Bovine Herpesvirus 4 Modulates Its β-1,6-N-Acetylglucosaminyltransferase Activity through Alternative Splicing

ABSTRACT

Carbohydrates play major roles in host-virus interactions. It is therefore not surprising that, during coevolution with their hosts, viruses have developed sophisticated mechanisms to hijack for their profit different pathways of glycan synthesis. Thus, the Bo17 gene of Bovine herpesvirus 4 (BoHV-4) encodes a homologue of the cellular core 2 protein β-1,6-N-acetylglucosaminyltransferase-mucin type (C2GnT-M), which is a key player for the synthesis of complex O-glycans. Surprisingly, we show in this study that, as opposed to what is observed for the cellular enzyme, two different mRNAs are encoded by the Bo17 gene of all available BoHV-4 strains. While the first one corresponds to the entire coding sequence of the Bo17 gene, the second results from the splicing of a 138-bp intron encoding critical residues of the enzyme. Antibodies generated against the Bo17 C terminus showed that the two forms of Bo17 are expressed in BoHV-4 infected cells, but enzymatic assays revealed that the spliced form is not active. In order to reveal the function of these two forms, we then generated recombinant strains expressing only the long or the short form of Bo17. Although we did not highlight replication differences between these strains, glycomic analyses and lectin neutralization assays confirmed that the splicing of the Bo17 gene gives the potential to BoHV-4 to fine-tune the global level of core 2 branching activity in the infected cell. Altogether, these results suggest the existence of new mechanisms to regulate the activity of glycosyltransferases from the Golgi apparatus.

IMPORTANCE Viruses are masters of adaptation that hijack cellular pathways to allow their growth. Glycans play a central role in many biological processes, and several studies have highlighted mechanisms by which viruses can affect glycosylation. Glycan synthesis is a nontemplate process regulated by the availability of key glycosyltransferases. Interestingly, bovine herpesvirus 4 encodes one such enzyme which is a key enzyme for the synthesis of complex O-glycans. In this study, we show that, in contrast to cellular homologues, this virus has evolved to alternatively express two proteins from this gene. While the first one is enzymatically active, the second results from the alternative splicing of the region encoding the catalytic site of the enzyme. We postulate that this regulatory mechanism could allow the virus to modulate the synthesis of some particular glycans for function at the location and/or the moment of infection.

INTRODUCTION

A lot of the key molecules involved in the innate and adaptive immune responses are glycoproteins (1, 2). These carbohydrates are classified by the nature of their linkage to the protein as either N-glycans or O-glycans. Both N- and O-glycans initiate various biological functions through interaction with receptors. Among O-glycans, mucin-type O-glycans can either be linear or branched (3). The central biosynthetic pathway used to synthesize complex-type O-glycans is through core 2 branching (4). Core 2 branched O-glycans are synthesized by core 2 β-1,6-N-acetylglucosaminyltransferases (Core2β1,6GnT) (5) and are involved in important immune mechanisms, such as T- and B-cell homing (6) or T-lymphocyte differentiation (7).

For millions of years, viruses have been coevolving with their hosts. During this coevolution process, viruses had to deal with the physiology of the host in order to overcome its barriers and defense mechanisms by mimicking or hijacking host biological processes in their favor. A growing list of studies has highlighted the importance of the glycome in the virus life cycle (8) and the ability of some viruses to manipulate the cellular and viral glycome (9). Viruses may modify glycomes by different mechanisms: (i) by regulating the expression of host glycosyltransferases or glycosidases, and/or (ii) by acquiring mutations that affect glycosylation sites or glycan binding specificities, and/or (iii) by encoding their own glycosyltransferases (10) or glycosidases (11). Very few viruses that infect vertebrates are known to encode glycosyltransferases (10, 12); one of these viruses is Bovine herpesvirus 4 (BoHV-4) (13).

BoHV-4 is a gammaherpesvirus which has been isolated throughout the world from healthy cattle as well as from those exhibiting a variety of diseases (14). While the BoHV-4 genome has a reduced set of open reading frames (ORFs) homologous to cellular genes (15, 16), its Bo17 gene encodes a functional homologue of the cellular Core2β1,6GnT type M (C2GnT-M) (13), which was acquired from a recent ancestor of the African buffalo (17). This gene is nonessential for virus replication, despite its contributing to posttranslational modification of virion proteins (18).

In the present study, we pursued the characterization of the BoHV-4 Bo17 gene. Surprisingly, we found that Bo17 undergoes an alternative splicing that generates two different products of expression. In contrast to the full-length form of the protein, the shortened spliced form is devoid of enzymatic activity and confers to BoHV-4 the potential to fine-tune the core 2 branching activity of the infected cell.

MATERIALS AND METHODS

Cells and viruses.

Madin-Darby bovine kidney cells (MDBK; ATCC CCL-22), embryonic bovine lung cells (EBL; DSMZ ACC192), BoMac (bovine macrophages [19]), embryonic bovine trachea cells (EBTr; ATCC CCL-44), bovine turbinate cells (BT; CRL-1390), bovine mammary epithelial cells (Mac-T [20]), Georgia bovine kidney cells (GBK [21]), and Chinese hamster ovary cells (CHO; ATCC CCL-61 [22]) were cultured in Dulbecco's modified Eagle Medium (DMEM; Invitrogen) containing 10% fetal calf serum (FCS), 2% penicillin-streptomycin (Invitrogen), and 1% nonessential amino acids (Invitrogen). Moreover, a primary cell culture of bovine bronchial epithelial (BBE) cells, obtained as previously described (23), was also used. The BoHV-4 V.test strain was initially isolated from a case of orchitis (24). BoHV-4 strains LVR 140, MOVAR 33/63, DN599, 66-P-347, 108, 130, and Buf representative of the BoHV-4 species isolated throughout the world and from different animal species were used in this study (18).

Reverse transcription-PCR (RT-PCR).

