Bovine Herpesvirus 1 UL49.5 Interacts with gM and VP22 To Ensure Virus Cell-to-Cell Spread and Virion Incorporation: Novel Role for VP22 in gM-Independent UL49.5 Virion Incorporation

ABSTRACT Alphaherpesvirus envelope glycoprotein N (gN) and gM form a covalently linked complex. Bovine herpesvirus type 1 (BHV-1) UL49.5 (a gN homolog) contains two predicted cysteine residues, C42 and C78. The C42 is highly conserved among the alphaherpesvirus gN homologs (e.g., herpes simplex virus 1 and pseudorabies virus). To identify which cysteine residue is required for the formation of the UL49.5/gM complex and to characterize the functional significance of the UL49.5/gM complex, we constructed and analyzed C42S and C78S substitution mutants in either a BHV-1 wild type (wt) or BHV-1 UL49.5 cytoplasmic tail-null (CT-null) virus background. The results demonstrated that BHV-1 UL49.5 residue C42 but not C78 was essential for the formation of the covalently linked functional UL49.5/gM complex, gM maturation in the Golgi compartment, and efficient cell-to-cell spread of the virus. Interestingly, the C42S and CT-null mutations separately did not affect mutant UL49.5 virion incorporation. However, when both of the mutations were introduced simultaneously, the UL49.5 C42S/CT-null protein virion incorporation was severely reduced. Incidentally, the anti-VP22 antibody coimmunoprecipitated the UL49.5 C42S/CT-null mutant protein at a noticeably reduced level compared to that of the individual UL49.5 C42S and CT-null mutant proteins. As expected, in a dual UL49.5 C42S/VP22Δ virus with deletion of VP22 (VP22Δ), the UL49.5 C42S virion incorporation was also severely reduced while in a gMΔ virus, UL49.5 virion incorporation was affected only slightly. Together, these results suggested that UL49.5 virion incorporation is mediated redundantly, by both UL49.5/gM functional complex and VP22, through a putative gM-independent novel UL49.5 and VP22 interaction. IMPORTANCE Bovine herpesvirus 1 (BHV-1) envelope protein UL49.5 is an important virulence determinant because it downregulates major histocompatibility complex class I (MHC-I). UL49.5 also forms a covalently linked complex with gM. The results of this study demonstrate that UL49.5 regulates gM maturation and virus cell-to-cell spread since gM maturation in the Golgi compartment depends on covalently linked UL49.5/gM complex. The results also show that the UL49.5 residue cysteine 42 (C42) mediates the formation of the covalently linked UL49.5-gM interaction. Furthermore, a C42S mutant virus in which UL49.5 cannot interact with gM has defective cell-to-cell spread. Interestingly, UL49.5 also interacts with the tegument protein VP22 via its cytoplasmic tail (CT). The putative UL49.5 CT-VP22 interaction is essential for a gM-independent UL49.5 virion incorporation and is revealed when UL49.5 and gM are not linked. Therefore, UL49.5 virion incorporation is mediated by UL49.5-gM complex interaction and through a gM-independent interaction between UL49.5 and VP22.

B ovine herpesvirus type 1 (BHV-1) is an important pathogen of cattle that can cause a severe respiratory tract infection, known as infectious bovine rhinotracheitis (IBR), and abortion in pregnant cows (1, 2). In addition, BHV-1 is an important component of the bovine respiratory disease complex (BRDC), also known as shipping fever (3,4). The BHV-1 gene product, envelope protein U L 49.5, a glycoprotein N (gN) homolog of alphaherpesviruses, forms a disulfide-linked complex with envelope glycoprotein M (gM). Both proteins are nonessential although in the absence of either U L 49.5 or gM, virus yield is reduced significantly (5,6). The U L 49.5 gene products of BHV-1, herpes simplex virus 1 (HSV-1), and equine herpesvirus 1 (EHV-1) are not glycosylated (6)(7)(8). The corresponding U L 49.5 gene product of pseudorabies virus (PRV) is glycosylated and is termed gN (9). In PRV, gN is not essential for gM maturation in the Golgi compartment and for gM virion incorporation, but gM is necessary for gN virion incorporation (10). In contrast, formation of the U L 49.5/gM complex is essential for BHV-1 gM maturation in the Golgi compartment (11,12). Currently, it is not known whether BHV-1 gM and/or U L 49.5 is necessary for each other's virion incorporation. Among the varicelloviruses, BHV-1, PRV, and EHV-1 U L 49.5 or its gN homologs bind to the transporter associated with antigen presentation (TAP) in virus-infected cells and thereby downregulates major histocompatibility complex class I (MHC-I) cell surface expression (11,13). However, unlike the PRV and EHV-1 proteins, BHV-1 U L 49.5 not only binds but also degrades TAP (13).
