Design and preliminary application of affinity peptide based on the structure of the porcine circovirus type II Capsid (PCV2 Cap)

Materials

Thirteen peptides, named L1∼L13, were synthesized and purified to at least 90% by GL Biochem. (Shanghai, China). The PCV2 Cap protein, porcine epidemic diarrhea virus S protein (PEDV-S), porcine reproductive and respiratory syndrome virus GP5 protein (PRRSV-GP5) and the polyclonal antibodies of PCV2 Cap (anti-Cap) were generated in our laboratory. A carboxyl sensor chip was purchased from Nicoya (Beijing, China). The streptavidin conjugated to horseradish peroxidase (SA-HRP) was purchased from Bioss Biological Engineering (Beijing, China). The goat anti-mouse IgG conjugated to horseradish peroxidase (IgG-HRP) was purchased from Abbkine (Wuhan, China). The NHS-activated magnetic agarose beads (NaMB) were purchased from Enriching Biotechnology (Shanghai, China). Bovine serum albumin (BSA) was purchased from Saibao Biotechnology (Yancheng, China).

Molecular docking virtual screening

(1) Preparation of protein crystal structure. The 3D structure of PCV2 Cap (PDB ID: 3R0R) (Khayat et al., 2011) was obtained from the protein data bank. First, the PCV2 Cap protein was analyzed by the option of Prepare protein structure of the SYBYL software (Tripos, USA). Backbone, side-chain, termini treatment, and protonation type were fixed following its suggesting. Hydrogens and charges were added. Second, in the stage minimization option, Biopolymer Hydrogens, Waters, and Side-chains were minimized 100 times per step. Third, in the Surflex-Dock-Define SFXC File option, the Threshold value was 0.5, the Bloat value was 10, and the remaining parameters were default values. Finally, the Surflex-Dock Screen (SFXC) File was generated for next docking.

(2) Virtual peptide library design. The initial conformation of the docking peptide was constructed using the Biopolymer/Build/Build Protein module in the SYBYL-X2.0 software. All peptides were designed as a series of linear peptides with different lengths. The peptide library was written to a single file with a library capacity of 16,000. Then, the hydrogen atoms were added and MMFF94 charges were loaded. The peptide was minimized by 1,000 cycles using the MMFF94 force region to obtain its optimum conformation (Liu et al., 2007).

(3) Molecular docking. Molecular docking simulation was carried out with the program FlexX Docking Suite embedded in the SYBYL package. PCV2 Cap was prepared by Preparing Protein Structure embedded in the SYBYL (Rarey et al., 1996).

(4) Evaluation of results. To assess the binding affinity in the docking of the ligand and protein, the Consensus Score (CScore) was referred to as the most criterion (Yu et al., 2018). CScore (Welch, Ruppert & Jain, 1996) is a consensus score function for ligand–receptor affinity. It integrates four scoring functions that include D Score (Jones et al., 1997), C Score (Eldridge et al., 1997), G Score (Kuntz et al., 1982), and PMF Score (Muegge & Martin, 1999) to somehow reduce the bias caused by scoring. Based on the criterion of virtual screening (CScore ≥ 5), we obtained 67 peptide sequences. In SYBYL, CScore always reports the output of the docking engine as total scores.

ELISA assay for affinity and specificity

The ELISA plate was coated with 50 µL of a 10 µg/mL purified Cap protein solution and incubated for 12 h at 4 °C. Blocking buffer containing 1% BSA was incubated for 1 h at 37 °C. After washing with PBST (PBS containing 0.5% Tween20), biotin (bio)-labeled peptides (1 µg/mL) were separately added and incubated for 30 min at 37 °C. The plate was washed as before, SA-HRP (1:1,000) and IgG-HRP (1:5,000) were separately added and incubated for 30 min at 37 °C. Meanwhile, Phosphate Buffer Saline (PBS) and anti-Cap antibodies were used as negative and positive controls, respectively. The reaction was developed by using 3,3,5,5-tetramethylbenzidine (TMB) as substrate. The OD value of each well was measured at 450 nm using an ELISA plate reader.

The P/N value and cross-reactivity of the candidate peptide were determined according to the ELISA method described above. When the ratio between P/N [(OD_sample-OD_blank)/(OD_negative-OD_blank)] is more than 2.1, the peptides exhibiting high specificity were chosen for posterior experiments using pET28a as negative control. Meanwhile, the cross-reactivity of affinity peptides was evaluated using proteins outside the circovirus genus, namely PEDV-S protein, PRRSV-GP5 protein, and BSA. Their coating concentration was 10 µg/mL.

