www.sciencemag.org/content/351/6270/285/suppl/DC1

Supplementary Materials for

Transcription factors LRF and BCL11A independently repress expression of

fetal hemoglobin

Takeshi Masuda, Xin Wang, Manami Maeda, Matthew C. Canver, Falak Sher, Alister P. W. Funnell, Chris Fisher, Maria Suciu, Gabriella E. Martyn, Laura J. Norton,

Catherine Zhu, Ryo Kurita, Yukio Nakamura, Jian Xu, Douglas R. Higgs, Merlin Crossley, Daniel E. Bauer, Stuart H. Orkin, Peter V. Kharchenko,* Takahiro Maeda*

*Corresponding author. E-mail: peter.kharchenko@post.harvard.edu (P.V.K.); tmaeda@partners.org (T.M.)

Published 15 January 2016, Science 351, 285 (2016) DOI: 10.1126/science.aad3312

This PDF file includes: Materials and Methods

Supplementary Text Figs. S1 to S14

References

Materials and MethodsMiceHematopoietic-specific conditional Zbtb7a knockout mice (Zbtb7aF/F Mx1-Cre+) were described previously (6, 26). Zbtb7aF/F Mx1-Cre+ mice were bred with the β-YAC mice (12) to establish Zbtb7a conditional KO mice harboring the human β-globin gene cluster as a yeast artificial chromosome transgene. pIpC was injected intraperitoneally at 8 weeks of age, when human γ-globin expression is silenced in this model (13).

Cell sortingTo collect splenic erythroblasts for RNA-Seq, mice were treated with phenylhydrazine and splenocytes harvested as described (27). After preparing single-cell suspensions using a cell strainer (BD), splenocytes were incubated with red cell lysis buffer and then incubated with a mixture of biotin-conjugated antibodies (anti-CD11b, Gr-1, CD19, B220, CD3, CD4 and CD8), followed by incubation with anti-biotin MicroBeads (Miltenyi Biotec). Cell suspensions were subsequently applied onto MACS separation LD columns (Miltenyi Biotec) for negative selection. Cells were then incubated with fluorochrome-conjugated anti-CD71, -TER119 and -CD44 antibodies and subjected to sorting. Cell sorting was performed at the Children's Hospital Boston Flow Cytometry Core Facility (CHBFCC) with an AriaIII cytometer using DAPI for live/dead discrimination

Primary human erythroid culture G-CSF-mobilized adult human PB CD34+ cells were obtained from the Hematopoietic Cell Processing Core at Fred Hutchinson Cancer Center, and erythroid differentiation induced as described with minor modifications (13, 22). The erythroid culture system consists of three phases using three different erythroid differentiation media (EDM-I, -II and -III) (fig. S3A). The composition of basal EDM is: IMDM (Cellgro) supplemented with 1% L-glutamine (Life Technologies), 2% penicillin-streptomycin (Life Technologies), holo-human transferrin (330 µg/ml, Sigma), heparin (2 IU/ml, Sigma), recombinant human insulin (10 µg/ml, Sigma), erythropoietin (3 IU/ml, Amgen) and 5% inactivated human plasma (Rhode Island Blood Center ). After thawing, cells were resuspended in EDM-I [EDM supplemented with hydrocortisone (10-6M, Sigma), hSCF (100 ng/ml, R&D) and hIL3 (5 ng/ml, R&D)] and cultured for 7 days (phase I). On day 7, medium was replaced with EDM-II [EDM supplemented with hSCF (100ng/ml)] (phase II) and, on day11 medium was switched to EDM-III (EDM with no supplement) (phase III).

HUDEP-2 cellsHUDEP-2 cells were maintained with StemSpan SFEM medium (Stem Cell Technologies) supplemented with hSCF (50 ng/ml), erythropoietin (3 IU/ml), dexamethasone (10-6M, Sigma) and doxycycline (dox: 1 µg/ml, Sigma) (15). To induce differentiation, medium was replaced with EDM-II and cells were cultured for 5 days in

2

the presence of dox (fig. S5A). If necessary, cells were cultured for 2 more days without dox and used for experiments (e.g. HPLC, Hb-FACS).

