Introduction

Foot-and-mouth disease (FMD) is one of the most frightening viral diseases, affecting approximately 70 wildlife species across three quarters of the world (Alexandersen and Mowat 2005; Robinson et al. 2011). This highly communicable viral disease causes great damage to the agricultural economy, with an estimated crisis cost of about US$6.5–21.0 billion in the endemic regions and over US$1.5 billion per year in FMD-free countries (Knight-Jones and Rushton 2013). The massive outbreak in South Korea in 2011 has been a cruel reminder of the devastating consequences on the global economy. The incursion of this highly contagious disease in FMD-free countries has resulted in dramatic decreases in livestock productivity, causing severe constraints in transboundary animal movement and restriction in the international trade of animal products (Sutmoller et al. 2003; OIE/FAO 2012; Smith et al. 2014). Thus, the availability of good diagnostic reagents coupled with rapid and sensitive diagnostic assays is one of the essential requirements for successful control of FMD.

The current established methods for the diagnosis of foot-and-mouth disease virus (FMDV) are based on the Office International des Epizooties (OIE) Terrestrial Manual. Various ELISA formats have been developed and employed and are the preferred procedures in the FMDV diagnosis. In recent years, ELISAs based on non-structural proteins (NSPs) of FMDV such as 3A, 3B, 3D, 2B, 2C, and 3ABC have been established for serological detection of past and present infections by the virus, regardless of the viral serotypes (Clavijo et al. 2004; Colling et al. 2014). They are also used to differentiate infected animals from those which had been vaccinated (Uttenthal et al. 2010; Ma et al. 2011). Nevertheless, one shortcoming of these diagnostic methods based on NSPs is the inability to rule out the heavily persistent infected animals in a vaccinated population, as they cannot be used to determine the animal vaccination status (OIE 2017). This is a challenge faced by many countries that desire to regain their FMD-free status following an outbreak. The diagnosis and certification of healthy uninfected animals into these FMD-free status countries as imposed by the World Trade Organisation (WTO) are crucial (OIE 2011). Thus, it is important to develop assays that can identify heavily persistent infected animals in a vaccinated animal population. This limitation can be solved using the FMDV structural proteins (SP). SP-based detection methods, which are serotype specific, are often used for the identification of the vaccination status of animals (Hamblin et al. 1986; Alexandersen et al. 2003). Therefore, SP-based detection methods coupled with NSP-based detection methods can be used to identify infected animals within a vaccinated population, where these heavily persistent infected animals serve as FMDV reservoir (Uttenthal et al. 2010). In view of this, we aimed to develop a SP-based detection method by displaying conserved epitopes of FMDV VP1 structural protein on phage T7. These recombinant bacteriophages were used to establish a phage ELISA.

VP1 has been intensively studied due to its functions in virus attachment and entry, protective immunity, and serotype specificity (Oem et al. 2005; Fowler et al. 2011). It also contributes to the formation of five main antigenic sites for neutralising antibodies in FMDV type O1 (Ptaff et al. 1988; Kitson et al. 1990). Despite the high variability in the nucleotide sequences of the VP1-coding region, approximately 26% of its residues were found to be highly conserved among serotypes (Carrillo 2012). These residues were found to reside in the prominent surface-exposed G-H loop of the VP1 protein and have been described as a structurally disordered major antigenic site (Acharya et al. 1990; Curry et al. 1996).

Previously, we displayed residues 131–170 of the VP1 capsid protein on phage T7, and the resultant recombinant phage, namely T7-FOVP1131–170, was used to establish a phage ELISA for detecting FMD antibodies in bovine sera (Wong et al. 2013). Within this 40-residue region (residues 131–170), two highly conserved regions at positions 145–152 and 159–170 were identified among different FMDV serotypes, postulated to be the binding sites for anti-FMDV antibodies. In the present study, the 8- and 12-residue conserved regions were displayed on phage T7 and used to develop phage ELISAs. The chimeric phage T7 displaying the 12-residue VP1159–170 epitope detected all the vaccinated bovine serum samples, suggesting the potential of this smaller functional epitope as a diagnostic reagent for the identification of vaccination status in cattle.

