PK15^CD163 (obtained from Prof. En-ming Zhou from Northwest A&F University), MARC-145, and PK15 cells were cultured at 5% CO₂ and 37°C in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, South Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, South Logan, UT, USA). The porcine alveolar macrophages (PAMs) were cultured at 5% CO₂ and 37°C in RPMI-1640 medium (Gibco, South Logan, UT, USA) supplemented with 10% FBS. PRRSV strain WUH3 (GenBank accession No.HM853673) was provided by Prof. Shaobo Xiao from Huazhong Agricultural University (31). Cells were infected with PRRSV for 1 h at 37°C, and then the culture medium was replaced with maintenance medium containing 2% FBS until cells were collected. To ensure the consistency of the expression levels of FAM134B protein in mock groups at different time points post infection, the DMEM medium containing 10% FBS was used. PRRSV infection experiments in vivo were performed according to the previously reported method (32). The PAMs were collected from PRRSV-infected pigs at 0 day post infection (dpi), 7 dpi, and 14 dpi, respectively.

SiRNAs targeting FAM134B (siFAM134B-1: CCACTGTTCGCAGAATCA, siFAM134B-2: GAAGGATACACTCCACAGA and siFAM134B-3: CCATCAAAGACCAGTTAGA) were synthesized by Ribobio (Guangzhou, China). The sequences of siRNA targeting TLR3, RIG-I, and ATG7 were synthesized, as previously described (33, 34). pGL3-IFNβ (IFNβ-Luc) was provided by Dr. Likai Ji from Shanghai Jiao Tong University (35). Trelief™ SoSoo Cloning kit (Tsingke Biotechnology, Wuhan, China) was used to construct plasmids pcDNA3.1-FAM134B, eGFP-FAM134B, mCherry-LC3B, mCherry-SEC61B, psi-Check2-FAM134B-3’UTR (FAM134B-3’UTR), and CMV-mCherry-eGFP-RAMP4-IRES-Puro. In addition, site-mutagenesis plasmids FAM134B-LIR-MUT and FAM134B-3’UTR-MUT were synthesized by Tsingke Biotechnology. The miR-142-5p (CAUAAAGUAGAAAGCACUACU) mimics, inhibitor (AGUAGUGCUUUCUACUUUAUG), and negative control (NC, UUCUCCGAACGUGUCACGUTT) were synthesized by GenePharma (Shanghai, China). The 25 nM siRNA/mimics or 2ug plasmids were transfected into cells using Lipofectamine™ 2000 transfection reagent (Invitrogen™, Carlsbad, CA, USA) for subsequent experiments.

The total protein of cells was extracted using RIPA lysis buffer (Beyotime, Shanghai, China), PMSF (Beyotime, Shanghai, China), and phosphatase inhibitor Cocktail I (MedChemExpress, Shanghai, China). Total protein was subjected to electrophores using SDS-PAGE (90 V for 0.5 h and 120 V for an additional hour), and then was transferred to polyvinylidene difluoride (PVDF) membrane (Millipore, MA, USA). PVDF membranes were blocked with 5% non-fat dry milk for 2.5 h at room temperature, and subsequently incubated with specific primary antibodies overnight at 4°C. After being washed with TBST (0.01% Tween-20 in TBS) for three times, the membranes were then incubated with corresponding secondary antibodies for 2 h at room temperature. Immunoblot was performed using ImageQuant LAS4000 mini (GE Healthcare Life Sciences, Piscataway, NJ, USA). All antibodies are shown in Table S1 .

For immunostaining, cells were fixed with 15% paraformaldehyde for 15 min, and the membranes were permeabilized with 0.3% Triton X-100 for 10 min. Then, the cells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature, and incubated with primary antibodies overnight at 4°C, followed by incubation with corresponding fluorescent secondary antibodies for 2 h at room temperature. The images were captured by confocal fluorescence microscopy (Zeiss LSM 800, Oberkochen, Germany) and analyzed by ZEN software (ZEISS, Oberkochen, Germany). The antibodies used are listed in Table S1 . MARC-145 cells or PAMs were fixed with Glutaraldehyde (Solarbio, Beijing, China) and subsequently observed with a transmission electron microscope (HITACHI H-7000FA, Tokyo, Japan).

