MBNL proteins repress ES-cell-specific alternative splicing and reprogramming

LETTER

doi:10.1038/nature12270

MBNL proteins repress ES-cell-specific alternative splicing and reprogramming Hong Han1,2*, Manuel Irimia1*, P. Joel Ross3, Hoon-Ki Sung4, Babak Alipanahi5, Laurent David6, Azadeh Golipour2,6, Mathieu Gabut1, Iacovos P. Michael4, Emil N. Nachman1,2, Eric Wang7, Dan Trcka6, Tadeo Thompson3, Dave O’Hanlon1, Valentina Slobodeniuc1, Nuno L. Barbosa-Morais1,8, Christopher B. Burge7, Jason Moffat1,2, Brendan J. Frey1,5, Andras Nagy4,9, James Ellis2,3, Jeffrey L. Wrana2,6 & Benjamin J. Blencowe1,2

Previous investigations of the core gene regulatory circuitry that controls the pluripotency of embryonic stem (ES) cells have largely focused on the roles of transcription, chromatin and non-coding RNA regulators1–3. Alternative splicing represents a widely acting mode of gene regulation4–8, yet its role in regulating ES-cell pluripotency and differentiation is poorly understood. Here we identify the muscleblind-like RNA binding proteins, MBNL1 and MBNL2, as conserved and direct negative regulators of a large program of cassette exon alternative splicing events that are differentially regulated between ES cells and other cell types. Knockdown of MBNL proteins in differentiated cells causes switching to an ES-cell-like alternative splicing pattern for approximately half of these events, whereas overexpression of MBNL proteins in ES cells promotes differentiated-cell-like alternative splicing patterns. Among the MBNL-regulated events is an ES-cell-specific alternative splicing switch in the forkhead family transcription factor FOXP1 that controls pluripotency9. Consistent with a central and negative regulatory role for MBNL proteins in pluripotency, their knockdown significantly enhances the expression of key pluripotency genes and the formation of induced pluripotent stem cells during somatic cell reprogramming. A core set of transcription factors that includes OCT4 (also called POU5F1), NANOG and SOX2, together with specific microRNAs and long non-coding RNAs, control the expression of genes required for the establishment and maintenance of ES-cell pluripotency1–3,10–12. Alternative splicing, the process by which splice sites in primary transcripts are differentially selected to produce structurally and functionally distinct messenger RNA and protein isoforms, provides a powerful additional mechanism with which to control cell fate7,8,13, yet its role in the regulation of pluripotency has only recently begun to emerge. In particular, the inclusion of a highly conserved ES-cell-specific ‘switch’ exon in the FOXP1 transcription factor changes its DNA binding specificity such that it stimulates the expression of pluripotency transcription factors, including OCT4 and NANOG, while repressing genes required for differentiation9. However, the trans-acting regulators of this and other alternative splicing events14–16 implicated in ES-cell biology are not known. These factors are important to identify, as they may control regulatory cascades that direct cell fate, and likewise they may also control the efficiency and kinetics of somatic cell reprogramming. To identify such factors, we used high-throughput RNA sequencing (RNA-seq) data to define human and mouse cassette alternative exons that are differentially spliced between ES cells and induced pluripotent stem cells (iPSCs), and diverse differentiated cells and tissues, referred to below as ‘ES-cell-differential alternative splicing’. A splicing code

