Identification of Closely Spaced but Distinct Transcription Initiation Sites for the Murine Gammaherpesvirus 68 Latency-Associated M2 Gene

Abstract

Murine gammaherpesvirus 68 (MHV68) infection of mice provides a tractable small-animal system for assessing viral requirements for establishment of and reactivation from latency. The M2 gene product has no homology to any known proteins but has been shown to play a role in both the establishment of MHV68 latency and reactivation from latency. Furthermore, we have recently shown that M2 expression in primary murine B cells leads to enhanced proliferation, survival, and differentiation toward a preplasma memory B-cell phenotype (A. M. Siegel, J. H. Herskowitz, and S. H. Speck, PLoS Pathog. 4:e1000039, 2008). Previous studies have characterized the structure of the M2 transcript, but to date there has been no characterization of the M2 promoter, additional open reading frames (ORFs) in the M2 region, or identified splice acceptor and splice donor sites present in the previously characterized M2 gene transcript. Here we report (i) the identification and disruption of a novel transcript that encodes a short, previously unreported ORF (M2b) located in the intron between exon 1 and exon 2 of the M2 transcript; (ii) the identification of clustered but distinct M2 gene transcription initiation sites suggesting the presence of multiple promoters involved in regulating M2 gene transcription; (iii) the characterization in vivo of recombinant MHV68 harboring deletions within the identified M2 promoter region; and (iv) the in vivo analysis of recombinant MHV68 harboring mutations that ablate either the identified M2 splice acceptor or splice donor site. Finally, our 5′ rapid amplification of cDNA ends in conjunction with splice acceptor mutation analyses confirmed that all detected M2 gene transcripts expressed during MHV68 infection in mice splice into the M2 ORF downstream of the first AUG codon, providing strong evidence that initiation of the M2 gene product arises from the second AUG codon located at residue 8 in the M2 ORF. This initial detailed analysis of M2 gene transcription in vivo will aid future studies on regulation of M2 gene expression.

Gammaherpesviruses establish lifelong latent infections within the lymphocytes of a wide variety of mammalian species. The species specificity of the gammaherpesviruses that infect humans (Epstein-Barr virus [EBV] and Kaposi's sarcoma-associated herpesvirus) has limited in vivo studies of these viruses. Murine gammaherpesvirus 68 (MHV68) is a gammaherpesvirus that naturally infects wild murid rodents and shares significant genome colinearity with EBV and Kaposi's sarcoma-associated herpesvirus (3, 21, 27, 35). Infection of mice with MHV68 provides a tractable model system for characterizing the genetic requirements of gammaherpesvirus infection and for the development of therapeutics against human herpesviruses (18).

Infection of mice with MHV68 leads to acute virus replication in a number of anatomical sites, including the lungs and spleen, with peak viral titers observed between days 5 and 10 postinfection and viral clearance by the host adaptive immune response between days 9 and 15 postinfection (4, 31). Upon resolution of the lytic infection, a latent infection in B lymphocytes, macrophages, and dendritic cells is apparent. Like with EBV infection, long-term latency is primarily restricted to germinal center and memory B lymphocytes (7-9, 28-30, 40).

The M2 open reading frame (ORF) is located at the left end of the MHV68 genome. It shares positional homology to latency-associated genes in other gammaherpesviruses (26, 34, 36). Studies of M2 translational stop mutants revealed that the M2 ORF is not required for acute virus replication in the lungs but plays a critical role in the establishment of splenic latency following low-dose intranasal inoculation (11, 13, 24). Increasing the inoculating dose partially restores establishment of splenic latency, but under these conditions there is a severe defect in virus reactivation from latently infected B cells (11, 13, 24). Thus, determining how the M2 gene is regulated during different stages of infection and in various cell types is important for determining the role of M2 in viral latency and for identifying possible therapeutic targets for interrupting latency in infected hosts.

The M2 gene product is a unique protein with no known homologs in viruses, mice, or humans. It contains an actively recognized CD8⁺ T-cell epitope that constitutes an important target for controlling the establishment of latent infection (12, 32, 33). Sequence analysis of M2 identified 9 PxxP motifs, characteristic of SH3 binding motifs, and a central positively charged region, leading to the hypothesis that M2 may manipulate signal transduction pathways. Indeed, specific key SH3 binding motifs and tyrosine phosphorylation motifs contribute significantly to the establishment of latency and reactivation in vivo (10). In vitro studies have demonstrated that the M2 protein has a cell-type-dependent localization and effectively inhibits interferon-mediated signal transduction by downregulation of STAT1/2 expression (15). Additionally, M2 interacts with the DDB1/COP9/cullin repair complex and ATM in fibroblast cultures, suppressing DNA damage-induced apoptosis (14). In B-cell lines M2 activates the Vav1/Rac1 pathway through a trimolecular complex with Vav1 and Fyn, leading to enhanced Vav1 phosphorylation (19, 22). M2 expression in primary B cells leads to interleukin-10 (IL-10)-dependent B-cell proliferation and secretion of IL-10, IL-2, MIP-1a, and IL-6 (23). Primary B cells expressing M2 differentiate into a preplasma memory B-cell phenotype, an intermediate differentiation state between plasma and memory B cells (23). Additionally, infection with an M2 null MHV68 (M2.Stop) leads to a significant decrease in serum IL-10 levels at day 16 postinfection, correlating with an increase in the frequency of MHV68-specific CD8 T cells (23).

Initial analysis of the M2 transcript revealed a spliced transcript composed of two exons, i.e., a noncoding 110-nucleotide (nt) 5′ exon and a 1,235-nt 3′ exon that contains the M2 ORF, and a 656-nt 3′ untranslated region (12). Here we report the analysis of M2 gene transcription in latently infected mice, along with the subsequent mapping of M2 gene promoter activity in vitro and analysis of M2 promoter mutants in vivo. In addition, we identified a novel transcript arising from the region of the M2 gene, which contains a short ORF (M2b). However, analysis of viruses lacking an intact M2b failed to identify a role for this putative gene product in MHV68 infection. Finally, we constructed M2 splice acceptor and splice donor mutants that demonstrate the importance of the observed M2 splice acceptor site and the flexibility of M2 splice donor sites.

