Deletion in the S1 Region of Porcine Epidemic Diarrhea Virus Reduces the Virulence and Influences the Virus-Neutralizing Activity of the Antibody Induced

Porcine epidemic diarrhea virus (PEDV) causes severe diarrhea and a high rate of mortality in suckling pigs. The epidemic of PEDV that occurred after 2013 was caused by non-insertion and deletion of S gene (S-INDEL) PEDV strains. During this epidemic, a variant of the non-S-INDEL PEDV strain with a large deletion of 205 amino acids on the spike gene (5-17-V) was also found to co-exist with a non-S-INDEL PEDV without deletion (5-17-O). Herein, we describe the differences in the complete genome, distribution, virulence, and antigenicity between strain 5-17-O and variant strain 5-17-V. The deletion of 205 amino acids was primarily located in the S1O domain and was associated with milder clinical signs and lower mortality in suckling pigs than those of the 5-17-O strain. The 5-17-V strain-induced antibody did not completely cross-neutralize the 5-17-O strain. In conclusion, the deletion in the S1 region reduces the virulence of PEDV and influences the virus-neutralizing activities of the antibody it induces.


Introduction
Porcine epidemic diarrhea (PED) is a swine disease that causes severe watery diarrhea and vomiting in pigs of all ages. The clinical signs in PED virus (PEDV)-infected pigs are age-dependent, wherein younger pigs have higher morbidity and mortality [1]. In weaned pigs, PEDV affects their innate immunity through a decrease in Peyer's patch M cells and lysozyme expression in Paneth cells [2].
The PED virus is an enveloped, single-stranded, positive-sense RNA virus with the spike (S) protein on the envelope. It belongs to the order Nidovirale, family Coronaviridae, subfamily Coronavirinae, and genus Alphacoronavirus [3,4]. The genome of PEDV is approximately 28 kilobases encoding four non-structural proteins (ORF1a, ORF1b, ORF3a, and ORF3b) and four structural proteins (S, envelope (E), membrane (M), nucleocapsid (N)) [3,4]. The ORF3 protein has been associated with PEDV pathogenicity in that deletions between positions 245 and 295 of ORF3 decrease the virulence of PEDV [5]. The S protein binds to the aminopeptidase N receptors and is an important epitope of the neutralizing antibody [6][7][8][9]. The M protein plays important roles in viral assembly and the induction PEDV: 4-2 strain; S-INDEL strain: CV777; non-S-INDEL: USA/Indiana/17846/2013) and two novel PEDV strains (5-17-O and 5-17-V) were compared. Furthermore, the multiplex alignment of the S genes of Taiwanese PEDV strains and PEDV reference strains obtained from GenBank of NCBI was analyzed in DNASTAR software using the Clustal W method. Phylogenetic analysis of the aligned sequences was performed by the neighbor-joining method and by bootstrap analysis with 1000 replicates in the DNASTAR software (DNASTAR Inc., Madison, WI, USA).

Distribution of Two Novel PEDV Strains in the Farms of Taiwan
From the 2013 epidemic, two novel PEDV strains (5-17-O and 5-17-V) were isolated from diarrhea samples. To understand the population distribution of these two novel PEDV strains in farms, a total of 105 diarrhea samples from 64 farms (herds) were collected in the year 2014, and the strains were detected by differential diagnostic RT-PCR. Based on the sequences of the S genes of the two novel PEDV strains, universal primers (SF1 and SR739) for PEDV were designed. The primer sequences of SF1 and SR739 were 5 -TGCTAGTGCGTAATAATGAC-3 and 5 -CCTTCTGGTATGTGGCCATT-3 , respectively. The total volume of the reaction, 25

Growth Curves of Two Novel PEDV Strains in the Various Concentration of Trypsin
To study the effect of trypsin concentration on the replication of novel PEDV strains, the kinetics of PEDV replication were determined. First, PIs with 0.1, 1, 5, and 10 µg/mL trypsin were prepared. Monolayer Vero cells in the 6-well plates were inoculated with 0.1 multiplicities of infection (MOI) of 5-17-O or 5-17-V strains at 37 • C for 2 h. Afterward, the inoculants were removed by washing with PI and replaced with fresh PI. The culture supernatant and cells were harvested at 3 h intervals between 0 and 27 h post-inoculation (HPI), and the PEDV titers in 50% tissue culture infectious doses (TCID 50 ) were determined.

