measurable impact on particle transitions without altering the dimer interface itself. The subunit contacts within the hexamers and pentamers are also largely unchanged although these capsomers have rotated slightly compared to the native state. In SS-CCMV, the N terminus of each hexamer subunit is involved in new contacts. These additional interactions would seem to favor the closed capsid state under swelling conditions. The N termini are likely to be more strongly anchored in the -hexamer, preventing distortion of the dimer interface, reducing the sampling of favorable subunit orientations, and averting dissolution of the -hexamer to the point necessary to achieve WT capsomer rotation and particle expansion. However, the swollen SS-CCMV particles are indistinguishable from swollen WT particles in sucrose sedimentation experiments when both are exposed to the swelling buffer overnight.
Stabilized particles retain dynamic pentamers. Past structural and biochemical studies indicate that the N-terminal region of the capsid protein is located inside the closed and swollen forms of WT-CCMV and inside the closed form of SS-CCMV (5, 42, 50, 54, 55). However, protease mass mapping experiments reveal that in solution the N terminus of the capsid subunit is readily accessible to trypsin. This agrees well with proteolysis studies of the BMV capsid, which produced primarily N-terminal peptides (52). For the protease to access these N-terminal residues, they must have intermittent exposure to the particle surface, implying a breathing motion for the capsid. The new interactions in SS-CCMV at the hexamer N termini could reduce their exposure at the particle surface and/or restrict the capsid breathing motion. Reductions in N termini exposure or capsid breathing would reduce the level of protease digestion leading to the greater protease resistance of the SS-CCMV particles.
The rate and sites of protease cleavage are a function of protein dynamics, allowing us to compare the solution phase properties of the particles (10). A highly biased distribution of amino acids creates chemically distinct regions and allows the differential probing of protein regions based on protease specificity. Proteases with basic or acidic specificity will, therefore, report primarily on the accessibility of the CCMV coat protein N- or C-terminal regions, respectively. Although SS-CCMV is dramatically stabilized, the sites and apparent early rate of protease cleavage in WT- and SS-CCMV are very similar, indicating that identical regions of the overall capsid and protein subunit are equally dynamic in the two particles (Fig. 6 and 7). All of the primary cleavages are within the first 42 residues of the protein, with one of the primary cleavage sites in both WT- and SS-CCMV being residue 42. Involvement of Arg42 in the new bonding network of the SS-CCMV particle apparently does not significantly reduce the preference for this site by trypsin. Indeed, protease cleavage sites map to position 42 at very early time points in both forms of the capsid. This suggests that the early appearance of this site in SS-CCMV is contributed by subunits within the pentamer. With no changes to the bonding of pentamer N termini, changes in dynamics of this region could indicate that the increased stability of hexamers propagates throughout the particle. This set of N termini can cycle back and forth to the particle surface at the WT rate without affecting the stabilizing interactions centered at the hexamers and be cleaved in the protease mapping experiments at nearly the same rate as WT-CCMV. This defines two sets of dynamic N termini in SS-CCMV. The set located at the center of the pentamers is unaffected by the K42R mutation and makes it appear that the stabilized particles have unaltered dynamics. The set located in the hexamers is contributing to a web of new interactions responsible for the stability and slightly abnormal swelling of the mutant capsids, and this set is probably cleaved more slowly than the equivalent set in WT particles. Alternatively, it may be that only subunits within pentamers are cleaved at the early time points in either type of particle. The difference in the stability of WT- and SS-CCMV would then be due mainly to an overall decrease in SS-CCMV dynamics connected to the increase in hexamer stability. The detailed hydrogen bonding network created by the K42R mutation supports the idea that initially only subunits in pentamers are cleaved.
