The genus Enterovirus encompasses important contaminants of water and food (e.g., coxsackieviruses), as well as viruses of acute public health concern (e.g., poliovirus). Depending on the properties of the surrounding matrix, enteroviruses exhibit different sensitivities to heat, which in turn influences their persistence in the environment, during food treatment, and during vaccine storage. Here, we determined the effect of NaCl and pH on the heat stability of different enteroviruses and related the observed effects to changes in protein interaction forces in the viral capsid. We demonstrate that NaCl renders enteroviruses thermotolerant and that this effect stems from an increase in van der Waals forces at different protein subunits in the viral capsid. This work sheds light on the mechanism by which salt enhances virus stability.
KEYWORDS: echovirus 11, coxsackievirus B1, coxsackievirus B5, thermostability, pentamer interface, protomer interface
ABSTRACT
Enteroviruses are common agents of infectious disease that are spread by the fecal-oral route. They are readily inactivated by mild heat, which causes the viral capsid to disintegrate or undergo conformational change. While beneficial for the thermal treatment of food or water, this heat sensitivity poses challenges for the stability of enterovirus vaccines. The thermostability of an enterovirus can be modulated by the composition of the suspending matrix, though the effects of the matrix on virus stability are not understood. Here, we determined the thermostability of four enterovirus strains in solutions with various concentrations of NaCl and different pH values. The experimental findings were combined with molecular modeling of the protein interaction forces at the pentamer and the protomer interfaces of the viral capsids. While pH only had a modest effect on thermostability, increasing NaCl concentrations raised the breakpoint temperatures of all viruses tested by up to 20°C. This breakpoint shift could be explained by an enhancement of the van der Waals attraction forces at the two protein interfaces. In comparison, the (net repulsive) electrostatic interactions were less affected by NaCl. Depending on the interface considered, the breakpoint temperature shifted by 7.5 or 5.6°C per 100-kcal/(mol·Å) increase in protein interaction force.
IMPORTANCE The genus Enterovirus encompasses important contaminants of water and food (e.g., coxsackieviruses), as well as viruses of acute public health concern (e.g., poliovirus). Depending on the properties of the surrounding matrix, enteroviruses exhibit different sensitivities to heat, which in turn influences their persistence in the environment, during food treatment, and during vaccine storage. Here, we determined the effect of NaCl and pH on the heat stability of different enteroviruses and related the observed effects to changes in protein interaction forces in the viral capsid. We demonstrate that NaCl renders enteroviruses thermotolerant and that this effect stems from an increase in van der Waals forces at different protein subunits in the viral capsid. This work sheds light on the mechanism by which salt enhances virus stability.
INTRODUCTION
Enteroviruses are common viruses infecting humans, and they can cause a spectrum of diseases ranging from mild to fatal (1). Vaccines against several enteroviruses are already available (e.g., poliovirus [2] and enterovirus 71 [3]) or are under development (e.g., enterovirus D68 [4] and coxsackievirus A16 [5]). To enable global immunization programs, such vaccines should be thermostable, to minimize the loss of potency over time (6). Enteroviruses are also common contaminants of water (7) and food (8). In these matrices, a high viral thermostability is detrimental to human health, as it can interfere with efforts at food preservation and consumption and enhance the environmental persistence of waterborne viruses (9).
Enteroviruses have a single-stranded RNA genome, surrounded by an icosahedral capsid composed of 60 protomer repeats, which are in turn composed of four structural proteins (VP1 to VP4) and organized in pentameric subunits (Fig. 1). For some enteroviruses, the mechanism of thermal inactivation involves the disintegration of the viral capsid. Specifically, the dissociation of the capsid into pentameric subunits with increasing temperature was observed for foot-and-mouth disease virus (FMDV) (10). This indicates that interaction forces between the capsid pentamers modulate this virus’ thermostability. Therefore, one option to enhance the thermoresistance of FMDV is to increase the interaction forces between the pentameric subunits, by selective mutation of the amino acids at the interface. For example, engineering a disulfide bond into the pentameric interface via the mutation of a single amino acid in VP2 efficiently increased the thermostability of FMDV and shifted the capsid melting temperature upward by several degrees (10, 11). A comparable shift was observed when noncovalent interactions between pentamers were enhanced, by introducing a VP2 mutation that causes hydrophobic stacking of aromatic side chains (12). And finally, engineered mutations that reduced electrostatic, carboxylate-mediated repulsion forces at the pentameric interface also enhanced the thermostability of FMDV (13).
FIG 1.
Structural features and interaction forces of the virus capsid. Green, van der Waals forces; red, electrostatic forces; black, overall interaction forces. (A) Schematic representation of the icosahedral capsid structure. A single pentamer is displayed in gray. (B) Schematic representation of pentameric interaction forces. A single protomer is displayed in blue. (C) Schematic representation of protomeric interaction forces. Each protomer contains a single copy of each structural protein (VP1 to VP4).
Other enteroviruses (e.g., poliovirus [14, 15], enterovirus 71 [EV71] [16], and coxsackievirus A7 [17]) have been shown to undergo conformational rearrangement rather than dissociation upon mild heating (50 to 56°C). This rearrangement is initiated by capsid expansion, which involves rotation of the protomer and leads to structural disruptions along the pentamer and protomer interfaces (14–17) and ultimately to the formation of a pore in the twofold axis of symmetry through which the viral RNA is released (15, 18, 19). It is reasonable to assume that the initial rotation, as well as the structural disruptions, may be inhibited by stronger interaction forces at the subunit interfaces. Experimental evolution experiments have shown that thermostable poliovirus variants carry mutations in the hydrophobic pocket region of VP1 (20–22). These mutations are distant from the pentameric interfaces but close to the interface of the protomeric subunits. They may thus increase capsid stability by enhancing the interaction forces at the protomer interface, though this effect has not been investigated.