Characterization of the Virulence In Vivo
To compare the virulence of the 5-17-O and 5-17-V strains, an animal experiment approved by the Institution Animal Care and Use Committee (IACUC) of the Animal Health Research Institute (AHRI Approval number: A04006) was conducted. Thirty-six PEDV-seronegative, one-week-old, healthy piglets were randomly assigned to 9 groups of 4 and housed separately in 9 rooms. The piglets in groups 1-4 were inoculated orally with 10 3 , 10 4 , 10 5 , and 10 6 TCID 50 of the 5-17-O strain, respectively. Those in groups 5-8 were inoculated orally with 10 4 , 10 5 , 10 6 , and 10 7 TCID 50 of the 5-17-V strain, respectively. Those in group 9 were negative controls inoculated with MEM. After inoculation, the clinical signs, including diarrhea, weight, and body temperature, were recorded during the experimental period. Based on the fecal formation, diarrhea was scored as follows: 0 = solid; 1 = soft to pasty; 2 = semi-liquid; 3 = complete liquid. The feces and serum were collected, respectively, daily and at two-day intervals at 0-11 days post-inoculation (DPI). The piglets that died during the experimental period were necropsied. The surviving piglets were necropsied at 14 DPI. The viral loads were detected by real-time reverse-transcription PCR (RRT-PCR). The antibody against PEDV was detected by IFA and a virus-neutralizing (VN) assay.

Detection of PEDV Loads by RRT-PCR
The PEDV loads in the feces were detected by RRT-PCR. The primer and probe in the RRT-PCR of PEDV were designed to target the nucleoprotein gene. The sequences of the primer and probe were 5 -GCTTCTCAGAACAGAGGA-3 , 5 -CATCGCGTGATGTTACAC-3 , and 5 -FAM-CAATAACAAGTCTCGTAACCAGTCCAA-BHQ1-3 . The total volume of 20 µL contained 3 µL of extracted nuclear acids, 1× Kappa Probe Fast qPCR Master mix (KAPA biosystems, Boston, MA, USA), 40 units of SuperScript TM III Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), 0.3 µM of PEDV probe, and 0.5 µM of each PEDV primer. The reaction conditions involved initial incubation at 42 • C for 30 min and 94 • C for 5 min followed by 45 cycles of denaturation at 95 • C for 20 s, annealing at 60 • C for 30 s, and extension at 72 • C for 30 s. Finally, the threshold line was set at a fluorescence level of 5.

Detection of Anti-PEDV Antibody
The anti-PEDV antibody in pig serum was detected by IFA. Briefly, Vero cells in 96-well plates were inoculated with 200 TCID 50 PEDV at 37 • C with 5% CO 2 for 16 h. The supernatant was disposed, and the cells were dried at 37 • C for 1 h before being fixed with 10% neutral-buffered formalin at room temperature for 10 min and then washed 3x with PBS. The two-fold serially diluted serum sample was added onto wells at 37 • C for 1 h and then washed again. Then FITC-labeled goat-anti-swine IgG (Jackson ImmunoResearch Inc., West Grove, PA, USA) was added, and the mixture was held at 37 • C for 1 h, washed again, and observed under fluorescence microscopy.

Virus-Neutralizing Assay for PEDV
Anti-PEDV-neutralizing antibody in serum was determined in Vero cells using 5-71-V or 5-17-O strains. In a separate 96-well plate, two-fold serially diluted (from 8 to 512 fold) serum samples were mixed with 200 TCID 50 /0.1 mL PEDV (strain 5-17-O or 5-17-V) at 37 • C with 5% CO 2 for 2 h. The pre-seeded Vero cells were washed 2 times with phosphate-buffered saline (PBS) before the mixture was transferred and incubated at 37 • C with 5% CO 2 for 2 h. The mixture was removed from the cells and washed again with PBS before being placed in fresh PI medium and incubated at 37 • C with 5% CO 2 for 2 days. The cutoff was based on the fusion CPE.