Materials and methods

Sera

The sera of FMDV-vaccinated (trivalent FMD vaccine consisting of inactivated FMDV serotypes O1 Manisa, A Malaysia 97 and Asia 1 Shamir) and FMDV-naïve bovine were kindly provided by Dr. Baljit Singh (Universiti Putra Malaysia) and the Department of Veterinary Services (DVS) in Kubang Kerian, Kelantan, Malaysia. The presence and absence of antibodies against FMDV serotype O in the collected bovine serum samples were confirmed using the PrioCHECK® FMDV type O solid-phase competitive ELISA test kit (Prionics, Schlieren-Zurich, Switzerland).

Homology modelling

A homologous model of the VP1 (residues 131–170) structural protein of the FMDV strain O1 Kaufbeuren [accession no. X00871] (Forss et al. 1984) was generated using the SWISS-MODEL programme (Biasani et al. 2014). A BLASTP search of the VP1 protein sequence against the proteins available in the PDB database was carried out (RID: FSCX5EJG015). Based on the sequence identity and query coverage, the FMDV capsid protein [PDB ID: 1FOD] (Logan et al. 1993) was chosen as the most suitable template for protein structure modelling. The primary sequences of these two proteins were stereo superimposed using the DeepView (Swiss-PdbViewer) programme (Guex and Peitsch 1997).

Amino acid sequence analysis of FMDV VP1131–170

A highly conserved sequence of VP1131–170 was deduced by comparing the amino acid sequences from groups with different FMDV serotype(s): -O, -A, -Asia, -C, -SAT-1, -SAT-2 and -SAT-3. The sequences of the isolates were obtained from the National Center for Biotechnology Information (NCBI, USA). Conserved amino acid residues (identical amino acid residues present at high frequency) of the FMDV isolates within each serotype (20 serotype O, 10 serotype A, 10 serotype Asia 1, 10 serotype C, 5 serotype SAT-1, 5 serotype SAT-2 and 5 serotype SAT-3) were analysed (Wong et al. 2013). The alignment of amino acid sequences was performed with the Multiple Sequence Alignment programme (Clustal-W) provided in the SDSC Biology Workbench software (http://workbench.sds.edu, Version 3.2, 1999; Board of Trustees of the University of Illinois).

DNA inserts of FMDV VP1145–152 and VP1159–170

All oligonucleotides and primers used in this study are summarised in Table 1. The corresponding oligonucleotides encoding the two highly conserved regions (residues 145–152: RGDLQVLA and residues 159–170: LPTSFNYGAIKA; Wong et al. 2013) of the VP1 protein of FMDV strain O1 Kaufbeuren, with EcoRΙ and HindΙΙΙ restriction endonuclease sites incorporated at the 5′ and 3′ ends, respectively, were synthesised by Ultramer™ DNA Oligo (IDT, Iowa, USA). These oligonucleotide pairs could self-anneal using the hybridisation method as described by Rémond et al. (2009).

Table 1 Oligonucleotides used for the construction of chimeric T7 phages
Full size table

Cloning of the FMDV-coding regions in T7 Select 415-1b vector arms and generation of chimeric T7 phages (T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170)

