Equine Herpesvirus 1 Multiply Inserted Transmembrane Protein pUL43 Cooperates with pUL56 in Downregulation of Cell Surface Major Histocompatibility Complex Class I

ABSTRACT

Herpesviruses have evolved an array of strategies to counteract antigen presentation by major histocompatibility complex class I (MHC-I). Previously, we identified pUL56 of equine herpesvirus 1 (EHV-1) as one major determinant of the downregulation of cell surface MHC-I (G. Ma, S. Feineis, N. Osterrieder, and G. R. Van de Walle, J. Virol. 86:3554–3563, 2012, http://dx.doi.org/10.1128/JVI.06994-11; T. Huang, M. J. Lehmann, A. Said, G. Ma, and N. Osterrieder, J. Virol. 88:12802–12815, 2014, http://dx.doi.org/10.1128/JVI.02079-14). Since pUL56 was able to exert its function only in the context of virus infection, we hypothesized that pUL56 cooperates with another viral protein. Here, we generated and screened a series of EHV-1 single-gene deletion mutants and found that the pUL43 orthologue was required for downregulation of cell surface MHC-I expression at the same time of infection as when pUL56 exerts its function. We demonstrate that the absence of pUL43 was not deleterious to virus growth and that expression of pUL43 was detectable from 2 h postinfection (p.i.) but decreased after 8 h p.i. due to lysosomal degradation. pUL43 localized within Golgi vesicles and required a unique hydrophilic N-terminal domain to function properly. Finally, coexpression of pUL43 and pUL56 in transfected cells reduced the cell surface expression of MHC-I. This process was dependent on PPxY motifs present in pUL56, suggesting that late domains are required for pUL43- and pUL56-dependent sorting of MHC class I for lysosomal degradation.

IMPORTANCE We describe here that the poorly characterized herpesviral protein pUL43 is involved in downregulation of cell surface MHC-I. pUL43 is an early protein and degraded in lysosomes. pUL43 resides in the Golgi vesicles and needs an intact N terminus to induce MHC-I downregulation in infected cells. Importantly, pUL43 and pUL56 cooperate to reduce MHC-I expression on the surface of transfected cells. Our results suggest a model for MHC-I downregulation in which late domains in pUL56 are required for the rerouting of vesicles containing MHC-I, pUL56, and pUL43 to the lysosomal compartment.

INTRODUCTION

The interplay between viruses and their hosts has led to the evolution of a number of strategies that facilitate evasion from the recognition and clearance of virus infection by the host immune system. Upon successful entry into the cell, viruses are uncoated. Structural components of the invading virus as well as newly produced proteins are polyubiquitinated and then fragmented into peptides by the proteasome (1). The processed antigenic peptides are transported into the endoplasmic reticulum (ER) and presented by major histocompatibility complex class I (MHC-I) molecules on the cell surface. Cytotoxic CD8⁺ T lymphocytes (CTL), whose T-cell receptor (TCR) specifically recognizes small peptides bound in the groove of MHC-I, ultimately eliminate the infected cell (2, 3). However, CTL-mediated immunity may fail or be delayed, because many viruses encode specific inhibitors that interfere with various stages of MHC-I antigen presentation (4). As a consequence, infected cells have reduced MHC-I expression and become less sensitive to patrolling CTL.

Equine herpesvirus 1 (EHV-1) is an important veterinary pathogen that poses a severe risk to the health of horse populations around the world. EHV-1 infection results in various clinical syndromes involving upper respiratory ailments, miscarriage, death of neonates, and neurological disease (5). Classified as a member of the Alphaherpesvirinae subfamily, EHV-1 is a double-stranded DNA virus featuring a large genome of ∼150 kbp. The EHV-1 genome contains at least 80 open reading frames (ORFs), of which at least 4 ORFs are duplicated in the inverted-repeat regions (6). Historically, the EHV-1 genome has been annotated in accordance with those of herpes simplex virus 1 (HSV-1) and varicella-zoster virus (VZV), prototype viruses of the Alphaherpesvirinae. This approach has also been applied to other closely related viruses, e.g., EHV-4 and pseudorabies virus (PRV) (7, 8). Hence, the role of a particular EHV-1 gene product can be deduced on the basis of its HSV-1 or VZV counterpart and extended to predict the function of the orthologues that are conserved in the genus, subfamily, or even family. Nevertheless, several genes and/or gene functions are unique to HSV-1, VZV, or EHV-1. For instance, HSV-1 ICP47 was the first protein identified in the Alphaherpesvirinae to induce downregulation of MHC-I at the cell surface by directly interacting with the transporter associated with antigen processing (TAP). ICP47 irreversibly prevents the transport of cytoplasmic peptides into the ER (9), but ICP47 homologues are absent in EHV-1 and other varicelloviruses, including VZV. EHV-1 also causes MHC-I downregulation in a species-independent fashion, and the pUL49.5 and pUL56 proteins have been shown to modulate cell surface MHC-I expression (10, 11); however, pUL49.5 and pUL56 of HSV-1 do not affect MHC-I levels (12).

The pUL56 and pUL49.5 homologues of various members of the Alphaherpesvirinae differ in their expression patterns and subcellular localizations (11, 13), indicating that they are mechanistically different. EHV-1 pUL49.5, a small type I transmembrane protein that interacts with viral glycoprotein M (gM) (14), inhibits the formation of peptide-loaded MHC-I molecules by preventing ATP binding to TAP (10). pUL56, a type II transmembrane protein, enhances the internalization of MHC-I through dynamin-dependent endocytosis (15). In our previous studies, we reported that pUL56 induced cell surface MHC-I reduction solely in the context of infection and that MHC-I levels were not completely restored when cells were infected with a mutant virus lacking both pUL49.5 and pUL56 (11). These findings gave rise to the hypothesis that another viral protein partners with pUL56 and participates in the downregulation of cell surface MHC-I.

The efforts of our work presented here focused on the identification of viral gene products that contribute to the downregulation of MHC-I in cells after EHV-1 infection. By screening a single-gene knockout library of EHV-1, the product of the HSV-1 UL43 homologue was identified to potently block surface MHC-I presentation in EHV-1-infected cells. Further characterization showed that pUL43 is nonessential for virus growth in vitro and is degraded in lysosomes at later times of infection. pUL43 also localizes to Golgi vesicles and requires its four C-terminal transmembrane (TM) domains for proper intracellular distribution. We also found that a unique hydrophilic domain of EHV-1 pUL43 is indispensable for reducing MHC-I levels. Finally, cotransfection of pUL43 and pUL56 resulted in robust inhibition of cell surface MHC-I expression. Taken together, these results suggest a novel mechanism by which alphaherpesviruses utilize a combination of viral transmembrane proteins to negatively regulate MHC-I antigen presentation and achieve immune evasion.

