See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236193316

Corepressor-dependent silencing of fetal hemoglobin expression by BCL11A

Article  in  Proceedings of the National Academy of Sciences · April 2013

DOI: 10.1073/pnas.1303976110 · Source: PubMed

CITATIONS

87READS

160

10 authors, including:

Some of the authors of this publication are also working on these related projects:

Historical (1990-2010) - signal transduction during hematopoiesis View project

Jian Xu

University of Texas Southwestern Medical Center

90 PUBLICATIONS   3,249 CITATIONS   

SEE PROFILE

Daniel E Bauer

Boston Children's Hospital

103 PUBLICATIONS   8,067 CITATIONS   

SEE PROFILE

Marc Kerenyi

Boehringer Ingelheim

66 PUBLICATIONS   1,745 CITATIONS   

SEE PROFILE

Gail Mandel

Howard Hughes Medical Institute

72 PUBLICATIONS   6,367 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Marc Kerenyi on 30 May 2014.

The user has requested enhancement of the downloaded file.

Corepressor-dependent silencing of fetal hemoglobinexpression by BCL11AJian Xua, Daniel E. Bauera, Marc A. Kerenyia, Thuy D. Voa, Serena Houa, Yu-Jung Hsua, Huilan Yaob,Jennifer J. Trowbridgea, Gail Mandelb, and Stuart H. Orkina,c,1

aDivision of Hematology/Oncology, Boston Children’s Hospital and Department of Pediatric Oncology, Dana-Farber Cancer Institute, Harvard Stem CellInstitute, Harvard Medical School, Boston, MA 02115; bVollum Institute, Howard Hughes Medical Institute, Oregon Health and Science University, Portland,OR 97239; and cHoward Hughes Medical Institute, Boston, MA 02115

Contributed by Stuart H. Orkin, March 4, 2013 (sent for review February 11, 2013)

Reactivation of fetal hemoglobin (HbF) in adults ameliorates theseverity of the common β-globin disorders. The transcription factorBCL11A is a critical modulator of hemoglobin switching and HbFsilencing, yet the molecular mechanism through which BCL11Acoordinates the developmental switch is incompletely understood.Particularly, the identities of BCL11A cooperating protein complexesand their roles in HbF expression and erythroid development re-main largely unknown. Here we determine the interacting partnerproteins of BCL11A in erythroid cells by a proteomic screen. BCL11Ais found within multiprotein complexes consisting of erythroid tran-scription factors, transcriptional corepressors, and chromatin-modifying enzymes. We show that the lysine-specific demethylase1 and repressor element-1 silencing transcription factor corepressor 1(LSD1/CoREST) histone demethylase complex interacts with BCL11Aand is required for full developmental silencing of mouse embryonicβ-like globin genes and human γ-globin genes in adult erythroidcells in vivo. In addition, LSD1 is essential for normal erythroiddevelopment. Furthermore, the DNAmethyltransferase 1 (DNMT1)is identified as a BCL11A-associated protein in the proteomic screen.DNMT1 is required to maintain HbF silencing in primary human adulterythroid cells. DNMT1 haploinsufficiency combined with BCL11Adeficiency further enhances γ-globin expression in adult animals.Our findings provide important insights into the mechanistic rolesof BCL11A in HbF silencing and clues for therapeutic targeting ofBCL11A in β-hemoglobinopathies.

gene regulation | globin switching | hematopoiesis

Fetal hemoglobin (HbF, α2γ2) is a major genetic modifier of thephenotypic heterogeneity in patients with the major β-globin

disorders sickle cell disease (SCD) and β-thalassemia (1). Thesynthesis of hemoglobin undergoes switching during ontogenysuch that HbF is the predominant hemoglobin produced duringfetal life and is gradually silenced and replaced by adult hemo-globin (HbA, α2β2) around birth. Because increased γ-globin ex-pression in adults can substitute for the mutant or absent β-globinin SCD and β-thalassemia, respectively, the fetal-to-adult globinswitch is critical to the pathogenesis of these conditions. As a re-sult, intense efforts have been aimed to elucidate the molecularmechanisms of HbF silencing and to develop target-based ther-apeutics to enhance HbF production (2).Recent genomewide association studies (GWAS) led to the

identification of a new HbF-associated locus on chromosome 2,located within the gene BCL11A (3). Subsequent studies demon-strated that BCL11A, a zinc-finger transcription factor, is a bonafide repressor of HbF expression (4–7). BCL11A protein is deve-lopmentally regulated and is required to maintain HbF silencing inhuman adult erythroid cells (4, 5). KO of BCL11A in mice carryinga human β-globin cluster transgene leads to profound delay inglobin switching and impaired HbF silencing in adult erythroidcells (5, 8). Previously silenced γ-globin genes can also be reac-tivated by loss of BCL11A in adult animals (8). Most importantly,inactivation of BCL11A alone is sufficient to ameliorate the he-matologic and pathologic defects associated with SCD throughhigh-level HbF induction in humanized mouse models (8). Thesegenetic studies provide persuasive evidence that BCL11A functions

as a central modulator of HbF expression and globin switchingin vivo. Further molecular studies revealed that BCL11A inter-acts with several erythroid regulators including GATA1, FOG1,and SOX6, and with the nucleosome remodeling and deace-tylase complex (NuRD) (4, 6). BCL11A acts within the β-globincluster by associating with several discrete regions, includingsequences specifically deleted in patients with hereditary per-sistence of fetal hemoglobin (HPFH). However, BCL11A doesnot bind detectably to the γ-globin promoter in erythroid chro-matin, suggesting that its mode of action is more complex thansimply blocking transcription at the proximal promoter. Thisobservation is consistent with a role of BCL11A in promotinglong-range chromosomal interactions within the β-globin locus(6, 7, 9). Therefore, in an effort to elucidate further the precisemolecular mechanisms by which BCL11A coordinates the switch,it is relevant to identify systematically BCL11A-interacting partnerproteins, and use functional and genetic approaches to assess theirindividual roles in HbF regulation and erythroid cell maturation. Inaddition to providing important insights into the molecular mech-anisms through which BCL11A controls hemoglobin switching,our data suggest opportunities, as well as challenges, for therapeuticinduction of HbF in patients with the major hematologic disorders.

ResultsBCL11A-Interacting Partner Proteins in Erythroid Cells. We charac-terized BCL11A-containing multiprotein complexes by a proteomicaffinity screen using a metabolic biotin tagging approach (10). Mu-rine erythroleukemia (MEL) cell lines that stably express the bac-terial biotin ligaseBirA and the recombinant BCL11A (XL isoform)containing a FLAG epitope tag together with a BirA recognitionmotif fused to its amino terminus were generated (FB-BCL11A; Fig.1A). Clones were selected that express the recombinant BCL11Aat levels similar to endogenous BCL11A (Fig. S1).Single streptavidin affinity purification or tandem anti-FLAG

immunoaffinity and streptavidin affinity purification was performedon nuclear extracts from MEL cells stably expressing BirA aloneor BirA together with FB-BCL11A. Copurified proteins werefractionated in SDS/PAGE, followed by liquid chromatographyand tandem MS peptide sequencing (Fig. 1A). We identified nu-merous peptides of BCL11A and nearly all previously establishedinteracting proteins, including FOG1, the entire NuRD complex,and the nuclear matrix protein MATRIN-3 (4). Besides previouslyrecognized partners, we also identified a panel of previously un-discovered interacting complexes, including the hematopoietic reg-ulators RUNX1 and IKZF1 (or IKAROS), several transcriptionalcorepressor complexes, and the SWI/SNF chromatin remodelingcomplex (Fig. 1B). The BCL11A-interacting corepressors include

Author contributions: J.X. and S.H.O. designed research; J.X., D.E.B., M.A.K., T.D.V., S.H.,Y.-J.H., and J.J.T. performed research; H.Y. and G.M. contributed new reagents/analytictools; J.X. and S.H.O. analyzed data; and J.X. and S.H.O. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. E-mail: stuart_orkin@dfci.harvard.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1303976110/-/DCSupplemental.

6518–6523 | PNAS | April 16, 2013 | vol. 110 | no. 16 www.pnas.org/cgi/doi/10.1073/pnas.1303976110

the histone demethylase complex lysine-specific demethylase 1(LSD1)/CoREST, the nuclear receptor corepressors NCoR/SMRT,the SIN3 deacetylase complex, the BCL6 corepressor BCOR,the DNA methyltransferase 1 (DNMT1), and histone demethylases(Fig. 1B).To identify BCL11A-interacting proteins in human erythroid

cells, we performed a similar proteomic screen in human eryth-roleukemia (K562) cells. Remarkably, the identified partner pro-teins are largely similar to those found in MEL cells (Fig. 1B),suggesting that BCL11A acts within functionally conserved pro-tein complexes. By coimmunoprecipitation (co-IP) analysis, weconfirmed that both the epitope-tagged and the endogenousBCL11A associate with the identified partner proteins in erythroidcell lines and primary human adult erythroid cells, respectively(Fig. S1; Fig. 1C).

Functional RNAi Screen in Primary Human Erythroid Cells. To assessthe functional roles of BCL11A-interacting proteins in HbF reg-ulation, we performed a lentiviral RNAi screen in primary adulthuman CD34+ hematopoietic stem/progenitor cells (HSPCs) un-dergoing ex vivo erythroid maturation (Fig. 2A). Primary CD34+HSPCs were transduced with lentiviruses containing shRNAs

targeting BCL11A or its interacting proteins during an expansionphase, followed by erythroid differentiation for 5–7 d. Knockdownof each gene was confirmed by qRT-PCR, and shRNAs that re-sulted in >75% decrease of target mRNA expression were selectedfor subsequent analyses (Table S1; Figs. S2 and S3). KLF1 (orEKLF) is a critical erythroid regulator required for adult β-globintranscription (11). Recently it was shown that KLF1 occupies theBCL11A promoter and activates its expression (12, 13). Consistentwith previous analysis, knockdown of BCL11A or KLF1 resultsin marked increase in γ-globin expression and serves as positivecontrols for the RNAi screen. We first ranked the genes accordingto the level of human γ-globin expression (Fig. 2B). Notably, thetop of the list consists of subunits of several transcriptional co-repressor complexes, including the NuRD complex (CHD4,HDAC1, HDAC2, and MBD2), DNMT1, SIN3A, NCOR1, andthe polycomb repressive complex 2 (PRC2; EZH2, EED, andSUZ12). shRNA-mediated depletion of BCL11A-interacting pro-teins also results in derepression of the human embryonic e-gene.We next ranked the genes according to the level of e-globin ex-pression (Fig. 2C). The highest ranked genes include subunits ofthe PRC2 complex (EED, EZH1, EZH2, and SUZ12), severalhistone demethylases (KDM7A, KDM3A, and KDM5D) andSIN3A. Importantly, depletion of KLF1, but not BCL11A, signif-icantly increases e-globin expression (Fig. 2C), indicating a distinctrole of KLF1 in regulating human embryonic globin expression.These results demonstrate that BCL11A-interacting proteins aredifferentially required for silencing of human embryonic or fetalglobin expression, suggesting that they may form distinct sub-complexes depending on the chromatin context.

