A Possible Action of Divalent Transition Metal Ions at the Interpentameric Interface of Inactivated Foot-and-Mouth Disease Virus Provides a Simple but Effective Approach To Enhance Stability

How to stabilize the inactivated FMDV without affecting virus infectivity and immunogenicity is a big challenge in the vaccine industry. The electrostatic repulsion induced by protonation of a large amount of histidine residues at the interpentameric interface of viral capsids is one of the major mechanisms causing the dissociation of capsids.

ABSTRACT

The structural instability of inactivated foot-and-mouth disease virus (FMDV) hinders the development of the vaccine industry. Here, we found that some transition metal ions like Cu²⁺ and Ni²⁺ could specifically bind to FMDV capsids at capacities of about 7,089 and 3,448 metal ions per capsid, respectively. These values are about 33- and 16-fold greater than the binding capacity of nontransition metal ion Ca²⁺ (about 214 per capsid). Further thermodynamic studies indicated that all of these three metal ions bound to the capsids in spontaneous enthalpy-driving manners (ΔG < 0, ΔH < 0, ΔS < 0), and the Cu²⁺ binding had the highest affinity. The binding of Cu²⁺ and Ni²⁺ could enhance both the thermostability and acid-resistant stability of capsids, while the binding of Ca²⁺ was helpful only to the thermostability of the capsids. Animal experiments showed that the immunization of FMDV bound with Cu²⁺ induced the highest specific antibody titers in mice. Coincidently, the FMDV bound with Cu²⁺ exhibited significantly enhanced affinities to integrin β6 and heparin sulfate, both of which are important cell surface receptors for FMDV attachment. Finally, the specific interaction between capsids and Cu²⁺ or Ni²⁺ was applied to direct purification of FMDV from crude cell culture feedstock by the immobilized metal affinity chromatography. Based on our new findings and structural analysis of the FMDV capsid, a “transition metal ion bridges” mechanism, which describes linkage between adjacent histidine and other amino acids at the interpentameric interface of the capsids by transition metal ion coordination action, was proposed to explain the stabilizing effect imposed on the capsid.

IMPORTANCE How to stabilize the inactivated FMDV without affecting virus infectivity and immunogenicity is a big challenge in the vaccine industry. The electrostatic repulsion induced by protonation of a large amount of histidine residues at the interpentameric interface of viral capsids is one of the major mechanisms causing the dissociation of capsids. In the present work, this structural disadvantage inspired us to stabilize the capsids through coordinating transition metal ions with the adjacent histidine residues in the FMDV capsid, instead of removing or substituting them. This approach was proven effective to enhance not only the stability of FMDV but also enhance the specific antibody responses, thus, providing a new guideline for designing an easy-to-use strategy suitable for large-scale production of FMDV vaccine antigen.

INTRODUCTION

Vaccines prepared from inactivated foot-and-mouth disease virus (FMDV) are the most effective product to prevent the highly contagious foot-and-mouth disease (FMD) in cloven-hoofed animals (1). The potency of the vaccine is largely determined by the structural integrity of inactivated FMDV. Unfortunately, the FMDV capsid is sensitive to heat and pH and easily dissociates into pentameric subunits even at neutral pH and 4°C (2, 3). The intact virus, also known as the 146S particle, has a spherical shape with a diameter of about 30 nm. The viral capsid is composed of 60 tightly packed asymmetrical protomers. Each protomer is composed of four types of structural proteins, VP1 to VP4 (Fig. 1a) (4, 5). The dissociation of the FMDV capsid starts from the adjacent pentameric subunits (3). The electrostatic repulsion induced by protonation of numerous histidine residues at the interpentameric interface of viral capsids is one of the major mechanisms causing poor acid-resistant stability of the capsids (4–7). The electrostatic repulsion between negatively charged amino acids and the weak interaction between adjacent pentamers further decrease the thermo- and long-term stability of capsids (8).

FIG 1.

(a) Structure of O serotype FMDV capsids (PDB accession number 1FOD). VP1, green; VP2, cyan; VP3, magenta; VP4, yellow. Two asymmetrical protomers located on two adjacent pentamers are marked with black solid lines. (b) Enlargement of two asymmetrical protomers colored as in panel a, VP1 to VP4 are shown in the image. GH-loop on VP1 and α-helix on VP2 are colored in red and blue, respectively. The RDG motif on the GH-loop is depicted with a red surface, and histidine residues adjacent to the interpentameric interface are depicted with a blue surface.

To improve the stability of FMDV capsids, many different strategies against the dissociation mechanisms mentioned above have been proposed. These strategies include substituting histidine or negatively charged residues involved in electrostatic repulsions (6, 7, 9), introducing disulfide bonds, or noncovalent interactions across adjacent pentamers to strengthen the link between pentamers (3, 8). Though these strategies have been proven to be effective in improving the stability of capsids (3, 10, 11), mutant virus with increased capsid stability would possibly reduce viral infectivity by impairing the uncoating process of the virus (10, 12), thus, hindering propagation of FMDV in the host cells. This will be disadvantageous to the large-scale production of inactivated FMDV vaccine antigen. Therefore, exploring other simple and effective stabilization strategies suitable for large-scale production of vaccine is still necessary and crucial for the FMD vaccine industry.