Although SS-CCMV has an isometric particle structure, its dynamics may be quite asymmetric. The K42R mutation can lead to a clear distinction between capsomer dynamics, where pentamer N termini have a faster surface exchange rate than those of hexamers, or at least more so than those of WT-CCMV. The hexamer N termini in the WT particle have more visibly ordered structure and interactions that could slow their surface exposure compared to those of the pentamers, hinting at partly asymmetric dynamics even in the WT particle. With the additional interactions created by the K42R mutation, the difference in exchange rates for pentamer versus hexamer N termini in SS particles can become even greater, resulting in increased salt stability and protease resistance. Increasingly asymmetric, dynamic movements that act to stabilize a highly symmetric virus particle seem to contradict the structural foundations of the particle itself; however, this is caused by a reduction of the dynamics that make the WT particle more vulnerable to proteases and high-salt conditions. Concurrently, the same reduction in dynamics does not seem to affect overall particle swelling and has little effect on infectivity of the SS-CCMV particle. Dynamic hexamers, at least of the level found in WT particles, appear not to be required for these large structural transitions associated with disassembly.
Thus, the viable increase in particle stability is a result of concentrated tethering of subunits at a highly symmetrical interface (i.e., quasi-sixfold axes) that does not interfere with the function of other key symmetry interfaces (i.e., fivefold, twofold, and quasi-threefold axes). Indeed, the relatively large quaternary rearrangements that lead to the reversible swelling of the capsid appear to have less of an effect on particle dynamics than the K42R point mutation. The effect is a particle that is still dynamic and insensitive to the high-salt conditions but is not disrupted due to a new, interleaving series of bonds that preserve the overall particle structure.
ACKNOWLEDGMENTS
The authors thank Jason Lanman and Lu Gan for insightful discussions, Tianwei Lin for assisting in X-ray data collection, and Gary Siuzdak for providing access to the TSRI Center for Mass Spectrometry.
This study was supported by Public Health Service grant GM54076 from the National Institutes of Health (J.E.J).
J.A.S. and B.B. contributed equally to this work.
REFERENCES
Agarwal, R. C. 1978. A new least-squares refinement technique based on the fast Fourier transform algorithm. Acta Crystallogr. A 34:791-809.
Ahlquist, P. 1992. Bromovirus RNA replication and transcription. Curr. Opin. Genet. Dev. 2:71-76.
Albert, F. G., J. M. Fox, and M. J. Young. 1997. Virion swelling is not required for cotranslational disassembly of cowpea chlorotic mottle virus in vitro. J. Virol. 71:4296-4299.
Allison, R. F., M. Janda, and P. Ahlquist. 1988. Infectious in vitro transcripts from cowpea chlorotic mottle virus cDNA clones and exchange of individual RNA components with brome mosaic virus. J. Virol. 62:3581-3588.
Annamalai, P., S. Apte, S. Wilkens, and A. L. Rao. 2005. Deletion of highly conserved arginine-rich RNA binding motif in cowpea chlorotic mottle virus capsid protein results in virion structural alterations and RNA packaging constraints. J. Virol. 79:3277-3288.
Bancroft, J. B. 1970. The self-assembly of spherical plant viruses. Adv. Virus Res. 16:99-134.
Bancroft, J. B., G. J. Hills, and R. Markham. 1967. A study of the self-assembly process in a small spherical virus: Formation of organized structures from protein subunits in vitro. Virology 31:354-379.
Bancroft, J. B., and R. W. Horne. 1977. Bromovirus (brome mosaic viurs) group, p. 287-302. In K. Maramorosch (ed.), The atlas of insect and plant viruses, vol. 8. Academic Press, New York, N.Y.
Bancroft, J. B., M. W. Rees, M. W. Johnson, and J. R. O. Dawson. 1973. A salt stable mutant of cowpea chlorotic mottle virus. J. Gen. Virol. 21:507-513.
Bothner, B., X. F. Dong, L. Bibbs, J. E. Johnson, and G. Siuzdak. 1998. Evidence of viral capsid dynamics using limited proteolysis and mass spectrometry. J. Biol. Chem. 273:673-676.
Bothner, B., A. Schneemann, D. Marshall, V. Reddy, J. E. Johnson, and G. Siuzdak. 1999. Crystallographically identical virus capsids display different properties in solution. Nat. Struct. Biol. 6:114-116.
Brumfield, S., D. Willits, L. Tang, J. E. Johnson, T. Douglas, and M. Young. 2004. Heterologous expression of the modified coat protein of cowpea chlorotic mottle bromovirus results in the assembly of protein cages with altered architectures and function. J. Gen. Virol. 85:1049-1053.
Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and G. L. Warren. 1998. Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54:905-921.