Total RNA was extracted by RNAiso Plus (TAKARA, Tokyo, Japan), and reverse transcribed into cDNA using a RevertAid First Strand cDNA Synthesis kit (Thermo Fisher Scientific, Carlsbad, CA, USA). The cDNA was subjected to quantitative real-time PCR analysis using TB Green^® Premix Ex Taq™ (TAKARA, Tokyo, Japan) on a BioRad CFX384 system (Bio-Rad, Richmond, CA, USA). The details of the primers are listed in Table S2 .

Lentiviral eukaryotic expression vector CMV-mCherry-eGFP-RAMP4-IRES-Puro and helper vectors PMG2.G and PSPAX were co-transfected into the 293T cells at the mass ratio 3:2:1. At 48 h post co-transfection, the cell supernatant was collected and filtered with a 0.45 um filter to remove dead cells or cell debris. The collected cell supernatant was ultracentrifugated at 25000 g for 2 h. Then, 300 ul serum-free DMEM was added to obtain diluted lentivirus, and the diluted lentivirus was stored at -80°C until subsequent use. MARC-145 cells were infected with lentivirus and added with 1μg/ml polybrene to increase infection efficiency. The infected MARC-145 cells were sorted by flow cytometry to prepare single cell suspensions. The positive cells stably expressing GFP/mCherry were selected for subsequent experiments.

Briefly, MARC-145 cells were inoculated onto 96-well plates. When the confluence reached 80%, the MARC-145 cells were infected with 10-fold serially diluted PRRSV. The cells were incubated for 72 to 120 h, and then virus titers were measured. PRRSV titers were expressed as TCID₅₀ per milliliter using the Reed-Muench method.

PK15 or PK15^CD163 cells were inoculated onto 24-well plates. The cells were then subjected to dual-luciferase assay using a Dual-Luciferase Reporter Assay System (Promega, Wisconsin, USA) with the Enspire reader (PerkinElmer, Singapore) at 24 h post co-transfection with corresponding plasmids (500 ng) and pRL-TK (50ng). After co-transfection, PK15^CD163 cells were infected with PRRSV for another 24 h for detecting the promoter activity of IFN-β.

After the cells were trypsinized and resuspended in complete medium without fixation, flow cytometry assay of cells was performed by ER autophagy tandem reporter (EATR). A total of 20000 cells were counted for each treatment. The acid gate was set according to previously-reported method (36). The data were then analyzed by eGFP fluorescence and mCherry fluorescence using FlowJo software.

PRRSV-infected PK15^CD163 cells were transfected with FAM134B or control overexpression vector for 24 h. Then, total RNA was extracted for RNA-seq library construction. Subsequently, hisat2 software (37) was used to align the high-quality reads to the pig reference genome (ENSEMBL Suscrofa11.1). Finally, differential expression of genes in each sample was analyzed using edgeR software (38). The raw RNA-seq data were submitted to SRA database (NO. PRJNA773967).

All the data had at least three biological replicates, and they were presented as the mean ± SD. Kruskal Wallis test was performed for multiple comparisons, and Wilcoxon test was conducted for two comparisons. P-value < 0.05 and < 0.01 were considered as statistically significant and highly significant, respectively.

PRRSV infection has been reported to change ER morphology and enlarge ER volume in the infected cells (30). Consistent with this report, through transmission electron microscopy, we observed ER expansion in the PRRSV-infected MARC-145 cells and the PAMs in vitro ( Figure 1A ) as well as in the PAMs from PRRSV-infected pigs in vivo ( Figure 1B ). Furthermore, we also observed that the staining of ER marker protein cytoskeleton-associated protein 4 (CKAP4) was enhanced in the PRRSV-infected MARC-145 cells ( Figure 1C ). Based on these observations, we hypothesized that ER-phagy, as the key step of ER-quality control, might affect ER morphology in the PRRSV-infected cells. To test this hypothesis, we constructed PRRSV-susceptible ER autophagy tandem reporter (EATR) system. PRRSV-susceptible EATR system was composed of MACR-145 cells stably expressing mCherry-eGFP-RAMP4 reporter (ER-colocalized RAMP4 fused with mCherry and eGFP) ( Figure 2A ). ER-phagy has been reported to deliver ER fragments and mCherry-eGFP-RAMP4 reporter into the acidic environment of the lysosomes, which deprives mCherry-eGFP-RAMP4 reporter of eGFP fluorescence but maintain its mCherry fluorescence (36). The flow cytometry assay showed that upon PRRSV infection, the number of cells with the acidified ER was significantly declined in the PRRSV-susceptible EATR system, but this number was significantly increased after treatment with thapsigargin (an ER-phagy inducer) ( Figure 2B ). In addition, the number of ER puncta with merely mCherry fluorescence was decreased in MARC-145 cells stably expressing mCherry-eGFP-RAMP4 reporter during PRRSV infection ( Figure 2C ). Furthermore, PRRSV-susceptible mCherry-Cleaved ER-phagy Reporter (CCER) system was also employed to investigate the effect of PRRSV infection on ER-phagy ( Figure 2D ). In CCER system, mCherry-RAMP4 will be cleaved into free mCherry in acidic environment of the lysosome (36). PRRSV-susceptible CCER system was constructed by transfecting mCherry-RAMP4 vector into MARC-145 cells. The immunoblot assay showed that the expression of free mCherry was significantly downregulated after PRRSV infection ( Figure 2E ). In addition, both PRRSV infection and BafA1 treatment promoted the accumulation of LC3-II, indicating that PRRSV could partially inhibit the fusion of autophagosomes and lysosomes ( Figure 2E and Figure S1 ). These results suggested that PRRSV could subvert ER-phagy and change ER morphology in the infected cells.