BHV-1 U L 49.5 is a 9-kDa type I membrane protein (6). The predicted U L 49.5 open reading frame (ORF) encodes 96 amino acids (aa) and is composed of an N-terminal signal sequence of 22 aa, an extracellular luminal domain of 32 aa, a transmembrane (TM) domain of 25 aa, and a short cytoplasmic tail (CT) of 17 aa (14) (Fig. 1A). There are two predicted cysteine residues in the BHV-1 U L 49.5 ORF, C42 and C78. Alignment of BHV-1 U L 49.5 amino acid sequences with the corresponding U L 49.5 sequences of other alphaherpesviruses showed that C42, located within the luminal domain of BHV-1 U L 49.5, is highly conserved among alphaherpesviruses (Fig. 1B). The complex between U L 49.5 and gM is thought to be linked via disulfide bonds between cysteine residues. Since the BHV-1 U L 49.5 C42 is highly conserved among herpesviruses (Fig. 1B), we hypothesized that the BHV-1 U L 49.5/gM complex is mediated by the predicted U L 49.5 residue C42 and a predicted cysteine residue in the gM ORF (Fig. 1C).
It was previously postulated that BHV-1 U L 49.5 binds to TAP through its TM domain (11). However, it has not been possible to map the BHV-1 U L 49.5/TAP binding domain within the TM because even a short deletion within the BHV-1 U L 49.5 TM domain resulted in degradation of the protein (11,15). Additionally, it was reported that in a stably transfected cell line, gM interferes with U L 49.5-mediated TAP inhibition and MHC-I downregulation function, indicating that gM might compete with U L 49.5 for TAP binding (12). Recently, we have reported that U L 49.5 residues 30 to 32 (RXE motif) within the luminal domain and the U L 49.5 CT residues together mediated maximum U L 49.5 TAP inhibition function without affecting the covalent U L 49.5/gM interaction (15). These findings raised the question of whether the C78 residue within the U L 49.5 TM domain, also conserved in the PRV gN, is important for U L 49.5-TAP interaction and thereby MHC-I downregulation.
The goal of this study was to determine whether one or both cysteine residues are required for the formation of covalently linked U L 49.5/gM complex, gM maturation, cell-to-cell spread of the virus, and U L 49.5 or gM virion incorporation. Additionally, we wanted to investigate whether the mutation of one or both cysteine residues affects the U L 49.5-mediated MHC-I downregulation function.
To this end, we have constructed several BHV-1 U L 49.5 mutants with residue C42 or C78 replaced individually or simultaneously with a serine (S) residue using a U L 49.5 CT-null or wild-type (wt) virus as a backbone. Further, we have constructed two BHV-1 VP22 deletion mutants, one with wt U L 49.5 and the other with a U L 49.5 C42S mutation, and analyzed their respective levels of U L 49.5 virion incorporation. Finally, we constructed a virus with a deletion of gM (gM-deleted) and determined its U L 49.5 virion incorporation. The results demonstrated the following: (i) that the U L 49.5 residue C42 but not C78 is essential for formation of the U L 49.5/gM covalently linked complex and gM maturation in the Golgi compartment; (ii) that the U L 49.5 C42S and U L 49.5 CT-null mutant proteins are incorporated in the respective mutant's virion envelope but that the U L 49.5 C42S lacking U L 49.5 CT residues 80 to 96 is not or markedly reduced; (iii) that covalently linked U L 49.5 and mature gM are incorporated in the virion of a VP22 deletion (VP22Δ) strain and that, however, unlinked U L 49.5 and immature gM require VP22 for their virion incorporation; and (iv) that in the absence of U L 49.5/gM complex, a gM-independent U L 49.5-VP22 interaction mediated probably by U L 49.5 CT residues 80 to 96 is essential for U L 49.5 virion incorporation.

RESULTS
BHV-1 U L 49.5 forms a disulfide-linked complex with gM, which is required for gM processing in the Golgi compartment (5). BHV-1 U L 49.5 also downregulates MHC-I cell surface expression by interacting with TAP in the endoplasmic reticulum (ER) (11). To investigate whether one or both of the cysteine residues in U L 49.5 affects U L 49.5-gM interaction and gM processing, the C42S and C78S mutants were generated using infectious BHV-1 wt and BHV-1 U L 49.5 CT-null bacterial artificial chromosome (BAC) clones. In addition, a C42S/C78S/CT-null triple mutant virus with a double cysteine substitution was constructed. These mutant viruses were characterized with respect to their plaque phenotypes and growth kinetics, mutant U L 49.5/gM complex formation, gM maturation, and U L 49.5/gM virion envelope incorporation. In addition, effects of (B) Alignment of predicted amino acid sequences of BHV-1, PRV, EHV-1, and HSV-1 U L 49.5/gN homologs. Note that C42 (boxed) is conserved in all four gN homologs and that C78 is conserved in BHV-1, PRV, and HSV-1 but not in EHV-1. (C) Alignment of predicted amino acid sequences of BHV-1, PRV, EHV-1, and HSV-1 gM homologs. Cysteines (C) that are not conserved are in bold. Conserved cysteines are boxed. Asterisks (*) indicate positions which have a single, fully conserved residue; colons (:) indicate conservation between groups of strongly similar properties; periods (.) indicate conservation between groups of weakly similar properties. C42S and/or C78S mutation on the mutant U L 49.5-mediated downregulation of MHC-I cell surface expression due to TAP inhibition were analyzed.
