Tracing dynamic changes of DNA methylation at single-cell resolution

Abstract

Mammalian DNA methylation plays an essential role in development. To date, only snapshots of different mouse and human cell types have been generated, providing a static view on DNA methylation. To enable monitoring of methylation status as it changes over time, we establish a reporter of genomic methylation (RGM) that relies on a minimal imprinted gene promoter driving a fluorescent protein. We show that insertion of RGM proximal to promoter-associated CpG islands reports the gain or loss of DNA methylation. We further utilized RGM to report endogenous methylation dynamics of non-coding regulatory elements, such as the pluripotency-specific super enhancers of Sox2 and miR290. Loci-specific DNA methylation changes and their correlation with transcription were visualized during cell-state transition following differentiation of mouse embryonic stem cells and during reprogramming of somatic cells to pluripotency. RGM will allow the investigation of dynamic methylation changes during development and disease at single-cell resolution.

Figure 1. An active minimal Snrpn promoter can be repressed in cis by means of spreading of DNA methylation into the promoter region

(A) Schematic representation of the sleeping-beauty based vectors. Endogenous CpG Islands (CGI) of Dazl and Gapdh genes were cloned upstream of a minimal Snrpn promoter region - driving GFP. Open circle lollipops schematically represent individual unmethylated CpG. (B) Flow cytometric analysis of V6.5 mESCs cultured in serum + LIF, following stable integration of unmethylated Gapdh and Dazl reporter vectors, demonstrating robust repression of GFP signal in the Dazl reporter cells over time. Shown are the mean percentages of GFP negative cells ± STD of two biological replicates. (C) Phase and fluorescence images of the sorted V6.5 mESCs, comprising stable integration of the Gapdh (left) and Dazl (right) vectors following prolonged culturing for 7 weeks. (D and E) Bisulfite sequencing analysis of the stably transfected Gapdh (D) and Dazl (E) reporter cell lines was performed on the gene promoter-associated CGI (left) and the downstream Snrpn promoter region (right). Open circles represent unmethylated CpGs; Filled circles - methylated CpGs. See also Figure S1.

Figure 2. An in vitro repressed Snrpn promoter can be reactivated in cis by means of spreading of DNA demethylation into the promoter region

(A) Schematic representation of an in vitro methylated sleeping-beauty based vectors. Closed circle lollipops schematically represent individual methylated CpG. (B) Phase and fluorescence images of the stably integrated V6.5 mESCs, harboring Gapdh (left) and Dazl (right) in vitro methylated vectors, following one week of antibiotics selection. (C and D) Flow cytometric analysis of the proportion of GFP positive cells in V6.5 mESCs, stably integrated with either Gapdh (C) or Dazl (D) in vitro methylated vectors, following 2 weeks in culture. (E and F) Bisulfite sequencing analysis of the stably transfected Gapdh (E) and Dazl (F) reporter cell lines, was performed on the gene promoter-associated CGI (left) and the downstream Snrpn promoter region (right). (G) Flow cytometric analysis of the proportion of GFP positive cells in V6.5 mESCs and Dnmt1 KO mESCs, stably integrated with in vitro methylated Dazl reporter vector. (H) Bisulfite sequencing analysis of sorted GFP positive Dnmt1 KO mESCs, stably integrated with in vitro methylated Dazl reporter vector. (I) Flow cytometric analysis of the proportion of GFP negative cells in control V6.5 mESCs, mESCs deficient for both Dnmt3a and Dnmt3b (Dnmt3ab KO) and V6.5 mESCs cultured in 2i + LIF, which were stably integrated with unmethylated Gapdh (upper panel) and Dazl (lower panel) reporter vectors. See also Figure S1.

Figure 3. Generation of DNA methylation reporter cell lines for endogenous gene promoters

(A) CRISPR/Cas-based strategy used to integrate the DNA methylation reporter into the endogenous promoter region of Gapdh and Dazl genes. TSS - transcription start site. Green sequence - endogenous CGI region; Black sequence - targeting CRISPR; Red sequence PAM recognition site. (B) Flow cytometric analysis depicting the mean GFP intensity of randomly picked clones following antibiotic selection of both Gapdh (upper panel) and Dazl (lower panel) targeted V6.5 mESCs. (C) Flow cytometric analysis of the proportion of GFP positive cells in two representative clones correctly targeted with the methylation reporter at the promoter region of Gapdh (D) Bisulfite sequencing analysis was performed on mESCs harboring the DNA methylation reporter in Gapdh promoter region. For each cell line, the PCR amplicon (marked with dashed line) includes both the endogenous CGI (left) and the downstream integrated Snrpn promoter region (right). (E) Flow cytometric analysis of the proportion of GFP positive cells in two representative clones correctly targeted with the methylation reporter at the promoter region of Dazl. (F) Bisulfite sequencing analysis was performed on mESCs harboring the DNA methylation reporter in Dazl promoter region. For each cell line, the PCR amplicon (marked with dashed line) includes both the endogenous CGI (left) and the downstream integrated Snrpn promoter region (right). See also Figure S2.