Statistical Analysis
The Student's t-test, carried out in Microsoft Office Excel 2007, was used to compare the differences between the two groups. The comparisons of the groups were analyzed by one-way analysis of variance (ANOVA). The ANOVA combined with Duncan's multiple range test was performed in SAS for Windows 6.12 (SAS Institute Inc., Cary, NC, USA). A p-value < 0.05 was considered significant. In the S protein, 5-17-V had a deletion of 205 amino acids ( Figure 1) and a difference of three amino acids from the 5-17-O strain. The deletion was from the aligned positions 23 to 229 of the amino acids. The three changes from the 5-17-O to the 5-17-V strains were from S to G, from P to L, and from P to Q in the aligned positions 254, 506, and 1063, respectively. From the comparison of the deleted locations among the variant non-S-INDEL strains (5-17-V, LC022792, and KM392229), the same gap was located between the aligned positions 34 and 220. Based on the same gap, the gap of each strain extended several amino acids. In the S protein, 5-17-V had a deletion of 205 amino acids ( Figure 1) and a difference of three amino acids from the 5-17-O strain. The deletion was from the aligned positions 23 to 229 of the amino acids. The three changes from the 5-17-O to the 5-17-V strains were from S to G, from P to L, and from P to Q in the aligned positions 254, 506, and 1063, respectively. From the comparison of the deleted locations among the variant non-S-INDEL strains (5-17-V, LC022792, and KM392229), the same gap was located between the aligned positions 34 and 220. Based on the same gap, the gap of each strain extended several amino acids.

Comparison of Difference of Various Genes between Reference PEDV and Novel Taiwanese PEDV
Phylogenetic analysis of the S proteins revealed that the 5-17-O and 5-17-V strains were both classified as genotype 2, similar to those of non-S-INDEL strains in the USA, Japan, South Korea, and China ( Figure 2). The similarity of the S gene between 5-17-O and the other non-S INDEL strains was from 98% to 100%. The similarity of the S gene between 5-17-V and the variable non-S INDEL strains (LC022792 and KM 392229) was from 99.3% to 99.7%. On the other hand, the traditional Taiwanese PEDV strain, 4-2, was classified as group 1 (S-INDEL) and had a similarity of 97.8% to the S gene from KP768390 (Taiwan/HC070225/2007). deleted patterns were similar to the non-S-INDEL PEDV genotype strains. The 4-2 strain was the traditional Taiwanese PEDV strain, and its inserted and deleted patterns were the same as the KP768390 (Taiwan/HC070225/2007) and the other S-INDEL PEDV genotype strains (AF353511 and KJ399978). The stars indicate the same amino acid in all strains. The red boxes reveal the inserted and deleted regions between non-S-INDEL and S-INDEL strains. The blue box shows the difference in amino acids between 5-17-O and 5-17-V.

Comparison of Difference of Various Genes between Reference PEDV and Novel Taiwanese PEDV
Phylogenetic analysis of the S proteins revealed that the 5-17-O and 5-17-V strains were both classified as genotype 2, similar to those of non-S-INDEL strains in the USA, Japan, South Korea, and China ( Figure 2). The similarity of the S gene between 5-17-O and the other non-S INDEL strains was from 98% to 100%. The similarity of the S gene between 5-17-V and the variable non-S INDEL strains (LC022792 and KM 392229) was from 99.3% to 99.7%. On the other hand, the traditional Taiwanese PEDV strain, 4-2, was classified as group 1 (S-INDEL) and had a similarity of 97.8% to the S gene from KP768390 (Taiwan/HC070225/2007).  and 5-17-O strains were the same for the 5 UTR, ORF1a, ORF1b, ORF3, E, M, N, and 3 UTR genes; the similarity of nucleotides was between 99.6% and 100% for the non-S-INDEL strain (KF452323) and between 96.7% and 99.4% for the S-INDEL strain (MW165329 and AF353511) (Table 1)

Comparison of Replication Kinetics between the 5-17-O and 5-17-V Strains in an In Vitro Experiment
The replication kinetics of the two novel PEDV strains in various concentrations of trypsin were determined by an in vitro study. The growth curves of the two PEDV strains in each condition were established and compared to each other at the same time ( Figure 3). The growth curves of the two PEDV strains in the same condition did not differ significantly. The PEDV titer in each condition did not significantly increase in the early 9 HPI. 1 μg/mL) at 12, 15, and 18 HPI. After 18 HPI, the titers of the two PEDV strains in the high trypsin concentrations gradually decreased. In the low trypsin concentration, the titer of the 5-17-O strain slowly increased from 10 4.5 to 10 5.81 TCID50/mL of 5-17-O/0.1 and from 10 4.54 to 10 5.81 TCID50/mL of 5-17-O/1 at 9 to 27 HPI, and the peak titer was at 27 HPI. However, the lower titer was revealed in the 5-17-V/0.1 and 5-17-V/1. The peak titers of the 5-17-V strain in 0.1 and 1 μg/mL trypsin were only 10 4.9 TCID50/mL at 15 HPI and 10 5.5 TCID50/mL at 18 HPI, respectively.

