Genetic responses of inbred chicken lines illustrate importance of eIF2 family and immune-related genes in resistance to Newcastle disease virus

Newcastle disease virus (NDV) replication depends on the translation machinery of the host cell; therefore, the eukaryotic translation initiation factor 2 (eIF2) gene family is a likely candidate for control of viral replication. We hypothesized that differential expression of host genes related to translation and innate immune response could contribute to differential resistance to NDV in inbred Fayoumi and Leghorn lines. The expression of twenty-one genes related to the interferon signaling pathway and the eIF2 family was evaluated at two- and six-days post infection (dpi) in the spleen from both lines, either challenged by NDV or nonchallenged. Higher expression of OASL in NDV challenged versus nonchallenged spleen was observed in Leghorns at 2 dpi. Lower expression of EIF2B5 was found in NDV challenged than nonchallenged Fayoumis and Leghorns at 2 dpi. At 2 dpi, NDV challenged Fayoumis had lower expression of EIF2B5 and EIF2S3 than NDV challenged Leghorns. At 6 dpi, NDV challenged Fayoumis had lower expression of EIF2S3 and EIF2B4 than NDV challenged Leghorns. The genetic line differences in expression of eIF2-related genes may contribute to their differential resistance to NDV and also to understanding the interaction between protein synthesis shut-off and virus control in chickens.


NDV challenge effect.
To evaluate the effect of NDV challenge, host responses were assessed by comparing expression of 21 genes between NDV challenged and nonchallenged spleens within the Fayoumi and the Leghorn chicken lines at 2 and 6 dpi ( Table 1; Fig. 1). At 6 dpi, Janus kinase 1 (JAK1) expression was higher in the spleen of NDV challenged than nonchallenged Leghorns (FDR = 0.06; Fig. 1d). NDV challenge effect was also observed in the expression of 2′-5′-oligoadenylate synthetase like (OASL), an ISG related to dsRNA degradation; higher expression of this gene in NDV challenged versus nonchallenged spleens was found in Leghorns at 2 dpi (FDR = 0.08; Fig. 1c). The NDV challenged Fayoumis had lower expression of the EIF2B5 gene than nonchallenged Fayoumis at 2 dpi (FDR = 0.02; Fig. 1e) and 6 dpi (FDR = 0.07; Fig. 1f). Expression patterns also showed lower expression of the EIF2B5 gene in NDV challenged than nonchallenged Leghorns at 2 dpi (FDR = 0.09; Fig. 1e).
Line effect. To further evaluate differences in the response of each line to NDV challenge, we compared the gene expression of receptors, cytokines and signaling molecules related to the early-antiviral activity between Fayoumis and Leghorns within challenge status and dpi (Table 1; Fig. 2). For these genes, line effect was only found for JAK1 expression where Leghorns had higher expression of this gene at 6 dpi in both challenge statuses (FDR = 0.07; Fig. 2d).
Because Fayoumis had faster NDV clearance than the Leghorns as previously reported 20 , we also evaluated the differences between lines in the expression of eIF2 family-related genes to better understand how NDV replication may be delayed by host protein synthesis shut-off (Table 1; Fig. 3). Within nonchallenged chickens, effect of chicken genetic line was found on the expression of two eIF2 genes: Fayoumis had higher expression of EIF2B5 (2 and 6 dpi; both FDR = 0.06; Fig. 3a,b), and highly significant lower expression of EIF2S3 (2 and 6 dpi; both FDR < 0.001; Fig. 3c,d) than the Leghorn chickens. Comparing genetic lines within NDV challenged chickens, the Fayoumis again had highly significant lower expression of EIF2S3 (2 and 6 dpi; both FDR < 0.001; Fig. 3c,d) than the Leghorns. NDV challenged Fayoumis also had significantly lower expression of EIF2B5 compared to the Leghorn chickens, but only at 2 dpi (FDR = 0.02; Fig. 3a). The highest fold changes were observed in expression of the EIF2S3 gene. The lower expression of eIF2 family genes in NDV challenged Fayoumis may illustrate a mechanism of resistance, in that NDV propagation may be delayed in Fayoumis compared to Leghorn due to a reduction in the host protein synthesis machinery. Different subunits of eIF2 and eIF2B with different functions can be regulated to help the innate immune response contain NDV replication.
Day post infection effect. The heat map (Fig. 4) shows comparisons between 6 and 2 dpi within line and challenge statuses. NDV challenged Fayoumis had higher expression of the JAK1 (FDR = 0.07), OASL (FDR = 0.08) and EIF2B4 (FDR = 0.04) genes at 2 than at 6 dpi (Table 1; Fig. 4). Within NDV challenged Leghorns, higher expression of EIF2B5 (FDR = 0.04) and lower expression of IFNB (FDR = 0.09) were observed at 2 than at 6 dpi. The patterns of gene expression for the NDV challenged Fayoumis and Leghorns can be helpful to understand how each line can repress the replication of NDV at early stages of infection.

