N-Linked Glycosylation and Expression of Duck Plague Virus pUL10 Promoted by pUL49.5

ABSTRACT Duck plague virus (DPV) is a member of the alphaherpesvirus subfamily, and its genome encodes a conserved envelope protein, protein UL10 (pUL10). pUL10 plays complex roles in viral fusion, assembly, cell-to-cell spread, and immune evasion, which are closely related to its protein characteristics and partners. Few studies have been conducted on DPV pUL10. In this study, we identified the characteristics of pUL10, such as the type of glycosylation modification and subcellular localization. The characteristic differences in pUL10 in transfection and infection suggest that there are other viral proteins that participate in pUL10 modification and localization. Therefore, pUL49.5, the interaction partner of pUL10, was explored. We found that pUL10 interacts with pUL49.5 during transfection and infection. Their interaction entailed multiple interaction sites, including noncovalent forces in the pUL49.5 N-terminal domains and C-terminal domains and a covalent disulfide bond between two conserved cysteines. pUL49.5 promoted pUL10 expression and mature N-linked glycosylation modification. Moreover, deletion of UL49.5 in DPV caused the molecular mass of pUL10 to decrease by approximately3 to 10 kDa, which suggested that pUL49.5 was the main factor affecting the N-linked glycosylation of DPV pUL10 during infection. This study provides a basis for future exploration of the effect of pUL10 glycosylation on virus proliferation. IMPORTANCE Duck plague is a disease with high morbidity and mortality rates, and it causes great losses for the duck breeding industry. Duck plague virus (DPV) is the causative agent of duck plague, and DPV UL10 protein (pUL10) is a homolog of glycoprotein M (gM), which is conserved in herpesviruses. pUL10 plays complex roles in viral fusion, assembly, cell-to-cell spread, and immune evasion, which are closely related to its protein characteristics and partners. In this study, we systematically explored whether pUL49.5 (a partner of pUL10) plays roles in the localization, modification, and expression of pUL10.

reported by the ICTV (4). The DPV genome has 78 open reading frames, and the proteins encoded by them are employed to take part in viral replication and morphology (4).
The envelope in alphaherpesviruses contains approximately 10 glycosylated proteins, which are vital for attachment, penetration, envelopment, and viral fusion (5)(6)(7)(8)(9). DPV protein UL10 (pUL10, glycoprotein M homolog) is encoded by the UL10 gene. It is conserved throughout the Herpesviridae family and nonessential for most alphaherpesviruses, such as herpes simplex virus 1 (HSV-1), bovine herpesvirus 1 (BHV-1), pseudorabies virus (PRV), and varicella-zoster virus (VZV) (10)(11)(12)(13). pUL10 plays complex roles in viral fusion, assembly, cell-to-cell spread, and immune evasion, which are closely related to its protein characteristics and numerous partners (11,(13)(14)(15)(16)(17)(18)(19)(20)(21). pUL10 homologs have been shown to effectively help equine herpesviruses 1 (EHV-1) virions invade primary equine respiratory epithelial cells (22). HSV-1 pUL10 has been targeted to the inner nuclear membrane at 4 h after virus infection, and it may modulate the entry of endonuclear capsids into the cytoplasm via primary envelopment and development (20). Moreover, pUL10 has been found to be localized to and to attract glycoprotein D (gD), gH, and gL transport to the Golgi apparatus in PRV and HSV-1 (23). gH/gL was shown to be relocalized to the sites of secondary envelopment that were vital for their incorporation into virions (24). Furthermore, the location transfer of the fusion core protein by pUL10 may result in the inhibition of viral fusion (23). Additionally, pUL10 homologs of PRV, EHV-1, and infectious laryngotracheitis virus have been shown to have inhibitory effects on viral fusion in transfection (25). pUL10 is a type III transmembrane protein with eight transmembrane domains that are sufficient for its localization and viral growth (26). Two amino acids (aa; valine 42 and glycine 301) of pUL10 were found to be responsible for pUL10 maturation in VZV infection (10). Moreover, deletion of the UL10 gene of BHV-1 has been shown to reduce the plaque morphology (12), and the pUL10 and pUL11 double deletion mutant has been found to exhibit additive growth defects and 70% plaque size reduction in HSV-1 (13). When a pUL10 deletion mutant of EHV-1 was used as a live vaccine to immunize foals, the host produced significantly more complement and neutralizing antibodies than other vaccination groups (27). pUL10 has multiple partners