According to the structure of FMDV capsids, there are numerous conserved histidine on VP2 and VP3 subunits at the interpentameric interface of capsids (7) (Fig. 1b). This feature inspired us to explore a new approach to stabilize the FMDV capsids through possible action of transition metal ions, such as Ni²⁺ and Cu²⁺. These ions incline to coordinate with histidine and some other amino acids containing aromatic nitrogen (e.g., tryptophan) and sulfur (e.g., cysteine), which are known as borderline Lewis bases according to the “hard-soft acid-base” principle (13, 14). In fact, the binding of metal ions has been widely reported to stabilize the viral capsids, mediate the self-assembly of virus, and regulate the virus-receptors recognition (2, 15, 16). However, the previous studies mostly focused on divalent nontransition metal ions such as Ca²⁺, Mg²⁺, and Mn²⁺, which are known as hard Lewis acids. They can coordinate with the amino acids containing carboxylate (e.g., glutamate and aspartate) and aliphatic nitrogen (e.g., asparagine and glutamine), known as hard Lewis bases (13, 14, 17), to stabilize the viral structure based on the so-called “ion bridge” mechanism. For FMDV, Ca²⁺ was also reported to stabilize the viral capsids through binding to Glu6 on the VP2 subunit (18). In addition, the binding of Ca²⁺ or Mg²⁺ was also found to regulate the specific recognition of FMDV by the integrin receptors on the cell surface (2). Nevertheless, possible interaction between histidine on FMDV capsids and transition metal ions is ignored, let alone stabilizing capsids through this action.

In the present work, we expected that the binding of transition metal ions on FMDV capsids might form “transition metal ion bridges” to strengthen the cross-link between adjacent pentameric subunits, thus, increasing the stability of capsids. To prove this assumption, the stoichiometry and thermodynamics of Cu²⁺ and Ni²⁺ binding to inactivated FMDV capsids were explored. Then, the effects of the binding of metal ions on the structure, stability, and immunogenicity of capsids were investigated. The further application of this interaction in purifying vaccine antigen was explored at last.

RESULTS

Stoichiometry of Cu²⁺ and Ni²⁺ binding to FMDV capsids.

According to the crystal structure and amino acids sequence of O serotype FMDV capsids (PDB accession number 1FOD), there are 1,200 histidine residues, 480 of which are located on the interpentameric interface, including His 21, His 65, His 87, His 157, and His 209 on VP2 and His 141, His 144, and His 191 on VP3. Most of these residues are relatively conserved among different serotypes. For example, His 157 on VP2 and His 141 and His 144 on VP3 exist in all 7 serotypes of FMDV (7). His 21 and His 209 on VP2 and His 191 on VP3 exist in O, A, and C serotypes, and His 87 on VP2 exists in A, O, Asia1, SAT1, and SAT3 serotypes (7). In addition, there are 420 tryptophan residues and 540 cysteine residues on each of the O serotype FMDVs. Because the histidine residues have the largest number and greatest influence on the stability of capsids (5, 7), they were taken as the main sites for exploring the binding of transition metal ions in this study.

To validate the existence of the interaction between transition metal ions and FMDV, the purified inactivated FMDV was mixed with Cu²⁺ or Ni²⁺ at 9 times the molar amount of total histidine residues on capsids. After removing the unbound metal ions, the number of metal ions binding on capsids was measured by inductively coupled plasma mass spectrometry (ICP-MS).

The binding capacity (number of metal ions per capsids) was calculated from the following equation:

$$\text{Binding capacity}=\frac{C_{\text{ICP}}\times M_{\text{FMDV}}}{C_{\text{FMDV}}\times M_{\text{metal}}}$$

where C_ICP is the mass concentration of metal ions in sample solution obtained by ICP-MS, and C_FMDV is the mass concentration of FMDV in sample solution determined by Micro BCA assay. M_metal is the molecular weight of metal ions, and M_FMDV is the molecular weight of FMDV capsids, which was estimated to be about 480 × 10⁴ Da, according to the amino acids sequence of capsids (PDB accession number 1FOD). The binding efficiency of the metal ions was defined as the ratio of the tightly bound amount determined by ICP-MS to the initially added amount of metal ions.

The ICP-MS results confirmed that Cu²⁺ and Ni²⁺ could bind to FMDV capsids at high binding capacities of about 7,089 and 3,448, corresponding to binding efficiencies of about 65% and 32%. We also measured Ca²⁺ binding as a comparison, and the binding capacity was only 214 and binding efficiency was as low as 2%.

The effects of the initially added metal ion amount on binding to FMDV were then investigated by mixing FMDV with each kind of metal ion with molar amounts equivalent to 100% (high level), 66% (medium level), and 33% (low level) of their corresponding binding capacities. Specifically, for Cu²⁺, the ratios were set as 7,089:1, 4,726:1, and 2,363:1 (metal ions/capsid). For Ni²⁺, the binding ratios were 3,448:1, 2,298:1, and 1,149:1. For Ca²⁺, the binding ratios of were 214:1, 142:1, and 71:1. After removing unbound metal ions, the FMDV-metal complexes were analyzed by ICP-MS. Results summarized in Table 1 show that Cu²⁺ and Ni²⁺ binding on capsids was positively correlated to the initially added number of ions. At each level, the binding efficiencies of Cu²⁺ and Ni²⁺ were about 70% to 90%, while binding for Ca²⁺ was too low to give an accurate binding efficiency value. In the following context, FMDV-metal complexes prepared by mixing each kind of metal ion at 100%, 66%, and 33% of their corresponding binding capacities were refereed as high, medium, and low metal ion binding levels.

TABLE 1.

Effects of added amount of metal ions on their binding to FMDV capsids

Metal ion	Binding level	Proportion equivalent to binding capacitya (%)	No. of metal ions per capsidb
			Added	Bound	Binding efficiency (%)
Cu2+	High	100	7,089	5,706	80.4
	Medium	66	4,726	3,767	79.7
	Low	33	2,363	1,725	73.0
Ni2+	High	100	3,448	3,104	90.0
	Medium	66	2,298	2,054	89.3
	Low	33	1,149	926	80.6
Ca2+	High	100	214	nsc	nsc
	Medium	66	142	nsc	nsc
	Low	33	71	nsc	nsc

^aBinding capacity refers to the maximal number of metal ions that could tightly bind to the FMDV capsid, which was determined to be 7,089 (Cu²⁺), 3,448 (Ni²⁺), and 214 (Ca²⁺), by mixing FMDV with metal ions of 9 times the molar amount of the total histidine residues in capsids.