Generation of ZBTB7AΔ/Δ, BCL11AΔ/Δ and ZBTB7AΔ/ΔBCL11AΔ/Δ HUDEP-2 cellsWe first established a HUDEP-2 line stably expressing Cas9 endonuclease (HUDEP-2_Cas9) using a lentivirus vector encoding Cas9 and a Blasticidin-resistance cassette (lentiCas9-Blast; Addgene plasmid ID 52962). To generate ZBTB7A or BCL11A KO HUDEP-2 cells, sgRNA sequences (ATCGGGATCCCGTTCCCCGA or TGAACCAGACCACGGCCCGT) were designed to target ZBTB7A or BCL11A, respectively (28). sgRNA-specifying oligos were subcloned into the lentiGuide-Puro vector (Addgene plasmid ID 52963). Lentivirus transduction was performed as previously described (27). Puromycin- and blasticidin-resistant cells were harvested 2 weeks after transduction and knockout was confirmed by Western blot. Independent LRF (or BCL11A) KO clones were established via limiting dilution as described (29). DKO HUDEP-2 cells (ZBTB7AΔ/ΔBCL11AΔ/Δ) were generated by targeting the BCL11A gene in ZBTB7AΔ/Δ HUDEP-2 cells.

Hemoglobin FACSCells were fixed with 4% paraformaldehyde at room temperature (RT) for 20 min, washed with PBS and permeabilized with 0.1% Triton X-100 (for 10 min at RT). After a PBS wash, cells were first incubated with primary antibody (anti-HbF or non-specific IgG) for 30 min, washed with PBS and then incubated with PE/Cy7-conjugated secondary antibody for 30 min. Incubation with antibodies was performed at RT on a rocking shaker. Antibodies used were: anti-Human Fetal Hemoglobin (2D12, BD), non-specific mouse IgG1κ (555748, BD) and PE/Cy7-conjugated anti-mouse IgG1 antibody (406613, Biolegend).

Hemoglobin HPLCThree to five million HSPC-derived erythroblasts or HUDEP cells were spun down, and cell pellets were subjected to HPLC. Over the course of this study, we measured HbF levels using a G7 HPLC Analyzer (TOSOH BIOSCIENCE, INC) in two different modes (hemoglobin A1c or beta-thalassemia programs). To compare HbF levels across experiments (performed in different modes), we defined %HbF as a proportion of HbF relative to HbA0.

Antibodies for FACS Antibodies were purchased from Biolegend, eBioscience, BD or Miltenyi Biotec. Fluorochrome-conjugated antibodies included: TER119 (TER-119), mouse CD44 (IM7), mouse CD71 (R17217), human CD235a (HIR2), human CD71 (CY1G4), human CD36 (5-271) and human α4-integrin (MZ18-24A9). The following biotin-conjugated antibodies were used for negative selection: Biotin-CD11b (M1/70), Biotin-Gr1 (RB6-8C5), Biotin-B220 (RA3-6B2), Biotin-CD19 (eBio1D3), Biotin-CD3 (145-2C11), Biotin-CD4 (L3T4) and Biotin-CD8 (eBio-H35-17.2).

3

Isoelectric focusing of PB hemolysates and MALDI-TOFMS analysis Hemoglobin was analyzed by isoelectric focusing (IEF) (Resolve; PerkinElmer) as per manufacturer’s instructions running for 45minutes at 300volts at 15°C (30). Focused hemoglobins were excised from the IEF gel, digested with trypsin (31) and analyzed using a matrix-assisted laser desorption/ionization time-of flight (MALDI-TOF/TOF) mass spectrometer (Ultralex Bruker Daltonics) (32).