Besides changes to the capsid composition, the thermostability of enteroviruses can be affected by the composition of the suspending matrix. Most importantly, it has long been recognized that the thermostability of enteroviruses is enhanced in the presence of dissolved salts and that the enhancement depends on the salt concentration and identity (23–25). The mechanism by which salts achieve this thermostabilizing effect, however, remains unclear. We can reasonably assume that salts modulate the interaction forces at the protein interfaces. Specifically, we expect that shifting the ionic strength of a virus-bearing solution results in shielding or promotion of electrostatic and van der Waals forces between protein subunits. Similar effects may be expected for shifts in the solution pH.
In this study, we investigated how changes in matrix composition modulate the protein interaction forces and the thermostability of enteroviruses. Specifically, we determined the thermal inactivation kinetics at 55°C, as well as the capsid breakpoint temperatures, in solutions with NaCl concentrations ranging from 0.01 to 3 M and pH values ranging from 3 to 9. Here, we define the capsid breakpoint temperature as the lowest temperature that causes a rapid loss of virus infectivity. Molecular modeling was applied to compute the corresponding electrostatic and van der Waals forces at different protein interfaces. As a representative of the genus Enterovirus, we used coxsackievirus group B5 (CVB5). CVB5 is an enteric human pathogen that is frequently detected in the aqueous environment (26) and that is notoriously resistant to disinfection (27). The most detailed analysis was conducted on the CVB5 Faulkner strain (CVB5-Faulkner). To generalize our findings to other Enterovirus strains, the results for CVB5-Faulkner were compared to those for a CVB5 environmental isolate (CVB5-L061815), an environmental isolate of the closely related serotype coxsackievirus group B1 (CVB1-L071615), and a more distantly related serotype (echovirus 11 Gregory strain [E11]) (28).
RESULTS AND DISCUSSION
Effects of salt and pH on breakpoint temperature and thermal inactivation kinetics.
To investigate the role of the matrix in virus stability, we determined the breakpoint temperature and inactivation kinetics in buffer solutions with various salt concentrations and different pH values. A linear, positive correlation was observed between the breakpoint temperature and the salt concentration up to 1.5 M NaCl (Pearson’s r = 0.98) (Fig. 2A). Specifically, the breakpoint temperature of CVB5-Faulkner increased from 45.0°C at a NaCl concentration of 10 mM to 64.7°C at 1.5 M, with an average breakpoint shift of 12.2°C/molar unit of NaCl. Beyond 1.5 M NaCl, the breakpoint continued to increase, albeit at a lower rate, reaching a maximal value of 67.2°C at 2.5 M.
FIG 2.
Effect of salt on the thermal stability of enteroviruses. (A) Capsid breakpoint temperatures of various Enterovirus B strains and NaCl concentrations. (B) Shift in breakpoint temperature between 10 mM and 1, 2, or 3 M NaCl solutions. Error bars indicate the 95% confidence intervals.
The shift in breakpoint temperature translated into pronounced changes in the sensitivity of CVB5-Faulkner to thermal inactivation. In a matrix containing 10 mM NaCl, the exposure of CVB5-Faulkner to 55°C resulted in an inactivation of 3 log10 within 15 s (Fig. 3). This finding was expected, given that the breakpoint of CVB5-Faulkner in this solution (45.0°C) (Fig. 2A) lay well below the treatment temperature. In contrast, if the matrix contained 1 M NaCl, the virus breakpoint (59.5°C) (Fig. 2A and Table 1) was greater than the treatment temperature, and hence, the virus remained stable over the time frame considered (Fig. 3).
FIG 3.
Effect of salt on the thermal inactivation kinetics of CVB5-Faulkner. The black line indicates inactivation at low salt condition (10 mM NaCl); the dotted line indicates inactivation at high salt concentrations (1 M NaCl). N/N0 is the residual fraction of infectious viruses.
TABLE 1.
