The Tip Region on VP2 Protein of Bluetongue Virus Contains Potential IL-4-Inducing Amino Acid Peptide Segments

Bluetongue is an infectious viral hemorrhagic disease of domestic and wild ruminants that has a considerable economic impact on domestic ruminants. There are currently at least 29 serotypes of bluetongue virus (BTV) in the world. Noteworthily, the pathogenesis among BTV serotypes is different, even in the same animal species. In this study, BTV2/KM/2003 and BTV12/PT/2003 were used to investigate the differential immunological effects on bovine peripheral blood mononuclear cells (PBMCs). The BTV viral load and the expression of cytokine messenger RNA (mRNA) in PBMCs were measured by fluorescence-based real-time reverse-transcription PCR (qRT-PCR). The immunofluorescence assay (IFA) was applied to detect BTV signals in monocyte-derived macrophages (MDMs). The SWISS-MODEL and IL-4pred prediction tools were used to predict the interleukin 4 (IL-4)-inducing peptides in BTV-coat protein VP2. Synthetic peptides of VP2 were used to stimulate PBMCs for IL-4-inducing capability. This study demonstrated that the cytokine profiles of BTV-induced PBMCs were significantly different between BTV2/KM/2003 and BTV12/PT/2003. BTV2 preferentially activated the T helper 2 (Th2) pathway, represented by the early induction of IL-4, and likely fed back to inhibit the innate immunity. In contrast, BTV12 preferentially activated the innate immunity, represented by the induction of tumor necrosis factor -α (TNF-α) and interleukin 1 (IL-1), with only minimal subsequent IL-4. The BTV nonstructural protein 3 antibody (anti-BTV-NS3) fluorescent signals demonstrated that monocytes in PBMCs and MDMs were the preferred targets of BTV replication. Bioinformatics analysis revealed that the capability to induce IL-4 was attributed to the tip region of the VP2 protein, wherein a higher number of predicted peptide segments on BTVs were positively correlated with the allergic reaction reported in cattle. Synthetic peptides of BTV2-VP2 induced significant IL-4 within 12–24 h post-infection (hpi) in PBMCs, whereas those of BTV12 did not, consistent with the bioinformatics prediction. Bovine PBMCs and synthetic peptides together seem to serve as a good model for pursuing the BTV-induced IL-4 activity that precedes the development of an allergic reaction, although further optimization of the protocol is warranted.


Introduction
Bluetongue (BT) is an important viral disease of domestic and wild ruminants. In the livestock industry, BT causes considerable economic loss, which affects national and international trade. It is caused by the bluetongue virus (BTV) and transmitted by Culicoides biting midges. Clinical signs of BT are commonly observed in sheep, white-tailed deer, cattle, and other species. At least 29 distinct serotypes currently exist worldwide [1]. Since 1998, several incursions of different BTV serotypes have occurred in the Mediterranean Basin. The appearance and spread of BTV8 also began in Northwestern Europe in 2006, causing severe disease not only in sheep but also in cattle populations [2]. as the cytomegalovirus immediate-early promoter and a promoter containing interferonstimulated response elements (ISRE) [16]. Noteworthily, the NS4 amino acid sequences of Taiwanese strains [10] shared 94.8% (BTV2) and 92.2% (BTV12) nucleotide similarities with the NS4 consensus sequence of the reference [16]. The effects of these Taiwanese strains on type I IFN production will be determined in the study.
The aim of this study was to explore whether the two BTV strains of BTV2/KM/2003 and BTV12/PT/2003 induce different immune responses in bovine PBMCs. The potential IL-4-inducing peptide segments on VP2 of BTV2 (allergy reported) and BTV12 (non-allergy reported) were synthesized to test for their ability to induce IL-4 in bovine PBMCs. This combination of protein immunogenicity prediction and experimental cytokine profiles may provide a new breakthrough in the study of BTV pathogenesis.

Experimental Design
To demonstrate the differential immune responses of bovine PBMCs to BTV2 and BTV12, PBMCs were infected and harvested at various hours post-infection (hpi). The expression of messenger RNA (mRNA) of IL-1β, tumor necrosis factor -α (TNFα), IFNα, and IL-4 was quantified by real-time reverse-transcription PCR (qRT-PCR). The protein levels of IL-1β, IFNα, and IL-4 in the culture supernatants were assayed by an enzyme-linked immunosorbent assay (ELISA). Bioinformatics analysis (through the IL-4pred database) revealed that the ability of BTV2 to induce significant IL-4 resides in the tip region of VP2, which has a higher number of potential IL-4-inducing peptide segments. Then synthetic IL-4-inducing peptide segments of BTV2 (and of BTV12 on the corresponding locations for comparison) were used to stimulate PBMCs. The culture supernatants of stimulated and infected PBMCs were quantified for the protein levels of IL-4. To confirm the identity of monocyte-derived macrophages (MDMs) and the infection yields, MDMs at day 7 of cultivation were infected with either BTV2 or BTV12 and harvested at 12 hpi. MDMs were labeled by using the macrophage surface marker antibody (anti-CD172a) and BTV nonstructural protein 3 antibody (anti-NS3) in an immunofluorescence assay (IFA) before being counted for the infection rates ( Figure 1).
Pathogens 2021, 10, x FOR PEER REVIEW 3 of 14 pathway downstream of retinoic acid-inducible gene I (RIG-I) and upstream of TANKbinding kinase 1/IκB kinase ε (TBK1/IKKε) activation [15]. The nonstructural protein NS4 inhibited the type I IFN by downregulating the activities of a variety of promoters, such as the cytomegalovirus immediate-early promoter and a promoter containing interferonstimulated response elements (ISRE) [16]. Noteworthily, the NS4 amino acid sequences of Taiwanese strains [10] shared 94.8% (BTV2) and 92.2% (BTV12) nucleotide similarities with the NS4 consensus sequence of the reference [16]. The effects of these Taiwanese strains on type I IFN production will be determined in the study. The aim of this study was to explore whether the two BTV strains of BTV2/KM/2003 and BTV12/PT/2003 induce different immune responses in bovine PBMCs. The potential IL-4-inducing peptide segments on VP2 of BTV2 (allergy reported) and BTV12 (non-allergy reported) were synthesized to test for their ability to induce IL-4 in bovine PBMCs. This combination of protein immunogenicity prediction and experimental cytokine profiles may provide a new breakthrough in the study of BTV pathogenesis.

