ArticlePDF Available

An epigenomic roadmap to induced pluripotency reveals DNA methylation as a reprogramming modulator

December 2014
Nature Communications 5(1):5619

December 2014
5(1):5619

Source
PubMed

License
CC BY 4.0

Authors:

Reprogramming of somatic cells to induced pluripotent stem cells involves a dynamic rearrangement of the epigenetic landscape. To characterize this epigenomic roadmap, we have performed MethylC-seq, ChIP-seq (H3K4/K27/K36me3) and RNA-Seq on samples taken at several time points during murine secondary reprogramming as part of Project Grandiose. We find that DNA methylation gain during reprogramming occurs gradually, while loss is achieved only at the ESC-like state. Binding sites of activated factors exhibit focal demethylation during reprogramming, while ESC-like pluripotent cells are distinguished by extension of demethylation to the wider neighbourhood. We observed that genes with CpG-rich promoters demonstrate stable low methylation and strong engagement of histone marks, whereas genes with CpG-poor promoters are safeguarded by methylation. Such DNA methylation-driven control is the key to the regulation of ESC-pluripotency genes, including Dppa4, Dppa5a and Esrrb. These results reveal the crucial role that DNA methylation plays as an epigenetic switch driving somatic cells to pluripotency.

Experimental and computational analysis overview of the study. (a) Establishment of secondary system and sample collection. (b) MethylC-Seq was performed on samples from secondary system. DMRs were identified. RNA-Seq and ChIP-Seq data were integrated with MethylC-Seq data based on transcripts. (c) Base-level visualization of DNA methylation and histone distribution around Dppa2.

…

DMRs and features affecting DNA methylation change during reprogramming. (a) Hierarchical clustering based on the DNA methylation level of DMRs in each sample. Each DMR was centred with the mean and normalized. DMRs were clustered into six groups based on pairwise correlations. (b) Base-level visualization of two DMRs from group DMR-3b in the promoter regions of Dppa4 and Dppa2, known ESC-pluripotency predictor genes. (c) DMR accumulation during reprogramming. DMRs were defined as hyper- and hypo-DMRs at each time point. Dark red and dark blue bars represent ESC-specific Hyper- and Hypo-DMRs. Other colours indicate Hyper- and Hypo-DMRs in the order of left to right. (d) Proportion of DMRs containing various genomic features. (e) Fold enrichment of examined genomic features within DMRs. (f) Percentage of DMRs containing H3K4me3 or H3K27me3 based on the methylation level (low-methylated ≤30%, high-methylated ≥70%).

…

Histone modification and DNA methylation change at transcription factor-binding sites. RNA expression level (FPKM) of transcription factors (line plots), average DNA methylation change (upper bar plots), average H3K4me3 change (blue bar plots) and average H3K27me3 change (red bar plots) at binding sites of each transcription factor. Selected transcriptionally active genes during high-dox treatment (blue box), transcriptionally silent genes during high-dox treatment (green box) and polycomb repressive complexes (red box) are shown.

…

Histone modification and DNA methylation change around transcription factor-binding sites. Average DNA methylation change (left), average H3K4me3 change (middle) and average H3K27me3 change (right) in the 80-kb neighbourhood of transcription factor-binding sites.

…

Epigenetic features of gene classes and model of gene expression control. (a) Genes were separated into clusters based on gene expression patterns and DNA methylation. The heatmap presents mRNA expression, DNA methylation level of promoter regions, normalized H3K4me3 level, normalized H3K27me3 level, CpG densities, pluripotency transcription factor-binding sites and binding sites of PRCs. (b) Base-level visualization of DNA methylation and histone modifications in the promoter regions of representative genes for each class across all samples. (c) Percentage of ESC-specific H3K4me3 mark for promoters with high and low initial methylation. (d) Percentage of ESC-specific H3K27me3 mark for promoters with high and low initial methylation.

…

Figures - uploaded by Thomas Preiss

Content may be subject to copyright.

Content uploaded by Thomas Preiss

Content may be subject to copyright.

Available via license: CC BY 4.0

Content may be subject to copyright.

ARTICLE

Received 30 Aug 2014 |Accepted 21 Oct 2014 |Published 10 Dec 2014

An epigenomic roadmap to induced pluripotency

reveals DNA methylation as a reprogramming

modulator

Dong-Sung Lee1,2,3,*, Jong-Yeon Shin1,4,*, Peter D. Tonge5, Mira C. Puri5,6, Seungbok Lee1,2,3, Hansoo Park1,2,3,

Won-Chul Lee1,4, Samer M.I. Hussein5, Thomas Bleazard7, Ji-Young Yun1,4, Jihye Kim1,4, Mira Li5,

Nicole Cloonan8,9, David Wood8, Jennifer L. Clancy10, Rowland Mosbergen11, Jae-Hyuk Yi1, Kap-Seok Yang4,

Hyungtae Kim4, Hwanseok Rhee12, Christine A. Wells11,13, Thomas Preiss10,14, Sean M. Grimmond8,15,

Ian M. Rogers5,16,17, Andras Nagy5,17,18 & Jeong-Sun Seo1,2,3,4

Reprogramming of somatic cells to induced pluripotent stem cells involves a dynamic

rearrangement of the epigenetic landscape. To characterize this epigenomic roadmap, we

have performed MethylC-seq, ChIP-seq (H3K4/K27/K36me3) and RNA-Seq on samples

taken at several time points during murine secondary reprogramming as part of Project

Grandiose. We ﬁnd that DNA methylation gain during reprogramming occurs gradually, while

loss is achieved only at the ESC-like state. Binding sites of activated factors exhibit focal

demethylation during reprogramming, while ESC-like pluripotent cells are distinguished

by extension of demethylation to the wider neighbourhood. We observed that genes with

CpG-rich promoters demonstrate stable low methylation and strong engagement of histone

marks, whereas genes with CpG-poor promoters are safeguarded by methylation. Such DNA

methylation-driven control is the key to the regulation of ESC-pluripotency genes, including

Dppa4, Dppa5a and Esrrb. These results reveal the crucial role that DNA methylation plays as

an epigenetic switch driving somatic cells to pluripotency.

DOI: 10.1038/ncomms6619 OPEN

1Genomic Medicine Institute (GMI), Medical Research Center, Seoul National University, Seoul 110-799, Korea. 2Department of Biomedical Sciences, Seoul

National University College of Medicine, Seoul 110-799, Korea. 3Department of Biochemistry, Seoul National University College of Medicine, Seoul 110-799,

Korea. 4Life Science Institute, Macrogen Inc., Seoul 153-781, Korea. 5Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, Ontario,

Canada M5G 1X5. 6Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada M5T 3H7. 7Faculty of Medical and Human Sciences,

University of Manchester, Manchester M13 9PT, UK. 8Queensland Centre for Medical Genomics, Institute for Molecular Bioscience, The University of

Queensland, St Lucia, Queensland 4072, Australia. 9QIMR Berghofer Medical Research Institute, Genomic Biology Lab, 300 Herston Road, Herston,

Queensland 4006, Australia. 10 Genome Biology Department, The John Curtin School of Medical Research, The Australian National University, Canberra,

Australian Capital Territory 2601, Australia. 11 Australian Institute for Bioengineering and Nanotechnology, The University of Queensland, Brisbane,

Queensland 4072, Australia. 12 Macrogen Bioinformatics Center, Macrogen, Seoul 153-781, Republic of Korea. 13 College of Medical, Veterinary and Life

Sciences, University of Glasgow, Glasgow, Scotland G12 8TA, UK. 14 Molecular, Structural & Computational Biology Division, Victor Chang Cardiac Research

Institute, Sydney, New South Wales 2010, Australia. 15Wolfson Wohl Cancer Research Centre, Institute for Cancer Sciences, University of Glasgow, Bearsden,

Glasgow Scotland G61 1BD, UK. 16 Department of Physiology, University of Toronto, Toronto, Ontario, Canada M5T 3H7. 17 Department of Obstetrics and

Gynaecology, University of Toronto, Toronto, Ontario, Canada M5T 3H7. 18 Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada M5T

3H7. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to J.-S.S. (email: jeongsun@snu.ac.kr).

NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications 1

Somatic cells can be reprogrammed into induced pluripotent

stem cells (iPSCs) by the expression of deﬁned transcription

factors1–5. During the reprogramming process, the global

epigenetic landscape has to be reset to establish the epigenetic

marks of the pluripotent state through DNA methylation and

chromatin-remodelling processes2,6–9. Through the development

of a secondary reprogramming system10, iPSC generation was

initially described as a multistep process characterized by

transcriptional, DNA methylation and chromatin changes11–14.

Genome-wide analysis of speciﬁc chromatin modiﬁcation

dynamics at early stages of reprogramming indicated that this

progress might be constrained by repressive epigenetic modi-

ﬁcations, such as H3K9me3 and DNA methylation15–18.

More recently, it has been proposed that DNA methylation

during iPSC generation functions in the silencing of genes

involved in differentiation, while also facilitating chromatin

remodelling18–20. DNA demethylation appears to play an

important role in reactivating pluripotency genes, which are

hypermethylated and silenced in somatic cells, particularly in the

late stages of the reprogramming process13. However, overall

understanding of the global dynamics of epigenetic modiﬁcation

at different stages during reprogramming remains poor.