MATERIALS AND METHODS

Viruses and tissue cultures.

The MHV68 bacterial artificial chromosome (BAC) was a kind gift from Ulrich Koszinowski and was the wild-type (wt) virus from which all mutants in this study were derived. Virus passage and titer determination were performed as previously described (6). NIH 3T12 cells and mouse embryonic fibroblast (MEF) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 100 U of penicillin per ml, 100 mg of streptomycin per ml, 10% fetal calf serum, and 2 mM l-glutamine (cMEM). Vero-Cre cells were maintained in cMEM supplemented with 300 μg of hygromycin B/ml. WEHI231 and S11 B lymphoma cell lines were maintained in RPMI medium 1640 supplemented with 100 U of penicillin per ml, 100 mg of streptomycin per ml, 2 mM l-glutamine, 50 μM β-mercaptoethanol, and 10% fetal calf serum. Cells were maintained in a 5% CO₂ tissue culture incubator at 37°C. Vero-Cre cells were a kind gift from David Leib, Washington University, St. Louis, MO. MEF cells were harvested as previously described (20).

Generation of virus mutants.

Mutations were introduced with overlap PCR mutagenesis using the wt MHV68 BAC (1) as a template. For the M2 promoter deletion viruses, a common set of outer primers, 5366_D_NsiI (5′ GATCATGCATGTCTTCTTCCCCACGCGCACCAG 3′) and 6569_C_NsiI (5′ GATCATGCATGCCAGATCCCCGGTCTGTAAGAAC 3′), was used in combination with unique inner primer pairs as follows: for Δ5925-6014, 5924-5898_C (5′ GAAGAGACTTTCAGCTTTCGGGAAGGG 3′) and 5910-5924_6015-6041_D (5′ GCTGAAAGTCTCTTCCTTTATTTTTAACAAAAAAATTTCCTC 3′); for Δ5925-5969, 5924-5898_C (5′ GAAGAGACTTTCAGCTTTCGGGAAGGG 3′) and 5910-5924_5970-5993_D (5′ GCTGAAAGTCTCTTCGGAGATGGGTCAAAACCACCTGAC 3′); for Δ5861-5925, 5860-5835_C (5′ GTCTCTTCTGGGTGAAGCGGTGTTAC 3′) and 5847-5860_5926-5953_D (5′ CACCCAGAAGAGACGCTAGTCAAGCGTGGTGCAGACCTCATG 3′); and for Δ5861-6014, 5860-5835_C (5′ GTCTCTTCTGGGTGAAGCGGTGTTAC 3′) and 5847-5860_6015-6046_D (5′ CACCCAGAAGAGACCTTTATTTTTAACAAAAAAATTTCCTCTTAAG 3′). For the M2 splice acceptor mutant, outer primers 3964-3987_D (5′ GTCACGCTTCTCCTTCCAGGCGTG 3′) and 5208-5184_C (5′ GGTGAAATCCTGGGGACTTCTAGAG 3′) were used with inner primers 4592-4625_D (5′ GGGTGTTGGGGCCATGCTTTGAAAACGAAACCTC 3′) and 4625-4592_C (5′ GAGGTTTCGTTTTCAAAGCATGGCCCCAACACCC 3′) to introduce the underlined 4-bp mutation, which disrupts the M2 splice acceptor and introduces an NlaIII restriction site. For the M2 splice donor mutant, outer primers 5199-5222_D (5′ GGATTTCACCTAGTCCTTGCCGGC 3′) and 6415-6391_C (5′ GCTACTACAAGTACAGCGTGAGCCC 3′) were used with inner primers 5799-5833_D (5′ GGCCCAGAATACATCGAGCCCTGTGCGTGAGAGTC 3′) and 5833-5799_C (5′ GACTCTCACGCACAGGGCTCGATGTATTCTGGGCC 3′) to introduce the underlined 4-bp mutation, which disrupts the M2 splice donor site and introduces a BanII restriction site. For the M2b.Stop mutation, M2bSTOP_5182_D (5′ GATCAGATCTCTCTCTAGAAGTCCCCAGGATTTC 3′) and M2bSTOP_6382_C (5′ GATCGCGGCCGCCCATGGACCACTTGATCCTAG 3′) were used with inner primers M2bSTOP_5782_D (5′ CCCACCATTCCCTCTTAGGTCTAGACTACATACCACCCTG 3′) and M2bSTOP_5821_C (5′ CAGGGTGGTATGTAGTCTAGACCTAAGAGGGAATGGTGGG 3′) to introduce the underlined mutations, which introduce a translational stop codon and an XbaI restriction site after the initiation codon. Each mutant PCR product was cloned into pCR Blunt vector (Invitrogen) and sequenced to verify mutations. Recombinant viruses were generated by allelic exchange in Escherichia coli as previously described (17, 25). Briefly, the NsiI restriction sites within the pCR Blunt vector were used to excise the MHV68 fragment. The fragment was cloned into the suicide vector pGS284, which contains an ampicillin resistance cassette and a levansucrase cassette for positive and negative selection, respectively. The resulting targeting vectors were transformed into S17λpir E. coli cells and mated to GS500 E. coli (RecA⁺) harboring wt MHV68 BAC. Cointegrants were selected on Luria-Bertani (LB) agar plates containing chloramphenicol at 17 μg per ml and carbenicillin at 50 μg per ml and were resolved following overnight growth in liquid LB medium containing only chloramphenicol. Bacterial cells were then plated on LB agar plates containing chloramphenicol and 7% sucrose to select for loss of the levansucrase cassette (and thus the pGS284 vector sequence). Colonies were screened by PCR for a reduction in size for the promoter mutants or for the introduction of the diagnostic restriction site for the M2b.Stop mutation, the M2 splice acceptor mutation, and the M2 splice donor mutation. Positive clones were grown in LB medium with chloramphenicol, and BAC DNA was purified with a MidiPrep kit (Qiagen, Hilden, Germany) as described by the modified manufacturer's protocol. The presence of the desired mutations was further confirmed by amplifying sequence outside of the targeting region (100 bp on either side of the homology arms) by PCR, cloning into the pCR Blunt vector, and sequencing with vector-specific primers.

RACE analysis.