Experimental Design
To demonstrate the differential immune responses of bovine PBMCs to BTV2 and BTV12, PBMCs were infected and harvested at various hours post-infection (hpi). The expression of messenger RNA (mRNA) of IL-1β, tumor necrosis factor -α (TNFα), IFNα, and IL-4 was quantified by real-time reverse-transcription PCR (qRT-PCR). The protein levels of IL-1β, IFNα, and IL-4 in the culture supernatants were assayed by an enzyme-linked immunosorbent assay (ELISA). Bioinformatics analysis (through the IL-4pred database) revealed that the ability of BTV2 to induce significant IL-4 resides in the tip region of VP2, which has a higher number of potential IL-4-inducing peptide segments. Then synthetic IL-4-inducing peptide segments of BTV2 (and of BTV12 on the corresponding locations for comparison) were used to stimulate PBMCs. The culture supernatants of stimulated and infected PBMCs were quantified for the protein levels of IL-4. To confirm the identity of monocyte-derived macrophages (MDMs) and the infection yields, MDMs at day 7 of cultivation were infected with either BTV2 or BTV12 and harvested at 12 hpi. MDMs were labeled by using the macrophage surface marker antibody (anti-CD172a) and BTV nonstructural protein 3 antibody (anti-NS3) in an immunofluorescence assay (IFA) before being counted for the infection rates ( Figure 1).

The Viral Loads and Titers of BTV2/BTV12 in Infected PBMCs
The yields of PBMCs were approximately 4.52 ± 1.7 × 10 6 cells/mL in a 25 mL suspension derived from 50 mL blood (n = 8). The standard curve of absolute viral loads in PBMCs was quantified by qRT-PCR using an external BTV VP7 cRNA standard [7]. The detection limit of 10 1 copies/µL was previously described [7]. The VP7 RNA copy numbers and titers in tissue culture infectious dose (TCID 50 ) are presented in Figure 2.

The Viral Loads and Titers of BTV2/BTV12 in Infected PBMCs
The yields of PBMCs were approximately 4.52 ± 1.7 × 10 6 cells/mL in a 25 mL suspension derived from 50 mL blood (n = 8). The standard curve of absolute viral loads in PBMCs was quantified by qRT-PCR using an external BTV VP7 cRNA standard [7]. The detection limit of 10 1 copies/μL was previously described [7]. The VP7 RNA copy numbers and titers in tissue culture infectious dose (TCID50) are presented in Figure 2. The intracellular loads of BTV2 were 10 3.35 copies/μL at 0 hpi and peaked at 10 4.04 copies/μL at 12 hpi. The supernatant progeny virus of BTV2 was 10 1.63 TCID50/mL at 0 hpi and peaked at 12 hpi with 10 2.69 TCID50/mL. The intracellular loads of BTV12 were 10 3.45 copies/μL at 0 hpi and peaked at 10 4.47 copies/μL at 12 hpi. The supernatant progeny virus of BTV12 was 10 1.63 TCID50/mL at 0 hpi and peaked at 12 hpi (10 3.22 TCID50/mL). Overall, the viral loads of both BTV2 and BTV12 were similar, with a higher load of BTV12 than BTV2 at 12 hpi, 2.69-fold higher VP7 RNA, and 3.39-fold higher TCID50 (Figure 2), consistent with the observation in Figure 3 (Section 2.3).