at different stages in viral infection. Glycoprotein N partners with pUL10 to modulate its activity in membrane fusion (17) and has been shown to act as an inhibitor of the transporter associated with antigen processing (TAP) to escape the recognition of cytotoxic T lymphocytes in BHV-1, PRV, and EHV-1 (28,29). pUL10 interacts with gE and pUL49, which play a role in secondary envelopment and pUL49.5 incorporation (30). Extended-synaptotagmin 1, as an interaction partner of pUL10, is a key mediator that triggers the secretion of cytosolic protein (31), and its knockdown in host cells results in increased virus-induced fusion and infectious virions in supernatants (19). The XPO6 exportin is required for pUL10 to leave the nuclear membrane in late infection (16). In Epstein-Barr virus, pUL10 interacts with the cellular protein p32, which is involved in nuclear egress and promotes viral morphology (32). Although it is known that pUL10 has many partners, the effect of interacting proteins on pUL10 itself has not been specifically explored. In this study, we found that the characteristic difference of pUL10 in transfection and infection was prompted by pUL49.5. Moreover, pUL10 modification and expression facilitated by pUL49.5, pUL10 (cysteine 54 [C54]), and pUL49.5 (cysteine 46 [C46]) were vital for the interaction between them, with other interaction sites strengthening their connection. This study provides a basis for future exploration of the effect of pUL10 on virus proliferation.

RESULTS
Characteristics of DPV pUL10 localization and modification. First, pUL10 localization and posttranscriptional modification were explored. Duck embryo fibroblast cells (DEFs) in 12-well plates were transfected with pCAGGS-UL10-3HA and collected at 36 h posttransfection, and an indirect immunofluorescence assay (IFA) was performed. pUL10 was localized in the cytoplasm, but it was not uniformly distributed, and the red fluorescence was intense around the nucleus (Fig. 1A). To determine the type of pUL10 glycosylation, peptide-N-glycosidase A (PNGase A), peptide-N-glycosidase F (PNGase F), endoglycosidase H (Endo H), and O-glycosidase were used. The first three enzymes were targeted for N-linked glycosylation, and the last enzyme was targeted for O-linked glycosylation. Proteins with mature N-linked glycosylation were sensitive to PNGase A and PNGase F, which can cleave between the innermost N-acetylglucosamine and asparagine residues of high-mannose, hybrid, and complex oligosaccharides. PNGase A differs from PNGase F in that it cleaves N-linked glycans with or without a-(1, 3)-linked core fucose residues. Immature N-linked glycosylation was formed in the endoplasmic reticulum (ER), which was mainly cleaved by Endo H. When pUL10 Microbiology Spectrum was expressed alone in DEFs, it was partially digested by PNGase F and Endo H (Fig. 1B), which was consistent with its localization (Fig. 1A). Next, the localization of pUL10 in DPV-infected cells was explored. Since there were no antibodies available for the IFA, a BAC-DPV-UL10-HA recombinant virus (where BAC is the bacterial artificial chromosome and HA is small peptide tag; see Materials and Methods for details) was constructed. DEFs were transfected with BAC-DPV-UL10-HA recombinant plasmids, and the changes in fluorescent plaques and cells over 2 days are shown in Fig. 1C. The HA tag sequence was successfully inserted into the carboxyl terminus of the UL10 gene, as expected by sequencing, which implied that the BAC-DPV-UL10-HA virus was generated. The pUL10-HA fusion protein was detected by mouse anti-HA tag MAb (M-HA), which showed that the HA tag was expressed correctly (Fig. 1D). The growth kinetics of BAC-DPV-UL10-HA were measured by a 50% tissue culture infective dose (TCID 50 ) assay, which is described in Materials and Methods. As shown in Fig. 1E, according to the multistep growth curve, the HA tag did not affect viral replication in DEFs. These experimental results suggested that BAC-DPV-UL10-HA could be used in subsequent experiments. DEFs were infected with BAC-DPV-UL10-HA (multiplicity of infection [MOI] of 0.1) and collected at 16, 24, and 36 h postinfection, and IFA was performed. pUL10 was localized in the perinuclear region with punctate aggregation (Fig. 1F), which was different from the results shown in Fig. 1A. Moreover, DEFs were infected with BAC-DPV-UL10-HA (MOI of 1), and cell debris was collected at 48 h postinfection and prepared for the glycosidase treatment assay. PNGase F worked more efficiently than the other three enzymes to digest pUL10 in viral infection (Fig. 1G), and VP22 was a nonglycosylated protein and used as a viral infection control.