Western blot analysisWestern blot analysis was performed using standard methodology with the following antibodies: LRF (13E9, eBioscience), BCL11A (ab19487, Abcam), GATAD2B (A301-283, Bethyl), MTA2 (ab8106, Abcam), HDAC1 (ab19845, Abcam), HDAC2 (ab32117, Abcam), GAPDH (sc-25778, Santa Cruz Biotechnology), HSP90 (clone: 68, BD), αTubulin (DM1A, Sigma), HBG1/2 (Hemoglobin γ: sc-21756, Santa Cruz Biotechnology) and Flag (M2, Sigma).

Wright-Giemsa stainWright-Giemsa staining was performed using PROTOCOL Wright-Giemsa Stain Solutions (Fisher Scientific) following the manufacturer’s specifications.

PlasmidsLentivirus vectors expressing shRNA-against human ZBTB7A (pLKO.1-shRNA vectors) were purchased from Sigma. Two independent shRNA clones were used: TRCN0000137891 (ZBTB7A clone #2) and TRCN0000136851 (ZBTB7A clone #4). The following plasmids were used for Co-IP experiments: pCMV-Tag2B-Flag-p66beta (human GATAD2B/p66β cDNA was obtained from the pcDNA3-mCherry-p66beta vector from Dr. Rainer Renkawitz), pcDNA3 N-Flag-CHD8 (from Dr. Keiichi Nakayama), pCI-neo3-Flag-hCHD3 (from Dr. Aaron Goodarzi), pcDNA3.1-hACF1-Flag (from Dr. Patrick Varga-Weisz) and pEF1-hLRF-V5His-Neo. For immunoprecipitation in MEL cells, we used the pMX-mCherry-FHC-mLRF retrovirus vector, in which Flag- and HA-tags were introduced at the 3’ end of mouse LRF cDNA (fig. S12C).

Real time PCR assayRNA was extracted using TRIZOL (Life Technologies) and treated with DNase I (Life Technologies). cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (ABI) and Real-time PCR was performed using Fast EvaGreen qPCR Master Mix (Biotium) and a ABI7900HT system (ABI). Transcript levels were calculated relative to corresponding β-actin (mouse) or RPS18 (human) levels in each sample. Primer sequences for qRT-PCR were: mouse β-actin FW: 5’-ACCCTAAGGCCAACCGTGA-3’, mouse β-actin RV: 5’-GTCTCCGGAGTCCATCACAA-3’, mouse Hbb-bh1 FW: 5’-CTCAAGGAGACCTTTGCTCA-3’, mouse Hbb-bh1 RV: 5’-AATCACCAGCTTCTGCCAGGC-3’, mouse Hbb-bs/bt FW: 5’-TTTAACGATGGCCTGAATCACTT-3’,

4

mouse Hbb-bs/bt RV: 5’-CAGCACAATCACGATCATATTGC-3’, human HBB FW: 5’-AGGAGAAGTCTGCCGTTACTG-3’, human HBB RV: 5’-CCGAGCACTTTCTTGCCATGA-3’, human HBG1/2 FW: 5’-AACCCCAAAGTCAAGGCACA-3’, human HBG1/2 RV: 5’-CATCTTCTGCCAGGAAGCCT-3’, human HBE1 FW: 5’-GAGAGGCAGCAGCACATATC-3’, human HBE1 RV: 5’-CAGGGGTAAACAACGAGGAG-3’,human RPS18 FW: 5’-GTAACCCGTTGAACCCCATT-3’, human RPS18 RV: 5’-CCATCCAATCGGTAGTAGCG-3’.

Retrovirus transductionTo transduce MEL cells, 293T cells were transfected with the pMX-mCherry-FHC-mLRF vector and the pMCV-Ecopac vector using Lipofectamine 2000, and virus-containing supernatants were collected. 2×105 MEL cells were spin-infected with 1 ml viral supernatant containing polybrene (5 µg/mL) and then selected in puromycin (2 µg/mL) starting 48 hours later.