BirA + FB-BCL11ACells Expressing

α-FLAG

Streptavidin

BCL11A

Nuclear Extract

FLAG Peptide Elution

BCL11A

SDS-PAGE Fractionation

LC-MS/MS

10%

Inpu

tM

ouse

IgG

α-BC

L11A

WBα-BCL11A

α-GATA1

α-LSD1

α-CoREST

α-SMRT

α-SIN3A

α-Mi-2β

α-HDAC1

α-NCOR1

α-HDAC2

IP

A

C

aSingle streptavidin affinity purificationbTandem α-FLAG and streptavidin affinity purification

BMEL K562

BCL11A 13a, 11a, 12b 17a, 14a, 11a, 12b

Transcription factors FOG1 (ZFPM1) 1, 2, 0 0, 0, 0, 0 RUNX1 1, 4, 0 0, 0, 0, 1 IKAROS (IKZF1) 1, 3, 0 0, 0, 0, 0 NuRD complex Mi-2β (CHD4) 8, 8, 1 21, 10, 21, 10 Mi-2α (CHD3) 0, 0, 0 4, 4, 9, 2 MTA1 0, 0, 0 6, 5, 2, 1 MTA2 0, 0, 1 23, 16, 13, 2 MTA3 0, 0, 0 2, 2, 0, 0 HDAC1 7, 7, 5 19, 16, 8, 4 HDAC2 5, 4, 2 12, 15, 6, 2 RBBP4 3, 3, 1 4, 5, 5, 1 RBBP7 2, 1, 1 7, 6, 6, 1 MBD3 0, 0, 0 7, 6, 0, 0 P66α (GATAD2A) 0, 0, 0 29, 10, 4, 1 P66β (GATAD2B) 2, 1, 0 12, 12, 5, 2 LSD1/CoREST complex LSD1 (KDM1A) 15, 27, 8 1, 5, 12, 5 CoREST (RCOR1) 8, 9, 3 8, 8, 3, 1 NCoR/SMRT complex NCoR (NCOR1) 6, 14, 3 0, 0, 2, 0 SMRT (NCOR2) 13, 16, 2 0, 0, 8, 0 TBLR1 0, 2, 0 4, 0, 2, 0 TBL1 0, 0, 0 3, 0, 0, 0 CORO2A 2, 1, 0 0, 0, 0, 0 KAISO (ZBTB33) 7, 7, 6 26, 21, 18, 8 SIN3 complex SIN3A 5, 1, 1 6, 4, 4, 2 SIN3B 0, 0, 0 1, 2, 1, 0 Other corepressors BCOR 17, 31, 6 1, 0, 6, 2 TRIM28 11, 12, 5 2, 11, 13, 4 SWI/SNF complex SNF5 (SMARCB1) 0, 1, 0 2, 3, 3, 1 BRG1 (SMARCA4) 7, 9, 2 8, 1, 12, 2 BRM (SMARCA2) 0, 0, 0 7, 1, 11, 0 BAF57 (SMARCE1) 2, 1, 0 5, 2, 3, 0 BAF60A (SMARCD1) 0, 0, 0 10, 3, 3, 0 BAF60B (SMARCD2) 0, 0, 0 7, 3, 5, 0 BAF155 (SMARCC1) 10, 6, 0 16, 5, 10, 2 BAF170 (SMARCC2) 3, 4, 0 15, 3, 10, 0 SNF2H (SMARCA5) 2, 5, 0 12, 2, 12, 0 BAF180 (PB1) 0, 0, 0 5, 1, 15, 0 ASH2L 0, 0, 0 2, 1, 5, 0 Other nuclear factors DNMT1 14, 22, 3 9, 1, 3, 1 JMJD1A (KDM3A) 1, 1, 0 0, 0, 0, 0 JMJD1B (KDM3B) 0, 1, 0 0, 0, 0, 0 JARID1A (KDM5A) 0, 1, 0 0, 0, 0, 0 CARM1 1, 1, 0 0, 0, 0, 0 YLPM1 16, 16, 0 0, 2, 9, 0 MSH2 3, 6, 0 0, 2, 3, 0 Nuclear matrix MATRIN-3 (MATR3) 15, 8, 5 3, 6, 13, 5

Protein Identity # of Peptides

Fig. 1. Identification of BCL11A-interacting proteins in erythroid cells bya proteomic affinity screen. (A) Schematic diagram of experimental design.(B) List of identified proteins is shown with the number of peptides obtainedin each trial. Official gene symbols of the identified proteins are shown inparentheses. (C) Confirmation of interactions between BCL11A and identifiedproteins by coimmunoprecipitation in primary human erythroid cells.

ProEsCD34+HSPCs

KnockdownHbF ExpressionDifferentiation

A

shG

FPB

CL1

1AK

LF1

CH

D4

DN

MT1

SIN

3AH

DA

C1

NC

oR1

EZH

2H

DA

C2

MB

D2

EE

DK

DM

3AS

UZ1

2K

DM

5BR

BB

P4

KD

M5A

EZH

1IK

ZF1

KD

M5D

KD

M4B

KD

M7A

RB

BP

7K

DM

4CB

AF1

55LS

D1

ZBTB

33N

CoR

2K

DM

3BK

DM

4AB

RG

1B

CoR

TRIM

28

B 100

80

60

40

20

0

Hum

an γ

-glo

bin

expr

essi

on (%

)

shG

FPB

CL1

1AK

LF1

EE

DE

ZH1

KD

M7A

KD

M3A

SU

Z12

KD

M5D

SIN

3AE

ZH2

HD

AC

1K

DM

3BH

DA

C2

KD

M5A

CH

D4

DN

MT1

KD

M5B

NC

oR1

KD

M4A

IKZF

1LS

D1

MB

D2

RB

BP

7R

BB

P4

BA

F155

ZBTB

33K

DM

4BTR

IM28

BR

G1

KD

M4C

NC

oR2

BC

oR

C 1.6

1.2

0.8

0.4

0

Hum

an ε

-glo

bin

expr

essi

on (%

)

D 100

80

60

40

20

0

Tota

l Hum

an β

-glo

bin

(rel

ativ

e to

GA

PD

H)

SCF, IL-3, IL-6, Flt-3 ligand

Expansion

Epo, SCF, IL-3,Dexamethasone,

β-estradiol

Differentiation

SO

X6

SO

X6

shG

FPB

CL1

1AK

LF1

CH

D4

DN

MT1

SIN

3AH

DA

C1

NC

oR1

EZH

2H

DA

C2

MB

D2

EE

DK

DM

3AS

UZ1

2K

DM

5BR

BB

P4

KD

M5A

EZH

1IK

ZF1

KD

M5D

KD

M4B

KD

M7A

RB

BP

7K

DM

4CB

AF1

55LS

D1

ZBTB

33N

CoR

2K

DM

3BK

DM

4AB

RG

1B

CoR

TRIM

28

SO

X6

Fig. 2. Functional RNAi screen of BCL11A-interacting proteins in primaryhuman erythroid cells. (A) Schematic diagram of experimental design. (B)Expression of human γ-globin mRNA was measured by qRT-PCR. Data areplotted as percentage of γ-globin over total β-like human globin gene levels(two to five shRNAs per gene). Genes are ranked based on the level ofγ-globin expression (highest to lowest). (C) Genes are ranked based on thelevel of e-globin expression (highest to lowest). (D) Expression of humantotal β-like globin mRNA normalized to GAPDH mRNA level. Results arethe means ± SD of at least three independent experiments.

Xu et al. PNAS | April 16, 2013 | vol. 110 | no. 16 | 6519

MED

ICALSC

IENCE

S

Besides the reactivation of HbF expression, depletion of sev-eral BCL11A-interacting proteins impairs erythroid cell matu-ration ex vivo, as revealed by reduced expression of erythroid cellsurface markers CD71 and CD235a (Fig. S3). Perturbation of totalglobin synthesis is a secondary consequence of defective redcell differentiation. Notably, depletion of BCL11A-interactingproteins results in variable effects on the synthesis of total humanβ-like globin mRNAs (Fig. 2D). Although knockdown of manygenes has little effect, knockdown of several genes leads to a greaterthan twofold increase (NCOR1 and KDM5D) or decrease (KLF1,KDM5B, SOX6, RBBP7, LSD1/KDM1A, ZBTB33, and KDM4A)in expression of total β-like globin mRNAs (Fig. 2D; Fig. S3),suggesting that their gene products influence processes requiredfor normal erythroid development.

Chromatin Occupancy of BCL11A-Interacting Proteins. The identifi-cation of BCL11A-interacting proteins and subsequent functionalRNAi screen led to prioritization of candidate HbF regulators fordetailed mechanistic studies. If BCL11A and its interacting part-ners regulate HbF expression in a functional protein complex,ablation of one critical component of the complex may lead todestabilization of the larger complex and/or its chromatin oc-cupancy. We next determined whether the chromatin occupancyof BCL11A-interacting proteins is dependent on the presence ofBCL11A. ChIP was performed in primary human adult erythroidcells transduced with control or BCL11A lentiviral shRNA. Incontrol cells, ChIP analysis revealed that BCL11A occupies sev-eral discrete regions within the human β-globin cluster, includingthe DNase I hypersensitive site 3 (HS3) within the locus controlregion (LCR), the e-globin promoter, and the γδ-intergenic region,consistent with previous analyses (4, 6). Knockdown of BCL11Aabolishes its occupancy at the above regions, consistent with thenear absence of BCL11A protein in the knockdown cells (Fig.3B). Interestingly, loss of BCL11A leads to a marked increase ofH3K4me2, a histone mark associated with active transcription, atthe Aγ-globin promoter and the γδ-intergenic region, with a con-comitant decrease of the H3K9me1 mark (Fig. 3 C and D). Be-cause both H3K4me2 and H3K9me1 are substrates for the histone

demethylase LSD1, which is identified as a BCL11A-interactingprotein (Fig. 1), we next examined the chromatin occupancy ofLSD1 and its cofactor CoREST. Loss of BCL11A also impairschromatin occupancy of the LSD1/CoREST complex within theβ-globin locus, particularly at the e-promoter and the γδ-intergenicregion (Fig. 3 E and F), suggesting that BCL11A is required forefficient recruitment or stable association of the LSD1/CoRESTcomplex within the β-globin cluster. Similarly, the binding of theNuRD complex (Mi-2β and HDAC1) and EZH2 is also decreasedon BCL11A knockdown (Fig. 3 G–I). Notably, RNA Pol II occu-pancy at the Aγ promoter is enhanced on knockdown, consistentwith transcriptional reactivation of the γ-globin genes in theBCL11A-depleted adult erythroid cells (Figs. 2B and 3J). Theseresults indicate that loss of BCL11A leads to impaired chromatinoccupancy of a subset of its interacting corepressors in humanadult erythroid cells.

Loss of LSD1/CoREST Reactivates Mouse Embryonic β-Like Globin Genes.We next determined whether the LSD1/CoREST complex is re-quired for the expression of β-like globin genes in vivo. Mousestrains carrying floxed alleles of Lsd1 or Rocr1 (encoding CoREST)were generated by gene targeting. We first examined the expressionof mouse embryonic β-like globin genes (ey- and βh1-globin) infetal livers of Lsd1- or Rcor1-deficient embryos. Erythroid-specificdeletion of LSD1 by EpoR-Cre results in a 6.2- and 26.7-foldincrease in expression of the ey- and βh1-globin mRNAs in em-bryonic day 12.5 (E12.5) fetal liver definitive erythroid cells, re-spectively (Fig. 4A; P < 0.05). Similarly, loss of CoREST results ina 13.0- and 15.4-fold increase in ey- and βh1-globin mRNAsin E13.5 fetal livers (Fig. 4B; P < 0.05). Because germ-line orerythroid-specific deletion of LSD1 is embryonic lethal, we studiedLSD1 conditional KO (cKO) adult mice using the hematopoietic-selective and IFN inducible Mx1-Cre allele (14). Acute deletionof LSD1 in adult bone marrow also results in a modest increase(5.8- and 10.6-fold) in expression of ey- and βh1-globin genes(Fig. 4D; P < 0.05). These results provide genetic evidence thatthe LSD1/CoREST complex is required to maintain full silencingof mouse embryonic β-like globin genes in definitive erythroid cells

0

0.04

0.08

0.12

0.16

0.2

% o

f Inp

ut

Control shBCL11A

LCR 3’HS1Embryonic Fetal Adult

0

0.5

1

1.5

2

2.5

3

% o

f Inp

ut

0

1

2

3

4

5

6

% o

f Inp

ut

0

0.03

0.06

0.09

0.12

0.15

% o

f Inp

ut

0

0.03

0.06

0.09

0.12

0.15

% o

f Inp

ut

0

0.05

0.1

0.15

0.2

0.25

% o

f Inp

ut

0.02

0.04

0.06

0.08

% o

f Inp

ut

0

0.3

0.6

0.9

1.2

% o

f Inp

ut

0

0.1

0.2

0.3

0.4

% o

f Inp

ut

0

0.02

0.04

0.06

0.08

0.1

% o

f Inp

ut

a b c d e f g h

a b c d e f g h

BCL11A

a b c d e f g h

IgG

a b c d e f g h

H3K4me2

a b c d e f g h

H3K9me1

a b c d e f g h

LSD1

a b c d e f g h

CoREST Mi-2β

0

HDAC1

EZH2 Pol II

a b c d e f g h a b c d e f g h

a b c d e f g h a b c d e f g h

A B

K

C E

F

D

G

I J

H

ε γG γA δ β ∗

∗

∗

∗

∗∗

∗

∗

∗ ∗

∗∗

∗∗

∗∗

∗

∗

∗

∗∗

∗

∗

∗

∗

Fig. 3. Chromatin occupancy of BCL11A-interactingproteins in primary human erythroid cells. (A) In vivochromatin occupancy in primary human erythroidcells transduced with lentiviral shRNAs against GFP(Control) or BCL11A (shBCL11A). Primers were de-signed to amplify discrete regions across the β-globinlocus, including (a) HS3 within the LCR, (b) e-globinpromoter, (c) Aγ-globin promoter, (d) +3-kb regiondownstream of Aγ-globin gene, (e) −1-kb region up-stream of δ-globin gene, (f) δ-globin promoter, (g)β-globin promoter, and (h) 3′HS1 site. (B–K ) ChIP-qPCR analysis of BCL11A, H3K4me2, H3K9me1, LSD1,CoREST, Mi-2β, HDAC1, EZH2, RNA Pol II, and rabbitIgG in primary human erythroid cells in the pres-ence or absence of BCL11A knockdown, respectively.Results are means ± SD; *P < 0.05.