Viral Shedding in Feces
No PEDV was detected in the feces of piglets in the negative controls (group 9) during the experimental period. The PEDV shedding from feces was found in all piglets inoculated with PEDV ( Table 2). In piglets inoculated with the 5-17-O strain (groups 1-4), PEDV shedding was detected within 2 DPI and continued to 6 DPI. After 7 DPI, some piglets recovered gradually, but in each group, one or more piglets continuously shed PEDV in feces until the end of the experiment. The earliest PEDV shedding time was detected at 1 DPI in piglets inoculated with 10 7 TCID50 of the 5-17-V strain (group 8) and 10 6 TCID50 of the 5-17-O strain (group 4). Comparison of average shedding days between the 5-17-O and 5-17-V strains showed no significant difference (Table 2)  The piglets in groups 5 to 8 were inoculated with 10 4 to 10 7 TCID 50 of 5-17-V (dotted and black lines), respectively. Group 9 was the negative control. The diarrhea was scored as follows: 0 = solid; 1 = soft to pasty; 2 = semi-liquid; 3 = complete liquid.

Viral Shedding in Feces
No PEDV was detected in the feces of piglets in the negative controls (group 9) during the experimental period. The PEDV shedding from feces was found in all piglets inoculated with PEDV ( Table 2). In piglets inoculated with the 5-17-O strain (groups 1-4), PEDV shedding was detected within 2 DPI and continued to 6 DPI. After 7 DPI, some piglets recovered gradually, but in each group, one or more piglets continuously shed PEDV in feces until the end of the experiment. The earliest PEDV shedding time was detected at 1 DPI in piglets inoculated with 10 7 TCID 50 of the 5-17-V strain (group 8) and 10 6 TCID 50 of the 5-17-O strain (group 4). Comparison of average shedding days between the 5-17-O and 5-17-V strains showed no significant difference (Table 2).

The Titer of Anti-PEDV Serum Antibody Detected by IFA and Neutralizing Assay in Experimentally Infected Pigs
The anti-PEDV antibody was not detected in any piglets by IFA before the in vivo experiment. At 6 DPI, anti-PEDV antibodies were detected by IFA in the piglets in the 5-17-V inoculation (groups [5][6][7][8], and the titer gradually increased from 6 to 10 DPI. However, in 5-17-O inoculated piglets (groups 1-4), only two piglets with inoculation of 10 6 TCID 50 (group 8) had anti-PEDV antibodies at 10 DPI. However, no neutralizing antibodies against the 5-17-V and 5-17-O strains were detected in the sera of all piglets at 10 DPI.

Discussion
The S protein is the target gene for evaluating the evolution of PEDV and the origin of PEDV transmission. The non-S-INDEL PEDV, with insertion and deletion in the S protein, has rapidly spread globally and possibly originated from the AH2012 strain [21,27]. In the 2013 PED epidemic, when non-S-INDEL PEDV first emerged in Taiwan [19], we found a variant of the non-S-INDEL