^bData were from experiments only conducted once.

^cThe binding amount was too low to be accurately determined by ICP-MS. ns, Not significant.

Thermodynamics of Ni²⁺ and Cu²⁺ binding to FMDV capsids.

To investigate the thermodynamics of the binding process, microscale thermophoresis (MST) analysis was applied to measure the binding affinities between FMDV capsids and the metal ions. MST was a powerful analytical technology for detecting the affinity between proteins and ligands with low molecular weight, such as small molecules and ions (19, 20). Our results showed that the affinity of Cu²⁺ and Ni²⁺ with FMDV capsids was 17 ± 5 μM and 41 ± 4 μM, respectively, which was significantly higher than that of Ca²⁺ (270 ± 18 μM) (Table 2 and Fig. 2a).

TABLE 2.

Summary of the capacities, affinities, and thermodynamic parameters of Cu²⁺, Ni²⁺, and Ca²⁺ binding to FMDV capsids measured by ICP-MS, MST, and ITC

Metal ion	Binding capacitya	Kdb (μM)	ΔHc (cal/mol)	ΔSc (cal/mol/deg)	ΔGc (cal/mol)
Cu2+	7,089 ± 603	17 ± 5	−(1.5 ± 0.3) × 107	−(5.0 ± 1.0) × 104	−(1.7 ± 0.4) × 104
Ni2+	3,448 ± 445	41 ± 4	−(6.1 ± 2.0) × 105	−(2.0 ± 0.6) × 103	−(4.0 ± 1.5) × 103
Ca2+	214 ± 39	270 ± 18	−(3.6 ± 2.9) × 105	−(5.1 ± 3.5) × 102	−(3.7 ± 0.1) × 103

^aBinding capacities refer to the maximal number of metal ions that could tightly bind to the FMDV capsid as determined by ICP-MS.

^bDetermined by MST analysis.

^cDetermined by ITC analysis. Values are averages and standard deviations obtained from two (binding capacities) or three (K_d, ΔH, ΔS, and ΔG) independent experiments.

FIG 2.

Thermodynamic studies of the binding of Cu²⁺, Ni²⁺, and Ca²⁺ to FMDV capsids. (a) Binding affinities measured by MST. (b) ITC titration of Cu²⁺, Ni²⁺, and Ca²⁺ into FMDV capsids at 25°C in 20 mM Tris buffer, pH 7.8. Solid red line is the best fit of the data using the one-site model. (c) ΔH and ΔS contribution to the ΔG of the binding process measured by ITC. For each sample, the average value obtained from 3 independent experiments and the corresponding error bars (standard deviation) are indicated.

Then, we further explored the thermodynamics of the binding process by isothermal calorimetric titration (ITC) (Fig. 2b). Table 2 and Fig. 2c summarized the contribution of the enthalpy (ΔH) and the entropy (ΔS) to the free energy (ΔG) of metal ion binding. The characteristic of thermodynamic parameters of metal-ion binding was similar to that of other metal-protein coordinated interactions reported by other researchers (Fig. 2c) (21). ΔG values of Cu²⁺ and Ni²⁺ binding to capsids were all negative, which indicated that the transition metal ions could spontaneously bind to capsid like Ca²⁺. ΔH and ΔS values were all negative and demonstrated that the binding process was an enthalpically driven process and inclined to be a specific interaction. The enthalpic component was the major contributor to the stability of the capsid-metal complex. ΔH of Cu²⁺ and Ni²⁺ binding was about 41- and 1.7-fold higher than that of Ca²⁺ (Table 2). These values suggested that the FMDV-Cu²⁺ complex was the most stable one and the FMDV-Ca²⁺ complex was the most unstable. The results of binding capacities, affinities, and thermodynamic parameters all indicated that the transition metal ions were more favorable for binding to FMDV capsids than the nontransition metal ions (Table 2).

Structural changes of FMDV capsids after binding metal ions.

To explore the effects of the binding of metal ions on the conformation of FMDV capsids, the morphology and structure of capsid binding with different levels of metal ions were characterized. The transmission electron microscopy (TEM) images showed that all capsids had a spherical shape, and no morphological change or dissociation of capsids occurred after binding high levels of Ni²⁺, Cu²⁺, or Ca²⁺ (Fig. 3a). The hydraulic diameters of the capsids were all centered at 29 nm no matter the binding levels of metal ions (Data not shown).

FIG 3.

Structural characterizations of FMDV after metal ion binding. (a) Negative-stain TEM images measuring the changes of morphology of FMDV capsids after binding Ni²⁺, Cu²⁺, or Ca²⁺ at their high binding levels. CD spectra (b) and fluorescence spectra (c) measuring the changes of size distribution, tertiary structure, and secondary structure of FMDV capsids after binding metal ions at their different binding levels. (d) SDS denaturation test performed by MST to analyze the mechanism for the decrease of fluorescence intensity in FMDV capsids after Cu²⁺ binding.