MATERIALS AND METHODS

Cells and viruses.

Equine dermal (NBL6) cells were propagated in Eagle's minimum essential medium (EMEM; Biochrom AG) supplemented with 20% fetal calf serum (FCS; Biochrom AG), 1% penicillin-streptomycin (100 U/ml penicillin and 100 μg/ml streptomycin; Sigma-Aldrich), 1 mM sodium pyruvate, and 1× nonessential amino acids (Biochrom AG). RK13 (rabbit kidney) cells, HeLa (human epithelial carcinoma) cells, and HEK293 (human embryonic kidney) cells were grown in EMEM containing 5% FCS and 1% penicillin-streptomycin. RK13 cells that constitutively express either the ER marker calreticulin or the Golgi marker β-1,4-galactosyltransferase fused to enhanced green fluorescent protein (EGFP) were generated by transfection of the plasmids pER-EGFP and pGolgi-EGFP, respectively, which were kindly provided by Michael Veit (Freie Universität Berlin, Berlin, Germany). The transfected cell lines were purified and maintained in the medium for RK13 cells supplemented with 500 μg/ml G418 disulfate salt (Sigma-Aldrich). The parental and mutant viruses were derived from EHV-1 strain Ab4, which was cloned as an infectious bacterial artificial chromosome (BAC). The BAC for Ab4 virus (pAb4) contains a mini-F cassette in which the egfp gene is driven by the human cytomegalovirus (HCMV) immediate early (IE) promoter (16). Viruses were reconstituted by transfection of BAC DNA into RK13 cells with polyethylenimine (PEI) (Polysciences), as previously described (11). To delete the egfp gene from the viral genome, RK13 cells expressing Cre were infected with the engineered virus at a multiplicity of infection (MOI) of 0.0001 (17). Nonfluorescent viral plaques were picked for purification on RK13 cells. Unless otherwise indicated, RK13 cells were used for virus propagation and titration.

Antibodies and reagents.

Rabbit anti-β-actin (13E5) monoclonal antibody (MAb) and rabbit and mouse anti-hemagglutinin (HA) tag MAbs were purchased from Cell Signaling Technologies. Rabbit polyclonal antibodies (PAbs) against EHV-1 pUL56 were prepared as described previously (11). Rabbit anti-EHV-1 IR6 PAb and anti-EHV-1 gM (F6) and anti-EHV-1 gC (2A2) MAbs were used as described in our previous studies (18–20). Mouse anti-MHC-I MAb specific for an equine haplotype (CZ3) was kindly provided by Douglas F. Antczak (Cornell University, Ithaca, NY). Mouse anti-HLA class I (W6/32) MAb was a gift from Hartmut Hengel (Universität Freiburg, Freiburg, Germany). Mouse anti-CD58 MAb was obtained from BioLegend. The mouse IgG isotype control was obtained from Santa Cruz Biotechnology. Alexa Fluor 647- or 568-labeled goat anti-mouse immunoglobulin G (IgG) and Alexa Fluor 488-labeled goat anti-rabbit and goat anti-mouse IgGs were produced by Invitrogen. Horseradish peroxidase (HRP)-conjugated goat anti-mouse and goat anti-rabbit IgGs were obtained from Southern Biotech. Phosphonoacetic acid (PAA), an inhibitor of viral DNA synthesis, was obtained from Alfa Aesar. Chloroquine and lactacystin were purchased from Sigma-Aldrich. Restriction enzymes were supplied by New England BioLabs.

Engineering of BAC mutants.

The pAb4 BAC was maintained in Escherichia coli GS1783 cells that were grown in Luria-Bertani (LB) broth in the presence of 30 μg/ml chloramphenicol (11). Genetic modification of the BAC was performed by en passant mutagenesis exactly as described previously (21). To start mutagenesis, a fragment flanked by homologous arms for the target region was PCR amplified by using a kanamycin resistance (Kan^r) gene present in plasmid pEP-Kan-S2, using the primers listed in Table 1. After gel electrophoresis, PCR products were purified and subjected to DpnI digestion to eliminate residual plasmid. GS1783 competent cells were electroporated with the PCR products and incubated at 32°C for 48 h. Kanamycin-resistant colonies were screened by restriction fragment length polymorphism (RFLP) analysis and compared to the predicted digestion pattern. Correct intermediates were used for a second round of Red-mediated recombination in the presence of 1% l-(+)-arabinose (Alfa Aesar) that induced the removal of the Kan^r gene sequence from the BAC construct. Following the resolution step, candidate colonies were examined by RFLP analysis and confirmed by DNA sequencing (LGC Sequencing Service). BAC DNA from picked colonies was prepared by standard alkaline lysis (22) and used for virus reconstitution.

TABLE 1.

List of primers for viral mutagenesis, plasmid construction, and DNA sequencing

Primer	Sequence (5′–3′)
BAC muatagenesis
43STOP_Fwa	CAAAGGTTGGCTTGCTACATCAAGGTTATCAATCATGATGTAACAGCCAGATAGAGAGCCCGGTAGGGATAACAGGGTAATCGAT
43STOP_Rv	GCACCAGACACGAGTCTTCACCGGGCTCTCTATCTGGCTGTTACATCATGATTGATAACCTTGCCAGTGTTACAACCAATTAACC
43Rev_Fwb	CAAAGGTTGGCTTGCTACATCAAGGTTATCAATCATGATGTACCAGCCAGATAGAGAGCCCGGTAGGGATAACAGGGTAATCGAT
43Rev_Rv	GCACCAGACACGAGTCTTCACCGGGCTCTCTATCTGGCTGGTACATCATGATTGATAACCTTGCCAGTGTTACAACCAATTAACC
HA43_Fwc	CAAAGGTTGGCTTGCTACATCAAGGTTATCAATCATGATGTACCCATACGACGTCCCAGACTACGCTTACCAGCCAGATAGAGAGCCCCAGTGTTACAACCAATTAACC
HA43_Rv	CCAGACACGAGTCTTCACCGGGCTCTCTATCTGGCTGGTAAGCGTAGTCTGGGACGTCGTATGGGTACATCATGATTGATAACCTTGTAGGGATAACAGGGTAATCGATT
HA_△N_43_Fw	ATCAATCATGATGTACCCATACGACGTCCCAGACTACGCTAAAGCTTTCGTTGGAATCGGTAGGGATAACAGGGTAATCGATT
HA_△N_43_Rv	TGAGGACGCAAGCTTGTAGTCCGATTCCAACGAAAGCTTTAGCGTAGTCTGGGACGTCGTCCAGTGTTACAACCAATTAACC
Expression vectors
pUL43s-EGFP
pUL43-EGFP_Fwc	GCGAATTCACCATGATGTACCCATACGACGTCCCAGACTACGCTTACCAGCCAGATAGAGAGCC
pUL43-EGFP_Rv	CGGGATCCATGTGTGATTATAGTTGCATAACAC
pUL43s-EGFP_Rv	CGGGATCCCGGCATCTCCTTGAAAAACTTGAAC
pUL43-UL56-P2A-EGFP
pHA43_Fwc	CGGGATCCACCATGATGTACCCATACGACGTCCCAGACTACGCTTACCAGCCAGATAGAGAGCC
pHA43_Rv	GCGAATTCTTTAATGTGTGATTATAGTTGCATAAC
P2A-EGFP_Fwd	CGGAATTCGGAAGCGGAGCTACTAACTTCAGCCTGCTGAAGCAGGCTGGAGACGTGGAGGAGAACCCTGGACCTATGGTGAGCAAGGGCGAGGAGC
P2A-EGFP_Rv	CGTCTAGATTACTTGTACAGCTCGTCCATG
Sequencing
UL43_Seq_Fw1	CACTTGTAGAAACACGCCCA
UL43_Seq_Fw2	CGTCATATGCTCAGCCAATG
UL43_Seq_Rv	AACATACCATGCACCAAAGG