6520 | www.pnas.org/cgi/doi/10.1073/pnas.1303976110 Xu et al.

in vivo. However, deletion of BCL11A in mouse results in muchgreater increases (31.2- to 232.9-fold) of embryonic globin genes indefinitive fetal liver and bone marrow erythroid cells (Fig. 4 C andE) (8), indicating that the extent of induction by LSD1 deletionis quite modest compared with that of BCL11A deletion (Fig. 4;compare A and B with C or D with E).

LSD1 Cooperates with BCL11A in Silencing γ-Globin Expression. Be-cause the repression of the endogenous ey- and βh1-globin genesis modestly and severely impaired in definitive fetal liver andbone marrow erythrocytes of Lsd1-deficient and Bcl11a-deficientmice, respectively (Fig. 4) (5), and LSD1 physically interacts withBCL11A (Fig. 1), we next examined whether LSD1 and BCL11Afunction collaboratively in silencing human γ-globin genes in vivo.We generated BCL11A and LSD1 compound KO mice carryingthe human β-locus transgene [β-yeast artificial chromosome(β-YAC)]. Because erythroid-specific loss of LSD1 by EpoR-Creresults in embryonic lethality, we obtained compound BCL11Ahomozygous LSD1 heterozygous (Bcl11a−/−::Lsd1+/−) cKO adultmice. Erythroid-specific loss of BCL11A alone leads to 383-foldincrease in γ-globin mRNA in the peripheral blood (from 0.022%in Bcl11a+/+::Lsd1+/+ control mice to 8.40% in Bcl11a−/−::Lsd1+/+

mice). The expression of γ-globin is further induced to 13.53% inBcl11a−/−::Lsd1+/− compound KO adult mice (Fig. 5A; P = 0.0006).To determine the effect of complete loss of BCL11A and LSD1 onγ-globin silencing, we generated compound homozygous knock-out adult mice using the inducible Mx1-Cre allele. Loss of LSD1alone leads to a slight increase of γ-globin mRNA in the periph-eral blood (from 0.025% in Bcl11a+/+::Lsd1+/+ mice to 0.04% inBcl11a+/+::Lsd1−/− mice; Fig. 5B). Loss of BCL11A alone reac-tivates γ-globin mRNA to 12.88%. Combined loss of BCL11A andLSD1 further induces γ-globin mRNA to 18.19% in Bcl11a−/−::Lsd1−/− compound KO adult mice (Fig. 5B; P = 0.03). These datademonstrate that depletion of LSD1 by itself has a modest effecton transgenic γ-globin expression in vivo. However, loss of LSD1further enhances the effect of BCL11A deficiency in HbF de-repression, suggesting that BCL11A serves as a major barrier toinduction by LSD1 deficiency.

Depletion of LSD1 Induces HbF Expression and Impairs Erythropoiesis.We initially observed that erythroid-specific loss of LSD1 resultedin embryonic lethality at E13.5 with severe anemia. We next ex-amined hematologic parameters in LSD1 and BCL11A compoundKO adult mice (by EpoR-Cre). Consistent with previous analysis(8), erythroid- or hematopoietic-specific loss of BCL11A has littleeffect on total red blood cell (RBC) number and hemoglobin(Hgb) level (Fig. S4). In contrast, the Bcl11a−/−::Lsd1+/− compoundKO mice are mildly anemic (Fig. S4A), indicating that LSD1 isrequired for normal erythroid development. Hematopoietic-selective loss of LSD1 also leads to profound decrease in totalnumber of white blood cells (WBCs; Fig. S4B), consistent with arole for LSD1 in terminal differentiation of other hematopoieticlineages (15).To examine more directly the role of LSD1 in HbF silencing

and erythropoiesis, we depleted its expression in primary humanadult erythroid progenitors. Transduction of two independentlentiviral shRNAs against LSD1 leads to efficient knockdown ofLSD1 protein in CD34+ HSPC-derived erythroid progenitors (Fig.5C). Although knockdown of BCL11A expression leads to a sub-stantial increase of human γ-globin mRNA (from 9.5% to 50.3%at day 7 of erythroid differentiation), depletion of LSD1 expressionresults in modest increases in γ-globin expression (sh1: 21.3%; sh5:32.3%; Fig. 5D). Of note, depletion of LSD1 also results in ∼65%decrease in total β-globin mRNAs in day 7 erythroid progenitors;therefore, much of the observed increase in relative γ-globin ex-pression is due to reduced β-globin expression rather than γ-globininduction per se (Fig. 5D, Lower). This finding is in contrast to thechanges in relative globin expression on BCL11A knockdown,which are predominantly characterized by induction of γ-globinitself. Consistent with decreased total globin production, LSD1-depleted cells retain expression of CD34 antigen and fail to ex-press maturing erythroid cell-specific markers CD71 and CD235a(Fig. S5A). In fact, the majority of LSD1-depleted cells exhibitproerythroblast and basophilic erythroblast morphology, whereasthe more mature polychromatophilic and orthochromatic eryth-roblasts are nearly absent at day 9 of differentiation (Fig. S5B).Collectively, these data indicate that LSD1 is required for bothfull HbF silencing and erythroid cell maturation.Several small molecule inhibitors have been used to target LSD1

by inhibiting its histone demethylase activity (16). We next testedtwo LSD1 inhibitors, pargyline and trans-2-phenylcyclopropylamine[tranylcypromine (TCP)], in reactivation of HbF expression in pri-mary human adult erythroid cells. Pargyline treatment by itself haslittle effect on HbF expression. Combining pargyline treatment andBCL11A knockdown results in modest increases in HbF expres-sion compared with BCL11A knockdown alone (Fig. S6A). Incontrast, TCP treatment induces HbF expression in both controland BCL11A knockdown cells (Fig. S6A). However, TCP treat-ment results in a profound decrease in production of total β-likeglobin mRNAs (Fig. S6B). Additionally, TCP-treated cells fail toexpress CD71 and CD235a (Fig. S6D), indicating that the inductionof HbF expression mediated by chemical inhibitors of LSD1 isalso associated with impaired erythroid maturation.

DNMT1 Is Required for Maintenance of HbF Silencing. The proteomicscreen identified the DNA methyltransferase DNMT1 as anotherprotein with intrinsic enzymatic activity in association with BCL11A(Fig. 1). DNA methylation plays an important role in globin geneexpression, and DNA demethylating agents have been shown toinduce HbF expression in various model systems and patients(8, 17, 18). However, the role of the methyltransferase DNMT1in γ-globin silencing in an intact animal has not been previouslyassessed by formal genetic experiments. Thus, we determinedwhether DNMT1 is directly involved in HbF silencing in vitroand in vivo. Upon shRNA-mediated knockdown of DNMT1expression in primary human adult erythroid cells, HbF expres-sion is significantly induced, whereas the amount of total β-likeglobin mRNAs is modestly reduced (Figs. 2B and 5E). Similarly,inhibition of DNMT1 activity by chemical inhibitors leads to en-hanced HbF expression in primary erythroid cells (Fig. S7).

A B C

D

0.1

1

10

100R

elat

ive

mR

NA

(%)

εy βh1

Lsd1+/+ Lsd1-/-

εy βh1 εy βh1

0.00001

Rel

ativ

e m

RN

A (%

)

εy βh1

0.0001

0.001

0.01

0.1 Mx1-Cre-Mx1-Cre+

Lsd1 cKO

0.00001

Rel

ativ

e m

RN

A (%

)

εy βh1

0.0001

0.001

0.01

0.1 Mx1-Cre-Mx1-Cre+

Bcl11a cKOE

6.2

26.713.0

15.4

31.255.8

5.3

10.6

50.9

232.9

Rcor1+/+ Rcor1-/-

0.1

1

10

100

Rel

ativ

e m

RN

A (%

)

Bcl11a+/+ Bcl11a-/-

0.1

1

10

100

Rel

ativ

e m

RN

A (%

)

Fig. 4. The LSD1/CoREST complex is required for silencing of mouse em-bryonic β-like globin genes. (A) Expression of mouse embryonic (ey and βh1)globin genes was monitored by qRT-PCR in E12.5 fetal livers of control (Lsd1+/+)and LSD1 KO (Lsd1−/−; by EpoR-Cre) embryos. All data are shown as percentage(%) of total mouse β-like globin expression. The fold changes of globin mRNAsare indicated. Results are means ± SD (n ≥ 5 per genotype; P < 0.05). (B) Ex-pression of ey- and βh1-globin genes in E13.5 fetal livers of control and CoRESTKO (Rcor1−/−) embryos (n ≥ 3 per genotype; P < 0.05). (C) Expression of ey- andβh1-globin genes in E12.5 fetal livers of control and BCL11A KO (Bcl11a−/−; byEpoR-Cre) embryos (n ≥ 3 per genotype; P < 0.01). (D) Expression of ey- and βh1-globin genes in PB of control and LSD1;Mx1-Cre cKO mice (3 wk post-pIpC,9–15 wk old; n ≥ 5 per genotype; P < 0.05). (E) Expression of ey- and βh1-globingenes in PB of control and BCL11A;Mx1-Cre cKO mice (3 wk post-pIpC, 9–15 wkold; n ≥ 3 per genotype; P < 0.01).

Xu et al. PNAS | April 16, 2013 | vol. 110 | no. 16 | 6521

MED

ICALSC

IENCE

S

To assess the role of DNMT1 in HbF silencing in vivo, wegenerated DNMT1 and BCL11A compound erythroid-specificcKO β-YAC mice (by EpoR-Cre). Although homozygous lossof DNMT1 results in embryonic lethality, compound BCL11Ahomozygous DNMT1 heterozygous (Bcl11a−/−::Dnmt1+/−) KOmice are viable despite a mild anemic phenotype (Fig. S8). Ofnote, DNMT1 haploinsufficiency leads to a 2.1-fold increase ofγ-globin mRNA in the absence of BCL11A (Fig. 5F; from 11.16%in Bcl11a−/−::Dnmt1+/+ mice to 23.58% in Bcl11a−/−::Dnmt1+/−

mice, P = 1.9 × 10−6). Collectively, these data provide importantgenetic evidence that DNMT1 haploinsufficiency in combinationwith BCL11A deficiency can further enhance HbF expression.