Circular dichroism (CD) spectra showed that there were approximate 14% α-helix and 19% β-turn on capsids. The secondary structure of capsids changed marginally after binding different levels of metal ions (Fig. 3b). The fluorescence spectra showed that the binding of all three kinds of metal ions led to a decrease in the fluorescence intensity of capsids, although the maximum fluorescence intensity all centered at a wavelength of about 332.4 ± 0.5 nm (Fig. 3c). To identify whether the decrease of fluorescence intensity was caused by the tertiary conformational change of capsids or by fluorescence quenching of metal ions, an SDS-denaturation test was performed in MST. The principle of this test is that if the fluorescence changes were caused by protein-ligand-specific interaction-induced structural change, then the fluorescence intensity in the protein and the complex sample will be equal after denaturation; otherwise, the fluorescence intensity will remain different in the case that nonspecific fluorescence loss occurs. As shown in Fig. 3d, the variation in fluorescence intensity between the FMDV and FMDV-Cu²⁺ complex was about 71.6%. After denaturation, the FMDV-Cu²⁺ interactions were disrupted, and the fluorescence variation between FMDV and FMDV-Cu²⁺ complex reduced to only 12.0%. Based on this SDS test, we speculated that the fluorescence change in FMDV was caused by a tertiary structural change induced by the specific binding of metal ions instead of the nonspecific quenching effects of metal ions.

Stability of FMDV capsids after binding metal ions.

After confirming the specific binding of Cu²⁺ and Ni²⁺ to FMDV capsids, whether the binding of transition metal ions could improve the stability of capsids was discussed. Here, Ca²⁺, which has been proven to improve the stability of many different kinds of viral capsids (22, 23), was investigated as comparison. Firstly, the thermal denaturation temperature (T_m) value related to the dissociation of capsids into pentameric subunits was determined by differential scanning fluorescence (DSF) analysis. As shown in Fig. 4a, the T_m values of capsids were all increased after binding metal ions, and the T_m values were positively correlated to the levels of metal ions binding. Particularly, the binding of Cu²⁺ at a high level increased the T_m value of capsids by 2.76°C (45.82 ± 0.03°C versus 43.06 ± 0.23°C). The T_m value of the capsid with binding of a high level of Ni²⁺ increased by 1.5°C, slightly lower than that of Ca²⁺ binding (Fig. 4a). This improvement in thermostability of FMDV capsids was comparable or even superior to results reported by using other strategies, such as virus mutation by amino acid substitution or adding stabilizers. For instance, the T_m values of two mutant O serotype FMDVs (S93Y and S93F) were all increased by 1.5°C (3). By adding stabilizers like dextrose, glycerol, or sucrose, the T_m value of O serotype FMDV was increased by about 1.5°C to 3°C (24). In both reports, the T_m values were determined by the same DSF analysis method used in this study.

FIG 4.

Stability of FMDV capsids after binding different metal ions. (a) T_m values of FMDV capsids after binding different levels of metal ions measured by DSF. (b) The change of T_m values of FMDV capsids after removing metal ions by EDTA. (c) Acid-resistant stability of FMDV after binding Ni²⁺, Cu²⁺, or Ca²⁺. All samples were separately stored at various pH solutions for 8 min and subsequently analyzed for the contents of remaining intact capsids by HPSEC. The pH value that induced 50% dissociation of the intact capsids was defined as pH₅₀. For each sample above, the average value obtained from 3 independent experiments and the corresponding error bars (standard deviation) are indicated.

Then, FMDVs binding Cu²⁺ or Ni²⁺ were mixed with metal chelating agent EDTA, and this was followed by desalting to remove possibly released metal ions. The T_m values of capsids before and after this treatment were presented in Fig. 4b. Apparently, the T_m values of capsids after removing metal ions were decreased to values close to those for no-metal capsids. This result indicated that the transition metal ions could tightly bind to capsids through coordination interactions, and the binding processes were reversible.

We further established a method based on high-performance size-exclusion chromatography (HPSEC) analysis to study the effects of metal ion binding on the acid-resistance stability of capsids. This method can directly observe the dissociation of viral capsids in buffers of different pH, thus, providing a quick and accurate method to quantify the acid resistance stability of capsids. Here, pH₅₀, defined as the pH value that induces 50% dissociation of the intact FMDV capsids, was adopted to give quantitative evaluation of the acid-resistance stability of capsids. As presented in Fig. 4c, the pH₅₀ of no-metal capsid was 6.79 ± 0.01. After binding Cu²⁺ and Ni²⁺, the pH₅₀ value shifted to 6.42 ± 0.02 and 6.69 ± 0.01, respectively, indicating that the acid-resistant stability was enhanced. An amino acid substitution strategy was reported to improve the acid-resistant stability of live FMDV by shifting pH₅₀ to as low as 5.4 (9, 25, 26). Compared with that result, the improvement of acid-resistant stability through Cu²⁺ and Ni²⁺ binding seemed less significant. However, it should be noted that for the live FMDV, the pH₅₀ value was defined as the pH that caused a loss of 50% of the viral infectivity, which was determined by viral infection experiments (25). In addition, stabilities of live virus and the inactivated virus are somewhat different, as the latter one is usually reported to be more unstable (26). Therefore, the pH₅₀ values determined by different methods for live virus or inactivated virus might not be comparable.

In contrast to Cu²⁺ and Ni²⁺ that can improve the acid-resistant stability of FMDV, a slight increase in pH₅₀ was observed for the Ca²⁺ binding capsids (from pH 6.79 ± 0.01 to 6.84 ± 0.01), indicating a slightly adverse effect on the acid-resistant stability (Fig. 4c).

The effects of metal ion binding on stimulating specific antibody responses.

