Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) (top) Immunoblot analysis of HIF1α in MCF7 cells exposed for 24 hours to 21% or 0.5% O 2 . ( b ) Comparison of the 7,153 HIF1β peaks detected in this study to a previously published dataset of ChIP-seq for HIF1β. ( c ) Frequency per basepair of the RCGTG motif inside and outside HIF1β peaks. On average, 4.6 HIF consensus sequences (RCGTG) are found in each HIF1β peak (average length 1,088 bps), whereas in the rest of the genome, 0.98 HIF consensus sequences (RCGTG) are found every 1,088 bps (Fisher exact test P<2.2 -16 ). ( d ) Ontology analysis of genes associated with the 7,153 HIF1β peaks detected in MCF7 cells by ChIP-seq for HIF1β. ( e ) Immunoblot analysis of HIF1β in MCF7 cells that are wild-type (WT) or knockout (KO) for HIF1β. ( f-g ) Heatmaps of HIF1β binding for WT MCF7 cells and of DNA methylation for WT and HIF1B -KO MCF7 cells, at regions surrounding the HIF1β ChIP-seq peak summit in WT MCF7 cells (±5 kb) ( f ), and violin and boxplots of methylation at the HIF1β ChIP-seq peak summit (±100 bps) for WT and HIF1B -KO MCF7 cells ( g ). Depicted are the data from the 4,794 regions having >30-fold methylation coverage for both genotypes (WT and KO) in the 100 bps flanking the binding summit, as assessed using SeqCapEpi BS-seq. HIF1β binding was assessed after 16 hours of 0.5% O 2 (hypoxia), and DNA methylation under 21% O 2 (normoxia). ( h ) Immunoblot analysis of HIF1β in mESCs that are wild-type (WT) or knockout (KO) for Hif1b . ( i, j ) Heatmaps of HIF1β binding for ESCs and of DNA methylation for murine WT and Hif1b -KO ESCs, at regions surrounding the HIF1β ChIP-seq peak summit in WT ESCs (±5 kb) ( i ), and violin and boxplots of methylation at the HIF1β ChIP-seq peak summit (±100 bps) for WT and Hif1b -KO ESCs ( j ). Depicted are the data from the 1,644 regions having >30-fold methylation-coverage for both genotypes (WT and KO) in the 100 bps flanking the binding summit, as assessed using whole-genome BS-seq. HIF1β binding was assessed after 16 hours of 0.5% O 2 (hypoxia), and DNA methylation under 21% O 2 (normoxia). ( k ) Ontology analysis of genes associated with HIF1 β peaks in 3 cell lines (SK-MEL-28, green; RCC4, yellow; MCF7, purple). ( l ) Cumulative frequency of distance to the nearest RCGTG motif, for HIF1 β binding peak summits detected in 3 cell lines and for a random set of genomic positions. ( m-n ) Sequencing read depth of HIF1β ChIP and its input, at all RCGTG sequences in RCC4 ( m ) and SK-MEL-28 ( n ) cells, stratified for methylation at the CG in the core R CG TG sequence. Shown are data for all RCGTG sequences in the human genome for which >10 × coverage was obtained after SeqCapEpi BS-seq.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Western Blot, ChIP-sequencing, Knock-Out, Binding Assay, DNA Methylation Assay, Methylation, Sequencing