Generation and sequencing of RNA-Seq librariesTotal RNA was extracted from FACS-sorted splenic erythroblasts (or HUDEP-2 cells) using RNeasy Mini Kit (QIAGEN). 125 ng of purified RNA was used for library construction. Libraries were generated using the Encore Complete RNA-Seq DR Multiplex System following the manufacturer’s specifications (Nugen). All libraries underwent 50 or 75 bp paired-end sequencing on an Illumina HiSeq 2000 or 2500.

RNA-Seq data analysisSequenced reads were aligned to the UCSC hg19 (human, mm9 mouse) annotation using tophat2 aligner, and read counts were quantified using HTSeq (33). Differential expression analysis was carried out using DESeq Bioconductor package (34). Hba-a1/Hba-a2 and Hbb-b1/Hbb-b2 reads were pooled during the alignment stage to estimate their combined expression magnitude. GSEA of differentially expressed genes was carried out using log2 fold-change values, restricting the gene set to those observed with an FPM (Fragments Per Million mapped reads) value of >1.0 in at least one condition, using a power factor of 1.

Generation and sequencing of ChIP-Seq librariesLRF-ChIP-Seq experiments were performed using HSPC-derived erythroblasts (Day8 erythroblasts) in duplicate and undifferentiated HUDEP-2 cells in triplicate. Five million cells were fixed with 1% formaldehyde at RT for 20 min and chromatin was collected using a truChIP High Cell Chromatin Shearing Kit with SDS (Covaris). Sonication was performed using a Covaris S2 instrument (Covaris). The chromatin solution was first incubated with 40 µl protein A/G Dynabeads (Life Technologies) at 4oC for 3h on a rotating shaker to prevent non-specific binding. ChIP was then performed using 1µg anti-LRF antibody (clone13E9, eBioscience) at 4oC overnight on a rotating shaker. The

5

antibody/protein complex was harvested following incubation with 15 µL of BSA-blocked protein A/G Dynabeads at 4oC for 1h. Beads were sequentially washed once with the following buffers: wash buffer A (10 mM Tris-HCl, pH7.4, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.1% SDC and 1 mM DTT), wash buffer B (10 mM Tris-HCl, pH7.4, 1 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.1% SDC, 300 mM NaCl and 1 mM DTT), wash buffer C (250 mM LiCl, 0.5% NP40 and 0.5% SDC) and finally TE buffer. Complexes were then eluted from beads in 50 mM Tris-HCl (pH8.0)/1% SDS/10 mM EDTA. After reverse cross-linking (at 65oC overnight) and proteinase K treatment, DNA was extracted using a QIAquick PCR purification kit (QIAGEN). ChIP-Seq libraries were generated using a Ovation Ultralow DR Multiplex System 1-8 (NuGEN) and sequenced. All libraries were sequenced by 50 bp single-end or paired-end reads using an Illumina HiSeq 2000 or 2500 system.

ChIP-Seq data analysisChIP-Seq reads were aligned (hg18 assembly for human, mm9 assembly for mouse) using bowtie2. Smoothed read density, enrichment profiles and peak calls were generated using the spp package (35). A consistent set of LRF binding positions was determined using IDR procedure (IDR=0.05). Sequence motifs were determined using the MEME package (JASPAR_CORE and uniprobe motifs) using +/-100bp regions around the top 500 sites. To better resolve LRF enrichment within the γ-globin locus, we performed paired-end alignment and estimated conservative enrichment profiles (lower bound of 95% confidence interval; see spp package (35)) using i) all mappable fragments and ii) only uniquely-mappable fragments (bottom tracks, Fig. 3).

Generation and sequencing of ATAC-Seq librariesATAC-Seq libraries were constructed using 7.5x104 HUDEP-2 cells per sample as described (18, 36). After a transposition reaction, transposed DNA fragments were purified using a MinElute Kit (Qiagen) and PCR-amplified using PCR primer1 (Ad1_noMX) and a barcoded PCR primer2 (Ad2) (18, 36). PCR conditions used were: 72ºC for 5 min, 98ºC for 30 s, followed by 11 cycles of 98ºC for 10 s, 63ºC for 30 s and 72ºC for 1 min. Amplified libraries were purified with a MinElute Kit and library quality was assessed using the TapeStation system (Agilent). All libraries were sequenced by 50 bp paired-end reads using the Illumina HiSeq 2500 system.