The effects of metal ion binding on stimulating specific antibody responses need to be evaluated when this stabilization strategy will be used in vaccines. Mice were immunized with no-metal FMDV and FMDV binding a high level of Ni²⁺, Cu²⁺, or Ca²⁺. The serum-specific IgG antibody titers were measured at 14 and 28 days after first immunization. As shown in Fig. 5a, 14 days after the first immunization, IgG titers of the FMDV-Cu²⁺ group were 2.5-fold higher than the other three groups, and IgG titers of FMDV-Ni²⁺ and FMDV-Ca²⁺ groups were similar to those of the no-metal group. Fourteen days after boosted immunization, IgG titers of all three metal ion binding groups were higher than those of the no-metal group. The FMDV-Cu²⁺ group also showed the highest antibody titers (about 3.6-fold higher than the no-metal group). And IgG titers of the FMDV-Ni²⁺ group and the FMDV-Ca²⁺ group were 2- and 2.5-fold higher than those of the no-metal group, respectively. The results indicated that the immunization of the capsids bound with metal ions, especially Cu²⁺, could induce a higher specific antibody response in mice.

FIG 5.

(a) FMDV-specific IgG antibody titers in mouse serum. Mice were subcutaneously injected with FMDV samples at 0 and 14 days. Blood samples were assayed at 14 and 28 days. (b) The affinities of FMDV capsids to integrin β6 receptor and heparin sulfate receptor determined by MST analyses. For each sample, the average value was obtained from 5 (IgG titers) or 3 (affinities) independent experiments, and the corresponding error bars (standard deviation) are indicated. Asterisks (*) denote statistically significant differences.

The ability of FMDV to attach to cells largely determines the subsequent immune responses (5, 27). Integrin is the most important surface receptor for FMDV and many other viruses entering cells (2). Among them, integrin αvβ6 has the highest affinity for FMDV, and the β subunit in general contributes to ligand specificity (2). Divalent metal ions such as Mg²⁺, Mn²⁺, and Ca²⁺ have been proven to regulate the conformation of integrin through binding to its metal-ion-dependent adsorption site and then to mediate the virus-receptor recognition by coordinating the RGD motif on FMDV (Fig. 1b) (2, 5, 16). In addition, some FMDV strains were reported to be able to directly recognize heparin sulfate (HS) as another high-affinity receptor even without the integrin receptor (5, 27).

Therefore, the possible mechanism of transition metal ions in improving the specific antibody response was explored from the perspective of receptor binding affinities. As shown in Fig. 5b, the binding of three metal ions could all increase the affinities between FMDV capsids and integrin β6 in vitro. The binding of Ca²⁺ led to a 4.4-fold increase in affinity compared to that of the no-metal binding capsids (dissociation constant [K_d], 0.017 ± 0.002 μM versus 0.075 ± 0.004 μM). Binding of Cu²⁺ or Ni²⁺ also increased the binding affinities but were weaker than that of Ca²⁺ (K_d, 0.024 ± 0.013 μM and 0.025 ± 0.002 μM, respectively). The effects of metal ion binding on FMDV-HS interactions showed different results. Only the binding of Cu²⁺ led to an enhancement in HS receptor affinity (about 6-fold higher than the no-metal capsids). There was no enhancement in affinities toward HS receptor by the binding of Ni²⁺ or Ca²⁺ (Fig. 5b).

From the results, we find that the binding of Cu²⁺ enhanced affinities of FMDV capsids to both receptors and, meanwhile, induced the highest antibody response in animal experiments. The binding of Ni²⁺ and Ca²⁺ enhanced affinities of FMDV capsids to the integrin β6 receptor only and also induced antibody level higher than the no-metal capsids did but less significantly than the Cu²⁺ binding capsids did.

One-step purification of FMDV based on the specific binding of FMDV with Ni²⁺ and Cu²⁺.

Based on the specific interaction between Ni²⁺, Cu²⁺, and the numerous transition metal-ion binding sites on FMDV capsids revealed in this study, we expected that immobilized-metal affinity chelating (IMAC) chromatography would be suitable for purification of FMDV antigen. To verify this hypothesis, agarose-based IMAC adsorbent chelated with Ni²⁺ or Cu²⁺ was used to capture FMDV from crude inactivated FMDV solution. HPSEC results showed that both of these two metal-chelated chromatography methods could specifically adsorb FMDV antigen and remove impure components. FMDV antigens with purity of >90% (based on HPSEC analysis) were obtained by this one-step purification (Fig. 6). Unlike IMAC purification of recombinant FMDV with extra inserted His tag in the VP1 GH-loop (28), our results reported for the first time purification of FMDV by using the natural transition metal-ion binding sites on the capsids, thus, simplifying the purification process.

FIG 6.

HPSEC analysis of crude inactivated FMDV solution (black line), purified FMDV by one-step Ni²⁺ chelating (red line) or Cu²⁺ chelating (blue line) chromatographic purification.

DISCUSSION

Different amino acids are involved in the binding of transition metal ions (Cu²⁺ and Ni²⁺) and nontransition metal ions (Ca²⁺).

Obtaining stable antigens is crucial to the success of the vaccine industry. In this study, inspired by the rich distribution of histidine residues at the interpentameric interface of FMDV capsids, we proposed to enhance the stability of FMDV by binding transition metal ions. The results of stoichiometry and thermodynamics of the interaction between FMDV capsids and Cu²⁺ and Ni²⁺ confirmed the existence of this interaction. As listed in Table 2, the binding capacities of Cu²⁺ and Ni²⁺ on capsids were 33- and 16-fold higher than that of the nontransition metal ion Ca²⁺, and the affinities of Cu²⁺ and Ni²⁺ to capsids were 15.9- and 6.6-fold higher than that of Ca²⁺. These differences were determined by the binding tendency of two kinds of metal ions to amino acids. Transition metal ions mainly incline to coordinate with amino acids containing aromatic nitrogen (e.g., histidine and tryptophan) and sulfur (e.g., cysteine) (13, 14, 17). In addition, they also have affinities for glutamate and aspartate, which contain carboxylate (29, 30). However, the nontransition metal ions like Ca²⁺ mainly coordinate with amino acids containing carboxylate or aliphatic nitrogen, known as hard Lewis bases.