Mutation of U L 49.5 residue C42 but not C78 resulted in a growth defect and small-plaque phenotype. To examine whether the C42S or C78S substitution within U L 49.5 affected viral replication kinetics and virus yield in infected MDBK cells, one-step growth curves of C42S, C78S/CT-null, C42S/CT-null, VP22Δ, C42S/VP22Δ, wt, and CT-null viruses were determined. The results showed that viral growth kinetics of CT-null and C78S/CT-null are almost identical to wt kinetics (Fig. 2). However, both the C42S and double C42S/CT-null mutant viruses replicated with a 10-fold-reduced virus yield compared to that of their parental wt and CT-null viruses (Fig. 2). As shown in Fig. 3A and B, average plaque sizes produced in MDBK cells by C42S and double C42S/CT-null mutant viruses were significantly smaller than those of their respective parental wt and CT-null viruses. However, C78S (Fig. 3B) and C78S/CT-null mutant viruses had approximately the same diameters as the parental wt and CT-null viruses ( Fig. 3A and B). The double cysteine C42S/C78S mutant virus produced plaques very similar to those produced by C42S and double C42S/CT-null mutant viruses, suggesting that the C42S mutation affected plaque size ( Fig. 3A and B). In the wt U L 49.5-expressing, stable MDBK cell line (MDBK-U L 49.5), the C42S and C42S/CT-null viruses produced plaques of wild-type size ( Fig. 3A and B), and they replicated with a 5-fold-higher titer than in noncomplementing MDBK cells (data not shown). However, the virus yield was still 5-fold lower than that of the wild-type virus, which could be due to a low level of U L 49.5 expression by the stable cell line (Fig. 3C). Therefore, these results indicated that the growth defects (smaller plaque phenotype and 10-fold lower yield) of the C42S, C42S/CT-null, and C42S/C78S mutant viruses were due to the replacement of the U L 49.5 C42 residue with a serine residue and not due to another mutation elsewhere in the genome. U L 49.5 residue C42 but not C78 is required for the formation of covalently linked U L 49.5/gM complex and gM maturation in the Golgi compartment. To determine whether U L 49.5 residues C42, C78, or both are essential for covalently linked U L 49.5-gM interactions and gM processing in the Golgi compartment, 35 S-labeled C42S, C78S, C42S/CT-null, C78S/CT-null, and C42S/C78S/CT-null mutant proteins expressed in the respective mutant virus-infected cells were immunoprecipitated with anti-U L 49.5 and anti-gM antibodies and analyzed by Western blotting. As controls, wt and CT-null virus-infected cell lysates were similarly analyzed. As shown in Fig. 4A, U L 49.5-specific antibody immunoprecipitated 9-kDa U L 49.5 wt, C42S, and C78S proteins, but 8-kDa U L 49.5 CT-null, C42S/CT-null, C78S/CT-null, and C42S/C78S/CT-null proteins were immunoprecipitated from the corresponding wt and mutant viruses. In addition, the antibody coimmunoprecipitated 43-kDa mature gM-specific proteins from wt, CT-null, C78S, and C78S/CT-null virus-infected cell lysates. However, the U L 49.5-specific antibody coimmunoprecipitated 36-kDa immature gM-specific proteins from the C42S, C42S/CT-null, and C42S/C78S/CT-null mutant virus-infected cell lysates unlike results with the wt and C78S mutant (Fig. 4A). Notably, a vastly reduced level of the 36-kDa immature gM was coimmunoprecipitated by the U L 49.5-specific antibody. As expected, gM-specific antibody immunoprecipitated the 43-kDa mature gM from wt, CT-null, C78S, and C78S/CT-null virus-infected cell lysates. Similar to results with immunoprecipitation with the anti-U L 49.5 antibody, a 36-kDa gM protein was also immunoprecipitated from the C42S, C42S/CT-null, and C42S/C78S/CT-null virus-infected cell lysates (Fig. 4B). In addition, the anti-gM-specific antibody coimmunoprecipitated the corresponding U L 49.5-specific 9-kDa C42S and C78S proteins and the 8-kDa CT-null, C42S/ CT-null, C78S/CT-null, and C42S/C78S/CT-null proteins. However, the levels of U L 49.5 C42S, C42S/CT-null, and C42S/C78S/CT-null proteins coimmunoprecipitated with the anti-gM antibody were reduced compared with the levels of the wt, CT-null, and C78S/CT-null proteins (Fig. 4B).