Cytoplasmic RNA was isolated from mock-infected and BoHV-4-infected cells 24 h after infection (multiplicity of infection [MOI], 2 PFU/cell). Contaminating DNA was removed by DNase treatment. cDNA was produced by using a Transcriptor high-fidelity cDNA synthesis kit (Roche Applied Science) with a poly(dT) primer. The cDNA products were amplified by PCR with Taq polymerase (New England BioLabs) using primers specific for Bo17. The forward primer 5′-ATGAAGATGGCTGGG-3′ and the reverse primer 5′-TCAAAGTTCAGTCCCATAGAT-3′ were used to amplify the entire coding sequence of Bo17 (primers ATG-STOP). The same PCRs were performed directly on cytoplasmic RNA to verify the absence of contaminant DNA.

Antibodies.

Anti-Bo17 422-441 polyclonal monospecific antibodies were produced as follows. On day 0, equal volumes of diluted keyhole limpet hemocyanin (KLH) peptide (amino acids 422 to 441 of the BoHV-4 V.test strain Bo17 protein, at 1 mg/ml) and Freund's complete adjuvant were emulsified and injected subcutaneously into rabbits at three different sites (200 μg/rabbit). On days 14, 28, 42, 56, and 70, each rabbit was immunized again with 100 μg of KLH peptide (1 mg/ml) emulsified with incomplete Freund's adjuvant. Serum was collected on day 77. Specificity of the immune sera against the Bo17 422-441 peptide was tested by enzyme-linked immunosorbent assay (ELISA), described previously (25). Briefly, Nunc Maxisorp ELISA plates (Nalge Nunc, NY) were coated (18 h, 4°C) with unfused Bo17 422-441 peptides, blocked in phosphate-buffered saline (PBS)–0.1% Tween 20–1% bovine serum albumin (BSA), and incubated with polyclonal antibody (pAb) sera. Bound antibodies were detected with horseradish peroxidase-conjugated goat anti-rabbit IgG pAb (Sigma). The plates were washed five times in PBS–0.1% Tween 20 after each step. The detection substrate was nitrophenylphosphatase (Sigma). Absorbance was read at 405 nm using a Benchmark ELISA plate reader (Thermo). For detection of gp180 protein (encoded by the Bo10 gene) via Western blotting, we used a rabbit monospecific polyserum raised against the C-terminal end of the protein (anti-Bo10-c15) (26). For some Western blot assays, we used serum of a rabbit infected intravenously with 10⁸ PFU of the BoHV-4 V.test strain and collected 63 days postinoculation (27).

Western blotting.

Infected cells or purified virions were lysed and denatured by heating (95°C, 5 min) in SDS-PAGE sample buffer (31.25 mM Tris-HCl [pH 6.8], 1% [wt/vol] SDS, 12.5% [wt/vol] glycerol, 0.005% [wt/vol] bromophenol blue, 2.5% [vol/vol] 2-mercaptoethanol). Proteins were resolved by electrophoresis on Mini-Protean TGX (Tris-glycine extended) precast 4 to 15% resolving gels (Bio-Rad) in SDS-PAGE running buffer (25 mM Tris base, 192 mM glycine, 0.1% [wt/vol] SDS) and transferred to polyvinylidene difluoride membranes (Immobilon-P transfer membrane; 0.45-μm pore size; Millipore). The membranes were blocked with 3% nonfat milk in PBS–0.1% Tween 20 and then incubated with anti-Bo17-422-441 or anti-Bo10-c15 rabbit antibodies or anti BoHV-4 rabbit polyserum in the same buffer. Bound antibodies were detected with horseradish peroxidase-conjugated goat anti-rabbit IgG pAb (Dako Corporation), followed by washing in PBS–0.1% Tween 20, development with enhanced chemiluminescence substrate (GE Healthcare) and exposure to X-ray film.

In vitro enzymatic assay.

To measure the activities of both forms of pBo17, soluble recombinant forms were generated as previously described (13). The sequences of pBo17 and pBo17 Spliced without their transmembrane regions were amplified by primers Bo17 soluble sense (5′-CCGGATCCCCTCAGGTTTAAATGTGATGTAG-3′; BamHI site in bold and nucleotides 99 to 121 of Bo17 in italics) and Bo17 soluble Rev (5′-CCATGCATTCAAAGTTCAGTCCCATAGATGG-3′). The fragments obtained were T/A cloned into the pGEM-T Easy vector (Promega Corporation) and then digested with BamHI and SpeI and inserted into pcDNA3-A opened with BamHI and XbaI. pcDNA3-A is a pcDNA3.1-derived vector harboring a sequence encoding a signal peptide and the IgG binding domain of protein A (28). Enzymatic assays were performed as described previously (28). Briefly, these enzyme assays were carried out with pcDNA3-A pBo17, pcDNA3-A pBo17 Spliced, or pcDNA3-A h-C2GnT-M that had been transiently transfected into CHO cells. The soluble chimeric enzymes were then adsorbed to IgG-Sepharose from the concentrated cell culture supernatant and used as the enzyme source for the assays. As acceptors, the reaction products Galβ1 → 3GalNAcα → p-nitrophenol and GlcNAcβ1 → 3GalNAcα → p-nitrophenol were used for assaying C2GnT and C4GnT activities, respectively. Galβ1 → 4GlcNAcβ1 → 3Galβ1 → 4GlcNAcβ1 → 6Manα-1 → 6Man-β → octyl and GlcNAcβ1 → 3Galβ-1→ 4GlcNAcβ1 → 6Manα-1 → Manβ1→ octyl were used for assaying IGnT activity. Because addition of N-acetylglucosamine occurs at the underlined galactose residues, the two acceptors serve for centrally acting IGnT (cIGnT) and predistally acting IGnT (dIGnT), respectively. CHO cells transfected with pcDNA3-A vector were used as a negative control. The radioactivity obtained with the negative control was subtracted from the values obtained in experiments with the enzymes. When detected, the activities were expressed relative to C2GnT activity, which was defined as 100%.

Production of BoHV-4 Bo17 mutant strains.