analysis17 was then performed to identify cis-elements that may promote or repress these exons. The RNA-seq data used to profile alternative splicing were also used to detect human and mouse splicing factor genes that are differentially expressed between ES cells/iPSCs and non-ES cells/tissues. By integrating these data sources, we sought to identify differentially expressed splicing regulators with defined binding sites that match cis-elements predicted by the code analysis to function in ES-cell-differential alternative splicing. We identified 181 human and 103 mouse ES-cell-differential alternative splicing events, with comparable proportions of exons that are $25% more included or more skipped in ES cells versus the other profiled cells and tissues (Fig. 1a, Supplementary Figs 1a and 2, and Supplementary Tables 1 and 2). When comparing orthologous exons in both species, 25 of the human and mouse ES-cell-differential alternative splicing events overlapped (P , 2.2 3 10216; hypergeometric test). The human and mouse ES-cell-differential alternative splicing events are significantly enriched in genes associated with the cytoskeleton (for example, DST, ADD3), plasma membrane (for example, DNM2, ITGA6) and kinase activity (for example, CASK, MARK2 and MAP2K7) (Supplementary Table 3). They also include the aforementioned FOXP1 ES-cell-switch alternative splicing event, and previously unknown alternative splicing events in other transcription or chromatin regulatory factor genes (for example, TEAD1 and MTA1) that have been implicated in controlling pluripotency18,19. These results suggest a considerably more extensive role for regulated alternative splicing in ES-cell biology than previously appreciated. The splicing code analysis revealed that motifs corresponding to consensus binding sites of the conserved MBNL proteins are the most strongly associated with ES-cell-differential alternative splicing in human and mouse. The presence of MBNL motifs in downstream flanking intronic sequences is associated with exon skipping in ES cells, whereas their presence in upstream flanking intronic sequences is associated with exon inclusion in ES cells (Fig. 1b, human code; Supplementary Fig. 1b, mouse code). To a lesser extent, features resembling binding sites for other splicing regulators, including polypyrimidine tract binding protein (PTBP) and RNA-binding fox (RBFOX) proteins, may also be associated with ES-cell-differential alternative splicing. From RNA-seq expression profiling of 221 known or putative splicing factors, 11 genes showed significant differential expression between human ES cells/iPSCs and other cells and tissues (Bonferronicorrected P , 0.05, Wilcoxon rank-sum test) (Fig. 1c and Supplementary Table 4). Notably, MBNL1 and MBNL2 had the lowest relative mRNA levels in ES cells/iPSCs compared to other cells and tissues

1

Banting and Best Department of Medical Research and Donnelly Centre, University of Toronto, Toronto, Ontario M5S 3E1, Canada. 2Department of Molecular Genetics, University of Toronto, Toronto, Ontario M5S 1A8, Canada. 3Developmental and Stem Cell Biology, The Hospital for Sick Children, 101 College Street, Toronto, Ontario M5G 1L7, Canada. 4Center for Stem Cells and Tissue Engineering, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario M5G 1X5, Canada. 5Department of Electrical and Computer Engineering, University of Toronto, Toronto, Ontario M5S 3G4, Canada. 6Center for Systems Biology, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario M5G 1X5, Canada. 7Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02142, USA. 8Instituto de Medicina Molecular, Faculdade de Medicina, Universidade de Lisboa, 1649-028 Lisboa, Portugal. 9Department of Obstetrics and Gynecology, University of Toronto, Toronto, Ontario, M5S 1A8, Canada. *These authors contributed equally to this work. 1 3 J U N E 2 0 1 3 | VO L 4 9 8 | N AT U R E | 2 4 1