DiscussionWe characterized interacting partner proteins of BCL11A in pri-mary human erythroid cells that may participate in HbF regulation.We infer that BCL11A acts within multiprotein complexes con-sisting of transcriptional corepressors and chromatin-modifyingsubunits. Knockdown of several BCL11A-interacting corepressorproteins induces HbF expression in primary human erythroidcells, whereby the effect on total hemoglobin production anderythroid maturation is variable. The chromatin occupancy of sev-eral BCL11A-interacting complexes, including the LSD1/CoRESTand NuRD complexes, within the β-globin cluster is dependent onthe presence of BCL11A in human erythroid cells. These resultssuggest that BCL11A coordinates the hemoglobin switch and HbFsilencing by assembling transcriptional corepressor complexeswithin the β-globin cluster.The identification of BCL11A-interacting proteins provides

clues to target BCL11A and/or its cofactors for therapeutic HbFinduction in β-hemoglobinopathies. To define favorable moleculartargets for HbF induction, it is imperative to evaluate severalcriteria, including quantitative effects on HbF expression, limitedeffects on expression of nonglobin genes and on erythroid mat-uration, minimal impact outside the erythroid lineage, and fea-sibility of therapeutic intervention (2). Several favorable features,such as potency in HbF silencing, dose dependence, and minimalinfluence on erythropoiesis, recommend BCL11A as a target.However, its roles in other cell lineages, including B lymphocytesand the central nervous system, may present challenges (19, 20).As transcription factors have traditionally been viewed as undrug-gable due to the lack of catalytic domains, alternative strategiesinvolving interference with protein–protein interactions or target-ing partner proteins with enzymatic activities may be considered.Our results show that BCL11A interacts with several chromatin-modifying enzymes, such as the LSD1/CoREST demethylase com-plex, DNMT1, and the NuRD complex. Small molecule inhibitorsfor these enzymes are in various phases of clinical development,

including those in current medical practice. In principle, targetingenzymatic partner proteins of BCL11A constitutes a strategy forindirectly targeting BCL11A function.Our in-depth characterization of the roles of two BCL11A-

interacting enzymatic partners, LSD1 and DNMT1, in HbF reg-ulation provides additional insights. KO of LSD1 in mice leads toderepression of mouse embryonic β-like globin genes and trans-genic human γ-globin genes in definitive erythroid cells. Interfer-ence with LSD1 by shRNA or small molecule inhibitors reactivatesHbF expression in primary human adult erythroid cells. Consistentwith a role of LSD1 in HbF silencing, while this paper was underreview, it was reported that RNAi or chemical inhibition of LSD1enhances γ-globin expression in human erythroid cells and quitemodestly β-locus transgenic mice (21). Although these findingsillustrate favorable features of LSD1 as a potential target, wealso demonstrate that LSD1 serves critical roles more broadly inerythroid cell maturation and overall globin expression. LSD1homozygous KO mice are embryonic lethal with severe anemia.Deletion of LSD1 in adult bone marrow results in impaired func-tion of hematopoietic stem cells, neutropenia, and markedly re-duced number of leukocytes (15), indicating that LSD1 is necessaryfor specification and terminal differentiation of several hema-topoietic lineages. Similarly, interference with LSD1 by shRNAsor inhibitors impairs erythroid maturation from human CD34+HSPCs ex vivo (Figs. S5 and S6). Accordingly, its potential as atherapeutic target for HbF induction is compromised by its multi-faceted roles in hematopoiesis. Our study also shows that DNMT1haploinsufficiency augments HbF expression elicited by BCL11Adeficiency. However, complete loss of DNMT1 is incompatiblewith normal hematopoietic development (22). Therefore, thetherapeutic window appears to be relatively narrow wherebyDNMT1 activity may be inhibited without perturbing normalerythroid functions.The nucleosome remodeling and histone deacetylase (NuRD)

complex is also identified as a BCL11A-interacting corepressorcomplex. The NuRD complex consists of several enzymatic sub-units, including the ATPase subunit Mi-2β and the histone deace-tylases (HDAC1 and HDAC2). Depletion of Mi-2β results in aprofound increase in γ-globin expression in primary human ery-throid cells (the top-ranked gene CHD4; Fig. 2B; Fig. S3). Con-sistent with a role in HbF silencing, knockdown or KO of Mi-2βreactivates γ-globin genes in the β-YAC transgenic mice (23, 24).However, the degree of γ-globin induction is substantially less thanthat seen with BCL11A deficiency (8). Additionally, depletion ofHDAC1 and HDAC2 reactivates γ-globin expression in primaryhuman erythroid cells (Fig. 2B). HDAC1 and HDAC2 have alsobeen identified as HbF inducers by a chemical genetic screen (25).These results collectively provide strong support for important

0

0.1

0.25

10

15

20

Lsd1

P = 0.0006

P = 0.0002P = 0.14

P = 0.001

γ-gl

obin

mR

NA

(%)

Bcl11a +/+ +/- +/- -/- -/-+/+ +/+ +/- +/+ +/-

0

0.05

0.110

15

20

25

P = 0.07

P = 0.03

Lsd1Bcl11a +/+ +/+ -/- -/-

+/+ -/- +/+ -/-

A B

LSD1

BCL11A

GAPDH

shBCL1

1A

LSD1 s

h1

LSD1 s

h5

shGFP

Day3Day7

0γ-gl

obin

mR

NA

(%)

102030405060

Day3Day7

shBCL1

1A

LSD1 s

h1

LSD1 s

h5

shGFP

0

Tota

l β-g

lobi

n m

RN

A

20

40

60

80

∗ ∗

∗∗ ∗∗

∗∗∗∗

∗∗ ∗∗

0

0.02

0.045

10

20

30

Dnmt1Bcl11a +/+ -/- -/-

+/+ +/+ +/-

P = 1.9e-6

P = 6.3e-10DNMT1

GAPDH

sh1

sh2

sh3

shG

FP

sh4

sh5

DNMT1

020406080

100

γ-gl

obin

mR

NA

(%)

sh1

sh2

sh3

shG

FP sh4

sh5

DNMT1

0

10

20

30

Tota

l β-g

lobi

nm

RN

A

∗∗ ∗

C FE

D

γ-gl

obin

mR

NA

(%)

γ-gl

obin

mR

NA

(%)

Bcl11a::Lsd1 cKO (by EpoR-Cre) Bcl11a::Lsd1 cKO (by Mx1-Cre) Bcl11a::Dnmt1 cKO (by EpoR-Cre)

Fig. 5. Compound KO of Bcl11a, Lsd1, or Dnmt1 enhances γ-globin expression. (A) Expression of γ-globin mRNA was measured by qRT-PCR in peripheralblood (PB) of Bcl11a::Lsd1 cKO animals (by EpoR-Cre; 8–12 wk old). Results are the means ± SD. P values were determined by a two-tailed t test. (B) Expressionof γ-globin in PB of Bcl11a::Lsd1 cKO animals (by Mx1-Cre; 8–12 wk old). (C) Lentiviral shRNA-mediated knockdown of BCL11A and LSD1 proteins in primaryhuman erythroid cells. (D) Knockdown of BCL11A or LSD1 results in increased γ-globin mRNAs (Upper). Knockdown of LSD1 reduces total β-like globin mRNAs(Lower). Data are shown as means ± SD; *P < 0.05, **P < 0.01. (E) shRNA-mediated knockdown of DNMT1 in primary human erythroid cells (Upper) resultsin increase of γ-globin mRNAs and modest decrease in total β-like globin mRNAs (Lower). Data are shown as means ± SD; *P < 0.05. (F) Expression of γ-globinin PB of Dnmt1::Bcl11a cKO animals (by EpoR-Cre; 8–12 wk old).

6522 | www.pnas.org/cgi/doi/10.1073/pnas.1303976110 Xu et al.

roles of several enzymatic subunits within the NuRD complexin globin expression, yet the extent to which various effects canbe ascribed directly to action in concert with BCL11A or inde-pendently are difficult to discern.Our studies used two model systems, including the primary

human erythroid culture and the β-YAC transgenic mice, to eval-uate HbF regulation and erythroid maturation. To interpret resultsmeaningfully, it is important to acknowledge the potential limi-tations unique to each system. Although cultured primary humanerythroid cells model several important aspects of human erythroiddevelopment such as an adult-stage pattern of globin profile, thebackground level of γ-globin mRNA is much higher than thatpresent in adult erythroid cells of normal individuals. Moreover,CD34+ HSPCs appear relatively permissive for HbF induction byvarious agents, thus rendering them sensitive but not necessarilyspecific indicators to assay for the effects of manipulations in-ducing HbF (2). Changes in growth kinetics and differentiationstatus of CD34+ HSPCs affect HbF expression. In contrast to pri-mary human cells, the baseline expression of transgenic γ-globingene in adult β-YAC mice is far lower than anticipated fromnormal individuals, possibly due to differences in the transactingenvironment between mouse and human (5). Thus, interventionsseeking to derepress γ-globin expression face a much greaterquantitative hurdle. Although some manipulations, such as in-hibition of LSD1 or DNMT1, may induce transgenic γ-globinexpression several fold above this low level, the effects are none-theless small compared with those elicited by BCL11A KO. Careneeds to be exercised in extrapolating findings in the availablemodel systems to nominate targets for therapeutic manipulation.

Materials and MethodsExperimental Animals. The β-globin locus transgenic (β-YAC) mouse strainwas created as described (26). Mice containing a Bcl11a floxed allele werecreated as described (8). Mice containing an Lsd1 floxed allele (M.A.K. andS.H.O.) or a Rocr1 (encoding CoREST) floxed allele (H.Y. and G.M.) weregenerated through gene targeting approaches. Dnmt1fl/fl mice were cre-ated as previously described (22, 27). To obtain the Bcl11a cKO mice, theBcl11afl/fl mice were crossed with the EpoR-Cre knock-in mice (28) or theMx1-Cre transgenic mice (14). These mice were crossed with Lsd1fl/fl orDnmt1fl/fl mice to create compound cKO mice. PolyI:ployC (pIpC) was pre-pared in PBS and administrated via i.p. injection daily at a dose of 25 μg/kgfor 7 d. Peripheral blood was isolated via the retroorbital plexus and

analyzed on a HEMAVET HV950 hematology system (Drew Scientific). Allexperiments were performed with the approval of the Children’s HospitalBoston Animal Ethics Committee.

Cell Culture. Primary adult human CD34+ HSPC-derived erythroid progenitorswere generated ex vivo as described (6). The MEL-BirA (MBirA), MEL-FLAG-Bio-BCL11A (MBB1.4), K562-BirA (KBirA), and K562-FLAG-Bio-BCL11A (KBB2.4)stable cell lines were generated as described (10).

Protein Affinity Purification and Proteomic Analysis. BCL11A-interacting multi-protein complexes were purified and characterized as described (29). Copurifiedproteins were separated by SDS/PAGE, followed by whole lane LC-MS/MSusing an LTQ linear ion-trap mass spectrometer. A subtractive approach in-cluding parallel pull-down in parental MEL-BirA (MBirA) or K562-BirA (KBirA)cells was used.

Lentiviral RNAi. Lentiviral shRNA clones in the pLKO.1-puro vector wereobtained from Sigma-Aldrich (two to five shRNAs per gene; Table S1).Primary human CD34+ HSPCs were transduced by spin infection at day 3 ofexpansion. Cells were washed three times with PBS and seeded in fresh media24 h after infection. Selection (1 μg/mL puromycin) was initiated at 48 h afterinfection, followed by erythroid differentiation for 5–7 d.

ChIP. ChIP was performed as described previously (29) using the followingantibodies: H3K4me2 (07-030; Millipore), H3K9me1 (ab9045; Abcam), BCL11A(ab19487 and ab18688; Abcam), LSD1 (ab17721; Abcam), CoREST (07-455; Mil-lipore), Mi-2β (provided by Stephen Smale, University of California, Los Angeles,CA), HDAC1 (06-720;Millipore), EZH2 (07-689;Millipore), RNAPol II (sc-899; SantaCruz Biotechnology), and normal rabbit IgG (sc-2027; Santa Cruz Biotechnology).

RNA Isolation and qRT-PCR. RNA was extracted using the QIAamp RNA BloodMini Kit or RNeasy Plus Mini Kit (Qiagen). RNA was reverse-transcribed andanalyzed with the iQ SYBR Green Supermix using the iCycler Real-Time PCRDetection System (Bio-Rad). Primer sequences are listed in Table S2.

ACKNOWLEDGMENTS. We thank K. Peterson for providing β-YAC mice. Wethank D. R. Higgs, D. A. Williams, L. I. Zon, A. B. Cantor, V. G. Sankaran, H.Huang, and A. Woo for helpful advice and discussions. This work was sup-ported by funding from the National Institutes of Health (to S.H.O. and G.M.),including a Center of Excellence in Molecular Hematology Award from theNational Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (toS.H.O.). S.H.O. and G.M. are Investigators of the Howard Hughes MedicalInstitute (HHMI). J.X. is an HHMI-Helen Hay Whitney Foundation fellowand is supported by NIDDK Career Development Award K01DK093543.

1. Stamatoyannopoulos G (2005) Control of globin gene expression during developmentand erythroid differentiation. Exp Hematol 33(3):259–271.

2. Bauer DE, Kamran SC, Orkin SH (2012) Reawakening fetal hemoglobin: Prospects fornew therapies for the β-globin disorders. Blood 120(15):2945–2953.

3. Thein SL, Menzel S, Lathrop M, Garner C (2009) Control of fetal hemoglobin: New insightsemerging from genomics and clinical implications. Hum Mol Genet 18(R2):R216–R223.

4. Sankaran VG, et al. (2008) Human fetal hemoglobin expression is regulated by thedevelopmental stage-specific repressor BCL11A. Science 322(5909):1839–1842.