Structural parameters and temperature breakpoints
| Virus | [NaCl] (M) | Value (SE) |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pentamer interface |
Protomer interface |
Breakpoint (°C) | ||||||||||
| Overall forces [kcal/(mol·Å)] | Electrostatic forces [kcal/(mol·Å)] | van der Waals forces [kcal/(mol·Å)] | No. of H bonds | No. of salt bridges | Overall forces [kcal/(mol·Å)] | Electrostatic forces [kcal/(mol·Å)] | van der Waals forces [kcal/(mol·Å)] | No. of H bonds | No. of salt bridges | |||
| CVB5-Faulkner | 0.01 | −111.2 (16.2) | 180.1 (7.1) | −291.3 (9.2) | 140 (1.1) | 4 (0.2) | −97.8 (63.2) | 242.7 (2.9) | −340.4 (64.7) | 148 (1.7) | 6 (0.0) | 44.9 (0.7) |
| 0.5 | −166.0 (22.9) | 156.3 (9.3) | −322.3 (13.8) | 136 (1.4) | 4 (0.0) | −116.6 (40.4) | 240.6 (4.9) | −357.2 (43.3) | 144 (1.6) | 6 (0.0) | 55.2 (0.6) | |
| 1 | −184.4 (24.5) | 153.0 (10.7) | −337.4 (14.3) | 130 (4.7) | 4 (0.0) | −186.5 (45.8) | 241.5 (4.6) | −431.4 (44.9) | 142 (1.6) | 6 (0.0) | 59.5 (0.3) | |
| 2 | −263.0 (17.5) | 117.8 (6.7) | −380.7 (11.3) | 118 (4.2) | 4 (0.0) | −279.0 (22.0) | 251.9 (2.8) | −531.9 (23.9) | 140 (2.7) | 6 (0.0) | 67.2 (0.2) | |
| 3 | −352.5 (14.7) | 93.7 (2.9) | −446.2 (12.3) | 96 (1.2) | 4 (0.2) | −297.5 (63.7) | 256.8 (3.8) | −554.3 (63.8) | 128 (2.5) | 6 (0.0) | 64.9 (1.5) | |
| CVB5-L061815 | 0.01 | −159.5 (23.6) | 174.9 (8.2) | −334.4 (15.8) | 138 (2.1) | 4 (0.1) | −86.4 (18.3) | 253.4 (1.2) | −339.7 (17.2) | 161 (2.3) | 6 (0.0) | 50.4 (0.6) |
| 1 | −245.0 (27.2) | 134.8 (12.5) | −379.8 (15.2) | 119 (5.3) | 4 (0.0) | −176.6 (27.7) | 255.1 (1.0) | −431.8 (26.7) | 145 (3.6) | 6 (0.0) | 61.4 (0.6) | |
| 2 | −360.3 (15.3) | 92.9 (3.1) | −453.2 (12.3) | 98 (2.0) | 4 (0.0) | −294.8 (15.7) | 252.5 (0.4) | −547.3 (15.6) | 133 (2.0) | 6 (0.0) | 66.8 (0.8) | |
| 3 | −354.5 (20.6) | 97.7 (7.1) | −452.2 (13.6) | 97 (3.1) | 4 (0.1) | −304.9 (14.1) | 251.3 (1.0) | −556.2 (13.7) | 132 (2.2) | 6 (0.0) | 68.5 (0.7) | |
| CVB1-L071615 | 0.01 | −103.6 (20.8) | 201.4 (6.4) | −305.0 (14.4) | 150 (3.4) | 4 (0.2) | −38.7 (21.9) | 237.1 (6.2) | −275.8 (20.5) | 142 (2.3) | 6 (0.1) | 49.5 (1.0) |
| 1 | −237.4 (29.2) | 156.2 (9.4) | −393.6 (20.0) | 127 (4.9) | 4 (0.0) | −159.7 (27.3) | 227.6 (2.0) | −387.3 (25.6) | 133 (3.3) | 7 (0.2) | 56.8 (1.4) | |
| 2 | −245.2 (26.8) | 163.0 (9.0) | −408.2 (18.0) | 125 (6.1) | 4 (0.0) | −170.3 (27.9) | 226.8 (2.0) | −397.1 (26.7) | 132 (3.4) | 6 (0.2) | 67.2 (0.7) | |
| 3 | −335.4 (12.7) | 129.1 (3.6) | −464.6 (9.2) | 104 (2.9) | 4 (0.0) | −270.2 (17.5) | 221.9 (1.6) | −492.1 (16.6) | 121 (2.4) | 7 (0.0) | 67.4 (0.3) | |
| E11 | 0.01 | −125.9 (26.5) | 278.2 (2.9) | −404.1 (23.5) | 160 (2.7) | 2 (0.0) | −212.5 (51.9) | 195.9 (4.5) | −408.4 (53.7) | 162 (2.1) | 4 (0.2) | 43.3 (0.3) |
| 1 | −203.8 (35.3) | 265.5 (7.2) | −469.2 (28.2) | 154 (2.5) | 2 (0.0) | −290.1 (64.4) | 190.2 (5.5) | −480.3 (61.7) | 147 (2.6) | 4 (0.1) | 57.0 (1.3) | |
| 2 | −341.5 (25.0) | 235.9 (4.8) | −577.4 (20.7) | 143 (0.4) | 2 (0.0) | −346.9 (31.3) | 194.2 (4.7) | −541.1 (29.4) | 139 (3.6) | 4 (0.3) | 65.2 (0.8) | |
| 3 | −372.7 (26.5) | 214.0 (5.1) | −586.8 (22.5) | 142 (0.5) | 2 (0.0) | −382.6 (30.5) | 180.4 (6.9) | −563.0 (25.8) | 136 (3.4) | 4 (0.1) | 62.8 (2.0) | |
Transmission electron microscopy (TEM) images revealed that in the 10 mM NaCl matrix, exposure of CVB5 to 55°C resulted in a decrease in the number of intact viral particles per grid area analyzed (162 versus 84 viral particles per 30 μm2 in the untreated and heated samples, respectively). This indicates that a portion of viruses disintegrated upon heating, though a subset of intact viral particles remained clearly visible (Fig. 4). Furthermore, compared to the observed extent of inactivation (3 log10), the decrease in particle numbers was modest. It is therefore reasonable to assume that inactivation is only partly due to capsid disintegration and that it also involves conformational rearrangement not visible by the TEM analysis performed.
FIG 4.
TEM images of CVB5-Faulkner prior to and after heating. Samples were held in a 10 mM NaCl–PBS matrix. Left, untreated sample; right, sample after exposure to 55°C for 1 min. Both samples exhibit intact viral particles, though the overall particle number was about 50% lower in the heat-treated sample.