We hypothesized that the 43-kDa proteins detected in the wt, C78S, and CT-null virus-infected lysates are the mature Golgi-processed gM proteins and that the 36-kDa band detected in the C42S virus-infected lysate is the immature gM. Therefore, we determined their endoglycosidase H (EndoH) sensitivity. As expected, results showed that the 43-kDa mature gM protein (Golgi apparatus-processed) was resistant to EndoH digestion (Fig. 4C), but the 36-kDa immature gM protein was EndoH sensitive. Interestingly, under both reducing and nonreducing conditions, anti-U L 49.5-specific antibody recognized two higher-molecular-mass proteins of approximately 92 kDa in all virus-infected cell lysates and an approximately 80-kDa protein in wt-and C42Sinfected lysates. The 80-kDa band was not detected in VP22Δ and U L 49.5 C42S/VP22Δ virus-infected cell lysates. These bands were more prominent under the nonreducing conditions. Both the 92-kDa and 80-kDa proteins were absent in the mock-infected cell lysate (Fig. 5A). When the identical blot was immunoblotted with the gM-specific antibody, both proteins were absent (Fig. 5B). Thus, the 92-kDa band might represent a heterodimeric complex of the approximately 9-kDa U L 49.5 protein plus the approximately 82-kDa TAP1 protein (UniProt accession number A6QPZ6). Currently, the identity of the 80-kDa protein is not known.
Further, as shown in Fig. 5B under nonreducing conditions, the anti-gM antibody recognized an additional approximately 72-kDa protein in the C42S and C42S/VP22Δ virus-infected cell lysates but not in the mock-, wt-, and VP22Δ-infected cell lysates. It is highly likely that the 72-kDa protein represents a covalently linked homodimer of the 36-kDa immature gM. Note that there is a nonspecific 43-kDa faint band in the mock-infected sample in both panels A and B; this band is also present in the wt-and mutant virus-infected lysate samples but is visible only when the gM (43 kDa) is not processed (C42S mutants). Also, in panel A anti-U L 49.5 antibody precipitated a nonspecific 9-kDa faint band in the mock-infected sample, and this band is also visible in the CT-null lysates. (C) 35 S-labeled lysates from various mutant virus-infected MDBK cells were immunoprecipitated with anti-gM-specific antibody and digested with EndoH (ϩ). The untreated samples (Ϫ) were included as controls. EndoH-sensitive, immature gM is marked by asterisks.
To validate further that the U L 49.5 C42S mutation alone can be attributed to the disruption of the U L 49.5/gM covalent interaction required for gM maturation, we analyzed gM maturation in the U L 49.5 C42S mutant-infected MDBK-U L 49.5-expressing cell line. As expected, the effect of the U L 49.5 C42S mutation on gM maturation was complemented to its mature 43-kDa molecular mass by the U L 49.5-expressing cell line. However, the rescue or complementation of the 43-kDa protein was at a reduced level (data not shown). As noted above with respect to U L 49.5 C42S mutant virus yield, the complementation at a reduced level is due to a lower level of U L 49.5 expression by the stable cell line, probably due to a lower copy number of the expressed U L 49.5 gene than during wt virus infection. Therefore, these results indicated that the C42S mutation alone is responsible for the defective gM maturation and growth defect (small-plaque phenotype and 10-fold-lower virus titer) of the mutant U L 49. 5 C42S virus. Taken together, the results indicated (i) that U L 49.5 residue C42 but not C78 is required for the formation of a covalent U L 49.5-gM complex, (ii) that gM was processed to the mature 43-kDa protein only when it was covalently linked to wt U L 49.5, and (iii) that in the absence of covalently linked U L 49.5/gM complex, the immature gM can form a covalently linked homodimer.   6A). Therefore, only the simultaneous C42S and CT-null mutations but not the individual C42S and CT-null mutations affected U L 49.5 virion incorporation (Fig. 6A). These results indicated that both U L 49.5 residue C42 and U L 49.5 CT residues 80 to 96 are essential for U L 49.5 virion incorporation. Both C42S and U L 49.5⌬ mutant viruses incorporated immature gM in the virion envelope. The results presented in Fig. 6B showed that in the case of wt, CT-null, C78S, and double C78S/CT-null mutant viruses, both mature and immature gM proteins were incorporated into the virion envelope. The immature gM was incorporated into the envelope of C42S, double C42S/CT-null, triple C42S/C78S/CT-null, and U L 49.5Δ viruses (Fig. 6B). Therefore, incorporation of the immature gM in the virion envelope appears to be independent of its covalently linked interaction with U L 49.5.