In this work, we have utilized a murine secondary

reprogramming system to sample cellular trajectories during

reprogramming and performed whole-genome bisulﬁte sequen-

cing, chromatin immunoprecipitation sequencing (ChIP-seq;

H3K4me3, H3K27me3 and H3K36me3), and RNA sequencing

(RNA-Seq) to characterize the epigenomic roadmap to pluripo-

tency at base resolution21,22. Our observations provide a deeper

understanding of the reprogramming process and reveal the

crucial role that DNA methylation plays in the epigenetic switch

that drives somatic cells to pluripotency.

Results and Discussion

Dynamic changes in DNA methylation during reprogramming.

The Project Grandiose secondary reprogramming samples

present a unique opportunity to proﬁle cellular state changes

at various time points during reprogramming10,21,22. These

consisted of secondary mouse embryonic ﬁbroblasts (2°MEF),

six intermediate time points at high doxycycline (dox)

concentrations (D2H, D5H, D8H, D11H, D16H and D18H),

three alternative intermediate time points collected for samples

treated with reduced dox concentrations (D16L, D21L

and D21Ø), the secondary iPSCs (2°iPSCs), the primary iPSCs

(1°iPSCs) used to generate the chimeric mouse and a mouse

Rosa rtTA embryonic stem cell line (ESC) for standard

comparison (Fig. 1a–c). As described in ref. 21, these samples

showed reprogramming to two distinct pluripotent states:

ESC-like cells and the ‘F-class’ consisting of stages D16H

and D18H.

In this manuscript, we describe base-resolution bisulﬁte

sequencing of the 13 Project Grandiose samples and investigation

of global DNA methylation changes during reprogramming

(Supplementary Data 1). The sample methylomes were scanned

using a sliding window of 30 CpGs, identifying 7,890 differen-

tially methylated regions (DMRs) covering 22 Mb, representing

0.81% of the mouse genome (Fig. 2a,b, Supplementary Data 2,

Epigenomic regulation of reprogramming

Chip-Seq

RNA-Seq

Integrative

analysis

H3K4me3

H3K27me3

H3K36me3

Differentially methylated regions

DNA methylation

change around TFBS

• Pattern analysis

• Feature enrichment

test in each pattern

• DNA methylation

accumulation

in gain and loss

Whole-genome bisulfite sequencing (MethylC-Seq)

1°iPSC

1°MEF

Tetraploid embryo

complementation

Primary

reprogramming

2°MEFs

Secondary

reprogramming

2°MEFs 2°iPSC

1°iPSC

ESC

D2H D5H D8H D11H D16H D18H

D16L

D21L

D21Ø

1,500

Dox

concn (ng)

Chr16qB5

2°MEF D18H 2°iPSCs ESC

Dppa2 Dppa2 Dppa2 Dppa2

Sample

RefSeq

DNA

methylation

H3K4me3

H3K27me3

H3K36me3

qA1 qB1qA2 qB2 qB3 qB4 qB5 qC1.1 qC1.3 qC3.1 qC3.3 qC4qC2

Figure 1 | Experimental and computational analysis overview of the study. (a) Establishment of secondary system and sample collection. (b) MethylC-

Seq was performed on samples from secondary system. DMRs were identiﬁed. RNA-Seq and ChIP-Seq data were integrated with MethylC-Seq data

based on transcripts. (c) Base-level visualization of DNA methylation and histone distribution around Dppa2.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619

2NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications

Supplementary Figs 1a–c, 6a). Unsupervised hierarchical cluster-

ing performed on the DNA methylation state of DMRs (Fig. 2a)

distinguished the intermediate states (D2H-D18H and D16L-

D21L) from the ESC-like pluripotent states (D21Ø, 1oiPSCs,

2oiPSCs and ESCs). DMRs were categorized into three groups

based on the changing pattern of DNA methylation (Fig. 2a).

The DMR-1 group exhibited increased methylation levels after

(DMR-1a) or during (DMR-1b) high-level reprogramming factor

expression and included genes related to development and cell

differentiation, such as the Hox family, Col25a1 and Meox2. The

DMR-2 group represented differential methylation changes

between two pluripotent states: either gradual demethylation to

F-class and methylation in the ESC-like state (DMR-2a) or

gradual methylation to F-class and acquired demethylation in the

ESC-like state (DMR-2b). A ﬁnal group (DMR-3) was identiﬁed

as exhibiting low methylation levels in the ESC-like state

Number of hyper-DMRs

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

3,000

1,500

Number of hypo-DMRs

3,000

Fold enrichment (x)

Low-methylated DMRs

(≤30%)

High-methylated DMRs

(≥70%)

K27me3

only 2.8%

No K4me3

or K27me3

3.1% K4me3

only 49.6%

Both

K4/K27me3

41.0%

K27me3 only

6.3%

No K4me3

or K27me3

79.7%

Both

K4/K27me3

0.9%

K4me3

only 16.6%

00.5 1

DNA methylation

level

DMR-1a

(n=1,819)

DMR-1b

(n=1,453)

DMR-2a

(n=553)

DMR-2b

(n=1,291)

DMR-3

(n=2,774)

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

RefSeq

DNAme

variation

DMR

TFBS

H3K4me3

H3K36me3

H3K27me3

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

51 kb

Dppa4 Dppa2

Nanog Esrrb Nanog

DMRs (%)

DMR_6651 DMR_6652

K4me3

TFBS

SINE

K27me3

Simple_repeat

CpG_shore

LTR

LINE

Intron

CpGi

exon

Promoter

K36me3

Enhancer

Meox1

Angpt2

Dmrt2

Erg

Hoxa1

Hoxd3

Mafb

Ptx3

Snai2

Abi3bp

Rapgef4

Sulf2

Clec14a

Col24a1

Col25a1

Il1rl2

Meox2

Nanog

Elf3

Gap43

Gfra1

Gli2

Icam1

Naprt1

Lhx6

100

Genes

within

DMR

Dnmt3l

Dppa4

Dppa5a

Esrrb

Gm11607

Itgb7

Rec8

Sox15

Sycp3

Trap1a

Triml2

Zfp42

Figure 2 | DMRs and features affecting DNA methylation change during reprogramming. (a) Hierarchical clustering based on the DNA methylation level

of DMRs in each sample. Each DMR was centred with the mean and normalized. DMRs were clustered into six groups based on pairwise correlations.

(b) Base-level visualization of two DMRs from group DMR-3b in the promoter regions of Dppa4 and Dppa2, known ESC-pluripotency predictor genes.

(c) DMR accumulation during reprogramming. DMRs were deﬁned as hyper- and hypo-DMRs at each time point. Dark red and dark blue bars represent

ESC-speciﬁc Hyper- and Hypo-DMRs. Other colours indicate Hyper- and Hypo-DMRs in the order of left to right. (d) Proportion of DMRs containing

various genomic features. (e) Fold enrichment of examined genomic features within DMRs. (f) Percentage of DMRs containing H3K4me3 or H3K27me3

based on the methylation level (low-methylated r30%, high-methylated Z70%).

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619 ARTICLE

NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications 3

(1°iPSCs, 2°iPSCs and ESCs), with stable methylation persisting

in the F-class state and intermediate reprogramming samples,

which included multiple pluripotency genes such as Dppa2,

Dppa4, Dppa5a,Esrrb,Tcl1 and Eras (Fig. 2a, Supplementary

Data 2).

We annotated the DMRs in each sample as Hyper- or Hypo-

DMRs where they differed from a corresponding 2°MEF baseline

by over 20% (Fig. 2c). We observed a widespread gradual increase

in methylation to generate Hyper-DMRs during reprogramming,

whereas limited demethylation was observed as cells repro-

grammed to the F-class state (D16H and D18H). The steady

increase in Hyper-DMRs during both high-dox and low-dox

reprogramming challenges the notion that most changes in DNA

methylation occur at a late stage when cells acquire stable

pluripotency13. A similar trend was observed for the average

methylation level of DMRs, as methylation occurred gradually,

while demethylation did not change signiﬁcantly during

transgene expression (Supplementary Fig. 2a,b). Almost all

Hypo-DMRs found in iPSCs were also observed in ESCs

(98.94%); however, this was not the case for Hyper-DMRs

(61.88%), suggesting that demethylation during reprogramming

occurred more conservatively.

Table 1 | Enrichment of TFBSs in each DMR group.

DMR groups DMR

Number

Sequence-speciﬁc transcription factors Transcription regulators

TET1 CTCF Oct4 SOX2 NANOG ESRRB ZFX KLF4 cMYC nMYC E2F1 TCFCP2L1 SMAD1 STAT3 p300 EZH2 SUZ12 RING1B

DMR-1

1a 1,819 NE NE   NE NE NE

1b 1,453 NE NE   þþþþþ

DMR-2

2a 553 NE NE NE NE NE NE þþ NE NE NE NE NE NE 

2b 1,291 þNE NE NE NE þþþNE NE þNE þNE þ þþþ þþþ þþþ

DMR-3 2,774 NE NE þþþ þþþ þþþ þþ þþ þþ þþþ þþþ þþþ þ þþþ þþþ þþþ   

TFBS enrichment of total

DMRs versus whole

genome

7,890 13.25 3.84 17.26 15.99 17.26 18.65 13.25 18.65 13.25 17.26 18.65 13.25 17.26 13.25 17.26 18.65 13.25 17.26

DMR, differentially methylated region; TFBS, transcription factor-binding sites.