A stability-enhancing effect of NaCl was observed for all enterovirus strains tested, though the extent of the enhancement varied between strains. While the breakpoints of E11 and CVB5-Faulkner shifted by approximately 22°C between 0.01 and 2 M NaCl, the increases were only 16.3 and 17.8°C for CVB5-L061815 and CVB1-L071615, respectively. At 3 M NaCl, all viruses tested exhibited similar breakpoint shifts of 18 to 20°C (Fig. 2B and Table 1).
Compared to the effect of salt, the effect of pH on the breakpoint of CVB5-Faulkner was subtler. At a salt concentration of 10 mM, the breakpoint temperature decreased with increasing pH, with values of 47.2, 45.0, and 40.9°C for pH 3, 7.4, and 9, respectively (Fig. 5). At all pH values tested, the breakpoint thus remained well below the temperature applied in the thermal inactivation experiment (55°C). Consequently, changing the pH did not result in a measurable difference in the thermal inactivation kinetics (data not shown). At a NaCl concentration of 1 M, the effects of pH were completely suppressed, and no differences in breakpoint were observed over the pH range considered (Fig. 5). The effect of pH on capsid thermostability is not well understood, but it may arise from changes in the protonation state of amino acid residues located at the capsid protein interfaces, which in turn modulate the protein interaction forces and hence the thermostability of the capsid. Alternatively, enteroviruses are known to undergo capsid conformational changes as a function of pH (29), which appeared to affect the inactivation kinetics under chlorine treatment (30). Those conformational changes may also affect the thermostability of the viral capsids. Given the comparatively mild effect of matrix pH on thermostability, however, this solution parameter was not included in our further investigation.
FIG 5.
Effect of pH on the breakpoint temperature of CVB5-Faulkner. In the presence of 10 mM NaCl (circles), an increase in pH lowered the breakpoint. In 1 M NaCl (squares), the breakpoint increases to close to 60°C for all pH conditions considered. The error bars correspond to the 95% confidence intervals.
Effect of salt on protein interaction forces.
To rationalize the pronounced effects of salt on thermostability, we determined how the forces that act at the protein interfaces are influenced by salt, using the 10 mM NaCl concentration as a reference. To this end, we determined the total electrostatic and van der Waals forces at the interface of viral proteins, and we analyzed the number of hydrogen bonds and salt bridges (Table 1). Specifically, we focused on the pentamer interface, which previous reports have described as the location of capsid disintegration during thermal inactivation of FMDV (10). In addition, we considered the protomer interface, which is affected by conformational rearrangement of poliovirus and EV71 (15, 16) and which is in close proximity to stability-enhancing mutations identified by experimental evolution (20–22) (Fig. 1).
The different protein interaction forces of CVB5-Faulkner are shown in Fig. 6. At 10 mM NaCl, the capsid exhibited strong attractive van der Waals forces of −291 ± 9 kcal/(mol·Å) at the pentamer and −340 ± 65 kcal/(mol·Å) at the protomer interface. Interestingly, however, the net electrostatic interactions at both interfaces were repulsive, with magnitudes of 180 ± 7 and 242 ± 3 kcal/(mol·Å) at the pentamer and the protomer, respectively. The stability of the viral capsid thus arises from the attractive van der Waals forces, which are sufficiently large in magnitude to overcome the destabilizing repulsive forces exerted by the electrostatic interactions at both interfaces. The overall attractive forces at both interfaces are in accordance with the rational fact that the virus needs a stable capsid.
FIG 6.
Pentamer and protomer interaction forces of CVB5-Faulkner as a function of the salt concentration. (A to C) Pentamer interface interaction forces. (D to F) Protomer interface interaction forces. The error bars correspond to the 95% confidence intervals, based on 7 computational replicas. Exact values are given in Table 1.
Increasing salt concentrations resulted in a strong increase in van der Waals attraction forces at the pentamer interface of CVB5-Faulkner. Concurrently, the repulsive electrostatic forces decreased, albeit to a lesser extent (Fig. 6A and B). Hence, the presence of salt resulted in an increase of overall attractive forces at this interface (Fig. 6C). The reduced repulsive forces could be attributed neither to a change in the number of salt bridges, which remained stable over all salt concentrations considered, nor to the number of H bonds, which decreased with increasing salt concentration (Table 1). Instead, this effect likely stems from the partial screening of the charged residues mediated by NaCl ions, which contributes to reducing the overall weight of the electrostatic interactions. Additionally, the partial neutralization of the charged residues leads to a reorientation of amino acids (in particular of their side chains), which allows the optimization of the van der Waals contacts and therefore contributes to the reduction of the interaction energy.
At the protomer interface, an even greater increase in attractive van der Waals forces with increasing salt concentration was found (Fig. 6D). In contrast, the electrostatic interactions were only minimally affected and remained roughly constant for all salt concentrations considered (Fig. 6E). Overall, the strong increase in van der Waals forces thus resulted in a marked increase of attractive forces at the protomer interface, similar in magnitude to the increase at the pentamer interface (Fig. 6F).
The four viruses tested exhibited some differences in absolute electrostatic and van der Waals forces (Table 1). The largest deviations were found for the pentamer interface of E11, the virus with the least sequence similarity among the viruses tested. At any given salt concentration, E11 exhibited stronger attractive van der Waals forces and repulsive electrostatic forces. Despite these differences, the overall pentamer interaction forces of E11 were equivalent to those of the other viruses. Furthermore, only minor differences were found among the viruses in the shift in interaction forces when the salt concentration was raised from 10 mM to higher values, suggesting comparable salt sensitivities across different strains of the Enterovirus B group.