RNA was harvested from pooled splenocytes of five mice harvested at day 16 post-intranasal infection with 100 PFU wt virus. Approximately 2 × 10⁸ cells were lysed with guanidine isothiocyanate-phenol solution (GITC-phenol, 2 M guanidine isothiocyanate, 0.05 M β-mercaptoethanol, 0.25% sarcosyl, 0.1 M Na acetate) followed by extraction with chloroform. RNA was precipitated with isopropanol and resuspended in RNase-free water. Rapid amplification of cDNA ends (RACE) was performed using the GeneRacer system (Invitrogen). The sequence of the gene-specific oligonucleotide primer used for round 1 5′ RACE, 4285-4309_D, was 5′ CTTCAGGACTTGGTACAGGACTCGG 3′; that for round 2 5′ RACE, 4391-4415_D, was 5′ CAAAGGCGGGCGCTGAGGTCTGCCC 3′. The sequence of the gene-specific oligonucleotide primer used for round 1 3′ RACE, 5918-5893_C, was 5′ GACTTTCAGCTTTCGGGAAGGGTTTAG 3′; that for round 2 3′ RACE, 5888-5864_C, was 5′ TTCCCCTCTCAAGCTGCTTCCTTAG 3′. PCR products were electrophoresed on 1% agarose gels, and specific bands were excised and prepared for ligation using a Qiagen gel extraction kit (Qiagen, Valencia, CA). Products were subcloned into pCR4-TOPO (Invitrogen) and analyzed by DNA sequencing. RACE analysis was repeated three times from independent mouse infections.

Plasmids.

M2 promoter reporter plasmids were constructed by PCR amplification of the indicated sequences from a wt MHV68 BAC DNA template with XhoI restriction sites incorporated in both the forward and reverse primers. PCR products were purified by gel extraction, digested with XhoI, and cloned into the XhoI site of the pGL4.10 luciferase reporter vector (Promega Corporation, Madison, WI). The orientation of the inserted sequence was verified by PCR with vector-specific primer pGL4.10_D (5′ CTAGCAAAATAGGCTGTCCCC 3′) or pGL4.10_C (5′ GCGTAGGTAATGTCCACCTCG 3′) with the insert-specific primer M2-103_6027C (5′ GATCCTCGAGGTTAAAAATAAAGTTATTTTTAAAAAAGATACGAG 3′). Upstream primer M2-1000_6924C (5′ GATCCTCGAGCTCTGTGGGTGGCACACCAGTG 3′), M2-500_6424C (5′ GATCCTCGAGCTTGCCCATGCTACTACAAGTAC 3′), M2-203_6127C (5′ GATCCTCGAGCTAACCTGGTTGGGGTCAGCTTGATG 3′), or M2-103_6027C (5′ GATCCTCGAGGTTAAAAATAAAGTTATTTTTAAAAAAGATACGAG 3′) was used with downstream primer M25′_5924D (5′ GATCCTCGAGCTGCTAGTCAAGCGTGGTGCAGACC 3′), M2 + 50_5875D (5′ GATCCTCGAGGCTTGAGAGGGGAAGTGCCTAAAC 3′), M2 + 75_5850D (5′ GATCCTCGAGCCAGAAGAGACTGGCTAAGGAAG 3′), or M2 + 100_5825D (5′ GATCCTCGAGGTGAGAGTCTGTAACACCGCTTC 3′) to construct the array of reporter fragment lengths.

Luciferase reporter assays.

WEHI231 or S11 cells were grown to confluence under the conditions described above. Reporter constructs were transfected with a Nucleofector I device (Amaxa, Inc., Gaithersburg, MD) according to manufacturer's protocol. Briefly, cells were counted to determine density, spun down, and resuspended at 1 × 10⁶ cells per 100 μl of the appropriate Nucleofector solution (solution V for S11 cells and solution R for WEHI231 cells). Five micrograms of reporter plasmid DNA in 5 μl water was mixed by gentle pipetting with 100 μl of cell suspension; the sample was placed in an Amaxa cuvette and subjected to the appropriate transfection program in the Nucleofector I device (program O-17 for both S11 and WEHI231 cells). Cells were transferred from the cuvette to a 12-well tissue culture plate and cultured in 2 ml RPMI 1640. The green fluorescent protein expression plasmid pmaxGFP (Amaxa) was used to measure nucleofection efficiency by fluorescence microscopy. At 48 h postnucleofection, cells were scraped from the wells and spun down. Cell pellets were lysed in 200 ml passive lysis buffer (Promega) for 20 min at room temperature. Fifty microliters of lysate was used to assay luciferase activity with a luciferase reporter assay system (Promega) according to the manufacturer's protocol. Transfections and luciferase assays were performed independently in triplicate at least two times.

Mice, infections, and organ harvests.

Female C57BL/6J mice (catalog no. 000664; The Jackson Laboratory, Bar Harbor, ME) were housed at the Yerkes or the Whitehead vivarium in accordance with university and federal guidelines. Mice between the ages of 8 and 12 weeks were placed under isofluorane anesthesia prior to intranasal inoculation with 100 PFU of virus in 20 μl of cMEM. Spleens were harvested into cMEM, homogenized, and filtered through a 100 μm-pore-size nylon strainer (Becton Dickson, Franklin Lakes, NJ). Erythrocytes were eliminated with red blood cell lysis buffer (Sigma, St. Louis, MO). Splenocytes were pooled from five mice in all experiments. Lungs were harvested in cMEM and stored at −80°C.

Plaque assay.

Plaque assays were performed as previously described (5) with the following modifications. Six-well plates were seeded with 2 × 10⁵ NIH 3T12 cells per well at 24 h prior to infection. Mechanical disruption of lungs was performed in four rounds of 1 minute each using 1.0-mm zirconia/silica beads (Biospec Products, Bartlesville, OK) in a Mini-Beadbeater-8 (Biospec Products). Serial 10-fold dilutions of lung homogenate were plated on NIH 3T12 monolayers in a 200-μl volume. Infections were performed at 37°C for 1 hour with rocking every 15 min. Following infection, monolayers were covered with 5 ml cMEM containing 2% methylcellulose. At 6 or 7 days following infection, plates were stained with a neutral red overlay, and plaques were counted the following day. The limit of detection for this assay is 50 PFU.

