Polyomavirus ALTOs, but not MTs, downregulate viral early gene expression by activating the NF-κB pathway

Abstract

Polyomaviruses are small, circular dsDNA viruses that can cause cancer. Alternative splicing of polyomavirus early transcripts generates large and small tumor antigens (LT, ST) that play essential roles in viral replication and tumorigenesis. Some polyomaviruses also express middle tumor antigens (MTs) or Alternate LT ORFs (ALTOs), which are evolutionarily related but have distinct gene structures. MTs are a splice variant of the early transcript whereas ALTOs are overprinted on the second exon of the LT transcript in an alternate reading frame and are translated via an alternative start codon. Merkel cell polyomavirus (MCPyV), the only human polyomavirus that causes cancer, encodes an ALTO but its role in the viral lifecycle and tumorigenesis has remained elusive. Here, we show MCPyV ALTO acts as a tumor suppressor and is silenced in Merkel cell carcinoma (MCC). Rescuing ALTO in MCC cells induces growth arrest and activates NF-κB signaling. ALTO activates NF-κB by binding SQSTM1 and TRAF2&3 via two N-Terminal Activating Regions (NTAR1+2), resembling Epstein-Barr virus (EBV) Latent Membrane Protein 1 (LMP1).. Following activation, NF-κB dimers bind the MCPyV non-coding control region (NCCR) and downregulate early transcription. Beyond MCPyV, NTAR motifs are conserved in other polyomavirus ALTOs, which activate NF-κB signaling, but are lacking in MTs that do not. Furthermore, polyomavirus ALTOs downregulate their respective viral early transcription in an NF-κB and NTAR dependent manner. Our findings suggest that ALTOs evolved to suppress viral replication and promote viral latency and that MCPyV ALTO must be silenced for MCC to develop.

Keywords: Epstein-Barr virus; Merkel cell carcinoma; NF-κB; alternate LT open reading frame (ALTO); polyomavirus.

Figure 1.. MCPyV ALTO is silenced in MCC.

A) MCPyV early transcripts with MCC LT-truncating mutations for WaGa, MS-1, MKL-1, and MKL-2 cell lines indicated. Reading frames of coding sequences are indicated by blue (+1), green (+2), and yellow (+3). Untranslated sequences are represented by a solid black line. B) T antigen expression in 293A cells transfected with wt MCPyV genome and MCC cell lines.

Figure 2.. ALTO induces growth arrest in in MCC cells and activates NF-κB signaling.

A&B) Growth curves of MKL-1 and MKL-2 cells expressing RFP or ALTO following doxycycline treatment. Each data point represents the mean ± standard deviation of three or more technical replicates. Statistical significance was calculated by 2-way ANOVA. **** P<0.0001. C) Western blots of MKL-1 and MKL-2 cells expressing RFP or ALTO. Cells were harvested after 6 days of doxycycline treatment. D) Volcano plot of differentially expressed genes in MKL-2 ALTO vs. RFP cells treated with doxycycline for 48 hours from RNA sequencing of extracts from three replicates. A subset of NF-κB target genes is highlighted. E) Enrichment plot for Hinata et al. (32) NF-κB targets upregulated in keratinocytes gene set. F) Growth curves of MKL-2 cells expressing RFP, HA-ALTO or HA-RelA following treatment with doxycycline. Each data point represents the mean ± standard deviation of three or more technical replicates. Statistical significance was calculated by 2-way ANOVA. ***** P<0.0001. G) Western blots of MKL-2 RFP, HA-ALTO and HA-RelA cells treated with doxycycline for 6 days.

Figure 3.. ALTO binds SQSTM1 and TRAF2&3 via NTAR1+2 motifs to activate NF-κB signaling.

