Loss of 5-methylcytosine alters the biogenesis of vault-derived small RNAs to coordinate epidermal differentiation

Abstract

The presence and absence of RNA modifications regulates RNA metabolism by modulating the binding of writer, reader, and eraser proteins. For 5-methylcytosine (m⁵C) however, it is largely unknown how it recruits or repels RNA-binding proteins. Here, we decipher the consequences of m⁵C deposition into the abundant non-coding vault RNA VTRNA1.1. Methylation of cytosine 69 in VTRNA1.1 occurs frequently in human cells, is exclusively mediated by NSUN2, and determines the processing of VTRNA1.1 into small-vault RNAs (svRNAs). We identify the serine/arginine rich splicing factor 2 (SRSF2) as a novel VTRNA1.1-binding protein that counteracts VTRNA1.1 processing by binding the non-methylated form with higher affinity. Both NSUN2 and SRSF2 orchestrate the production of distinct svRNAs. Finally, we discover a functional role of svRNAs in regulating the epidermal differentiation programme. Thus, our data reveal a direct role for m⁵C in the processing of VTRNA1.1 that involves SRSF2 and is crucial for efficient cellular differentiation.

Fig. 1

Methylation of VTRNA1.1 by NSUN2 determines the biogenesis of svRNA4. a–c Correlation of site-specific methylation (m⁵C) levels (a, b) and methylation level at all covered cytosines in VTRNA1.1 (c) in NSUN2-/- cells infected with the empty (e.) vector (ctr), the enzymatic dead construct K190M or the wild-type NSUN2 construct. NSUN2-specific sites are highlighted in red. d, e Heatmaps (upper panels) and methylation level (bottom panels) of VTRNA1.1, VTRN1.3, RPPH1, and HECTD1 in infected NSUN2-/- cells (d), human embryonic fibroblasts (H9) and HEK293 cells (e). Shown are five independent bisulfite conversion experiments. f Schematic illustration of NSUN2-dependent methylation (CH₃) of VTRNA1.1 and the small regulatory non-coding fragments svRNA4. g Abundance of svRNA4 in the presence (+/−) and absence (−/−) of NSUN2. Methylation dead NSUN2-mutant constructs (C271A; C321A) fail to rescue svRNA4 levels in NSUN2−/− cells. Error bars indicate s.d. (n = 3 qRT-PCR reactions). ***p < 0.001 unpaired student’s t-test. h Log₂ fold-change of NSUN2 in the indicated cells compared to NSUN2+/+control cells measured by ribosome profiling. Source data are provided as a Source Data file

Fig. 2

SRSF2 preferentially binds un-methylated human VTRNA1.1. a Of the 144 common proteins binding to VTRNA1.1 in two different RP-SMS experiments, a small number bound methylated (red) or unmethylated (blue) VTRNA1.1 with higher affinity. b Gene Ontology (GO) analyses of the 144 commonly bound proteins. c Western blot and Coomassie stain for SRSF2 in HeLa cell lysates pulled-down with agarose beads coupled to methylated (m⁵C69) or un-methylated (C69) Vault-RNA1.1 (upper panel). hnRNP A1 serves as a loading and RNA-binding control (lower panel). Numbers indicate band intensity vs. loading control. d Location of the putative SRSF2 RNA-binding motifs (RRM1 and RRM2) in VTRNA1.1 (upper panel) and RNA pulldowns using wildtype or mutated (C69A; C88U) VT-RNA1.1-constructs to confirm both putative SRSF2 binding sites are necessary for SRSF2 binding. Shown is mean and range (n = 2 independent experiments). Quantification in (c, d) was done using ImageJ. e EMSA assay using methylated (m⁵C69) and unmethylated (C69) VTRNA1.1 to measure binding of recombinant SRSF2. f Quantification of (e). Error bars indicate s.d. (n = 3 experiments). **p < 0.01, *p < 0.05 students t-test. Source data are provided as a Source Data file

Fig. 3

Methylation-guided VT-RNA1.1 processing is altered in the absence of SRSF2. a, b Western blot for endogenous SRSF2 (a) and NSUN2 (b) in NSUN2-expressing (+/+, +/−) and -lacking (−/−) human fibroblasts (from two patients). Tubulin served as a loading control. c qRT-PCR measuring SRSF2-bound VTRNA1.1 normalised to the control (ctr; Rabbit serum conjugated with Dynabeads) after recovering the pulled down RNA in the SRSF2 immunoprecipitation. d Western blot for SRSF1 and SRSF2 in NSUN2−/− human fibroblasts infected with shRNAs (1.1, 1.6, 2.4, 2.5). Tubulin served as a loading control. e Fold-change (FC) of svRNA1 and 4 abundances after knock-down of SRSF1 and 2, relative to NSUN2−/− cells infected with the empty vector (e.V.). Shown are the pooled values using the two shRNA constructs shown in (d). Error bars indicate s.d. (n = 3-6 qRT-PCRs). **p < 0.01 students t-test. Source data are provided as a Source Data file