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Heatmaps of HIF1β binding and DNA methylation for 7,153 regions (identified using a stringent threshold of P<10 -15 in MACS) surrounding the HIF1β ChIP-seq peak summit (± 5 kb). Heatmaps depict reads per kb per million reads (RPKM) of HIF1β ChIP-seq and of 5mC DNA IP-seq (mDIP), and % DNA methylation as estimated by SeqCapEpi BS-seq or whole-genome BS-seq (respectively, SeqCapEpi and WGBS). HIF1β binding was assessed after 16 hours of 0.5% O2 (hypoxia) and DNA methylation under 21% O2 (normoxia). ( b ) Violin plots of the methylation level inside and outside HIF1β binding peaks, as estimated by SeqCapEpi BS-seq. ( c ) Sequencing read depth of HIF1β ChIP and its input, at all RCGTG sequences in MCF7 cells, stratified for methylation at the CG in the core R CG TG sequence. Shown are boxplots for all RCGTG’s in the human genome for which >10 × coverage was obtained after SeqCapEpi BS-seq, with dark red dots denoting averages. See for additional QC of ChIP-seq data and Supplementary Table 1 for more details about HIF1β binding peak locations. ( d ) Venn diagram of 20,613 shared and unique HIF1β binding sites detected across 3 cell lines. Only stringent binding sites (P<10 -15 ) are shown. Binding sites showing intermediate levels of HIF1β ChIP-seq enrichment in 1 or 2 cell lines are unclassified and not shown here (n=445, 2,812 and 887 peaks, detected in SK-MEL-28, RCC4 and MCF7 respectively). ( e ) Heatmaps of HIF1β binding ( red ) and DNA methylation as estimated using SeqCapEpi BS-seq ( blue ) at regions flanking the HIF1β ChIP-seq peak summit (± 5 kb). ( top ) HIF1β binding peaks shared between the 3 cell lines. ( bottom ) HIF1β binding peaks unique to each cell line. Heatmaps depict RPKM of HIF1β ChIP-seq and % DNA methylation. HIF1β binding was assessed after 16 hours of 0.5% O2 (hypoxia) and DNA methylation under 21% O2 (normoxia). ( f ) Quantification of the methylation level at HIF1β binding peak summits ±100 bps, for peaks that are shared between or unique to one of the 3 cell lines. ( g ) Enrichment of gene expression (observed/expected) upon hypoxia per cell line, for genes associated with HIF1β binding sites (within 50 kb) that are shared between or unique to one of the 3 cell lines, as labelled on the X-axis. ( h ) Fraction of HIF1β peaks overlapping with the binding peaks of individual transcription factors , or with any of the 11 transcription factors profiled in MCF7 cells (“ combined ”). ( i-j ) ( i ) Overlap between HIF1β binding peaks and other transcription factor binding sites detected in MCF7 cells. Shown are fractions of HIF1β binding peaks shared between (grey) or unique for a cell line (coloured). ( j ) mRNA expression level of transcription factors in each cell line, as determined using RNA-seq. Transcription factors expressed in all 3 cell lines are highlighted as “ shared TFs” with a light grey box.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Binding Assay, DNA Methylation Assay, ChIP-sequencing, Methylation, Sequencing, Expressing, RNA Sequencing Assay

