EBV-BILF1 and the PLHV1-3 BILF1 receptors with an N-terminal FLAG-tag were cloned into the pcDNA3.1+ vector. Gα_Δ6qi4myr recombinant G protein was kindly provided by Evi Kostenis (Institute for Pharmaceutical Biology, University of Bonn, Germany).

Human embryonic kidney 293 (HEK-293) cells were cultured at 37°C and 10% CO₂ in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Porcine kidney 15 (PK-15) cells were cultured at 37°C and 5% CO₂ in minimum essential medium (MEM; Invitrogen) containing 10% heat inactivated FBS and 1% penicillin-streptomycin. CRISPR/Cas9 modified HEK-293A pan knock-out (KO) cells (ΔGs/olf/q/11/12/13/z) (54) and HEK-293A parental cells were kindly provided by Asuka Inoue (Tohoku University, Japan). They were cultured at 37°C and 5% CO₂ in DMEM supplemented with 10% heat inactivated FBS, 1% penicillin-streptomycin and 1% L-glutamine. Cells were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s recommendation. For immunohistochemistry, cells were transfected using Fugene 6 (HEK-293) or FuGeneHD (PK-15) transfection reagent (Promega) according to the manuacturer’s instructions.

Pig tissue samples were obtained from earlier published in vivo experiments that have been described in detail (26, 32, 55, 56). These animals were derived from partially inbred MHC-defined miniature swine herd (Massachusetts General Hospital) and all received 1000cGy thymic irradiation (TI) and 0.05mg/kg pCD3-CRM9 two days prior to cytokine mobilized peripheral blood cell transplantation (57). Oral cyclosporine (Neoral) was started at a dose of 15mg/kg one day before transplantation and was continued BID for 30-60 days (or until death). Samples from three pigs were collected at the time of the transplantation and after the onset of PTLD disease, whereas samples from two PTLD diseased pigs were collected only after disease onset. Samples from one pig without signs of disease were used as controls.

HEK-293 and PK-15 cells were seeded in Poly-D-lysine pre-coated 96-well plates and after 24 hours transiently transfected with various concentrations of receptor DNA (0, 1, 2, 5, 10, 25, 35, 50ng) using Lipofectamine 2000. In order to achieve comparable transfection efficiency in both cell types, we increased the amount of DNA in PK-15 cells. Twenty-four hours after transfection, cells were fixed with 3.7% paraformaldehyde in PBS/CaCl₂ (pH 7.3) for 10 minutes. After three washes in PBS, cells were blocked with 2% bovine serum albumin (BSA) in PBS/CaCl₂ for 30 minutes at RT and subsequently incubated with primary mouse M1 anti-FLAG antibody (Sigma-Aldrich) at 1:2250 in 1% BSA/PBS/CaCl₂ for 1 hour at RT. Following three washes in PBS/CaCl₂, cells were incubated with secondary goat anti-mouse horseradish peroxidase-conjugated IgG antibody (Sigma-Aldrich) at 1:1000 in PBS/CaCl₂ for 1 hour at RT. Peroxidase activity was determined by addition of 3,3’-5,5’-tetramethyl benzidine substrate (TMB) (Sigma-Aldrich) for 5 minutes. The reaction was terminated by addition of 0.2M H₂SO₄ and absorbance was measured at 450 nm using FlexStation3^® Benchtop Multi-Mode Microplate Reader (Molecular devices).