ATAC-Seq data analysisSequenced reads were aligned using bowtie2, and PCR duplicates were removed using the “samtools rmdup” command. Aligned fragments with length

Yeast two-hybrid screen (Y2H)Α Y2H screen (ULTimate Y2HTM, Hybrigenics) was performed with the human LRF-BTB domain (aa 1-131) as bait using a cDNA library constructed from human B-cell lymphoma cell lines (SUDHL-4, SUDHL-6 and RCK8). The human LRF-BTB domain coding sequence (aa 1-131) was PCR-amplified and subcloned into the pB27 (N-LexA-bait-C fusion) vector (Hybrigenics). The interaction assay uses a His3 reporter that allows yeast to grow in medium lacking histidine (37). Autoactivation of the bait fragment was tested in the presence of 3-aminotriazole (3-AT). A total of 52.1 million interactions were analyzed and 360 positive clones processed for analysis.

ImmunoprecipitationImmunoprecipitation was performed using nuclear protein extracts of HSPC-derived erythroblasts (day 8) or MEL cells. Nuclei were isolated by incubating cell pellets (3x107) in lysis buffer A (10 mM HEPES-NaOH, pH7.9,10 mM KCl, 1.5 mM MgCl2 and 1 mM DTT) for 10 min on ice. During incubation, the solution was mixed by 10 strokes of gentle pipetting. Nuclear pellets were fixed with 3% formaldehyde (30 min at RT), washed with PBS 4 times, and incubated 10 min on ice in lysis buffer B (10 mM HEPES-NaOH, pH7.9, 0.2% SDS, 0.1% SLS, 2 mM EDTA, 1 mM EGTA and 50 mM NaCl). After brief sonication (Bioruptor, Diagenode), nuclear extracts were collected and subjected to immunoprecipition using the following antibodies: LRF (clone 13E9, eBioscience), GATAD2B (A301-282, Bethyl) and MTA2 (ab8106, Abcam).

7

Supplementary TextEffects of LRF depletion on human erythroid differentiationEffects of LRF knockdown/knockout on human terminal erythroid differentiation (HSPC-derived erythroblasts and HUDEP-2 cells) were milder than those seen in LRF conventional (Zbtb7a-/-) or conditional (Zbtb7aF/F Mx1-cre+) (14) knockout mice. There are a few plausible explanations for this outcome. For example, effects of LRF depletion on erythroid differentiation may be more severe in vivo due, in part, to non-cell autonomous effects. In fact, we observed only marginal overlap of differentially-expressed genes between mouse erythroblasts (in vivo) and HUDEP-2 cells (in vitro) in our RNA-Seq data sets. Only fetal-type β-globin (Hbb-bh1 or HBG1/2), embryonic α-globin (Hba-x or HBZ) and few other genes were differentially expressed in both systems. For instance, Dll4 upregulation, which causes aberrant T-cell development in bone marrow of Zbtb7aF/F Mx1-cre+ mice (6), was only observed in LRF-deficient mouse splenic erythroblasts. Collectively, differences between two experimental systems (a mouse model versus erythroid culture) could account for the milder phenotype observed in human erythroid cells. Mouse and human LRF proteins exhibit high amino acid homology (100% identical in the BTB domain and 89% in remaining amino acid sequences) and bind a very similar DNA sequence. Specifically, LRF-CAST analysis, which we performed using recombinant mouse LRF protein (16), and LRF-ChIP-Seq, performed in human erythroid cells (Fig. S9 and S10), identified an almost identical LRF consensus sequence. Thus, mouse and human LRF may recruit similar transcriptional modulators through the BTB domain and bind to homologous sequences in the genome. Differences in mouse and human DNA sequence, namely those found in regulatory sequences (38-40), could account for observed phenotypic differences. We observed incomplete differentiation in ShRNA-transduced HPSC-derived erythroblasts, including those transduced with scrambled ShRNA, by FACS (Fig. S4A) and morphology (Fig. S4B). However, morphological differences among ShRNA-transduced cells (scrambled, ShRNA#2 and #4) were not evident. HbF levels in LRF KD HSPC-derived erythroblastsWe consistently observed relatively high HbF levels in scrambled shRNA-transduced cells (%HbF=15-26%), presumably due to non-specific shRNA effects (41) and/or cellular stress induced by Puromycin (42, 43). Nonetheless, the effect of LRF KD (%HbF=45-70%), was as significant as those seen in BCL11A KD erythroblasts (21). In that study, erythroblasts subjected to ShRNA-mediated BCL11A KD reportedly exhibited HbF levels 25.6-35.9%, while non-specific effects of shRNA expression were not determined. Since LRF KD efficiencies were not 100% (Fig. S3B), LRF remaining after knockdown may still repress γ-globin expression to some degree.