^aLetters in boldface italic type indicate the stop codon introduced into the UL43 gene.

^bLetters in boldface italic type indicate the original codon of the UL43 gene.

^cUnderlined nucleotides are the sequence for a hemagglutinin epitope.

^dUnderlined letters represent the P2A sequence with a GSG linker (italic).

Construction of expression vectors.

Full-length UL43 or a truncated form was amplified by PCR using appropriate primers (Table 1). For cloning, PCR products were digested with EcoRI and BamHI and inserted into plasmid pEGFP-N3 (Clontech) to obtain fusion proteins with an EGFP tag at the C terminus and were termed pUL43-EGFP (wild type) and pUL43s-EGFP (truncated mutant), respectively. In addition, a construct that allows coexpression of pUL43 and pUL56 was generated. First, a transfer vector containing pUL56 and EGFP was created based on pcDNA3 (Invitrogen), with the two genes connected by a P2A linker. Using the linker, pUL56 and EGFP should be expressed with similar stoichiometry due to the cotranslational “ribosome skipping” event mediated by the P2A peptide. This property and the P2A sequence used for this study were reported previously (23). Specifically, PCR products were amplified by using a forward primer containing the P2A sequence and a reverse primer for the egfp gene (Table 1). After digestion with EcoRI and XbaI, the purified P2A-EGFP fragment was ligated into the corresponding sites of a pcDNA3 plasmid harboring the UL56 plasmid, which resulted in the transfer vector pcUL56-P2A-EGFP. Second, the UL56-P2A-EGFP fragment was released from the transfer plasmid by cutting with BamHI and XbaI. The chimeric sequence was then inserted into the first cloning site of pVITRO2-mcs-Hygro (InvivoGen), where pUL43 was placed in the second cloning site. Similarly, a coexpression vector for pUL43 and pUL56(AY) was constructed, in which all the PPxY motifs in pUL56 had been mutated to AAxY.

Virus growth properties and plaque morphology.

The role of the UL43 gene in the production of progeny virus was evaluated by one-step growth kinetics. Confluent NBL6 cells were infected with the parental virus vAb4G (16), the mutant virus vUL43STOP, or its revertant virus vUL43STOP_R at an MOI of 3. Cells were incubated at 4°C for 1 h to allow attachment of the viruses and then shifted to 37°C for virus entry. After 1.5 h of penetration, infected cells were treated with ice-cold citrate buffer (0.062 M Na₂HPO₄, 0.132 M citric acid, 0.5% bovine serum albumin [BSA] [pH 3.0]) for 30 s to remove surface-bound virions that had not entered the cells. After washing with phosphate-buffered saline (PBS) three times, fresh medium was added, and the cells were kept at 37°C. To determine the extracellular and intracellular virus titers, supernatants and infected-cell pellets were separately harvested at 0, 4, 8, 12, 24, and 36 h postinfection (p.i.). Samples were titrated on RK13 monolayers and overlaid with 1.5% (wt/vol) methylcellulose in EMEM. At 72 h p.i., when viral plaques were clearly visible, the methylcellulose was discarded, and the cells were fixed with 3.5% paraformaldehyde (PFA) in PBS for 5 min. Plaques were counted after staining with a 0.1% (wt/vol) crystal violet solution. The data are presented as PFU per milliliter from three independent experiments.

The influence of UL43 on viral cell-to-cell spread was investigated by measuring plaque sizes. To this end, each virus at an MOI of 0.0001 was inoculated onto 6-well plates where RK13 cells were seeded and grown to confluence. After incubation for 2 h at 37°C, residual virus was removed, and cell monolayers were covered with 1.5% (wt/vol) methylcellulose in EMEM. At 72 h p.i., viral plaques were inspected under a Zeiss Axiovert S100 microscope. For each virus, at least 50 plaques were randomly acquired with a digital camera. Diameters of the plaques were measured by using the ImageJ program (http://imagej.nih.gov/ij/), using a line tool. The measurements were normalized to those from the parental virus, which were set as 100%. Three independent assays were conducted.

Immunofluorescence and microscopy.

Cells were seeded onto coverslips and grown to 70 to 80% confluence. After transfection or infection, cells were fixed in 3.5% PFA in PBS for 5 min and permeabilized with 0.1% Triton X-100 lysis buffer for 10 min. Coverslips were blocked in PBS supplemented with 3% BSA for 2 h at room temperature (RT). Later, cell samples were probed with primary antibodies, including anti-HA MAb (1/200), anti-EHV-1 gC MAb (1/100), and anti-pUL56 PAbs (1/200), for 1 h. After three washes with PBS, cells were reacted with Alexa Fluor 568-conjugated goat anti-mouse IgG (1/500) or Alexa Fluor 488-conjugated goat anti-rabbit IgG (1/500) for another hour at RT. After washing, coverslips were mounted onto glass slides, stained with 4′,6-diamidino-2-phenylindole (DAPI) medium (Vector Laboratories), and examined under a Zeiss Axio Imager M1 microscope or a confocal laser scanning microscope (LSM510; Zeiss).

Western blotting.