Discussion
The severity of infection with NDV is determined by many different factors such as tissue tropism, efficiency of replication, concurrent diseases, and host resistance to the virus 26 . Some of the NDV response traits are known to have a component of genetic control 27 . Gene expression changes in the immune response to the NDV challenge in two highly inbred lines and their sublines, the Fayoumi (M5. 1 Table 1. Gene expression in the spleen of Fayoumi and Leghorn chickens NDV challenged or nonchallenged at 2 and 6-days post infection (dpi). Pairwise comparisons between groups were based on a linear model and tested effects of line, challenge status, dpi. Significant line differences are shown as bold capitalized letters (FDR < 0.05) and trends as capitalized letters (FDR < 0.1). Significant differences due to challenge status are indicated by bold small letters (FDR < 0.05) and trends by small letters (FDR < 0.1). Significant differences across dpi are illustrated with *(FDR < 0.05) and trends with † (FDR < 0.1). 1 ISG, Interferon stimulated genes. 2 NDV, NDV challenged. 3 Non, nonchallenged. www.nature.com/scientificreports www.nature.com/scientificreports/ (Ghs6 and Ghs13), have shown that the Major Histocompatibility Complex (MHC) can contribute to the genetic resistance against NDV 28 . Evaluation of the transcriptome of these two genetically distinct lines also suggests that NDV infection may cause upregulation of different genes in a line-dependent manner 19 . These results were also observed in the spleen, where few interferon stimulated genes were upregulated by NDV challenge 21 .
In the current study, the observed time effect on the expression of IFNB may help to explain the different rates of viral clearance between lines, as the Fayoumi had lower viral load than the Leghorn at 6 dpi 19,20 . Recognition of viral RNA by TLR7 triggers signaling for cytokines with antiviral activity to be upregulated 29 . Infection with NDV causes a strong immune response as showed by the upregulation of different cytokines and chemokines 4 . Type I interferon production is also induced by NDV infection 30 and favors an effective host response. Interferon α and β bind to a common receptor comprising IFN-α receptor 1 (IFNAR1) and IFNAR2 subunits and signal through the JAK-STAT pathway to induce the transcription of several ISGs related to the antiviral response 31 . Although, IFNα and β play an important role against viral infection, each one may signal through different ISGs and may have different antiviral activities 32 . IFNβ may be mainly committed to signaling and immune modulation, while IFNα could have the strongest antiviral activity against vesicular stomatitis virus, avian influenza virus, and NDV  www.nature.com/scientificreports www.nature.com/scientificreports/ infection in chickens 32 . Because interferons act in the first line of defense against viruses, the expression pattern for IFNB after infection, i.e. higher expression in Leghorn at the later time, could indicate later signaling to effector molecules in the Leghorns.
OAS is one of the main IFN effectors that acts in the early phase of viral infection to initiate degradation of viral RNA, resulting in the inhibition of viral replication 33 . After being activated, OAS promotes the synthesis of 2-5 oligoadenylate, which activates RNase L. Upon binding 2′-5′-linked oligoadenylates (2-5 A), activated RNase L degrades cellular and viral RNA 34 . This RNase L-dependent pathway is best characterized; however, recent studies have described OAS antiviral activity through a RNase L-independent pathway 35,36 . In chickens, OAS has been shown to be encoded by only one gene (OASL) 37 and to be upregulated after infection with different viruses 38,39 including NDV 3,4,21 . The role of OASL in the control of NDV replication was confirmed in vitro 8 , as the overexpression of OASL reduced the replication of NDV and the absence of OASL significantly enhanced viral replication. In our study, OASL was upregulated by NDV challenge in Leghorns at 2 dpi. Upregulated expression of the OASL gene was also observed in challenged Fayoumi and Leghorn in our previous study using RNAseq. 21 . Besides the function described above, OAS also is related to programed cell death 40 . Genetically controlled apoptosis in response to viral infection can help to reduce the spread of progeny virus 41 ; however, the role of the OAS/apoptosis interaction in the response to NDV infection is unclear and needs to be further investigated. Lower expression of OASL in the spleen of NDV challenged Fayoumis was found at 6 than at 2 dpi. Because the OAS-RNase L pathway requires production of both viral dsRNA and type I IFN together to upregulate OAS and activate RNase L 42 , we suggest the decreased expression of OASL at 6 dpi in the Fayoumis is due to the lower type I interferon expression and the lower viral load 20 in Fayoumis at this time point.