Considering that FAM134B is the best-characterized ER-phagy receptor regulating ER turnover, we hypothesized that PRRSV might disrupt ER-phagy by suppressing FAM134B expression. To test this hypothesis, we investigated the expression of FAM134B protein in PAMs isolated from pigs at 7 or 14 days post PRRSV infection (dpi) in vivo. The results showed that the expression level of FAM134B protein was significantly decreased in the PRRSV-infected cells ( Figure 3A ). Consistent with the FAM134B expression pattern in vivo, PRRSV infection significantly downregulated the expression of this protein in PAMs ( Figure 3B ), PK15^CD163 ( Figure 3C ), and MARC-145 cells ( Figure 3D ) in vitro at 36 hpi or 48 hpi. Additionally, PRRSV infection downregulated the expression of FAM134B protein in a dose-dependent manner ( Figures 3B–D ). Porcine FAM134B protein is homologous to human and mouse FAM134B containing typical RHD and LIR domain ( Figure 3E ). Porcine FAM134B was co-localized with ER markers CKAP4 and SEC61B ( Figure S2A ), and we observed starvation-induced degradation of FAM134B, CKAP4, and RTN4 ( Figure S2B ), indicating that FAM134B might be a major factor responsible for ER-phagy. The transient overexpression results of GFP-FAM134B and mCherry-LC3B in PK15 cells showed that mCherry and GFP puncta were co-localized. However, mutation of porcine FAM134B LIR motif abolished its colocalization with mCherry-LC3B in PK15 cells ( Figure 3F ), which was in line with the previous report in mouse model (14).

PRRSV infection significantly downregulated FAM134B protein expression in vivo and in vitro. The miRNA-mediated gene regulation has been reported to be closely related to PRRSV pathogenesis (39). Our prediction by Targetscan software indicated that FAM134B gene was the best target of miR-142-5p, and that the binding sites of miR-142-5p to 3’ untranslated regions (UTR) of FAM134B were conserved across different species ( Figure 4A ). Our previous transcriptomic data have shown that the expression of miR-142-5p was significantly upregulated in PAMs in response to PRRSV infection (40). The expression of miR-142-5p was also upregulated in PRRSV-infected PK15^CD163 cells and MARC-145 cells in vitro ( Figures 4B, C ). Dual-luciferase reporter assays confirmed that FAM134B gene was directly targeted by miR-142-5p. The luciferase activity of FAM134B-3’UTR reporter was significantly decreased when this reporter was co-transfected with miR-142-5p mimics into PK15 cells, while that was enhanced when this reporter was co-transfected with miR-142-5p inhibitor. The luciferase activity of FAM134B-3’UTR-MUT reporter remained unchanged when this mutant reporter was co-transfected with miR-142-5p mimics or inhibitor into PK15 cells ( Figure 4D ). The immunoblot assay showed that miR-142-5p mimics inhibited the expression of FAM134B in PK15 cells, whereas miR-142-5p inhibitor promoted it ( Figure 4E ). Furthermore, miR-142-5p had no effect on mRNA expression of FAM134B ( Figure 4F ).