Cell lysates were extracted with radioimmunoprecipitation assay (RIPA) buffer (25 mM Tris [pH 7.5], 150 mM NaCl, 1% Nonidet P-40, 0.25% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 1 mM EDTA) containing a protease inhibitor cocktail (Roche) and Benzonase (Novagen). Samples were separated by SDS-12% polyacrylamide gel electrophoresis (PAGE), as described previously (15). After fractionation, proteins were transferred onto polyvinylidene difluoride (PVDF) membranes (Carl Roth). To prevent unspecific binding, membranes were blocked in PBS containing 0.05% Tween 20 (PBST) with 5% (wt/vol) skim milk powder overnight at 4°C. Membranes were then probed with the following primary antibodies for 1 h at RT: anti-HA MAb (1/1,000), anti-pUL56 PAb (1/500), anti-EHV-1 IR6 PAb (1/500), anti-EHV-1 gM MAb (1/500), and anti-β-actin MAb (1/1,000). Membranes were washed with PBST three times and incubated with suitable HRP-conjugated secondary antibodies (1/10,000) for 1 h at RT. After three washing steps with PBST, reactive proteins on membranes were visualized by using an enhanced chemiluminescence (ECL) detection kit (Amersham ECL Prime; GE Healthcare).

Flow cytometry.

Mock-infected, infected, or transfected cells were trypsinized and suspended in PBS with 2% FCS. To determine cell surface MHC-I levels, NBL6 cells were incubated with anti-MHC-I (CZ3) MAb (1/50) or isotype control IgG (1/100); alternatively, HeLa or HEK293 cells were incubated with mouse anti-HLA class I (W6/32) MAb (1/50). After 30 min of incubation on ice, cells were washed with PBS plus 2% FCS three times and reacted with Alexa Fluor 647-conjugated goat anti-mouse IgG (1/1,000) for 30 min. At least 10,000 live cells from each sample were analyzed by using a FACSCalibur flow cytometry system according to the manufacturer's instructions (BD Biosciences). The data from flow cytometry experiments are presented as mean fluorescence intensity values.

RESULTS

Identification of EHV-1 pUL43 as a novel inhibitor of MHC-I presentation.

Our previous studies showed that pUL56 failed to reduce cell surface expression of MHC-I molecules in the absence of virus infection (11), suggesting that there is at least one additional viral gene product that directly or indirectly cooperates with pUL56 to mediate MHC-I downregulation. To identify viral proteins functionally cooperating with pUL56, we engineered a library of 26 single-gene knockout mutants of EHV-1 strain Ab4. The library was based on predictions of gene functions, and we particularly focused on early genes predicted to be nonessential for virus growth. Our selection list was shortened only for those genes that are functionally unclear or expressed with early kinetics and possibly located in the Golgi vesicles. The mutants were generated by the insertion of a positive selection marker (Kan^r) and removal of the respective open reading frames (ORFs) using en passant mutagenesis. The reconstituted viruses were tested for their potential to downregulate expression levels of cell surface MHC-I in equine NBL6 cells (Table 2). Among the generated mutant viruses, only the ORF17-negative virus was unable to significantly induce downmodulation of cell surface MHC-I levels. The ORF17 gene product is homologous to HSV-1 pUL43, as established by sequence and structure comparisons (6). Given that the selection marker may interfere with the expression of neighboring ORFs, we minimized the alterations in the viral genome and abolished expression of pUL43 by replacing the third codon of the ORF (TAC) with a stop codon (TAA) (Fig. 1A and B). Prior to virus reconstitution, the pAb4 mutant BAC (pUL43STOP) was confirmed by RFLP analysis (Fig. 1E) and DNA sequencing (data not shown). Transfection of pUL43STOP DNA resulted in viable virus, and the mutant was termed vUL43STOP. In comparison to the parental virus, the ability of vUL43STOP to induce MHC-I downregulation was substantially attenuated (P < 0.05), consistent with the results of the library screening (Fig. 2A and B). When the stop codon introduced at position 3 of the ORF was repaired to the original codon TAC, the resulting virus mutant, vUL43STOP_R, was able to trigger a reduction of MHC-I with kinetics and efficiency indistinguishable from those of the parental virus (P > 0.05) (Fig. 2B). This effect was mediated exclusively by the UL43 gene, as we confirmed that the expression of the downstream UL44 (gC) gene was not affected by the stop codon (data not shown). Our results demonstrated that pUL43 plays an important role in regulating cell surface expression of MHC-I during virus infection. Furthermore, we found that pUL43 not only exerted this function in equine cells but also induced a considerable decrease of cell surface MHC-I expression in human cells (Fig. 2C), indicating that pUL43 might govern a conserved pathway in mammalian cells to redistribute MHC-I.

TABLE 2.

Deletion mutants tested for MHC-I downregulation

Deletion mutant	Homologue of HSV-1, function	Virus growtha	MHC-I recoveryb
ORF3	NA, unknown	+	N
ORF4	UL55, unknown	+	N
ORF5	UL54, transcriptional activator	+	N
ORF8	UL51, unknown	+	N
ORF11	VP22, tegument protein	−	NA
ORF12	UL48, tegument protein	+	N
ORF13	UL47, tegument protein	−	NA
ORF14	UL46, unknown	+	N
ORF15	UL45, virion protein	+	N
ORF17	UL43, multiply hydrophobic protein	+	Y
ORF19	UL41, host shutoff virion protein	+	N
ORF23	UL37, unknown	−	NA
ORF34	NA, ubiquitinated virion protein	+	N
ORF38	UL23, Thymidine kinase	+	N
ORF40	UL21, unknown	−	NA
ORF41	UL20, multiply hydrophobic protein	−	NA
ORF46	UL16, unknown	−	NA
ORF48	UL14, unknown	−	NA
ORF51	UL11, myristoylated virion protein	−	N
ORF55	UL7, unknown	−	NA
ORF58	UL4, unknown	+	N
ORF59	NA, cytosolic virion protein	−	NA
ORF63	ICP0, transcriptional activator	+	N
ORF68	US2, unknown	+	N
ORF67	NA, nucleocapsid egress	+	N
ORF76	US9, tegument protein	+	N

^a+ means that the deletion mutant could be reconstituted and grew efficiently, while − indicates that the deletion mutant could not be rescued.

^bAt 6 h p.i., levels of cell surface MHC-I expression in equine NBL6 cells were measured and compared between the individual mutant and parental viruses; Y, responsible for MHC-I downregulation; N, not responsible for MHC-I downregulation; NA, not assessed.

FIG 1.

Schematic for virus mutagenesis and RFLP analysis of the infectious BAC mutants. (A) Parental EHV-1 strain Ab4 was derived from a BAC, which harbors an egfp gene driven by the HCMV IE promoter. U_L, unique long region; U_S, unique short region; IR, internal repeats; TR, terminal repeats. (B to E) En passant mutagenesis was performed to generate the following mutants for the present study: UL43STOP, in which a stop codon, TAA, in the reverse direction was introduced into the ORF17 gene downstream of the second codon (B); HA-UL43, in which an HA tag (amino acid sequence YPYDVPDYA) was fused to the N terminus of ORF17 (C); HA-ΔN-UL43, in which amino acids 3 to 40 at the N terminus of ORF17 were deleted (D); and UL43STOP-ΔUL56, a mutant based on our previously reported mutant Ab4GΔ1 (11) with the introduction of a stop codon in the ORF17 gene, as described above (E). All mutants were confirmed by DNA sequencing. (F) Representative gel from RFLP analysis. DNA from the parental strain or the indicated BAC mutants was digested with SmaI or XhoI and separated by electrophoresis using an 0.8% agarose gel. Changes in the digestion profile are consistent with those predicted in silico.