Fold enrichment versus total DMRs: o0.75 rNE (not enriched) r1.25 rþo1.5 rþþo1.75rþþþ.

100

150

200

500

2,000

H3K27me3

difference

H3K4me3

difference

TF expression

(FPKM)

Average

TFBS

methylation

difference

OCT4

n= 3,761

SOX2

n= 4,526

KLF4

n= 10,875

NANOG

n= 10,343

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

TF expression

(FPKM)

TF expression

(FPKM)

ESRRB

n= 21,647

TCFCP2L1

n=26,910

SUZ12

n= 4,215

EZH2

n= 5,185

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1iPSC

2iPSC

ESC

2°MEF

D11

D16

D18

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

–0.1

–0.2

0.1

0.2

–0.1

Average

TFBS

methylation

difference

H3K4me3

difference

H3K27me3

difference

Average

TFBS

methylation

difference

H3K4me3

difference

H3K27me3

difference

0.1

0.2

–0.1

0.1

–0.1

0.1

Figure 3 | Histone modiﬁcation and DNA methylation change at transcription factor-binding sites. RNA expression level (FPKM) of transcription factors

(line plots), average DNA methylation change (upper bar plots), average H3K4me3 change (blue bar plots) and average H3K27me3 change (red bar

plots) at binding sites of each transcription factor. Selected transcriptionally active genes during high-dox treatment (blue box), transcriptionally silent

genes during high-dox treatment (green box) and polycomb repressive complexes (red box) are shown.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619

4NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications

TFBSs and histone modiﬁcation are enriched in the DMRs.To

assay the distribution of histone marks, we performed ChIP-Seq

for H3K4me3, H3K27me3 and H3K36me3 (see Methods). We

determined the distribution and enrichment of these histone

marks within DMRs, as well as other genomic features including

ESC-TFBSs from published data23–26 (Supplementary Data 3).

Notably, we found that 98% of DMRs contained H3K4me3

clusters and 68% contained ESC-TFBSs (Fig. 2d). When we

assessed enrichment of each feature relative to the whole genome,

H3K4me3 marks, ESC-TFBSs, CpG islands, CpG shores and

enhancers showed more than 10-fold enrichment, followed by

promoters and H3K27me3 clusters (Fig. 2e).

Our ﬁnding that histone marks were highly enriched within

DMRs led us to explore the relationship between DNA

methylation levels and H3K4me3/H3K27me3 marks within

DMRs (Fig. 2f, Supplementary Fig. 2c, Supplementary Table 1).

DMRs exhibiting low-level methylation (less than 30%) were

frequently associated (96.9%) with H3K4me3 and H3K27me3. In

contrast, the absence of both histone marks was most frequently

associated (79.7%) with DMRs with high levels of methylation

(Z70%), supporting the inverse relationship between DNA

methylation and these two histone modiﬁcations. Furthermore,

CpGs inside H3K4me3 and H3K27me3 marks exhibit more

methylation change, in comparison with CpGs inside H3K36me3

mark (Supplementary Fig. 2d).

To investigate the involvement of ESC-TFBSs in reprogram-

ming, we performed separate enrichment analysis for each DMR

group (Table 1). Polycomb-repressive complex (PRC)-binding

sites, including SUZ12, EZH2 and RING1B, were enriched in

DMR-1 and DMR-2b. On the other hand, sequence-speciﬁc

pluripotency-associated ESC-TFBSs such as Nanog, Oct4 and

Klf4 (but not CTCF and TET1)-binding sites were enriched in

DMR-3, the group of DMRs that are demethylated only in the

ESC-like state. These results demonstrate the dynamic changes in

DNA methylation at TFBSs, and the connection between the

pattern of changes and TFBS enrichment.

Dynamic changes of TFBS methylation during reprogram-

ming. Interrogating methylation changes at ESC-TFBSs resulted

in the detection of methylation depletion during high-dox treat-

ment, which was not apparent by examining DMRs (Fig. 3,

Supplementary Fig. 3; Methods). This was most obvious at the

binding sites for activated or overexpressed transcription factors

during early time points, such as OCT4, SOX2, KLF4 and

NANOG. These TFBSs also accumulated H3K4me3 modiﬁca-

tions that proceed after the methylation depletion. H3K27me3

marks diminished at binding sites of expressed transcription

factors early in reprogramming. In contrast, ESC-TFBSs for genes

that were not activated during high-dox reprogramming but are

known to play critical roles in ESC-like pluripotent state, such as

ESRRB and TCFCP2L1 (refs 14,27,28), showed no change in

DNA methylation and were demethylated only in the ESC-like

state. The PRC (SUZ12 and EZH2)-binding sites underwent a

gain of DNA methylation during reprogramming but showed

baseline levels of methylation in ESC.

We assessed DNA methylation changes occurring within

±40 kb of ESC-TFBSs (Fig. 4, Supplementary Fig. 4). At the

40 –40 0–40

Position around TFBS (kb)

OCT4

n= 3,761

SOX2

n= 4,526

0–40

Position around TFBS (kb)

040

Position around TFBS (kb)

SUZ12

n= 4,215

ESRRB

n= 21,647

–1 0 1

Normalized average

CpG methylation c hange versus 2MEF

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

Normalized average

H3K4me3 change versus 2MEF

Normalized average

H3K27me3 change versus 2MEF

–1 0 1 –1 0 1

Figure 4 | Histone modiﬁcation and DNA methylation change around transcription factor-binding sites. Average DNA methylation change (left),

average H3K4me3 change (middle) and average H3K27me3 change (right) in the 80-kb neighbourhood of transcription factor-binding sites.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619 ARTICLE

NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications 5

Expr-1a

(87)

Expr-1b

(38)

Expr-1c

(55)

Expr-2a

(113)

Expr-2b

(109)

Expr-3a

(41)

Expr-3b

(35)

100% 50% 0% 50% 100%

CpG density CountCount

00.02 0 30

010≤

FPKM

G met

0101

K4me3

K36me3

Pluripotency TFBS: OCT4, SOX2, KLF4, NANOG, ESRRB, TCFCP2L1,

PRC-binding sites: EZH2, SUZ12, RING1B

ESC specific H3K4me3 in promoters

2°MEF met ≥0.7

n= 673

2°MEF met ≤0.3

n= 93

2°iPSC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

ESC

ESC specific H3K27me3 in promoters

2°MEF met ≥0.7

n= 86

2°MEF met ≤0.3

n= 659

% Of promoters with H3K4me3 engagement

2°iPSC

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

ESC

% Of promoters with H3K27me3 engagement

2°MEF

D2H

D5H

D8H

D11H

D16H

D18H

D16L

D21L

D21Ø

1°iPSC

2°iPSC

ESC

CpGI

DNA methylation

H3K4me3

H3K27me3

RefSeq

Dnmt3l

Dppa4

Dppa5a

Eras

Esrrb

Nr5a2

Pecam1

Tcl1

Triml2

Zfp42

Cdh1

Epcam

Fgfbp1

Kit

Lef1

Lin28b

Mycl1

Nanog

Sall4

Slain1

Car4

Ina

Kis2

Lin28a

Nodal

Ooep

Pou3f1

Sall1

Tcf15

Utf1

Cbx4

Clec14a

Col14a1

Egfl6

Meox2

Mylk

Rspo2

Shox2

Spock3

Wnt16

Abi3bp

Col10a1

Emr1

Fabp4

Igf1

Lrrn3

Mfap2

Slc16a4

Stab1

Tyrobp

Drp2

Elavl4

Esx1

Klrg2

Mst1

Prox1

Shisa3

Sost

Sybu

Tmc3

Adamts9

Bmp4

Cgnl1

Des

Enox1

Fblim1

Gbp2

Icam1

Mdk

Smarca1

PRC-binding sites

CpG density

mRNA

expression

DNA

methylation

H3K4me3

H3K27me3

Pluripotency TFBS

H3K36me3

D18H

2°MEF

ESC

CpGI

D18H

2°MEF

ESC

D18H

2°MEF

ESC

D18H

2°MEF

ESC

D18H

2°MEF

ESC

K27me3

20 kb 20 kb 20 kb 20 kb 20 kb

Expr-1a Expr-1b Expr-1 c Expr-2a Expr-2b

Slain1 Car4 Pecam1 Rspo2 Abi3bp

Figure 5 | Epigenetic features of gene classes and model of gene expression control. (a) Genes were separated into clusters based on gene expression

patterns and DNA methylation. The heatmap presents mRNA expression, DNA methylation level of promoter regions, normalized H3K4me3 level,

normalized H3K27me3 level, CpG densities, pluripotency transcription factor-binding sites and binding sites of PRCs. (b) Base-level visualization of

DNA methylation and histone modiﬁcations in the promoter regions of representative genes for each class across all samples. (c) Percentage of

ESC-speciﬁc H3K4me3 mark for promoters with high and low initial methylation. (d) Percentage of ESC-speciﬁc H3K27me3 mark for promoters with high

and low initial methylation.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619

6NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications

binding sites of core ESC-pluripotency transcription factors

(OCT4, SOX2, KLF4 and NANOG), we observed rapid focal

demethylation during high-dox treatment (D2H-D18H) if the

factors were expressed. On the other hand, ESC-like cells (1°iPSC,

2°iPSC and ESC) exhibited extensive demethylation, up to 20 kb

distal from the binding sites. A similar but more delayed process

was also observed for H3K4me3 modiﬁcations. The broad

neighbourhoods around PRC-binding sites were hypermethylated

in all samples examined. Interestingly, although methylation

accumulated broadly around PRC (SUZ12, EZH2 and RING1B)-

binding sites (Fig. 4, Supplementary Fig. 4), these underwent

focal renormalization at the ESC-like pluripotent state. These sites

also demonstrate bivalent marks of H3K4me3 and H3K27me3

in ESC-like state24. The patterns of change to DNA methylation

and histone marks were distinct for the three types of

transcription factor shown (Figs 3 and 4). Our results show an

interesting contrast between the focal demethylation induced

early in reprogramming and broader demethylated regions at

ESC-like pluripotent state, perhaps representing a key

distinguishing feature of the pluripotent state where broader

demethylation is required for completion of the reprogramming

to ESC-like state.