Infection Rate of BTV2 and BTV12 in MDMs Detected by IFA
After 7 days of cultivation, the yields of MDMs were approximately 1 ± 0.4 × 10 5 cells/mL/well in a 12-well plate. Uninfected bovine MDMs expressed membranous CD172a (red) with a more irregular cell shape (Figure 3a). The infected MDMs exhibited both membranous CD172a (red) and an intracytoplasmic viral NS3 signal (orange), with a more rounded cell shape (Figure 3b,c). The infection rate of BTV2, at multiplicity of infection (MOI) 1, was 75% (n = 80 cells counted, Figure 3b). When MDMs were infected by BTV12, at MOI 1, almost all cells died and detached. This was consistent with the finding in Figure 2. The infection rate of BTV12 was only countable at MOI 0.5, which was 76.1% (n = 46, Figure 3c). The intracellular loads of BTV2 were 10 3.35 copies/µL at 0 hpi and peaked at 10 4.04 copies/µL at 12 hpi. The supernatant progeny virus of BTV2 was 10 1.63 TCID 50 /mL at 0 hpi and peaked at 12 hpi with 10 2.69 TCID 50 /mL. The intracellular loads of BTV12 were 10 3.45 copies/µL at 0 hpi and peaked at 10 4.47 copies/µL at 12 hpi. The supernatant progeny virus of BTV12 was 10 1.63 TCID 50 /mL at 0 hpi and peaked at 12 hpi (10 3.22 TCID 50 /mL). Overall, the viral loads of both BTV2 and BTV12 were similar, with a higher load of BTV12 than BTV2 at 12 hpi, 2.69-fold higher VP7 RNA, and 3.39-fold higher TCID 50 (Figure 2), consistent with the observation in Figure 3 (Section 2.3).

Infection Rate of BTV2 and BTV12 in MDMs Detected by IFA
After 7 days of cultivation, the yields of MDMs were approximately 1 ± 0.4 × 10 5 cells/mL/well in a 12-well plate. Uninfected bovine MDMs expressed membranous CD172a (red) with a more irregular cell shape (Figure 3a). The infected MDMs exhibited both membranous CD172a (red) and an intracytoplasmic viral NS3 signal (orange), with a more rounded cell shape (Figure 3b,c). The infection rate of BTV2, at multiplicity of infection (MOI) 1, was 75% (n = 80 cells counted, Figure 3b). When MDMs were infected by BTV12, at MOI 1, almost all cells died and detached. This was consistent with the finding in Figure 2. The infection rate of BTV12 was only countable at MOI 0.5, which was 76.1% (n = 46, Figure 3c).

Expression of mRNA in BTV-Infected PBMCs
BTV2 induced marked IL-4 expression early at 6 hpi, and it was significantly higher than that of BTV12 and negative groups at 6-48 hpi (Figure 4a,b), with a low significant expression of TNF-α only being observed at 6 hpi. Both results were consistent with our previous data [7].
In contrast, the expression of IL-1β in the BTV12 group peaked at 6 hpi (56.554-fold, p < 0.01), and was significantly higher than the values of other groups, and then gradually declined up to 48 hpi ( Figure 4c). Meanwhile, IL-4 increased only slightly, with its maximum at 6 hpi (6.031-fold, Figure 4c). The level of TNF-α induced by BTV12 peaked at 6 hpi (31.947-fold, p < 0.05), and was significantly higher than the values of the BTV2 group and negative control, and then gradually declined by 48 hpi (Figure 4c).
Overall, the mRNA expression of IL-1β and TNF-α of the BTV12 group was strongly activated, whereas that of the IL-4 was minimal and significantly lower than that of the BTV2 group. IFN-α of the BTV2 group peaked at 24 hpi (3.465-fold), and that of the BTV12 group also peaked at 24 hpi (2.759-fold) (Figure 4b,c).

The Cytokine Production Induced by BTV2 and BTV12
The IL-1β in the BTV12 group showed a sustained significantly high production at 24-48 hpi and then peaked at 1162.02 pg/mL (p < 0.01) at 24 hpi ( Figure 5a). The BTV-12induced IL-1β production was significantly greater than that in BTV2 and the negative control (Figure 5a), which was consistent with the mRNA expression ( Figure 4c). On the Macrophage surface marker CD172a was labeled with the mouse anti-bovine CD172a antibody, with a secondary antibody of anti-mouse IgG1 conjugated with PerCP/Cyanine5.5 (Red). The BTV signal was labeled with the rabbit anti-BTV-NS3 antibody with a secondary antibody of goat anti-rabbit IgG conjugated with Alexfluor 546 (Orange). The cell nucleus was labeled with 4 ,6-diamidino-2-phenylindole (DAPI, Blue). Uninfected MDMs express membranous CD172a with a more irregular cell shape. The BTVinfected MDMs express cytoplasmic NS3 protein with a more rounded shape.

Expression of mRNA in BTV-Infected PBMCs
BTV2 induced marked IL-4 expression early at 6 hpi, and it was significantly higher than that of BTV12 and negative groups at 6-48 hpi (Figure 4a,b), with a low significant expression of TNF-α only being observed at 6 hpi. Both results were consistent with our previous data [7]. Macrophage surface marker CD172a was labeled with the mouse anti-bovine CD172a antibody, with a secondary antibody of anti-mouse IgG1 conjugated with PerCP/Cyanine5.5 (Red). The BTV signal was labeled with the rabbit anti-BTV-NS3 antibody with a secondary antibody of goat anti-rabbit IgG conjugated with Alexfluor 546 (Orange). The cell nucleus was labeled with 4′,6-diamidino-2-phenylindole (DAPI, Blue). Uninfected MDMs express membranous CD172a with a more irregular cell shape. The BTV-infected MDMs express cytoplasmic NS3 protein with a more rounded shape.