LD ex vivo reactivation analyses.

Limiting-dilution (LD) analysis to determine the frequency of cells harboring virus capable of reactivating from latency was performed as previously described (37, 39). Briefly, splenocytes were resuspended in cMEM as described above and plated in serial twofold dilutions (starting with 1 × 10⁵cells) onto MEF monolayers in 96-well tissue culture plates. Twelve dilutions were plated per sample, and 24 wells were plated per dilution. Wells were scored for cytopathic effect (CPE) at 21 days postplating. The presence of preformed infectious virus was determined by plating parallel samples of mechanically disrupted cells (latent virus cannot reactivate from killed cells) onto MEF monolayers (37-39). The level of sensitivity of this assay is 0.2 PFU (37).

LD-PCR detection of the frequency of cells harboring the viral genome.

LD analysis to determine the frequency of cells harboring the viral genome was performed using a nested PCR assay with single-copy sensitivity as previously described (38, 39). Briefly, splenocyte samples were counted, washed, resuspended in isotonic buffer, and plated in serial threefold dilutions in a background of 1 × 10⁴ uninfected NIH 3T12 cells in 96-well plates (MWG Biotech, High Point, NC). Plates were covered with PCR foil (Eppendorf Scientific, Westbury, NY). Cells were lysed with proteinase K for 6 h at 56°C. Ten microliters of round 1 PCR mix was added to each well by foil puncture. Upon completion of the first round of PCR, 10 μl of round 2 PCR mix was added to each well by foil puncture. Following the second round of PCR, products were resolved by ethidium bromide staining on 2% agarose gels. Cell lysis and PCR were performed on a PrimusHT thermal cycler (MWG Biotech, High Point, NC). Twelve PCRs were performed for each sample dilution, and six dilutions were performed for each sample. Each PCR plate contained control reactions (uninfected cells or 10 copies, 1 copy, or 0.1 copy of target plasmid DNA in a background of 1 × 10⁴ cells) as previously described (38, 39). All LD-PCR assays demonstrated approximately single-copy sensitivity without false positives.

Statistical analyses.

Data were analyzed with GraphPad Prism software (GraphPad Software, San Diego, CA). Data were subjected to nonlinear regression analysis to determine the single-cell frequency for each LD analysis. From the Poisson distribution, the frequencies of reactivation and viral genome-positive cells were obtained from the nonlinear regression fit of the data where the regression line intersected 63.2%.

RESULTS AND DISCUSSION

Characterization of M2 gene transcripts in latently infected splenocytes.

Previous studies have failed to identify M2 transcripts by northern analyses in lytically infected cells (12, 36), consistent with a role for the M2 gene product during viral latency. The M2 gene transcript was initially characterized in the MHV68 latently infected S11 B lymphoma tumor cell line as a spliced transcript composed of two exons, i.e., a short noncoding 5′ exon of 110 nt in length and a 1,235-nt 3′ exon which contains the entire M2 coding sequences, and a 656-nt 3′ untranslated region (9). To determine the structure of M2 gene transcripts in vivo, we performed nested RACE analyses on poly(A) RNA isolated from splenocytes of five mice infected with 100 PFU of wt MHV68 and harvested at 16 days postinfection. RACE products were generated and subcloned using appropriate RACE primers (see Materials and Methods). Multiple independent clones were sequenced to determine the 5′ end for the M2 transcript. A total of seven clones had 5′ initiation sites at bp 5924 or 5925 in the viral genome, consistent with the mapping studies reported by Husain et al. (12). However, the majority of the 5′ RACE clones (24 of 41 5′ RACE products analyzed) initiated at bp 5861 (Fig. 1). A third distinct cluster of initiation sites was identified at bp 5892 and 5893 in the viral genome, and 3 additional 5′ RACE clones that initiated at bp 5870 were identified (Fig. 1). Notably, these likely represent full-length transcripts (i.e., bona fide 5′ ends), because the 5′ RACE protocol employed should amplify only 5′-capped transcripts (see Materials and Methods). However, it is possible that some of the RACE products obtained reflect truncated transcripts (e.g., those transcripts initiating at bp 5870). The presence of three distinct clusters of transcription initiation raises the possibility that there may be multiple promoters involved in driving M2 gene transcription.

FIG. 1.

(A) Identification of M2 gene transcription start sites. 5′ and 3′ RACE analyses were performed using nested primer sets on 5′-capped poly(A) RNA prepared from bulk splenocytes harvested at day 16 postinfection as described in Materials and Methods. The locations of the RACE primers are indicated by black arrows beneath the representation of the M2 transcript structure. 5′ initiation sites are indicated by vertical arrows beneath the primary sequence, and the number of sequenced clones originating at the indicated sequence is listed below the arrow. (B) In silico analysis (http://www.cbil.upenn.edu/tess/) of potential transcription factor binding sites upstream of the proximal and distal transcription initiation sites.

We also examined the 3′ end of the M2 transcript using RACE primers. The majority of the clones utilized the previously identified splice acceptor and donor sites, as well as the 3′ termination site reported by Husain et al. (12). However, we also detected a novel unspliced transcript that initiated at bp 5861 and terminated at bp 5498. The latter transcript contained a short ORF that we have termed M2b (Fig. 2). Notably, this short ORF does not exhibit significant homology to any known viral or cellular proteins. As discussed below, we have addressed the significance of the M2b ORF in MHV68 pathogenesis to determine whether it contributes to the phenotype of the M2 null virus.

FIG. 2.

Schematic overview of the region of the MHV68 genome carrying the M2 ORF. The locations of the M1, M2, M2b, and M3 ORFs are indicated by arrows. The three M2 transcripts, two spliced and one unspliced, are shown below the schematic of the viral genome. 3′ RACE identified the termination site for the spliced M2 transcripts at bp 3378 and that for the unspliced M2b transcript at bp 5498. All the detected spliced M2 gene transcripts utilized the splice acceptor site at bp 4609 and the splice donor site at bp 5815.

Identification of sequences required for M2 gene promoter activity.