HEK-293 and PK-15 cells were seeded on fibronectin coated (10µg/mL) coverslips in 24-well plates. The next day cells were transiently transfected using FuGene (HEK-293) or FugeneHD (PK-15). Twenty-four hours after transfection, cells were washed with PBS and fixed in 3.7% paraformaldehyde for 10 minutes on ice and additional 10 minutes at RT. Following three washing steps, cells were permeabilized using 0.02% saponin in 1% donkey serum/PBS and additionally blocked with 10% donkey serum in PBS for 20 minutes. For cell surface receptor visualization, cells were incubated with primary mouse M1 anti-FLAG antibody (Sigma-Aldrich) at 1:2250 in 1% donkey serum/PBS/CaCl₂ for 1 hour at RT. Following three washes with PBS/CaCl₂, cells were incubated with secondary donkey anti-mouse Alexa Fluor 594 antibody at 1:100 in PBS for 1 hour at RT. For the last 10 minutes of the incubation, wheat germ agglutinin (WGA) conjugated to Alexa488 (Invitrogen) was added to the cells as a membrane marker (5μg/mL). Cells were additionally incubated with PBS containing Hoechst 33342 stain (Invitrogen) (1μg/mL) and samples were mounted with 8μL of Fluorescence mounting medium (Dako) before imaging on a fluorescence microscope (LSM700).

Luciferase assays were performed on HEK-293 cells or CRISPR/Cas9 genetically engineered HEK-293A cells depleted of different G proteins and parental HEK-293A cells. Cells were seeded at 35,000 cells per well on white 96-well plates pre-treated with poly-D-lysine. The following day, cells were transiently transfected with receptor constructs at indicated concentrations and co-transfected with transcription reporter plasmids CRE-Luc, NFAT-Luc or NF-κB-Luc at 30ng/well using Lipofectamine 2000 (Invitrogen). Additional co-transfection with Gα_Δ6qi4myr, which is recognized by GPCRs as a Gα_i protein but elicits Gα_q-dependent signaling (58) was performed in CRE and co-transfection with Gα_q or Gα₁₁ was performed in NFAT luciferase assay. After 24 hours, cells were washed with PBS and incubated for 30 minutes with a mixture of SteadyLite (50μL/well, PerkinElmer) and PBS (50μL/well). Plates were read on an EnVision Multilabel Plate Reader (PerkinElmer) using the luciferase program. The experiment was performed three times in triplicates.

To determine the internalization properties of BILF1 orthologues, we performed two separate assays, both relying on the principal of antibody uptake over time. With time-course cell-based ELISA assay, we quantitatively assessed the amount of surface expressed receptors over time. Using microscopy approach, surface expressed and internalized receptors, can be visualized in different time points after induction of internalization.

HEK-293 cells were seeded in a 6-well plate and transfected using FuGeneHD reagent. Fourty-eight hours after transfection, 5x10⁵ cells/sample were transferred to a tube and stained using APC-labelled anti human HLA-A,B,C (W6/32) antibody at the recommended concentration (Biolegend). Receptors were detected using anti-FLAG FITC conjugated antibodies (Genscript) and isotype controls were used to differentiate non-specific background signal from specific antibody signals. Samples were analysed on a BD FACSCanto™ II instrument equipped with 488 nm and 633 nm lasers using FITC (530/30) and APC (660/20) filters. Data were analysed using Kaluza software.

For analysis of MHC-I downregulation, cells were seeded, transfected and fixed as described above (paragraph 2.5). Afterwards, cells were incubated with primary rabbit anti-FLAG antibody (Sigma-Aldrich) at 1:300 and mouse anti human HLA-A, B, C class I antibody (1:100) in HEK-293 cells (Santa Cruz) or mouse anti-pig SLA class I antibody (1:100) for PK-15 cells (R&D systems) for 1 hour at RT. Following three washes with PBS/CaCl₂, cells were incubated with secondary goat anti-rabbit Alexa Fluor 647 antibody and goat anti-mouse Alexa Fluor 555 antibody at 1:500 in PBS for 1 hour at RT. For the last 10 minutes of the incubation, WGA conjugated to Alexa488 (Invitrogen) was added to the cells as a membrane marker (5μg/mL). Following two wash steps, cells were incubated with PBS containing Hoechst 33342 stain (Invitrogen) (1μg/mL). Before imaging on a fluorescence microscope (LSM700) cells were mounted with 8μL of Fluorescence mounting medium (Dako). Equivalent numbers of transfected and non-transfected cells were blindly selected and we measured the intensity of 555 channel (MHC-I) using FIJI software.