The U L 49.5 CT residues 80 to 96 most likely interact with VP22. VP22 in HSV-1 and PRV is well known for its interaction with a number of envelope proteins (gM, gE, and gD) (16) and tegument protein VP16 (17,18). We hypothesized that U L 49.5 CT might be interacting with VP22 and thus play a role in the U L 49.5 C42S virion incorporation. Therefore, we determined whether the mutant U L 49.5 proteins expressed by C42S, CT-null, and double C42S/CT-null mutant viruses are coimmunoprecipitated with anti-VP22 antibody (Fig. 7). Since VP22 also interacts with gE, which was not manipulated in the U L 49.5 mutants, we compared the levels of U L 49.5, gE, and VP22 that are coimmunoprecipitated by anti-VP22 antibody from the corresponding virusinfected cell lysates. The results showed that in both wt and U L 49.5 mutant virusinfected lysates, VP22-specific antibody immunoprecipitated or coimmunoprecipitated similar levels of VP22 and gE ( Fig. 7B and C). However, the anti-VP22 antibody coimmunoprecipitated a reduced level of dual U L 49.5 C42S/CT-null mutant protein compared with the corresponding level of C42S and CT-null mutant proteins (Fig. 7A). Since VP22 also interacts with gM in alphaherpesviruses (16, 17), we determined additionally the level of VP22 coimmunoprecipitated with anti-gM antibody in the wt, U L 49.5 C42S, CT-null, and C42S/CT-null mutant virus-infected cell lysates with the anti-gM antibody. As shown in Fig. 7, regardless of gM maturation status (Fig. 7E), the levels of VP22 coimmunoprecipitated by the anti-gM antibody from the corresponding virus-infected cell lysates were very similar (Fig. 7F). Taken together, these data indicated (i) that U L 49.5 CT residues 80 to 96 most likely interact with VP22, which is revealed only in the absence of covalent U L 49.5/gM complex, and (ii) that neither C42S nor double C42S/CT-null mutations affected the gM-VP22 interaction.
In the absence of covalent U L 49.5/gM complex, incorporation of U L 49.5 in the virion envelope is probably mediated by interaction of U L 49.5 CT residues 80 to 96 and VP22. To determine whether VP22 interaction with U L 49.5 CT residues 80 to 96 plays an essential role in mutant U L 49.5 C42S protein virion incorporation, a double C42S/VP22Δ and VP22Δ mutant virus were constructed. We predicted (i) that mutant U L 49.5 C42S protein virion incorporation in the presence of U L 49.5 CT residues 80 to 96 and VP22 will not be affected, and (ii) that the C42S mutant protein containing U L 49.5 CT residues 80 to 96 but expressed by a VP22Δ mutant virus would be defective. As predicted, a significantly reduced level of U L 49.5 C42S, expressed in the backbone of a VP22Δ mutant virus, was incorporated into the virion envelope (Fig. 8). However, the results also showed that the 36-kDa immature gM expressed by the C42S mutant virus was incorporated into the virion, but the immature gM in context of the U L 49.5 C42S/VP22Δ virus was not. This raised the alternative possibility that lack of VP22immature gM interactions led to the defective U L 49.5 C42S incorporation into the virion. To exclude this possibility, we constructed a gM-deleted virus and determined whether U L 49.5 expressed in the absence of gM is incorporated in the virion envelope. As shown in Fig. 9, U L 49.5 expressed in the backbone of a gM-deleted virus was incorporated in the virion though at a slightly reduced level.

Neither the individual U L 49.5 cysteine residue mutations nor the combined mutations had an effect on MHC-I cell surface expression in mutant virus-infected cells.
To determine the effects of U L 49.5 residue C42 and C78 substitutions on U L 49.5mediated TAP inhibition or MHC-I downregulation, we compared MHC-I cell surface expression in the C42S, C78S, and double C42S/C78S mutant virus-infected cells with that of wt virus-infected cells. Fluorescence-activated cell sorting (FACS) analysis results clearly showed that the C42S and C78S mutations, either individually or combined, did not abrogate U L 49.5-mediated MHC-I downregulation (Fig. 10).