The construction of the chimeric T7 phages was performed as described previously by Wong et al. (2013) with slight modifications. In brief, the annealed double-stranded oligonucleotides were first cloned into the pGEM®-T vector system Ι (Promega, Wisconsin, USA). Two pairs of primers [pGEM-8-VP1145–152-for/ pGEM-8/12-VP1-rev and pGEM-12-VP1159–170-for/ pGEM-8/12-VP1-rev] (Table 1) were used to amplify the coding fragments of the FMDV epitopes by PCR. PCR was performed in reaction mixtures (50 μl) containing 1× buffer [2 mM Tris-HCl (pH 8.0), 10 mM KCl, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% (v/v) Triton X-100, 0.1 mg/ml nuclease-free BSA, 2 mM MgCl2], dNTPs (0.2 mM each; 5 μl), forward primer (100 μM; 0.5 μl), reverse primer (100 μM; 0.5 μl) and DreamTaq™ DNA Polymerase (5 U/μl; 0.25 μl: Fermentas, York, UK). The PCR cycle conditions were denatured at 94 °C for 3 min followed by a 30-cycle reaction (94 °C, 30 s; 48 °C, 1 min; 72 °C, 42 s) and a final elongation at 72 °C for 5 min. The PCR products were digested with EcoRΙ and HindΙΙΙ restriction endonucleases. The digested products were ligated to the T7 Select 415-1b vector (Novagen Merck KGaA, Darmstadt, Germany), which had been digested with EcoRΙ and HindΙΙΙ. Ligation [0.5 μg insert, 0.5 μg vector arms, 40 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 10 mM DTT, 5 mM ATP, and 3 U of T4 DNA ligase (Fermentas, York, UK)] was carried out at 16 °C for 16 h. The ligation mixture (5 μl) was mixed with 25 μl of packaging extract at room temperature (RT) for 2 h. The reaction was stopped by adding Luria Bertani (LB) broth (Pronadisa, Basel, Switzerland; 270 μl), and the packaging efficiency was determined by phage titration assay using E. coli BL21 as a host (Adams 1959). The nucleotide sequences of the inserts were confirmed by sequencing using two sets of primers, FMD-8-VP1145–152-for/FMD-8/12-VP1-rev and FMD-12-VP1159–170-for/FMD-8/12-VP1-rev (Fig. S1).

Preparation and purification of chimeric phages

T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170 phages were propagated in the log phase culture of the E. coli strain BL21 [F−omp hsdS B (r B −m B ) gal dcm] at the multiplicity of infection (MOI) of 1, with T7-FOVP1131–170 and wild-type T7 phages served as the positive and negative controls. The mixture was further incubated with vigorous shaking at 37 °C until the complete lysis of cells was observed. DNase I (0.2 μg/ml; AMRESCO, Solon, USA) was added 15 min prior to removing the culture from the incubator shaker. The lysed culture was precipitated with polyethylene glycol 8000 [10% (w/v) AMRESCO, Solon, USA] and further purified by ultracentrifugation on a four-step gradient of caesium chloride (CsCl) as described by Tan et al. (2005). The band containing phage particles was isolated, and residual CsCl was removed by dialysis in TBS buffer [100 mM Tris-HCl (pH 8.0), 100 mM NaCl] with four consecutive changes. Purified T7 phages were quantified by phage titration assay as described below and stored at 4 °C.

Phage titration assay

The E. coli strain BL21 was inoculated into LB broth and was incubated with shaking at 37 °C to an OD600 of approximately 0.6. Purified phage dilutions were prepared, and 100 μl was added to log-phase E. coli host cells (250 μl). Molten top agarose (1% Bacto Tryptone, 0.5% yeast extract, 1% NaCl, 0.6% agarose; 3 ml) that had been melted and kept at 45 °C in a water bath was then added to the mixture, vortexed, and dispensed uniformly over the surface of pre-warmed LB agar plates. The plates were inverted and incubated for 3–4 h at 37 °C. The plaques were counted, and the phage titre was expressed in plaque forming units (pfu) per unit volume.

