Structural Features of a Conformation-dependent Antigen Epitope on ORFV-B2L Recognized by the 2E4 mAb

Previously, we successfully prepared a monoclonal antibody (mAb) named 2E4, that directly recognizes the major envelope protein B2L of the orf virus (ORFV), but there is little information about its epitope. Here, we meticulously mapped the 2E4 epitope through combinatorial programs and identified the functional binding domain and a key amino acid residue. Briefly, the simulated epitope peptide closely resembles 84VDVQSKDKDADELR97 located at the N-terminus of B2L, strongly suggesting that the epitope is conformationally or spatially structure-dependent. Subsequently, we combined these findings with the results from the antigenicity prediction of B2L to design three truncated fragments of B2L (F1, F2 and F3) selected using 2E4, and only the F1 fragment was found to be eligible for the advanced stage. Alanine-scanning mutagenesis suggested that the D94 residue is structurally crucial for the 2E4 epitope. The other participating residues, including K61, E62, and D92, together with D94 were responsible for enabling 2E4 binding and served as factors that synergistically enabled binding to the whole 2E4 epitope. In this paper, we describe, for the first time, the architecture of an ORFV conformational epitope, and it is also expected that mAb 2E4 and its epitope can be used for applications relating to orf control.

The method used to map antigenic determinants is important for investigating the mechanism by which antigens and antibodies externally assemble in a biological process. Significantly, epitope mapping is helpful in vaccine design. A B-cell epitope or paratope often acts as basic data and is defined by its complementary and adaptive potential as well as its activity 16 . In 1993, a series of synthetic peptides derived from proteins encoded by open reading frames 2 and 3 (ORF2 and ORF3) of the hepatitis E virus were used in an enzyme immunoassay to determine the localization of an epitope 17 . Soon after that, phage-displayed random peptide libraries were shown to be a promising technique for the investigation of protein-protein interaction. By screening peptide mimics from a random biological protein molecule library, researchers have broken through many barriers in the methodology of epitope mapping. This process facilitates the discovery of antibody-binding fractions or epitopes [18][19][20] . In 1994, Wang et al. 21 used a recombinant phage expression library to screen for an mAb-defined linear epitope. Until 2011, a conformation-dependent epitope was precisely mapped by a potent neutralizing mAb 3E11, which could neutralize the foot-and-mouth disease virus (FMDV) Asia1/YS/CHA/05 in vitro 22 . Other applications of the immunodominant epitopes have been reported in serodiagnosis and vaccine development 23 .
In our previous study, we prepared three mAbs against ORFV B2L, and only the mAb 2E4 was found to exhibit certain neutralization activity to ORFV infection in vitro. In addition, 2E4 was able to specifically recognize ORFV in an indirect enzyme-linked immunosorbent assay (ELISA) and an indirect immunofluorescent assay (IFA) but did not recognize epitopes according to western blotting (in which it did not recognize denatured protein). Given that there may be further applications of 2E4, this study aimed to better characterize the epitope of interest to which the mAb for 2E4 (mAb 2E4) binds and employed a set of integrated approaches to address its unclear topographic profile. Detailed information related to the 2E4 epitope will assist us by providing with a better ability to combat this viral disease in the future.

Results
Mock epitope recognized by mAb 2E4 via biopanning. Biopanning of a phage display 12-mer random peptide library was initially performed to determine the mimotope recognized by mAb 2E4 by using the affinity-purified mAb. After four rounds using the biopanning program, individual phage clones were isolated, and the reactivity of these phage clones was assessed 24 by phage ELISA (Fig. 1a). Ten positive clones were submitted to Sangon Biotech (Shanghai) for sequencing, and final sequences of seven clusters of the phage inserts were however, the BSA controls displayed lower specificity to mAb 2E4. (b) Sequence comparisons of positive phage clones. Conserved amino acid motifs are colored in various kinds of square blocks, the consensus motif VKVNPPQYDLE/RR is shown at the bottom using light red letters, and E/R is indeterminate. Frequency represents the sequence specificity that occasionally repeats within all of the 10 probable samples.
collected. Among these, P3, P6 and P10 were repeated samples. Sequence alignment of random peptides resulted in a derived sequence motif: VKVNPPQYDLE/RR (Fig. 1b), where this motif was highly inclined to represent a veritable 2E4 epitope.