FIG 2.

pUL43 induces downregulation of cell surface MHC-I. (A) Equine NBL6 cells were infected with vAb4G or vUL43STOP virus at an MOI of 3. At 6 h p.i., cells were analyzed by flow cytometry after incubation with mouse anti-MHC-I (CZ3) MAb and Alexa Fluor 647-labeled goat anti-mouse IgG. Representative dot plots are derived from three independent experiments. (B) vAb4G, vUL43STOP, or vUL43STOP_R virus was used to infect NBL6 cells at an MOI of 3. Surface MHC-I levels were measured after 6 h p.i., as described above. (C) At 16 h p.i., HEK293 and HeLa cells infected with vAb4G or vUL43STOP virus were probed with mouse anti-HLA class I (W6/32) MAb and subjected to flow cytometry. (D) NBL6 cells were exposed to infection with the vHA-UL43 mutant. At 6 h p.i., cells were analyzed after incubation with mouse anti-MHC-I (CZ3) MAb and staining with Alexa Fluor 647-labeled goat anti-mouse IgG. All experiments were independently performed in triplicate and analyzed by using the Student t test. Data are presented as means ± standard deviations (error bars). Asterisks represent statistical significance (P < 0.05). ns, not significant.

To facilitate detection of pUL43 in further experiments and due to the lack of a specific antibody, we constructed a mutant virus, named vHA-UL43, in which the N terminus of pUL43 was tagged with an HA epitope (YPYDVPDYA) (Fig. 1C). With the inclusion of the HA sequence, the mutant virus was still able to induce a reduction of cell surface expression of MHC-I and did so as efficiently as the parental virus (P > 0.05) (Fig. 2D).

Abrogation of pUL43 does not impair virus production but slightly affects cell-to-cell spread.

It was unknown whether pUL43 is required for EHV-1 replication in horse fibroblasts. We therefore performed one-step growth assays with the parental virus vAb4G, the vUL43STOP mutant, and the vUL43STOP_R revertant on equine NBL6 cells. At the indicated times after infection, extracellular and intracellular titers were determined. With respect to the production of virus progeny, there was no significant difference between the mutant virus lacking UL43 (ORF17) and the parental virus (Fig. 3A), suggesting that pUL43 is nonessential for virus replication. In parallel, we analyzed the effects of pUL43 on viral cell-to-cell spread by measuring plaque sizes. The average diameter of plaques induced by vUL43STOP was ∼20% smaller than those of the parental virus (P < 0.05). Repair of the stop codon in pUL43 restored plaque formation to the morphology seen for the parental virus (Fig. 3B). These results demonstrate that pUL43 is nonessential for virus growth but is involved in virus spread between cells. Our findings are in agreement with data from previous studies on PRV and HSV-1, in which deletion of UL43 proved dispensable for virus growth in vitro (24–26).

FIG 3.

Effects of pUL43 on virus replication and cell-to-cell spread. (A) Single-step growth kinetics. NBL6 cells were infected with the parental virus vAb4G, vUL43STOP, or the vUL43STOP_R revertant at an MOI of 3. At the indicated times following infection, supernatants and cell pellets were harvested for determination of extracellular and intracellular titers, respectively. Data are from triplicate measurements and expressed as means ± standard deviations (error bars). (B) Comparison of plaque sizes. Individual viruses were used to infect RK13 cells at an MOI of 0.0001 and overlaid with methylcellulose. Three days after infection, images of at least 80 plaques for each virus were acquired with a camera. The plaque diameter of vAb4G was set as 100%, and the relative plaque sizes for other viruses were then calculated. A representative phenotype of a viral plaque is shown (green). Statistical significance (P < 0.05) is indicated by the asterisk.

Determination of the pUL43 expression pattern and degradation in lysosomes.

In previous reports on HSV-1 and PRV UL43 genes, mRNA was detected as early as 2 h p.i. even in the presence of phosphonoacetic acid (PAA), an inhibitor of viral DNA synthesis (25, 27). However, the expression kinetics of pUL43 throughout an infection cycle has not been addressed in detail. Detection of pUL43 homologues by SDS-PAGE and Western blot analysis is complicated by the fact that they are predicted to be highly hydrophobic polypeptides, which presents problems for the design and generation of specific antibodies. After repeated failure to raise antibodies to EHV-1 pUL43, we decided to fuse an HA tag with the N terminus of the target protein. As mentioned above, insertion of this epitope into the mutant virus (vHA-UL43) did not impair pUL43 function in MHC-I downregulation or virus growth. In infected cells, the gene product of pUL43 was expressed as a specific moiety of 34 kDa that could be detected from 2 h p.i., and the amount of protein continued to increase until the 8-h time point; afterwards, the levels of detectable protein began to decline (Fig. 4A). The single band was considered specific, although its apparent molecular weight is considerably lower than the predicted M_r of 43,000. To determine the temporal class of pUL43 expression, viral DNA synthesis was chemically inhibited with PAA, which substantially inhibited the expression of pUL43 as well as the production of gM, which was used as a control. In contrast, the levels of the product of the early IR6 gene were not affected by PAA treatment (Fig. 4A). To determine whether pUL43 is degraded at later times of infection, we used different inhibitors to block the pathways responsible for the breakdown of cellular proteins. Our results showed that expression levels of pUL43 were stabilized when infected cells were incubated with chloroquine, an inhibitor of lysosomes. In contrast, lactacystin, which inhibits the proteasome, did not protect pUL43 from degradation (Fig. 4B). Next, a combination of PAA and chloroquine was used to confirm whether pUL43 is indeed expressed with early kinetics. Compared to treatment with PAA alone, treatment of infected cells with a combination of chloroquine and PAA resulted in increased levels of pUL43 in infected cells (Fig. 4C). From these data, we concluded that pUL43 is an early gene product and directed to the lysosomal pathway for degradation at later times of infection.

FIG 4.