Expression of mRNA in BTV-Infected PBMCs
BTV2 induced marked IL-4 expression early at 6 hpi, and it was significantly higher than that of BTV12 and negative groups at 6-48 hpi (Figure 4a,b), with a low significant expression of TNF-α only being observed at 6 hpi. Both results were consistent with our previous data [7].
In contrast, the expression of IL-1β in the BTV12 group peaked at 6 hpi (56.554-fold, p < 0.01), and was significantly higher than the values of other groups, and then gradually declined up to 48 hpi ( Figure 4c). Meanwhile, IL-4 increased only slightly, with its maximum at 6 hpi (6.031-fold, Figure 4c). The level of TNF-α induced by BTV12 peaked at 6 hpi (31.947-fold, p < 0.05), and was significantly higher than the values of the BTV2 group and negative control, and then gradually declined by 48 hpi (Figure 4c).
Overall, the mRNA expression of IL-1β and TNF-α of the BTV12 group was strongly activated, whereas that of the IL-4 was minimal and significantly lower than that of the BTV2 group. IFN-α of the BTV2 group peaked at 24 hpi (3.465-fold), and that of the BTV12 group also peaked at 24 hpi (2.759-fold) (Figure 4b,c).

The Cytokine Production Induced by BTV2 and BTV12
The IL-1β in the BTV12 group showed a sustained significantly high production at 24-48 hpi and then peaked at 1162.02 pg/mL (p < 0.01) at 24 hpi (Figure 5a). The BTV-12induced IL-1β production was significantly greater than that in BTV2 and the negative control (Figure 5a), which was consistent with the mRNA expression (Figure 4c). On the In contrast, the expression of IL-1β in the BTV12 group peaked at 6 hpi (56.554fold, p < 0.01), and was significantly higher than the values of other groups, and then gradually declined up to 48 hpi ( Figure 4c). Meanwhile, IL-4 increased only slightly, with its maximum at 6 hpi (6.031-fold, Figure 4c). The level of TNF-α induced by BTV12 peaked at 6 hpi (31.947-fold, p < 0.05), and was significantly higher than the values of the BTV2 group and negative control, and then gradually declined by 48 hpi (Figure 4c).
Overall, the mRNA expression of IL-1β and TNF-α of the BTV12 group was strongly activated, whereas that of the IL-4 was minimal and significantly lower than that of the BTV2 group. IFN-α of the BTV2 group peaked at 24 hpi (3.465-fold), and that of the BTV12 group also peaked at 24 hpi (2.759-fold) (Figure 4b,c).

The Cytokine Production Induced by BTV2 and BTV12
The IL-1β in the BTV12 group showed a sustained significantly high production at 24-48 hpi and then peaked at 1162.02 pg/mL (p < 0.01) at 24 hpi (Figure 5a). The BTV-12induced IL-1β production was significantly greater than that in BTV2 and the negative control (Figure 5a), which was consistent with the mRNA expression (Figure 4c). On the other hand, the protein levels of IFN-α were consistently low (the detection limit was 15 pg/mL, Figure 5b), in line with their low mRNA expressions (Figure 4b,c).
Pathogens 2021, 10, x FOR PEER REVIEW 6 of 14 other hand, the protein levels of IFN-α were consistently low (the detection limit was 15 pg/mL, Figure 5b), in line with their low mRNA expressions (Figure 4b,c). In contrast, the IL-4 production in the BTV12 group remained below the detection limit at all time points (Figure 5c), whereas that of the BTV2 group was significantly and persistently high during 6-48 hpi and peaked at 32.60 pg/mL (p < 0.05) at 24 hpi ( Figure  5c), consistent with the mRNA expression (Figure 4b).

Prediction of Potential IL-4-Inducing Peptides on the VP2 of Various BTV Strains
IEDB analysis showed that those BTVs (serotypes 2, 8, 11, and 17) with records of an allergy, manifested by IL-4 expression, had significantly higher numbers of potential IL-4-inducing peptides (segments) on the whole length of VP2 than those BTVs (serotypes 12, 1) without allergy records (Table 1 and Table S1), averaging 24.80 ± 2.713 segments (n = 20) versus 19.80 ± 2.926 segments (n = 15) (p < 0.01), respectively. Note: a Number of potential IL-4-inducing peptide segments in the tip region of VP2. b Number of potential IL-4-inducing peptides on the whole length of VP2. c Ratio of number in the tip region divided by number on the whole length of VP2.
The raw data are provided in Table S1.
Modeling of the 3D structure of the VP2 of BTV2 and BTV12 showed their folding in the tip region to be distinct from one another (Figure 6a,b), while the rest of the VP2 mainframe structure was similar. This distinct folding in the tip region may be related to their capability to interact with the host receptor, in this case, on PBMCs. There were higher numbers of potential IL-4-inducing peptides, averaging 7.20 ± 0.748 segments (n = 20), in this VP2 tip region of allergic BTVs, compared to the 2.60 ± 0.800 segments (n = 15) (p < 0.01) recorded for BTVs without allergy records ( Table 1). That of the non-allergy strain BTV10 was graded intermediate (Table 1 and Table S1). In contrast, the IL-4 production in the BTV12 group remained below the detection limit at all time points (Figure 5c), whereas that of the BTV2 group was significantly and persistently high during 6-48 hpi and peaked at 32.60 pg/mL (p < 0.05) at 24 hpi (Figure 5c), consistent with the mRNA expression (Figure 4b).