Molecular mechanisms underlying LRF-mediated γ-globin repressionWe used paired-end alignment to estimate LRF enrichment near γ-globin genes. Furthermore, given the high sequence similarity of the HBG1 and HBG2 genes, we

8

analyzed LRF occupancy sites around the γ-globin genes using two different read alignment methods: one mapping all mappable fragments (”LRF all” track in Fig.3) and the other mapping only uniquely-mappable fragments (“LRF uniquely-mapped” track in Fig.3). In the “LRF all” track, fragments mappable to either HBG1 or HBG2 were randomly distributed between both genes (depicted with asterisks in Fig. 3). We observed prominent LRF ChIP-Seq signals in the bodies of genes and immediately downstream of HBB, HBG1 and the LCR when we used either alignment method. However, we cannot rule out that LRF may bind to HBG2. Although we don't know exactly how LRF represses γ-globin expression following binding to immediately downstream of HBG1, there are a few plausible explanations. Since γ-globin induction occurs with concomitant reduction in levels of adult β-globin transcripts in LRF KO HUDEP cells (fig S7A), γ-globin reactivation presumably occurs through a LCR-dependent mechanism. LRF depletion from downstream HBG1 sequences may facilitate a long-range interaction between γ-globin loci and the LCR (44-46). γ-globin expression in LRF KO HUDEP cells becomes more pronounced upon erythroid differentiation (Fig. 3 and Fig. S7A), suggesting that it is positively regulated by erythroid-specific transcription factors. We propose that LRF depletion is prerequisite for activators to gain access to this locus by opening-up local chromatin. Consistent with this idea, there is a small, but noticeable ATAC-Seq peak downstream of HBG1 that overlaps with the LRF occupancy site in undifferentiated WT HUDEP cells (Fig. 3, top row). That region exhibits highest ATAC-Seq signals at γ-globin region in LRF KO HUDEP cells under undifferentiated conditions (Fig.3, second row from top), suggesting that local chromatin downstream of HBG1 would first become accessible upon LRF inactivation. Since that region was previously identified as the “HBG1 3’enhancer” (47, 48), it is tempting to speculate that LRF resides at the HBG1 3’ enhancer, presumably together with the NuRD complex and interferes with enhancer activity in adult erythroid cells (e.g, by physically blocking looping between the LCR and the enhancer). It would also be interesting to determine localization of NuRD components in the presence or absence of LRF. To our knowledge, there are no reports describing ChIP-Seq experiments using antibodies against GATAD2B or CHD3, two LRF-binding NuRD components. Furthermore, few reports describe ChIP-seq using antibodies against NuRD components (49, 50). It should be noted that knockdown of CHD4, a NuRD component, reportedly exhibit strongest HbF reactivation activity in HSPC-derived erythroblasts (22). In addition to the γ-globin gene, LRF occupies the HBB gene and LCR HS sites at the β-globin cluster (Fig. 3). It is possible that LRF suppresses γ-globin expression on the one hand and positively regulates the HBB gene on the other. The NuRD complex may mediate such “dual” activities of LRF as previously reported (51). Alternatively, LRF may regulate HBB transcription through chromatin remodeling factors such as SRCAP and BAZ1A, two LRF-interacting proteins identified by our Y2H screen (Fig. S11).