The BoHV-4 V.test strain deleted for the Bo17 ORF (Bo17 Del) and the corresponding revertant strain (Bo17 Rev) have been described previously (18). The recombinant strains Bo17 MuDir and Bo17 Spliced were derived from a cloned BoHV-4 bacterial artificial chromosome (BAC) (29). First, the Bo17 MuDir sequence was generated by site-directed mutagenesis as follows. Two mutations were inserted in the 5′ splice site and one mutation in the 3′ splice site. Briefly, a first PCR with sense primer HindIII sense.native (5′-GAAGAAGCTTTGCCCGGGAC-3′; native HindIII restriction site in bold) and reverse primer MuDir XhoI (5′-GGCTCGAGAGGCCACAAAATAGGCATTCCCGGTGAAC-3′; native XhoI restriction site in bold and mutation in the sequence that was incorporated shown in bold underline) allowed us to obtain a mutation in the 3′ splicing site. This fragment was cloned into the vector pGEM-T Easy/EcoRII (18) instead of the wild-type (WT) sequence via HindIII and XhoI restriction sites. To introduce two other mutations at the 5′ splicing site, a second PCR was performed with the sense primer HindIII sense.native and the reverse primer MuDir KpnI (5′-GCGGTACCTCAGACTCCATACTGTTTTTGCCCTTCAAC-3′, KpnI restriction site in bold and two mutations in the sequence in bold underline). This second fragment was then HindIII/KpnI restricted and cloned into the corresponding sites in pGEM-T Easy/EcoRII already mutated for the 3′ splicing site. We obtained a pGEM-T Easy with the Bo17 MuDir sequence. The Bo17 Spliced sequence was directly generated from the spliced product obtained by amplification of Bo17 with the ATG-STOP primers on cDNA from BoHV-4-infected cells (Fig. 1A). The spliced band was then HindIII/XhoI restricted and cloned into the corresponding sites of the pGEM-T Easy/EcoRII. Mutagenesis of the WT V.test strain BAC G plasmid was performed by a two-step mutagenesis procedure in bacteria via the shuttle plasmid pST76KSR (29). Plasmids used to induce homologous recombination were constructed as follows. Each construct was subcloned as a HindIII/XhoI fragment into the same sites of the pST76K-SR shuttle vector and recombined into the BoHV-4 BAC (29). We also isolated, in the same way, revertants of the Bo17 MuDir and Bo17 Spliced BACs, in which the Bo17 locus was restored to its WT form. The correct construction of each mutant was confirmed by DNA sequencing. Reconstitution of infectious virus from BAC plasmids was obtained by transfection in MDBK cells.

FIG 1.

Bo17 mRNA undergoes alternative splicing. (A) RT-PCR analysis results for the BoHV-4 Bo17 reading frame. MDBK cells were infected with BoHV-4 V.test strain (MOI of 0.5 PFU/cell). At 24 h postinfection, expression of Bo17 was studied by using primers amplifying the entire Bo17 coding sequence (primers ATG-STOP, 1,323 bp). The same PCR was performed directly on cytoplasmic RNA to verify the absence of contaminant DNA. Sizes are indicated on the left. (B) The two PCR products obtained by RT-PCR were purified and sequenced. Then, alignments were performed using CLUSTAL W. The first sequence corresponds to the entire coding sequence of the Bo17 gene. The second sequence displayed a disappearance of 138 bp potentially generated by splicing. The gray boxes indicate the putative splice donor and acceptor sites. The sequence in red indicates the intron. (C) MDBK, EBL, BOMAC, EBTr, BT, BBE, MAC-T, and GBK cells were infected with the BoHV-4 V.test strain (MOI of 0.5 PFU/cell). At 24 h postinfection, expression of Bo17 was studied using primers amplifying the entire Bo17 coding sequence (primers ATG-STOP, 1,323 bp). The same PCR was performed directly on cytoplasmic RNA to verify the absence of contaminant DNA (data are only shown for the MDBK sample). Sizes are indicated on the left.

Southern blotting.

Southern blot analysis (30) of viral DNA digested with EcoRI was performed with a probe corresponding to nucleotides 26 to 774 of the V.test strain Bo17 ORF.

Growth curves.

The growth kinetics of mutant and revertant viruses were compared to those of the WT as described previously (31). Cell cultures were infected at an MOI of 0.01 (multistep assay). After 1 h of adsorption, the cells were then washed and overlaid with DMEM containing 10% FCS. Supernatants of infected cultures and infected cells were harvested at successive intervals, and the amount of infectious virus was determined by plaque assay on MDBK cells. Plaques were visualized after immunostaining with monoclonal antibody (MAb) MAb35 recognizing gB (27).

Virus purification.

BoHV-4 virions grown on MDBK cells were purified as described previously (32). Briefly, after removal of the cell debris by low-speed centrifugation (1,000 × g, 10 min), virions present in the infected cell supernatant were harvested by ultracentrifugation (100,000 × g, 2 h) through a 30% (wt/vol) sucrose cushion, then centrifuged through two successive 20 to 50% (wt/vol) potassium tartrate gradients in PBS (100,000 × g, 2 h). Virions were finally washed and concentrated in PBS (100,000 × g, 2 h).

O-Glycomic profile analysis.