Figure 7 schematically depicted the structure of three adjacent pentameric subunits of FMDV capsids, as well as the binding of different metal ions on the capsid. Obviously, the binding sites of two kinds of metal ions on intersubunits of FMDV capsids are different. As shown in Fig. 7a and b, the Glu6 in VP2 located on the icosahedral 3-fold axes of the FMDV capsid are the main binding sites for Ca²⁺. These Ca²⁺ binding sites are conserved in all of the picornaviruses (18). According to Acharya et al., the binding of Ca²⁺ would tighten the icosahedral 3-fold axes of the capsid structure (18, 31) (Fig. 7b). For transition metal ions, they can coordinate with histidine, aspartic acid, and glutamic acid located at the interpentameric interface of FMDV (6, 7) (e.g., E11, H65, H87, D96, and H209 on VP2 and H144, D148, and H191 on VP3) (see Fig. 7c and d). Cu²⁺ can even bind to single histidine, while Ni²⁺ requires coordination with more than one amino acid, such as histidine clusters (13). Therefore, the binding of Cu²⁺ on capsids had the maximum binding sites and the strongest affinity, which was conducive to forming the most stable capsids-metal complex (Table 2).

FIG 7.

(a) Structure of three adjacent pentameric subunits of FMDV capsids. Subunits are shown in spheres and colored as in Fig. 1. The interpentameric interface is marked with a black dashed line. Ca²⁺ binding sites and transition metal ion binding areas on capsids are circled in red and blue, respectively. (b) Binding modes of Ca²⁺ at icosahedral 3-fold axes of capsids as reported by Acharya et al. (18). (c) Putative binding sites and modes of Ni²⁺ at the interpentameric interface of capsids. (d) Putative binding sites and modes of Cu²⁺ at the interpentameric interface of capsids. Amino acids involved in binding with metal ions are depicted with sticks. Ca²⁺, Ni²⁺, and Cu²⁺ are depicted with green balls, red balls, and blue balls. Coordination bonds are depicted with black dash lines in panels b, c, and d.

Transition metal ion bridges were formed at the interpentameric interfaces to stabilize the FMDV.

Combined with the results of stability study and the speculative binding modes of each metal ion on FMDV capsids, the stabilization mechanism was proposed. Firstly, the binding of Ni²⁺ and Cu²⁺ at the interpentameric interface of capsids could tighten the link between adjacent subunits, which was similar to the mechanisms of Ca²⁺ stabilizing capsids. However, the binding of Ni²⁺ or Cu²⁺ introduced coordination bonds, thus, forming unique “transition metal ion bridges” to tighten the cross-link of adjacent pentamers (Fig. 7c and d). Previously, studies have shown the dissociation of FMDV capsids starting from the adjacent pentameric subunits. Enhancing the interaction between adjacent pentamers has been proven to stabilize capsids (3, 8). Therefore, it is also possible to stabilize capsids by these transition metal ion bridges. As mentioned above, Cu²⁺ was easier to coordinate with more amino acids than Ni²⁺. The binding of Cu²⁺ could form more transition metal ion bridges (Fig. 7c and d), thus, showing the strongest stabilization effects.

Secondly, the binding of the positively charged Cu²⁺ or Ni²⁺ would also partially counteract the electrostatic repulsion between the negatively charged amino acids at the interpentameric interface. These amino acids were considered to be the main cause of thermal dissociation of FMDV capsids (6, 12). Furthermore, the binding of Cu²⁺ or Ni²⁺ could reduce the protonation of histidine at the interpentameric interface of capsids and attenuate the electrostatic repulsion effect, thus, enhancing the acid-resistant stability. In contrast, the interaction between Ca²⁺ and histidine was too weak to inhibit the protonation of histidine. Therefore, we observed that the binding of Cu²⁺ and Ni²⁺ could both enhance the thermo- and acid-resistant stability of FMDV capsids, while the binding of Ca²⁺ only enhanced the thermostability (Fig. 4a and c).

FMDV capsids bound with metal ions could induce higher specific antibody responses in mice.

In addition to enhancing stability, the binding of transitional metal ions has also shown potential to improve the immunogenicity of the FMDV capsids (Fig. 5a), though the overall immunogenicity of FMDV should be further evaluated through neutralizing antibody detection and in vivo virus challenge experiments. The mechanisms of metal ions improving antibody responses would involve two aspects as follows: effect of metal ion binding on FMDV-receptor interactions and on the stability FMDV capsids.

Ca²⁺ was reported to bind with aspartate in the RGD motif on the GH-loop of FMDV and the serine in the integrin β subunits (2). As discussed above, Cu²⁺ and Ni²⁺ could also bind to the carbonyl group of serine and the carboxyl group of aspartates. Therefore, the binding of Cu²⁺ and Ni²⁺ enhanced the affinity of FMDV to integrin β6, similar to Ca²⁺, but was slightly weaker (Fig. 5b).

The HS binding sites on FMDV include His195 on VP1; Lys134, Arg135, and Tyr138 on VP2; and Arg56, Gly59, Gly60, Ser87, and Asn88 on VP3 (27). Unlike integrin receptor, the interaction between HS and FMDV is related to sulfated sugar binding and does not need to be mediated by metal ions. However, a significantly enhanced affinity was observed between HS and the FMDV capsids binding Cu²⁺ (Fig. 5b). Considering that the binding of Cu²⁺ induced significant tertiary structural changes in FMDV capsids (Fig. 3c), we speculated that some conformational changes occurred at amino acids close to the HS binding sites, thus, enabling the capsids to be more favorable for HS binding.