DPV pUL49.5 interacts with pUL10 and promotes its glycosylation. The difference in pUL10 characteristics in transfection and infection suggested that other viral proteins participated in assisting in the modification and localization of pUL10 (Fig. 1). Our laboratory previously reported that DPV pUL49.5 colocalized with pUL10 (33). Therefore, pUL49.5, as the first candidate partner of pUL10, and its subcellular localization were explored. DEFs were transfected with pDsRed2-ER and pCAGGS-UL10-3HA, and pUL10 was mainly localized in the ER. Moreover, pUL49.5 was localized in the ER when it was expressed alone, but pUL49.5 and pUL10 were able to leave the ER and accumulate in the perinuclear region when they were coexpressed ( Fig. 2A). Therefore, we inferred that there would be an interaction between DPV pUL49.5 and pUL10. pCAGGS-UL49.5-Flag and pCAGGS-UL10-3HA plasmids were cotransfected into HEK-293T cells to explore their interaction. pUL49.5 was immunoprecipitated by mouse anti-Flag tag MAb (M-Flag), and pUL10 was also coimmunoprecipitated ( Fig. 2B) but not in the mouse IgG (M-IgG) group. Additionally, mouse anti-HA tag mAb (M-HA) was immunoprecipitated with pUL10 and coimmunoprecipitated with pUL49.5 but not the M-IgG control group (Fig. 2C) in BAC-DPV-UL10-HA-infected (MOI of 1) DEFs. Then, when pUL10 was coexpressed with pUL49.5 in DEFs, it was mainly digested by PNGase F (Fig. 2D), which was similar to the results shown in Fig. 1G. Taken together, these data showed that the interaction between pUL49.5 and pUL10 was present in transfection and viral infection. Relocalization and mature N-linked glycoprotein of pUL10 were facilitated by pUL49.5.
DPV pUL10 (C54) and pUL49.5 (C46) were crucial residues for complex formation. To explore the interaction between pUL49.5(30-60 aa) and pUL10, we compared their amino acid sites in 12 herpesviruses (Table 1). There were two conserved cysteine residues in pUL49.5 and pUL10, cysteine residue 46 and cysteine residue 54 (C46 and C54), respectively ( Fig. 4A and B). The interaction between DPV pUL49.5 and pUL10 may form covalent disulfide bonds because of the two conserved cysteines; therefore, b-mercaptoethanol (b-ME) was used to break intramolecular disulfide bonds. Cell lysates were collected at 36 h posttransfection and treated with or without b-ME ( Fig. 4C and D). As shown in Fig. 4D, the molecular weight of pUL10 in the pCAGGS- UL10-3HA and pCAGGS-UL49.5-Flag cotransfection groups was approximately 40 to 45 kDa with b-ME treatment, but it had bands at 40 to 55 kDa (indicated by a red asterisk in the upper panel) without b-ME treatment and not in the other groups. An approximately 50-kDa band of pUL49.5 (indicated by a red asterisk in the lower panel) was detected in the pUL10/pUL49.5 cotransfection group without b-ME treatment, and its molecular mass corresponded with the uppermost image of pUL10 (red asterisk in upper panel). The conserved cysteines were mutated in pUL49.5 and pUL10, and the higher band disappeared in the b-ME-untreated group. The results showed that pUL49.5 C46 and pUL10 C54 formed a covalent disulfide bond. The influence of the two cysteine sites on pUL10 glycosylation was explored. The mature N-linked glycosylation of pUL10 was reduced when pUL49.5 C46 or pUL10 C54 was mutated (Fig. 4C). Additionally, other interaction sites in pUL49.5 still maintained a degree of N-linked glycosylation of pUL10 when the covalent disulfide bond was broken. pUL10 (C54) and pUL49.5 (C46) were vital sites for the formation of the pUL10/pUL49.5 complex.