MDBK cells (3 × 10⁶ cells) were infected with Bo17 MuDir, Bo17 MuDir Rev, Bo17 Spliced, or Bo17 Spliced Rev strains at an MOI of 1 PFU per cell. At 36 h postinfection, cells were harvested, lysed in PBS containing 0.5% (vol/vol) SDS, and treated as described previously (33). Briefly, all samples were subjected to homogenization in an extraction buffer (25 mM Tris, 150 mM NaCl, 5 mM EDTA, and 1% 3-[3-cholamidopropyl dimethyl-ammonio]-1-propane sulfonate [CHAPS] at pH 7.4). After reduction and carboxymethylation, samples were sequentially digested with trypsin (Sigma-Aldrich) and PNGase F (Roche). N-Glycans were separated from peptides and/or glycopeptides on Sep-Pak cartridges (Waters, Hertfordshire, United Kingdom), and O-glycans were released from the latter by reductive elimination. After purification on Dowex 50W-X8 (H) 50- to 100-mesh columns (Sigma-Aldrich), the O-glycan samples were permethylated and then further purified with Sep-Pak cartridges. O-Glycans were eluted in aqueous acetonitrile fractions and then lyophilized. Permethylated samples were dissolved in 10 μl of methanol, and 1 μl of the dissolved sample was premixed with 1 μl of matrix (20 mg/ml 2,5-dihydroxy-benzoic acid in 70% [vol/vol] aqueous methanol) and spotted onto a target plate. Matrix-assisted laser desorption ionization (MALDI) MS and tandem MS (MS-MS) data were acquired using a 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems) in positive ion mode. The collision energy was set to 1 kV, and argon was used as collision gas. The data obtained were viewed and processed using Data Explorer 4.9 (AB Sciex UK Ltd.). Comparison of the relative abundance of core 2 and core 1 O-glycan (core 2/core 1 ratio) was done based on the total relative intensity of the core 2 peaks and the total relative intensity of the core 1 peaks, with the maximum peak being defined as 100%. A core 2/core 1 ratio was obtained for each replicate, and mean values were finally compared between samples.

Details of experimental data utilized to define glycan structures.

To define glycan structures, (i) all O-glycans were assumed to have a reducing-end GalNAc, based on known biosynthetic pathways. (ii) Monosaccharide compositions in terms of numbers of Hex, HexNAc, etc., molecules were derived from the MALDI-MS results, run in the positive ion mode of molecular ions of reductively eliminated permethylated species, which were assigned manually. (iii) MALDI-TOF/TOF MS/MS fragmentation of the following molecular ions was performed for m/z 534, 779, 895, 983, 1,187, 1,256, 1,344, 1,548, and 1,705. Fragment ions were identified manually and with the assistance of the Glycoworkbench tool (version 1.2). Details provided were guided by using Mirage (34).

Lectin neutralization assay.

Artocarpus integrifolia lectin (jacalin; recognizing core 1 structures) was purchased from Vector Laboratories. Briefly, virus was incubated (2 h, 37°C) with jacalin at various concentrations, and the amount of infectious lectin-treated virus was determined by plaque assay on MDBK cells.

RESULTS

Bo17 mRNA undergoes alternative splicing.

The initial DNA sequence analysis of the BoHV-4 genome demonstrated that the Bo17 gene is highly homologous to the cellular C2GnT-M (13). Then, RT-PCR experiments identified a product of expression corresponding to the full-length coding sequence of the gene and encoding a functional β1,6-GnT that formed core 2, core 4, and I structures (13). Expression of this active Bo17 product was confirmed for all available BoHV-4 strains by RT-PCR and a subsequent enzymatic activity assay (18).

To investigate deeper the transcription of this Bo17 message during the BoHV-4 cycle, we repeated the RT-PCR on mRNA from BoHV-4-infected MDBK cells with another set of primers located at both ends of the Bo17 open reading frame (ATG-STOP primers). As previously described, a product corresponding to the expected size of a full-length Bo17 mRNA was observed. Surprisingly, a second band was also visible (Fig. 1A). DNA sequencing revealed that this band corresponded to the BoHV-4 Bo17 gene sequence but displayed the disappearance of 138 bp (Fig. 1B), suggesting that Bo17 mRNA could undergo alternative splicing. We therefore inspected both ends of the 138-bp deleted sequence for potential splice donor and acceptor sites. By comparing consensus splice donor and acceptor sequences (35) with these Bo17 sequences, we observed that the 5′ GT and 3′ AG as well as other important nucleotides were present at both ends of the 138-bp deleted sequence (Fig. 1B). Finally, as the splicing patterns for a single gene in different cells are quite different, we tested Bo17 splicing in cells of different origins, including kidney (MDBK, GBK), mammary epithelium (MAC-T), respiratory epithelium (EBL, EBTr, BT, BBE), and macrophage (BoMac). Interestingly, Bo17 splicing was observed in all tested cells (Fig. 1C). Altogether, these results show that BoHV-4 Bo17 mRNA undergoes alternative splicing in cells of various origins, at least those of the V.test strain.

Bo17 splicing is conserved among BoHV-4 strains but not in its cellular counterparts.

To estimate the conservation of these splice donor and acceptor sites, we compared sequences of eight contemporary BoHV-4 strains representative of different phylogenetic clades (17). This comparison revealed that both the splice donor and the splice acceptor sites are conserved in these BoHV-4 strains (Fig. 2A). Moreover, RT-PCR on mRNA from MDBK cells infected with these eight different strains confirmed that two expression products are encoded by the Bo17 genes of all tested strains (Fig. 2B). In contrast, sequence analysis of cellular C2GnT-M from different animal species showed that none of these sequences displayed both of these consensus regions (Fig. 2A). It has to be noted that the weaker band observed in the mock-infected cells corresponds to the cellular C2GnT-M. Indeed, the BoHV-4 Bo17 gene displays 94.6% identity with the gene encoding bovine C2GnT-M, and the primers used to amplify the entire sequence of Bo17 have only one mismatch (at base 10 of the forward primer) with the cellular C2GnT-M sequence. It has to be noted that, as expected, this gene does not undergo alternative splicing of its mRNA. However, based on this homology, it was not possible to set up a quantitative RT-PCR assay in order to estimate the amounts of the different transcripts.

FIG 2.

The Bo17 gene displays consensus splice sites that are conserved among BoHV-4 strains but not in their cellular counterparts. (A) The consensus sites for Bo17 splicing (GT and AG in boxes) are conserved in all the BoHV-4 strains that were available. Sequences of cellular, homologous β-1,6-N-acetylglucosaminyltransferases from different species were aligned with Bo17 sequences of BoHV-4 strains. The cellular sequences did not contain the consensus sites for splicing. Nucleotides that are not conserved are highlighted in red. (B) MDBK cells were infected with different strains of BoHV-4 (MOI, 0.5 PFU/cell). At 24 h postinfection, expression of Bo17 was studied by RT-PCR using primers amplifying the entire Bo17 coding sequence (primers ATG-STOP, 1,323 bp). The same PCRs were performed directly on cytoplasmic RNA to verify the absence of contaminant DNA (data not shown). The weak band observed in the mock-infected sample corresponds to the expression of cellular C2GnT-M. Sizes are indicated on the left.