5. Sankaran VG, et al. (2009) Developmental and species-divergent globin switching aredriven by BCL11A. Nature 460(7259):1093–1097.

6. Xu J, et al. (2010) Transcriptional silencing of gamma-globin by BCL11A involves long-range interactions and cooperation with SOX6. Genes Dev 24(8):783–798.

7. Jawaid K, Wahlberg K, Thein SL, Best S (2010) Binding patterns of BCL11A in theglobin and GATA1 loci and characterization of the BCL11A fetal hemoglobin locus.Blood Cells Mol Dis 45(2):140–146.

8. Xu J, et al. (2011) Correction of sickle cell disease in adult mice by interference withfetal hemoglobin silencing. Science 334(6058):993–996.

9. Sankaran VG, et al. (2011) A functional element necessary for fetal hemoglobin si-lencing. N Engl J Med 365(9):807–814.

10. Kim J, Cantor AB, Orkin SH, Wang J (2009) Use of in vivo biotinylation to studyprotein-protein and protein-DNA interactions in mouse embryonic stem cells. NatProtoc 4(4):506–517.

11. Miller IJ, Bieker JJ (1993) A novel, erythroid cell-specific murine transcription factorthat binds to the CACCC element and is related to the Krüppel family of nuclear proteins.Mol Cell Biol 13(5):2776–2786.

12. Borg J, et al. (2010) Haploinsufficiency for the erythroid transcription factor KLF1causes hereditary persistence of fetal hemoglobin. Nat Genet 42(9):801–805.

13. Zhou D, Liu K, Sun CW, Pawlik KM, Townes TM (2010) KLF1 regulates BCL11A expressionand gamma- to beta-globin gene switching. Nat Genet 42(9):742–744.

14. Kühn R, Schwenk F, Aguet M, Rajewsky K (1995) Inducible gene targeting in mice.Science 269(5229):1427–1429.

15. Sprüssel A, et al. (2012) Lysine-specific demethylase 1 restricts hematopoietic progenitorproliferation and is essential for terminal differentiation. Leukemia 26(9):2039–2051.

16. Lee MG, Wynder C, Schmidt DM, McCafferty DG, Shiekhattar R (2006) Histone H3lysine 4 demethylation is a target of nonselective antidepressive medications. ChemBiol 13(6):563–567.

17. DeSimone J, Heller P, Hall L, Zwiers D (1982) 5-Azacytidine stimulates fetal hemo-globin synthesis in anemic baboons. Proc Natl Acad Sci USA 79(14):4428–4431.

18. Ley TJ, et al. (1982) 5-azacytidine selectively increases gamma-globin synthesis in apatient with beta+ thalassemia. N Engl J Med 307(24):1469–1475.

19. Liu P, et al. (2003) Bcl11a is essential for normal lymphoid development. Nat Immunol4(6):525–532.

20. Leid M, et al. (2004) CTIP1 and CTIP2 are differentially expressed during mouse em-bryogenesis. Gene Expr Patterns 4(6):733–739.

21. Shi L, Cui S, Engel JD, Tanabe O (2013) Lysine-specific demethylase 1 is a therapeutictarget for fetal hemoglobin induction. Nat Med 19(3):291–294.

22. Trowbridge JJ, SnowJW,Kim J,Orkin SH (2009)DNAmethyltransferase1 is essential for anduniquely regulates hematopoietic stem and progenitor cells. Cell Stem Cell 5(4):442–449.

23. Costa FC, Fedosyuk H, Chazelle AM, Neades RY, Peterson KR (2012) Mi2β is requiredfor γ-globin gene silencing: Temporal assembly of a GATA-1-FOG-1-Mi2 repressorcomplex in β-YAC transgenic mice. PLoS Genet 8(12):e1003155.

24. Amaya ML, et al. (2013) Mi2β-mediated silencing of the fetal γ-globin gene in adulterythroid cells. Blood, 10.1182/blood-2012-11-466227.

25. Bradner JE, et al. (2010) Chemical genetic strategy identifies histone deacetylase 1(HDAC1) and HDAC2 as therapeutic targets in sickle cell disease. Proc Natl Acad SciUSA 107(28):12617–12622.

26. Peterson KR, et al. (1995) Use of yeast artificial chromosomes (YACs) in studies ofmammalian development: Production of beta-globin locus YAC mice carrying humanglobin developmental mutants. Proc Natl Acad Sci USA 92(12):5655–5659.

27. Jackson-Grusby L, et al. (2001) Loss of genomic methylation causes p53-dependentapoptosis and epigenetic deregulation. Nat Genet 27(1):31–39.

28. Heinrich AC, Pelanda R, Klingmüller U (2004) A mouse model for visualization andconditional mutations in the erythroid lineage. Blood 104(3):659–666.

29. Xu J, et al. (2012) Combinatorial assembly of developmental stage-specific enhancerscontrols gene expression programs during human erythropoiesis. Dev Cell 23(4):796–811.

Xu et al. PNAS | April 16, 2013 | vol. 110 | no. 16 | 6523

MED

ICALSC

IENCE

S

Supporting InformationXu et al. 10.1073/pnas.1303976110SI Materials and MethodsStreptavidin Immunoprecipitation and Western Blot. Streptavidinimmunoprecipitation (SA-IP) experiments in MEL or K562 stablecell lines were performed as described previously (1). Briefly,nuclear extracts were prepared from MEL-BirA (MBirA), MEL-FLAG-Bio-BCL11A (MBB1.4), K562-BirA (KBirA), and K562-FLAG-Bio-BCL11A (KBB2.4) stable cell lines, immunoprecipitatedwith streptavidin agarose beads (Invitrogen), and processed forWestern blot analysis. The following antibodies were used forcoimmunoprecipitation (co-IP) andWestern blot analysis: BCL11A(ab18688; Abcam), GATA1 (ab28839; Abcam), FOG1 (sc-9362;Santa Cruz Biotechnology), lysine-specific demethylase 1 (LSD1;ab17721; Abcam), CoREST (07-455; Millipore), NCOR1 (ab24552;Abcam), SMRT (04-1551; Millipore), KAISO (ab12723; Abcam),SIN3A (sc-994; Santa Cruz Biotechnology), TRIM28 (ab22553;Abcam), Mi-2β (provided by Stephen Smale, University ofCalifornia, Las Angeles, CA), MTA2 (sc-9447; Santa Cruz Bio-technology), HDAC1 (06-720; Millipore), HDAC2 (sc-7899; SantaCruz Biotechnology), MBD3 (sc-9402; Santa Cruz Biotechnology),SNF5, BRG1, BAF155 (provided by Charles Roberts, Dana-FarberCancer Institute, Boston), EZH2 (612666; BC Biosciences), SUZ12(ab12073; Abcam), SP1 (sc-14027; Santa Cruz Biotechnology),DNA methyltransferase 1 (DNMT1; 39204; Active Motif), andGAPDH (sc-26778; Santa Cruz Biotechnology).

Co-IP and Western Blot.Co-IP experiments of endogenous BCL11Aprotein in primary human erythroid cells were performed asdescribed (2). Briefly, 1–5 mg of nuclear extract proteins was

incubated with 5–25 μg of BCL11A or mouse IgG antibody over-night at 4 °C. The protein complexes were collected by protein G/A-agarose beads (Invitrogen), followed by four washes withBC139K buffer. The beads were boiled for 5 min in sample buffer(Bio-Rad), and the eluted material was used for Western blotanalysis.

Flow Cytometry.Cells were analyzed by flow cytometry as describedpreviously (3). Live cells were identified and gated by exclusion of7-amino-actinomycin D (7-AAD; BD Pharmingen). The cells wereanalyzed for expression of cell surface antigens with antibodiesspecific for CD34, CD71, and CD235a conjugated to phycoerythrin(PE), fluorescein isothiocyanate (FITC), or allophycocyanin (APC;BD Pharmingen). Data were analyzed using FlowJo software.

Cytospin. Cytocentrifuge preparations from cells at various stagesof differentiation were stained with May-Grunwald-Giemsa asdescribed previously (4).

Chemicals. Two LSD1 inhibitors, pargyline (Cayman Chemical)and tranylcypromine (TCP; Sigma-Aldrich), and three DNAmethylation inhibitors, 5-aza-2′-deoxycytidine (5-azaD; Sigma-Aldrich), Zebularine (Tocris Bioscience), and RG108 (TocrisBioscience), were used to treat primary human CD34+ hema-topoietic stem/progenitor cell–derived erythroid progenitors.Cells were incubated with inhibitors for 72 h before harvest foranalysis.

1. Xu J, et al. (2012) Combinatorial assembly of developmental stage-specific enhancerscontrols gene expression programs during human erythropoiesis. Dev Cell 23(4):796–811.

2. Xu J, et al. (2010) Transcriptional silencing of gamma-globin by BCL11A involves long-range interactions and cooperation with SOX6. Genes Dev 24(8):783–798.

3. Xu J, et al. (2011) Correction of sickle cell disease in adult mice by interference withfetal hemoglobin silencing. Science 334(6058):993–996.

4. Sankaran VG, et al. (2009) Developmental and species-divergent globin switching aredriven by BCL11A. Nature 460(7259):1093–1097.

Xu et al. www.pnas.org/cgi/content/short/1303976110 1 of 11

MB

irAM

BB

1.4

MB

irAM

BB

1.4

Input IP

WBα-BCL11A

α-GATA1

α-FOG1

α-LSD1

α-CoREST

α-SMRT

α-KAISO

α-SIN3A

α-TRIM28

α-Mi-2β

α-MTA2

α-HDAC1

α-HDAC2

α-MBD3

α-SNF5

α-BRG1

α-BAF155

α-EZH2

α-SUZ12

KB

irAK

BB

2.4

KB

irAK

BB

2.4

Input IP

WB

α-BCL11A

α-GATA1

α-LSD1

α-CoREST

α-SMRT

α-KAISO

α-SIN3A

α-TRIM28

α-Mi-2β

α-MTA2

α-HDAC1

α-HDAC2

α-MBD3

α-SNF5

α-BRG1

α-BAF155

α-EZH2

α-SUZ12

α-NCOR1α-NCOR1

LSD

1/C

oRE

ST

NC

oR/S

MR

TO

ther

Co-

Rep

.N

uRD

SW

I/SN

FP

RC

2

LSD

1/C

oRE

ST

NC

oR/S

MR

TO

ther

Co-

Rep

.N

uRD

SW

I/SN

FP

RC

2

α-SP1 α-SP1

BA

Fig. S1. Validation of interactions between BCL11A and identified partner proteins in erythroid cells. (A) SA-IP experiments were performed in MEL-BirA(MBirA) and MEL-FLAG-Bio-BCL11A (MBB1.4) stable cells. BCL11A-interacting protein complexes were purified by streptavidin immunoprecipitation followedby Western blot analysis; 2% of input nuclear extracts were analyzed as loading controls. (B) SA-IP experiments were performed in K562-BirA (MBirA) andK562-FLAG-Bio-BCL11A (KBB2.4) stable cells, followed by Western blot analysis.

Xu et al. www.pnas.org/cgi/content/short/1303976110 2 of 11

A

shG

FPB

CL1

1AK

LF1

CH

D4

DN

MT1

SIN

3AH

DA

C1

NC

oR1

EZH

2H

DA

C2

MB

D2

EE

DK

DM

3AS

UZ1

2K

DM

5BR

BB

P4

KD

M5A

EZH

1IK

ZF1

KD

M5D

KD

M4B

KD

M7A

RB

BP

7K

DM

4CB

AF1

55LS

D1

ZBTB

33N

CoR

2K

DM

3BK

DM

4AB

RG

1B

CoR

TRIM

28

B

30

20

10

0

γ-gl

obin

mR

NA

(rel

ativ

e to

GA

PD

H)

C

0.16

0.12

0.08

0.04

0

50

40

30

20

10

0

SO

X6

shG

FPB

CL1

1AK

LF1

CH

D4

DN

MT1

SIN

3AH

DA

C1

NC

oR1

EZH

2H

DA

C2

MB

D2

EE

DK

DM

3AS

UZ1

2K

DM

5BR

BB

P4

KD

M5A

EZH

1IK

ZF1

KD

M5D

KD

M4B

KD

M7A

RB

BP

7K

DM

4CB

AF1

55LS

D1

ZBTB

33N

CoR

2K

DM

3BK

DM

4AB

RG

1B

CoR

TRIM

28

SO

X6

ε-gl

obin

mR

NA

(rel

ativ

e to

GA

PD

H)

shG

FPB

CL1

1AK

LF1

CH

D4

DN

MT1

SIN

3AH

DA

C1

NC

oR1

EZH

2H

DA

C2

MB

D2

EE

DK

DM

3AS

UZ1

2K

DM

5BR

BB

P4

KD

M5A

EZH

1IK

ZF1

KD

M5D

KD

M4B

KD

M7A

RB

BP

7K

DM

4CB

AF1

55LS

D1

ZBTB

33N

CoR

2K

DM

3BK

DM

4AB

RG

1B

CoR

TRIM

28

SO

X6

β-gl

obin

mR

NA

(rel

ativ

e to

GA

PD

H)

Fig. S2. Functional RNAi screen of BCL11A-interacting partner proteins. Expression of human β-like globin mRNAs was measured by qRT-PCR on shRNA-mediated knockdown of each gene. Data are plotted as relative mRNA level normalized to GAPDH mRNA level for human (A) fetal γ-globin, (B) embryonice-globin, and (C) adult β-globin genes, respectively. Results are the means ± SD of at least three independent experiments. The same data are plotted aspercentage of each globin gene over total β-like human globin gene levels and shown in Fig. 2.