Expression of DPV pUL10 was promoted by pUL49.5. The expression and molecular mass of pUL10 in the pUL10/pUL49.5 coexpression group were higher than those in the pUL10 and pCAGGS vector cotransfection group, even when the number of pCAGGS-UL10-3HA plasmids in each group was the same (Fig. 4D and E). Therefore, the expression of DPV pUL10 promoted by pUL49.5 was explored. The experimental design is described in Materials and Methods. DPV pUL10 expression was increased when it was coexpressed with pUL49.5 (Fig. 5A). Following densitometric analysis by ImageJ, the ratio of pUL10 and pUL49.5 to GAPDH (glyceraldehyde-3-phosphate dehydrogenase) density was calculated (Fig. 5B). We found that the mature N-linked glycosylation of pUL10 required pUL49.5 participation and that a certain amount of pUL49.5 was needed for more efficient promotion through dose-dependent experiments (Fig. 5C). Additionally, pUL10 expression increased in the presence of pUL49.5 over time, while pUL10 expression was maintained in the pCAGGS group (Fig. 5D). In addition, degradation of pUL10 and pUL49.5 persisted when cycloheximide (CHX) was added to DEFs that were coexpressed with pUL10/pUL49.5 (Fig. 5E). Furthermore, their degradation was prevented by MG132 (proteasomes inhibitor), which means that they were degraded by   (Fig. 5F). Combined with the results above, these results showed that the expression of pUL10 were boosted by pUL49.5. Deleting UL49.5 affects the N-linked glycosylation of pUL10. Based on the abovedescribed experimental results, the effect of pUL49.5 on the N-linked glycosylation of pUL10 in viral infection was explored. Deletion constructs of UL49.5 of BAC-DPV (BAC-DPV-DUL49.5) and its revertant virus (BAC-DPV-DUL49.5 R) were prepared in our laboratory (34).
DEFs were infected with equivalent titers of BAC-DPV-DUL49.5, BAC-DPV, and BAC-DPV-DUL49.5 R. Cell lysates were collected at 48 h postinfection and treated with or without PNGase F to detect the changes in pUL10 molecular mass. pUL10 was concentrated at approximately 36 kDa after treatment with PNGase F in the three strain virus infection groups. Differences were reflected in the untreated PNGase F group. Compared with the molecular weights of the other two virus strain infection groups, the pUL10 molecular weight was decreased by approximately 3 to 10 kDa in BAC-DPV-DUL49.5-infected cells, but it was still higher by approximately 2 to 3 kDa than that in the PNGase F-treated group (Fig. 6A). Following densitometric analysis by ImageJ, the ratios of pUL10 to VP22 density and pUL10 to b-tubulin density were calculated, and the results shown in Fig. 6B and C suggest that the expression of pUL10 was reduced in the BAC-DPV-DUL49.5-infected cell group compared to the other two groups. According to these results, pUL49.5 affected the expression and mature N-glycosylation of pUL10.
We detected the localization of pUL10 in the cytoplasm, while it was localized in the perinuclear region in DEFs that were infected with BAC-DPV-UL10-HA. In transfection, pUL10 is mainly digested by Endo H and PNGase F, but during infection, pUL10 is mainly digested by PNGase F (Fig. 1), which suggests that other viral proteins promote the mature N-linked glycoprotein, and the site of pUL10 aggregation in the perinuclear region should be the Golgi apparatus, which is the site for maturation of N-linked glycans (35). Interestingly, the bands of pUL10 digested by PNGase F are different in the presence or absence of UL49.5 ( Fig. 1B and 2D). pUL10/pUL49.5 complex formation may change the folding of pUL10 and impact the diversity of the glycosylation modification degree in pUL10. Additionally, the pUL10 bands consisted of at least four bands, Each group of samples was divided into two parts: one was treated with PNGase F, and the other was not treated. The reactants were analyzed by WB. Rabbit anti-UL10 polyclonal antibodies, rabbit anti-UL49.5 polyclonal antibodies, and rabbit anti-VP22 polyclonal antibodies were used as primary antibodies. HRP-conjugated goat anti-rabbit IgG and HRPconjugated goat anti-mouse IgG were used as secondary antibodies. VP22 was used as a viral infection control, and b-tubulin was used as a loading control. (B) Following densitometric analysis by ImageJ, the ratio of pUL10 to VP22 density was calculated. (C) Following densitometric analysis by ImageJ, the ratio of pUL10 to b-tubulin density was calculated.