Specificity of the anti-LRF antibody

9

The antibody we used for LRF ChIP-Seq is a monoclonal hamster anti-LRF antibody (clone 13E9) that we generated (16). It recognizes aa1-20 (MAGGVDGPIGIPFPDHSSDIC) within the LRF-BTB domain, a region conserved between mouse and human. Two strong pieces of evidence attest to its specificity. First, we did not detect any signals in LRF knockout mouse cells, including in MEFs (16), erythroblasts (6, 14), myeloid cells (14, 26) or lymphocytes (26, 52) based on western blot, IP/western or immunohistochemistry analysis. Also we detected no signals with this antibody in LRF-deficient cells analyzed in this study, namely, LRF knockout mouse splenic erythroblasts (Fig. S1A) or HUDEP-2 cells (Fig. S7E and S12D). Collectively, our previous reports (6, 14, 16, 26, 52) and this study consistently demonstrate specificity of this antibody. Antibody specificity was also validated by the fact that the most enriched sequence in LRF ChIP-Seq experiments with high statistical significance (e-value 4.1e-71) was concordant with the one we previously identified by CAST (cyclic amplification and selection of targets) analysis (16). CAST analysis (in vitro) and LRF ChIP-seq (in vivo) are completely independent experiments that employed two different antibodies, yet they identified a nearly identical sequence. This finding further confirms the specificity of the anti-LRF antibody that we used for ChIP-Seq.

Effects of LRF depletion in non-globin genesWe previously showed that LRF represses expression of the pro-apoptotic factor Bim in fetal liver (FL) erythroblasts and splenic erythroblasts (14). We also reported that the tumor suppressor p19Arf is upregulated in LRF-deficient mouse embryonic fibroblasts (MEFs) (16) and germinal center B (GCB) cells (52). Importantly, LRF suppresses p19Arf expression in a cell-type specific fashion: p19Arf upregulation was not observed in LRF-deficient erythroblasts, and neither p19Arf loss nor Tp53 deletion prevents LRF-deficient FL erythroblasts from undergoing apoptosis (14). Consistent with our previous report, we observed upregulation of Bim (Bcl2l11) transcript levels in our RNA-Seq data set, while p19Arf transcripts remained unchanged. While we did not examine cell cycle status specifically, we are not aware of significant defects in proliferation or viability in human erythroid cultures upon LRF knockdown/knockout. In fact, gene expression signatures relevant to apoptosis or proliferation were not evident from GSEA analysis (Fig. S14).

LRF as a therapeutic target for hemoglobinopathiesTargeting LRF for treatment of hemoglobinopathies will be challenging given its broad function in normal hematopoietic development observed in LRF knockout mice (6, 14, 26, 52). However, since our study identifies major factors and/or complexes necessary for fetal globin silencing, our findings may be relevant to future drug design. For example, one approach to target LRF for HbF reactivation would be to interfere with interaction between the LRF-BTB domain and GATAD2B/CHD3, as previously reported for the BCL6-BTB domain (53-56). Dr. Merlnick’s group reported a series of peptides and small molecules that physically block protein-protein interactions between BCL6-BTB and

10

SMRT/BCoR corepressors (53-56). Although it is possible that blocking LRF/GATAD2B/CHD3 interaction could broadly alter LRF function, it is worth exploring this strategy in the future.