Xu et al. www.pnas.org/cgi/content/short/1303976110 3 of 11

A

100 101 102 103 1040

30

60

90

120

97.5

100 101 102 103 1040

30

60

90

120

98.9

CD71

Cou

nts

100 101 102 103 1040

20

40

60

80

97.1

100 101 102 103 1040

20

40

60

80

96.6

CD235a

Cou

nts

100 101 102 103 1040

20

40

60

7.97

100 101 102 103 1040

20

40

60

80

3.71

CD34

Cou

nts

shGFP shBCL11A CHD4 sh4 CHD4 sh5

100 101 102 103 1040

1

2

3

4

100 101 102 103 1040

20

40

60

80

100

4.24 9.21

100 101 102 103 1040

10

20

30

40

87.4

100 101 102 103 1040

20

40

60

80

100

90.7

100 101 102 103 1040

10

20

30

83.1

100 101 102 103 1040

20

40

60 84.9

100 101 102 103 1040

30

60

90

120

97.5

CD71

Cou

nts

100 101 102 103 1040

20

40

60

80

97.1

CD235

Cou

nts

100 101 102 103 1040

20

40

60

7.97

CD34

Cou

nts

shGFP NCoR1 sh3

4.04

100 101 102 103 1040

50

100

150

100 101 102 103 1040

10

20

30

100 101 102 103 1040

1

2

3

4

100 101 102 103 1040

50

100

150

200

250

98.3

100 101 102 103 1040

10

20

30

40

85.8

100 101 102 103 1040

2

4

6

72

NCoR1 sh4 NCoR1 sh5

2.08 4.12

100 101 102 103 1040

30

60

90

120

97.5

CD71

Cou

nts

100 101 102 103 1040

20

40

60

80

97.1

CD235

Cou

nts

100 101 102 103 1040

20

40

60

7.97

CD34

Cou

nts

shGFP SIN3A sh1 SIN3A sh2 SIN3A sh3

100 101 102 103 1040

50

100

150

200

100 101 102 103 1040

50

100

150

100 101 102 103 1040

50

100

150

100 101 102 103 1040

30

60

90

120

90.5

100 101 102 103 1040

20

40

60

80

100

89.1

100 101 102 103 1040

50

100

150

200

250

98.7

100 101 102 103 1040

20

40

60

75

100 101 102 103 1040

20

40

60 77.5

100 101 102 103 1040

30

60

90

120

98.1

0.21 0.54 0.32

Mi-2β

GAPDH

sh1

sh2

sh3

shG

FP

sh4

sh5

CHD4

shB

CL1

1A0

20406080

100γ-

glob

in m

RN

A(%

)

01020

30

Tota

l β-g

lobi

nm

RN

A

40

50

shG

FP sh4

sh5

CHD4

shB

CL1

1A

B

NCOR1

GAPDH

sh1

sh2

sh3

shG

FP

sh4

sh5

NCOR1

shB

CL1

1A

020406080

100

γ-gl

obin

mR

NA

(%)

01020

30

Tota

l β-g

lobi

nm

RN

A

40

50

shG

FP sh4

sh5

NCOR1

shB

CL1

1A sh3

C D

SIN3A

GAPDH

sh1

sh2

sh3

shG

FP

sh4

sh5

SIN3A

shB

CL1

1A

020406080

100

γ-gl

obin

mR

NA

(%)

0

20

Tota

l β-g

lobi

nm

RN

A 40

60

shG

FP sh2

sh3

SIN3A

shB

CL1

1A sh1

100 101 102 103 1040

50

100

150

200

FL2-H: CD235

98.5

100 101 102 103 1040

10

20

30

80.2

100 101 102 103 1040

1

2

3

4

72

E F

Fig. S3. Representative genes in functional RNAi screen in primary human adult eryrhoid cells. (A) Lentiviral shRNA-mediated knockdown of CHD4 (encodingMi2β) reactivates human γ-globin expression. Expression of Mi2βwas monitored by Western blot analysis (Upper). Expression of human γ-globin mRNA (as % oftotal β-like globin mRNAs) and total β-like globin mRNA (relative mRNA value normalized to GAPDH mRNA) was measured by qRT-PCR (Lower). Results are themeans ± SD of at least three independent experiments. (B) Erythroid maturation was assessed by flow cytometry analysis of cell surface marker (CD71 andCD235a) expression in primary human CD34+ HSPC-derived erythroid progenitor cells at day 5 of differentiation. (C) Lentiviral shRNA-mediated knockdown ofNCOR1 reactivates human γ-globin expression. (D) Erythroid maturation was assessed by flow cytometry analysis of cell surface marker expression onknockdown of NCOR1 expression. (E) Lentiviral shRNA-mediated knockdown of SIN3A reactivates human γ-globin expression. (F) Erythroid maturation wasassessed by flow cytometry analysis of cell surface marker expression on knockdown of SIN3A expression.

Xu et al. www.pnas.org/cgi/content/short/1303976110 4 of 11

A

B

RB

Cs

(x10

/mL)

9

0

3

6

9

12

Lsd1Bcl11a +/+ -/- -/-

+/+ +/+ +/-

Hgb

(g/d

L)

0

4

8

12

16

+/+ -/- -/-+/+ +/+ +/-

WB

Cs

(x10

/mL)

6

0

3

6

9

12

+/+ -/- -/-+/+ +/+ +/-

RB

Cs

(x10

/mL)

9

0

3

6

9

12

Lsd1Bcl11a +/+ -/- -/-

+/+ +/+ -/-H

gb (g

/dL)

0

4

8

12

16

+/+ -/- -/-+/+ +/+ -/-

WB

Cs

(x10

/mL)

6

0

3

6

9

12

+/+ -/- -/-+/+ +/+ -/-

∗

∗∗

∗

∗∗

∗

Bcl11a::Lsd1 cKO (by EpoR-Cre)

Bcl11a::Lsd1 cKO (by Mx1-Cre)

Fig. S4. Differential peripheral blood analysis of LSD1 and BCL11A compound KO mice. (A) Differential peripheral blood (PB) counts in control (EpoR-Cre−),Bcl11a KO, and Bcl11a::Lsd1 compound KO (by EpoR-Cre) β-YAC mice. Results are shown as means ± SEM (n ≥ 5 per genotype; *P < 0.05). (B) Differential PBcounts in control (Mx1-Cre−), Bcl11a KO, and Bcl11a::Lsd1 compound KO (by Mx1-Cre) β-YAC mice (n ≥ 5 per genotype; *P < 0.05, **P < 0.01).

Xu et al. www.pnas.org/cgi/content/short/1303976110 5 of 11

100 101 102 103 1040

30

60

90

120

97.5

100 101 102 103 1040

30

60

90

120

98.9

100 101 102 103 1040

10

20

30

35.2

100 101 102 103 1040

20

40

60

31.9

CD71

Cou

nts

100 101 102 103 1040

20

40

60

80

97.1

100 101 102 103 1040

20

40

60

80

96.6

100 101 102 103 1040

10

20

30

26.9

100 101 102 103 1040

20

40

60

80

17

CD235a

Cou

nts

100 101 102 103 1040

20

40

60

7.97

100 101 102 103 1040

20

40

60

80

3.71

100 101 102 103 1040

10

20

30

40

30.8

100 101 102 103 1040

20

40

60

30.5

CD34

Cou

nts

shGFP shBCL11A LSD1 sh1 LSD1 sh5A

shGFP shBCL11A LSD1 sh1 LSD1 sh5

Day

5D

ay 9

B

Fig. S5. Depletion of LSD1 expression in primary human CD34+ HSPCs leads to impaired erythroid differentiation. (A) Erythroid maturation was assessed byflow cytometry analysis of cell surface marker (CD34, CD71, and CD235a) expression in primary human CD34+ HSPC-derived erythroid progenitor cells at day 5of differentiation. Cells were transduced with lentiviruses containing shRNAs again GFP (shGFP, negative control), BCL11A (shBCL11A), and LSD1 (sh1 and sh5),respectively. (B) Representative cytospin images are shown for cells at days 5 and 9 of differentiation. Under normal culture conditions, the majority of thedifferentiating erythroid progenitors acquired proerythroblast morphology at day 5 of differentiation. At day 9 of differentiation, the majority of cells werepolychromatophilic and orthochromatic. Depletion of BCL11A expression by lentiviral shRNA had no effect on erythroid maturation. Depletion of LSD1 re-sulted in marked decrease in total cell number and polychromatophilic/orthochromatic erythroid progenitors.

Xu et al. www.pnas.org/cgi/content/short/1303976110 6 of 11

Control 0.3 1 3 5 10 25Pargyline (mM) TCP (μM)

0

40

60

80

100

20

Tota

l β-g

lobi

n m

RN

AB

100 101 102 103 104

CD71

Cou

nts

CD235a

Cou

nts

100 101 102 103 104

0 mM Pargyline0.3 mM Pargyline1 mM Pargyline3 mM Pargyline

100 101 102 103 104

CD71

Cou

nts

CD235a

Cou

nts

100 101 102 103 104

0 μM TCP5 μM TCP10 μM TCP25 μM TCP

C

D

0

γ-gl

obin

mR

NA

(%)

10

20

30

40

50

60

Control 0.3 1 3 5 10 25Pargyline (mM) TCP (μM)

Control shBCL11A

∗∗ ∗∗ ∗∗

∗ ∗∗∗

∗

Control shBCL11A

∗∗

∗∗∗∗ ∗∗ ∗∗ ∗∗

A

Fig. S6. Inhibition of LSD1 activity leads to increased γ-globin expression and impaired erythroid differentiation in primary human erythroid cells. (A) Ex-pression of human γ-globin mRNA was measured by qRT-PCR in primary human erythroid progenitor cells treated with two LSD1 inhibitors: pargyline and TCP.Data are shown as means ± SD; *P < 0.05. (B) Inhibition of LSD1 results in decrease in total mRNA level of β-like globin genes (Lower) in primary humanerythroid progenitor cells. Data are shown as means ± SD; *P < 0.05, **P < 0.01. (C) Erythroid maturation was assessed by flow cytometry analysis of cell surfacemarker (CD71 and CD235a) expression in primary human erythroid progenitor cells at day 5 of differentiation in the presence or absence of pargyline. (D)Erythroid maturation was assessed by flow cytometry analysis of cell surface marker (CD71 and CD235a) expression in primary human erythroid progenitor cellsat day 5 of differentiation in the presence or absence of TCP.

Xu et al. www.pnas.org/cgi/content/short/1303976110 7 of 11

5-azaD (μM)

0

20

40

60

80

100

γ-gl

obin

mR

NA

(%)

0 0.001 0.01 0.1 1 10

5-azaD (μM)

020406080

100

Tota

l β-g

lobi

n m

RN

A

0 0.001 0.01 0.1 1 10

120

Zebularine (μM)0 1 10 100 1000

0

20

40

60

80

100

γ-gl

obin

mR

NA

(%)

Zebularine (μM)0 1 10 100 1000

20406080

100

Tota

l β-g

lobi

n m

RN

A 120

0

RG108 (μM)0 1 10 100 1000

0

20

40

60

80

100

γ-gl

obin

mR

NA

(%)

RG108 (μM)0 1 10 100 1000

20406080

100

Tota

l β-g

lobi

n m

RN

A 120

0

ControlshBCL11A

ControlshBCL11A

ControlshBCL11A

ControlshBCL11A

ControlshBCL11A

ControlshBCL11A

A

B

∗∗ ∗

Fig. S7. Inhibition of DNMT1 activity leads to increased γ-globin expression in primary human erythroid cells. (A) Expression of human γ-globin mRNA wasmeasured by qRT-PCR in primary human erythroid progenitor cells treated with three DNAmethylation inhibitors in the presence or absence of BCL11A shRNA.(B) Expression of total human β-like globin mRNA was measured by qRT-PCR in primary human erythroid progenitor cells treated with DNA methylationinhibitors in the presence or absence of BCL11A shRNA. Data are shown as means ± SD; *P < 0.05.

