Fig 1.
Identification of Merkel cell polyomavirus circRNAs.
(A) (Top panel) Schematic representation of potential circRNAs identified by the vircircRNA for MCPyV. V-shaped lines above the map indicate forward splicing events. Elliptical arcs below the map indicate predicted backsplicing. (Bottom panel) Red rectangles indicate ALTO open reading frames. Bold numbers above the map indicate the first base of ALTO start and stop codons, italic numbers below the map indicate the positions of exon boundaries. (B) (Left panel) RT-PCR analysis of four VP-MCC cell lines (MKL1, MKL-2, MS-1 and WaGa) with and without RNase R treatment. (Middle panel) Sanger sequencing of inverse PCR products from WaGa cells confirmed the backsplice junction of the predicted circRNAs. (Right panel) Schematic of the two different circALTO forms, we termed circALTO1 and circALTO2 based on the product length. (C) qRT-PCR analysis for circALTOs from VP-MCC (MKL-1, MKL-2, MS1, and WaGa) and VN-MCC (MCC13, MCC26 and UISO) after RNase R treatment. ACTB served as the normalization control. Values are the mean of three technical replicates; bars represent standard deviation. (D) Northern blot of total RNA after with and without RNase R treatment from VP-MCC cell line WaGa and VN-MCC cell line UISO using an ALTO-specific probe. Arrow indicates RNase R resistant band consistent with circALTO2. (E) Endpoint PCR confirmed the presence of circALTO2 in patient MCC tumors (upper panel), but not a non-malignant skin control or VN-MCC cell lines. circHIPK3 served as a control for RNA integrity (bottom panel).
Fig 2.
(A) The expression of circALTOs and its linear mRNA counterparts after treatment with Actinomycin D at the indicated time points in WaGa cells by qRT–PCR analysis. Both circALTO and linear ALTO were normalized to 18S and then compared to levels at the pre-treatment (0 h) time point. Error bars calculated the SD from 3 biological replicates. The P value was determined using two-tailed t-test. (B) WaGa were fractionated and analyzed by qRT-PCR. MALAT1 and ACTB served as nuclear and cytoplasmic fractionation controls respectively. Values are normalized to the enriched fraction. Error bars were calculated from 3 biological replicates. (C) qRT-PCR of m6A or IgG control RNA immunoprecipitation (RIP) of VP-MCC cell lines WaGa. SON, m6A RNA IP control. Error bars represent SD (n = 3 biological replicates).
Fig 3.
MCPyV circALTO can be translated.
(A) CircALTOs do not inhibit MCPyV miRNA function in vitro. 293 cells were transfected with Renilla luciferase reporter with a 300 nucleotide region complementary to the MCPyV miRNA and the indicated plasmids including a MCPyV-miRNA-M1 expression plasmid (derived from isolate MCV350), pCDNA3.1 control vector, pCDNA circALTO1 (pcircALTO1) or pCDNA circALTO2 (pcircALTO2) expression plasmid. Firefly luciferase served as a transfection control and Renilla luciferase levels are plotted normalized relative to firefly luciferase levels (n = 2 biological replicates). The P value was determined by unpaired two-tailed t-test. (B) Western blot (WB) for FLAG from 293T cells co-transfected with a pCDNA vector control plasmid or pCDNA3.1-Flag-circALTO1/2. HSP90, loading control. (C) 293T cells were transfected with vector, the Flag-circALTO1/2 plasmids. After 48 hours, cells were fixed and stained for FLAG (green) and DAPI (blue). Scale bar = 50 μm. (D) WB for ALTO protein from 293T cells co-transfected with circALTO1 (untagged) and siRNAs targeting the body of the ALTO ORF, a segment unique to linear ALTO transcripts, or an siRNA targeting the unique junction of circALTO1. (E) WB for ALTO protein from 293T cells co-transfected with circALTO2 (untagged). (F) Immunoblotting analysis of 293T cells were transfected with empty vector, pCDNA-circALTO1 (Flag), or pcircALTO2 (Flag) with or without MCPyV miRNA expression vector. Co-transfection with the MCPyV miRNA expression vector results in decreased expression of ALTO. (G) Endpoint inverse RT-PCR detection of analysis for circALTO from 293T cells transfected with pcircALTO1/2 (WT) and the indicated constructs harboring splice sites mutations (SASD) (see S2A Fig). (H) Inverse qRT-PCR analysis of circALTO from cells with WT constructs and constructs harboring SASD mutant (see S2A Fig). ACTB served as the internal control. Error bar represents the standard deviation of 3 independent experiments. (I) Western blots for Flag-ALTO from cells with WT constructs and constructs harboring the SASD mutation. HSP90, loading control. The P value was determined by unpaired, two-tailed t-test, *<0.05, **<0.01.