We attempted to show that the dynamics of methylation

change at transcription factor-binding sites (TFBSs) could act as a

predictor of importance to the reprogramming process. We

proposed criteria for DNA-binding transcription factors of

41.2 enrichment and 410% overlap in DMR-3, implying

over-representation in DMRs that underwent demethylation at

transition to the ESC-like state, but little change early in

reprogramming. We tested a set of 118 transcription factors with

computationally predicted binding sites against these criteria29,30.

We found only three transcription factors (SOX2, MYC and

OCT4) that fulﬁlled our criteria, all of which are known to be

important in reprogramming to iPSCs (Supplementary Data 4).

This suggests a high speciﬁcity for the prediction criteria,

although sensitivity is low as other factors known to be

involved in reprogramming were not identiﬁed. Transcription

factors whose binding sites show signiﬁcant change in

methylation late in a transition can be called important to that

transition with high conﬁdence. We believe that methylome-

based tests of this nature could have useful application in

prediction of transcription factors involved in other cellular

transitions.

Demethylation leads to precise control of gene expression.We

integrated corresponding RNA expression data22 with our DNA

methylation and histone modiﬁcation data sets (Supplementary

Tables 2–4, Supplementary Data 5; Methods). Activation of

genes was associated with H3K4me3 occupancy in promoter

regions and repression was associated with either H3K27me3

occupancy or no histone mark (Supplementary Fig. 5a).

Moreover, as we observed in DMRs, engagement of both

H3K4me3 and H3K27me3 marks in promoters was dependent

on DNA methylation levels with a strong inverse relationship

(Supplementary Fig. 5b).

We selected 477 genes segregating into seven clusters on the

basis of expression and epigenetic change over the course of

reprogramming (Fig. 5a,b, Supplementary Table 5; Methods).

These groups represent: activated early in reprogramming (Expr-

1a), activated late in reprogramming with either low- (Expr-1b)

or full- (Expr-1c) DNA methylation in 2°MEF and repressed

during reprogramming with either low- (Expr-2a) or full-

(Expr-2b) DNA methylation in ESC. Genes in Expr-3a were

turned on, while those in Expr-3b were turned off in high-dox;

therefore, they were differentially expressed between D16H/D18H

(F-class cells) and ESC-like cells. Expression changes of genes in

Expr-1a and Expr-2a/b are likely responsible for pluripotency, as

they were differentially expressed between 2°MEF and pluripo-

tent cells21. Finally, the presence of genes in Expr-1b/c explains

why F-class cells are distinct from ESC-like state cells.

The expression dynamics through reprogramming of these

genes was clear upon visualization of the categories and

representative genes from each class (Fig. 5a,b, Supplementary

Fig. 5a–d). Genes repressed by H3K27me3 with low-methylated

promoters in 2°MEF tended to be activated early in reprogram-

ming and had CpG-rich promoters (Expr-1a/b). These loci were

enriched in genes involved in cell adhesion, such as Epcam and

Cdh1 (Fig. 5a (Expr-1a)). In contrast, quiescence of Expr-1c genes

was initially safeguarded by DNA methylation of CpG-poor

promoters, and H3K4me3 was only acquired after late demethy-

lation. The same two modes of control were observed for the

genes repressed by reprogramming. However, as in the analysis of

DMRs, DNA methylation in promoter regions happened early in

reprogramming (Expr-2b), whereas demethylation was detected

exclusively in the ESC-like state, revealing that a gain of

methylation is kinetically favoured over demethylation. This is

also true for histone marks in relation to changes in gene

expression, where histone modiﬁcations, speciﬁcally the modula-

tion of H3K27me3, occurred early during reprogramming

(Expr-2a) within low-methylated promoters. Interestingly, the

dynamic process of histone modiﬁcation alterations during

reprogramming was strongly inﬂuenced by the starting methyla-

tion state of gene promoters (Fig. 5c,d). Genes with low-

methylated promoters at 2°MEF showed a signiﬁcantly higher

rate of transition to the ESC-like state for both ESC-speciﬁc

histone marks compared with those with fully methylated

promoters. This suggests that DNA methylation presents a major

barrier during somatic cell reprogramming to ESC-like cells and

that the methylation status of a given region determines its

control by histone modiﬁcations.

We propose a model that describes the key mechanism of

epigenetic control of gene expression during reprogramming

Epigenomic control of gene expression

Off

Late activation

(Expr-1c)

Late activation (Expr-1c)

Repression (Expr-2b)

Repression

(Expr-2b)

Repression (Expr-2a)

Early activation (Expr-1a)

Off

Late activation

(Expr-1b)

Histone modification-driven

(CpG rich)

Off

DNA methylation-driven

(CpG poor)

Off

H3K27me3

H3K4me3

Un-meC

5meC

(Expr-3a)

(Expr-3b)

Figure 6 | A model summarizing DNA methylation and histone

modiﬁcation-driven control of gene expression. Dashed arrow represents

the strict control of demethylation. Gene classes affected by changes

are shown in brackets accompanying arrows.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619 ARTICLE

NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications 7

(Fig. 6). In genes with CpG-poor promoters, control is driven by

DNA methylation. Such genes may be activated by demethylation

followed by H3K4me3 engagement, producing expression proﬁles

characteristic of class Expr-1c/2b. In genes with CpG-rich

promoters, low methylation levels allow histone modiﬁcation-

driven control. This model is supported by data showing the role

of initial methylation status as a modulator of the dynamic

changes to histone modiﬁcation, and the sequential modiﬁcation

of DNA methylation followed by histone marks in TFBSs. The

model also accounts for characteristic gene expression classes

(detailed in Figs 5 and 6). We predict that this mechanism may

not only apply to iPSC reprogramming but also to lineage

speciﬁcation of cells. Therefore, our insights into how DNA

methylation controls the epigenetic landscape in reprogramming

to pluripotency could be crucial to a better understanding of the

mechanisms underlying general cell fate change, and could have

ramiﬁcations for stem cell-based therapies.

Methods

Cell culture and secondary reprogramming.ROSA26-rtTA-IRES-GFP mouse

ESC, iPSCs and mouse embryonic ﬁbroblasts were cultured as previously descri-

bed31. ESCs and iPSCs were cultured in 5% CO

at 37 °C on irradiated MEFs in

DMEM containing 15% FCS, leukaemia-inhibiting factor, penicillin/streptomycin,

L-glutamine, nonessential amino acids, sodium pyruvate and 2-mercaptoethanol.

1B 1°iPS cells were aggregated with tetraploid host embryos as described10 and

MEFs established from E13.5 embryos. High-dox cell samples were collected on

days 0, 2, 5, 8, 11, 16 and 18 (D2H, D5H, D8H, D11H, D16H and D18H). A

subculture of the reprogramming cells was established from day 19 and cultured in

the absence of dox, to develop a factor-independent 2°iPS cell line by day 30

(2°iPSC). Low-dox samples were maintained from day 8 to day 14 cells in 5 ng dox.

On day 14 the culture was diverged into two, with some of the cells being cultured

until day 21 in the absence of dox (D21Ø) and the remainder being cultured in

5ngml1of dox and collected on day 16 (D16L) and (D21L). Rosa26rtTA ESCs

and 1B 1o iPSCs were collected as controls.

MethylC-Seq library generation.For all 13 samples (2°MEF, D2H, D5H, D8H,

D11H, D16H, D18H, D16L, D21L, D21Ø, 1°iPSC, 2°iPSC and rtTA ESC), 5 mg of

genomic DNA was mixed with 25 ng unmethylated cl857 Sam7 Lambda DNA

(Promega, Madison, WI, USA). The DNA was fragmented by sonication to

300–500 bp with a Covaris S2 system (Covaris) followed by end repair with the

End-It DNA End-Repair Kit (Epicenter). Paired-end universal library adaptors

provided by Illumina were ligated to the sonicated DNA as per the manufacturer’s

instructions for genomic DNA library construction. Ligated products were puriﬁed

with AMPure XP beads (Beckman, Brea, CA, USA). Adaptor-ligated DNA was

bisulﬁte-treated using the EpiTect Bisulﬁte Kit (QIAGEN) following the manu-

facturer’s instructions and then PCR-ampliﬁed using PfuTurboCx Hotstart DNA

polymerase (Agilent, Santa Clara, CA, USA) with the following PCR conditions

(2 min at 95 °C, 4 cycles of 15 s at 98 °C, 30 s at 60 °C, 4 min at 72 °C and then

10 min at 72 °C). The reaction products were puriﬁed using the MinElute gel

puriﬁcation kit (QIAGEN). The sodium bisulﬁte non-conversion rate was

calculated as the percentage of cytosines sequenced at cytosine reference positions

in the lambda genome.