Cerf, R., B. Michels, J. A. Schulz, J. Witz, P. Pfeiffer, and L. Hirth. 1979. Ultrasonic absorption evidence of structural fluctuations in viral capsids. Proc. Natl. Acad. Sci. USA 76:1780-1782.
Chidlow, J., and J. H. Tremaine. 1971. Limited hydrolysis of cowpea chlorotic mottle virus by trypsin and chymotrypsin. Virology 43:267-278.
Collaborative Computational Project, Number 4. 1994. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50:760-763.
Da Poian, A. T., J. E. Johnson, and J. L. Silva. 2002. Protein-RNA interactions and virus stability as probed by the dynamics of tryptophan side chains. J. Biol. Chem. 277:47596-47602.
Douglas, T., E. Strable, D. Willits, A. Aitouchen, M. Libera, and M. Young. 2002. Protein engineering of a viral cage for constrained nanomaterials synthesis. Adv. Mater. 14:415-418.
Fersht, A. R., J. P. Shi, J. Knill-Jones, D. M. Lowe, A. J. Wilkinson, D. M. Blow, P. Brick, P. Carter, M. M. Waye, and G. Winter. 1985. Hydrogen bonding and biological specificity analysed by protein engineering. Nature 314:235-238.
Fontana, A., M. Zambonin, P. Polverino de Laureto, V. De Filippis, A. Clementi, and E. Scaramella. 1997. Probing the conformational state of apomyoglobin by limited proteolysis. J. Mol. Biol. 266:223-230.
Fox, J. M., G. Wang, J. A. Speir, N. H. Olson, J. E. Johnson, T. S. Baker, and M. J. Young. 1998. Comparison of the native CCMV virion with in vitro assembled CCMV virions by cryoelectron microscopy and image reconstruction. Virology 244:212-218.
Fox, J. M., X. Zhao, J. A. Speir, and M. J. Young. 1996. Analysis of a salt stable mutant of cowpea chlorotic mottle virus. Virology 222:115-122.
Jacrot, B. 1975. Studies on the assembly of a spherical virus. II. The mechanism of protein aggregation and virus swelling. J. Mol. Biol. 95:433-446.
Johnson, J. E. 2003. Virus particle dynamics. Adv. Protein Chem. 64:197-218.
Jones, T. A., J. Y. Zou, S. W. Cowan, and M. Kjeldgaard. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A. 47:110-119.
Kleywegt, G. J., and T. A. Jones. 1999. Software for handling macromolecular envelopes. Acta Crystallogr. D 55:941-944.
Lane, L. C. 1981. Bromovirus, p. 333-336. In E. Kurstak (ed.), Handbook of plant virus infection and comparative diagnosis. Elsevier/North-Holland, Amsterdam, The Netherlands.
Lane, L. C. 1974. The bromoviruses. Adv. Virus Res. 19:151-220.
Lanman, J., T. T. Lam, M. R. Emmett, A. G. Marshall, M. Sakalian, and P. E. Prevelige, Jr. 2004. Key interactions in HIV-1 maturation identified by hydrogen-deuterium exchange. Nat. Struct. Mol. Biol. 11:676-677.
Lanman, J., and P. E. Prevelige, Jr. 2004. High-sensitivity mass spectrometry for imaging subunit interactions: hydrogen/deuterium exchange. Curr. Opin. Struct. Biol. 14:181-188.
Laskowski, R. A., M. W. MacArthur, D. S. Moss, and J. M. Thornton. 1993. PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26:283-291.
Lewis, J. K., B. Bothner, T. J. Smith, and G. Siuzdak. 1998. Antiviral agent blocks breathing of the common cold virus. Proc. Natl. Acad. Sci. USA 95:6774-6778.
Li, Q., A. G. Yafal, Y. M. Lee, J. Hogle, and M. Chow. 1994. Poliovirus neutralization by antibodies to internal epitopes of VP4 and VP1 results from reversible exposure of these sequences at physiological temperature. J. Virol. 68:3965-3970.
Liu, H., C. Qu, J. E. Johnson, and D. A. Case. 2003. Pseudo-atomic models of swollen CCMV from cryo-electron microscopy data. J. Struct. Biol. 142:356-363.
Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276:307-326.