The BoHV-4 Bo17 gene encodes two proteins.

In order to test if both of these mRNA products encode proteins, Bo17-specific polyclonal antibodies were generated by immunizing rabbits with a KLH-conjugated peptide (VLQCLEEYLRHKAIYGTEL) corresponding to the shared C-terminal end of the predicted Bo17 gene products (amino acids 422 to 441) (Fig. 3A). Immunization against the peptide was confirmed by ELISA (Fig. 3B). To identify proteins recognized by the anti-Bo17–422-41 serum, Western blotting was performed on WT-, Bo17 Del-, or Bo17 Rev-infected MDBK cells. In BoHV-4 WT and Bo17 Rev samples, proteins of apparent molecular masses (MM) of 52 kDa and 45 kDa were detected under reducing conditions (Fig. 3C). No such proteins were detected in cells infected by the BoHV-4 Bo17 Del strain. The predicted MM of the full-length Bo17 gene product was 50.8 kDa (estimated by the Compute pI/MW tool of the Expasy server), so this result suggests that the full-length Bo17 mRNA encodes the observed 52-kDa protein, here called pBo17 full length, and that the spliced Bo17 mRNA encodes a 45-kDa protein, here called pBo17 Spliced.

FIG 3.

Two proteins are encoded by the BoHV-4 Bo17 gene. (A) Amino acid comparisons of the spliced and unspliced Bo17 products. The gray box indicates the predicted transmembrane region of both Bo17 expression products. Putative N-glycosylation sites are highlighted with asterisks. Amino acids 422 to 441, used to immunized rabbits, are highlighted in a black box. (B) Immunization of rabbits with the Bo17–422-441 peptide elicited antibody response against the peptide. The antipeptide antibody responses were determined for the two immunized rabbits via an anti-Bo17 422-441 ELISA as described in Materials and Methods. (C) Detection of specific BoHV-4 proteins by the anti-Bo17–422-441 serum. MDBK cells were infected with the WT V.test strain or the Bo17 Del or the Bo17 Rev strain of BoHV-4 (1 PFU/cell). After 48 h, cells were scraped off the plate, and Western blotting was carried out as described in Materials and Methods. The positions of molecular mass (MM) standards are shown to the left of the gel. The open and filled triangles indicate the two proteins encoded by the BoHV-4 Bo17 gene.

The spliced form of pBo17 is not enzymatically active.

The impact of the splicing of pBo17 on its enzymatic activity was first assessed by aligning the sequences for different mammalian C2GnT-L and C2GnT-M proteins with the sequence of pBo17 full length and pBo17 Spliced (Fig. 4A). While five of the six amino acids found to be involved in direct acceptor substrate interactions (36) are conserved in pBo17 full length, the splicing of Bo17 affected three (E243, K251, and R254 [Fig. 4A, green asterisks]) of these five amino acids (Fig. 4A, other two amino acids identified by magenta asterisks) playing a role in acceptor substrate. In particular, the K251 and R254 residues were found to make key hydrogen bonds and a van der Waals interaction with the bound disaccharide acceptor (37). Thus, alternative splicing removes key amino acids involved in the acceptor substrate binding site and could therefore affect the activity of pBo17. The three-dimensional (3D) structure of pBo17 (Fig. 4B) was then estimated on the basis of the X-ray crystal structure of the murine C2GnT-L (36) in order to better visualize the spliced region (Fig. 4B, in red) and the three amino acids involved in catalytic activity that disappear after splicing (Fig. 4B, in green).

FIG 4.

The spliced form of the pBo17 enzyme is inactive. (A) Multiple-sequence alignment of β-1,6-N-acetylglucosaminyltransferases involved in core 2 O-glycan biosynthesis. Residues that are absolutely conserved among murine C2GnT-L, human C2GnT-L, human C2GnT-M, bovine C2GnT-M, pBo17, and pBo17 Spliced sequences are highlighted in gray. The murine C2GnT-L secondary structure (36) is shown above the alignment and is colored to highlight the putative stem region (gray) and the catalytic domain (cyan). In the catalytic domain, the region without pBo17 Spiced is highlighted in red. Residues forming direct interactions with the acceptor substrate are marked with an asterisk. Green asterisks show residues that disappear with the splicing, and purple asterisks show those that are not affected by the splicing. (B) The 3D structure of pBo17 was generated by using SWISS-MODEL and visualized with ICM-Browser according to the colors described above. (C) Soluble chimeric forms of h-C2GnT-M, pBo17, and pBo17 Spliced were assayed for C2GnT (C2), C4GnT (C4), cIGnT (cI), and dIGnT (dI) activities as described in Materials and Methods. The soluble form of pBo17 full-length has core 2, core 4, and I branching activities, whereas the pBo17 Spliced form has none of these activities.

Finally, we tested the activity of the pBo17 Spliced protein in an in vitro enzymatic assay. As described previously, the full-length form of pBo17 was shown to have core 2, core 4, and I branching activities (13), as does its cellular counterpart (28). In contrast, the chimeric pBo17 spliced protein did not exhibit core 2, core 4, or I activities (Fig. 4C).

Generation of the Bo17 MuDir and Spliced BoHV-4 mutants.