As protein synthesis shut-off in chicken cells can be used as a defense mechanism in response to viral infection 10 , we evaluated the effect of NDV challenge on the expression of genes related to eIF2 family. eIF2 is a protein composed of a regulatory α-subunit (encoded by the EIF2S1 gene), a tRNA-binding β-subunit (EIF2S2 gene) and a GDP/GTP-binding γ-subunit (EIF2S3 gene) that plays a critical role in protein synthesis regulation 43 . The initiation step of translation requires the ternary complex formed by eIF2-GTP and tRNA Met to assemble the pre-initiation complex. To set up the ternary complex, eIF2:GDP binds to eIF2B, a guanine nucleotide exchange factor, which makes the exchange of GDP by GTP 44 . Although the precise mechanism and function of eIF2 subunits has not been fully described in chickens, in mammals this process can be regulated to prevent viral replication and spread to other tissues through protein synthesis shut-off 45 .
PKR is another ISG to be activated after viral infection and is been related to the reduction in protein synthesis 46 . In chicken cells challenged by NDV, PKR undergoes autophosphorylation and phosphorylates the eIF2α subunit 10 . Phosphorylation of eIF2α turns it into an inhibitor of the guanine nucleotide exchange factor eIF2B and prevents the formation of the translation initiation complex eIF2:GTP:tRNA Met that is required for the initiation of protein synthesis 15 . In vitro NDV infection caused PKR activation, eIF2α phosphorylation and viral replication inhibition in HeLa cells 18 . Phosphorylation of eIF2α after NDV infection in DF-1 chicken fibroblast cells also induces higher translation of ATF4 and an increase in growth arrest and DNA-damage-inducible protein 34 (GADD34) expression 10 . GADD34 has been described as the link between eIF2α phosphorylation and immune system, however, the mechanism by which GADD34 controls cytokine synthesis after viral infection remains unclear 47 . The phosphorylation of eIF2α can also occur by Protein kinase R-like endoplasmic reticulum kinase (PERK) induced by endoplasmic reticulum stress 16 , however the role of PERK in viral containment in chickens has not been investigated.
Because protein translation is required for viral replication, viruses have developed strategies to avoid protein synthesis shut-off 48 . NDV can manipulate the PKR/eIF2a signaling cascade to favor viral replication by arresting  Table 1  www.nature.com/scientificreports www.nature.com/scientificreports/ cellular mRNA inside stress granules 49 . We hypothesize that the host cell can use different mechanisms to overcome these viral strategies. Here we show for the first time the relationship of NDV infection on the expression of genes encoding subunits of eIF2 and eIF2B. We observed that NDV challenge downregulated the expression of EIF2B5 in the Fayoumis at 2 and 6 dpi and in Leghorns at 2 dpi. This suggests that this gene could be downregulated to facilitate protein synthesis shut-off at an early stage of infection.
In mammals, eIF2B contains five subunits. The eIF2ε subunit has the catalytic domain and forms the catalytic subcomplex together with the eIF2γ subunit. They are encoded by EIF2B5 and IEF2B3 genes, respectively. The subunits eIF2Bα (encoded by EIF2B1 gene), β (EIF2B2 gene) and δ (EIF2B4 gene) form the regulatory subcomplex 43 . In this study, we assumed that the chicken genes encoding these subunits have the same interactions as in the mammalian model because eIF2 and eIF2B structures have not been fully investigated in chicken. We present the novel finding that infection with NDV could result in the inactivation of the translation initiation factors eIF2 and/or eIF2B in a chicken tissue. We demonstrated that NDV challenged Fayoumis had lower expression than NDV challenged Leghorns of a gene encoding a eIF1B subunit with regulatory (EIF2B5) activity, as well as a gene encoding a subunit of eIF2 (EIF2S3). Both of these genes could contribute to the difference in relative resistance between these lines. It is important to notice that lower expression of EIF2S3 gene was also observed in Fayoumis than Leghorns for nonchallenged chickens, which shows a difference between the genetic lines under homeostatic conditions. The activity of eIF2B appears to depend on the phosphorylation of eIF2α as well the phosphorylation in the eIF2B itself 50 . Other results demonstrate that the catalytic and regulatory subcomplexes in eIF2B can both bind eIF2 independently in vitro 51 ; the interaction between eIF2 and the regulatory subcomplex is increased by phosphorylation of eIF2, whereas the interaction of catalytic subcomplex and eIF2 it is not dependent on eIF2 phosphorylation. Dysregulation of eIF2Bα may be sufficient to make cells more susceptible to viral infection by neutralizing the consequences of eIF2α phosphorylation 52 . Finally, reduced expression of the eIF2Bε subunit has been related to reduced protein synthesis in vitro 53 . All these results show the relation between different eIF2B subunits and the control of protein synthesis. Because inhibition of eIF2B can cause the down-regulation of global protein synthesis by reducing the cellular level of ternary complexes that are available for translation initiation 13 , the differential expression of EIF2B5 observed between the two challenged lines could be related to the higher ability of Fayoumis to reduce the replication efficiency of NDV. It should be noted that our results were observed in the spleen which could suggest that the eIF2/eIF2B pathway could act as systemic mediators of interferons.