SDS-PAGE and western blotting

Chimeric T7 phages were boiled for 5 min in 6× sample buffer [62.5 mM Tris-HCl (pH 6.8), 30% glycerol, 5% SDS, 0.01% bromophenol blue, 5% β-mercaptoethanol]. The protein samples were electrophoresed on a 12% (w/v) polyacrylamide gel and were stained with Coomassie brilliant blue (CBB) R-250. For western blotting, the proteins were transferred onto a nitrocellulose membrane using a semidry blotting device (Bio-Rad Laboratories, Hercules, USA). The membrane was blocked with 5% (w/v) skimmed milk in TBS for 2 h and was probed with the T7-tag monoclonal antibody (1:10,000 dilution, Novagen Merck KGaA, Darmstadt, Germany) followed by the anti-mouse IgG conjugated to alkaline phosphatase (1:5000 dilution, Chemicon, Massachusetts, USA). Protein bands were visualised by adding a substrate mixture of nitro-blue tetrazolium (NBT) and 5-bromo-4-chloro-3′-indolyl phosphate toludinium salt (BCIP) in alkaline phosphatase buffer [100 mM Tris-HCl (pH 9.5), 100 mM NaCl, 1 mM MgCl2].

Antigenicity of FMDV epitopes displayed on T7 phages

Microtitre-plate wells (Thermo Scientific, Massachusetts, USA) were coated with different amounts of the purified chimeric T7 and wild-type T7 phages (10–5000 ng in sodium bicarbonate buffer, pH 9.6; 100 μl) and were blocked with 5% (w/v) skimmed milk in TBS for 2 h. The coated wells were washed three times with TBST [TBS (pH 7.4); 0.05% Tween 20], and the displayed epitope was detected with a primary antibody [vaccinated bovine sera or naïve bovine sera; 1:25 dilution with 2% (w/v) skimmed milk in TBS]. After 1 h incubation at RT, the wells were washed with TBST, and a secondary antibody [goat anti-bovine; 1:5000 dilution (Chemicon, Massachusetts, USA)] conjugated to alkaline phosphatase was added and incubated at RT for 1 h. Unbound antibodies were then washed off, and a substrate solution [0.001% (w/v) p-nitrophenyl phosphate (pNPP), 9.55% (v/v) diethanolamine, 0.5 mM MgCl2 (pH 9.8)] was added and incubated at RT for colour development. Absorbance was determined at 405 nm (A405) using a microtitre plate reader (BioTek, Vermont, USA). All serum samples were measured in triplicate.

Phage ELISA to detect FMDV antibodies

FMDV antibodies in bovine sera that reacted with the coated chimeric T7 phages were determined by indirect ELISA. Chimeric T7 (T7/FMD8-VP1145–152, T7/FMD12-VP1159–170, T7-FOVP1131–170) and wild-type T7 phages in sodium bicarbonate buffer (pH 9.6) (20 μg/ml; 100 μl) were coated on the wells. The performance of the assay was tested with 66 (50 vaccinated; 16 naïve) diluted bovine sera (1:25 dilution; 100 μl). The sensitivity of the assays was determined as described in Eshaghi et al. (2005). The PrioCHECK® FMDV type O solid-phase competitive ELISA test kit (Prionics, Schlieren-Zurich, Switzerland) was used as a gold standard as recommended by the manufacturer’s protocol. The absorbance was then measured with a microtitre plate reader. The cut-off point was calculated as the mean value of the negative controls (naïve serum samples) plus 3 times its standard deviation (Mackay et al. 2001; Wattanaphansak et al. 2008; Colling et al. 2014).