In addition, we also observed that the stabilization effects of metal ions on FMDV capsids were consistent with their ability to induce a specific antibody response (Fig. 4 and 5). Although there is still a lack of clear explanation on how the in vitro stability of FMDV affects its in vivo immunogenicity, both mathematical modeling and some experimental studies showed that the high stability of FMDV did benefit to induce higher specific antibody responses (3, 10). Therefore, it was believed that both the enhanced capsid stability and the increased receptor binding affinities mediated by metal ions, at least partially, are responsible for the ability to elicit specific antibody immune response of the FMDV capsids.

When used in vaccine, the safety of using Cu²⁺ and Ni²⁺ should be considered. Cu²⁺ and Ni²⁺ are essential trace elements for animals and play important roles in maintaining enzyme and protein activities in vivo (32, 33). Although excessive intake of Cu²⁺ and Ni²⁺ can induce some adverse effects, the amount of Cu²⁺ or Ni²⁺ in each dose of vaccine sample (contains FMDV ca. 4 μg) is about 375 ng and 169 ng, even at their highest binding levels. This intake amount is far below the normal daily intake (33, 34). Therefore, there is no safety concern for using Cu²⁺ and Ni²⁺ in vaccines.

In addition to Cu²⁺ and Ni²⁺, we also investigated the possible binding and stabilizing effect of Zn²⁺ on the FMDV capsids, as Zn²⁺ is usually considered more biocompatible. Nevertheless, Zn²⁺ will precipitate at a pH of >6.2, under which pH FMDV will dissociate immediately. Therefore, we were not able to directly study the interaction between Zn²⁺ and FMDV in solution. Conversely, our group has recently proven that FMDV can specifically bind to Zn²⁺ chelated chitosan nanoparticles. This novel particulate delivery system based on FMDV-Zn²⁺ coordination interaction was found to stabiliz the antigen on the nanoparticles and induce higher cellular and humoral immune responses in immunized mice (35).

The used of metal ions such as Ca²⁺ or Mg²⁺ to stabilize viral capsids has been widely reported (2, 15, 16), but the strategy of stabilizing FMDV by transition metal ions is designed for the key mechanism of the dissociation of FMDV capsids. For other viruses, this strategy may not be directly used. However, we can use the idea proposed in this study to select specific metal ions or stabilizers according to the specific amino acid sites of other virus particles.

Although the exact binding modes of Cu²⁺, Ni²⁺, and other transition metal ions on FMDV capsids need to be further elucidated by utilizing technologies like cryo-TEM, X-ray diffraction (XRD), and molecular simulation, the interesting results reported in the present work do provide new guidelines for designing an easy-to-use strategy suitable for large-scale production of FMDV vaccine antigen with higher stability and higher ability to elicit specific antibody responses.

MATERIALS AND METHODS

Materials.

The 2-amino-2-hydroxymethyl-propane-1,3-diol (Tris), polyethylene glycol 6000 (PEG 6000), CuSO₄·5H₂O, NiSO₄·6H₂O, CaCl₂·2H₂O, NaCl, NaOH, and HCl were purchased from Sinopharm Chemical Reagent (China). Fluorescent dye Sypro green was purchased from Sigma-Aldrich (USA). Heparin sulfate was purchased from Guchen Biological (China). Recombinant mouse integrin β6 was purchased from ImmunoClone (USA). All other reagents were analytical grade and all solutions were prepared using Milli-Q grade water (Millipore, USA).

Purification of FMDV.

Inactivated O strain FMDV (Lanzhou Veterinary Research Institute, Lanzhou, China) was purified under well-established condition as reported previously (36). Purity and contents of FMDV capsids were measured by high-performance size-exclusion chromatography (HPSEC) (37).

Structural and sequence analyses.

The protein data bank atomic coordinates for crystal structure of FMDV-type O (PDB accession number 1FOD) were analyzed using PyMOL software (Schrodinger, USA). The theoretical molecular weight of FMDV capsids was estimated to be about 480 × 10⁴ Da according to the amino acid sequence of capsids (PDB accession number 1FOD).

Determination of stoichiometry and thermodynamics of the interaction between metal ions and FMDV.

Purified FMDV (1 ml, 0.055 μM in 20 mM, pH 7.8 Tris buffer) were mixed with 10 μl 60 mM CuSO₄·5H₂O, NiSO₄·6H₂O, or CaCl₂·2H₂O solution prepared in the same buffer, respectively, which corresponded to a 9-fold excess of the total histidine residues in capsids to ensure supersaturated binding of the metal ions. The binding lasted for 12 h on an incubator at 4°C. After removing unbound metal ions by using a G-25 desalting column (GE Healthcare, USA), the capsids were collected and analyzed by ICP-MS to precisely quantify the amount of metal ions tightly bound with the FMDV capsids.

The thermodynamic parameters, including ΔG, ΔH, and ΔS, describing the metal ion binding process were determined by isothermal calorimetric titration (ITC; Malvern Panalytical, UK). Briefly, 3 mM CuSO₄, NiCl₂, or CaCl₂ were titrated into 0.63 μM, 0.83 μM, or 0.73 μM FMDV capsid solution at a time interval of 150 s, respectively. Three measurements were performed for each sample. The reported experiment values were the average of individual best-fit values determined by one-site model in Origin software provided by MicroCal ITC. The results were presented by showing the baseline-adjusted experimental titration data on the top and the peak-integrated background-subtracted concentration-normalized molar heat flow per aliquot versus the titrant-to-sample molar ratio on the bottom. ΔG was calculated from the following equation:

$$△G=△H-\text{\hspace{0.17em}T}△S$$

where ΔH and ΔS were determined from the best fit of the ITC experimental data.