RB

Cs

(x10

/mL)

9

0

3

6

9

12

Dnmt1Bcl11a +/+ -/- -/-

+/+ +/+ +/-

WB

Cs

(x10

/mL)

6

0

3

6

9

12

+/+ -/- -/-+/+ +/+ +/-

∗

Hgb

(g/d

L)

0

4

8

12

16

+/+ -/- -/-+/+ +/+ +/-

Bcl11a::Dnmt1 cKO (by EpoR-Cre)

Fig. S8. Differential peripheral blood analysis of DNMT1 and BCL11A compound KO mice. Differential PB counts in control (EpoR-Cre−), Bcl11a KO, andBcl11a::Dnmt1 compound KO (by EpoR-Cre) β-YAC mice (n ≥ 5 per genotype; *P < 0.05).

Xu et al. www.pnas.org/cgi/content/short/1303976110 8 of 11

Table S1. List of lentiviral shRNA sequences used in this study

Gene name shRNA ID TRC number shRNA sequences

GFP shGFP SHC005 CCGGTACAACAGCCACAACGTCTATCTCGAGATAGACGTTGTGGCTGTTGTATTTTTBCL11A sh49 TRCN0000033449 CCGGCGCACAGAACACTCATGGATTCTCGAGAATCCATGAGTGTTCTGTGCGTTTTTG

sh51 TRCN0000033451 CCGGCCAGAGGATGACGATTGTTTACTCGAGTAAACAATCGTCATCCTCTGGTTTTTGsh53 TRCN0000033453 CCGGGCATAGACGATGGCACTGTTACTCGAGTAACAGTGCCATCGTCTATGCTTTTTG

KLF1 sh1 TRCN0000230814 CCGGTGCACATGAAGCGCCACCTTTCTCGAGAAAGGTGGCGCTTCATGTGCATTTTTGsh4 TRCN0000230812 CCGGCCCTCCTTCCTGAGTTGTTTGCTCGAGCAAACAACTCAGGAAGGAGGGTTTTTGsh5 TRCN0000230813 CCGGCAGAGGATCCAGGTGTGATAGCTCGAGCTATCACACCTGGATCCTCTGTTTTTG

CHD4 sh4 TRCN0000021362 CCGGGCTGACACAGTTATTATCTATCTCGAGATAGATAATAACTGTGTCAGCTTTTTsh5 TRCN0000021363 CCGGGCGGGAGTTCAGTACCAATAACTCGAGTTATTGGTACTGAACTCCCGCTTTTT

DNMT1 sh2 TRCN0000021891 CCGGGCCCAATGAGACTGACATCAACTCGAGTTGATGTCAGTCTCATTGGGCTTTTTsh3 TRCN0000021893 CCGGCGACTACATCAAAGGCAGCAACTCGAGTTGCTGCCTTTGATGTAGTCGTTTTTsh5 TRCN0000021892 CCGGCGAGAAGAATATCGAACTCTTCTCGAGAAGAGTTCGATATTCTTCTCGTTTTT

SIN3A sh1 TRCN0000021774 CCGGCGTGAACATCTAGCACAGAAACTCGAGTTTCTGTGCTAGATGTTCACGTTTTTsh2 TRCN0000021775 CCGGCCCTGAGTTGTTTAATTGGTTCTCGAGAACCAATTAAACAACTCAGGGTTTTTsh3 TRCN0000021776 CCGGGCTACGTCTCAAAGAACCTATCTCGAGATAGGTTCTTTGAGACGTAGCTTTTT

HDAC1 sh2 TRCN0000004814 CCGGCGTTCTTAACTTTGAACCATACTCGAGTATGGTTCAAAGTTAAGAACGTTTTTsh3 TRCN0000004816 CCGGGCCGGTCATGTCCAAAGTAATCTCGAGATTACTTTGGACATGACCGGCTTTTTsh5 TRCN0000004818 CCGGGCTGCTCAACTATGGTCTCTACTCGAGTAGAGACCATAGTTGAGCAGCTTTTT

NCOR1 sh3 TRCN0000060655 CCGGCGCAGTATTGTCCAAATTATTCTCGAGAATAATTTGGACAATACTGCGTTTTTGsh4 TRCN0000060656 CCGGGCCATCAAACACAATGTCAAACTCGAGTTTGACATTGTGTTTGATGGCTTTTTGsh5 TRCN0000060657 CCGGGCTCTCAAAGTTCAGACTCTTCTCGAGAAGAGTCTGAACTTTGAGAGCTTTTTG

EZH2 sh1 TRCN0000286227 CCGGTATTGCCTTCTCACCAGCTGCCTCGAGGCAGCTGGTGAGAAGGCAATATTTTTGsh2 TRCN0000040074 CCGGGCTAGGTTAATTGGGACCAAACTCGAGTTTGGTCCCAATTAACCTAGCTTTTTGsh4 TRCN0000286290 CCGGCGGAAATCTTAAACCAAGAATCTCGAGATTCTTGGTTTAAGATTTCCGTTTTTG

HDAC2 sh1 TRCN0000004822 CCGGGCAGACTCATTATCTGGTGATCTCGAGATCACCAGATAATGAGTCTGCTTTTTsh2 TRCN0000004823 CCGGGCAAATACTATGCTGTCAATTCTCGAGAATTGACAGCATAGTATTTGCTTTTTsh3 TRCN0000004819 CCGGCAGTCTCACCAATTTCAGAAACTCGAGTTTCTGAAATTGGTGAGACTGTTTTTsh4 TRCN0000004820 CCGGCCAGCGTTTGATGGACTCTTTCTCGAGAAAGAGTCCATCAAACGCTGGTTTTT

MBD2 sh1 TRCN0000013319 CCGGGCCTAGTAAATTACAGAAGAACTCGAGTTCTTCTGTAATTTACTAGGCTTTTTsh2 TRCN0000013320 CCGGGTAGCAATGATGAGACCCTTTCTCGAGAAAGGGTCTCATCATTGCTACTTTTTsh3 TRCN0000013321 CCGGGTACGCAAGAAATTGGAAGAACTCGAGTTCTTCCAATTTCTTGCGTACTTTTTsh5 TRCN0000013318 CCGGGCTTAATGAAAGGGTTTGTAACTCGAGTTACAAACCCTTTCATTAAGCTTTTT

EED sh1 TRCN0000021204 CCGGGCAAACTTTATGTTTGGGATTCTCGAGAATCCCAAACATAAAGTTTGCTTTTTsh2 TRCN0000021205 CCGGCCAGAGACATACATAGGAATTCTCGAGAATTCCTATGTATGTCTCTGGTTTTTsh3 TRCN0000021206 CCGGGCAGCATTCTTATAGCTGTTTCTCGAGAAACAGCTATAAGAATGCTGCTTTTTsh4 TRCN0000021207 CCGGCCTATAACAATGCAGTGTATACTCGAGTATACACTGCATTGTTATAGGTTTTTsh5 TRCN0000021208 CCGGCCAGTGAATCTAATGTGACTACTCGAGTAGTCACATTAGATTCACTGGTTTTT

KDM3A sh1 TRCN0000021149 CCGGCCCAAGATGTATAATGCTTATCTCGAGATAAGCATTATACATCTTGGGTTTTTsh2 TRCN0000021150 CCGGCCCTAATAACTGTTCAGGAAACTCGAGTTTCCTGAACAGTTATTAGGGTTTTTsh3 TRCN0000021151 CCGGGCTGGTATTTAGACCGATCATCTCGAGATGATCGGTCTAAATACCAGCTTTTTsh4 TRCN0000021152 CCGGGCTTTGATTGTGAAGCATTTACTCGAGTAAATGCTTCACAATCAAAGCTTTTTsh5 TRCN0000021153 CCGGCCATACGTTTAACAGCACAATCTCGAGATTGTGCTGTTAAACGTATGGTTTTT

SUZ12 sh2 TRCN0000038725 CCGGGCTTACGTTTACTGGTTTCTTCTCGAGAAGAAACCAGTAAACGTAAGCTTTTTGsh3 TRCN0000038726 CCGGCCAAACCTCTTGCCACTAGAACTCGAGTTCTAGTGGCAAGAGGTTTGGTTTTTGsh4 TRCN0000038727 CCGGCGGAATCTCATAGCACCAATACTCGAGTATTGGTGCTATGAGATTCCGTTTTTGsh5 TRCN0000038728 CCGGGCTGACAATCAAATGAATCATCTCGAGATGATTCATTTGATTGTCAGCTTTTTG

KDM5B sh1 TRCN0000014759 CCGGGCTCCCTTACTTTAGATGATACTCGAGTATCATCTAAAGTAAGGGAGCTTTTTsh2 TRCN0000014760 CCGGCCTCTCCAAGATGTGGATATACTCGAGTATATCCACATCTTGGAGAGGTTTTTsh3 TRCN0000014761 CCGGCCTGAGGAAGAGGAGTATCTTCTCGAGAAGATACTCCTCTTCCTCAGGTTTTTsh4 TRCN0000014762 CCGGCGAGATGGAATTAACAGTCTTCTCGAGAAGACTGTTAATTCCATCTCGTTTTTsh5 TRCN0000014758 CCGGCCCACCAATTTGGAAGGCATTCTCGAGAATGCCTTCCAAATTGGTGGGTTTTT

RBBP4 sh2 TRCN0000115869 CCGGCCCTTGTATCATCGCAACAAACTCGAGTTTGTTGCGATGATACAAGGGTTTTTGsh3 TRCN0000115870 CCGGGCCTTTCTTTCAATCCTTATACTCGAGTATAAGGATTGAAAGAAAGGCTTTTTGsh4 TRCN0000115868 CCGGCGGCAGTAGTAGAAGATGTTTCTCGAGAAACATCTTCTACTACTGCCGTTTTTGsh5 TRCN0000115871 CCGGGCAGACTGAATGTCTGGGATTCTCGAGAATCCCAGACATTCAGTCTGCTTTTTG

KDM5A sh1 TRCN0000014628 CCGGCCAGGTACTTAATGCCCTAAACTCGAGTTTAGGGCATTAAGTACCTGGTTTTTsh2 TRCN0000014629 CCGGCCAGACTTACAGGGACACTTACTCGAGTAAGTGTCCCTGTAAGTCTGGTTTTTsh3 TRCN0000014630 CCGGCGGACCGACATTGGTGTATATCTCGAGATATACACCAATGTCGGTCCGTTTTTsh4 TRCN0000014631 CCGGCCCATGCAGAAGAAATGTCTTCTCGAGAAGACATTTCTTCTGCATGGGTTTTTsh5 TRCN0000014632 CCGGCCTTGAAAGAAGCCTTACAAACTCGAGTTTGTAAGGCTTCTTTCAAGGTTTTT

SOX6 sh1 TRCN0000017990 CCGGCCAGTGAACTTCTTGGAGAAACTCGAGTTTCTCCAAGAAGTTCACTGGTTTTTsh2 TRCN0000017988 CCGGCCAACACTTGTCAGTACCATTCTCGAGAATGGTACTGACAAGTGTTGGTTTTTsh3 TRCN0000017989 CCGGGCCACACATTAAGCGACCAATCTCGAGATTGGTCGCTTAATGTGTGGCTTTTT

Xu et al. www.pnas.org/cgi/content/short/1303976110 9 of 11

Table S1. Cont.