PRRSV replication and assembly have been reported to be closely associated with the ER (7). Considering this, we speculated that FAM134B-dependent ER-phagy might inhibit PRRSV replication. To test this speculation, FAM134B was transiently overexpressed in PRRSV-infected PK15^CD163 cells to rescue PRRSV-induced downregulation of FAM134B. Additionally, the porcine FAM134B gene-targeting siRNA was transfected into PK15^CD163 cells to enhance the downregulation of FAM134B. The results showed that the expressions of PRRSV N protein and ORF7 as well as the titers of virus were significantly decreased in FAM134B-overexpressed PK15^CD163 cells ( Figures 5A, B and Figure S3B ), whereas they were significantly increased in FAM134B-silenced PK15^CD163 cells ( Figures 5A, B and Figures S3A–D ). In addition, overexpression of FAM134B-LIR-MUT protein failed to inhibit PRRSV replication, indicating that FAM134B could inhibit PRRSV replication by enhancing ER-phagy ( Figures 5C, D and Figures S3E, F ). Furthermore, the immunoblot assay showed that miR-142-5p promoted PRRSV replication ( Figure 5E ). Overexpression and silencing of FAM134B blunted or enhanced the effect of miR-142-5p on PRRSV replication, respectively ( Figure 5E ). Further, we tested the virulence of PRRSV produced by these cells, and found that TCID₅₀ was negatively correlated with FAM134B, but positively with miR-142-5p ( Figures 5F ). In conclusion, all above data demonstrated that miR-142-5p promoted PRRSV replication by inhibiting FAM134B expression.

To reveal the molecular mechanism by which FAM134B as a host cell restriction factor inhibited PPRSV replication, we conducted transcriptomics analysis of PK15^CD163 cells transfected with pcDNA3.1-FAM134B or control vector and infected with PRRSV ( Figure S4A ). A total of 230 differentially expressed genes (DEGs) were identified, of which 115 DEGs were upregulated and 105 were downregulated (Fold change >1.5 and p<0.05) ( Figure S4B ). Among 115 upregulated DEGs, the expressions of numerous interferon-stimulated genes (ISGs) such as IFIT1, IFIT2, IFIT3, ISG15, ISG12(A), OAS1, OAS2, and MX1 were significantly upregulated ( Figure 6A ). Furthermore, GO and KEGG pathway analyses showed that most DEGs were enriched in antiviral- and immune-related pathways ( Figure S4C ), suggesting that the overexpression of FAM134B activated type I interferon signaling pathway to inhibit PRRSV replication. Dual luciferase assay confirmed that overexpression of FAM134B in PRRSV-infected PK15^CD163 cells activated IFNβ luciferase activity ( Figures 6B, C ). Consistently, quantitative PCR confirmed that the expressions of ISG15, IFIT1, IFIT2, and IFIT3 were significantly increased in PRRSV-infected PK15^CD163 cells transfected with FAM134B-overexpressed vector, and that the expressions of these ISGs were inhibited after the treatment with the autophagy inhibitor 3-methyladenine (3-MA) ( Figure 6D ). In addition, the treatments with 3-MA or siRNA targeting autophagy-related 7 (ATG7) inhibited the expressions of type I interferon signaling in PRRSV-infected PK15^CD163 cells transfected with FAM134B-overexpressed vector ( Figure 6B and Figure S4D ), indicating that autophagy-related processes could regulate type I interferon production. The phosphorylation of IFN regulatory factor 3 (IRF3) plays a critical role in activating type I interferon signaling pathway. In this study, we found that the overexpression of FAM134B activated IRF3 phosphorylation in PRRSV-infected PK15^CD163 cells, whereas siRNA-mediated depletion of FAM134B inhibited its phosphorylation ( Figure 6E ). Taken together, these results suggested that FAM134B could inhibit PRRSV replication by enhancing type I interferon signaling pathway. Autophagy mechanism regulates downstream TLR signaling and innate immune response (41). Further, we investigated whether FAM134B-mediated ER-phagy regulated host innate immune response by activating Toll-like 3 receptor (TLR3) and retinoic acid-inducible gene I (RIG-I) signaling pathways. The results showed that upon PRRSV infection, FAM134B could not activate host immune response after silencing expression of TLR3 or RIG-I ( Figures 6C, D and Figure S4D ). In addition, the expression level of IFNβ in the group with both TLR3 and RIG-I silenced was significantly lower than that in the group with TLR3 or RIG-I silenced alone ( Figure 6C ). Overall, these finding suggested that FAM134B-mediated ER-phagy could promote PRRSV recognition by TLR3 and RIG-I to activate type I interferon signaling, eventually inhibiting PRRSV replication.