Antigenicity prediction and homology modelling suggested that the 2E4 epitope is located on the B2L-F1 segment. In our previous study, the full length of B2L was artificially partitioned into three truncated peptide fragments: F1 (at aa58-112), F2 (at aa205-235) and F3 (aa at 338-375), which were used for determining the location of the 2E4 epitope. This strategy was based on the results from the antigenicity prediction of B2L obtained using the DNAStar Protean program (data not shown here). When the amino acid sequence of F1 (at aa58-112) was submitted to the ABCpred Prediction Server, the program gave us some predicted candidate epitope information with various score values. Among them, we found that the first 58 SSTKEGVDVKDKLCTL 73 and the second 88 SKDKDADELRAAGINY 103 represented the residues located near the N-terminal domain of the F1 segment (Fig. 2a, No.1-2). Subsequently, these two oligopeptides were displayed on the forepart of the 2D structure using the SWISS MODEL server (Fig. 2b1-2). 58 SSTKEGVDVKDKLCTL 73 and 88 SKDKDADELRAAGINY 103 happen to be adjacent to each other topographically, and have some similarity to the VKVNPPQYDLE/RR mimotope. Therefore, we assumed that the authentic epitope recognized by mAb 2E4 was in the B2L-F1 segment.
B2L-F1 is responsible for 2E4 binding at a conformation-dependent epitope. E.coli BL21 (DE3) cells containing F1, F2 or F3 recombinant plasmids were induced, and expressed recombinant proteins were purified from a biomass of 0.3 L culture using Ni-NTA agarose (Qiagen, Germany). Approximately equal amounts of E.coli cell lysates of expressed GST-fusion proteins were prepared to be electrophoretically separated using a 12% polyacrylamide gel for SDS-PAGE. Nevertheless, no band was found on the nitrocellulose membrane by western blot analysis (Fig. 3e). Fortunately, in ELISA, B2L (Fig. 3a) and F1 (Fig. 3b-d) were recognized by 2E4, Figure 2. B-cell epitope prediction and homology modelling. (a) B2L-F1 is composed of 55 amino acid residues ranging from S58 to E112, and the sequence was submitted to the ABCpred Prediction Server for predicting probable B-cell epitopes of mAb 2E4.The server provided 5 referenced options, and their score values are ranked from 0.85 (the highest) to 0.62 (the lowest) via strict selection using an artificial neural network. Among them, No.2 SKDKDADELRAAGINY closely resembles our expectations. The red letters represent a highly putative epitope that includes the candidate residues. (b) Homology modelling and 2D structure of B2L via the SWISS MODEL server. F1 contains 2α-helical structures in this model. (b-1) includes S58~E112, which is displayed ahead of B2L, marked in yellow, and (b-2) includes the peptide No. 2 derived from (a) at the most prominent location in the F1 peptide fragment marked in pink, which may be responsible for binding to 2E4.
whereas both F2 and F3 showed lower activity to 2E4 (Fig. 3f). These results indicated that the 2E4 epitope was located in the F1 region, with a novel discontinuous conformational antigenic determinant, corresponding to our expectations.