Prediction of Potential IL-4-Inducing Peptides on the VP2 of Various BTV Strains
IEDB analysis showed that those BTVs (serotypes 2, 8, 11, and 17) with records of an allergy, manifested by IL-4 expression, had significantly higher numbers of potential IL-4-inducing peptides (segments) on the whole length of VP2 than those BTVs (serotypes 12, 1) without allergy records (Table 1 and Table S1), averaging 24.80 ± 2.713 segments (n = 20) versus 19.80 ± 2.926 segments (n = 15) (p < 0.01), respectively. Note: a Number of potential IL-4-inducing peptide segments in the tip region of VP2. b Number of potential IL-4-inducing peptides on the whole length of VP2. c Ratio of number in the tip region divided by number on the whole length of VP2. The raw data are provided in Table S1.
Modeling of the 3D structure of the VP2 of BTV2 and BTV12 showed their folding in the tip region to be distinct from one another (Figure 6a,b), while the rest of the VP2 mainframe structure was similar. This distinct folding in the tip region may be related to their capability to interact with the host receptor, in this case, on PBMCs. There were higher numbers of potential IL-4-inducing peptides, averaging 7.20 ± 0.748 segments (n = 20), in this VP2 tip region of allergic BTVs, compared to the 2.60 ± 0.800 segments (n = 15) Pathogens 2021, 10, 3 7 of 14 (p < 0.01) recorded for BTVs without allergy records ( Table 1). That of the non-allergy strain BTV10 was graded intermediate (Table 1 and Table S1).
To test whether this segment number difference accounted for the different immune responses observed above, cocktails of four segments of potential IL-4-inducing peptides (Figure 6c) were synthesized to stimulate the PBMCs.

Induction of IL-4 Production in PBMC Stimulated with Synthetic Peptide
According to the instructions from the manufacturer, standard curves of IL-4 protein with the cover range of the IL-4 concentration ranging from 8.19 to 800 pg/mL were determined in our previous experiments (the R 2 value was 0.9921) [7].  Table 1 and Table S2)

Predicted IL-4-Inducing Peptide Segments in the VP2 Tip Region of BTV2 and BTV12
The tip region of VP2 of BTV2, having significant IL-4 expression (Figures 4b and 5c), contained seven segments (numbers 1-2-3-4-5-6-7, Figure 6a,c) of the potential IL-4inducing peptide (yellow shaded, Figure 6c), whereas that of BTV12 contained only two segments (numbers 1-2, shaded green, Figure 6b,c; Table S2). The ratio of the total copy number of synthetic peptides to PBMC was set at 100 copies (=400 segments) per PBMC, which approximated the dose of 10 2 TCID 50 at 6 hpi (Figure 2), and each segment was assumed to have an equal IL-4-inducing capability.
To test whether this segment number difference accounted for the different immune responses observed above, cocktails of four segments of potential IL-4-inducing peptides (Figure 6c) were synthesized to stimulate the PBMCs.

Induction of IL-4 Production in PBMC Stimulated with Synthetic Peptide
According to the instructions from the manufacturer, standard curves of IL-4 protein with the cover range of the IL-4 concentration ranging from 8.19 to 800 pg/mL were determined in our previous experiments (the R 2 value was 0.9921) [7].
The IL-4 in the BTV2-pep group displayed a significantly high expression of 18.437 pg/mL at 12 hpi, as did that induced by BTV2 virus, which also peaked at 12 hpi (20.827 pg/mL, Figure 7). Both were significantly higher than those of the other four experimental groups (p < 0.05), consistent with the prediction (Figure 6; Table 1) and mRNA expression profile (Figure 4). There were no significant differences between the responses to BTV2-pep and BTV2 virus (p = 0.128). The IL-4 in the BTV2-pep group displayed a significantly high expression of 18.437 pg/mL at 12 hpi, as did that induced by BTV2 virus, which also peaked at 12 hpi (20.827 pg/mL, Figure 7). Both were significantly higher than those of the other four experimental groups (p < 0.05), consistent with the prediction (Figure 6; Table 1) and mRNA expression profile (Figure 4). There were no significant differences between the responses to BTV2pep and BTV2 virus (p = 0.128).
The IL-4 in the BTV12-pep group also peaked at 12 hpi (15.023 pg/mL), and IL-4 in the BTV12 virus group peaked at 72 hpi (15.023 pg/mL, Figure 7). Both were significantly lower than those of the BTV2-pep group.