ChIP library generation.ChIP was carried out as described in ref. 32. In all,

40–150 million cells were ﬁxed with 1% formaldehyde for 10 min at room

temperature, and scraped and stored as pellets ( 80 °C). Samples were lysed at

20 million cells per ml Farnham lysis buffer for 10 min and subsequently at

10 million cells per ml nuclear lysis buffer. The released chromatin was sheared to

100–500 bp (250 bp average) on ice using a SonicsVibraCell Sonicator equipped

with a 3-mm probe. For each sample, 50 ml of solubilized chromatin was used as

input DNA to normalize sequencing results and the remaining chromatin was

immunoprecipitated with 10 mg of H3K4me3 (ab8580)33,10mg H3K27me3

(Millipore 07-449)16 or 10 mg H3K36me3 (ab9050)16 antibodies, separately.

Antibody–chromatin complexes were pulled down with 100 ml magnetic Protein G

Dynal beads (Invitrogen) and washed six times. The chromatin was then eluted,

reverse crosslinked at 65 °C overnight and subjected to RNaseA/proteinase K

treatment. ChIP and input DNA were puriﬁed using a Qiagen Puriﬁcation Column

and quantiﬁed using a Quant-it dsDNA High Sensitivity Assay (Invitrogen). For

ChIP sequencing, ChIP-seq libraries were prepared according to the protocols

described in the Illumina ChIP-seq library preparation kit. Brieﬂy, 50 ng of

immunopuriﬁed DNA or 100 ng of genomic DNA from an input sample was end-

repaired, followed by the 30addition of a single adenosine nucleotide and ligation

to universal library adapters. Ligated material was separated on a 2.0% agarose gel,

followed by the excision of a 250- to 350-bp fragment and column puriﬁcation

(QIAGEN). DNA libraries were prepared by PCR ampliﬁcation (18 cycles).

High-throughput sequencing.MethylC-Seq DNA and ChIP DNA libraries were

sequenced using the Illumina HiSeq 2000 as per the manufacturer’s instructions.

Sequencing of libraries was performed up to 2 101 cycles. Image analysis and

base calling were performed with the standard Illumina pipeline version RTA 2.8.0.

Processing and alignment of MethylC-Seq data.MethylC-Seq sequencing data

were processed using the Illumina analysis pipeline, and FastQ format reads were

aligned to the NCBI37/mm9 mouse reference using the Bismark/Bowtie alignment

algorithm18,34,35. Paired-read MethylC-Seq sequences produced by the Illumina

pipeline in FastQ format were trimmed with trim threshold 1,500; we removed the

last two bases from sequences that were not trimmed and removed three bases

from sequences that were trimmed. The Bismark package version 0.7.7 was used as

the aligner using the following parameters: -e 90 -n 2 -l 32 -X 550. As up to six

independent libraries from each biological replicate were sequenced, we ﬁrst

removed duplicate reads. Subsequently, the reads from all libraries of a particular

sample were combined. Unique read alignments were then subjected to post-

processing. The number of calls for each base at every reference sequence position

and on each strand was calculated. All results of aligning a read to both the Watson

and Crick converted genome sequences were combined. The CpG methylation

levels were calculated using bisulﬁte conversion rates by (Number of not converted

Cs per read depth) for each position (Supplementary Data 1).

RNA-Seq library generation and sequencing.Total RNA was subjected to two

rounds of on column DNAseI treatment to remove contaminating DNA using the

RNase-Free DNase set (Qiagen PN 79254) as per the manufacturer’s protocol. The

total RNA was then analysed using the Agilent RNA 6000 Nano Kit (PN 5067-

1511) on the Agilent Bioanalyzer 2100 (PN G2939AA) to quantify yield, qualify

integrity and conﬁrm removal of DNA contamination.

Following DNAseI treatment, 5 mg total RNA from each sample was depleted

of ribosomal RNA using the Ribo-ZerorRNA Removal Kit (Epicenter PN

RZH110424) as per the manufacturer’s instructions. The ribosomal-depleted RNAs

were then run on an Agilent RNA 6000 Pico Kit (PN 5067-1513) on the Agilent

Bioanalyzer 2100 to conﬁrm ribosomal RNA depletion. Sequencing libraries where

generated from the ribosomal-depleted RNA using the SOLiD Transcriptome

Multiplexing Kit (PN 4427046) from Applied Biosystems following the

manufacturer’s publication. Final libraries were quantiﬁed and qualiﬁed using the

Agilent High Sensitivity DNA Kit (PN 5067-4626) on the Agilent Bioanalyzer 2100.

Sequencing libraries were subsequently pooled in equimolar ratios (four

libraries per pool) and clonally ampliﬁed on SOLiD nanobeads. Clonal

ampliﬁcation was completed via emulsion PCR using the SOLiD EZ Bead System

(PN 4448419, 4448418 and 4448420) coupled with SOLiD EZ Bead N200

ampliﬁcation reagents (PN 4467267, 4457185, 4467281, 4467283 and 4467282).

Following emulsion PCR, clonally ampliﬁed nanobeads were enriched using the

SOLiD EZ Bead Enricher Kits (PN 4467276, 4444140 and 4453073) before being

deposited into SOLiD 6-Lane FlowChip (PN 4461826) using the SOLiD Flowchip

Deposition Kit v2 (PN 4468081) as per the manufacturer’s recommendations.

In total, two ﬂowchips were sequenced yielding a total of eight lanes of data,

with sequencing reads generated using the SOLiD 5500xl platform generating

paired 75 bp forward and 35 bp reverse reads. To allow de-convolution of the

pooled libraries, a single 5-bp index read was generated. A total of 1,204,676,394

fragments (2,409,352,788 reads) were generated post deconvolution, ranging from

35,714,748 to 147,282,580 fragments per library.

Processing and alignment of RNA-Seq data.Sequence mapping was performed

using Applied Biosystems LifeScope v2.5 whole transcriptome (paired-end) ana-

lysis pipeline against the NCBIM37 (mm9) genome and exon-junction libraries

constructed from the Ensembl v64 gene model. Brieﬂy, this pipeline ﬁrst removes

potential contaminant reads by aligning to a ﬁlter set containing rRNA, tRNA,

adaptor sequences and retrotransposon sequences. Following ﬁltering, LifeScope

then aligns all reads to the genome and F3 reads to the junction library. F5 reads

are additionally aligned at a higher sensitivity to exonic sequences within insert size

distance from the paired (F3) read alignment. Read alignments are merged and

disambiguated, and a single BAM (binary alignment/mapped) ﬁle output per

library.

BAM ﬁles were then additionally ﬁltered to remove reads with a mapping

quality (MAPQ)o9 and all mitochondrial reads. Alignments were then assembled

using Cufﬂinks (v2.0.2) using the –G parameter to quantify gene and isoform

FPKM expression values against the reference gene model (Ensembl v67).

Identiﬁcation of methylated cytosines.At each reference cytosine, the binomial

distribution was used to identify whether at least a subset of the genomes within the

sample were methylated, using a 0.01 FDR-corrected Pvalue. We identiﬁed methyl

cytosines while keeping the number of false-positive methylcytosine calls below 1%

of the total number of methyl cytosines we identiﬁed. The probability Pin the

binomial distribution B(n, P) was estimated from the number of cytosine bases

sequenced in reference cytosine positions in the unmethylated Lambda genome

(referred to as the error rate: nonconversion plus sequencing error frequency). We

interrogated the sequenced bases at each reference cytosine position one at a time,

where read depth refers to the number of reads covering that position. For each

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619

8NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications

position, the number of trials (n) in the binomial distribution was the read depth.

For each possible value of nwe calculated the number of cytosines sequenced (k)at

which the probability of sequencing kcytosines out of ntrials with an error rate of

pwas less than the value M, where M* (number of unmethylated cytosines) o0.01*

(number of methylated cytosines) and if the error rate of pwas over 0.01, we

assumed that the cytosine was not methylated. In this way, we established the

minimum threshold number of cytosines sequenced at each reference cytosine

position at which the position could be called as methylated, so that out of all

methyl cytosines identiﬁed no more than 1% would be because of the error rate.

Calculation of DNA methylation level.If the error rate is less than 0.01 we

calculated adjusted DNA methylation level for cytosine as follow:

Adjusted cytosine methylation level ¼ab



að1Þ

(a¼total Cs, b¼number of converted Cs, cr ¼bisulﬁte conversion rate).