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Patterns of DNA methylation and of the indicated chromatin modifications in MCF7 cells, at HIF1β binding sites detected in MCF7 cells at CpG islands ( top ), not at CpG islands ( middle ) or in SK-MEL-28 and/or RCC4 cells but not MCF7 ( bottom ). DNA methylation was determined using SeqCapEpi BS-seq or WGBS; occupancy of chromatin modifications was determined using the ENCODE project. ( b ) Estimates of the contribution of individual epigenetic marks in predicting the presence of HIF1β binding peaks, as determined by partial R 2 analysis of linear regression models. Shown are the results of a linear model using all available marks, a linear model using all available marks but not DNA methylation, and the results of linear models using each modification in an individual analysis. All 3 available FAIRE-seq datasets , show similar results. ( c ) Functional genome annotation of MCF7 cells using ChromHMM, at shared and cell-type-specific HIF1β binding peaks. ( d ) Fraction of regions in MCF7 cells showing high versus low methylation in regions of open chromatin. 10,805 regions of open chromatin that showed high GpC methylation levels were assessed (>10-fold coverage) for their CpG methylation. Indicated are the fraction of these 10,805 regions showing at least 25%, 50% or 80% CpG methylation. GCG trinucleotides were not considered in these analyses. ( e ) Fraction of HIF1β binding peaks, shared between or unique for a cell line, that overlaps with other TF binding sites detected in MCF7 cells. Note how HIF1β binding peaks unique to MCF7 cells ( purple ) predominantly overlap with TFs uniquely expressed in MCF7 cells (TFASP2C, ESR1, …), while HIF1β binding peaks shared between cell lines ( dark grey ) more frequently overlap with commonly expressed TFs (NR2F2, STAG1, …). ( f ) mRNA expression level of indicated transcription factors in each cell line, as determined using RNA-seq. These data enable stratification of TFs into those that are shared in expression ( light grey box ) and those that are uniquely expressed ( white box ). ( g ) Fraction of HIF1β binding peaks, shared between or unique for a cell line, that overlaps with other transcription factor binding sites detected in A549 cells. Note how HIF1β binding peaks unique to A549 cells ( blue ) predominantly overlap with TFs expressed in A549 cells (GATA3 and FOXA1), while HIF1β binding peaks shared between cell lines ( dark grey ) more frequently overlap with binding of CTCF, a commonly expressed TF. ( h ) Comparison of the HIF1β peaks detected to a previously published dataset of ChIP-seq for HIF1β, HIF1α and HIF2α. ( i ) ChIP-seq signal for the indicated epitopes at the 200 bps surrounding the HIF1β peak summits that are uniquely co-bound by HIF1α or HIF2α. ( j ) Functional genome annotation of MCF7 cells using ChromHMM, at HIF1β binding peaks that are uniquely co-bound by HIF1α or HIF2α. ( k ) Fraction of HIF1β binding peaks, uniquely co-bound by HIF1α or HIF2α, that overlaps with other transcription factor binding sites detected in MCF7 cells.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: DNA Methylation Assay, Binding Assay, Modification, Functional Assay, Methylation, CpG Methylation Assay, Expressing, RNA Sequencing Assay, ChIP-sequencing

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a-b ) Boxplot ( left ) and scatter plot ( right ) of methylation levels of HIF1β-bound immunoprecipitated DNA fragments obtained by ChIP-BS-seq (ChIP-BS) compared to input by SeqCapEpi BS-seq (SeqCapEpi) in MCF7 cells (a) , or of HIF1β-bound immunoprecipitated DNA fragments obtained by ChIP-BS compared to input by whole-genome BS-seq (WGBS) in mouse Tet -triple-knockout ( Tet- TKO) ESCs (b) . The red dotted line in the scatter plot indicates the theoretical value of equal methylation in immunoprecipitated and input DNA. P values by t-test. ( c-d ) Recombination-mediated cassette exchange. (c) A human HIF binding site (chr16: 30,065,212-30,065,711 on hg38) was cloned between 2 L1 Lox sites and in vitro methylated ( blue ) or not ( red ) prior to co-transfection with a CRE recombinase-encoding plasmid into mESCs transformed to contain an L1 Lox -flanked thymidine kinase (TK). (d) Following successful cassette exchange, these ESCs were incubated in hypoxia (0.5% O2 for 16 hours) and probed using HIF1β ChIP-qPCR for HIF binding at the differentially methylated cassette. Shown is the fold enrichment over background (n=3 independent ChIP pairs; *P<0.05 by t-test). ( e-f ) Microscale thermophoresis-based assessment of sensitivity of HIF1α-HIF1β (e) and HIF2α -HIF1β (f) heteroduplexes to methylation at HIF binding sites in physiological buffer (PBS). RCGTG sequences in the double-stranded DNA oligonucleotides were either absent ( grey ), methylated ( blue ) or unmethylated ( red ) at the CpG site. Calculated KD values are shown under each graph. ( g ) Excerpt from the crystal structure of HIF2α-HIF1β in complex with a DNA duplex containing the core HIF binding sequence 5’-ACGTG-3’ (PDB code 4ZPK) . ( h ) Modelling of methylation of CpG cytosines in A CG TG reveals severe steric hindrance. The two views show hard-sphere models of methylated cytosines modelled at position 5 (including bonding hydrogen atoms) and how they severely violate the van der Waals envelopes (2.5Å width) of Arg27 in HIF2α ( left ) and Arg102 in HIF1β ( right ).