Discussion
This study demonstrates the different responses of bovine PBMCs (summarized in Table 2 (Figures 4b,c and 5a,c). The infection rate of BTV12 can only be calculated at MOI 0.5, suggesting that BTV12 may have a stronger affinity or virulence to PBMC monocytes and MDMs (Figure 2), in which infected monocytes/MDMs activated the innate immunity (Figure 4c), as is normally expected for sequential immune responses [17,18]. These infected monocytes/MDMs then underwent necrosis/lysis (Figure 3c), resulting in a reduced number of intact cells available for analysis (Figure 3c and data not shown).
As in regular viral infection, BTV12 activated monocytes normally and strongly, indicated by the TNF-α and IL-1β expressions, but activated the Th2 pathway poorly, indicated by the IL-4 (Figures 4c, 5c and 8). In contrast, BTV2 infection activated innate immunity poorly but activated the Th2 pathway significantly (Figure 4b). On the other hand, BTV2 seemed able to infect monocytes (Figure 3b) while activating the Th2 pathway early on (Figure 4b), which then fed back to inhibit the innate immunity (Figures 4b and 8) [7,19]. These differences observed between BTV2/KM/2003 and BTV12/PT/2003 ( Figure 8) were also found in [4], wherein BTV1 stimulated a significant IL-1β expression, but BTV8 did not. It is speculated that the abolition of IL-1β activation might be key for BTV survival in infected animals [20]. The IL-4 in the BTV12-pep group also peaked at 12 hpi (15.023 pg/mL), and IL-4 in the BTV12 virus group peaked at 72 hpi (15.023 pg/mL, Figure 7). Both were significantly lower than those of the BTV2-pep group.

Discussion
This study demonstrates the different responses of bovine PBMCs (summarized in Table 2) induced by the two Taiwanese strains BTV2/KM/2003 and BTV12/PT/2003. The replication curve ( Figure 2) and infection yield (Figure 3) of both BTV2 and BTV12 were largely parallel in both PBMCs and MDMs, yet the cytokine profiles they induced were significantly different (Figure 4b,c and Figure 5a,c). The infection rate of BTV12 can only be calculated at MOI 0.5, suggesting that BTV12 may have a stronger affinity or virulence to PBMC monocytes and MDMs (Figure 2), in which infected monocytes/MDMs activated the innate immunity (Figure 4c), as is normally expected for sequential immune responses [17,18]. These infected monocytes/MDMs then underwent necrosis/lysis (Figure 3c), resulting in a reduced number of intact cells available for analysis (Figure 3c and data not shown).  Association with allergy ( Table 1, Supplement Tables S1 and S2) Yes Low IL-4-inducing capacity of synthetic peptides (Figure 7) Yes Low Since the two BTV Taiwanese strains have a striking similarity in all proteins, expect VP2 [10], we speculated that VP2, rather than VP5, of BTV2 may be related to the significant IL-4 induction (Figures 4b, 5c and 8) [7]. However, despite the amino acid similarity of only 25.8% on VP2 between the two Taiwanese BTVs, their 3D structures were strikingly similar in the main frame (body, hairpin, and hub) among the serotypes analyzed, and only their tip regions differed significantly ( Figure 6) [9]. This allowed us to focus on analyzing the difference between the tip regions, 191-407 a.a., of the two Taiwanese BTVs (Figure 6c).  Association with allergy ( Table 1, Supplement Tables S1 and S2) Yes Low IL-4-inducing capacity of synthetic peptides (Figure 7) Yes Low Figure 8. Different immune pathways induced by BTV2 and BTV12 in bovine PBMCs. BTV2 induces significant IL-4 expression and then feeds back to inhibit innate immunity (upper part) [7]. In contrast, BTV12 activates innate immunity strongly, without inducing significant IL-4 expression. Figure 8. Different immune pathways induced by BTV2 and BTV12 in bovine PBMCs. BTV2 induces significant IL-4 expression and then feeds back to inhibit innate immunity (upper part) [7]. In contrast, BTV12 activates innate immunity strongly, without inducing significant IL-4 expression.
As in regular viral infection, BTV12 activated monocytes normally and strongly, indicated by the TNF-α and IL-1β expressions, but activated the Th2 pathway poorly, indicated by the IL-4 (Figures 4c, 5c and 8). In contrast, BTV2 infection activated innate immunity poorly but activated the Th2 pathway significantly (Figure 4b). On the other hand, BTV2 seemed able to infect monocytes (Figure 3b) while activating the Th2 pathway early on (Figure 4b), which then fed back to inhibit the innate immunity (Figures 4b  and 8) [7,19]. These differences observed between BTV2/KM/2003 and BTV12/PT/2003 ( Figure 8) were also found in [4], wherein BTV1 stimulated a significant IL-1β expression, but BTV8 did not. It is speculated that the abolition of IL-1β activation might be key for BTV survival in infected animals [20].
The type I interferon was minimally expressed at both the mRNA and protein levels in both BTV2-and BTV12-infected PBMCs (Figure 4b,c and Figure 5b) in the present study setting. The result contradicts the finding in [14] but is consistent with others showing that BTV utilizes NS3 [15] and NS4 [16] to inhibit type I interferon expression. The NS4 amino acid sequences of Taiwanese strains [10] were similar to the NS4 consensus sequence of [16], suggesting that the NS4 of BTV Taiwanese strains may have contributed to the inhibition of IFNα mRNA transcription in the present study.
Since the two BTV Taiwanese strains have a striking similarity in all proteins, expect VP2 [10], we speculated that VP2, rather than VP5, of BTV2 may be related to the significant IL-4 induction (Figures 4b, 5c and 8) [7]. However, despite the amino acid similarity of only 25.8% on VP2 between the two Taiwanese BTVs, their 3D structures were strikingly similar in the main frame (body, hairpin, and hub) among the serotypes analyzed, and only their tip regions differed significantly ( Figure 6) [9]. This allowed us to focus on analyzing the difference between the tip regions, 191-407 a.a., of the two Taiwanese BTVs (Figure 6c).
The number of predicted IL-4-inducing peptide segments contained in the tip region of VP2 seems to be positively correlated with the capability to induce the bovine allergy reported in the literature (Table 1), as well as with the induction of IL-4 expression in the current experimental setting (Figure 4b,c and Figure 5c) [7]. BTV2, which contains seven segments in the tip region ( Figure 6, Table 1 and Table S2), is closer to that in BTV11 (with five segments) with reported bovine IgE induction [6]. In contrast, the BTV12 tip region, which only has two segments ( Figure 6), is closer to that in BTV1 (with three segments, Table S2), without an allergy record.
In general, the IL-4 expressions of PBMCs induced by synthetic peptides of VP2 are similar to those stimulated by virus particles (Figure 7). IL-4 can be induced in both synthetic peptides and virus infection groups. Because the peptides have no replication ability and are only given once, the results of the peptide stimulation test suggested the following three possibilities: 1. IL-4 stimulation may not need a full replication of the virus; 2. synthetic peptides may serve as a good model for inducing hypersensitivity quickly in PBMCs, without requiring large animals; and 3. most events were evident during the early infection. It was easier to distinguish the Th2-inducing capacity of BTV2 and BTV12 by the total IL-4 expression at both mRNA and protein levels, as well as in the peptide induction model. Together, the IL-4-inducing peptides and bovine PBMCs seem to be a good model for pursuing BTV-induced allergic reactions, although the protocol needs further optimization.