DISCUSSION
We conducted these studies to determine the following: (i) which of the two U L 49.5 cysteine residues (C42 and C78) is required for formation of the covalently linked U L 49.5/gM complex and gM maturation; (ii) whether one or both of the proteins play a role in each other's virion incorporation; and (iii) how U L 49.5 residue C42S and C78S mutations affect MHC-I downregulation. The results of this study determined (i) that the covalently linked U L 49.5/gM complex is necessary for BHV-1 gM processing in the Golgi compartment; (ii) that the U L 49.5 residue C42S substitution mutation and not the C78S mutation affected the formation of the covalently linked U L 49.5/gM complex; and (iii) that the covalently linked U L 49.5/gM complex is also necessary for efficient cell-to-cell spread of the virus and efficient virus replication. Notably, C42S mutant virus produced 30% smaller plaques and replicated with more than 10-fold-reduced virus yield. The results also showed the following: (iv) that in the absence of the covalently linked   Fig. 11. Previously, by using stable cell lines expressing gM or both U L 49.5 and gM, Lipinska et al. (12) reported that the U L 49.5/gM complex formation was required for gM maturation in the Golgi compartment. They also reported that in cells infected with a virus with a deletion of the U L 49.5 TM domain (BHV-1 U L 49.5ΔTM), gM was not processed in the Golgi compartment. Since the U L 49.5 TM domain contains one of the two cysteine residues of U L 49.5 (C78) and since U L 49.5/gM complex formation involves covalently linked disulfide bonds, they suggested that the U L 49.5 C78 residue is essential for U L 49.5/gM complex formation and gM processing. However, in that study, the status of the mutant U L 49.5 protein expressed by U L 49.5Δ TM virus was not analyzed. Recently, we reported that deletion of the U L 49.5 TM domain resulted in degradation of the mutant U L 49.5 protein and that gM expressed by the mutant U L 49.5Δ TM virus was not processed (15). Here, we have characterized the U L 49.5 C42S and C78S mutant viruses for U L 49.5/gM complex formation and gM processing. Our results demonstrate that the U L 49.5 residue C42 and not C78 was required for the formation of the covalently linked U L 49.5/gM complex and that in the absence of the covalently linked U L 49.5/gM complex, the C42S mutant virus produced smaller plaques and replicated with reduced virus yield. Nevertheless, in agreement with Lipinska et al. (12), we found that gM processing in the Golgi compartment is dependent on the covalently linked U L 49.5/gM complex formation. Therefore, in BHV-1, U L 49.5 (gN homolog) is a dominant determinant of gM maturation in the Golgi compartment. However, the opposite is true for HSV and PRV because gM is required for transport and/or processing of gN in the Golgi compartment (10,19,20).
Our results also indicate the following in BHV-1: (i) that U L 49.5 and gM incorporation into the virion may occur without a covalently linked U L 49.5/gM complex and that the uncomplexed U L 49.5 and gM virion incorporation require VP22; (ii) that in the absence of covalently linked U L 49.5/gM complex, U L 49.5 CT residues 80 to 96 are essential.
In alphaherpesviruses, the tegument protein VP22 is known to interact with multiple viral proteins and thereby regulate their cellular translocations (17,18). In HSV-1 (17) and PRV (16,18), VP22 binds to both gE and gM and bridges a complex between gE and gM. Hence, we predicted that VP22 also interacts (i) with U L 49.5 in a gM-independent manner and (ii) with the immature gM. Therefore, VP22 may play a redundant role in U L 49.5 and gM virion incorporation. Additionally, we hypothesized that U L 49.5 CT residues most likely interact with VP22, and this interaction may be essential for gM-independent U L 49.5 virion incorporation. We proved these possibilities in five different ways: (i) by showing that the levels of U L 49.5 C42S/CT-null coimmunoprecipitated by an anti-VP22 antibody from the mutant virus-infected cell lysates is reduced; (ii) by showing that the level of U L 49.5 C42S virion incorporation in a double U L 49.5 C42S/VP22Δ mutant virus is vastly reduced; (iii) by showing that in the absence of gM (gMΔ virus) U L 49.5 is incorporated in the virion; (iv) by showing that the immature gM expressed by the C42S and C42S/CT-null mutant viruses was incorporated into the virion but that the immature gM in the backbone of the C42S/VP22Δ virus was not; and (v) by determining that both the mature and immature gM proteins interacted with VP22 with similar efficiencies.
In summary, these results revealed a redundant role of VP22 in the virion incorporation of wt U L 49.5/gM complex and a novel but essential role for both U L 49.5 C42S and immature gM virion incorporation when they are not covalently linked.