Gene name shRNA ID TRC number shRNA sequences

EZH1 sh1 TRCN0000002439 CCGGGCTACTCGGAAAGGAAACAAACTCGAGTTTGTTTCCTTTCCGAGTAGCTTTTTsh2 TRCN0000002440 CCGGGCTCTTCTTTGATTACAGGTACTCGAGTACCTGTAATCAAAGAAGAGCTTTTTsh3 TRCN0000002441 CCGGCCGCCGTGGTTTGTATTCATTCTCGAGAATGAATACAAACCACGGCGGTTTTTsh4 TRCN0000002442 CCGGCAACAGAACTTTATGGTAGAACTCGAGTTCTACCATAAAGTTCTGTTGTTTTTsh5 TRCN0000010708 CCGGGCTTCCTCTTCAACCTCAATACTCGAGTATTGAGGTTGAAGAGGAAGCTTTTT

IKZF1 sh2 TRCN0000107871 CCGGGCGGAGGATTTACGAATGCTTCTCGAGAAGCATTCGTAAATCCTCCGCTTTTTGsh3 TRCN0000107872 CCGGCCGTTGGTAAACCTCACAAATCTCGAGATTTGTGAGGTTTACCAACGGTTTTTGsh4 TRCN0000107873 CCGGGCCGAAGCTATAAACAGCGAACTCGAGTTCGCTGTTTATAGCTTCGGCTTTTTGsh5 TRCN0000107874 CCGGCGCCAAACGTAAGAGCTCTATCTCGAGATAGAGCTCTTACGTTTGGCGTTTTTG

KDM5D sh1 TRCN0000022114 CCGGCGATCACATTACGAACGCATTCTCGAGAATGCGTTCGTAATGTGATCGTTTTTsh2 TRCN0000022115 CCGGGCCACATTGGAAGCCATAATTCTCGAGAATTATGGCTTCCAATGTGGCTTTTTsh3 TRCN0000022116 CCGGCCAGTGCTAGATCAGTCTGTTCTCGAGAACAGACTGATCTAGCACTGGTTTTTsh4 TRCN0000022117 CCGGCGCGTCCAAAGGCTAAATGAACTCGAGTTCATTTAGCCTTTGGACGCGTTTTTsh5 TRCN0000022118 CCGGCAGCCCTTTCTTGAAAGGAAACTCGAGTTTCCTTTCAAGAAAGGGCTGTTTTT

KDM4B sh1 TRCN0000018014 CCGGGCCCATCATCCTGAAGAAGTACTCGAGTACTTCTTCAGGATGATGGGCTTTTTsh2 TRCN0000018013 CCGGCCGGCCACATTACCCTCCAAACTCGAGTTTGGAGGGTAATGTGGCCGGTTTTTsh3 TRCN0000018015 CCGGGCGGCATAAGATGACCCTCATCTCGAGATGAGGGTCATCTTATGCCGCTTTTTsh4 TRCN0000018016 CCGGGTGGAAGCTGAAATGCGTGTACTCGAGTACACGCATTTCAGCTTCCACTTTTT

KDM7A sh1 TRCN0000253856 CCGGTTAGACCTGGACACCTTATTACTCGAGTAATAAGGTGTCCAGGTCTAATTTTTGsh2 TRCN0000253855 CCGGTATGGGATCAACAGGTATTTACTCGAGTAAATACCTGTTGATCCCATATTTTTGsh3 TRCN0000253854 CCGGTGGATTTGATGTCCCTATTATCTCGAGATAATAGGGACATCAAATCCATTTTTGsh4 TRCN0000253852 CCGGGCAGTTGTATCGCTATGATAACTCGAGTTATCATAGCGATACAACTGCTTTTTGsh5 TRCN0000253853 CCGGAGGCTCCCTTCACCTACATTTCTCGAGAAATGTAGGTGAAGGGAGCCTTTTTTG

RBBP7 sh1 TRCN0000038885 CCGGCCTCCAGAACTCCTGTTTATTCTCGAGAATAAACAGGAGTTCTGGAGGTTTTTGsh2 TRCN0000038886 CCGGCGTGTCATCAATGAAGAATATCTCGAGATATTCTTCATTGATGACACGTTTTTGsh3 TRCN0000038887 CCGGGCACAGTTTGATGCTTCCCATCTCGAGATGGGAAGCATCAAACTGTGCTTTTTGsh5 TRCN0000038884 CCGGCGTTTCTATATGACCTGGTTACTCGAGTAACCAGGTCATATAGAAACGTTTTTG

KDM4C sh1 TRCN0000022054 CCGGGCCCAAGTCTTGGTATGCTATCTCGAGATAGCATACCAAGACTTGGGCTTTTTsh2 TRCN0000022055 CCGGCCTTGCATACATGGAGTCTAACTCGAGTTAGACTCCATGTATGCAAGGTTTTTsh3 TRCN0000022056 CCGGGCCTCTGACATGCGATTTGAACTCGAGTTCAAATCGCATGTCAGAGGCTTTTTsh4 TRCN0000022057 CCGGGCACCTATCTATGGTGCAGATCTCGAGATCTGCACCATAGATAGGTGCTTTTT

BAF155 sh1 TRCN0000015628 CCGGGCAGGATATTAGCTCCTTATACTCGAGTATAAGGAGCTAATATCCTGCTTTTTsh2 TRCN0000015629 CCGGCCCACCACATTTACCCATATTCTCGAGAATATGGGTAAATGTGGTGGGTTTTTsh3 TRCN0000015630 CCGGGCTATGATACTTGGGTCCATACTCGAGTATGGACCCAAGTATCATAGCTTTTTsh5 TRCN0000015632 CCGGCCTAGCTGTTTATCGACGGAACTCGAGTTCCGTCGATAAACAGCTAGGTTTTT

LSD1 (KDM1A) sh1 TRCN0000046068 CCGGGCCTAGACATTAAACTGAATACTCGAGTATTCAGTTTAATGTCTAGGCTTTTTGsh5 TRCN0000046072 CCGGCCACGAGTCAAACCTTTATTTCTCGAGAAATAAAGGTTTGACTCGTGGTTTTTG

ZBTB33 sh1 TRCN0000017838 CCGGCCCTTCCATGTTAGCACTTTACTCGAGTAAAGTGCTAACATGGAAGGGTTTTTsh2 TRCN0000017840 CCGGCGGTGAAGATACTTATGATATCTCGAGATATCATAAGTATCTTCACCGTTTTT

NCOR2 sh2 TRCN0000060704 CCGGCCTCTATTACTACCTGACTAACTCGAGTTAGTCAGGTAGTAATAGAGGTTTTTGsh5 TRCN0000060707 CCGGGCAGTGTAAGAACTTCTACTTCTCGAGAAGTAGAAGTTCTTACACTGCTTTTTG

KDM3B sh1 TRCN0000017093 CCGGCCCTAGTTCATCGCAACCTTTCTCGAGAAAGGTTGCGATGAACTAGGGTTTTTsh2 TRCN0000017095 CCGGGCGATCTTTGTAGAATTTGATCTCGAGATCAAATTCTACAAAGATCGCTTTTTsh3 TRCN0000017096 CCGGGCTGTTAATGTGATGGTGTATCTCGAGATACACCATCACATTAACAGCTTTTTsh4 TRCN0000017097 CCGGCCTTGTAGATAAACTGGGTTTCTCGAGAAACCCAGTTTATCTACAAGGTTTTT

KDM4A sh1 TRCN0000013493 CCGGGCTGCAGTATTGAGATGCTAACTCGAGTTAGCATCTCAATACTGCAGCTTTTTsh3 TRCN0000013495 CCGGGCACCGAGTTTGTCTTGAAATCTCGAGATTTCAAGACAAACTCGGTGCTTTTTsh4 TRCN0000013496 CCGGCCGAAACTTCAGTAGATACATCTCGAGATGTATCTACTGAAGTTTCGGTTTTTsh5 TRCN0000013497 CCGGGCCTTGGATCTTTCTGTGAATCTCGAGATTCACAGAAAGATCCAAGGCTTTTT

BRG1 sh1 TRCN0000015548 CCGGCCATATTTATACAGCAGAGAACTCGAGTTCTCTGCTGTATAAATATGGTTTTTsh2 TRCN0000015549 CCGGCCCGTGGACTTCAAGAAGATACTCGAGTATCTTCTTGAAGTCCACGGGTTTTTsh4 TRCN0000015551 CCGGCCGAGGTCTGATAGTGAAGAACTCGAGTTCTTCACTATCAGACCTCGGTTTTTsh5 TRCN0000015552 CCGGCGGCAGACACTGTGATCATTTCTCGAGAAATGATCACAGTGTCTGCCGTTTTT

BCOR sh1 TRCN0000033460 CCGGGCCAAATAAGTATTCACTGAACTCGAGTTCAGTGAATACTTATTTGGCTTTTTGsh2 TRCN0000033461 CCGGCCACGAAACTTATACTTTCAACTCGAGTTGAAAGTATAAGTTTCGTGGTTTTTGsh3 TRCN0000033462 CCGGGCTCTATATTTCTGTCTCCAACTCGAGTTGGAGACAGAAATATAGAGCTTTTTGsh5 TRCN0000033459 CCGGCCCGCATATTTCGCTGCAATTCTCGAGAATTGCAGCGAAATATGCGGGTTTTTG

TRIM28 sh2 TRCN0000017999 CCGGCCTGGCTCTGTTCTCTGTCCTCTCGAGAGGACAGAGAACAGAGCCAGGTTTTTsh3 TRCN0000018001 CCGGCTGAGACCAAACCTGTGCTTACTCGAGTAAGCACAGGTTTGGTCTCAGTTTTT

The gene name, TRC number (Sigma-Aldrich), and sequences of shRNAs are shown.

Xu et al. www.pnas.org/cgi/content/short/1303976110 10 of 11

Table S2. List of primers used in this study

Primer name Species Region/gene Application Primer sequence (5′→3′)Productsize (bp)

a-fwd Human HS3 ChIP ATAGACCATGAGTAGAGGGCAGAC 142a-rev TGATCCTGAAAACATAGGAGTCAAb-fwd Human HBE1 ChIP GCCAGAACTTCGGCAGTAAA 98b-rev GGCCTGAGAGCTTGCTAGTGc-fwd Human HBG1 ChIP TTACTGCGCTGAAACTGTGG 130c-rev CAGTGGTTTCTAAGGAAAAAGTGCd-fwd Human HBG1 +3kb ChIP AATGACCTAATGCCCAGCAC 80d-rev AGTGTTGGGGGAGAAGTGTGe-fwd Human HBD -1kb ChIP GCAACAGAAGCCCAGCTATT 104e-rev GTGGCATGGTTTGATTTGTGf-fwd Human HBD ChIP TGTAGAGGAGAACAGGGTTT 181f-rev CTGCCTTTTATGCTGGTCCTg-fwd Human HBB ChIP TGCTCCTGGGAGTAGATTGG 161g-rev TGGTATGGGGCCAAGAGATAh-fwd Human 3′HS1 ChIP TCTTCAGCCATCCCAAGACT 137h-rev TGGTCTTTTCTGGACACCACHBE1-RT2-fwd Human HBE1 RT-PCR GCAAGAAGGTGCTGACTTCC 142HBE1-RT2-rev ACCATCACGTTACCCAGGAGHBG-RT-fwd Human HBG1/HBG2 RT-PCR TGGATGATCTCAAGGGCAC 209HBG-RT-rev TCAGTGGTATCTGGAGGACAHBB-RT-fwd Human HBB RT-PCR CTGAGGAGAAGTCTGCCGTTA 146HBB-RT-rev AGCATCAGGAGTGGACAGAThGAPDH-RT-fwd Human GAPDH RT-PCR ACCCAGAAGACTGTGGATGG 125hGAPDH-RT-rev TTCAGCTCAGGGATGACCTTey-RT-fwd Mouse Hbb-y RT-PCR TGGCCTGTGGAGTAAGGTCAA 120ey-RT-rev GAAGCAGAGGACAAGTTCCCAbh1-RT-fwd Mouse Hbb-bh1 RT-PCR TGGACAACCTCAAGGAGACC 145bh1-RT-rev ACCTCTGGGGTGAATTCCTTbmaj/min-RT-

fwdMouse Hbb-b1/Hbb-

b2RT-PCR TTTAACGATGGCCTGAATCACTT 132

bmaj/min-RT-rev CAGCACAATCACGATCATATTGCmGapdh-RT-fwd Mouse Gapdh RT-PCR TGGTGAAGGTCGGTGTGAAC 123mGapdh-RT-rev CCATGTAGTTGAGGTCAATGAAGG

fwd, forward; rev, reverse.

Xu et al. www.pnas.org/cgi/content/short/1303976110 11 of 11

View publication statsView publication stats