Kinetic dissociation measurements

The equilibrium binding constant (K_D) of peptides was determined using an Open SPR instrument (Nicoya, Canada). First, purified Cap protein was covalently immobilized on a carboxyl Au colloidal nanoparticles chip. Second, the running buffer (PBS, pH 7.4) was flowing through the chip (for preparation of the PCV2 Cap protein biochip; see ESI 1) until a smooth testing baseline was obtained. Then, all peptides were respectively dissolved in PBS at predetermined concentrations and injected to detect the resonance signal changes. In each cycle, 300 µL of peptide solution was injected into the flow cell of the Cap chip for 5 min. Finally, the data were assessed to determine the association constant (k_a, 1/Ms), the dissociation constant (k_d, 1/s), the equilibrium binding constant (K_D, M), and K_D = kd/ka via Trace Drawer 1.7 (Nicoya, Canada).

Effect of pH

First, the NaMB-L11 affinity adsorbent was prepared (see ESI 2). Next, the NaMB-L11 (100 µL) was added into 100 µL of recombinant PCV2 Cap solution (for the expression of the recombinant PCV2 Cap protein; see ESI 3) with different protein concentrations (0.5∼6 mg/mL) at various pH values (10, 9, 8, 7, 6, 5, and 4). The mixtures were then shaken 2 h at 25 °C. The amount of adsorbed protein was calculated according to a mass balance Langmuir equation was fitted to the adsorption isotherm. The formula is as follows: ${\displaystyle {\mathrm{Q}}^{\mathrm{\ast }}=\frac{{\mathrm{Q}}_{\mathrm{m}}\times {\mathrm{c}}^{\mathrm{\ast }}}{{\mathrm{K}}_{\mathrm{d}}+{\mathrm{c}}^{\mathrm{\ast }}}}$

Where Q* and C* are the equilibrium adsorption capacity of resin (mg/g resin) and the equilibrium protein concentration in the liquid phase (mg/mL), respectively. Q_m represents the saturated adsorption capacity (mg/g resin), while Kd is the apparent dissociation constant (mg/mL).

Effect of ligand density

First, the affinity adsorbent NaMB-L11 with different ligand densities was prepared (see ESI 2). Second, the effect of different ligand densities on the adsorption of PCV2 Cap protein was measured by the same method as above. Third, the NaMB-L11 (100 µL) with different ligand densities (see ESI 4) was added to 100µL of recombinant PCV2 Cap solution (PBS, pH 9) containing different concentrations (0.5∼6 mg/mL). Subsequently, the adsorption operation until reaching equilibrium was carried out as described above. Finally, SDS–PAGE was used to monitor the purification effect. The purity of PCV2 Cap protein was analyzed by gel-pro Analyzer. The recovery rate was calculated using a micro-BCA Kit (Thermo Fisher Scientific).

Analysis of PCV2 Cap protein immunoactivity after elution

Frist, the PCV2 Cap protein sample was eluted using 100 mM glycine-hydrochloric acid (pH 2.4), and then the sample was neutralized with 100 mM Tris–HCl (pH 8). Second, protein samples were separated by 12% SDS-PAGE. Third, the protein samples were transferred to a polyvinylidene fluoride (PVDF) membrane that was blocked with 5% non-fat milk for 1 h at 37 °C. Then, the anti-Cap was added and incubated for 12 h at 4 °C. Subsequently, the PVDF membrane was washed 5 times with PBST and incubated with a secondary antibody for 1 h at room temperature. Finally, the detection was performed using an Enhanced Chemiluminescent (ECL) kit (NCM, Suzhou, China).

Statistical analysis

All analyses were conducted using GraphPad Prism version 5.0 software and Excel. Values were expressed as mean ± SEM. Cross-reactivity between groups was performed using a two-way ANOVA. Statistical significance was determined at P ≤ 0.05 (*), P ≤ 0.01 (**), or P ≤ 0.001 (***).