Even though the covalently linked U L 49.5/gM complex was not essential for U L 49.5 or gM virion incorporation, the complex was essential for cell-to-cell spread of virus. Recently, El Kasimi and Lippe reported that U L 49.5 regulates gM translocation and cell-to-cell fusion at the basolateral cell surface (19). In human herpesvirus 6 (HHV-6), gN was required for gM maturation, and the gN-gM complex interacted with v-SNARE protein vesicle-associated membrane protein 3 (VAMP3) in infected cells (21), which is known to facilitate membrane fusion (22). It is noteworthy that while U L 49.5 downregulates MHC-I cell surface expression during BHV-1 infection to evade cellular immune responses, it may also regulate the post-Golgi transport of gM and/or the U L 49.5/gM complex to promote viral cell-to-cell spread and to avoid the circulating neutralizing antibodies. Therefore, in light of the above reports of gN/gM complex in HSV-1 and HHV-6, it could be interesting to determine whether the BHV-1 U L 49.5/gM complex also interacts with v-SNARE, with or without VP22. In conclusion, our U L 49.5 mutational study determined that the U L 49.5/gM functional complex was necessary for efficient cell-to-cell spread of the virus but not for U L 49.5 and gM virion incorporation and MHC-I downregulation. Importantly, the U L 49.5 mutational study revealed a previously unidentified gM-independent novel, functional interaction of VP22 with U L 49.5.

MATERIALS AND METHODS
Cells and wt U L 49.5-expressing cell line. The MDBK cell line was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5 to 10% heat-inactivated fetal bovine serum (FBS). The MDBK cell line expressing wt U L 49.5 was generated as described previously (15) and maintained in DMEM UL49.5/gM Complex and a Novel UL49.5-VP22 Interaction Journal of Virology supplemented with FBS as above, but supplemented additionally with blasticidin as described earlier (15).
Virus and bacterial strains. The BHV-1 Cooper (Colorado-1) strain was obtained from the American Type Culture Collection (ATCC VR-864) and low-passage-number viral stocks were maintained. Reconstituted BHV-1 Cooper BAC-excised virus and BHV-1 U L 49.5 CT-null BAC-excised virus were generated previously (15,23). BHV-1 virus with a deletion of U L 49.5 was a kind gift from E. J. Wiertz (Leiden University, The Netherlands). Infectious BHV-1 wt and BHV-1 U L 49.5 CT-null BAC clones were maintained in Escherichia coli strain DH10B. E. coli strain SW105 (kindly provided by N. G. Copeland) was used for Red recombination.
Construction of mutant viruses. Table 1 Table 2, primer pairs specific for serine substitutions at BHV-1 U L 49.5 residues C42 (C42S), C78 (C78S), and both C42 and C78   Table 4). The 1,191-bp EcoRI/KpnI fragment was then cloned into the corresponding EcoRI/KpnI sites of the pBHV-1 VP22Δ GFP plasmid construct (Fig. 13A) described above. In the resulting plasmid clone, pU L 49.5 C42S/VP22Δ GFP, the UL49.5 C42S mutation (TGC ¡TCG) was incorporated, and the nucleotide sequences coding for VP22 aa 34 to 149 were deleted. The C42S mutation in the plasmid pU L 49.5 C42S VP22Δ GFP was verified by amplifying the entire U L 49.5 ORF by PCR using U L 49.5 upstream and downstream sequence-specific forward and reverse primers (Table 4) and sequencing. Subsequently, a U L 49.5 C42S/VP22Δ GFP virus was generated by homologous recombination of pU L 49.5 C42S/VP22Δ GFP DNA with full-length BHV-1 wt virus DNA. A recombinant double U L 49.5 C42S/VP22Δ GFP mutant virus was plaque purified two times and verified further by sequencing and immunoblotting. Construction of a gM-deleted BHV-1 mutant. The U L 10 gene encoding the envelope glycoprotein gM is transcribed from the complementary strand of the BHV-1 genome and is flanked on the left by U L 11 and U L 12 (3= end) and on the right (5= end) by U L 9 (Fig. 13B). To generate a gM-deleted BHV-1 mutant (BHV-1 gMΔ), a gM deletion vector (pBHV-1 gMΔ) was generated. Briefly, a chimeric 2,100-bplong DNA fragment was synthesized (Genscript) to include the following (5= to 3=): an EcoRI site followed by a 1,000-bp sequence comprising a partial U L 12 ORF, the full U L 11 ORF, the authentic stop codon of the gM ORF (nt 83819 to 84818; GenBank accession number JX898220,), a HindIII site, six additional nucleotides (CCGCGC), and a KpnI site followed by a 1,099-bp sequence comprising partial U L 10 and partial U L 9 ORFs (nt 85188 to 86287) and a BamHI restriction site. This 2,100-bp EcoRI-BamHI fragment was cloned into the corresponding EcoRI/BamHI sites of pUC57 vector (GenScript). In the resulting plasmid, nt 84819 to 85187 coding for gM residues 289 to 411 were deleted and replaced with HindIII/KpnI sites, which allowed insertion of an approximately 2-kb HindIII/KpnI fragment containing an eGFP expression cassette (27), resulting in plasmid pBHV-1gMΔ eGFP (Fig. 13B). BHV-1 gMΔ virus was generated by homologous recombination of pBHV-1 gMΔ eGFP with full-length BHV-1 wt virus DNA. A recombinant BHV-1 gMΔ virus, verified by sequencing and immunoblotting analyses, was selected for further study.