as indicated by asterisks in Fig. 6A. The blue asterisk indicates the same band as that of BAC-DPV-DUL49.5. We speculate that there are other viral proteins that promote band formation, because little difference from the PNGase F treatment group is evident when pUL10 is expressed alone in transfection (Fig. 1B). HA tag causes a little different migration in PNGase F treatment group between Fig. 1G and Fig. 2D and 6A.
DPV pUL49.5 contains three or even more sites to build the connection with pUL10, which is crucial for the conformational stability, transportation, and modification of pUL10. The N-linked glycosylation process is very complicated, and many factors can influence it (36)(37)(38). First, the protein requires correct conformation. In some cellular transporters, certain glycosylation sites depend on the formation of intramolecular disulfide bonds, which are also important for maintaining the protein conformation; otherwise, they may lead to functional changes (39). The proteins that can enter the secretory pathway require proper folding in the ER to ensure the native conformation, such as the formation of disulfide bonds (40). As shown in Fig. 5D, pUL10 was labeled with M-HA, and there was a weak band between 100 and 130 kDa (red asterisk), which could be the trimer of pUL10. The trimer would disband due to the intervention of pUL49.5. Therefore, pUL49.5 can affect the conformation and even the activity of pUL10 to boost its expression. N-linked glycosylation on the viral protein surface helps viruses invade host cells and avoid antibody reactions to some viruses (41)(42)(43)(44)(45). Correct localization is also critical for protein modification, such as N-linked glycosylation. In HSV-1, pUL10 was present as punctate aggregates around the nucleus in the absence of other viral proteins, and pUL49.5 reached the Golgi apparatus from the ER, which was promoted by pUL10 (17,46). However, DPV pUL10 is slightly different from HSV-1 pUL10. DPV pUL10 was predominantly localized in the ER when it was expressed alone in DEFs, as was pUL49.5. However, the pUL10/pUL49.5 complex was formed and transported to the perinuclear space. The sensitivity of pUL10 to Endo H and PNGase F changed to a sensitivity mainly to PNGase F ( Fig. 1B and 2D), which demonstrated that the mature N-linked glycosylation of pUL10 was completed. In HSV-1, pUL10(1-342 aa) is an Endo H-sensitive protein, but pUL10(1-361 aa) and pUL10 are Endo H-resistant proteins and the localization of these truncated proteins were significantly different from each other (26). N-linked glycosylation requires the N-X-T/S (asparagine-any residue except proline-threonine/serine) sequence. Mutation at any amino acid (N, T/S) would result in the failure of protein glycosylation modification (45). Conversely, the predicted N-linked glycosylation site may not be glycosylated in the natural state, and nonglycosylated sites will be glycosylated when the protein suffers some changes (39). DPV pUL10 contained two N-linked glycosylation sites according to bioinformatics analysis, and pUL10 was glycosylated during transfection and infection. Kaposi9s sarcoma-associated herpesvirus encodes a homolog of interleukin-6, and it contains two glycosylation sites, asparagine 78 and asparagine 89. The former is present in highmannose-type N-linked sugars and is dispensable for cytokine function, while the latter affects protein conformation and is critical for receptor binding (47). The HSV-1 pUL10 immature form has been shown to cause limited syncytium formation (48). In BAC-DPV-infected cells, the bands of pUL10 were composed of four parts (marked with red and blue asterisks in Fig. 6A) in the untreated PNGase F group. It showed that pUL10 has various degrees of glycosylation and is mainly dependent on pUL49.5. Each form of pUL10 may serve a different function for virus survival. Therefore, we need to further explore the function of hyperglycosylated pUL10 during viral infection and the influence of pUL10 localization in the Golgi apparatus on viral envelopment.
Data availability. The data sets used during the present study are available from the corresponding author upon reasonable request.