Molecular docking virtual screening

Based on the detection of active pockets, the amino acid region at positions 63-120 of the PCV2 Cap protein was predicted as docking pockets (Fig. 1A). In this article, the affinity peptide length was randomly assigned to 2∼9 amino acids because peptide ligands are rarely outside these boundaries (Jeziorna et al., 2014). Meanwhile, through preliminary screening of the experiments (see ESI 5), we obtained and synthesized 13 potentially high affinity candidate peptides from 67 peptide sequences.The 13 peptide sequences and CScore values are listed in Table 1. The L11 affinity peptide consists of two uncharged amino acids (Ser and Gln), two acidic amino acids (Asp and Glu), and two hydrophobic amino acids (Tyr and Trp). Its 2D structure is shown in Fig. 1B. The hydrogen bond (yellow dotted line) of the interaction between L11 peptide and PCV2 Cap is shown in Fig. 1C. The key amino acids at the binding site of PCV2 Cap protein are 64-Val, 65-Asp, and 73-Asp. From the analysis of the charge interaction between L11 peptide and the active site of PCV2 Cap protein (Fig. 1D), it was found that Glu of peptide L11 had strong intermolecular electrostatic attraction with His-68 and Lys-174 of PCV2 Cap, respectively, and there was also strong intermolecular electrostatic attraction between Asp and Lys-127 of PCV2 Cap. Therefore, the electrostatic interaction may play an important role in the binding of peptide L11 and PCV2Cap.

ELISA analysis affinity and specificity

Known from the ELISA results (Fig. 2), the OD₄₅₀ values of 13 candidate peptides were obtained ranging from 0.836 to 2.534. Among them L4, L7, L9, L11, and L13 had the best ability to bind with Cap. As a result, the above five peptides were selected for further analysis.

The specificity of these five peptides was evaluated by P/N values and cross-reactivity. When the P/N value is greater than 2.1, the specificity of the peptide is good (Zhang et al., 2018). Figure 3A shows that the five peptides for Cap had P/N values greater than 2.1. Importantly, peptides L4 and L11 are the most promising due to their higher P/N values, 4.96 and 5.57; respectively. For the cross-reactivity of three proteins (Fig. 3B), L9 was significantly different from L4, L7, and L11 peptides (P ≤ 0.01), and was also significantly different from L13 (P ≤ 0.001). The results indicated that L9 had the lowest cross-reactivity.

LSPR of affinity peptides analysis

Based on the fitted different curves by the 1:1 binding interaction model (Figs. 4A–F), the relevant kinetic data of five peptides associated with the PCV2 Cap, were obtained (Table 2). As can be seen, all the peptides exhibited different binding strength for Cap with various K _D values. The L11 peptide had the highest binding affinity compared to other peptides, with a K _D of 103 nM. These results demonstrate that the affinity of the five peptides corresponds well with the result in the ELISA. From the mentioned description, we have found that L11 has the highest affinity and specificity, thus it is considered as the best peptide ligand.

Effect of pH

In Fig. 5A, it is presented that the adsorption capacities of the NaMB-L11 greatly relied on pH. At different pH values, with the increase of sample concentration, the adsorption capacity of the Cap protein increased gradually. The maximum binding capacity could be found at pH 9 (Fig. 5B). The values of Qm and Kd are listed in Table 3. SDS-PAGE analysis that: at pH 9, NaMB-L11 could specifically adsorb Cap protein with the maximum adsorption capacity, while the non-specific binding of contaminants is shown at pH 8-4. The control BSA-blocked NaMB had almost no adsorption of the PCV2 Cap protein. In order to avoid non-specific adsorption, the NaMB need to be blocked with 1% BSA. Because the NaMB will carry some BSA, there is a weak protein band between 63–75 kDa in Fig. 5B corresponding to BSA.

Effect of ligand density

From Fig. 6A, it is evident that NaMB with higher ligand density at the range tested had higher adsorption capacities. The values of Qm and Kd are listed in Table 4. Figure 6B showed that for the high ligand density region (5∼2.5 mmol/L), increasing the adsorption of Cap protein may also lead to an increase in the adsorption of non-specific substances. For the low ligand density range (0.08∼0.31 mmol/L), the recovery of protein was significantly reduced. For the medium ligand density range (1.25∼0.63 mmol/L), the NaMB-L11 adsorbent presented a better compromise between purity and recovery with a recovery yield of 90% and protein purity of 98%, especially at 1.25 mmol/L. The recovery and purity of PCV2 Cap at each ligand density are shown in Table 5. Thus 1.25 mmol/L was chosen as the best ligand density.

Analysis of PCV2 Cap Protein immunoactivity after elution

As shown in Fig. 7A, SDS-PAGE results showed that the PCV2 Cap protein was efficiently dissociated in a glycine-hydrochloric acid solution and a single target band appeared at 26 kDa. As shown in Fig. 7B, WB results showed that the PCV2 Cap protein reacted specifically with anti-PCV2 antibody; indicating that the purified PCV2 Cap had immunogenicity. The results suggested that the NaMB-L11 affinity adsorbents can be used to purify recombinant PCV2 Cap.