The critical residue of mAb 2E4 epitope was D 94 . We were particularly interested in determining whether amino acid changes in the 2E4 epitope could hinder antibody binding. A panel of GST-fusion proteins expressing variants derived from the F1 (aa at 58-112) was generated according to the results of homology modelling of B2L by SWISS-MODEL (Fig. 3a) to determine if a particular residue was critical for mAb 2E4 to recognize the epitope, and each individual residue of interest, including K 61 , K 69 , K 75 , D 94 , L 95 or E 96 , was substituted with an alanine (Fig. 4a1-3; Table 1). PAGE and anti-GST detection suggested all target proteins were successfully expressed (Fig. 4b,c). Among these mutants, GST-D 94 was found by ELISA to show obviously decreased reactivity to mAb 2E4 (Fig. 4d), suggesting that D 94 is important for 2E4 binding. The reaction activity of the D 94 mutant decreased by approximately 70%, while the other mutants showed no impact on binding action compared to the D 94 mutant, where all of the other mutants corresponded to the F1 original. The GST-fused peptides F1, F2 and F3 were separated by 12% SDS-PAGE. Western blot analysis with 2E4 corresponded with the fusion proteins shown by PAGE, but none of them showed an obvious band on the nitrocellulose membrane. (f) F1 was found to specifically recognize mAb 2E4 by ELISA. OD450 nm values less than 0.5 were considered non-specific binding to 2E4 except the full-length B2L antigen and F1 in this ELISA assay. (**p < 0.01, represents a statistically highly significant difference; Mean ± SD, n = 3). Approximately equal amounts of E. coli cell lysates with the expressed GST-fusion proteins were electrophoresed on a 12% SDS-PAGE gel and stained with Coomassie blue. 6p-1 represents the pGEX-6p-1 vector; K 61 , K 69 , K 75 , D 94 , E 95 , and L 96 represent individual mutants in which the indicated amino acids were substituted with alanines; "All" represents the mutant in which K 61 , E 62 , D 92 and D 94 were simultaneously substituted. (c) Expression products are exhibited at approximately 28 kD. The purified expression products were in accord with the situation in (b) and used to investigate the reactivity of the variants if the interesting residues were deficient. (d) Individual substitutive mutants with reactivity to mAb 2E4 were evaluated by ELISA. The pGEX-6p-1 vector was used as a negative control. Only D 94 showed remarkably attenuated reactivity to 2E4, contrasting with F1 and other mutant objects. (e) "ALL" fully abolished the reactivity of the epitope to 2E4, in contrast with the D 94 mutant. F1 is a positive control in this test (*p < 0.05, represents a statistically significant difference; **p < 0.01, represents a statistically highly significant difference; Mean ± SD, n = 3).

Mutants
The Synthesized oligonucleotides www.nature.com/scientificreports www.nature.com/scientificreports/ K 61 , e 62 and D 92 are natural auxiliary components of the whole 2E4 epitope. Additionally, based on the B2L crystal structure (homology modelling) and amino acid position of D 94 , we selected K 61 , E 62 and D 92 as doubtful contributors to the 2E4 epitope. These three residues, together with D 94 , were simultaneously substituted with alanines, and the mutant K 61 /A-E 62 /A-D 92 /A-D 94 A of F1 was named "ALL" (Fig. 4b). The result showed "ALL" most lacked reactivity compared with the D 94 mutant alone (Fig. 4d), indicating that the K 61 , E 62 and the D 92 residues were natural auxiliary components for mAb 2E4 binding. Summarily, these candidates were evaluated for their ability to affect 2E4 binding directly by alanine substitution. In the N-terminus region of F1, at least 10 exposed amino acids of the F1 surface adjacent to the 2E4 epitope were generated to evaluate their contribution to the 2E4 reactivity of the mutant F1. In contrast, the expressed GST fusion proteins of the mutant F1 peptides, including K 61 , K 69 , K 75 , E 95 and L 96 , had less impact on binding activity to 2E4 (Fig. 4e). These results indicated that the residue D 94 was a critical residue of the 2E4 epitope, and taken together, the 2E4 epitope was mapped into an exquisite region containing several original functional residues (Fig. 5a-c).

Discussion
Epitope information is essential for understanding antigen-antibody interactions 25 . In the present study, we have clarified an accurate conformational epitope for mAb 2E4 binding via several rounds of biopanning 26 . Meanwhile, an epitope MODEL was formulated by online servers and by using a visualization toolkit and was used to help identify the probability distribution of the 2E4 epitope.
Briefly, the biopanning of a random phage display peptide library led to a novel mimotope VKVNPPQYDLE/ RR motif (Fig. 1) that was recognized by the mAb 2E4 in phage ELISA. Subsequently, a series of experiments of epitope mapping were designed and achieved. Firstly, segmental expression of B2L was completed according to a B-cell epitope prediction generated beforehand, and every part was singly selected by 2E4. The reactivity of three truncate peptides (the F1, F2 and F3 fragments) were evaluated by western blot and ELISA, followed by alanine scanning mutagenesis combined with online prediction and homology modelling. Finally, mAb 2E4 was www.nature.com/scientificreports www.nature.com/scientificreports/ experimentally demonstrated to recognize a discontinuous conformational antigenic determinant (binding site) at the N-terminus of the B2L protein of ORFV (Fig. 2). Moreover, D 94 was identified as a critical residue of this epitope, and could be interpreted as a major factor influencing immunity recognition. Additionally, the K 61 , E 62 , and D 92 residues within the F1 fragment are required for constructing the epitope of the mAb 2E4, where these residues with D 94 could likely form the binding region of 2E4 in B2L (Fig. 5b,c).