Protein interaction forces correlate with breakpoint temperature.
Across all viruses and salt concentrations considered, the absolute overall forces at the pentamer interface were strongly correlated with the capsid breakpoint temperatures measured under the corresponding solution conditions (Pearson’s r = 0.87) (Fig. 7A). This correlation indicates that the breakpoint temperature changes by 7.5°C if the overall pentamer interaction forces change by 100 kcal/(mol·Å). A similar correlation was also found for the protomer interface (Pearson’s r = 0.68) (Fig. 7B). At this interface, the predicted change in breakpoint temperature corresponded to 5.6°C per 100-kcal/(mol·Å) increase in the protein interaction force. The strength of these correlations further increased if the change in interface forces relative to the values at 10 mM NaCl, rather than the absolute values, was considered (Pearson’s r = 0.87 and 0.92 for the pentamer and protomer interfaces, respectively). Such correlations may allow us to predict how changes in capsid protein interaction forces induced by solution conditions alter the thermostability of an enterovirus. As such, they may serve as tools to guide the design of matrices that preserve or destabilize enteroviruses. Such predictions, however, remain to be validated for additional viruses and solution conditions (e.g., different salts).
FIG 7.
Correlation between enterovirus breakpoint temperature and protein interaction forces. Breakpoint temperatures were measured over a NaCl concentration ranging from 0 to 3 M (with each point on the graph representing a different salt concentration) and were correlated against the overall interaction forces at the pentamer interface (Pearson’s r = −0.87) (A) and the protomer interface (Pearson’s r = −0.68) (B). Interaction forces correspond to those shown in Fig. 6C and F. The error bars show 95% confidence intervals.
Overall, this study demonstrates a major impact of simple matrix variation on virus thermostability and unravels the underlying changes in protein interaction forces. We demonstrate that even simple matrix modifications, such as changing the pH or the salinity, can have a strong capsid-stabilizing effect and that the main driving forces for the stabilization are van der Waals forces at protein interfaces. Finally, our findings were consistent across several Enterovirus strains, indicating that they may also apply to strains of clinical relevance or under vaccine development.
MATERIALS AND METHODS
Chemicals.
Na2HPO4 (99.0%) was obtained from Fluka (Honeywell International Inc.), glycine (puriss.) from Bio-Rad (Hercules, CA), and HCl (1 N) and NaCl (99.5%) from Acros Organics (Geel, Belgium). Cell culture media consisted of modified Eagle medium supplemented with 1% penicillin-streptomycin per ml and 10% (growth medium) or 2% (maintenance medium) heat-inactivated fetal bovine serum (all purchased from Gibco, Frederick, MD).
Cells and viruses.
CVB5 strain Faulkner (ATCC VR-185) and echovirus 11 strain Gregory (ATCC VR-41) were obtained from LGC Standards (Molsheim, France). Virus environmental isolates (CVB5-L061815 and CVB1-L071615) were isolated from untreated domestic sewage, as described elsewhere (28). Buffalo green monkey kidney (BGMK) cells were kindly provided by the Spiez Laboratory (Spiez, Switzerland). Virus stock solutions were prepared by amplification in BGMK cells and were purified by polyethylene glycol (PEG) precipitation as described previously (28), and were stored in phosphate-buffered saline (PBS; 5 mM Na2HPO4, 10 mM NaCl, pH 7.4). Virus infectivity was determined by endpoint dilution with most-probable-number (MPN) statistics (31) as described previously (28). In brief, virus samples were inoculated on confluent BGMK cells on 96-well plates, with five replicates and eight dilutions for each experimental sample. The cytopathic effect (CPE) was determined through microscopy 5 days after infection and incubation at 37°C with 5% CO2. The infectivity was then reported as most probable number of cytopathic units (MPNCU) per milliliter.
Thermal inactivation experiments.
Thermal inactivation experiments were performed in aqueous buffers with different pH values and ionic strengths. Experiments were conducted in 25 mM glycine-HCl (pH 3), in 5 mM Na2HPO4 (pH 7.4), and in 25 mM glycine-NaOH (pH 9). Different buffer compositions had to be used to optimize pH stability at each targeted pH value. The ionic strength was varied by addition of 10 mM to 1 M NaCl.
Kinetic experiments were conducted in a PCR thermocycler (Applied Biosystems; GeneAmp PCR system 9700). A 10-μl portion of virus stock was used to spike PCR tubes containing 90 μl of preheated buffer to reach an initial concentration around 107 MPNCU ml−1. Samples were maintained at 55°C for various amounts of time ranging from 0 to 2 min and were then quickly cooled by placing them on an aluminum PCR cooling block on ice. The residual infectious virus concentration in each sample was enumerated on the same day. Each sample was then mixed with 900 μl of cell culture medium and stored at –20°C prior to enumeration by the MPN assay described above. Each experiment was conducted at least in duplicate.
Determination of the capsid breakpoint temperature.
The breakpoint temperature, or melting temperature, of the viral capsids was determined by a thermal-shift assay. The assay was performed in an PCR thermocycler as described above, but each PCR tube was held at a different temperature ranging from 25 to 70°C at 2° intervals and was incubated for 1 min at each temperature. The breakpoint was identified as the temperature at which the inactivation rate shifted from low to high. It was determined by fitting a segmental linear regression to a plot of ln(N/N0) versus temperature applied, where N/N0 represents the residual fraction of infectious virus titers after the 1-min inactivation period (Fig. 8). The breakpoint temperature, reflected by the intersection of the two linear regression lines, were determined together with the respective standard error. Each breakpoint was derived based on at least two pooled experimental replicates.