We suggest that the Fayoumis are more resistant to the NDV challenge than the Leghorns due in part to differences in the expression of eIF2 and eIF2B genes, which act on host protein synthesis machinery and, thus, can regulate translation to contain viral replication. Our results describe for the first time the differential expression of eIF2 and eIF2B subunits during NDV challenge in chickens. Because the application of information about resistance mechanisms in disease genetic control studies may lead to better resistance to Newcastle disease, our results suggest that eIF2 family should be the focus of further studies evaluating host genetic resistance to NDV.

Methods
Ethics statement. The activities of this experiment were approved by the Iowa State University Institutional Animal Care and Use Committee (IACUC log number 1-13-7490-G). All animal use and procedures were performed according to the approved protocol and in accordance with the relevant guidelines and regulations.
Animals and experimental design. The effect of NDV challenge on the gene expression in the spleen of two different chicken lines at two-and six-days post infection (dpi) was evaluated. The Fayoumis (M15.2) and the Leghorns (GHs 6), two inbred lines (inbreeding coefficients = 99.95% 54 ) were divided in two treatments regarding NDV challenge. At 21 days of age, half of the chickens were inoculated with 200 µl of 10 7 EID 50 of La Sota NDV (NDV challenged group, n = 49) through nasal and ocular inoculation routes. The other chickens received 200 µl of phosphate-buffered saline (PBS) (nonchallenged group, n = 40) through same conditions, as described in [19][20][21] . The chickens from both lines and challenge statuses (NDV challenged or nonchallenged) were euthanized with sodium pentobarbital solution at 2 or 6 dpi. At each dpi, chickens from each treatment were randomly selected for gene expression analysis. Groups are: Fayoumi NDV challenged (n = 4), Fayoumi nonchallenged (n = 3), Leghorn NDV challenged (n = 4) and Leghorn nonchallenged (n = 3). Samples were collected from spleen and placed into RNAlater solution (ThermoFisher Scientifc, Waltham, MA) and stored in a −80 °C freezer until the RNA isolation procedure. The samples used for gene expression analysis in the current study were previously used by 21 . The detailed description for the experimental design and animal activities can be found in 20 .
Gene expression. The spleen tissues were homogenized using mechanical disruption and total RNA was isolated using the Ambion RNAqueous Total RNA Isolation Kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer's protocol. All the samples were treated with DNase using the DNA-free kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer's recommendations. Proper quantity and quality of the RNA samples were ensured respectively through assessment by NanoDrop ND-1000 UV-vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA) and the RNA 6000 Nano kit on the Agilent 2100 Bioanalyzer (Agilent Technology, Santa Clara, CA). The RNA integrity numbers are average = 8.  www.nature.com/scientificreports www.nature.com/scientificreports/ Tool. IFNA is a multi-copy gene family, and since primers used for IFNA in this study will bind to more than one of these gene copies, IFNA expression should be considered as representing the gene family. Primers sequences for the housekeeping gene (28 s gene) were previously reported 55 . All of the reactions were performed in triplicate and each plate contained both a negative and a no reverse transcriptase control. Data analysis. Gene expression data were evaluated as the adjusted Ct (cycle threshold) value using the following formula: 40 − [(mean test gene Ct) + (median 28S Ct − mean 28S Ct) × (test gene slope/28S slope)]. Slopes were determined with five points of a 10-fold dilution series using pooled RNA for the housekeeping gene and target-specific amplicons as the template to determine PCR efficiency of each primer/gene set. Median 28S Ct represents the median Ct value of all individual samples for the housekeeping gene. Additionally, the fold change was used to display the differential gene expression in Figs. 1-3.
The expression of each gene (expressed as adjusted Ct values) was fit to a linear model (adjusted Ct ~ group, where group incorporates line, challenge status and dpi) using R 3.6.1 and analyzed by ANOVA using the car package 3.0-2. The multcomp package 1.4-10 was used to test pairwise contrasts between groups, testing the effect of line, challenge status, and dpi (12 total comparisons) and included False Discovery Rate (FDR) correction based on the Benjamini-Hochberg procedure. FDR-adjusted p-values were considered significant at FDR < 0.05; FDR < 0.1 was used to show results of genes that had an expression pattern similar to the others in a pathway. All FDR-adjusted p-values for each gene from all 12 contrasts are reported in Supplementary Table S1.