Fig 2: SND1 is physically associated and colocalized with MHC-I HC.(A) Immunopurification and mass spectrometry of SND1-containing protein complexes. Cellular extracts from HeLa cells stably expressing SND1-FLAG were immunopurified with anti-FLAG affinity beads and eluted with FLAG peptide. The elutes were resolved on SDS-PAGE and silver-stained. The protein bands on the gel were recovered by trypsinization and analyzed by mass spectrometry. (B) Co-IP analysis of the association between SND1 and HLA-A. Whole-cell extracts from HeLa cells with SND1-FLAG expression were immunoprecipitated with anti-FLAG beads, followed by Western blot with antibodies against the HLA-A. (C) Co-IP analysis of the association between SND1 and HC10. Whole-cell extracts from HeLa cells were immunoprecipitated with anti-HC10, followed by immunoblot (IB) with antibodies against the SND1. (D) Cellular extracts from HeLa cells were immunoprecipitated with anti-SND1 antibody, followed by Western blot with antibodies against the indicated proteins. (E) Duolink in situ PLA was adopted for detecting the association between SND1 and HLA-A. Two PLA probes were designed to respectively recognize either mouse or rabbit antibody against SND1 or HLA-A. Immunoglobulin G (IgG) was used as staining control. Scale bar, 20 µm. (F) Immunostaining and confocal microscopic analysis of subcellular colocalization of SND1 and HLA-A–FLAG (C terminus) in HeLa cells. HeLa cells were fixed and immunostained with antibodies against the indicated proteins. Scale bar, 10 µm. (G) GST pull-down analysis of the bacterially produced GST-fusion protein containing full-length SND1 (GST-SND1), SN domain (GST-SN), and TSN domain (GST-TSN) involved in the interaction with in vitro–translated HLA-A from rabbit reticulocytes. Coomassie blue staining for GST-fusion proteins refers to fig. S1A. aa, amino acid. (H) GST pull-down analysis of the different domains of HLA-A involved in the interaction with SND1. The His-SND1 and sample of GST-tagged different domains of HLA-A were purified from E. coli bacteria cells. Coomassie blue staining for GST-fusion proteins refers to fig. S1B. (I) Immunoprecipitation analysis of the domains involved in the interaction between SND1 and HLA-A with FLAG-tagged deletion mutants of SND1 purified from HeLa SND1-KO cells. The immunoprecipitation of FLAG refers to fig. S1C. (J) The spatial conformation of SND1-HLA-A complex predicted by the database of ZDOCK (http://zdock.umassmed.edu/) was further analyzed using the Gromacs package. The structural stability and binding energy refer to fig. S1 (E and F).

Fig 3: Loss of SND1 in murine tumors limits tumor size with more CD8+ T cell infiltration in vivo.(A) Three clones of B16F10 cells with stable depletion of SND1 by CRISPR-Cas9 system were collected, followed by IB for murine MHC-I (H2Kb). (B) Two clones of B16F10 cells with stable depletion of SND1 by CRISPR-Cas9 system were analyzed by flow cytometry for murine MHC-I (H2Kb/H2Db) using antibody simultaneously against H2Kb/H2Db. (C to E) 5 × 105 of either WT or SND1-KO B16F10 cells were subcutaneously transplanted into C57BL/6 mice. The tumor growth was monitored at the indicated times. C57BL/6 mice were sacrificed at day 11. Tumors were removed and photographed. The tumor tissues were weighed and plotted. Data are presented as means ± SD; n = 5 tumors for each group. *P < 0.05, two-tailed t test. (F) Immunofluorescence images of CD4+ T and CD8+ T cells in B16F10 tumor sections (scale bar, 20 μm). (G) C57BL/6 mice injected with equal numbers of WT or SND1-KO B16F10 cells were sacrificed at day 11. The digested tumor suspensions stained with antibodies against CD8 and CD45.2 (pan-leukocyte marker) were subjected to flow cytometry. (H to J) Percentages of infiltrating CD45.2+ cells and CD8+ T cells among total tumor tissue–derived cells and the percentage of infiltrating CD8+ T cells among total CD45+ leucocytes. n = 5 tumors for each group. *P < 0.05 and **P < 0.01, by unpaired t test. The experiments were performed and repeated at least three times, independently. (K) The percentage of infiltrating PD-1+ CD8+ T cells among total CD8+ T cells. n = 5 tumors for each group. n.s., not significant.