A BoHV-4 Bo17 knockout strain has previously been described (18). To unravel the function of both the spliced and the unspliced Bo17 expression products, we generated two additional Bo17 mutant viruses. In the Bo17 MuDir mutant, we punctually mutated the Bo17 splicing donor site (T to C) and the Bo17 splice acceptor site (A to C) in order to only express the unspliced form (pBo17 full length) without affecting its amino acid sequence. On the opposite, the Bo17 Spliced strain was generated in order to only encode the Bo17 spliced sequence (Fig. 5A). Enzymatic restriction of viral DNA and subsequent Southern blotting (Fig. 5B) confirmed the expected genomic structures of the two mutants and of the revertants of both these strains. The expected mutations were further confirmed by DNA sequencing (data not shown). As expected, RT-PCR analysis showed that the Bo17 MuDir and the Bo17 Spliced strains expressed the unspliced and the spliced mRNA, respectively (Fig. 5C). Finally, to confirm that the two Bo17-encoded proteins observed in BoHV-4 WT-infected cells are encoded by the two Bo17-derived mRNAs, we performed Western blotting on lysates of cells infected with these different Bo17 mutants. Immunoblotting with the anti-Bo17–422-441 rabbit polyserum confirmed that the Bo17 MuDir viral strain only expresses pBo17 full length and that the Bo17 Spliced strain only expresses pBo17 spliced, while WT and revertant strains expressed both (Fig. 5D). As expected, the Bo17 Del strain expressed none of the Bo17-encoded proteins.

FIG 5.

Generation of the Bo17 MuDir and Spliced BoHV-4 mutants. (A) Schematic representation of the strategies followed to produce the recombinant BoHV-4 strains. The recombinant strains Bo17 MuDir and Bo17 Spliced were derived from a cloned BoHV-4 BAC by the aid of pST76KSR plasmid. First, the Bo17 MuDir sequence was generated by site-directed mutagenesis. Two mutations were inserted in the 5′ splice site and one mutation was inserted in the 3′ splice site (the two consensus sites are highlighted by gray boxes). The mutated nucleotides are highlighted in red. The Bo17 Spliced sequence was directly generated from the spliced product obtained by amplification with Bo17 ATG-STOP primers on cDNA from BoHV-4 infected cells. The Bo17 MuDir mutant only encodes the long form (enzymatically active) of pBo17 with no amino acid substitution. The Bo17 Spliced mutant only encodes the spliced form of pBo17 (inactive). (B) Verification of the molecular structure. The DNAs of the different BoHV-4 strains were analyzed by EcoRI restriction (left panel) and further tested by Southern blotting using a probe corresponding to nucleotides 26 to 774 of the V.test Bo17 ORF (right panel). The white triangle and white star indicate the restriction fragments containing the enhanced green fluorescent protein ORF and the BAC cassette, respectively. The black triangle shows the restriction fragment that contains Bo17 Spliced. Marker sizes are indicated on the left. (C) RT-PCR analysis of BoHV-4 Bo17 expression by the different mutants. MDBK cells were infected with different mutants (MOI, 0.5 PFU/cell). At 24 h postinfection, expression of Bo17 was studied using primers specific for the Bo17 sequence (primers ATG-STOP, 1,323 bp). Only the entire transcript or the spliced transcript was detected for the recombinant Bo17 MuDir virus and Bo17 Spliced virus, respectively. RT reactions were executed (+) or omitted (-) to verify the absence of contaminant DNA. The weak band observed in the Bo17 Del sample corresponds to expression of the cellular C2GnT-M. Sizes are indicated on the left. (D) Detection of specific BoHV-4 protein by the anti-Bo17–422-441 serum. MDBK cells were infected with the different BoHV-4 strains as indicated (1 PFU/cell). Forty-eight hours later, cells were scraped off the plate, and Western blotting was carried out as described in Materials and Methods. The positions of molecular mass (MM) standards are shown on the left of the gel. The open and filled triangles indicate the bands containing, respectively, pBo17 full length and pBo17 Spliced.

Effect of Bo17 mRNA splicing on BoHV-4 replication in vitro.

We previously showed that the Bo17 knockout strain does not display any growth deficit in vitro. However, sole expression of either of the two Bo17-encoded proteins could influence BoHV-4 growth. To address this question, multistep growth assays were performed on MDBK, EBTr, and BT cells with the different Bo17 mutants. None of the Bo17 mutants displayed any in vitro growth deficit in any of these different cell lines (Fig. 6A to C).

FIG 6.

Replication of BoHV-4 MuDir and Spliced strains in vitro. MDBK (A), EBTr (B), or BT (C) cells grown in 6-well cluster dishes were infected at an MOI of 0.05 in a multistep assay as described in Materials and Methods with different BoHV-4 strains. Supernatants of infected cultures and infected cells were harvested at different times (days) postinfection, and the amount of infectious virus was determined by plaque assay on MDBK cells. Time zero postinfection data are from retitration of the inocula to ensure that similar amounts of virus were put on the cells. Plaques were visualized by immunofluorescence staining as described in Materials and Methods.

Effect of BoHV-4 infections on core 2 O-glycan content of infected cells.

To identify the influence of the expression of the different forms of pBo17 on the glycomes of infected cells, we characterized the structures of O-glycans isolated from cells mock infected or infected with the Bo17 Del, Bo17 MuDir, Bo17 MuDir Rev, Bo17 Spliced, and Bo17 Spliced Rev strains of BoHV-4. First, MDBK cells were infected or not with the different Bo17 mutants at an MOI of 1 for 36 h. Then, these infected cells were processed through a glycomic workflow. Consistent and reproducible data were obtained from two repeats of experiments. Assignment of core 1 or core 2 structures were based on MS/MS fragmentation patterns. In comparison with the mock- and Bo17 Del-infected samples, the results revealed an increase in core 2 O-glycan structures detected in Bo17 MuDir-, MuDir Rev-, and Spliced Rev-infected cells (Fig. 7A to F). For instance, the peaks at m/z 1,187.5 and 1,548.7 had higher relative abundances in those cells than in mock- or Bo17 Del-infected cells. In order to quantify these differences, the core 2/core 1 ratios were calculated (Fig. 7G). While the mean core 2/core 1 ratios were 0.075 and 0.07 in mock- and Bo17 Del-infected cells, respectively, the mean ratios observed in Bo17 MuDir-, MuDir Rev-, and Spliced Rev-infected cells were significantly higher, indicating that there are proportionally more core 2 O-glycan structures in those cells. In contrast, the mean core 2/core 1 ratio measured in Bo17 Spliced-infected cells did not differ significantly from those of mock- and Bo17 Del-infected cells (Fig. 7G). Altogether, these results show that the expression of either the full-length pBo17 or the inactive spliced pBo17 determines the global core 2 branching activity of BoHV-4 infected cells.