Cell viability assay was performed as described in the manufacturer’s manual. Briefly, CRISPR/Cas9 genetically engineered HEK-293A cells depleted of different G proteins and parental HEK-293A cells were seeded at 35,000 cells per well and transfected the next day. 24 hours after transfection, plates were incubated at RT for approximately 30 minutes. 100 µL of CellTiter Glo reagent was carefully added to each well. Plates were incubated for 12 minutes on shaker before the luminescence was measured on EnVision Multilabel Plate Reader (PerkinElmer).

Sequencing libraries were prepared from DNA extracted from four PTLD diseased pig samples using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs), according to manufacturer’s instructions. Libraries were subsequently multiplexed and sequenced using a 300 cycle mid-output kit (2x150 paired-end mode) on a NextSeq 550 (Illumina Technologies).

Prior to competitive alignment, sequence read datasets were quality and adapter trimmed using TrimGalore (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). For the competitive alignment step, a hybrid reference genome was constructed by merging FASTA files containing the Sus scrofa (Sscrofa11.1) genome and partial genome sequences for PLHV1 (NC_038264.1), PLHV2 (NC_038265.1), and PLHV3 (AY170316). Sequence read datasets were aligned against the hybrid genome using bbmap (https://sourceforge.net/projects/bbmap/) with minid=0.9 and ambiguous=random before post-processing with SAMtools (59) and BEDtools (60). Coverage plots were generated using the R 4.0 bioconductor packages Gviz (61) and GenomicFeatures (62).

RNA was extracted from samples using QIAzol^® lysis reagent (Qiagen). RNA (500ng) was used to prepare cDNA using Superscript III reverse transcriptase (Thermo Fischer Scientific). RT-qPCR was performed on 96-well plates using TaqMan^® Fast Advanced Master mix (2x) (Thermo Fischer Scientific) and costumed TaqMan^® gene expression assay probes (Thermo Fischer Scientific), designed for specific detection of PLHV1-BILF1 (forward primer: 5’TCTGATACTCACTGTTGCTAACTTTGTT3’, reverse primer: 5’ACTTATAGCTTGGCGACACTTGAA3’, probe: 5’FAM-ACCTTCAAAGCTCAAAGTAGT-MGB3’), PLHV2-BILF1 (forward primer: 5’GCTGTTGCTAACTTTCTTAvGTTTTGGA3’, reverse primer: 5’ACTTATACCTTGGCGACACTTGAAG3’, probe: 5’FAM-CCTTCAAAGCTCAATACAGC-MGB3’), PLHV3-BILF1 (forward primer: 5’GTCTTTTGGCAGTTGTTGCTAATCT3’, reverse primer: 5’ACTTGTAGCCTGGCGACATT3’, probe: 5’FAM-CAAGTGCAGCATAATTC-MGB3’) and a reference gene RPL4 (assay identification number: Ss03374063_g1). RT-qPCR was performed on a QuantStudio 6 Flex Real-Time PCR system (Thermo Fisher Scientific) using a standard protocol. Ct (cycle threshold) values were acquired using QuantStudio Real-Time PCR software. To present the relative fold change of cDNA levels compared to a reference gene (RPL4) we used 2^ΔCt calculation.

Compared PDB entries: 3BO8 (human) (https://www.rcsb.org/structure/3bo8) & 5NPZ (porcine) (https://www.rcsb.org/structure/5NPZ).

Data were analysed using Graph Pad Prism (8.3.0), Kaluza analysis software and FIJI software and reported as mean ± SEM (standard error of the mean). For all microscopy experiments, images were visualized using an LSM700 microscope and ZEN blue software. Statistical analysis was performed with Graph Pad Prism using ANOVA or student t-test as indicated. P value < 0.05 was considered statistically significant.