Identiﬁcation of DMRs.DMRs (Fig. 2) were identiﬁed using a sliding window

approach (Supplementary Fig. 6a, Fig. 2b). A window size of 30 CpGs less than 6 kb

with coverage more than 5 in 15 CpGs per window in all samples were con-

sidered, progressing one CpG per iteration. Total of 20,214,978 windows were

assessed. Windows showing maximum difference and fold enrichment of 30% and

fourfold with Benjamini–Hochberg-corrected FDR from analysis of variance

(ANOVA) test Pvalues of less than 1% were identiﬁed as differentially methylated

windows. In all, 188,529 differentially methylated windows were then joined if

regions were overlapped or progressing region and the succeeding regions were

covering more than 60% of the region. This set of 7,890 DMRs covering

21,618,964 bp of the whole genome are reported in Fig. 2 and Supplementary Data 2.

DMRs were then deﬁned as Hyper-DMRs and Hypo-DMRs if the average

methylation level difference of each DMR in each sample was higher or lower by

more than 20% relative to 2°MEF.

Mapping and enrichment analysis of ChIP-Seq reads.Paired-end ChIP-Seq

data were processed using the Illumina analysis pipeline, and mapping was con-

ducted using Bowtie version 0.12.8 with the following parameters: --pairtries 100 -y

-k 1 -n 3 -l 50 -I 0 -X 1000. Enrichment analysis was conducted using MACS36

with parameters of --nomodel -S -w –n –space 30.

ChIP-Seq data analysis.Enriched peaks from ChIP-Seq data were joined into

clusters where at least one sample has a peak for each modiﬁcation (H3K4me3,

H3K27me3 and H3K36me3; Supplementary Fig. 6b). The total peak width of each

sample within the cluster was calculated as histone mark score within clusters.

TFBS epigenomic change analysis.ESC-TFBSs of mouse ESCs were obtained

from different studies23–25. CpG methylation level of each TFBS in each sample

was calculated. The average CpG methylation change of each TFBS was than

calculated in each sample relative to 2°MEF (Fig. 3). For calculating CpG

methylation change around ESC-TFBSs, the same procedure was applied for 200 bp

400 bins around each ESC-TFBS. The same procedure using enrichment score for

30-bp window was applied for calculating average histone modiﬁcation change

(Fig. 4).

Genome annotation.Genomic regions and CpG islands were deﬁned based on

NCBI37/mm9 coordinates downloaded from the UCSC website (http://geno-

me.ucsc.edu/). Promoters were arbitrarily deﬁned as 5 kb upstream and 1 kb

downstream of transcriptional start site for each Ensembl release-67 transcript.

Gene bodies are deﬁned as from transcription start to end sites for each transcript.

Histone modiﬁcation clusters and DMRs were annotated if they overlap with their

promoters.

Fold-enrichment test.Fold enrichment was calculated as follows: (Observed

number of Xin examining region/total length of examining region (bp))/(total

number of Xin reference region/reference region length (bp)), X¼genomic

feature)).

Gene expression pattern separation.We selected genes of expression patterns as

described in Supplementary Table 5.

Data integration and normalization.DNA methylation levels of promoters were

calculated from 5 kb upstream and 1 kb downstream of the transcription start site.

H3K4me3 and H3K27me3 marks were considered if their cluster of peaks were

overlapped with promoters. Overlapped H3K36me3 peaks were calculated for

whole gene. In Fig. 5, for calculating normalized histone modiﬁcation scores,

maximum peak width was considered as 1 and relative widths were calculated for

each sample in each gene.

Accession codes.Methylome sequencing data are available under the European

Nucleotide Archive accessions no. ERP004116 (http://www.ebi.ac.uk/ena/data/

view/PRJEB4795). Long RNA-seq and Chip-seq sequencing data are available

under the NCBI Sequence Read Archive (SRA) accessions no. SRP046744 (http://

www.ncbi.nlm.nih.gov/sra). Analysed data sets can be obtained from Stemfor-

matics (www.stemformatics.org)37.

References

1. Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse

embryonic and adult ﬁbroblast cultures by deﬁned factors. Cell 126, 663–676

(2006).

2. Maherali, N. et al. Directly reprogrammed ﬁbroblasts show global epigenetic

remodeling and widespread tissue contribution. Cell Stem Cell 1, 55–70 (2007).

3. Takahashi, K. et al. Induction of pluripotent stem cells from adult human

ﬁbroblasts by deﬁned factors. Cell 131, 861–872 (2007).

4. Yu, J. et al. Induced pluripotent stem cell lines derived from human somatic

cells. Science 318, 1917–1920 (2007).

5. Park, I. H. et al. Reprogramming of human somatic cells to pluripotency with

deﬁned factors. Nature 451, 141–146 (2008).

6. Kang, L., Wang, J., Zhang, Y., Kou, Z. & Gao, S. iPS cells can support full-term

development of tetraploid blastocyst-complemented embryos. Cell Stem Cell 5,

135–138 (2009).

7. Zhao, X. Y. et al. iPS cells produce viable mice through tetraploid

complementation. Nature 461, 86–90 (2009).

8. Onder, T. T. et al. Chromatin-modifying enzymes as modulators of

reprogramming. Nature 483, 598–602 (2012).

9. Singhal, N. et al. Chromatin-remodeling components of the BAF complex

facilitate reprogramming. Cell 141, 943–955 (2010).

10. Woltjen, K. et al. piggyBac transposition reprograms ﬁbroblasts to induced

pluripotent stem cells. Nature 458, 766–770 (2009).

11. Samavarchi-Tehrani, P. et al. Functional genomics reveals a BMP-driven

mesenchymal-to-epithelial transition in the initiation of somatic cell

reprogramming. Cell Stem Cell 7, 64–77 (2010).

12. Mikkelsen, T. S. et al. Dissecting direct reprogramming through integrative

genomic analysis. Nature 454, 49–55 (2008).

13. Polo, J. M. et al. A molecular roadmap of reprogramming somatic cells into iPS

cells. Cell 151, 1617–1632 (2012).

14. Buganim, Y. et al. Single-cell expression analyses during cellular

reprogramming reveal an early stochastic and a late hierarchic phase. Cell 150,

1209–1222 (2012).

15. Chen, J. et al. Vitamin C modulates TET1 function during somatic cell

reprogramming. Nat. Genet. 45, 1504–1509 (2013).

16. Wang, T. et al. The histone demethylases Jhdm1a/1b enhance somatic cell

reprogramming in a vitamin-C-dependent manner. Cell Stem Cell 9, 575–587

(2011).

17. Plath, K. & Lowry, W. E. Progress in understanding reprogramming to the

induced pluripotent state. Nat. Rev. Genet. 12, 253–265 (2011).

18. Lister, R. et al. Hotspots of aberrant epigenomic reprogramming in human

induced pluripotent stem cells. Nature 471, 68–73 (2011).

19. Papp, B. & Plath, K. Epigenetics of reprogramming to induced pluripotency.

Cell 152, 1324–1343 (2013).

20. Surani, M. A., Hayashi, K. & Hajkova, P. Genetic and epigenetic regulators of

pluripotency. Cell 128, 747–762 (2007).

21. Tonge, P. D. et al. Divergent reprogramming routes lead to alternative stem cell

states. Nature doi: 10.1038/nature14047 (2014).

22. Hussein, S. M. I. et al. Genome-wide characterization of the routes to

pluripotency. Nature doi: 10.1038/nature14046 (2014).

23. Chen, X. et al. Integration of external signaling pathways with the core

transcriptional network in embryonic stem cells. Cell 133, 1106–1117

(2008).

24. Ku, M. et al. Genomewide analysis of PRC1 and PRC2 occupancy identiﬁes two

classes of bivalent domains. PLoS Genet. 4, e1000242 (2008).

25. Wu, H. et al. Dual functions of Tet1 in transcriptional regulation in mouse

embryonic stem cells. Nature 473, 389–393 (2011).

26. Visel, A., Minovitsky, S., Dubchak, I. & Pennacchio, L. A. VISTA Enhancer

Browser--a database of tissue-speciﬁc human enhancers. Nucleic Acids Res. 35,

D88–D92 (2007).

27. Feng, B. et al. Reprogramming of ﬁbroblasts into induced pluripotent stem cells

with orphan nuclear receptor Esrrb. Nat. Cell Biol. 11, 197–203 (2009).

28. Fischedick, G. et al. Zfp296 is a novel, pluripotent-speciﬁc reprogramming

factor. PloS ONE 7, e34645 (2012).

29. Stormo, G. D. DNA binding sites: representation and discovery. Bioinformatics

16, 16–23 (2000).

30. Ho Sui, S. J. et al. oPOSSUM: identiﬁcation of over-represented transcription

factor binding sites in co-expressed genes. Nucleic Acids Res. 33, 3154–3164

(2005).

31. Nagy, A. & Gertsenstein, M. Manipulating the Mouse Embryo: A Laboratory

Manual (Cold Spring Harbor Press, 2003).

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619 ARTICLE

NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications 9

32. O’Geen, H., Echipare, L. & Farnham, P. J. in Epigenetics Protocols 791, 265–286

(Humana Press, 2011).

33. Gaspar-Maia, A. et al. Chd1 regulates open chromatin and pluripotency of

embryonic stem cells. Nature 460, 863–868 (2009).

34. Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-

efﬁcient alignment of short DNA sequences to the human genome. Genome

Biol. 10, R25 (2009).