Figure 4. Generation of DNA methylation reporter cell lines for the pluripotent-specific miR290 and Sox2 SE regions

(A) Regional view depicting the DNA methylation (upper panel) and chromatin (lower panel) landscape of miR290 upstream pluripotent-specific SE. Shown are average methylation levels and enrichment of chromatin marks in mouse undifferentiated cells (green) and in adult tissues (gold), with respect to the genomic organization of the genes. DNA methylation varies from 1-hypermethylated to 0- hypomethylated; Characteristic clusters of typical enhancer marks and binding of tissue-specific TF determine the SE region (light blue). (B) CRISPR/Cas-based strategy used to integrate the DNA methylation reporter into the endogenous SE region. HR - homologous recombination. Green sequence - endogenous miR290 CpG region; Black sequence - targeting CRISPR; Red sequence PAM recognition site. (C) Phase and fluorescence images of correctly integrated DNA methylation reporter cell lines for miR290 (upper panel) and Sox2 (lower panel) endogenous SE regions. GFP marks endogenous expression levels of Nanog, whereas tdTomato reflects the endogenous DNA methylation levels at both miR290 and Sox2 SE regions. (D) Bisulfite sequencing analysis was performed on undifferentiated mESCs harboring the DNA methylation reporter in either miR290 SE region (upper panel) or Sox2 SE region (lower panel). For each cell line, the PCR amplicon (marked with dashed line) includes both the endogenous CGI (left) and the downstream integrated Snrpn promoter region (right). See also Figure S3.

Figure 5. Dynamics of de novo DNA methylation of miR290 and Sox2 SE regions upon in vitro differentiation

(A) Schematic representation of the RA-based differentiation protocol used on miR290 and Sox2 reporter cell lines. GFP marks endogenous expression levels of Nanog, whereas tdTomato reflects the endogenous DNA methylation levels at both miR290 and Sox2 SE regions. (B) Flow cytometric analysis of the proportion of Nanog-GFP positive cells (X axis) and tdTomato positive cells (Y axis) during 7 days of differentiation of miR290 #21 (upper panel) and Sox2 #2 (lower panel) reporter cell lines. (C) Bar graph summarizing the proportion of the different cell populations during the course of 7 days RA differentiation for both miR290 #21 (upper panel) and Sox2 #2 (lower panel) reporter cell lines. Data represents two biological replicates. R - tdTomato ; G - GFP. (D and E) Bisulfite sequencing analysis on the three main cell populations - sorted at 48 hours following initial treatment with RA. For both miR290 #21 (D) and Sox2 #2 (E) cell lines, the PCR amplicon (marked with dashed line) includes the endogenous CGI (left) and the downstream integrated Snrpn promoter region (right). R - tdTomato ; G - GFP. See also Figure S4.

Figure 6. Dynamics of DNA demethylation of miR290 and Sox2 SE regions during cellular reprogramming

(A) miR290 (upper panel) and Sox2 (lower panel) reporter chimeric experimental embryos (right embryo in each panel). As controls, Gapdh CGI reporter mESCs driving GFP and constitutively expressing tdTomato (Control Gapdh-GFP and tdTomato, respectively) were injected into host blastocysts (left embryo in each panel). (B) Schematic representation of the experimental procedure to monitor the dynamics of demethylation during reprogramming of miR290 and Sox2 reporter cell lines. GFP marks endogenous expression levels of Nanog, whereas tdTomato reflects the endogenous DNA methylation levels at both miR290 and Sox2 SE regions. (C) Flow cytometric analysis of the proportion of GFP positive cells (X axis) and tdTomato positive cells (Y axis) in P0 MEFs derived from miR290 #21 (left) and Sox2 #2 (right) chimeric embryos. (D) Bisulfite sequencing analysis was performed on P0 MEFs derived from miR290 #21 (upper panel) and Sox2 #2 (lower panel) chimeras. For each cell line, the PCR amplicon (marked with dashed line) includes both the endougenous CGI (left) and the downstream integrated Snrpn promoter region (right). (E) Analysis of the proportion of GFP positive cells (X axis) and tdTomato positive cells (Y axis) during the course of reprogramming of MEFs derived from miR290 #21 (upper panel) and Sox2 #2 (lower panel) chimeras. Shown are flow cytometric data from different time points following addition of dox supplemented with 3C culture condition. (F) Representative images of established miR290 and Sox2 iPSC lines, derived from sorted double positive (tdTomato⁺ / GFP⁺) colonies. (G) Bisulfite sequencing analysis was performed on P2 iPSCs derived from miR290 #21 (upper panel) and Sox2 #2 (lower panel) MEFs. For each cell line, the PCR amplicon (marked with dashed line) includes both the endogenous CGI (left) and the downstream integrated Snrpn promoter region (right). See also Figure S5.