FIG 7.

MALDI-TOF MS profile of O-glycans from MDBK cells infected with BoHV-4 mutant strains. (A to F) We determined the glycomic profiles of reduced and permethylated O-glycans detected in MDBK cells infected with Bo17 MuDir (A), Bo17 MuDir Rev (B), Bo17 Spliced (C), or Bo17 Spliced Rev (D) BoHV-4 mutants. Structural assignments were based on MS and MS/MS data and knowledge of O-glycan biosynthetic pathways. These data are representative of two independent experiments. (G) The relative intensities of the core 1 and core 2 peaks were calculated relative to the core 1 peak at m/z 895, defined as 100%. The core 2/core 1 ratio corresponds to the total of relative intensities of the core 2 peaks divided by the total of relative intensities of the core 1 peaks. These ratios were calculated for each replicate of each sample. The data presented are the averages ± standard deviations for 2 measurements and were analyzed by Welsh's t test relative to the mean value obtained for the mock-infected sample. **, P < 0.01; *, P < 0.05; ns, nonsignificant.

Effect of Bo17 splicing on BoHV-4 virion O-glycosylation.

It has been shown that Bo17 expression contributes to posttranslational modifications of BoHV-4 structural proteins (18). Sequence analysis using the NetOGlyc 4.0 server (38) revealed that in the BoHV-4 strain, most of the potential virions O-glycans (>98%) are located on gp180 (∼70%) and on glycoprotein B (gB) (∼28%). This was confirmed by mass spectrometry analysis (32). Moreover, we showed that O-glycans associated with gp180 account for most of the sensitivity to neutralization by the O-glycan-specific lectin jacalin (26). We therefore investigated how Bo17 expression and splicing affect glycosylation of gp180 in virions. Briefly, we purified virions of the different strains on tartrate gradients. Then, we subjected these virion preparations to Western blotting with a polyserum recognizing gp180 (39). Interestingly, while we observed a clear band at approximately 180 kDa in WT, Bo17 Rev, Bo17 MuDir, Bo17 MuDir Rev, and Bo17 Spliced Rev virion samples, the apparent molecular mass of the protein was reduced in the Bo17 Del and the Bo17 Spliced samples (Fig. 8A).

FIG 8.

Effect of Bo17 splicing on BoHV-4 glycosylation. (A) Purified virions were subjected to Western blotting with a rabbit monospecific polyserum raised against the gp180 glycoprotein (anti-Bo10-c15 serum) or whole BoHV-4 virions. The glycoprotein gp180 has a lower molecular mass for the viruses devoid of the enzyme Bo17 (Bo17 Del) as well as for the viruses expressing an inactive form of pBo17 (Spliced). The position of an MM standard is shown on the left. (B) Sensitivity of different BoHV-4 mutants to jacalin-mediated neutralization. BoHV-4 mutant strains were incubated (2 h, 37°C) with increasing amounts of jacalin (core 1 O-glycan-specific lectin) and then assayed for infectivity on MDBK cells. BoHV-4 titers are expressed relative to virus without lectin. The data presented are the averages ± standard errors of the means for 3 measurements and were analyzed by 2-way analysis and Bonferroni posttests. ***, P < 0.001.

A lectin neutralization assay is another strategy to highlight the type of O-glycans present at the virion surface of different mutant strains. As described previously (26), we used the lectin from jackfruit seeds (jacalin). This lectin binds to a GalNAcα1 peptide, in which the C-6 OH of αGalNAc is free (i.e., core 1), whereas it cannot recognize a GalNAcα1 peptide with a substitution at the C-6 position (i.e., core 2) (40). Jacalin can block entry of another gammaherpesvirus, Murid herpesvirus 4, by steric hindrance (41), especially as the amount of core 1 structures is important at the surface of the virion. Therefore, we tested the capacity of jacalin to inhibit MDBK infection by the different mutant viruses. Although the infection by WT, Bo17 MuDir, and the revertant strains was moderately inhibited by jacalin, this neutralization was significantly stronger for the Bo17 Del strain and even more for the Bo17 Spliced strain (Fig. 8B). Altogether, these results show that Bo17 expression and splicing may affect the structure of O-glycans on BoHV-4 virions and in particular on glycoprotein gp180.

DISCUSSION

During coevolution with their hosts, viruses have developed sophisticated mechanisms in order to hijack some important cellular functions to promote infection. Deciphering these interactions could not only allow a better understanding of the virus biology but also could reveal unknown cellular pathways, as previously shown (42). Polysaccharides play central roles in many important biological processes. It is therefore not surprising that several viruses have exploited glycosylation to their benefit (9).

Protein O-glycosylation involves the addition of monosaccharides in a stepwise process requiring no glycan template. Thus, each unique oligosaccharide structure reflects the sequential order in which multiple enzymes act in glycan synthesis. Control of the synthesis of specific carbohydrate structures is therefore achieved by regulating the expression of the enzymes required for their construction. In many cases, this is accomplished by controlling the expression of key enzymes. Interestingly, BoHV-4 encodes a functional homologue of cellular C2GnT-M (13), which is a central enzyme for the synthesis of complex-type O-glycans (4). Many biologically significant oligosaccharide structures are constructed on this branch, including some important cell recognition and adhesion molecules involved in immune responses (3, 5, 43). Expression of this enzyme is therefore likely important in the BoHV-4 infection cycle.