Available strategies, such as homologue scanning or hybrid gene expression, can be applied to determine the general conformational epitope 21,27 . In the current study, binding assays using an individual phage on a purified target confirmed that the phage interacted with 2E4 specifically in a pattern consistent with its epitope. Therefore, random libraries might be advantageous when recombinant phages displaying infermation of epitopes are used for diagnosis or vaccination purposes 21 , as in our application. Using this approach for the mapping of epitope of a mAb, Wang and his assistants mapped an epitope on the bluetongue virus outer capsid protein VP5 in 1994-1995. In this paper, when the mimic epitope consensus motif VKVNPPQYDLE/RR was obtained, it was a fantastic and exciting challenge. Since very low sequence identity was found between the motif and the B2L gene, the 2E4 epitope was inferred to be a non-continued conformational pattern in the B2L protein. To identify the location of this conformational antigenic site, we expressed a set of truncated genes of B2L including F1, F2 and F3, and the interactivity of these products with 2E4 were assessed under the same reaction conditions. As expected, F1 displayed a unique capability for recognizing 2E4 through ELISA. None of these were detected on a nitrocellulose membrane by western blot, which corresponded to our expectations.
Epitope mapping is high consuming in terms of time and cost, making in silico methods a good complementary accelerant 28 . In silico studies and simulations can accelerate these research processes due to their time-saving manner 29 . High-accuracy predictions should result in a faster handling of the project being studied 30 . In silico approaches that showed the 3D-structure of B2L vastly facilitated our study of the 2E4 epitope. The 3D model of F1 showed 55 amino acid residues spatially in the front of B2L that contained 2α-helixes (Fig. 2b1), involved in five potential candidate peptides in the predicted epitope (Fig. 2a). Interestingly, the No.1:SSTKEGVDV-KDKLCTL (Score:0.85) and No.2:SKDKDADELRAAGINY (Score:0.83) that were generated via the ABCpred Prediction Server prediction appeared very close to 84 VDVQSKDKDADELR 97 of B2L. Perhaps, we merely need an eventual demonstration of these showing binding in a subsequent 2E4 binding test.
Fortunately, among the three truncated peptides, F1 was observed to be specifically recognized by 2E4 by ELISA rather than by western blot (Figs 3e,f and s2). Experimental results have demonstrated that the mAb 2E4-recognized epitope is conformation-dependent, which may explain the low identity between the consensus motif VKVNPPQYDLE/RR and 84 VDVQSKDKDADELR 97 in the B2L protein (Fig. s1). Here, our 3D model highlighted that F1 is in the front of the B2L model, with all geographical possible residues of the epitope region (Fig. 5a-c). Around the tripeptide 94 DEL 96 , we have found that D 92 has structural significance, as well as that K 61 , K 69 and K 75 are related to its topological structure shown by alanine scanning mutagenesis (Fig. s3). Residue substitutions of K 61 , K 69 , K 75 , D 92 , E 95 and L 96 , had not directly affected the F1 binding activity to mAb 2E4 (Fig. 4d). Remarkably, the D 94 residue mutant showed a distinctive detection value that impaired the binding activity of F1, and therefore was determined as a crucial residue for the ability of the epitope to bind to mAb 2E4. Furthermore, we wondered if K 61 , E 62 , D 92 and D 94 became an intact structure (namely "ALL") responsible for 2E4 binding due to their very adjacent topological distribution (Fig. 4e). The ALL mutant nearly abolished the binding activity of F1 to mAb 2E4, which further showed this to be the most important element in the recognition of the epitope by mAb 2E4 (Figs 4e and 5c). Meanwhile, amino acid substitution confused the original structure of B2L, and the new model (on the right) is slightly smaller than the original structure (Fig. 6a,b). Additionally, the hydrogen bond number between atoms significantly increased in the new B2L model (Fig. 7a,b). However, the essential structure of B2L persisted, though many additional loops emerged compared with the original component ( Fig. 8a-c). Mutagenesis of "ALL" resulted in extensive locus shifting in the B2L mutant, which was particularly serious in the C-terminus, and a structure-shifting trend was found (Fig. 8d).