Determination of the pUL43 expression pattern and degradation by lysosomes. (A) Expression profile of pUL43 after infection. Cells were mock infected or infected with vHA-UL43 at an MOI of 3 in the absence or presence of PAA. Samples were harvested at different times postinfection. (B) Lysosomes rather than proteasomes are responsible for degradation of pUL43. After infection, cells were maintained in culture medium supplemented with 150 μM chloroquine or 5 μM lactacystin, respectively, for 12 h and 16 h. (C) pUL43 is an early gene product and subjected to lysosomal degradation during infection. Infected cells were treated with PAA, chloroquine, or both inhibitors for 12 h. To detect proteins, cell lysates were prepared in RIPA buffer. After separation by SDS-12% PAGE, proteins were transferred onto PVDF membranes and incubated with anti-HA MAb, anti-gM MAb, or anti-pIR6 PAbs. The immunoblots were developed by enhanced chemiluminescence. β-Actin was included as a loading control. Molecular mass markers were run for each blot in parallel, and sizes are indicated in kilodaltons.

pUL43 is present in the Golgi vesicles and requires the transmembrane domains at the C terminus for correct localization.

As the intracellular distribution of pUL43 was poorly defined, we performed indirect immunofluorescence microscopy to monitor the pUL43 subcellular distribution. RK13 cell lines that constitutively express EGFP-conjugated β-1,4-galactosyltransferase or calreticulin were generated by transfection and selection in the presence of G418. Calreticulin and β-1,4-galactosyltransferase are commonly used markers for the ER and Golgi compartments, respectively (28, 29). To avoid conflicts of EGFP signals during visualization, the mini-F cassette present in vHA-UL43 was excised by Cre-mediated recombination, resulting in a modified virus, vHA-UL43_M, that lacks EGFP expression in infected cells. RK13 cells expressing the individual compartment markers were infected with vHA-UL43_M. After reaction with anti-HA monoclonal antibody, pUL43 was found to predominantly localize to the Golgi apparatus. Moreover, the Golgi complex seemed partially fragmented, and pUL43 became dispersed into vesicles in infected cells (Fig. 5A). In contrast, pUL43 was not apparently associated with structures that expressed the ER marker (Fig. 5B). In transiently transfected cells, pUL43 was still distributed in a vesicular fashion, which is consistent with the observations of infected cells (Fig. 5D).

FIG 5.

Subcellular localization of pUL43 and role of TM domains at the C terminus. (A) Cells expressing EGFP-labeled β-1,4-galactosyltransferase, a Golgi marker (green). (B) Cells expressing EGFP-labeled calreticulin, an ER marker (green). The two cell lines were mock infected or infected with vHA-UL43_M virus. At 6 h p.i., samples were fixed with 3.5% PFA in PBS and then permeabilized with 0.1% Triton X-100. A confocal laser scanning microscope was used to visualize the expression of pUL43 after sequential incubation with anti-HA MAb and Alex Fluor 568-conjugated goat anti-mouse IgG (red). Bar, 10 μm. (C) Domains of pUL43 homologues and phylogenic analysis. Putative TM domains were predicted with SOSUI (http://harrier.nagahama-i-bio.ac.jp/sosui/sosui_submit.html). Identical amino acids are highlighted in black after multiple-sequence alignment with ClustalW 2.0 (http://www.ebi.ac.uk/Tools/msa/clustalw2). The RxR and RLAA motifs are underlined with solid and dotted lines, respectively. Sequences of pUL43 proteins were derived from GenBank, including EHV-1 (GenBank accession number YP_053062), EHV-4 (accession number NP_045234), PRV (accession number AFI70808), VZV (accession number NP_040138), Marek’s disease virus (MDV) (accession number YP_001033972), HSV-1 (AER37980), and HSV-2 (NP_044513) proteins. The phylogeny tree was constructed by using the unweighted-pair group method using average linkages, after 10,000 bootstraps were executed by using the MEGA 6.0 toolkit. The identity and similarity of the amino acid (aa) sequences were calculated by using the SIAS online tool with default settings (http://imed.med.ucm.es/Tools/sias.html). (D) Expression of pUL43 and the truncated mutant lacking 4 TM domains at the C terminus. At 24 h after transfection with the indicated plasmids (green), cells were fixed and stained with DAPI (blue). Images were acquired by using an upright fluorescence microscope under a 100× oil objective. Bar, 5 μm.

Structurally, the C terminus of the pUL43 homologue contains 4 putative TM domains, which are more conserved than those at the N terminus in terms of their relative positions of amino acids. Additionally, two motifs, RxR and RLAA, were identified within the C terminus, which are conserved among related viruses (Fig. 5C). When the protein was truncated by the removal of these four predicted TM regions, the shortened pUL43 protein appeared with a diffuse pattern in the cytoplasm after transfection of an expression plasmid (Fig. 5D). Taken together, these data demonstrated that pUL43 primarily localizes to the Golgi network and suggested that at least one of the four predicted TM domains at the C terminus are required for correct subcellular localization.

The hydrophilic domain at the N terminus plays a critical role in MHC-I downregulation mediated by pUL43.

The pUL43 homologues of HSV-1 and PRV are predicted to represent type III transmembrane proteins (24, 25) (Fig. 5C). Apart from 10 putative TM regions, the topology of EHV-1 pUL43 is characterized by a unique hydrophilic domain at its N terminus and an extremely short C-terminal domain. In addition, in silico analyses revealed that no cleavable signal sequence is present, which led us to investigate the possible involvement of the flexible domain in pUL43-mediated downregulation of MHC-I. Using en passant mutagenesis, the region encompassing amino acids 3 to 40 was deleted, resulting in the vHA-ΔN-UL43 mutant virus (Fig. 1D). In infected cells, mutant pUL43 lacking the N terminus appeared more focally localized and was present in fewer vesicles than the full-length protein (Fig. 6A). Flow cytometry showed that the expression level of cell surface MHC-I was higher in cells infected with vHA-ΔN-UL43 or vUL43STOP, whereas a dramatic reduction of MHC-I levels was induced by vHA-UL43 (P < 0.05) (Fig. 6B). These results suggest that the hydrophilic N terminus is essential for the function of pUL43, possibly by modulating intracellular sorting and/or trafficking of pUL43.

FIG 6.

The hydrophilic domain at the N terminus plays a critical role in MHC-I downregulation mediated by pUL43. (A) Putative structures of pUL43 and its mutant lacking the N-terminal 40 amino acids. The orientation of proteins was simulated by using SOSUI (http://harrier.nagahama-i-bio.ac.jp/sosui/sosui_submit.html). Cells were infected with vHA-UL43 or mutant vHA-ΔN-UL43 virus. At 6 h p.i., cells were incubated with anti-HA MAb and visualized after staining with Alex Fluor 568-conjugated goat anti-mouse IgG (red) and DAPI (blue). Coverslips were inspected with a 100× oil objective. Bar, 5 μm. (B) vHA-UL43, vHA-ΔN-UL43, or vUL43STOP was used to infect NBL6 cells for 6 h. Levels of cell surface MHC-I were measured by flow cytometry, and triplicate assays were performed independently. Data are expressed as means ± standard deviations (error bars). Differences between various treatments were evaluated by Student's t test. The asterisk represents a significant level (P < 0.05). ns, no significant difference.

pUL43 and pUL56 cooperate to downregulate cell surface MHC-I in transfected cells.