Viral growth kinetics and plaque size determination. One-step growth curve assays were performed twice as described earlier (27). Briefly, for each virus and time point (see below), 20 T25 flasks containing 4 ϫ 10 6 MDBK cells/flask were seeded. The prechilled (4°C) cells were infected with various viruses at a multiplicity of infection (MOI) of 5 and adsorbed for 1 h at 4°C. Following adsorption and washing, 4 ml of medium was added to each flask, and one flask was frozen immediately for each virus sample (0 h) at Ϫ80°C. The remaining flasks were incubated further at 37°C in a CO 2 incubator, and samples were frozen as described above at 3, 6, 12, 18, 24, 30, 36, and 42 h postinfection (hpi). Virus titers at these time points were determined by standard plaque assay as described earlier (28). In plasmid pBHV-1 VP22Δ GFP, coding sequences for amino acids 34 to 149 of VP22 are deleted (nt 9499 to 9845) and two stop codons (uppercase residues in TAAcTGA) and KpnI and BamHI restriction sites are incorporated immediately downstream of the deletion site for the insertion of the eGFP gene cassette. (B) Localizations of the U L 10 gene (gM) and its flanking U L 12, U L 11, and U L 9 genes are shown. In plasmid pBHV-1 gMΔ GFP, nt 84819 to 85187 coding for gM amino acid residues 289 to 411 were deleted, and HindIII and KpnI restriction sites are incorporated at the deletion locus for the insertion of the eGFP gene cassette.
To determine the average plaque size of each mutant, two wells of a six-well plate containing confluent monolayers of MDBK cells or MDBK cells expressing wt U L 49.5 were infected with 80 to 100 PFU of mutant viruses and overlaid with 1.6% carboxymethyl cellulose (CMC) at 2 hpi. At 48 hpi, the cells were fixed (10% formaldehyde) and stained with crystal violet. Average plaque size of wt and mutant viruses was calculated by measuring approximately 50 randomly selected plaques of each virus under a microscope with a graduated ocular objective, as described previously (15).
Radiolabeling of mock-or virus-infected MDBK cell proteins, SDS-PAGE, and immunoprecipitation/immunoblotting analysis. The method for [ 35 S]methionine-cysteine labeling of mock-or virusinfected MDBK cells and immunoprecipitation of virus-specific proteins using protein A-Sepharose/virus protein-specific antibody was described previously (29). For the analysis of gM, virus-infected cell lysates and immunoprecipitates were incubated at 60°C in reducing sample buffer containing 100 mM dithiothreitol (DTT) as described previously (15) and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Unless otherwise mentioned in the figure legend, the SDS-PAGE was performed under reducing conditions. For SDS-PAGE under nonreducing conditions, sample buffer without DTT was used. For all other samples, cell lysates were prepared as described previously (29), and immunoprecipitates were boiled for 5 min in reducing sample buffer containing ␤-mercaptoethanol and separated by SDS-PAGE. Immunoprecipitated/SDS-PAGE-separated proteins were visualized by autoradiography or by immunoblotting as described earlier (15).
EndoH digestion. Endoglycosidase H (EndoH) digestion was performed as described previously (30). The digested samples were subjected to SDS-PAGE, and labeled proteins were visualized by autoradiography.
FACS analysis of MHC-I cell surface expression. MDBK cells either mock infected or infected with BHV-1 wt, U L 49.5 C42S, or C78S mutant virus were collected at 18 hpi, blocked with IgG-free bovine serum albumin (BSA), incubated with mouse anti-bovine MHC-I antibody (Ab), and subsequently stained with FITC-conjugated rat anti-mouse Ab and analyzed by flow cytometry as described previously (15). MDBK cells infected with the respective viruses were stained by FITC-conjugated mouse IgG2a and used as isotype controls.
Statistical analysis. Normality of distribution of the examined variables was evaluated by a D'Agostino-Pearson omnibus normality test. Statistical significance of plaque size variations between the mutant and wt viruses was determined using a one-way analysis of variance (ANOVA) followed by Tukey's multiple-comparison test using Graph Pad Prism (GraphPad Software, La Jolla, CA, USA). A P value of Յ0.05 was considered statistically significant.