FIG 8.
Determination of the breakpoint temperature. The figure shows the examples of CVB5-Faulkner and CVB5-L061815 at a NaCl concentration of 0.01 M. The ln of the residual infectivity (N/N0) after 1 min of treatment is plotted as a function of treatment temperature. The data are fitted to a segmental linear regression (black lines). The intersection of the two linear portions defines the breakpoint temperature of each virus.
TEM analysis.
To investigate if capsid disintegration occurs upon heating, we evaluated the number of intact CVB5 particles after exposed to different heat treatments. A PEG-purified solution of CVB5 (∼108 MPNCU ml−1 in PBS) was split into two aliquots of 100 μl. One aliquot was held at room temperature, and one was exposed to 55°C for 1 min. For TEM analysis, each sample was adsorbed on a glow-discharged carbon-coated copper grid (400 mesh; EMS, Hatfield, PA), washed with deionized water, and stained with 0.75% uranyl formate. Observations were made using an F20 electron microscope (Thermo Fisher, Hillsboro, OR) operated at 200 kV. Digital images were collected using a Falcon III direct detector camera (4,098 by 4,098 pixels; Thermo Fisher, Hillsboro, OR) at magnifications of ×29,000 (pixel size = 0.35 nm) and ×100,000 (pixel size = 0.10 nm), with a defocus range between −1.5 μm and −2.5 μm. For each sample, 16 images were acquired, which jointly covered 30 μm2 of the grid. Intact viral particles were enumerated visually.
Molecular modeling.
Three-dimensional (3D) capsid models were built using the known crystal structure of coxsackievirus B3 as a scaffold (PDB accession code 1cov) (32). Modeller (33) was used to obtain reliable 3D capsid structures (by homology modeling) of strains for which a crystal structure is not currently available in literature. After obtaining the basic unit for each strain, we used the Visual Molecular Dynamics (VMD) software (34) to recreate the pentamer, by alignment with the available crystal structure. We also added a portion of the neighboring pentamers, in order to obtain a full model of the pentamer-pentamer interface (Fig. 1). The protonation states of the amino acids that compose the proteins were assigned with PROPKA (35). All the models were solvated using a TIP3P water model (36) and ionized.
In order to assess the influence of salt concentration on the structural properties of each strain, we tested different NaCl concentrations (10 mM, 500 mM, 1 M, 2 M, and 3 M). The structures were minimized with 1,500 steps of conjugate gradient algorithm, in order to remove eventual clashes encountered during the homology modeling and aligning procedures. The presence of hydrogen bonds within the different models was detected using 3.0 Å and 20° as the distance and the angle thresholds, respectively; similarly, the salt bridges were assigned based on the distance between the deprotonated oxygen of acidic residues and the protonated nitrogen of basic ones. Specifically, we assumed the presence of a salt bridge when this distance was below a threshold of 3.2 Å. In both cases, we examined the total number of bonds both within a pentamer and within two specific interfaces, i.e., the pentamer interface and the protomer interface (Fig. 1). Hydrogen bonds and salt bridges are considered to be part of an interface when the two residues that compose the bond are not part of the same domain (pentamer or protomer) but belong to two separate domains.
The NAMD molecular dynamics software engine (37) was used to calculate the van der Waals forces (VDW) and electrostatic forces (ELEC) acting at these interfaces, with the unit of kilocalories per mole·Ångstrom. These forces were decomposed along three main axes: x and z for shearing forces parallel to the interface and y for the perpendicular forces (Fig. 1). The overall interface force was calculated by calculating the sum of VDWy and ELECy force vectors. These forces were determined separately for the pentamer and protomer interfaces. The calculations for the interaction forces were performed in seven replicas for each viral strain. For each replica, ions were randomly displaced within the solvent. Afterwards, the system was equilibrated through the minimization procedure described above. After the minimization, interaction forces were calculated for each replica, and the aberrant outlier values were removed (<±1,000 kcal/mol·Å).
Data analysis.
Data handling, MPN calculation, and statistical analysis were performed in R (38). The CRAN packages ggplot2 (39), gridExtra (40), and segmented (41) were used. The breakpoint temperature was calculated using the segmented function from the segmented package using the linear model function (lm) as object.
ACKNOWLEDGMENTS
This work was funded by the Swiss National Science Foundation (project number 31003A_163270).
We thank Virginie Bachmann for laboratory assistance and Davide Demurtas for the TEM analyses.
S.M. and T.K. conceived and planned the experiments. S.M. carried out the experiments and analyzed the data. A.P. and M.D.P. planned the molecular modeling. A.P. carried out the molecular modeling. T.K. and S.M. wrote the manuscript in consultation with A.P. and M.D.P.
REFERENCES
- 1.Pons-Salort M, Parker EPK, Grassly NC. 2015. The epidemiology of non-polio enteroviruses: recent advances and outstanding questions. Curr Opin Infect Dis 28:479–487. doi: 10.1097/QCO.0000000000000187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Centers for Disease Control and Prevention. Polio vaccination. https://www.cdc.gov/vaccines/vpd/polio/index.html.