Following the identification of three distinct clusters of 5′ transcription initiation sites over 64 nt in our 5′ RACE analyses (Fig. 1), we sought to identify the minimal region(s) required for M2 promoter activity. To this end, a series of luciferase reporter plasmids containing putative M2 regulatory sequences, spanning the region from 1,000 bp upstream of the previously identified M2 5′ initiation site at bp 5925 to 100 bp downstream of this transcription initiation site, were constructed (Fig. 3). Luciferase activity was assayed in the non-MHV68-infected murine B lymphoma cell line WEHI231 231 and in the latently MHV68-infected S11 B lymphoma cell line, which expresses the M2 gene product. The results from both cell lines correlated very well, although higher activity was observed in the S11 cell line. Importantly, reporter constructs which contained only sequence upstream of the previously identified 5′ end of the M2 transcript (bp 5925) did not exhibit detectable activity in either WEHI231 or S11 cells (Fig. 3). Reporter constructs extending downstream to include sequences to either bp 5875 or bp 5850 exhibited low-level activity in both B-cell lines (Fig. 3). Finally, reporters that extended downstream of all the identified 5′ transcription initiation sites to bp 5825, which includes all but 10 nt of the first exon, exhibited the highest level of luciferase activity (Fig. 3). Inclusion of sequences upstream of bp 6027 slightly increased promoter activity (Fig. 3). Taken together, these data indicate that the critical cis elements regulating M2 promoter activity in the WEHI231 and S11 B-cell lines map between bp 5825 and 6027.

FIG. 3.

Identification of an M2 promoter active in WEHI231 and S11 cells. Schematic illustrations of the M2 gene promoter constructs generated are shown beneath a map of the M3 gene and genomic sequences upstream of the M2 exon 1. The promoter constructs cloned were upstream of the firefly luciferase gene in the pGL4.10 reporter vector (Promega). The corresponding genomic coordinates, as well as the locations of M3 ORF and both M2 exon 1 species, are indicated above the reporter constructs. M2 promoter activity was assayed in WEHI231 and S11 cells. Each construct, or the control empty luciferase vector, was nucleofected into WEHI231 or S11 cells (see Materials and Methods). Fold induction in luciferase activity was calculated by dividing the luciferase activity observed with the indicated promoter constructs by that of empty vector. Error bars represent standard errors of the means from at least three independent experiments.

Characterization of M2 promoter mutant viruses.

The 5′ RACE data indicated the presence of two or three distinct transcriptional start sites, raising the possibility that perhaps two or three overlapping promoters are involved in driving M2 gene transcription. Indeed, it is difficult to envision a single promoter being responsible for the transcripts initiating at bp 5863 and those initiating at bp 5930/5931. As discussed above, those 5′ RACE products mapping between these transcription initiation sites may reflect truncated transcripts that were not efficiently removed during the cap-dependent 5′ RACE protocol. Notably, the data obtained with established B-cell lines transiently transfected with luciferase reporters supported the presence of only a single promoter, which is most consistent with the major transcription initiation site at bp 5863 identified from our 5′ RACE analyses (Fig. 3). However, since the M2 gene product appears to play a role in both establishment of virus latency and reactivation from latency, it is certainly possible that transcription initiation from an upstream promoter was not revealed in the latter assays. For example, the reporter gene assays would likely have failed to detect an M2 promoter whose activity is specifically regulated in response to reactivation stimuli. To more rigorously address the importance of sequences immediately upstream of the M2 transcription initiation sites, a panel of MHV68 mutant viruses containing small deletions in this region of the MHV68 genome were generated and verified by Southern blotting and DNA sequencing (Fig. 4). Notably, for all but one of the M2 promoter deletions (Δ5925-5969), two independently generated mutants were isolated to ensure that observed phenotypes were not due to spurious mutations at distal sites in the viral genome (Fig. 4).

FIG. 4.

Construction of M2 promoter mutant viruses. (A) The MHV68 BAC developed by Adler et al. (1) was used introduce the indicated deletions by RecA-mediated recombination (see Materials and Methods). (B) Southern blot analysis of M2 promoter mutant viruses. Deletion of the sequences from bp 5861 to 6014 and bp 5925 to 6014 removed both a KpnI and a DraI restriction site. Digests with these enzymes demonstrate the absence of these sites in the recombinant viruses. Deletion of the sequence from bp 5925 to 5969 resulted in the loss of the KpnI restriction site. A digest of this virus with KpnI does not cut, but the DraI site remains intact. Deletion of the sequence from bp 5861 to 5925 does not ablate either the DraI or KpnI site but does result in smaller fragments. This reduction in size is distinguishable in both digests. Sequencing 1.5 kb of sequence flanking both sides of the mutations confirmed specific deletion of the indicated sequence as well as the absence of second-site mutations. The probe contained sequence from genomic coordinates bp 4652 to 6559. Molecular size markers (1-kb DNA ladder; New England Biolabs, Beverly, MA) were run in the last lane.

The minimal region identified upstream of the M2 gene transcription initiation sites that exhibited strong promoter activity in the luciferase reporter assay extended from bp 6027 to 5825. Notably, the M3 polyadenylation signal lies at bp 6021 to 6016 in the MHV68 genome. In an effort to preserve M3 transcription, none of the 5′ boundaries of the M2 promoter deletion mutations extended upstream of bp 6014. Based on our working hypothesis that there are at least two distinct promoters involved in driving M2 gene transcription, we introduced deletions that would likely completely disrupt either only the proximal promoter (Δ5861-5925), the distal promoter (Δ5925-5969 and Δ5925-6014), or both promoters (Δ5861-6014) (Fig. 4).

To determine if the deletions in the M2 promoter region affected acute-phase virus replication, a multistep growth curve was performed (Fig. 5A and B). As expected, there was no statistical difference in the in vitro growth kinetics of any of the M2 promoter mutants compared to wt MHV68. Furthermore, no significant difference in acute virus replication in the lungs at day 9 postinfection was observed with any of the M2 promoter mutant viruses following intranasal inoculation with 100 PFU of virus (Fig. 5C). The latter results were expected since, as we have previously shown, there was no defect in acute replication of M2 null virus (M2.Stop) (13, 24).

FIG. 5.