Results

Homology modelling of FMDV VP1 antigenic loops

Homology modelling was used to construct a three-dimensional (3D) model of the VP1 protein of the FMDV strain O1 Kaufbeuren. From the BLASTP search (RID: FSCX5EJG015) against the PDB database, the FMDV capsid protein with PDB ID: 1FOD (Logan et al. 1993) was revealed as the closest homologue, sharing 97.5% identity and 100% query coverage. Thus, it was selected as a template structure for homology modelling using the SWISS-MODEL programme (Biasani et al. 2014). The 40-residue polypeptide located at positions 131–170 of the VP1 protein was superimposed on the generated model using the PyMOL visualisation software (PyMOL v1.8.6.0) (Fig. 1a). Both structures were found to be strikingly similar in topology and were aligned well with each other, with an RMSD value of 0.11 Å. There is only one replacement of the amino acid residue at position 137 (Ser to Asn) in the VP1 protein. The overall low RMSD value reflects the high structural conservation between the constructed models. From the constructed 3D model of the VP1 protein, the positions of the two highly conserved epitopes, VP1145–152 and VP1159–170, were determined. The former, with the amino acid sequence RGDLQVLA, was found to adopt part of the α-helical path in the structure of the FMDV loop, while the latter, bearing the sequence LPTSFNYGAIKA, was observed to be positioned at the C-terminal region of the protein model (Fig. 1b).

Fig. 1
figure 1

Homology model of the FMDV VP1 structural protein generated by the SWISS-MODEL and viewed using the PyMOL Molecular Graphics System. a Stereo depiction of the capsid protein of the FMDV strain O1 Kaufbeuren (Accession no.: X00871). Superimposition of the VP1 G-H loop (residues 131–170) of the FMDV strain O1 Kaufbeuren (Accession no: X00871; magenta) with the capsid protein of FMDV (PDB ID: 1fod.1; blue). The whole capsid protein motif is shown in green. b Stereoplot of a stick-cartoon representation of the VP1 of FMDV O1 K (Forss et al. 1984). The whole peptide is shown in green. The VP1 residues 145–152 (RGDLQVLA) are highlighted in blue, and residues 159–170 (LPTSFNYGAIKA) are highlighted in magenta

Full size image

Amino acid sequence analysis of the VP1 of different FMDV serotypes

The highly conserved regions within the FMDV VP1 protein indicate that these regions are maintained in evolution and have functional values. Therefore, the identical amino acids in VP1131–170 were identified by comparing seven different serotypes (Table 2). The alignment analysis showed that residues 145–152 (RGDLQVLA) have an average identity of 85.7%, with a highly conserved RGD motif (100%) at positions 145–147. The amino acids 159–170 (LPTSFNYGAIKA) shared 72.6% identity, with 82.1% identity at positions 159–166 (LPTSFNYG). In addition, it was observed that the SAT serotypes exhibited great variabilities in the amino acid sequences, with the presence of up to four extra amino acid residues between positions 137 and 138, as well as 158 and 159 (Table 2).

Table 2 Percent identity of amino acid sequences of VP1 protein of FMDV serotypes O, A, Asia, C, SAT-1, SAT-2, and SAT-3
Full size table

Generation of chimeric T7 phages (T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170)

Chimeric T7 phages were generated by inserting the coding regions of the VP1 proteins (residues 145–152 and 159–170) at the 3′ end of the 10B capsid gene in the T7 Select 415-1b vector. A high yield of chimeric T7 phages (~ 1.7–2.7 × 108 pfu/ml) was observed after the successful packaging of the recombinant DNA with the T7 extracts. The genomic DNA of recombinant clones was extracted, and the presence of the DNA insert was verified by PCR. The positive transformants encoding the VP1145–152 and VP1159–170 peptides produced PCR products of approximately 458 and 470 bp, respectively (Fig. 2a). The nucleotide sequences of the inserts were confirmed by DNA sequencing (Fig. S1). As high as ~ 1011–1012 pfu/ml phage progenies were obtained following the infection of E. coli BL21 culture at an MOI of 1 after 1.5 to 2 h of incubation. The T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170 chimeric phages analysed by SDS-polyacrylamide gel electrophoresis (Fig. 2b) and western blot analysis using the T7-tag monoclonal antibody revealed a major band of approximately 38 and 42 kDa, respectively (Fig. 2c). This shows that VP1145–152 and VP1159–170 were successfully fused at the C-terminal end of the T7 10B capsid protein.