Although pUL56 was shown to be an inhibitor of cell surface MHC-I presentation at early times of EHV-1 infection, this effect could not be achieved by transfection of a UL56 expression vector (11). As pUL43 and pUL56 share similarities in terms of the intracellular localization and expression patterns and the abrogation of both genes in one virus did not have an additive effect (Fig. 1E and 7A), we hypothesized that the two proteins cooperate to trigger the downregulation of MHC-I molecules. To test our hypothesis, an expression vector was constructed, which allows expression of pUL43 and pUL56 from a single vector. In addition, we introduced a self-cleavable EGFP marker that was separated from pUL56 by a P2A sequence to facilitate detection of transfected cells (Fig. 7B). We also replaced all PPxY motifs of pUL56 with AAxY elements, which resulted in the pUL56(AY) mutant (Fig. 7C). At 24 h after transfection, pUL43 and pUL56 or pUL56(AY) were detectable by Western blotting. As expected, some of the pUL56/UL56(AY)-P2A-EGFP protein was not completely cleaved and migrated more slowly, as determined by SDS-PAGE (Fig. 7C). Expression of pUL43 or pUL56 individually did not induce downregulation of cell surface MHC-I (Fig. 7D, left). However, a dramatic reduction of cell surface MHC-I molecules was observed in cells expressing both wild-type pUL43 and pUL56. In contrast, levels of cell surface MHC-I remained stable in cells coexpressing pUL43 and pUL56(AY) (Fig. 7D, middle). Cell surface levels of CD58, which was used as a negative control, were not affected by the expression of both pUL43 and pUL56 (Fig. 7D, right). To further assess the cooperation of pUL43 and pUL56, we performed indirect immunofluorescence followed by confocal microscopy and found that pUL43 colocalized with pUL56 in transfected as well as infected cells (Fig. 7E). These results are consistent with the conclusion that pUL43 cooperates with pUL56 to specifically induce the downregulation of MHC-I on the cell surface in transfected and infected cells.

FIG 7.

Downregulation of MHC-I is induced by coexpression of pUL43 and pUL56 in transfected cells. (A) Deletion of both pUL43 and pUL56 does not induce an additional increase of MHC-I expression. NBL6 cells were infected with individual viral mutants at an MOI of 3. At 6 h p.i., levels of cell surface MHC-I were measured by flow cytometry. There is no significant difference between single- and double-deletion mutants. The asterisk indicates statistical significance (P < 0.05). (B) Strategy of engineering the coexpression vectors with a self-cleavable EGFP marker. The P2A sequence was inserted between pUL56 and EGFP. The fragment of pUL56-P2A-EGFP was subcloned into the pVITRO-UL43 vector after digestion with BamHI and XbaI. (C) Sequence alignment and detection of the coexpression vectors. Mutations are highlighted in boldface italic type. HEK293 cells were transfected with the indicated vectors. At 24 h, expression of individual genes was detected by immunoblotting (IB) following incubation with anti-HA MAb and anti-pUL56 PAbs, respectively. β-Actin was used as a loading control. (D) Downregulation of MHC-I caused by pUL43 and pUL56. HEK293 cells were transfected with 500 ng of different expression plasmids. At 24 h, cells were harvested and processed for flow cytometry analysis. Anti-HLA class I MAb (W6/32) was used to react with cell surface MHC-I. CD58 was included as a negative control. Representative histograms are from three independent assays. (E) Colocalization of pUL43 with pUL56. (Top) HEK293 cells were transfected with plasmid pUL43-pUL56-P2A-EGFP for 24 h. (Bottom) NBL6 cells were infected with vHA-UL43 virus for 6 h. Samples were fixed with ice-cold acetone until the EGFP fluorescence was quenched. After reaction with anti-HA MAb and anti-pUL56 PAbs, the coverslips were stained with Alexa Fluor 488-labeled goat anti-mouse IgG (green) and Alexa Fluor 568-labeled goat anti-rabbit IgG (red). Images were captured by a confocal microscope with a 60× oil objective. Bar, 10 μm.

DISCUSSION

Antagonizing the MHC-I presentation pathway is an effective strategy to achieve immune evasion and a result of the long coevolution of herpesviruses and their respective hosts. Different members of the Herpesviridae family encode a variety of viral proteins to reduce MHC-I molecules on the cell surface, and they achieve this by exploiting diverse mechanisms. In the Alphaherpesvirinae, the ICP47 homologue of HSV-1 was the first protein identified that blocks peptide binding by direct interaction with TAP (9). Likewise, the pUL49.5 homologues of the varicelloviruses, including bovine herpesvirus 1 (BoHV-1), PRV, EHV-1, and EHV-4, interfere with the activity of TAP by proteasomal degradation and/or by inhibiting the affinity of ATP for TAP (10, 12). It is noted that ICP47 homologues are absent in varicelloviruses, but the pUL49.5 homologue of HSV-1 cannot mediate MHC-I downregulation (12), suggesting that suppression of the MHC-I pathway by a particular gene product can be restricted to specific viruses and that it is not the action of an individual protein but the result that is conserved. Moreover, the US3 kinases of VZV, PRV, and HSV-1 were also demonstrated to be necessary for induction of the downregulation of MHC-I during productive infection (30–32). The mechanisms by which pUS3 homologues achieve downregulation vary greatly between viruses and were shown to be highly dependent on the particular cell type infected. Therefore, it is necessary to explore the entire repertoire of viral genes that hold the potential of blocking MHC-I presentation based on the virus species of interest.

In our previous studies, the pUL56 homologue of EHV-1 was identified as a novel viral protein that modulates the presentation of MHC-I molecules at the cell surface by accelerating dynamin-dependent endocytosis (15). However, pUL56 alone is not sufficient to induce the downregulation of MHC-I, as it did so only in the context of viral infection, implying that either a direct or indirect interaction of a viral protein(s) with pUL56 is required for MHC-I depletion. To test this hypothesis, we screened a single-gene knockout library of EHV-1 and focused on genes that were predicted to be nonessential for virus growth and not well defined with respect to function. These efforts led to the identification of a viral ORF17 gene encoding the pUL43 homologue that we then showed to be involved in the downregulation of MHC-I at early times of infection, coinciding with the time when pUL56 was shown to exert its function.