35. Krueger, F. & Andrews, S. R. Bismark: a ﬂexible aligner and methylation caller

for bisulﬁte-Seq applications. Bioinformatics 27, 1571–1572, 2011).

36. Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9,

R137 (2008).

37. Wells, C. A. et al. Stemformatics: visualisation and sharing of stem cell gene

expression. Stem Cell Res. 10, 387–395 (2013).

Acknowledgements

This work has been supported by the Korean Ministry of Knowledge Economy (grant no.

10037410 to J.-S.S.), by the SNUCM research fund (grant no. 0411-20100074 to J.-S.S.)

and by Macrogen Inc. (grant no. MGR03-11 and 12).

Author contributions

J.-S.S. and A.N. conceived and designed the experiments. J.-Y.S., J.-Y.Y., J.K., K.-S.Y. and

H.K. performed MethylC-Seq and ChIP-Seq experiments. P.D.T derived iPSC lines.

M.C.P., M.L., S.M.I.H. and I.M.R. performed pull downs for ChIP-Seq. N.C. and

S.M.G. performed RNA-Seq. D.-S.L. performed sequencing data processing. D.-S.L.,

S.L., W.-C.L. and H.R. conducted bioinformatic and statistical analyses. J.-S.S., D.-S.L.,

J.-Y.S., H.P., T.B. and J.-H.Y. wrote the manuscript.

Additional information

Accession codes: Methylome sequencing data are available under the European Nucleo-

tide Archive accessions no ERP004116 (http://www.ebi.ac.uk/ena/data/view/PRJEB4795).

Long RNA-seq and Chip-seq sequencing data areavailable underthe NCBI SequenceRead

Archive (SRA) accessions no. SRP046744 (http://www.ncbi.nlm.nih.gov/sra).

Supplementary Information accompanies this paper at http://www.nature.com/

naturecommunications

Competing ﬁnancial interests: The authors declare no competing ﬁnancial interests.

Reprints and permission information is available online at http://npg.nature.com/

reprintsandpermissions/

How to cite this article: Lee, D.-S. et al. An epigenomic roadmap to induced

pluripotency reveals DNA methylation as a reprogramming modulator. Nat. Commun.

5:5619 doi: 10.1038/ncomms6619 (2014).

This work is licensed under a Creative Commons Attribution 4.0

International License. The images or other third party material in this

article are included in the article’s Creative Commons license, unless indicated otherwise

in the credit line; if the material is not included under the Creative Commons license,

users will need to obtain permission from the license holder to reproduce the material.

To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms6619

10 NATURE COMMUNICATIONS | 5:5619 | DOI: 10.1038/ncomms6619 | www.nature.com/naturecommunications

Lineage Reprogramming: Genetic, Chemical, and Physical Cues for Cell Fate Conversion with a Focus on Neuronal Direct Reprogramming and Pluripotency Reprogramming

Article

Full-text available

Apr 2024

Although lineage reprogramming from one cell type to another is becoming a breakthrough technology for cell-based therapy, several limitations remain to be overcome, including the low conversion efficiency and subtype specificity. To address these, many studies have been conducted using genetics, chemistry, physics, and cell biology to control transcriptional networks, signaling cascades, and epigenetic modifications during reprogramming. Here, we summarize recent advances in cellular reprogramming and discuss future directions.

Publisher Correction: Mechanisms, pathways and strategies for rejuvenation through epigenetic reprogramming

Article

Full-text available

Jan 2024

Mechanisms, pathways and strategies for rejuvenation through epigenetic reprogramming

Article

Full-text available

Dec 2023

Over the past decade, there has been a dramatic increase in efforts to ameliorate aging and the diseases it causes, with transient expression of nuclear reprogramming factors recently emerging as an intriguing approach. Expression of these factors, either systemically or in a tissue-specific manner, has been shown to combat age-related deterioration in mouse and human model systems at the cellular, tissue and organismal level. Here we discuss the current state of epigenetic rejuvenation strategies via partial reprogramming in both mouse and human models. For each classical reprogramming factor, we provide a brief description of its contribution to reprogramming and discuss additional factors or chemical strategies. We discuss what is known regarding chromatin remodeling and the molecular dynamics underlying rejuvenation, and, finally, we consider strategies to improve the practical uses of epigenetic reprogramming to treat aging and age-related diseases, focusing on the open questions and remaining challenges in this emerging field.

Dynamics of cell-type transition mediated by epigenetic modifications

Article

Nov 2023
J THEOR BIOL

multiomics: A user-friendly multi-omics data harmonisation R pipeline

Article

Aug 2023

Data from multiple omics layers of a biological system is growing in quantity, heterogeneity and dimensionality. Simultaneous multi-omics data integration is of immense interest to researchers as it has potential to unlock previously hidden biomolecular relationships leading to early diagnosis, prognosis, and expedited treatments. Many tools for multi-omics data integration are developed. However, these tools are often restricted to highly specific experimental designs, types of omics data, and specific data formats. A major limitation of the field is the lack of a pipeline that can accept data in unrefined form to preserve maximum biology in an individual dataset prior to integration. We fill this gap by developing a flexible, generic multi-omics pipeline called multiomics , to facilitate general-purpose data exploration and analysis of heterogeneous data. The pipeline takes unrefined multi-omics data as input, sample information and user-specified parameters to generate a list of output plots and data tables for quality control and downstream analysis. We have demonstrated its application on a sepsis case study. We enabled limited checkpointing functionality where intermediate output is staged to allow continuation after errors or interruptions in the pipeline and generate a script for reproducing the analysis to improve reproducibility. Our pipeline can be installed as an R package or manually from the git repository, and is accompanied by detailed documentation with walkthroughs on three case studies.

Compromised Mitotic Fidelity in Human Pluripotent Stem Cells

Preprint

Full-text available

Jun 2023

Human pluripotent stem cells (PSCs), which include both embryonic and induced pluripotent stem cells, are widely used in fundamental and applied biomedical research. They have been instrumental for better understanding development and cell differentiation processes, disease origin and progression, and can aid in the discovery of new drugs. PSCs also hold great potential in regenerative medicine to treat or diminish the effects of certain debilitating diseases, such as degenerative disorders. However, some concerns have recently been raised over their safety for the use in regenerative medicine. One of the major concerns is the fact that PSCs are prone to errors in passing the correct number of chromosomes to daughter cells, resulting in aneuploid cells. Aneuploidy, characterised by an imbalance in chromosome number, elicits the upregulation of different stress pathways that are deleterious to cell homeostasis, impair proper embryo development and can potentiate cancer development. In this review we will summarise known molecular mechanisms recently revealed to impair mitotic fidelity in human PSCs and the consequences of the decreased mitotic fidelity of these cells. We will finish with speculative views on how the physiological characteristics of PSCs can affect the mitotic machinery and how their suboptimal mitotic fidelity may be circumvented.

Compromised Mitotic Fidelity in Human Pluripotent Stem Cells

Article

Full-text available

Jul 2023
INT J MOL SCI

Human pluripotent stem cells (PSCs), which include both embryonic and induced pluripotent stem cells, are widely used in fundamental and applied biomedical research. They have been instrumental for better understanding development and cell differentiation processes, disease origin and progression and can aid in the discovery of new drugs. PSCs also hold great potential in regenerative medicine to treat or diminish the effects of certain debilitating diseases, such as degenerative disorders. However, some concerns have recently been raised over their safety for use in regenerative medicine. One of the major concerns is the fact that PSCs are prone to errors in passing the correct number of chromosomes to daughter cells, resulting in aneuploid cells. Aneuploidy, characterised by an imbalance in chromosome number, elicits the upregulation of different stress pathways that are deleterious to cell homeostasis, impair proper embryo development and potentiate cancer development. In this review, we will summarize known molecular mechanisms recently revealed to impair mitotic fidelity in human PSCs and the consequences of the decreased mitotic fidelity of these cells. We will finish with speculative views on how the physiological characteristics of PSCs can affect the mitotic machinery and how their suboptimal mitotic fidelity may be circumvented.

A Rejuvenation Signature in Skeletal Muscle That Is Mediated By Late-Life Exercise

Article

Full-text available

May 2023

The objectives of this study were to measure to what extent exercise stimulates partial molecular reprogramming in skeletal muscle. We hypothesized that exercise training, beyond its known functional improvements in aged tissues, could recapitulate some of the same effects that are seen with in vivo partial reprogramming by Yamanaka factors. Using transcriptome profiling from 1) a skeletal muscle-specific in vivo Oct3/4, Klf4, Sox2, and Myc (OKSM) reprogramming-factor expression murine model, 2) an in vivo inducible muscle-specific Myc induction murine model, 3) a translatable high-volume hypertrophic exercise training approach in aged mice, and 4) human exercise muscle biopsies, we collectively defined exercise-induced genes common to partial reprogramming. Late-life exercise training lowered murine DNA methylation age according to several contemporary muscle-specific clocks. A comparison of the murine soleus transcriptome after late-life exercise training to the soleus transcriptome after OKSM induction revealed an overlapping signature. Within this signature, downregulation of specific mitochondrial and muscle-enriched genes was conserved in skeletal muscle of long-term exercise-trained humans; among these was muscle-specific Abra/Stars. Myc is the OKSM factor most induced by exercise in muscle and was elevated following exercise training in aged mice. A pulse of MYC rewired the global soleus muscle methylome, and the transcriptome after a MYC pulse partially recapitulated OKSM induction. A common signature also emerged in the murine MYC-controlled and exercise adaptation transcriptomes, including lower muscle-specific Melusin and reactive oxygen species-associated Romo1. With Myc, OKSM, and exercise training in mice as well habitual exercise in humans, the complex I accessory subunit Ndufb11 was lower; low Ndufb11 is linked to longevity in rodents. Our data collectively suggest: 1) a biological age-mitigating effect on the epigenetic landscape by late-life exercise-training in murine skeletal muscle, 2) a common gene expression signature of partial reprogramming by OKSM and exercise training in muscle of humans and aged mice, and 3) that Myc is an exercise-responsive factor that contributes to a rewired molecular profile at the transcriptome and methylome levels. National Institutes of Health: Kevin A. Murach, AG063994; American Federation for Aging Research (AFAR): Kevin A. Murach, Junior Investigator Grant This is the full abstract presented at the American Physiology Summit 2023 meeting and is only available in HTML format. There are no additional versions or additional content available for this abstract. Physiology was not involved in the peer review process.