EBV-BILF1 is known to localize primarily to the cell surface, which is based on results from a number of studies performed in a variety of cell types [e.g. human embryonic kidney (HEK-293) cells and human melanoma (MJS) cells] (41, 44, 46, 49, 51). To examine the subcellular localization of the three uncharacterized PLHV1-3 BILFs in comparison to EBV-BILF1 in the present study, we cloned BILFs into eukaryotic expression plasmids followed by transient expression in HEK-293 and porcine kidney epithelial (PK-15) cells. Localization was then analysed using both a quantitative cell-based ELISA and fluorescence microscopy. The cell-based ELISA using increasing concentrations of EBV-BILF1 and PLHV1-3 BILF1 expression plasmids, showed cell surface localization of all BILFs in both HEK-293 ( Figure 1A ) and PK-15 cells ( Figure 1B ). To achieve a comparable dose response in both cell lines, higher amounts of plasmid DNA for each BILF1 orthologue was required for transfection in PK-15 cells ( Supplementary Figure 1 ). PLHV1-3 BILF1 showed ~20-40% of surface expression levels compared to EBV-BILF1 in HEK-293 cells, and ~60-70% (PLHV1- and 2-BILF1) and ~20% (PLHV3-BILF1) of EBV-BILF1 levels in PK-15 cells. Subcellular distribution determined by fluorescence microscopy using WGA as a cell surface (plasma membrane) marker was consistent with the cell-based ELISA results, as BILFs co-localized with WGA at the cell surface of HEK-293 ( Figure 1C ) and PK-15 cells ( Figure 1D ). In summary, analysis of subcellular distribution shows that similar to EBV-BILF1, PLHV1-3 BILFs are localized at the plasma membrane, and this appears to be independent of cell type.

Next, we focused on the membrane proximal signaling pattern of PLHV1-3 BILFs in comparison to the previously described Gα_i-dependent signaling of EBV-BILF1 (41, 50). For these studies we used HEK-293 cells, which is a cell-type commonly used for studying signaling properties of GPCRs (63) also based on their expression of full repertoire of G protein subunits (64). We first investigated receptor-mediated regulation on the activity of the downstream cyclic adenosine monophosphate (cAMP) response element (CRE) transcription factor ( Figure 2 ), which responds to receptor activity via Gα_s (increased activity) and Gα_i (decreased activity) (65). Cells were pre-treated with forskolin, an inducer of cAMP formation, which enables measurement of Gα_i-coupled receptor activity ( Figures 2A, B ). Similar to EBV-BILF1, PLHV1-3 BILFs inhibited forskolin-induced CRE transcription factor activity in a dose dependent manner ( Figure 2A ). Among PLHV BILFs, PLHV2-BILF1 showed a moderate (30%) decrease in forskolin induced CRE transcription factor activity in comparison with BILF1 from PLHV1 and PLHV3, where the level of inhibition was approximately 2-fold higher (approx. 60% decrease) and comparable to EBV-BILF1 levels. In the absence of forskolin, receptor-mediated CRE transcription factor activity was not detected for any BILF1 orthologue, excluding Gα_s dependent activation. To support these findings, we co-transfected HEK-293 cells with the different BILFs together with the chimeric G protein Gα_Δ6qi4myr ( Figures 2C, D ). Activation of this chimeric protein leads to increased CRE activity through phospholipase C (PLC). Using this approach, BILF1 from both PLHV1 and PLHV2 increased CRE activity in a gene dose-dependent manner to a similar level observed for EBV-BILF1; PLHV3-BILF1 showed lower (~46%) activity. Together, our analysis showing inhibition of both forskolin-dependent CRE transcription factor activity and increased CRE activity following co-transfection with Gα_Δ6qi4myr indicate that constitutive membrane proximal signaling via Gα_i is generally conserved across PLHV1-3 BILFs similar to EBV-BILF1 (41, 50).