Animals, and Preparation of Bovine PBMC and Monocyte-Derived Macrophage (MDMs)
Eight 5-year-old female Holstein cows (C1 to C8) in the dry period of the lactation cycle raised in the National Taiwan University (NTU) experimental herd (non-closed environment, the experiment permission number in NTU is IACUC#A02041) were used for blood collection. The whole blood was used for PBMC isolation and BTV infection. All cows were healthy, without clinical signs or medication. All cow sera were free of anti-BTV antibody production, tested by commercial cELISA (VMRD, Pullman, WA, USA) during the 3-year experiment period. Heparinized blood was collected for PBMC isolation and quantitated as previously described [7]. Differentiation of the MDMs in vitro was carried out as previously described [7]. The morphology of MDMs was confirmed via bright field inverted microscopy. The identity of MDMs was demonstrated by macrophage surface marker CD172a (mouse anti-bovine CD172a antibody, Bio-Rad, Hercules, CA, USA) [7] and observed via confocal microscopy.

Virus Infection and Quantitation of BTV Viral Loads in PBMCs
The BTV2/KM/2003 and BTV12/PT/2003 Taiwanese isolates [21] were propagated in seven passages in baby hamster kidney (BHK-21) cells and adjusted to the titer of 10 6 TCID 50 /mL.
The MDMs and PBMCs were infected with either BTV2 or BTV12 at a multiplicity of infection (MOI) of 1 in triplicate or served as a negative control. The PBMCs were harvested at various hpi and quantitated by qRT-PCR for the mRNA expression levels [7] with primer sequences designed for housekeeping gene ribosomal protein S9 (RPS9) [22] and BTV VP7 [7]. The BTV viral loads in PBMCs at various time points were quantified by qRT-PCR for intracellular VP7 RNA expression as in previous research [7]. The BTV-VP7 cRNA standard preparation protocols were also as described in previous research [23].

TCID 50 Assay to Determine the Titer of Progeny Virus in the PBMC Supernatant
The culture supernatants of BTV2-and BTV12-infected PBMCs, harvested at various similar time points, were titrated by the TCID 50 assay to confirm the progeny virus of BTVs [24,25]. The BHK-21 cells pre-seeded in 96-well plates, at 8 × 10 4 cells/well, were inoculated with serially diluted samples, from 10 −1 to 10 −7 , with each being placed into 6 wells (hexaplicate). After 1 h incubation, cells were washed with PBS, fed with DMEM, and cultured during 72 hpi, at which point, cells were washed with PBS, fixed with 95% ethanol, and stained with 0.2% crystal violet in 2% ethanol. To confirm the titer of progeny virus, the cytopathic effects during 1 h incubation (CPE, i.e., necrosis, swelling, and rounding) were scored under 20× magnification using a dissecting microscope. The difference in logarithms of the TCID 50 assay was calculated by the following formula: [(mortality at dilution next above 50%) − 50%]/[(mortality next above 50%) − (mortality next below 50%)] [24,25].