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Methylation, Immunoprecipitation, Triple Knockout, Binding Assay, Clone Assay, In Vitro, Cotransfection, Plasmid Preparation, Transformation Assay, Incubation, Microscale Thermophoresis, Sequencing

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Example plots of regions on chromosomes 6 and 16, showing HIF1β ChIP-seq read depths, and methylation levels as determined by whole-genome BS-seq and HIF1β ChIP-BS-seq in MCF7 cells. Coordinates: human genome build hg19. Grey boxes in HIF1β ChIP-BS-seq tracks indicate regions where insufficient sequences were recovered. ( b ) As in ( a ), but in murine Tet -TKO ESCs, and for regions on chromosomes 5 and 3. Coordinates follow genome build Mm10. In each example plot, HIF1β peaks (fragment per million) are shown on top; percentage of DNA methylation (calculated in bulk normoxic cells by WGBS) at the same position of the HIF1β peaks is shown in the middle; percentage of DNA methylation of the immunoprecipitated DNA fragment from HIF1β ChIP-BS-seq is shown at the bottom. ( c-d ) Methylation levels of a human DNA fragment inserted in mESCs, as detected using amplicon bisulfite sequencing. A human HIF binding site (chr16: 30,065,212-30,065,711 on hg38) was cloned between 2 L1 Lox sites and in vitro methylated ( d ) or not ( c ) prior to insertion into mESCs transformed to contain an L1 Lox -flanked thymidine kinase.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: ChIP-sequencing, Methylation, DNA Methylation Assay, Immunoprecipitation, Amplification, Methylation Sequencing, Binding Assay, Clone Assay, In Vitro, Transformation Assay

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Heatmaps of HIF1β binding (RPKM) and DNA methylation as determined using WGBS at regions flanking the summit of HIF1β binding peaks (± 5 kb) either shared with WT or TKO-specific ESCs. ( b ) % methylation at shared and TKO-specific HIF1β binding sites in WT ESCs. See for scatter plots illustrating the correlations between HIF1β ChIP-seq replicates in Dnmt- WT versus Dnmt -TKO ESCs. ( c ) Cumulative frequency of distance to the nearest RCGTG motif for shared, TKO-specific and randomized HIF1β binding peaks. ( d ) Observed/expected frequency of upregulated genes associated with shared and TKO-specific HIF1β binding peaks in WT and Dnmt -TKO ESCs exposed to 24 hours of hypoxia (0.5% O2). ( e ) Distance of shared and TKO-specific HIF1β binding peaks in ESCs to the nearest TSS. A bimodal peak was detected indicating proximal and distal binding events. ( f ) Functional genome annotation using ChromHMM of shared and TKO-specific HIF1β binding peaks in ESCs. ( g ) Distance of shared and TKO-specific HIF1β binding peaks to open chromatin regions in ESCs. A bimodal peak was detected indicating proximal and distal binding events. ( h ) Ontology analysis of genes associated with shared and TKO-specific HIF1β binding peaks in ESCs. ( i ) HIF1β binding sites in LINEs, LTRs and SINEs after 10,000 random permutations and as observed by HIF1β ChIP-seq (actual HIF1β binding) for all HIF1β sites ( top panel ) and only for distal HIF1β sites ( bottom panel ). *** P<0.001 by Fisher’s exact test. ( j ) Distribution of HIF1β binding peaks detected in murine Dnmt -TKO ESCs for the retrotransposon families, colour-coded by retrotransposon class (green: LTR; violet: LINE; yellow: SINE).