Fig. 2
figure 2

Characterisation of recombinant T7 phages. a PCR amplification of the T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170 phages. Lane M is the DNA markers in base pairs. b SDS-PAGE of the recombinant T7 and wild-type T7 phages. The phages (as labelled on top of the gel) were electrophoresed on 12% (w/v) SDS-polyacrylamide gel, and were stained with CBB R-250. c Western blot analysis of the recombinant T7 and wild-type T7 phages. The protein bands on the polyacrylamide gel were transferred to a nitrocellulose membrane and detected with the T7-tag monoclonal antibody. Lane M: molecular weight markers in kilodaltons

Full size image

Phage ELISA

The chimeric phages displaying the FMDV antigenic loops T7/FMD8-VP1145–152 and T7/FMD12-VP1159–170 were employed as a diagnostic reagent to detect the anti-FMDV antibodies in bovine sera via phage-ELISA. Figure 3 shows the results of the minimal amount of chimeric phages required to be used as coating antigens for the sufficient determination of the anti-FMDV antibody response. In this small-scale assay, the T7/FMD12-VP1159–170 phage was found to react significantly with the vaccinated bovine sera compared with the chimeric T7/FMD8-VP1145–152, T7-FOVP1131–170, and wild-type T7 phages. The 12-residue epitope displayed on the capsid of the T7/FMD12-VP1159–170 phage could provide good discrimination of antibodies detected in different pools of bovine sera when 25 ng of the phage was used. The amount of T7/FMD12-VP1159–170 phage required in this assay was found to be approximately twofold lower than that in the 40-residue immunodominant region displayed on the T7-FOVP1131–170 phage. However, the eight-residue epitope VP1145–152 appears to have the lowest antigenicity compared with both the VP1159–170 and VP1131–170 epitopes displayed on phage T7. Meanwhile, the naïve bovine serum constantly showed low absorbance readings for all the chimeric and wild-type T7 phages.

Fig. 3
figure 3

Antigenicity of the FMDV epitopes displayed on T7 phages. To determine the minimal amount of antigen required for the detection of FMDV antibodies, microtitre plate wells were coated with either recombinant T7 phages (T7/FMD8-VP1145–152, T7/FMD12-VP1159–170, and T7-FOVP1131–170), wild-type T7 phage or skimmed milk (negative control) at a range of 10–5000 ng. The wells were blocked, washed and probed with either vaccinated bovine serum or naïve bovine serum. Assays were performed in triplicate, and error bars represent the standard deviation from the triplicate measurements

Full size image

A panel of 66 serum samples from vaccinated and naïve cattle was tested at the same time using the PrioCHECK® FMDV type O test kit. The absorbance readings obtained from phage ELISA were subtracted with the background absorbance from unspecific binding towards the wild-type T7, and the results are summarised in Fig. 4a for T7/FMD12-VP1159–170 and Fig. 4b for T7-FOVP1131–170. From the results, both T7/FMD12-VP1159–170 and T7-FOVP1131–170 could discriminate between the vaccinated and naïve bovine serum samples, up to a sensitivity of 100 and 92%, respectively, compared with the gold standard assay (Table 3). The VP1159–170 epitope appears to be highly antigenic by producing higher absorbance values than the VP1131–170 epitope. By contrast, the T7/FMD8-VP1145–152 phage did not provide sufficient discrimination towards the serum samples, where the absorbance readings are equivalent to those of the wild-type T7 phage, which served as a negative control (Fig. S2).