Dynamics of cell type transition mediated by epigenetic modifications

Preprint

Full-text available

Apr 2023

Maintaining tissue homeostasis requires proper regulation of stem cell differentiation. The Waddington landscape suggests that gene circuits in a cell form a potential landscape of different cell types, with cells developing into different cell types following attractors of the probability landscape. However, it remains unclear how adult stem cells balance the trade-off between self-renewal and differentiation. We propose that random inheritance of epigenetic states plays a crucial role in stem cell differentiation and develop a hybrid model of stem cell differentiation induced by epigenetic modifications. Our model integrates a gene regulation network, epigenetic state inheritance, and cell regeneration to form multi-scale dynamics ranging from transcription regulation to cell population. Our simulation investigates how random inheritance of epigenetic states during cell division can automatically induce cell differentiation, dedifferentiation, and transdifferentiation. We show that interfering with epigenetic modifications or introducing extra transcription factors can regulate the probabilities of dedifferentiation and transdifferentiation, revealing the mechanism of cell reprogramming. This \textit{in silico} model offers insights into the mechanism of stem cell differentiation and cell reprogramming.

Novel Approaches to Studying SLC13A5 Disease

Article

Full-text available

Jan 2024

Adriana S Beltran

The role of the sodium citrate transporter (NaCT) SLC13A5 is multifaceted and context-dependent. While aberrant dysfunction leads to neonatal epilepsy, its therapeutic inhibition protects against metabolic disease. Notably, insights regarding the cellular and molecular mechanisms underlying these phenomena are limited due to the intricacy and complexity of the latent human physiology, which is poorly captured by existing animal models. This review explores innovative technologies aimed at bridging such a knowledge gap. First, I provide an overview of SLC13A5 variants in the context of human disease and the specific cell types where the expression of the transporter has been observed. Next, I discuss current technologies for generating patient-specific induced pluripotent stem cells (iPSCs) and their inherent advantages and limitations, followed by a summary of the methods for differentiating iPSCs into neurons, hepatocytes, and organoids. Finally, I explore the relevance of these cellular models as platforms for delving into the intricate molecular and cellular mechanisms underlying SLC13A5-related disorders.

Corrigendum: Genome-wide characterization of the routes to pluripotency

Article

Full-text available

Dec 2014

Somatic cell reprogramming to a pluripotent state continues to challenge many of our assumptions about cellular specification, and despite major efforts, we lack a complete molecular characterization of the reprograming process. To address this gap in knowledge, we generated extensive transcriptomic, epigenomic and proteomic data sets describing the reprogramming routes leading from mouse embryonic fibroblasts to induced pluripotency. Through integrative analysis, we reveal that cells transition through distinct gene expression and epigenetic signatures and bifurcate towards reprogramming transgene-dependent and -independent stable pluripotent states. Early transcriptional events, driven by high levels of reprogramming transcription factor expression, are associated with widespread loss of histone H3 lysine 27 (H3K27me3) trimethylation, representing a general opening of the chromatin state. Maintenance of high transgene levels leads to re-acquisition of H3K27me3 and a stable pluripotent state that is alternative to the embryonic stem cell (ESC)-like fate. Lowering transgene levels at an intermediate phase, however, guides the process to the acquisition of ESC-like chromatin and DNA methylation signature. Our data provide a comprehensive molecular description of the reprogramming routes and is accessible through the Project Grandiose portal at http://www.stemformatics.org.

Corrigendum: Divergent reprogramming routes lead to alternative stem-cell states

Article

Full-text available

Dec 2014

Pluripotency is defined by the ability of a cell to differentiate to the derivatives of all the three embryonic germ layers: ectoderm, mesoderm and endoderm. Pluripotent cells can be captured via the archetypal derivation of embryonic stem cells or via somatic cell reprogramming. Somatic cells are induced to acquire a pluripotent stem cell (iPSC) state through the forced expression of key transcription factors, and in the mouse these cells can fulfil the strictest of all developmental assays for pluripotent cells by generating completely iPSC-derived embryos and mice. However, it is not known whether there are additional classes of pluripotent cells, or what the spectrum of reprogrammed phenotypes encompasses. Here we explore alternative outcomes of somatic reprogramming by fully characterizing reprogrammed cells independent of preconceived definitions of iPSC states. We demonstrate that by maintaining elevated reprogramming factor expression levels, mouse embryonic fibroblasts go through unique epigenetic modifications to arrive at a stable, Nanog-positive, alternative pluripotent state. In doing so, we prove that the pluripotent spectrum can encompass multiple, unique cell states.

Vitamin C modulates TET1 function during somatic cell reprogramming

Article

Full-text available

Oct 2013
Nat Genet

Vitamin C, a micronutrient known for its anti-scurvy activity in humans, promotes the generation of induced pluripotent stem cells (iPSCs) through the activity of histone demethylating dioxygenases. TET hydroxylases are also dioxygenases implicated in active DNA demethylation. Here we report that TET1 either positively or negatively regulates somatic cell reprogramming depending on the absence or presence of vitamin C. TET1 deficiency enhances reprogramming, and its overexpression impairs reprogramming in the context of vitamin C by modulating the obligatory mesenchymal-to-epithelial transition (MET). In the absence of vitamin C, TET1 promotes somatic cell reprogramming independent of MET. Consistently, TET1 regulates 5-hydroxymethylcytosine (5hmC) formation at loci critical for MET in a vitamin C-dependent fashion. Our findings suggest that vitamin C has a vital role in determining the biological outcome of TET1 function at the cellular level. Given its benefit to human health, vitamin C should be investigated further for its role in epigenetic regulation.

Ultrafast and memory-efficient alignment of short DNA sequences to the human genome

Article

Jan 2009

Corrigendum: Hotspots of aberrant epigenomic reprogramming in human induced pluripotent stem cells

Article

Oct 2014
NATURE

Induction of pluripotent stem cells from adult human fibroblasts by defined factors

Article

Jan 2007
CELL

K. Takahashi

Model-based analysis of ChIP-seq (MACS)

Article

Jan 2008

Yuji Zhang

Manipulating the Mouse Embryo: A Laboratory Manual

Article

Jan 1994

Epigenetics of Reprogramming to Induced Pluripotency

Article

Mar 2013

Reprogramming to induced pluripotent stem cells (iPSCs) proceeds in a stepwise manner with reprogramming factor binding, transcription, and chromatin states changing during transitions. Evidence is emerging that epigenetic priming events early in the process may be critical for pluripotency induction later. Chromatin and its regulators are important controllers of reprogramming, and reprogramming factor levels, stoichiometry, and extracellular conditions influence the outcome. The rapid progress in characterizing reprogramming is benefiting applications of iPSCs and is already enabling the rational design of novel reprogramming factor cocktails. However, recent studies have also uncovered an epigenetic instability of the X chromosome in human iPSCs that warrants careful consideration.

Stemformatics: Visualisation and sharing of stem cell gene expression

Article

Dec 2012

Genome-scale technologies are increasingly adopted by the stem cell research community, because of the potential to uncover the molecular events most informative about a stem cell state. These technologies also present enormous challenges around the sharing and visualisation of data derived from different laboratories or under different experimental conditions. Stemformatics is an easy to use, publicly accessible portal that hosts a large collection of exemplar stem cell data. It provides fast visualisation of gene expression across a range of mouse and human datasets, with transparent links back to the original studies. One difficulty in the analysis of stem cell signatures is the paucity of public pathways/gene lists relevant to stem cell or developmental biology. Stemformatics provides a simple mechanism to create, share and analyse gene sets, providing a repository of community-annotated stem cell gene lists that are informative about pathways, lineage commitment, and common technical artefacts. Stemformatics can be accessed at stemformatics.org.

An epigenomic roadmap to induced pluripotency reveals DNA methylation as a reprogramming modulator

Abstract and Figures

Recommended publications

Epigenetic plasticity: A central regulator of epithelial-to-mesenchymal transition in cancer

Global DNA methylation remodeling during direct reprogramming of fibroblasts to neurons

Does transcription factor induced pluripotency accurately mimic embryo derived pluripotency?

Bmi1 Is a Key Epigenetic Barrier to Direct Cardiac Reprogramming