Fig 4.
MCPyV circALTO increases transcription from specific promoters and can modulate host cell gene expression.
(A) 293 cells were co-transfected with Renilla luciferase reporter or Firefly luciferase reporter and with either pCDNA vector control, pcircALTO1, or pcircALTO1-SASD (n = 3 biological replicates). The P value was determined by unpaired two-tailed t-test. (B) Western blot for GFP and FLAG from 293T cells which were co-transfected with pCDNA-GFP and indicated plasmids including pCDNA control vector, pcircALTO1 (Flag), or pircALTO2 (Flag) plasmids. HSP90 is the loading control. (C) 293T cells were co-transfected with pCDNA-GFP vector and indicated plasmids. IF images are representative of GFP expression after 48 hours. Scale bar = 200 μm. (D) qRT-PCR analysis for GFP from cells with pCDNA-GFP constructs and indicated plasmids. Expression was normalized against ACTB. Error bar representative of 3 independent experiments. (E) A construct in which all potential ATG start codons have been mutated, circALTO1-ΔATG, was generated. Western blot for GFP and FLAG from 293T cells were transfected with pCDNA-GFP reporter vector and pCDNA control vector, Flag-circALTO1, and pcircALTO1-ΔATG plasmids. HSP90 is the loading control. (F) qRT-PCR analysis for GFP transcripts from 293T cells as described in (E). Expression was normalized against ACTB. Error bars = SD from one biological replicate. Results are representative of 3 independent experiments. (G) Western blot for GFP and FLAG from 293T cells which were co-transfected with the SV40 early or late promoter cloned into pU-5864 and the indicated plasmids including pCDNA control vector or pcircALTO1 (Flag). HSP90 is the loading control. (H) qRT-PCR analysis for GFP from cells co-transfected with the SV40 early or late promoter cloned into pU-5864 and the indicated plasmids including pCDNA control vector or pcircALTO1 (Flag). HSP90 is the loading control. Expression was normalized against ACTB. Error bar representative of 3 biological replicates. (I) Principal component analysis (PCA) of pCDNA control and pcircALTO1 transcriptomes (48 hours) reveals distinct clustering of ALTO compared with vector control. (J) Volcano plot (single gene) of differentially expressed genes in pcircALTO1 compared with pCDNA vector. Significantly upregulated genes are indicated in red (log2 FC>1.5, adjusted p<0.001). A subset of genes previously implicated in viral infections are labeled. (K) KEGG pathway signatures induced in ALTO1 expressing cells compared to vector control. The P value was determined by unpaired, two-tailed t-test, *<0.05, **<0.01.
Fig 5.
Characterization of exosomes from WaGa cells.