Fig 4: SND1 RNA-binding sites specifically overlap m6A peaks (a, b) Genome sequencing tracks from latent (0 hr) and lytic (8 hr and 20 hr post-reactivation) TREx BCBL1-Rta cells depicting input (green) and RIP (blue) coverage.m6A-IP reads (red) are also shown. Blue boxes indicate exons while blue lines represent introns. SND1 overlaps with m6A peaks are evident in introns (ARL17A) and 5’UTRs (ABCD1). Note that in ABCD1, SND1 binding and m6A peaks are both reduced at 8 hr post-reactivation while at 20 hr post-reactivation methylation and SND1 binding signal are both lost with decreasing expression in inputs. (c, d) Dynamic SND1 binding to m6A-modified regions in SND1 target transcripts during KSHV infection. Genome tracks depicting sequencing read coverage from input (green) and RIP (blue) samples. m6A-IP reads (pink) are also shown. Note that in contrast to ABCD1, RHOU/DUSP5P1 remain highly expressed even at 20 hr post-reactivation, suggesting the coupled loss of methylation/SND1-binding is independent of the ability to detect these events due to loss of expression. (e) SND1 regions consistently enriched across the three time points studied (0 hr, 8 hr and 20 hr) were defined by applying a fold change RIP/input >2 and>50 mean reads per region (FDR < 1%). Consistently SND1-unbound regions throughout KSHV infection were defined using a fold change RIP/input <0.5 and>50 mean reads per region (FDR < 1%). For all m6A overlap analysis, the cut-off for an m6A peak was defined as having 50 read paired at the tallest point in the m6A peak and a 2-fold enrichment of m6A-IP reads over input reads using a FDR < 1%.

Fig 5: SND1 deletion promotes specific antigen presentation and enhances antitumor immunity.(A) B16F10 cells with stable depletion of SND1 by CRISPR-Cas9 system were stably transfected with OVA vector, followed by IB. (B) B16F10-OVA with SND1 deficiency was analyzed by flow cytometry for murine MHC-I (H2Kb/H2Db). (C to E) OT-I mice were injected with equal numbers of WT or SND1-KO B16F10-OVA cells, and tumor growth was observed over time. Then tumors were removed, photographed, and weighted. *P < 0.05 and **P < 0.01. (F) Flow cytometry was used for the analysis of CD45.2+ leucocyte and CD8+ T cell infiltration in tumor tissues. (G to I) Percentages of infiltrating CD45.2+ leucocytes and CD8+ T cells among total tumor tissue–derived cells and the percentage of infiltrating CD8+ T cells among total CD45.2+ leucocytes. n = 5 tumors for each group. **P < 0.01 and ***P < 0.001, by unpaired t test. (J) CD8+ T cells were purified from spleens of tumor-bearing OT-I mice and stimulated with 257 to 264 (SIINFEKL) peptide of OVA for 24 hours. Percentages of IFNγ+CD8+ T cells among total CD8+ T cells in the culture system were measured by flow cytometry. (n = 5, **P < 0.01). The experiments were repeated two times independently. (K) CD8+ T cells recognizing specific peptide of OVA (SIINFEKL) were purified from spleens of OT-I and then cocultured with WT or SND1-KO B16F10 cells stably expressing OVA (CD8+ T:B16F10-OVA, 10:1). Representative images were taken under a bright field at different time points. Scale bar, 20 μm. (L) In vitro comparison of cytolysis rates against CD8+ T cells purified from spleens of OT-I mice between WT and SND1-KO B16F10-OVA cells at different cell rates of CD8+ T (effector cells) to B16F10 (target cells) with/without anti–MHC class I antibodies (Ab) (E:T, 5:1, 10:1, 15:1, or 20:1). A lactate dehydrogenase–releasing cytotoxicity assay was performed to measure the cytolysis efficiency of CD8+ T cells on tumor cells. Each bar represents mean ± SD for biological triplicate experiments. ****P < 0.0001, two-way analysis of variance (ANOVA).