Pettersen, E. F., T. D. Goddard, C. C. Huang, G. S. Couch, D. M. Greenblatt, E. C. Meng, and T. E. Ferrin. 2004. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25:1605-1612.
Pfeiffer, P. 1980. Changes in the organization of bromegrass mosaic virus in response to cation binding as probed by changes in susceptibility to degradative enzymes. Virology 102:54-61.
Phelps, D. K., B. Speelman, and C. B. Post. 2000. Theoretical studies of viral capsid proteins. Curr. Opin. Struct. Biol. 10:170-173.
Read, R. J. 1986. Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallogr. A 42:140-149.
Sheu, S. Y., D. Y. Yang, H. L. Selzle, and E. W. Schlag. 2003. Energetics of hydrogen bonds in peptides. Proc. Natl. Acad. Sci. USA 100:12683-12687.
Speir, J. A., S. Munshi, T. S. Baker, and J. E. Johnson. 1993. Preliminary X-ray data analysis of crystalline cowpea chlorotic mottle virus. Virology 193:234-241.
Speir, J. A., S. Munshi, G. Wang, T. S. Baker, and 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.
Tama, F., and C. L. Brooks III. 2002. The mechanism and pathway of pH-induced swelling in cowpea chlorotic mottle virus. J. Mol. Biol. 318:733-747.
TenEyck, L. F. 1973. Crystallographic fast Fourier transforms. Acta Crystallogr. A 29:183-191.
Tong, L., and M. G. Rossmann. 1997. Rotation function calculations with GLRF program. Methods Enzymol. 276:594-611.
Tremaine, J. H., and H. O. Agrawal. 1972. Limited proteolysis of potato virus X by trypsin and plant proteases. Virology 49:735-744.
Verduin, B. J., B. Prescott, and G. J. Thomas, Jr. 1984. RNA-protein interactions and secondary structures of cowpea chlorotic mottle virus for in vitro assembly. Biochemistry 23:4301-4308.
Vriend, G., M. A. Hemminga, B. J. M. Verduin, J. L. De Wit, and T. J. Schaafsma. 1981. Segmental mobility involved in protein-RNA interaction in cowpea chlorotic mottle virus. FEBS Lett. 134:167-171.
Vriend, G., M. A. Hemminga, B. J. M. Verduin, and T. J. Schaafsma. 1982. Swelling of cowpea chlorotic mottle virus studied by proton nuclear magnetic resonance. FEBS Lett. 146:319-321.
Vriend, G., B. J. M. Verduin, and M. A. Hemminga. 1986. Role of the N-terminal part of the coat protein in the assembly of cowpea chlorotic mottle virus. A 500 MHz proton nuclear magnetic resonance study and structural calculations. J. Mol. Biol. 191:453-460.
Vriend, G., B. J. M. Verduin, M. A. Hemminga, and T. J. Schaafsma. 1982. Mobility involved in protein-RNA interaction in spherical plant viruses, studied by nuclear magnetic resonance spectroscopy. FEBS Lett. 145:49-52.
Wang, L., L. C. Lane, and D. L. Smith. 2001. Detecting structural changes in viral capsids by hydrogen exchange and mass spectrometry. Protein Sci. 10:1234-1243.
Willits, D., X. Zhao, N. Olson, T. S. Baker, A. Zlotnick, J. E. Johnson, T. Douglas, and M. J. Young. 2003. Effects of the cowpea chlorotic mottle bromovirus beta-hexamer structure on virion assembly. Virology 306:280-288.
Wolfs, C. J., M. van der Graaf, and M. A. Hemminga. 1993. Conformational changes in oligo-nucleotides upon binding to a peptide representing the N-terminal region of a viral coat protein. An optical spectroscopy study. Biochim. Biophys. Acta 1172:1-4.
Zhao, X., J. M. Fox, N. H. Olson, T. S. Baker, and M. J. Young. 1995. In vitro assembly of cowpea chlorotic mottle virus from coat protein expressed in Escherichia coli and in vitro transcribed viral cDNA. Virology 207:486-494.
Zlotnick, A., R. Aldrich, J. M. Johnson, P. Ceres, and M. J. Young. 2000. Mechanism of capsid assembly for an icosahedral plant virus. Virology 277:450-456.(Jeffrey A. Speir, Brian B)