Fig. 4
figure 4

Phage ELISA to detect anti-FMDV antibodies in bovine sera with a T7/FMD12-VP1159–170 and b T7-FOVP1131–170 phages. The reactivity of all the sera from the FMDV-vaccinated and FMDV-naïve groups was assessed using phage ELISA with 2 μg of recombinant T7 phage (T7/FMD12-VP1159–170, T7-FOVP1131–170), wild-type T7 phage or skimmed milk (negative control) coated onto the well. Assays were performed in triplicate, and the error bars represent the standard deviation from triplicate measurements. The line across the figure represents the cut-off point (0.1561 for T7-FOVP1131–170 and 0.1907 for T7/FMD12-VP1159–170). The cut-off point was calculated as the mean value of the negative controls (naïve serum samples) plus three times its standard deviation (Colling et al. 2014; Mackay et al. 2001; Wattanaphansak et al. 2008). V1–V50 indicate the vaccinated bovine sera, and N1–N16 indicate the naïve bovine sera

Full size image
Table 3 Comparative analysis of ELISA performance based on recombinant T7 phages and inactivated FMDV-type specific antigen (PrioCHECK® FMDV type O)
Full size table

Discussion

FMD exists in seven distinct serotypes (O, A, Asia 1, C, SAT1, SAT2, and SAT3) with multiple variants within each serotype (Klein 2009; Jamal and Belsham 2013). The extensive antigenic variation of this virus is due to high spontaneous mutations that have hampered the effectiveness of the FMD control programme. The variations are unequally distributed among the four structural proteins, particularly the VP1 protein (Ptaff et al. 1988). We have previously reported two highly conserved VP1 sequences at positions 145–152 and 159–170 of the seven serotypes that could constitute the antigenic loops recognised by the FMDV antibodies. In the present study, we aligned the conserved amino acid residues of the FMDV isolates within each serotype and further delineated that the VP1145–152 and VP1159–166 regions are highly conserved among different serotypes, with a sequence identity more than 80%.

Determination of the three-dimensional structure of viral proteins and prediction of potential antibody-binding domains play a vital role in defining the molecular basis of the antigenic variation of viruses. In the present study, we utilised the SWISS-MODEL server to generate a homologue model from the peptide sequence of VP1131–170, strain O1 Kaufbeuren, to identify possible antigenic epitopes. VP1145–152 and VP1159–170 of FMDV were then located from the model generated and were viewed using the PyMOL software. The model indicates that VP1145–152 containing the arginine-glycine-aspartic acid (RGD) tripeptide motif, a highly antigenic region, is located at the α-helical loop exposed on the surface of the viral capsid. This highly conserved triplet at positions 145–147 has been shown to be involved in cell receptor binding and virus infectivity through interaction with integrin-binding proteins (Baxt and Becker 1990; Lazarus and McDowell 1993; Krezel et al. 1994).

VP1145–152 is a part of the highly disordered loop located between the β-strands G and H of the VP1 capsid protein known as the G-H loop, which, in turn, forms part of the major antigenic site (site 1) involved in FMDV neutralisation (Kitson et al. 1990; Crowther et al. 1993). From the findings of Acharya et al. (1990), the electron density map revealed that the disordered sequence between amino acid residues 134 and 157 might occur as a prominent loop. All these findings provide evidence that VP1145–152 is a possible antigenic loop that reacts with anti-FMDV antibodies in cattle. To our surprise, the VP1145–152 epitope did not show any antigenicity towards the bovine sera. This could be due to the fact that the VP1145–152 epitope displayed on phage T7 was too short or lacked some flanking amino acids required to form a native conformational epitope.

In contrast to the highly disordered site of the short α-helical coils, the antigenic region of VP1159–170 was located at the C-terminal region of the depicted homologue model. This region is highly flexible as shown in Fig. 1b. The 12-residue epitope, VP1159–170, appears to be a promising candidate as a diagnostic reagent because it is highly antigenic when probed with bovine sera. Both the predicted antigenic loops (VP1145–152 and VP1159–170) in this study were highly exposed on the surface of the virus particle and, thus, were expected to detect antibodies directed towards these epitopes. The calculated molecular masses for VP1145–152 (T7/FMD8-VP1145–152) and VP1159–170 (T7/FMD12-VP1159–170) fused with T7 10B protein are 38 and 39 kDa, respectively. However, the size of VP1159–170 appeared to be ~ 3 kDa larger when analysed with SDS-PAGE (Fig. 2b). This “gel shifting” is common in SDS-PAGE due to different interactions between proteins and SDS (Shi et al. 2012), and that even one single amino acid difference in a protein can result in a big shift (kDa difference) in its migration in the gel (de Jong et al. 1978; Rae and Elliott 1986; Panayotatos et al. 1993; Prudencio et al. 2009). In addition, possible phosphorylation of serine, threonine and tyrosine residues in VP1159–170 epitope (LPTSFNYGAIKA) could attribute to the slow migration of the polypeptide in SDS-PAGE (Smith and Hightower 1981; Kho et al. 2002). However, the VP1145–152 epitope (RGDLQVLA) does not contain S, T or Y residues which could be phosphorylated.