- 3.Li Y, Zhou Y, Cheng Y, Wu P, Zhou C, Cui P, Song C, Liang L, Wang F, Qiu Q, Guo C, Zeng M, Long L, Cowling BJ, Yu H. 2019. Effectiveness of EV-A71 vaccination in prevention of paediatric hand, foot, and mouth disease associated with EV-A71 virus infection requiring hospitalisation in Henan, China, 2017–18: a test-negative case-control study. Lancet Child Adolesc Heal 3:697–704. doi: 10.1016/S2352-4642(19)30185-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dai W, Zhang C, Zhang X, Xiong P, Liu Q, Gong S, Geng L, Zhou D, Huang Z. 2018. A virus-like particle vaccine confers protection against enterovirus D68 lethal challenge in mice. Vaccine 36:653–659. doi: 10.1016/j.vaccine.2017.12.057. [DOI] [PubMed] [Google Scholar]
- 5.Mao Q, Wang Y, Yao X, Bian L, Wu X, Xu M, Liang Z. 2014. Coxsackievirus A16. Epidemiology, diagnosis, and vaccine. Hum Vaccin Immunother 10:360–367. doi: 10.4161/hv.27087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kartoglu U, Milstien J. 2014. Tools and approaches to ensure quality of vaccines throughout the cold chain. Expert Rev Vaccines 13:843–854. doi: 10.1586/14760584.2014.923761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gerba CP, Pepper IL. 2019. Microbial contaminants, p 191–217. In Brusseau ML, Pepper IL, Gerba CP (ed), Environmental and pollution science, 3rd ed Academic Press, New York, NY. [Google Scholar]
- 8.Larkin EP. 1981. Food contaminants—viruses. J Food Prot 44:320–325. doi: 10.4315/0362-028X-44.4.320. [DOI] [PubMed] [Google Scholar]
- 9.Bertrand I, Schijven JF, Sánchez G, Wyn-Jones P, Ottoson J, Morin T, Muscillo M, Verani M, Nasser A, de Roda Husman AM, Myrmel M, Sellwood J, Cook N, Gantzer C. 2012. The impact of temperature on the inactivation of enteric viruses in food and water: a review. J Appl Microbiol 112:1059–1074. doi: 10.1111/j.1365-2672.2012.05267.x. [DOI] [PubMed] [Google Scholar]
- 10.Kotecha A, Seago J, Scott K, Burman A, Loureiro S, Ren J, Porta C, Ginn HM, Jackson T, Perez-Martin E, Siebert CA, Paul G, Huiskonen JT, Jones IM, Esnouf RM, Fry EE, Maree FF, Charleston B, Stuart DI. 2015. Structure-based energetics of protein interfaces guides foot-and-mouth disease virus vaccine design. Nat Struct Mol Biol 22:788–794. doi: 10.1038/nsmb.3096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Porta C, Kotecha A, Burman A, Jackson T, Ren J, Loureiro S, Jones IM, Fry EE, Stuart DI, Charleston B. 2013. Rational engineering of recombinant picornavirus capsids to produce safe, protective vaccine antigen. PLoS Pathog 9:e1003255. doi: 10.1371/journal.ppat.1003255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Scott KA, Kotecha A, Seago J, Ren J, Fry EE, Stuart DI, Charleston B, Maree FF. 2017. SAT2 Foot-and-mouth disease virus structurally modified for increased thermostability. J Virol 91:e02312-16. doi: 10.1128/JVI.02312-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rincón V, Rodríguez-Huete A, López-Argüello S, Ibarra-Molero B, Sanchez-Ruiz JM, Harmsen MM, Mateu MG. 2014. Identification of the structural basis of thermal lability of a virus provides a rationale for improved vaccines. Structure 22:1560–1570. doi: 10.1016/j.str.2014.08.019. [DOI] [PubMed] [Google Scholar]
- 14.Belnap DM, Filman DJ, Trus BL, Cheng N, Booy FP, Conway JF, Curry S, Hiremath CN, Tsang SK, Steven AC, Hogle JM. 2000. Molecular tectonic model of virus structural transitions: the putative cell entry states of poliovirus. J Virol 74:1342–1354. doi: 10.1128/jvi.74.3.1342-1354.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Levy HC, Bostina M, Filman DJ, Hogle JM. 2010. Catching a virus in the act of RNA release: a novel poliovirus uncoating intermediate characterized by cryo-electron microscopy. J Virol 84:4426–4441. doi: 10.1128/JVI.02393-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wang X, Peng W, Ren J, Hu Z, Xu J, Lou Z, Li X, Yin W, Shen X, Porta C, Walter TS, Evans G, Axford D, Owen R, Rowlands DJ, Wang J, Stuart DI, Fry EE, Rao Z. 2012. A sensor-adaptor mechanism for enterovirus uncoating from structures of EV71. Nat Struct Mol Biol 19:424–429. doi: 10.1038/nsmb.2255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Seitsonen JJT, Shakeel S, Susi P, Pandurangan AP, Sinkovits RS, Hyvönen H, Laurinmäki P, Ylä-Pelto J, Topf M, Hyypiä T, Butcher SJ. 2012. Structural analysis of coxsackievirus A7 reveals conformational changes associated with uncoating. J Virol 86:7207–7215. doi: 10.1128/JVI.06425-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bostina M, Levy H, Filman DJ, Hogle JM. 2011. Poliovirus RNA is released from the capsid near a twofold symmetry axis. J Virol 85:776–783. doi: 10.1128/JVI.00531-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Shingler