M2 promoter mutants replicate with wt MHV68 kinetics in vitro and in vivo. (A and B) NIH 3T12 monolayers were infected with 0.05 PFU of the indicated virus. Wells were harvested at the indicated time points, and virus titers were determined on NIH 3T12 cells (see Materials and Methods). Error bars represent standard errors of the means from three independent experiments. (C) C57BL/6J mice were infected with 100 PFU of indicated virus. Lungs were harvested at day 9 postinfection, and viral titers were determined by plaque assay (see Materials and Methods). The data were compiled from two independent experiments, and each symbol represents the viral titer determined from an individual mouse infected with the indicated virus. The horizontal dashed line indicates the limit of detection of the assay (50 PFU/ml).

While the M2.Stop virus did not exhibit a defect in acute virus replication in the lungs following intranasal inoculation, it was severely impaired for establishment of latency (13, 24). To determine if deletion of portions of the M2 promoter recapitulated the M2.Stop phenotype, we measured the frequency of cells harboring the viral genome, as determined by LD-PCR on bulk splenocytes at day16 postinfection following intranasal inoculation with 100 PFU of the indicated virus (Table 1 and Fig. 6A and C). wt MHV68 established latency at a frequency of ca. 1 in 142 splenocytes, while the M2.Stop mutant established latency at a frequency of ca. 1 in 46,423 splenocytes. All the M2 promoter mutants exhibited a significant defect in establishment of latency, which ranged from an 8-fold to a 40-fold reduction compared to wt MHV68 (Table 1 and Fig. 6A and C). However, none of the M2 promoter mutants was as impaired as the M2.Stop virus (ca. 300-fold defect compared to wt MHV68) (Table 1 and Fig. 6A and C). Not surprisingly, the M2 promoter mutant with the largest deletion, Δ5861-6014, established latency at a lower frequency than the mutants with smaller deletions. Notably, none of the deletions clearly distinguished the importance of the proximal and distal transcription initiation sites. This may reflect the presence of cis-regulatory elements that are required for both upstream and downstream transcription initiation sites. Importantly, even the largest M2 promoter deletion (Δ5861-6014), although significantly reduced in establishment of latency compared to wt virus, did not exhibit as severe a defect as the M2.Stop mutant (Table 1 and Fig. 6C). This raises the possibility that there may be an alternatively mechanism for generating M2 gene expression, which could reflect either transcription from a cryptic promoter(s) that is revealed when the identified M2 promoter(s) is deleted or an alternative bona fide M2 gene promoter. Importantly, to date none of our 5′ RACE analyses of wt MHV68-infected splenocytes have identified the presence of any alternatively initiated M2 gene transcripts.

TABLE 1.

Frequencies of viral genome-positive and reactivating splenocytes at day 16 postinfection

Virus	Total no. of splenocytes recovered	Reciprocal frequency of viral genome-positive cells	Genome frequency relative to wt virus frequency	Estimated no. of viral genome-positive cells	Reciprocal frequency of cells reactivating	Reactivation frequency relative to wt virus frequency	Estimated no. of reactivating cells	Reactivation efficiency
wt MHV68	1.95 × 108	142	1.000	1,373,239	2,925	1.000	66,667	4.85
M2.Stop	9.52 × 107	46,423	0.003	2,051	735,437	0.004	129	6.31
Δ5861-5925(1)	1.40 × 108	1,608	0.088	86,857	135,336	0.022	1,032	1.19
Δ5861-5925(2)	1.70 × 108	2,158	0.066	78,622	160,882	0.018	1,055	1.34
Δ5925-5969(1)	1.61 × 108	1,161	0.122	138,961	626,969	0.005	257	0.19
Δ5925-6014(1)	1.76 × 108	1,858	0.076	86,832	509,488	0.006	346	0.40
Δ5925-6014(2)	3.32 × 108	1,742	0.082	190,586	656,555	0.004	506	0.27
Δ5861-6014(1)	1.48 × 108	4,699	0.030	31,425	696,639	0.004	185	0.59
Δ5861-6014(2)	1.63 × 108	7,241	0.020	22,465	762,698	0.004	213	0.95
ΔM2SA(1)	1.34 × 108	16,881	0.008	7,918	880,909	0.003	152	1.92
ΔM2SA(2)	1.40 × 108	13,285	0.011	10,513	679,656	0.004	205	1.95
ΔM2SD(1)	1.37 × 108	598	0.237	229,097	93,049	0.031	1,472	0.64
ΔM2SD(2)	1.54 × 108	893	0.159	172,826	95,769	0.031	1,612	0.93
M2b.Stop(1)	5.99 × 107	84	1.690	713,393	4,961	0.590	12,079	1.69
M2b.Stop(2)	4.71 × 107	57	2.491	825,439	3,509	0.834	13,408	1.62

FIG. 6.

Analysis of virus latency and reactivation of M2 promoter mutants reveals significant defects. Bulk splenocytes were harvested from MHV68-infected C57BL/6J mice at day 16 and analyzed by LD analyses to determine the frequency of splenocytes harboring the viral genome and capable of spontaneously reactivating virus upon explant. (A and C) LD-PCR determination of the frequency of cells harboring the viral genome (see Materials and Methods). (B and D) LD ex vivo reactivation determination of the frequency of cells reactivating virus (see Materials and Methods). Plated in parallel, splenocytes that were mechanically disrupted allowed the presence of preformed infectious virus to be distinguish from reactivation from latency (see Materials and Methods). Notably, no preformed infectious virus was detected in these experiments (data not shown). For LD-PCR and reactivation assays, curve fit lines were derived from nonlinear regression analysis, and symbols represent mean percentages of wells positive for virus (DNA or CPE) ± the standard error of the mean. The dashed line represents 63.2%, from which the frequency of viral genome-positive cells or the frequency of cells reactivating virus was calculated based on a Poisson distribution. The data shown represent at least three independent experiments with cells pooled from five mice per experimental group.

When reactivation from latency was examined, all but the Δ5861-5925 deletion mutant were nearly as impaired as the M2.Stop virus (<1 in 500,000 splenocytes reactivating virus) (Table 1 and Fig. 6B and D). The Δ5861-5925 deletion mutant was also significant impaired for virus reactivation (ca. 1 in 150,000 splenocytes reactivating virus versus 1 in 2,925 for wt virus) (Table 1), although when the decreased establishment of latency (ca. 13-fold) is taken into consideration, the reduction of reactivation appears to be only ca. 4-fold compared to wt virus (Table 1). This suggests that the proximal transcription initiation site may play an important role in establishment of MHV68 latency and a less prominent role in virus reactivation from latency. With respect to reactivation of the other M2 promoter mutants, as well as the M2.Stop virus, comparison of establishment of latency and reactivation cannot be as rigorously addressed because the extremely low frequency of cells reactivating virus makes it difficult to experimentally determine this frequency (reactivation frequencies for these mutants were estimated based on extrapolations of LD analyses [Fig. 6]).