Determination of structural changes of FMDV capsids after binding metal ions.

Morphology of FMDV capsids was observed using transmission electron microscopy (TEM; Royal Philips Electronics, Amsterdam). The hydraulic diameters of the capsids were analyzed using the DynaPro NanoStar module of multiangle laser light scattering (MALLS) detector (Wyatt Technology, USA) connected with HPSEC. Fluorescence spectra (Hitachi F-4500, Japan) and circular dichroism spectra (Jasco J-810, USA) were applied to measure the tertiary and secondary structures of the capsids. Three measurements were performed for each sample.

SDS denaturation test by MST.

The SDS denaturation test (SDS test) in MST was applied to distinguish between fluorescence changes caused by the structural change of FMDV capsids and those caused by quenching effects of metal ions. First, a Cu²⁺ solution was prepared with a concentration of 10 mM, 5 mM, 2.5 mM, 1.2 μM, 0.6 μM, and 0.3 μM, respectively. Then, 40 nM RED-NHS-labeled FMDV solution was mixed with equal volume of Cu²⁺ solution. Finally, dip a Monolith NT.115 capillary (Nanotemper Technologies, Germany) into each sample and start the SDS test measurement. At the same time, mix capsids-metal solution with equal volume of SDS mix (4% SDS, 40 mM dithiothreitol [DTT]). After incubating the reactant mixture for 5 min at 95°C, the same measurement was performed. If the variations of fluorescence intensity of samples before denaturation were higher than 20% and those after denaturation were lower than 20%, this indicates that the decrease of fluorescence intensity was due to the structural changes of samples.

Analyzing the stability of FMDV capsids after binding metal ions.

Differential scanning fluorescence (DSF) was applied to investigate the effect of the binding of metal ions on the thermostability of FMDV capsids. DSF was performed according to our previous report (24). Briefly, 18 μl 52 nM capsids solution was mixed with 2 μl 1:100-fold diluted Sypro green dye. DSF experiments were performed with a real-time PCR instrument (Applied Biosystem, USA). A scan rate of 1°C min⁻¹ from 25°C to 95°C was used. Three measurements were performed for each sample. The T_m values were calculated according to the maximum of the first derivative [d(RFU)/dT] plot of the fluorescence curve.

The acid-resistant stability of FMDV capsids was expressed by pH₅₀ (9), which was defined as the pH leading to the 50% dissociation of intact capsids. To measure pH₅₀ of capsids, 1 volume of capsids (52 nM in 20 mM Tris buffer, pH 8.0) was diluted with 3 volumes of 200 mM phosphate buffer with pH ranging from 8.0 to 6.0. After mixing at 25°C for 8 min, the remaining intact capsids were determined by HPSEC immediately without neutralization. Three measurements were performed for each sample. The pH₅₀ was obtained by fitting the curve of the contents of intact capsids against pH values by a four-parameter equation using GraphPad Prism 7.0 software.

Animal experiment and specific antibody response detection.

Female BALB/c mice (6 weeks old) were purchase from Beijing Vital River Laboratory Animal Technology Co., Ltd. The mice were divided into four groups with five mice in each group. Mice were immunized subcutaneously with various samples at 100 μl. Samples were prepared by binding 8 nM (corresponding to about 40 μg/ml) inactivated FMDV with maximum ratio of three metal ions respectively. The sample without binding metal ions was used as comparison. Then, samples were emulsified with Montanide ISA 206 (Seppic, France) at a volumetric ratio of 1:1. Mice were immunized on days 0 and 14, and serum samples were collected on days 14 and 28. The level of FMDV-specific IgG antibody in mouse serum was determined by enzyme-linked immunosorbent assay (ELISA) as described previously (38).

Protocols for animal experiments were performed in strict accordance with the Experimental Animal-Guidelines for Ethical Review of Animal Welfare (GB/T 35892-2018) and were approved by the Committee on the Ethics of Animal Experiments of the Institute of Process Engineering at the Chinese Academy of Sciences (Beijing, China).

Determination of affinities of FMDV to cell surface receptors mediated by metal ions.

To investigate the possible effects of metal ion binding on affinity of FMDV to cell surface receptor, the K_d values of interaction between β6 subunits of integrin or heparin sulfate (HS) and FMDV capsids, with or without metal ion binding, were analyzed by MST. The concentration ranges of integrin and HS were set from 2.5 μM to 0.076 nM and 52 μM to 1.6 nM, respectively. Three measurements were performed for each sample, and the data were analyzed by MO. Control software (NanoTemper Technology, Germany) and the dissociation constant, K_d, were determined.

Purification of FMDV by immobilized metal affinity chelating chromatography.

Crude inactivated FMDV solution was directly loaded onto a chromatographic column (GE Healthcare, USA) filled with homemade Ni-Sepharose medium, which has been preequilibrated in equilibrium buffer (0.02 mol/liter phosphate buffer, pH 7.5, containing 0.15 mol/liter NaCl), followed by eluting with 0.2 mol/liter imidazole in equilibrium buffer. By replacing the Ni²⁺ ligand with Cu²⁺, the above experiments were repeated following the same procedure. The purity of FMDV was measured by high-performance size-exclusion chromatography (HPSEC) (37).

Statistical analysis.

Data analysis was performed with Origin 8.0 and GraphPad Prism 7.0 software and presented as mean with standard deviation (SD). Statistical significance of difference was determined by the one-way analysis of variance (ANOVA). P values of less than 0.05 were considered statistically significant. Asterisks indicate significant differences (*, P < 0.05; **, P < 0.01).