To evaluate the antigenicity of the predicted epitopes, ELISA based on the T7-phage displayed peptides was established. The application of this T7 phage display system as a diagnostic reagent has several advantages. The T7 phage can display 415 copies of a peptide, up to 50 amino acids, on a single particle without any secretion problems (Rosenberg et al. 1996). This is particularly crucial in maximising the antigenicity and immunogenicity properties of an epitope. Viral epitopes should be presented in multiple copies, closely resembling their native structures to mimic their native viruses (Netter et al. 2001). The small icosahedral structure of the T7 phage particle (~ 60 nm in diameter) can facilitate antibody access to the displayed peptides due to its 360° display. A high yield of chimeric phages (~ 1011–1012 pfu/ml) can be obtained from a simple production in bacterial culture and a rapid purification procedure due to their particulate form.

In this study, the antigenicity of the purified chimeric T7 phages was explored. It is evident that the VP1159–170 and VP1131–170 antigenic loops displayed on the T7 phage could detect anti-FMDV antibodies in bovine sera. The T7/FMD12-VP1159–170 phage could distinguish between the vaccinated and naïve bovine sera with the minimum amount of antigen, as low as 25 ng of the purified phage. The T7/FMD12-VP1159–170 phage demonstrated a significantly higher detection limit of anti-FMDV antibodies with a higher absorbance value (A405 = 0.91) in wells coated with 100 ng of antigen, approximately twofold higher than that observed for T7-FOVP1131–170 phage. This antigenic variation between the displayed FMDV antigens could be due to the steric hindrance of the highly antigenic region (VP1159–170) by a less antigenic region (possibly VP1131–158) in the T7-FOVP1131–170 phage that lowered the overall signal. This is supported by the low antigenicity of the VP1145–152 epitope, a fragment within the VP1131–158 that was displayed on the capsid of phage T7. In addition, it is important to identify a shorter peptide as a candidate antigen with higher antigenicity. Thus, the 12-residue epitope of the VP1 protein (VP1159–170) displayed on the T7 phage serves as a better diagnostic reagent in ELISA.

Sixty-six bovine serum samples (FMDV vaccinated and naïve) were used to evaluate the performance of the phage ELISA. Both the T7/FMD12-VP1159–170 and T7-FOVP1131–170 phage ELISAs showed high sensitivities for FMDV vaccinated and naïve serum samples. The sensitivity of the assay has increased from 92% (T7-FOVP1131–170) to 100% (T7/FMD12-VP1159–170) based on the ELISA performance of the two different recombinant T7 phages, as compared with the PrioCHECK® FMDV type O solid-phase ELISA test kit. This will greatly reduce the chance of false-negative diagnosis, whereby a vaccinated animal is mistakenly diagnosed as naïve, which eventually leads to the importation of vaccinated animals into FMD-free countries where the vaccination is not practiced; hence, this threatens the recognition of the FMD-free status of member countries. This demonstrates the potential of the T7/FMD12-VP1159–170 phage as a bio-diagnostic marker for the screening of FMDV antibodies in bovine sera. This study has successfully delineated a smaller 12-residue epitope located at the C-terminal end of the FMDV VP1 protein.