Discussion
The S protein is the target gene for evaluating the evolution of PEDV and the origin of PEDV transmission. The non-S-INDEL PEDV, with insertion and deletion in the S protein, has rapidly spread globally and possibly originated from the AH2012 strain [21,27]. In the 2013 PED epidemic, when non-S-INDEL PEDV first emerged in Taiwan [19], we found a variant of the non-S-INDEL PEDV strain (5-17-V) with a deletion of 205 amino acids in the S1 region during the routine diagnosis of swine diseases. The PEDV with large deletions in the S1 region was also found in Japan (deletion of 194 amino acids) [28] and the USA (deletions of 194 to 204 amino acids) [29]. Nevertheless, the size and locations of the deletions in the S1 region differed slightly among these various PEDV strains. The deletion size ranged from 194 to 205 amino acids, and the segment between residues 34 and 212 of the S protein was completely absent. The deletion of 205 amino acids also occurred in the 5-17-V strain and JKa-295/CS1de205 (KU363101) [30] but in different locations. The variant non-INDEL PEDV strain, 5-17-V, was first found in Taiwan.
The phylogenetic analysis of the S gene revealed that 5-17-O was a Taiwanese non-S-INDEL PEDV strain. The inserted and deleted patterns and sequences in the non-S-INDEL found in Taiwan were highly similar to those of the non-S-INDEL strains found in the USA, Japan, South Korea, and China and was different from that of the traditional Taiwanese PEDV strain. These results indicated that the Taiwanese non-S-INDEL was, in fact, a new invasive strain. The same discourse was demonstrated in a recent study showing that the Taiwanese non-S-INDEL PEDV strains were highly related with the non-S-INDEL strain in the USA [19]. In addition, the variable non-S-INDEL strain, similar to the 5-17-V strain with a large deletion, was reported in the USA and Japan. Our investigation found that the 5-17-V and 5-17-O usually co-existed in the same sample. Therefore, the piglets might be simultaneously infected with both the 5-17-V and 5-17-O strains, suggesting that the 5-17-V invasion followed that of the 5-17-O invasion at the same stage. Viral evolution allows viruses to adapt to changes in host environments, leading to the development of various strains. The pathways of viral evolution include mutation, insertion, deletion, and recombination of genes. The non-S-INDEL PEDV strain is a new PEDV strain of genotype 2 with mutation, insertion, and deletion as compared with the traditional PEDV strain of genotype 1. In this study, we found that the variant 5-17-V strain coexisted in 14% of cases (10/73 samples, see Section 3.3) of non-S-INDEL PEDV infection. This large deletion of the S protein has also occurred in the evolutions of other coronaviruses such as transmissible gastroenteritis virus (TGEV) [31] and Middle East respiratory syndrome coronavirus [32]. The gene deletion of TGEV led to a change in the virus tropism from intestinal enterocytes to respiratory tissues, indicating that the deletion of the S protein is one pathway in the process of coronavirus evolution.
Trypsin plays an important role during PEDV replication. Recent studies and this study demonstrate that insufficient load and lack of trypsin decrease the efficacy of PEDV production in cells [33,34]. Furthermore, we also found that the replication of the 5-17-V strain was significantly more trypsin dependent than the 5-17-O strain. The difference in trypsin dependence between these two strains is presumed to be associated with the deletion of the PEDV S protein. Strain 5-17-V had the deletion of 205 amino acids of the S protein located mainly in the S1 O domain and partially in the S1 A domain. During PEDV infection, trypsin can cleave the S protein into S1 and S2 subunits, leading the S1 subunit to bind to the aminopeptidase N receptor [33,34]. If the S1 subunit is thus incomplete, it could possibly lower the binding capability of PEDV. In addition, trypsin is able to enhance the viral budding and fusion formation of CPE in infected cells [33,34]. The effect of both a lower trypsin concentration and an incomplete S protein of PEDV could significantly decrease the efficacy of PEDV replication.
The S protein is one of the determinants of virulence in the pathogenesis of PEDV [35], wherein high virulence in the suckling pigs occurred with the non-S-INDEL PEDV strains. When the S1 region of S-INDEL PEDV strains was experimentally replaced with that of non-S-INDEL strains, more severe clinical signs were observed in suckling pigs infected with the recombined strain [33].
In this study, we demonstrated milder clinical signs and lower mortality in suckling pigs with 5-17-V infection, indicating that the deletion of 205 amino acids decreased the virulence of PEDV. The variant non-S-INDEL (5-17-V), a low-virulent PEDV strain, has the advantage of easier attenuation, but its antigenicity to induce the antibody production is not completely characterized. For the development of a vaccine candidate, the reduced antigenicity of the variant 5-17-V strain would be improved by an adjuvant to further induce the adaptive immunity. In addition, higher infection doses induced more severe diarrhea and higher mortality in suckling pigs, as found in other studies [36,37], implying that the infectious dose directly influences the PEDV pathogenesis.
The receptor-binding domain of PEDV, which resides in the S protein, is also associated with the induction of a neutralizing antibody [7,8]. The S1 B domain of PEDV attaches to the porcine aminopeptidase N (pAPN) receptor during virus entry [9,38]. The S1 O domain, which binds to Sia, is associated with the hemagglutinating activities of PEDV [39,40]. The large deletion of S in the 5-17-V strain, located mainly in the S1 O domain and partially in the S1 A domain, induced a neutralizing antibody that has a lower capacity to cross-neutralize 5-17-O ( Figure 5), giving the virus a survival advantage. The neutralizing epitopes of the S1 O domain were also determined [41], suggesting that positions 23-227 of the S protein contain the potent neutralizing epitopes, and this deletion affected the neutralizing activity of PEDV-induced serum.

Conclusions
In conclusion, the variant PEDV (5-17-V) with a large deletion in S was first found in Taiwan and commonly coexists in herds infected with PEDV without a large deletion in S. This large deletion, primarily located in the S1 O domain, is associated with the reduced virulence of PEDV in vivo and influences the neutralizing activities of the antibody induced by the variant PEDV to cross-neutralize those PEDVs without a large deletion on S, providing the virus a survival advantage.