From these results, we began to observe that the accuracy of the resulting 2E4 epitope constructs was verified by visualization servers (Fig. 5a-c). The prediction tools used here were computationally validated for identifying 2E4 epitope, contributing to the experimentally validated antigenic determinants. Many of the experimentally identified regions correctly overlapped with the computational predictions for the 2E4 epitope on F1. These findings emphasized the utility of in silico in mapping epitopes of mAbs.

Cells, virus and the mAb 2E4. MDBK cells were maintained in Dulbecco's modified Eagles, medium
(DMEM) supplemented with 10% heat-inactivated foetal bovine serum (FBS) (Gibco, Grand Island, NY) at 37 °C with 5% CO 2 . Hybridomas and mAbs to the B2L protein of ORFV were stored in our laboratory. The mAb 2E4 was preliminarily purified using the caprylic acid-ammonium sulfate method, followed by purification with the NAb TM Protein G Spin Purification Kit (Pierce, USA).

Biopanning. Selection of peptides from random peptide phage display libraries (Ph.D-12 Phage Display
Peptide Library Kit, New England Biolabs) was described previously 22 . The affinity selection of the phage clones was conducted following the manufacturer's recommendations with minor modifications. Briefly, in the first round, microwell plates were coated with 10 µg/ml of mAb 2E4 in 0.1 M NaHCO3 buffer (pH 8.6) at 4 °C for 12 h. The coated wells were washed with Tris-buffered saline (50 mM Tris-Cl pH 7.5, 150 mM NaCl) containing 0.1% Tween 20 (TBST) followed by blocking with 1 mg/ml bovine serum albumin (BSA) in 0.1 M NaHCO 3 buffer. We added 2 × 10 10 phages into the coated wells and incubated them with mAb-2E4 for 1 h at room temperature. Unbound phages were removed with 10 washes of TBST. The bound phages were eluted by 0.2 M glycine-HCl containing 1 mg/mL BSA (pH 2.2) and immediately neutralized with 1 M Tris-HCl (pH 9.1). The eluted phages were amplified in E. coli (ER2738) and titred on LB/IPTG/Xgal plates. The second and third rounds of selection www.nature.com/scientificreports www.nature.com/scientificreports/ were performed similarly to the first round, except the concentration of Tween-20 was raised to 0.5% while the concentrations of mAb 2E4 were reduced to 5 µg/mL, and 1 µg/mL, respectively. For the fourth round of biopanning, 300 ng of purified mAb 2E4 was mixed with the amplified third-round eluate (5 × 10 11 phages) and incubated for 20 min at room temperature. Then, 50 mL of protein G-Agarose suspension (Gibco, NY, USA) was added, and the incubation was continued for 15 min at room temperature. The incubation mixture was centrifuged (30 s, 3,000 × g), and the supernatant was discarded while the pellet was washed 10 times with TBST. The final pellet was resuspended in 1 mL of 0.2 M glycine-HCl containing 1 mg/mL BSA (pH 2.2). The final eluted phages were plated, and individual clones were picked randomly for phage ELISA and DNA sequencing.
In vitro phage binding assay. 96-well plates were coated with 100 ng of purified mAb 2E4 (100 mL/well mAb in 0.1 M NaHCO 3 , pH 8.6) overnight at 4 °C. BSA solution was used as a negative control. The coated wells were blocked for 2 h at room temperature, and then the phages (10 10 pfu/100 mL/well) diluted in blocking solution were added. The plates were incubated for 1 h at room temperature then washed 10 times with TBST. A phage clone was detected using horseradish peroxidase (HRP)-conjugated anti-M13 mAb (GE Healthcare, USA) at a 1:5000 dilution. The reaction was quantified using the ABTS [2,20-azinobis(3-ethylbenzthiaz -olinesulfonic acid)] substrate, and the absorbance at 450 nm was measured. Monophages exhibiting high absorbance were selected for sequencing. Single-stranded DNA was prepared according to the methods 24 described previously and sequenced using the 96gIII sequencing primer.
Sequencing and computer-based analyses of structure. After four additional rounds of selection of amplified phages, DNA from well-separated plaques was sequenced, and the corresponding peptide sequence was then generated from the DNA sequences. Computer analyses of DNA and protein sequences were performed using the DNAStar Protean software program (USA). The consensus sequence motif was defined by the peptide library screening. Molecular coordinates for the B2L protein used in structural analysis and modelling were obtained from the Protein Data Bank (PDB), then visualized and analysed using Accelrys DS Visualizer 1.7.