M2 splice acceptor and M2 splice donor viruses exhibit defects in establishment of and reactivation from latency.

Every spliced transcript isolated in our RACE analyses contained the same splice from bp 5815 to 4609. Based on the observation that even the large M2 promoter deletion (Δ5861-6014) was not as impaired for establishment of latency as the M2.Stop mutant (see above), we sought to determine whether the strict use of the splice donor at bp 5814 and the splice acceptor at bp 4607 is required for efficient latency establishment and reactivation. To this end, we constructed M2 splice acceptor and M2 splice donor mutant viruses incorporating base substitutions and restriction sites that were successfully utilized in the mutagenesis of the HSV-1 genome (2) (Fig. 7A). The splice acceptor mutant incorporated an NlaIII restriction site at the acceptor, which disrupted the acceptor sequence and changed four base pairs but kept the initiation codon intact. The splice donor mutant incorporated a BanII restriction site, which disrupted the donor sequence and changed five base pairs. Two independent isolates of the M2 splice acceptor and M2 splice donor viruses were verified by Southern blotting (Fig. 7B) and DNA sequencing.

FIG. 7.

Construction of M2 splice acceptor (SA) and splice donor (SD) mutant viruses. (A) The MHV68 BAC developed by Adler et al. (1) was used to disrupt the M2 splice acceptor or splice donor sequences and incorporate an NlaIII or BanII restriction site, respectively, by RecA-mediated recombination (see Materials and Methods). The two independently generated mutants for each virus were termed M2 SA(1), M2 SA(2), M2 SD(1), and M2 SD(2). Also shown is the position for the M2 exon 2 splice acceptor site relative to the two closely spaced ATG codons, which are underlined. (B) Southern blot analyses of splice acceptor and splice donor viruses. For the splice acceptor mutants, incorporation of the NlaIII site resulted in two unique bands of 134 and 148 nt for the mutant virus isolates and a single band of 282 nt for the wt virus. For the splice donor mutants, incorporation of the BanII site resulted in two unique bands at 576 and 366 nt for the mutant virus isolates and a single band of 942 nt for the wt virus. Digests with PstI and BanII for the SA mutants or PstI and NlaIII for the SD mutants demonstrate the expected restriction pattern, demonstrating appropriate targeting of the mutation into the viral genome. Complete sequencing of the region confirmed that only the intended mutations were present (see Materials and Methods). For the SA mutant blot, the probe contained sequence from genomic coordinates bp 3822 to 5301; the probe for the SD mutant blot contained sequence from genomic coordinates bp 5090 to 6517. DNA stnds, DNA standards (100-bp ladder and 1-kb DNA ladder; New England Biolabs, Beverly, MA).

To determine if the M2 splice acceptor and M2 splice donor mutations altered virus replication, a multistep growth curve was performed (Fig. 8A). As expected, there was no difference between the in vitro growth kinetics of the M2 splice acceptor mutants or the M2 splice donor mutants compared to wt MHV68. When establishment of latency was assessed, following intranasal inoculation with 100 PFU of virus, the splice donor mutant exhibited a modest defect (<10-fold reduction) (Table 1 and Fig. 8B). However, the splice acceptor mutants were significantly more impaired for establishment of latency (ca. 100-fold reduction) (Table 1 and Fig. 8B), with a frequency only slightly greater than that observed with the M2.Stop mutant (ca. 300-fold reduced) (Table 1 and Fig. 8B). Notably, although reactivation of the splice donor mutants could readily be measured (ca. 1 in 95,000 splenocytes reactivating virus) (Table 1 and Fig. 8C), reactivation of the splice donor mutants was very similar to that of the M2.Stop mutant (<1 in 650,000 splenocytes). Taken together, these analyses suggest that there are alternative splice donor sites that either compensate or are used to generate M2 transcripts. However, it appears that utilization of the known splice acceptor site is required for functional expression of M2 during virus infection in vivo.

FIG. 8.

M2 splice donor and splice acceptor mutants are impaired for latency establishment and reactivation. (A) M2 splice donor and acceptor mutants replicate with wt MHV68 kinetics in an in vitro multistep growth analysis. NIH 3T12 monolayers were infected with 0.05 PFU of the indicated virus (see Materials and Methods). Error bars represent standard errors of the means from three independent experiments. (B) Frequency of cells harboring viral genome. The frequency of viral genome-positive cells was determined using an LD-PCR analysis, as described in Materials and Methods. (C) Frequency of cells reactivating virus. Serial dilutions of live, intact splenocytes were plated on MEF indicator monolayers in parallel with samples that were mechanically disrupted to distinguish between virus reactivation from latency and virus replication resulting from preformed infectious virus (see Materials and Methods). No preformed infectious virus was detected in these experiments (data not shown). For LD-PCR and reactivation assays, curve fit lines were derived from nonlinear regression analysis, and symbols represent mean percentages of wells positive for virus (DNA or CPE) ± the standard error of the mean. The dashed line represents 63.2%, from which the frequency of viral genome-positive cells or the frequency of cells reactivating virus was calculated based on a Poisson distribution. The data shown represent at least three independent experiments with cells pooled from five mice per experimental group.

The M2b ORF is not required for acute virus replication in vivo or for establishment and reactivation of MHV68 latency.