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Binding Assay, DNA Methylation Assay, Methylation, ChIP-sequencing, Functional Assay

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Heatmaps of HIF1β binding (RPKM) and DNA methylation as determined using SeqCapEpi BS-seq at regions flanking the summit of HIF1β binding peaks (± 5 kb). Shown are HIF1β binding peaks that are shared between vehicle-treated and aza-treated MCF7 cells, or that are specific to aza-treated cells. 12,782 HIF1β binding peak positions were detected across vehicle- and aza-treated MCF7 using a P<10 -15 threshold. ( b ) Violin plots of methylation detected by SeqCapEpi BS-seq at HIF1β binding peaks that are shared between vehicle-treated and aza-treated MCF7 cells, or that are specific to aza-treated MCF7 cells. ( c ) HIF1β binding sites in LINEs, LTRs and SINEs after 10,000 random permutations and as observed by HIF1β ChIP-seq (actual HIF1β binding) for HIF1β binding peaks that are shared between vehicle- and aza-treated MCF7 cells, or specific to aza-treated MCF7 cells for all HIF1β sites ( top panel ) and only for distal HIF1β sites ( bottom panel ). ( d ) Distribution of HIF1β binding peaks detected in aza-treated MCF7 cells at retrotransposon families, colour-coded by retrotransposon class (green: LTR; violet: LINE; yellow: SINE). ( e ) Volcano plots showing differential expression of HIF-bound cryptic transcripts, as determined by CREDENToR in MCF7 cells exposed to vehicle (DMSO) or 5-aza-2’-deoxycytidine (aza; 1 μM) for 4 days, hypoxia (0.5% oxygen, 1 day) or normoxia. Significantly upregulated and downregulated transcripts are highlighted in red and blue, respectively. The associated numbers refer to how many transcripts are up- or downregulated at a 1% FDR and a 0.001% FDR, as indicated by the horizontal line. ( f ) dsRNA formation potential of all cryptic transcripts ( grey ) and of HIF-bound cryptic transcripts ( red ). Shown are the fraction of all RNAs for which transcription overlaps with a transcript expressed from the complementary strand (“sense-antisense”), and RNAs containing the same retrotransposon repeat element in sense and antisense orientation (“palindromic”). P values from chi-square test. ( g-h ) Expression of HIF-bound ( g ) and non-HIF-bound ( h ) cryptic transcripts relative to vehicle-treated controls (vehicle normoxia) in MCF7 cells wild-type (WT) ( g upper panels and h ) or knockout (KO) ( g bottom panels ) for HIF1B . Shown are expression changes as assessed using CREDENToR (g left panels and h) and RepEnrich (g right panels) , with error bars indicating geometric mean ± s.e.m. n.s. not significant, ***P<0.001 by t-test.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Binding Assay, DNA Methylation Assay, Methylation, ChIP-sequencing, Expressing, Knock-Out

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a ) Recurrent detection of cryptic transcripts in different cancer types profiled in TCGA. Shown are the number of cryptic transcripts detected in 1 cancer type or in different cancer types, as per the legend. ( b ) Volcano plot showing the effect of the interaction between cryptic transcript DNA methylation and tumour oxygenation on the expression of individual cryptic transcripts, tested by DESeq. Positive coefficients represent the cooperative enhancement of cryptic transcript expression in low-methylation, hypoxic conditions. Highlighted in red and blue are significantly positive and negative coefficients (5% FDR) respectively. ( c ) Immunogenicity estimates for TCGA tumours responsive or non-responsive to immunotherapy as described by Turajlik and colleagues ( red and white ). Shown are number of somatic mutations extracted from the TCGA database, mRNA expression of PDL1 , PD1 and LAG3 as log 2 RPKM, cytolytic activity (CYT) defined as the log 2 -average (geometric mean) of GZMA and PRF1 expression in RPKM, cell fraction for CD8 + T-cells estimated with respect to the total cells in the sample as defined by quanTIseq (The Cancer Immunome Atlas, TCIA ), and cryptic transcript expression as log 10 counts per million. **P<0.01 and ***P<0.001 by t-test. ( d ) The number of cryptic transcripts expressed in cancer cells, T-cells and other stromal cells isolated from non-small cell lung tumours (left, n=5) or from breast tumours (right, n=22) and profiled using single-cell RNA-sequencing. Note that this expression dataset was generated using unique molecular identifiers at the 3’-end or 5’-end of poly-adenylated transcripts (resp. left and right panel), enabling the counting of transcripts per cell.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: DNA Methylation Assay, Expressing, Methylation, Activity Assay, Isolation, RNA Sequencing Assay, Generated