©2013 Macmillan Publishers Limited. All rights reserved

RESEARCH LETTER

4

a siControl 1:1 1:2 1:4 1:8

b

FOXP1-ES code features MBNL

Anti-MBNL1

RBFOX

Anti-tubulin

PTBP

Anti-MBNL2

d

FOXP1-ES

l1 +2

6

l2

5

bn

4

siM

3

l1

c

2

siM

1

bn

Anti-tubulin

+ +1 1 +200 +300 0 –3 0 –200 0 –1 0 00 –1

(Fig. 1c, Supplementary Fig. 3a and Methods). Quantitative RT–PCR (polymerase chain reaction with reverse transcription; qRT–PCR) assays confirmed this observation (Supplementary Fig. 3b). Similar results were obtained when analysing mouse expression data (Supplementary Fig. 3c–e and Supplementary Table 4). PTBP, RBFOX and other splicing factors potentially associated with ES-cell-differential alternative splicing by the splicing code analysis did not exhibit significant differences in mRNA levels between ES cells/iPSCs and other cells or tissues. Collectively, these results suggest a conserved and prominent role for MBNL1 and MBNL2 in ES-cell-differential alternative splicing. Because MBNL proteins are expressed at minimal levels in ES cells compared to other cell types, we proposed that they may repress EScell-differential exons in non-ES cells, and/or activate the inclusion of exons in non-ES cells that are skipped in ES cells. Indeed, previous studies have shown that in differentiated cells, MBNL proteins suppress exon inclusion when they bind upstream flanking intronic sequences, and they promote inclusion when binding to downstream flanking intronic sequences20,21. The results of the splicing code analysis are consistent with this mode of regulation, when taking into account that MBNL proteins are depleted in ES cells relative to differentiated cells and tissues (Fig. 1b and Supplementary Fig. 1b). To test the above hypothesis, we used short interfering RNAs (siRNAs) to knock down MBNL1 and MBNL2 (to ,10% of their endogenous levels), individually or together, in human (293T and HeLa) and mouse (neuro2A (N2A)) cells (Fig. 2a and Supplementary Fig. 4a; see below). For comparison, knockdowns were performed in human (H9) and mouse (CGR8) ES cells. RT–PCR assays were used to monitor the ES-cell-switch exon of FOXP1/Foxp1 (human exon 18b/ mouse exon 16b), which is partially included in ES cells and fully skipped in differentiated cell types9. The splicing code analysis suggested that this exon is associated with conserved regulation by MBNL proteins, through possible direct disruption of splice-site recognition (Fig. 2b; see legend and below). Knockdown of MBNL2 in 293T or HeLa cells resulted in a ,1% increase in FOXP1 exon 18b inclusion, whereas

–4 0

higher (bottom) inclusion in ES cells/iPSCs. Dashed lines indicate 300 nucleotide intervals from splice sites. c, Heat map of Z scores of mRNA expression (cRPKM) levels for splicing factors. Twenty-five splicing factors with the lowest or highest relative mRNA expression levels in ES cells/iPSCs compared to non-ES cells/tissues are shown. cRPKM, corrected reads per kilobase exon model per million reads30.

l

Figure 1 | Identification of regulators of ES-cell-differential alternative splicing. a, Heat map of per cent spliced in (PSI) values for 95 representative ES-cell-differential alternative splicing events in transcripts that are widely expressed across human ES cells/iPSCs, non-ES-cell lines and differentiated tissues. b, Splicing code features that are significantly associated with ES-celldifferential alternative splicing. Features are ranked according to Pearson correlation P values (y axis) for alternative exons with either lower (top) or

Z score

bn

ES-cell inclusion

tro

MBNL

C on

10–9

siM

QKI

PTBP RBFOX

si

10–6

–5 0 –4 0 0 –3 0 0 –2 0 0 –1 0 00 –1

RBFOX

SRSF6

2

QKI

10–3

RBM26 HNRNPAB SFPQ TIA1 SART3 RBM27 SNRPD1 TRA2B SNRPF PRPF40A MSI1 ESRP1 RBMX RBM12B KHDRBS1 HNRNPD HNRNPA2B1 FUBP1 SRRM1 SFRS15 RBM12 HNRNPA1 CPSF6 AQR PSIP1

2

P value (correlation)

10–6

0 50 100

MBNL1 MBNL2 CELF6 RBM18 SRSF8 LSMD1 BICC1 RBMS3 UHMK1 SRPK3 ISY1 RBM22 RBMXL2 RBMS2 RBM23 RBM20 RBM24 RBM43 U2AF1L4 PCBP4 MBNL3 LSM10 RBM42 SNRNP35 RBM3

C on siM tro BN l L1 +

MBNL

Differentiated tissues

si

ES-cell exclusion PTBP

Cell lines

tro l BN L siM 1 BN L siM 2 BN L1 +

10–9

iPS

on

b

ES cell

siM

PSI

c

si C

Differentiated tissues

H1-a H1-b H9-a H9-b hES2 iPS Fibroblast HNEK HUVEC MCF7 293T HeLa Adipose Adrenal Brain Cerebellum Breast Colon Muscle Kidney Liver Lung Lymph N. Ovary Prostate Testis Thyroid

Exons with ES-cell-differential alternative splicing

ES cell iPS Cell lines

H1-a H1-b H9-a H9-b hES2 iPS Fibroblast HNEK HUVEC MCF7 293T HeLa Adipose Adrenal Brain Cerebellum Breast Colon Muscle Kidney Liver Lung Lymph N. Ovary Prostate Testis Thyroid

a

0.0

9.5