KL, Yoder JL, Carnegie MS, Ashley RE, Makhov AM, Conway JF, Hafenstein S. 2013. The enterovirus 71 A-particle forms a gateway to allow genome release: a cryoEM study of picornavirus uncoating. PLoS Pathog 9:e1003240. doi: 10.1371/journal.ppat.1003240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Adeyemi OO, Nicol C, Stonehouse NJ, Rowlands DJ. 2017. Increasing type 1 poliovirus capsid stability by thermal selection. J Virol 91:1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Shiomi H, Urasawa T, Urasawa S, Kobayashi N, Abe S, Taniguchi K. 2004. Isolation and characterisation of poliovirus mutants resistant to heating at 50°C for 30 min. J Med Virol 74:484–491. doi: 10.1002/jmv.20202. [DOI] [PubMed] [Google Scholar]
- 22.Nguyen Y, Jesudhasan PR, Aguilera ER, Pfeiffer JK. 2018. Identification and characterization of a poliovirus capsid mutant with enhanced thermal stability. J Virol 93:e01510-18. doi: 10.1128/JVI.01510-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Majer M, Thomssen R. 1965. Thermal inactivation of 32P-poliovirus at 37°C and 50°C in the presence of NaCl with high molarity. Arch Gesamte Virusforsch 17:585–593. doi: 10.1007/bf01262235. [DOI] [PubMed] [Google Scholar]
- 24.Wallis C, Melnick JL. 1961. Stabilization of poliovirus by cations. Tex Rep Biol Med 19:683–700. [PubMed] [Google Scholar]
- 25.Wallis C, Melnick JL, Rapp F. 1965. Different effects of MgCl2 and MgSO4 on the thermostability of viruses. Virology 26:694–699. doi: 10.1016/0042-6822(65)90332-6. [DOI] [PubMed] [Google Scholar]
- 26.Mena KD, Gerba CP, Haas CN, Rose JB. 2003. Risk assessment of waterborne coxsackievirus. J Am Water Works Assoc 95:122–131. doi: 10.1002/j.1551-8833.2003.tb10413.x. [DOI] [Google Scholar]
- 27.Payment P, Tremblay M, Trudel M. 1985. Relative resistance to chlorine of poliovirus and coxsackievirus isolates from environmental sources and drinking water. Appl Environ Microbiol 49:981–983. doi: 10.1128/AEM.49.4.981-983.1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Meister S, Verbyla ME, Klinger M, Kohn T. 2018. Variability in disinfection resistance between currently circulating enterovirus B serotypes and strains. Environ Sci Technol 52:3696–3705. doi: 10.1021/acs.est.8b00851. [DOI] [PubMed] [Google Scholar]
- 29.Butler M, Medien AR, Taylor GR. 1985. Electrofocusing of viruses and sensitivity to disinfection. Water Sci Technol 17:201–210. doi: 10.2166/wst.1985.0111. [DOI] [Google Scholar]
- 30.Young DC, Sharp DG. 1985. Virion conformational forms and the complex inactivation kinetics of echovirus by chlorine in water. Appl Environ Microbiol 49:359–364. doi: 10.1128/AEM.49.2.359-364.1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kott Y. 1966. Estimation of low numbers of Escherichia coli bacteriophage by use of the most probable number method. Appl Microbiol 14:141–144. doi: 10.1128/AEM.14.2.141-144.1966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Muckelbauer JK, Kremer M, Minor I, Diana G, Dutko FJ, Groarke J, Pevear DC, Rossmann MG. 1995. The structure of coxsackievirus B3 at 3.5 A resolution. Structure 3:653–667. doi: 10.1016/s0969-2126(01)00201-5. [DOI] [PubMed] [Google Scholar]
- 33.Webb B, Sali A. 2014. Comparative protein structure modeling using MODELLER. Curr Protoc Bioinformatics 47:5–6. doi: 10.1002/0471250953.bi0506s15. [DOI] [PubMed] [Google Scholar]
- 34.Humphrey W, Dalke A, Schulten K. 1996. VMD: visual molecular dynamics. J Mol Graph 14:33–38. doi: 10.1016/0263-7855(96)00018-5. [DOI] [PubMed] [Google Scholar]
- 35.Rostkowski M, Olsson MH, Søndergaard CR, Jensen JH. 2011. Graphical analysis of pH-dependent properties of proteins predicted using PROPKA. BMC Struct Biol 11:6. doi: 10.1186/1472-6807-11-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML. 1983. Comparison of simple potential functions for simulating liquid water. J Chem Phys 79:926–935. doi: 10.1063/1.445869. [DOI] [Google Scholar]
- 37.Phillips JC, Braun R, Wang W, Gumbart J, Tajkhorshid E, Villa E, Chipot C, Skeel RD, Kalé L, Schulten K. 2005. Scalable molecular dynamics with NAMD. J Comput Chem 26:1781–1802. doi: 10.1002/jcc.20289. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.R Development Core Team. 2017. R: a language and environment for statistical computing. R Foundation, Vienna, Austria: R-project.org. [Google Scholar]
- 39.Wickham H. 2009. ggplot2: elegant graphics for data analysis. Springer, Dordrecht, The Netherlands. [Google Scholar]
- 40.Auguie B. 2016. gridExtra: miscellaneous functions for “grid” graphics. https://rdrr.io/cran/gridExtra/.
- 41.Muggeo VMR. 2008. Segmented: an R package to fit regression models with broken-line relationships. R News 8:20–25. [Google Scholar]