With respect to the M2 promoter mutants, it was important to determine whether potential alterations in transcription of the M2b ORF contributed to the observed phenotypes. To determine if the M2b ORF plays a role in MHV68 infection, we constructed a recombinant virus in which the M2b ORF was interrupted with a translational stop codon and XbaI restriction site, and this recombinant virus was designated M2b.Stop (Fig. 9). Two independent isolates were confirmed by Southern blotting and DNA sequencing across the region containing the introduced mutations. Notably, no defect in M2b.Stop virus growth was observed in vitro (data not shown). To assess whether the M2b ORF-encoded gene product plays a role in virus replication in vivo, mice were infected via intranasal inoculation with 100 PFU of either wt MHV68 or the two M2b.Stop isolates. At 9 days postinfection, there was no statistically significant difference in acute titer in the lung (data not shown). These findings indicate that M2b is not required for acute-phase replication in the lungs following intranasal inoculation. Notably, previous studies demonstrated that the M2 ORF is dispensable for acute replication in vivo and in vitro (13, 31).

FIG. 9.

Construction of M2b mutant virus. (A) The MHV68 BAC developed by Adler et al. (1) was used to introduce a translational stop codon, along with an XbaI restriction site and frameshift mutation downstream of the putative M2b initiation codon by RecA-mediated recombination (see Materials and Methods). The two independently generated isolates of this mutant virus were termed M2b.Stop(1) and M2b.Stop(2). (B) Southern blot analysis of M2b.Stop viruses. Insertion of the Stop/XbaI cassette resulted in two unique bands of 2,251 and 616 nt for the M2b.Stop viruses and a single band of 2,867 nt for the wt virus. Digestion with BamHI and EcoRI demonstrated the absence of spurious mutations, and complete sequencing of the M2 gene region confirmed that only the intended mutations were present. The probe contained sequence from MHV68 genomic coordinates bp 4652 to 6559. Molecular size markers (1-kb DNA ladder; New England Biolabs, Beverly, MA) were run in the last lane.

To assess the ability of the M2b.Stop mutants to establish latency, we determined splenic latency in C57BL/6J mice following intranasal inoculation with 100 PFU M2b.Stop(1), M2b.Stop(2), M2.Stop, or wt MHV68. At 16 days postinfection, both M2b.Stop isolates and wt virus established latency at a frequency of ca. 1 in 100 splenocytes, and M2.Stop established latency at a frequency of ca. 1 in 46,423 splenocytes (Table 1 and Fig. 10A). We also determined the reactivation efficiencies of these viruses at day 16 postinfection. M2b.Stop(1) reactivated at a frequency of ca. 1 in 4,961, M2b.Stop(2) at ca. 1 in 3,509, wt virus at ca. 1 in 2,925, and M2.Stop at ca. 1 in 735,437 splenocytes (Table 1 and Fig. 10B). Taken together, these data indicate that the M2b ORF is dispensable for both the establishment and reactivation of MHV68 latency. At this point it remains unclear whether the M2b ORF actually encodes a protein and, if so, what role it plays in MHV68 infection. However, we can conclude that alterations in M2b transcription do not contribute to the phenotypes observed with the M2 gene promoter mutants.

FIG. 10.

In vivo analysis of M2b mutant viruses. (A) Frequency of cells harboring the viral genome. The frequency of viral genome-positive cells was determined by using an LD-PCR assay. Serial dilutions of splenocytes were plated into a background of 1 × 10⁴ uninfected cells, lysed, and analyzed with a nested PCR to detect viral genomes (see Materials and Methods). (B) Frequency of cells reactivating virus. Serial dilutions of live, intact splenocytes were plated on MEF indicator monolayers in parallel with samples that were mechanically disrupted to distinguish between virus reactivation from latency and virus replication resulting from preformed infectious virus (see Materials and Methods). No preformed infectious virus was detected in these experiments (data not shown). For LD-PCR and reactivation assays, curve fit lines were derived from nonlinear regression analysis, and symbols represent mean percentages of wells positive for virus (DNA or CPE) ± the standard error of the mean. The dashed line represents 63.2%, from which the frequency of viral genome-positive cells or the frequency of cells reactivating virus was calculated based on a Poisson distribution. The data shown represent at least three independent experiments with cells pooled from five mice per experimental group.

Conclusions.

Here we have addressed M2 gene transcription in vivo. These analyses have revealed a previously unappreciated complexity with respect to the sites of M2 gene transcription initiation, as well as identifying a novel transcription containing the M2b ORF. Our mutagenesis studies did not reveal a role for the M2b ORF in viral pathogenesis; however, this does not rule out a role for the putative M2b gene product in MHV68 infection. More-detailed studies, including determining whether M2b encodes a protein expressed during virus infection, would be required to adequately address this issue. However, based on the results presented here, there is no evidence that the short M2b encodes a protein that plays a role in virus infection.

With respect to M2 gene transcription, the 5′ RACE analyses provide evidence that multiple, closely linked promoters are involved in driving M2 gene transcription. The observation that deletion of sequences just upstream of the proximal M2 transcription initiation site (Δ5861-5925) led to a greater impact on establishment of latency than reactivation from latency suggests the possibility that distinct transcription regulation of the M2 gene may be involved in the role of the M2 protein in establishment of latency versus its role in reactivation from latency. Consistent with this hypothesis, the Δ5925-5969 mutant, which deletes sequences upstream of the identified distal M2 transcription initiation site, exhibited a slightly milder establishment defect than the Δ5861-5925 deletion but a more severe reactivation defect. Of course, the difficulty with such promoter deletion analyses, particularly in the case of closely spaced promoters as proposed here, is the likely possibility of deleting cis-regulatory elements that play a role in regulating transcription from both the proximal and distal promoters. Therefore, mutating/deleting these cis elements will likely lead to complex/mixed phenotypes. Notably, computer analyses of the regions upstream of the proximal and distal M2 gene transcription initiation sites reveal the presence of several potentially interesting transcription factor binding sites (Fig. 1B). Upstream of the proximal promoter there is a candidate NF-κB binding site, along with three closely spaced Ets-2 binding sites. Upstream of the distal promoter are candidate SP-1, c-Jun, and 2 MEF-2 binding sites, reminiscent of cis-regulatory elements present in the EBV immediate-early BZLF1 promoter. Thus, the organization of candidate cis elements upstream of the distal promoter is consistent with the hypothesis that this promoter may be differentially utilized during virus reactivation from latency. Future mutagenesis studies, along with analyses of specific cellular factor binding, will be required to gain a complete understanding of how M2 gene transcription is regulating during establishment of and reactivation from latency. The analyses of M2 gene transcription presented here should provide an important framework for future studies on this important MHV68 latency-associated gene product.