A) Western blots of MKL-2 cells expressing RFP, ALTO, TID or TID-ALTO. B) Western blots of streptavidin pulldown using lysates from MKL-2 TID and TID-ALTO cells biotinylated for 24 h prior to lysis. C) Volcano plot of proteins identified by mass spectrometry enriched in MKL-2 TID-ALTO vs TID samples. Proteins associated with LMP1’s CTAR1 and CTAR2 are highlighted in blue and red, respectively. D) Western blots of MKL-2 HA-ALTO variant cell lysates and anti-HA immunoprecipitations. E) Cartoon of EBV LMP1 and MCPyV ALTO variants. LMP1 CTAR2 and ALTO NTAR1 are highlighted in red. LMP1 CTAR1 and ALTO NTAR2 are highlighted in blue. Hydrophobic domains are highlighted in purple.

Figure 4.. ALTO downregulates MCPyV early gene expression via NF-κB signaling.

A) DNA pulldowns with MKL-2 HA-ALTO cell lysates incubated with biotinylated MCPyV NCCR or AmpR dsDNA coupled to streptavidin beads. B) MCPyV early luciferase in 293A cells transfected with HA-ALTO wt, HA-ALTO ΔNTAR1+2 or HA-RelA plasmids. Luciferase activity was normalized to CMV-driven renilla activity. Statistical significance was calculated by one-way ANOVA. * P<0.03, ** P<0.002. C) Western blots from 293A cells transfected with either HA-ALTO wt; HA-ALTO ΔNTAR1+2; HA-RelA; HA-RelB and FLAG-p52; or HA-RelA, HA-RelB and FLAG p52 plasmids. D) qRT-PCR analysis of relative LT and ST expression normalized to 36B4 expression in MKL-2 RFP, HA-ALTO wt or ΔNTAR1+2 cells treated with doxycycline for 6 days. The data show the mean ± the standard deviation of three technical replicates. Statistical significance was calculated by 2-way ANOVA. * P<0.03, ** P<0.003, *** P<0.0002, **** P<0.0001. E) Western blots of MKL-2 cells expressing RFP, HA-ALTO wt or ΔNTAR1+2 harvested after 6 days doxycycline induction. F) Growth curves of MKL-2 cells expressing RFP, HA-ALTO wt or ΔNTAR1+2. Each data point represents the mean ± standard deviation of three or more technical replicates. Statistical significance was calculated by 2-way ANOVA. **** P<0.0001.

Figure 5.. ALTO/NF-κB/early gene signaling network is conserved in ALTO-encoding polyomaviruses.

A) EBV LMP1, MCPyV ALTO, MuPyV MT and MuPyV ALTO constructs transfected in 293A cells in B). NTAR1/CTAR2 and NTAR2/CTAR1 motifs are highlighted in red and blue, respectively. C terminal hydrophobic domains are highlighted in purple. B) Western blots of 293A cells transfected with plasmids encoding EBV HA-LMP1, MCPyV HA-ALTO, MuPyV MT-HA or MuPyV HA-ALTO. C) Cartoon schematic comparing ALTOs and MTs from almipolyomaviruses. Amino acid sequences were aligned using Clustal Omega to create a phylogenetic tree. NTAR1 and NTAR2 motifs are highlighted in red and blue, respectively. C terminal hydrophobic domains are highlighted in purple. D) Western blots of 293A cells transfected with plasmids encoding PyV HA-ALTOs. E) Western blots of 293A cells transfected with plasmids encoding PyV HA-ALTO wt and ΔNTAR1 mutants. F) PyV early luciferase assays in 293A cells transfected with HA-ALTO wt or ΔNTAR1 plasmids. Luciferase activity was normalized to CMV-driven renilla activity. Statistical significance compared to empty vector control was determined by 2-way ANOVA. *** P<0.0002, **** P<0.0001.

Fig 6.. MCPyV ALTO activates the NF-κB pathway and suppresses tumorigenesis by downregulating LT and ST.

A) MuPyV MT, MCPyV ALTO, and EBV LMP1 binding partners, activated pathways, and their effect on tumorigenesis. B) ALTO activates NF-κB signaling by binding SQSTM1 and TRAF2&3 via two N-Terminal Activating Regions (NTAR1&2). Canonical NF-κB signaling via RelA appears to indirectly downregulate MCPyV early transcription. Meanwhile, non-canonical NF-κB signaling directly inhibits MCPyV early transcription with RelB binding to MCPyV NCCR.