Pair-Wise Comparison of the Cytokine mRNA Expression by qRT-PCR
The PBMCs were harvested at various hpi and quantitated by qRT-PCR for the mRNA expression levels [7] with primer sequences designed for TNF-α, IL-1β [22], IFN-α [26], and IL-4 [27]. SYBR Green qRT-PCR is a widely used method for investigating cytokine mRNA expression in ruminants [27]. The extraction of total RNA and data analysis were carried out as in our previous research [7]. The Ct values of RPS9 in qRT-PCR were stable from 16.96 to 19.71. All cytokines of the uninfected negative control group were under the detected limit. Every cytokine Ct was double normalized by housekeeping gene ribosomal protein S9 (RPS9) and its 0 hpi to monitor variations in cell health and the validity of the experimental procedures. The "relative" fold change in the cytokine mRNA level was calculated by the 2 −∆∆Ct method [28]. The first normalization was employed to calculate ∆Ct: each target cytokine mRNA Ct − RPS9 Ct = ∆Ct. The second normalization was applied to calculate ∆∆Ct: The ∆Ct of 6, 24, 48 hpi − the ∆Ct of 0 hpi. The relative fold change value was 2 −∆∆Ct .

IFA of BTV-Infected MDMs by Confocal Microscopy
All MDMs were cultivated on round coverslips within 12-well plates in the 5% CO 2 incubator at 37°C. The MDMs were infected with either BTV2 (MOI = 1) or BTV12 (MOI = 0.5) for 2 h. After 2 h incubation, the culture supernatants were replaced with fresh medium and harvested 12 hpi. To decrease the non-specific binding of antibodies with Fc receptors on MDMs, azide-free Fc receptor blocker (20 µL/10 5 cells, Innovex Biosciences Inc., Richmond, CA, USA) was used for blocking for 30 min. After blocking, all cells were fixed with 4% paraformaldehyde (PFA, Merck, Darmstadt, Germany) in PBS for 3 min.
The IFA staining protocol was as described in the previous study [7]. Macrophage surface marker CD172a was labeled with the mouse anti-bovine CD172a antibody, with a secondary antibody of anti-mouse IgG1 conjugated with PerCP/Cyanine5.5 (Rat IgG, BioLegend, San Diego, CA, USA). BTV signals were labeled by rabbit anti-BTV-NS3 antibody (GenScript, Piscataway, NJ, USA), with a secondary antibody of goat anti-rabbit IgG conjugated with Alexfluor 546 (Thermo, Waltham, MA, USA). After staining, DAPI-Fluoromount-G Mounting Medium (Invitrogen, Carlsbad, CA, USA) was used to mount the slide. Cells were microphotographed with a Zeiss LSM 780 confocal microscope by using ZEN 2009 Light Edition software (Carl Zeiss, Oberkochen, Germany). 4.6. Enzyme-Linked Immunosorbent Assay (ELISA) for IL-1β, IFN-α, and IL-4 The supernatants of PBMCs infected with BTVs were harvested at each time point to monitor cytokine production. The commercial bovine IL-1β ELISA kit was purchased from Thermo Inc. (Waltham, MA, USA) and bovine IL-4 ELISA kit was purchased from Sigma-Aldrich™ (St Louis, MO, USA). Their standard curves were established as previously described [7]. The commercial IFN-α ELISA kit was purchased from Sunlong Inc. (Hangzhou, Zhejiang, PRC) and its standard curve was established as shown in Figure S1. Potential IL-4-inducing peptides in the VP2 tip region were predicted by the IL4pred online platform for IL-4 motif scanning (https://webs.iiitd.edu.in/raghava/il4pred/). The BTV-VP2 protein sequences in FASTA format [29] were entered in the "Scanning of IL-4 Motifs" field, and the Koolman-Röhm class of motifs [30] was selected for searching and visualizing the motifs tabulated (Table 1; Tables S1 and S2). After the database server replied, repeat sequence items were removed, and both the total number of IL-4-inducing peptides in the tip region (191-407 a.a.) and the full length of VP2 were calculated manually.
Three-dimensional protein structure models of VP2 was generated by the SWISS-MODEL homology-modeling pipeline (https://swissmodel.expasy.org/). The IL-4 motif sequence predicted by the IL4pred was located in the 3D model. The IL-4 motif location in the outermost area of the tip region was selected for ex vivo testing on bovine PBMCs.
All PBMC groups were harvested at 0, 6, 12, 24, 48, and 72 hpi. The supernatants of each hpi were tested for IL-4 protein levels by ELISA. All data were represented by the averages from nine wells (three cows each with triplicate wells).

Statistical Analysis
The data in Figures 2, 4, 5 and 7 were analyzed by two-way ANOVA in SPSS version 20.0 (IBM ® , Armonk, NY, USA) as previously described [7]. The data in Table 1 were analyzed by one-way ANOVA in SPSS version 20.0 (IBM ® , Armonk, NY, USA).

Conclusions
Although both growth curves of BTV2 and BTV12 are very similar in infected PBMCs, the cytokine profiles they induce differ significantly. BTV2 preferentially activates the Th2 pathway, represented by IL-4 expression, but it poorly activates the innate immunity. In contrast, BTV12 preferentially activates the innate immunity, represented by TNF-α and IL-1β, with only minimal IL-4 expression. The capability to induce IL-4 production is attributed to the tip region on the VP2 protein, and the number of predicted IL-4-inducing peptide segments contained therein is positively correlated with the allergic reaction reported in cattle or the more significant IL-4 expression in ex vivo settings.