In addition to CRE, EBV-BILF1 activates more distally located downstream NF-κB and NFAT transcription factor pathways (45, 46, 50, 51). To test the capacity of PLHV1-3 BILFs for activation of these additional signaling pathways, we first examined constitutive activation of NF-κB and NFAT in HEK-293A (parental) cells. Figure 3 and Supplementary Figure 2 show the activity of PLHV1-3 BILF1 receptors relative to the maximal NF-κB or NFAT activity elicited by EBV-BILF1 in parental cells (designated as 100% activity). EBV-BILF1 activated both NF-κB and NFAT transcription factors. BILF1 from PLHV1 and PLHV2 showed ~130% and ~240% NF-κB activity, respectively ( Figure 3A and Supplementary Figure 2A ) compared to EBV-BILF1 and either low or no NFAT activity, respectively. PLHV3-BILF1 failed to activate NF-κB, but showed comparable NFAT activity (~120%) to EBV-BILF1 in these cells ( Figure 3B and Supplementary Figure 2B ). In summary, these results from parental cells show that only PLHV1- and PLHV2-BILF1 can mediate NF-κB activity, whereas NFAT activity is only induced by PLHV3-BILF1 in comparison to EBV-BILF1, which is able to activate both transcription factors.

To determine the dependence of the observed NF-κB and NFAT activity on Gα_i, we then used HEK-293A cells genetically engineered by using CRISPR/Cas9 to express only Gα_i proteins (ΔG_{s/olf/q/11/12/13/z}; “pan KO” cells). As shown in Figure 3A , BILF1 from EBV, PLHV1 and PLHV2 retained NF-κB activity in pan KO cells (50%, 50% and 80%, respectively, compared to EBV-BILF1 activity in parental cells), showing important but not exclusive Gα_i mediated NF-κB activation for these receptors. Again, comparable to its behaviour in parental cells, PLHV3-BILF1 did not show any significant activity in pan KO cells ( Figure 3A and Supplementary Figure 2A ). NFAT activity of EBV and PLHV3 -BILF1 ( Figure 3B and Supplementary Figure 2B ) was conserved in pan KO cells (100% and ~75%, respectively) implicating Gα_i as the primary G protein involved in NFAT signaling through these receptors. Importantly, cell surface expression of BILFs was approximately 20% lower in pan KO cells and comparable to our results shown on Figure 1 with EBV-BILF1 showing highest cell surface expression ( Supplementary Figure 3 ). Surprisingly, in pan KO cells PLHV1-BILF1 showed ~30% activity, whereas in parental cells it did not show any activity. Consistent with NFAT signaling in parental cells, PLHV2-BILF1 did not activate NFAT ( Figure 3B and Supplementary Figure 2B ). In summary, only PLHV3-BILF1 showed NFAT activity comparable to EBV-BILF1.

Together, results from these studies show that Gα_i-coupling is conserved between all BILFs at membrane proximal levels, but differs for the activation of more distal levels of NF-κB and NFAT transcription factors.

Constitutive internalization of EBV-BILF1 has previously been proposed both as a means to function as an immunoevasin downregulating MHC-I molecules, and as an important regulatory mechanism to enable constitutive signaling during the virus replication cycle (44). We next determined the internalization of PLHV1-3 BILFs by using antibody feeding experiments in combination with fluorescence microscopy and cell-based ELISA. Microscopy analysis confirmed that all BILFs were located at the cell surface at the beginning of the experiment (t=0). Additionally, the constitutively internalized fraction of the receptors were shown to localize to a comparable perinuclear endocytic vesicular site after 30 minutes in both HEK-293 and PK-15 cells ( Figures 4A, B ). Time-dependent receptor internalization was analysed by cell-based ELISA in HEK-293 cells where BILF1 surface expression was measured following incubation at 37°C for 0, 5, 10, 20 and 30 minutes ( Figure 4C ). Internalization of PLHV1 and PLHV2 BILF1 receptors showed comparable rates of internalization as EBV-BILF1, with ~20% of these receptors being internalized after 30 minutes. Internalization of PLHV3-BILF1 was slower with only 8% being internalized over the same time period. Thus, in terms of rate and total level of internalization as well as subcellular distribution following internalization, BILFs from PLHV1 and 2 are similar to the constitutively internalizing EBV-BILF1 receptor.