In this study, we showed that, in contrast to its cellular counterparts, BoHV-4 modulates its C2GnT-M activity through alternative splicing (Fig. 1 to 4), with changes of the glycome composition not only at the global cellular level (Fig. 7) but also at the viral level (Fig. 8). The splicing of Bo17 was observed in cells of various origins, including kidney (MDBK, GBK), mammary epithelium (MAC-T), respiratory epithelium (EBTr, BT, EBL, and BBE), and macrophage (BoMac) cells, suggesting that in vivo expression of the spliced form is very likely. Interestingly, we showed that the consensus sites for splicing are conserved across all the available BoHV-4 strains (Fig. 2), but not in the cellular homologues from different species among which the C2GnT-M gene of Syncerus caffer, which is phylogenetically the closest (17). This therefore indicates a strong selection pressure for the capacity of Bo17 splicing. Interestingly, Markine-Goriaynoff et al. pointed out that, in comparison with cellular homologues, the molecular clock of the viral sequences increased (17). In contrast, among the viral sequences, the rate of evolution was similar to what is observed among the cellular sequences. They hypothesized that this could either reflect a relaxation of the constraint to preserve the protein sequence or, more interestingly, a positive selection for critical changes in the sequence after gene acquisition. Acquisition of the capacity to undergo splicing is certainly one of these changes. A central question about this mechanism is why BoHV-4 uses alternative splicing of the Bo17 gene instead of regulating the expression of an active form of pBo17. One of the reasons could be that the pBo17 spliced protein, though enzymatically inactive, has its own function.

For many glycosyltransferases, the information that instructs Golgi complex localization is found within a relatively short sequence of amino acids in the N termini of these proteins comprising the cytoplasmic tail, the transmembrane-spanning region, and the stem region (CTS) (44, 45). Moreover, it has been observed that glycosyltransferases form homo- or hetero-oligomers that are too large to be incorporated in recycling vesicles (46). As the pBo17 Spliced CTS is identical to the CTS of the full-length form (Fig. 1B) and highly homologous to that of the cellular C2GnT-M (13), pBo17 Spliced could replace its active counterparts in Golgi complex glycosyltransferases rafts. Expression of the spliced inactive form could therefore have a dominant negative effect. As C2GnT-M is a central player in complex O-glycan biosynthesis, alternative splicing of Bo17 would allow the virus to affect either positively or negatively the kind of carbohydrate produced by the cell. This effect could be regulated over time or in some cellular contexts. Unfortunately, our glycomic analysis was not able to highlight such an effect, probably due to the very low content of core 2 O-glycan structures in mock-infected MDBK cells (Fig. 7). However, this hypothesis might explain why the sensitivity of virions to neutralization by jacalin, which recognizes core 1 structures, was the highest for the Bo17 Spliced strain. Indeed, Bo17 Spliced virions were more sensitive to jacalin neutralization than Bo17 Del virions (Fig. 8B), suggesting therefore that the expression of the inactive form of the Bo17 enzyme could turn off the C2GnT activity of the cell. This will have to be tested in the future under other experimental conditions. Indeed, this property could be particularly important on the time scale of a virus, as a replicative cycle takes less than 20 h, a period which is similar to the half-life of most Golgi apparatus proteins (47).

Modulation of core 2 structures could have several implications in the BoHV-4 life cycle. Thus, we have shown that pBo17 activity affects gp180 O-glycans (Fig. 8A). We recently showed that gp180 and its glycans provide part of a glycan shield for otherwise-vulnerable viral epitopes on other viral glycoproteins, among which is glycoprotein L (26). However, we did not observe any difference of sensitivity to antibody neutralization between the strains described in this study (data not shown). One reason could be that the pBo17 enzymatic activity affects mainly the kind of O-glycans at the virion surface but not their presence. Key neutralizing epitopes would therefore remain hidden. Interestingly, while the dense shield of glycans that decorate different viruses are believed to be refractory to antibody recognition, several recent reports have shown that it can be recognized by rare neutralizing antibodies (48–52). Bo17 alternative splicing could therefore provide some antigenic variation of these protecting carbohydrates. The effect of Bo17 splicing on antibody neutralization could therefore only be observed over time, for example, when the virus reactivates from an immune host to allow transmission, or at the scale of a population. Interestingly, recent studies have highlighted remarkable variability between individual antibody responses with respect to their reactivity against glycopeptides of Epstein-Barr virus and herpes simplex virus 2 (53, 54). The capacity of transmission of the different viral strains used in this study should therefore be compared. However, herpesvirus transmission models are difficult to set up (55), and none is available for BoHV-4.

BoHV-4 has been shown to persist in CD11b cells of infected animals (56). Expression of core 2 structures has been shown to regulate monocyte extravasation and migration of monocyte-derived dendritic cells to lymph nodes (57–59). BoHV-4 could therefore regulate extravasation of infected monocytes and migration of these cells in tissues or toward draining lymph nodes by modulating glycan structures at the infected cell surface. Interestingly, another gammaherpesvirus has been shown to hijack myeloid cells for host colonization (60, 61), and therefore sophisticated mechanisms of control of those cells are possible. Further studies will be required to test these hypotheses.

Altogether, our results showed that the BoHV-4 Bo17 gene sequence evolved to allow alternative splicing and expression of a spliced inactive protein. To our knowledge, a similar phenomenon has never been observed for any other Golgi complex glycosyltransferase. Only one similar phenomenon has been observed for human UDP-glucuronosyltransferase enzymes (UGTs) and particularly for the UGT1A gene (62). Indeed, shorter UGT1A isoforms are deficient in glucuronic acid transferase activity and function as negative regulators of enzyme activity through protein-protein interactions. As Golgi complex localization and activity of some glycosyltransferases is also mediated by enzyme oligomerization, we propose that pBo17 Spliced expression could allow BoHV-4 to regulate C2GnT-M activity of the infected cells. This hypothesis will have to be tested in the future.

In conclusion, the present study of the BoHV-4 Bo17 gene revealed, similar to what has been observed for enzymes of other families, that inactive isoforms of Golgi complex glycosyltransferases might be expressed after mRNA alternative splicing. This new regulatory mechanism could have implications for our understanding not only of the BoHV-4 life cycle but also more generally of cellular biology.