Journal: bioRxiv

Article Title: DNA M ethylation R epels B inding of HIF T ranscription F actors to M aintain T umour I mmunotolerance

doi: 10.1101/2020.02.07.931071

Figure Lengend Snippet: ( a-c ) Cryptic transcript expression in tumours characterized by TCGA. Shown is cryptic transcript expression in tumours with high or low methylation of cryptic transcript promoter regions (blue or red; > or ≤ the median methylation level of each tumour type), and in normoxic or hypoxic (light or dark colour) tumours. Data are shown for ( a ) all tumour types combined, ( b ) stratified into those that are responding or non-responding to immunotherapy following the classification described by Turajlik and colleagues , and ( c ) for each tumour type separately. P values by t-test, red values indicating inverse correlations. ( d ) DNA methylation levels at cryptic transcript promoters ( left ) and cryptic transcript expression ( right ) in tumours profiled in TCGA, stratified into tumour types that are responsive (n=2,280) or non-responsive (n=2,214) to checkpoint immunotherapy. ***P<0.001 by t-test. ( e ) Heatmap showing the expression (Z score, blue to red) of the 59 cryptic transcripts associated with cytolytic activity in tumours responsive to immunotherapy from TCGA. The boxplot on the right depicts the log fold change in expression of the same 59 cryptic transcripts in hypoxic versus normoxic MCF7 cells (24 hours, 0.5% O2), and of MCF7 cells after 4-day exposure to aza versus vehicle-treated hypoxic MCF7 cells (P<0.05 for all cryptic transcripts, either for hypoxia versus vehicle, or for hypoxia plus aza versus aza alone). At the bottom, cytolytic activity of each TCGA sample is depicted. LUAD; lung adenocarcinoma; LUSC, lung squamous cell carcinoma; HNSC, head and neck squamous cell carcinoma; BLCA, bladder urothelial carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; SKCM, skin cutaneous melanoma.

Article Snippet: ChIP-BS-seq was done as ChIP-seq, except that methylated adaptors (NEB) were ligated, and DNA libraries were bisulfite-converted using the EZ DNA Methylation-Lightning™ kit (Zymo) prior to library amplification using HiFi Uracil+ (KAPA).

Techniques: Expressing, Methylation, DNA Methylation Assay, Activity Assay

Journal: Genome Biology

Article Title: Age and sun exposure-related widespread genomic blocks of hypomethylation in nonmalignant skin

doi: 10.1186/s13059-015-0644-y

Figure Lengend Snippet: DNA methylation in epidermal but not dermal samples clusters by age and sun exposure. (A) In epidermis, DNA methylation segregates old versus young individuals, and also segregates sun-exposed and sun-protected anatomical regions, shown by multidimensional scaling of pairwise distances derived from methylation levels assayed on the HumanMethylation450 BeadChip (450k). (B) In dermis, DNA methylation does not segregate samples by age or anatomical region, shown by multidimensional scaling of pairwise distances derived from methylation levels assayed on the 450k.

Article Snippet: After purification, samples were bisulfite converted and purified using Zymo EZ DNA Methylation Gold.

Techniques: DNA Methylation Assay, Derivative Assay, Methylation