(A) Western blot analysis of exosomal marker proteins in extracellular vesicles isolated from VP-MCC cell line WaGa and VN-MCC cell lines MCC26. CD63 and CD81 are exosomal marker proteins. The absence of calnexin and β-tubulin exclude contamination from the cytoplasm and cellular organelles. (B) Nanoparticle tracking analysis of extracellular vesicles isolated from WaGa cells are consistent with exosomes. (C) Representative TEM images of exosomes isolated from WaGa cells show the expected size and cup-shaped morphology described for exosomes. (D) RT-PCR analysis of RNase R treated total RNA prepared from exosomes of WaGa and MCC26 cells reveals the presence of circALTO in the exosomes. (E) qRT-PCR analysis the ratio of circRNA to mRNA between WaGa cells and cell-derived exosomes. Both circALTO and linear ALTO were first normalized to ACTB, then the ratio of circALTO:linear ALTO was set as 1. Error bars represent the SD (n = 3 biological replicates). Two-tailed t-test was performed, **<0.01. (F) Fluorescence of UISO cells incubated with DiI labeled exosomes from WaGa or Dil labeled PBS control by fluorescent microscopy. Scale bar = 50 μm.
Fig 6.
Exosomal circALTO can mediate ALTO expression in HeLa cells and keratinocytes.
(A) Western blot analysis of exosomal marker in exosomes extracted from 293 T cells transfected with pCDNA vector control or Flag-circALTO1/2. CD63 and CD81 are exosomal marker proteins, while calnexin and β-tubulin exclude contamination from the cytoplasm and cellular organelles. (B) Size distribution analysis of total exosomes isolated from the Flag-circALTO1 (upper panel) and Flag-circALTO2 (bottom panel) expressed 293T cells. (C) TEM images of exosomes isolated from circALTO2 expressed 293T cells confirm expected cup-shaped morphology. (D) RT-PCR analysis of total RNA from exosomes of transfected with pCDNA vector control or Flag-circALTO1/2 293 T cells with and without RNase R treatment. (E) Fluorescence of HeLa cells incubated with DiI labeled exosomes from indicated cells by fluorescent microscopy. Scale bar = 50 μm. (F) Purified exosomes from 293T cells transfected with Flag-circALTO1/2 were transferred to HeLa cell. Expression of Flag was assessed by WB. HSP90 is the loading control. (G) The indicated purified exosomes were transferred to HeLa cell. After 70 hours, cells were transfected with pCDNA-GFP (48 hours), then lysates were generated to determine expression of Flag and GFP by WB. HSP90 is the loading control. (H) HeLa cells treated as previously described (G) were imaged by IF. Scale bar = 200 μm. (I) The indicated purified exosomes were transferred to HeLa cell. After 70 hours, cells were transfected with pCDNA-GFP (48 hours), then lysates were generated to determine expression of Flag and GFP by WB. HSP90 is the loading control.
Fig 7.
Conservation of circALTO in TSPyV.
(A) Schematic representation of a predicted circALTO in TSPyV. Diagram highlights the location of the divergent primers used to detect TSPyV circALTO. (B). RT-PCR analysis of cDNA prepared from total RNA prepared with random hexamer primers. 293T cells were transfected with pCDNA vector or a construct containing the potentially backspliced early region of TSPyV. Total RNA was first treated with RNase R. (C) Western blot for TSPyV ALTO from 293T cells transfected with a pCDNA vector or the TSPyV circALTO construct. Lysates were probed with a polyclonal antibody generated against 2 peptides specific to the TSPyV ALTO, pAbMT [14]. HSP90, loading control. (D) Images from cases of TS identified in this study. Clinical images (top) from TS patients 1, 2, and 5. The photos show the characteristic folliculocentric papules on the face and neck. Histological images (bottom) show the H&E stained image of lesional skin from patients 1, 2, and 5. Affected skin shows dysmorphic hair follicles. Patient 1 inset shows expansion of inner root sheath (IRS). Patient 2 inset shows enlarged trichohyaline granules in IRS cells. (E) Total RNA was extracted from lesional biopies and cDNA prepared with random hexamer primers. RT-PCR analysis using indicated primers revealed the presence of linear small T (sT) in 4/5 TS samples and circALTO in 3/5 TS samples. A TSPyV negative squamous cell carcinoma was used as a negative control. 293T cells transfected with TSPyV circALTO was used as a positive control for circALTO. (F) Sanger sequencing of PCR products from 293T transfected cells and PCR product from patient 2 confirmed the expected backsplice junction.