Given that this is the first description of EHV-1 pUL43, we characterized its role in virus growth and examined its expression pattern and subcellular localization. We found that the insertion of a stop codon within the open reading frame of pUL43 had little effect on virus growth in vitro and only mildly inhibited virus spread between cells. These findings on EHV-1 pUL43 are in agreement with findings for its counterparts in HSV-1 and PRV (24–26). Due to the difficulty in the generation of specific antibodies, the expression profiles of pUL43 homologues have not been well determined, and most of the available studies focused only on the detection of mRNA transcripts. Treatment with PAA to inhibit viral DNA synthesis revealed that mRNA transcripts of pUL43 homologues in HSV-1 and PRV are produced with early kinetics and are detectable as early as 2 h p.i. (25, 27). However, our expression kinetics showed that the production of pUL43 in cells after EHV-1 infection was reduced when viral DNA synthesis was blocked, suggesting that pUL43 could also be expressed as a late protein. However, in light of our observation that the pUL43 protein is degraded in lysosomes at later times of infection, we currently surmise that pUL43 represents a bona fide early and not a late protein and that the reduced expression levels under PAA treatment are a result of degradation. This conclusion is supported by an experiment in which the addition of lysosome inhibitors resulted in levels of pUL43 that were only marginally affected in the presence of PAA (Fig. 4C). It is noteworthy that EHV-1 pUL43 is expressed as a single species without any detectable posttranslational modifications, regardless of infection or transfection; however, the protein migrates faster than predicted from its predicted molecular weight. This anomalous mobility, as assessed after SDS-PAGE, is commonly seen in all studied pUL43 homologues (26, 27). A reasonable interpretation of this migration anomaly is that pUL43 exhibits extraordinary hydrophobicity and, consequently, may not be fully accessible to the detergent (33). In this case, nondenatured pUL43 protein could aggregate and migrate with a mobility that is faster than expected. This abnormal mobility would certainly make it difficult to identify posttranslational modifications, including ubiquitination.

Phylogenic analysis predicts that pUL43 homologues are conserved in the Alphaherpesvirinae (26). pUL43 homologues commonly consist of multiple TM regions but have various numbers of TM domains and share low amino acid identity (Fig. 5C). To ascertain the intracellular localization of pUL43, we applied confocal microscopy and observed that pUL43 is located primarily at the Golgi apparatus, which appears to be critically dependent on the most conserved TM domains at the C terminus of the protein. The pUL43 homologue of PRV is present in vesicles and inhibits syncytium formation, indicating that it might be involved in the trafficking of membrane proteins and vesicles (26). Consistent with this localization, PRV pUL43 is incorporated into virions, which was also shown for the EHV-1 homologue (Antonie Neubauer-Juric, personal communication). The Golgi apparatus is an organelle that directs sorting and trafficking of proteins (34, 35), such as mature MHC-I. Evidenced by the localization of pUL56 to the Golgi network and its role in inhibiting MHC-I presentation (11), it is conceivable that localization of viral proteins to the Golgi compartment to obstruct MHC-I presentation would be an optimal strategy for immune evasion. Our current model predicts that, similar to pUL56, pUL43 is targeted to Golgi and endocytic vesicles. To this end, pUL43 specifies two pivotal domains. First, the association of pUL43 with the Golgi complex is maintained by the TM domains at the C terminus; second, the N-terminal hydrophilic domain determines the localization in vesicles that are involved in directing the intracellular transport of MHC-I molecules.

Although both pUL56 and pUL43 are Golgi-associated viral proteins, neither of them is able to cause downregulation of MHC-I independently. Arguably, the most significant finding of this report is our demonstration that pUL43 and pUL56 collaborate in decreasing the amount of cell surface MHC-I molecules (Fig. 7). This process requires the PPxY motifs present in the cytoplasmic domain of the type II transmembrane protein pUL56. Similar to HSV-2 pUL56 (36), the mutated EHV-1 protein also migrated faster, as assessed by SDS-PAGE, when AAxY motifs were introduced instead of PPxY (Fig. 7C). We cannot exclude the possibility that pUL56 is structurally altered, but we currently surmise that the change in mobility is caused by a difference in electric charge. Like the HSV-2 orthologue, this change may also be triggered by the presumably absent interaction between pUL56 and Nedd4. Due to technical limitations, we did not further investigate the mechanisms that regulate MHC-I reduction after cotransfection of pUL43 and pUL56, but we favor a model in which pUL43 and pUL56 orchestrate the sorting of MHC-I to and its degradation in endolysosomes, which is based mainly on the colocalization of pUL43 and pUL56 (Fig. 7E) and the following previously reported experimental evidence: (i) PPxY motifs are required for the interaction of pUL56 with the cellular E3 ubiquitin ligase Nedd4 (36); (ii) trafficking of viral proteins containing PPxY motifs toward the endolysosomal pathway requires Nedd4.1-mediated ubiquitination and recruitment of TSG101, a component of ESCRT-1 (endosomal sorting complex required for transport) (37); and (iii) Nedd4 is known to ubiquitinate proteins with multiple transmembrane domains, such as ion channels, thereby facilitating endocytosis and degradation (38). It is conceivable that pUL43 is modified similarly and acts as an adaptor for MHC-I endocytosis where pUL56 is the recruiter for Nedd4. This interpretation is supported at least by the colocalization of pUL56 and MHC-I in the Golgi and endosomal vesicles during EHV-1 infection (15).

It has been shown that EHV-1 strains differ in their potential to reduce cell surface MHC-I expression. Infection with the Ab4 strain caused a severe downregulation of MHC-I molecules, while the levels of surface MHC-I were moderately reduced by infection with the RacL11 strain (11). This difference in the modulation of MHC-I largely depends on the presence of pUL56, but our results demonstrate that pUL43 and pUL56 cooperate to decrease the expression of cell surface MHC-I both during virus infection and after transient transfection. We surmise that EHV-1 strains, which contain full-length pUL43 and pUL56, are likely to cause MHC-I downregulation through interaction of the two proteins. In contrast, RacL11 or other strains lacking pUL56 and/or pUL43 are unable to adopt this immune evasion strategy. Apart from MHC-I, a variety of cell surface molecules might be affected by the cooperation of pUL43 with pUL56, as pUL56 was recently shown to modulate a selection of cell surface markers in equine mesenchymal stem cells after EHV-1 infection (39). In the future, we will address questions on the involvement of pUL43 in MHC-I downregulation by various EHV-1 strains and the spectrum and functional consequences of the pUL43-pUL56 interaction.

In summary, our present study identified a new function for the poorly understood pUL43 homologue of EHV-1. pUL43 has an important role in modulating MHC-I presentation, although it is dispensable for virus growth and is subjected to lysosomal degradation in the course of infection. Interestingly, the combination of pUL56 and pUL43 induces the downregulation of cell surface MHC-I independently of viral infection. Our findings open a new aspect of the complex landscape of viral immune evasion and may provide useful insights into the rational design of immunotherapies against EHV-1 infection.