Given the suggested link between constitutive internalization and immunoevasive properties shown for EBV-BILF1 (45, 46) and the observed conserved levels of constitutive internalization for at least two of the PLHV BILF1 receptors, we next determined their immunoevasive properties. Initially, we performed FACS analysis, measuring the expression of endogenous MHC-I molecules at the surface of BILF1 transfected HEK-293 cells in comparison to control (pcDNA) transfected HEK-293 cells as described previously for EBV-BILF1 (44). Our decision to use human HEK293 cells, was supported by comparing the sequence relationship between human and porcine MHC-I with a homology model ( Figure 5A ) showing a 74% aa sequence homology and high structure identity. Our FACS results showed a comparable downregulation of MHC-I in EBV-BILF1 transfected cells as had been previously observed (44) ( Supplementary Figure 4 ). Ratio between BILF1 transfected and non-transfected cells showed significantly lower MHC-I expression (0,61 and 0,67) for EBV-BILF1 and PLHV3-BILF1 respectively, whereas PLHV1- and PLHV2-BILF1 showed a tendency towards MHC-I downregulation (0,84 and 0,83 respectively) ( Supplementary Figure 4 ). Considering low transfection efficiency observed for all PLHV BILFs which persisted despite optimization using higher concentrations of DNA, different transfection protocols or different vector systems (pcDNA and bicistronic IRES vector) we looked at the single cell MHC-I expression using a new methodology based on fluorescence microscopy ( Figures 5B, C ). This approach allowed us to determine both the receptor expression and to evaluate MHC-I surface expression on transfected (coloured bars) compared to non-transfected (grey bars) cells. Notably, this approach was not limited by transfection efficiency enabling analysis in both human (HEK-293) and porcine (PK15) cells. Although not definitive, the level of homology conservation for human and porcine MHC-I molecules shown ( Figure 5A ), supports the possibility for cross-species MHC-I recognition. As shown in Figure 5 , all BILF1 receptors significantly reduced surface MHC-I expression in HEK-293 cells. EBV-BILF1 and PLHV3-BILF1 showed approximately 50% downregulation of MHC-I molecules, whereas PLHV1-BILF1 and PLHV2-BILF1 showed 30-35% downregulation of surface MHC-I molecules compared to non-transfected cells ( Figure 5B ). Importantly, FACS data indicated the conserved MHC-I downregulation for PLHV BILF1 receptors, which was further confirmed by microscopy studies in HEK-293 cells. In contrast to HEK-293 cells, in PK-15 cells all BILFs including EBV-BILF1 showed higher expression of MHC-I molecules at the cell surface in transfected cells in comparison to non-transfected cells ( Figure 5C ). This is in contrast to what has been published for EBV-BILF1 and our observations for EBV- and PLHV1-3 BILF1 in HEK-293 cells, therefore we conclude that downregulation of MHC-I cell surface expression mediated by BILF1 is cell type dependent.

Previously, Huang et al. (2001) reported the development of PTLD in miniature pigs undergoing SOT or HSCT transplantation and immunosuppression, which was associated with PLHV1 infection (26). Notably, histological and pathological properties of PTLD in these pigs resembled the disease in humans, with B-cell lymphoproliferation in peripheral blood, lymph nodes and tonsils. We therefore examined PLHV1-3 expression in samples obtained from four miniature pigs with PTLD by using Illumina sequencing. Further supporting the association of PLHV1 with PTLD, as only PLHV1 was detected in the diseased pigs ( Figures 6A, B ). By contrast, neither PLHV2 nor PLHV3 genomes were detected in any of the sequence data for these samples ( Figure 6B ). Next, to examine the association of PLHV1-BILF1 expression with PTLD in these pigs, we designed specific probes and performed PLHV1-BILF1 qPCR. Notably, in five of five samples collected after the PTLD onset, we detected high expression levels of PLHV1-BILF1 ( Figure 6C ). Importantly, PLHV1-BILF1 was not detected in any of three pre-transplant samples from these same animals that developed disease, nor in a sample from a pig lacking PTLD symptoms, showing specific link between PLHV1 infection and PTLD development. BILF1 from PLHV2 or PLHV3 were not detected in any of the pigs ( Figure 6C ).