Fig. 4

VTRNA1.1 methylation and processing are altered during cell differentiation. a Treatment regime of keratinocytes using calcium switch assay. b, c qRT-PCR to measure RNA levels of up-regulated (b) and down-regulated (c) markers at 2 and 6 days after calcium treatment compared to the 0 day control. Error bars indicate s.d. (n = 3 qRT-PCRs) ****p < 0.0001, ***p < 0.001, **p < 0.01 multiple t-tests. d Methylation levels (n = 5 BS conversion reactions) at cytosines in RNA isolated from undifferentiated (undiff) and differentiated (diff) primary HK shown as box plots showing all points with minimum to maximum values. ****p < 0.0001 Mann Whitney test. e Log₂ coverage (n = 5 BS conversion reactions) of sites in RNA isolated from undifferentiated (undiff) and differentiated (diff) primary HK. f Correlation between methylation levels at cytosines in undifferentiated and differentiated primary HK. Elevated methylation levels at tRNAs (examples in blue) and VTRNA1.1 (red). g, h Methylation levels in VTRNA1.1 (g) and tRNA Leu CAA (h) in the indicated cells (left hand panels) and heat maps (right hand panels) showing methylation levels in the individual replicates. i Light microscope image comparing the morphology of primary HK transfected with a control siRNA (Ctr) or svRNA4 after 4 days of differentiation in high CaCl₂. Scale bar: 50 μm. j Western blot detecting KRT10 and OVOL1 in HK transfected with Ctr siRNA or svRNA4 four days after calcium-induction. Tubulin served as loading control. k, l Treatment regimes and transfection (upper panels) of svRNA4 (k) or anti-svRNA4 (l) and qRT-PCR (lower panels) to measure RNA levels of the indicated markers 4 days after calcium treatment. Error bars indicate s.d. (n = 3 qRT-PCRs). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 student’s t-test. Source data are provided as a Source Data file

Fig. 5

SRSF2 is required for cell cycle and survival of undifferentiated cells. a Treatment regime and transfection of differentiating primary human keratinocytes. b Western blot detecting SRSF2 after treatment with Srsf2 siRNA. Tubulin (TUBB) serves as loading control. c Quantification of RNA expression levels of Ovol1, Vtrna1.1, Tgm, Krt10, and Ivl after 6 days of calcium-induced differentiation vs. untreated control (0 days). FC: Fold-change. Error bars indicate s.d. (n = 3 qRT-PCRs). **p < 0.01 two-way ANOVA. d, e Log₂ RNA fold-change (FC) of differentiation markers (d) and cell cycle regulators (e) after 48 and 72 h of Srsf2 knock-down. Data shown as box plot with mean showing all data from minimum to maximum (n = 4). ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05 Two-way ANOVA. f Cell cycle distribution of primary HK transfected with the Srsf2 siRNA for 72 h. Error bars Data shown as box plot showing all data from minimum to maximum (n = 10 Flow sorts). ****p < 0.0001 Two-way ANOVA. g Light microscope images of HK transfected with a control (ctr) siRNA (upper panels) and a Srsf2 siRNA (lower panels) after 48 and 72 h in low calcium medium. Scale bar: 50 μm. h Small RNA qRT-PCR measuring the abundance of svRNA4 in primary HK transfected with the indicated siRNA constructs. Data shown as box plot with mean showing all data from minimum to maximum. Error bars represent s.d. (n = 4 qRT-PCRs). **p < 0.01 unpaired student’s t-test. i Illustration how levels of NSUN2, svRNA4, SRSF2, and OVOL1 change upon terminal differentiation in keratinocytes. Source data are provided as a Source Data file

Fig. 6

Summary of VTRNA1.1 processing into svRNA4. a Expression of both NSUN2 and SRSF2 (e.g. in progenitor cells) result in high levels of VTRNA1.1 methylation (CH₃) and high levels of svRNA4 and repressed differentiation. b No NSUN2 in the presence of SRSF2 suppresses formation of svRNA4 and allows differentiation. c Lack of expression of both NSUN2 and SRSF2 release VTRNA1.1 from SRSF2 binding and increases the levels of svRNA4