Protokaryon expression plasmids and expression of antigenic peptides. For further epitope determination, we generated a series of antigenic fragments derived from B2L or B2L mutants by amino acid substitution to assess whether the reaction against the 2E4 mAb in a molecular context was different from that of the B2L protein. The three representative antigenic fragments F1 (aa at 58-112), F2 (aa at 205-235) and F3 (aa at 338-375) were expressed as recombinant fusions to GST, according to the antigenicity predication, B2L protein was used as www.nature.com/scientificreports www.nature.com/scientificreports/ a positive control, and a GST empty vector was used as a negative control, prepared using the Gateway ® Cloning System (Thermo Fisher Scientific) according to manufacturer's instructions. Briefly, the DNA sequences of interest were amplified using the primers containing the sequences shown in Table 1 (amino acid substitution) and Table 2 (fragments), followed by purification of the amplified products ready for gene expression according to manufacturer´s instructions (Thermo Fisher Scientific).
Homology modelling and B-cell conformational epitope prediction. The amino acid sequences of B2L (GenBank accession number KU523790) were retrieved from the GenBank database and used as targets for homology modelling using the SWISS MODEL server (https://swissmodel.expasy.org/). The latter performed the target-template sequence alignment after searching the putative X-ray template proteins in PDB to generate the 3D models for all target sequences. The best homology models were selected according to Global Model Quality Estimation (GMQE) and QMEAN statistical parameters. Further, the established models were exported to the SAVES server Version 4 and their overall stereochemical quality, including backbone torsional angles through the Ramachandran plot, was checked according to PROCHECK. Lastly, each model was refined by an energy minimization protocol, using the Smart Minimizer algorithm in Discovery Studio version 4.1. Additionally, B2L-F1 was subjected to select probable epitopes using the ABCpred Prediction Server online (http://www.imtech.res.in/ raghava/abcpred/index.html). The predicted B cell epitopes are ranked according to their score values obtained by a trained recurrent neural network. The higher peptide score means a higher likelihood of being an epitope. All peptides shown here are above the threshold value chosen.

Western blot analyses on GST fusion proteins.
For western blot analysis, approximately equal amounts of each GST fusion protein were subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12% SDS-PAGE). The gel was either stained with Coomassie blue staining solution or electrophoretically To investigate the effectiveness of mutants in regard to their ability to affect the 2E4 epitope, clustered alanine substitution was used to test their contribution to the epitope. Mutation to alanine resulted in an obvious deformation of B2L involving the conformation of the F1 fragment. The hydrogen bond number between atoms was significantly increased in the new B2L model. (b) Deformation of B2L led to denaturation of the structure responsible for the binding activity to 2E4. Note: the red letters indicate replaced residues in the 2E4 epitope, while the yellow letters indicate nonvalent residues employed in this investigation. www.nature.com/scientificreports www.nature.com/scientificreports/ transferred to a nitrocellulose membrane. After blocking the membrane with 5% nonfat milk in PBS overnight at 4 °C, we incubated the membrane with mAb 2E4 (diluted 1:1,000 in PBS) at 37 °C for 1 h, then washed it three times with PBST, and probed it with a 1:2,000 dilution of HRP-conjugated goat anti-mouse IgG (Sigma) at 37 °C for 1 h. The reactivity was visualized with the substrate 3, 3′-diaminobenzidine (DAB; Sigma). The variant slightly influenced the molecular conformation of B2L via the four alanine residues. (c) Superimposed graph derived from the UCSF server suggesting that the essential structure of B2L persisted despite many additional loops emerging compared with the original structure. (d) Primary amino acid structure comparison between natural B2L and the mutant. We found that the mutant had not changed the folding pattern of the N-terminus that was occupied by the 2E4 epitope (the mutant residue is indicated by ⋆). However, the mutation resulted in extensive shifting of the locus in the latter part, particularly the C-terminus. Note: the two red rectangle frames indicate conserved functional regions; the multiple blue rectangle frames indicate the regions of the shifting locus. The coloured arrow indicates the variation tendency in structure stability.  Table 2. Primer sequences of recombinant protein F1, F2 and F3 (restriction site is italic and underlined).