Author ContributionsContribution: A.P.W.F and M.C. initiated this project. T.M., J.X., M.C., D.R.H., S.H.O., D.E.B. and T.M. designed mouse studies and human erythroid culture experiments. T.M., C.Z., G.E.M., L.J.N., A.P.W.F. and M.M. executed mouse experiments, RNA-Seq, ChIP-Seq, ATAC-Seq, western blot and immunoprecipitation. T.M., M.M., M.C.C. and F.S. executed human erythroid culture experiments and CRISPR-mediated LRF/BCL11A KO supervised by S.H.O., D.E.B. and T.M. C.F. and M.S. executed mouse globin protein analysis supervised by D.R.H. R.K. and Y.N. provided the HUDEP-2 cell line. X.W, T.M. and P.V.K. designed and performed data analysis for ChIP-Seq, RNA-Seq and ATAC-Seq experiments. T.M., P.V.K. and T.M. wrote the manuscript with help from all authors.

11

Fig. S1. Efficient Zbtb7a deletion in splenic erythroblasts of Zbtb7aF/F Mx1-Cre+ mice. (A) LRF protein was not detected by Western blot in splenic erythroblasts of Zbtb7aF/F Mx1-Cre+ mice. Splenic erythroblasts were obtained one month after pIpC injection. Hdac1: loading control. (B) RNA-Seq read profiles at the Zbtb7a locus. No signals from exon2 were detected in LRF KO samples, verifying efficient Zbtb7a deletion. Two loxp sites were introduced to flank Zbtb7a exon2 in our Zbtb7a conditional knockout model (26).

12

AZbtb7aF/F Mx1-Cre+Control

#1 #2 #3 #4 #5 #6

LRF

Hdac1

B

mouse ID

Zbtb7aF/F Mx1-Cre+ #1

Zbtb7aF/F Mx1-Cre+ #2

Zbtb7aF/F Mx1-Cre- #1

Zbtb7aF/F Mx1-Cre- #2

loxp

21 3 Floxed

Deleted31

frt

RNA-Seq (Basophilic erythroblasts)

Zbtb7a

Fig. S2. RNA-Seq analysis of LRF-deficient mouse splenic erythroblasts. (A) Bar graphs show levels of embryonic- and adult-globin transcripts in control (F/F) and LRF KO (F/F Cre+) basophilic- and polychromatophilic (CD71+TER119+CD44medFSCmed)-erythroblasts. y-axis represents mean FKPM values obtained from RNA-Seq experiments (two independent samples per genotype). Values in parentheses indicate mean FKPM values of less abundant transcripts. P values: unadjusted P values determined using the DESeq Bioconductor package (34). (B) Relative Hbb-bh1 mRNA levels in splenic erythroblasts as determined by q-PCR. Hbb-bh1 transcript levels in LRF-deficient erythroblasts were 1714x higher than those seen in controls. (C) LRF inactivation in β-YAC mice. (D) Levels of human embryonic β-globin (HBE1) transcripts were monitored by q-PCR as shown in Fig. 1C.

13

Human β-globin YAC

pIpC x3at 8wks old

Harvest PB samples before and after pIpC treatment

Week1 6 12

0

0

5

10

20

15

Hbb-bh1

1714x

Rel

ativ

e m

RN

A le

vels

Mea

n FP

KM (x

104)

5

10

15

20

0

1

2

5

Mea

n FP

KM (x

106)

Hbb-y Hbb-bh1

0

Mea

n FP

KM (x

105)

5

10

15

20

0

2

4

6

Mea

n FP

KM (x

105)

Hbb-s Hbb-t

Embryonic β-globin

Embryonic α-globin

Adult β-globin

Adult α-globin

(56) (99)

(377) (362)

A

B C

Zbtb7aF/F

Zbtb7aF/F Mx1-Cre+

Zbtb7aF/F

Zbtb7aF/F Mx1-Cre+

CD71+Ter119+FSChighCD44high (basophilic erythroblasts)

CD71+Ter119+FSCmedCD44med (polychromatophilic erythroblasts)

Zbtb7aF/F Zbtb7aF/F

Mx1-Cre+

12.1

0.007

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

F/F F/F

Cre+

(8.6) (10) (28) (40)

Weeks

after pIpC0 1 6 120 1 6 120 1 6 12

F/F Cre-

β-YAC+F/+ Cre+

β-YAC+F/F Cre+

β-YAC+

0

0.02

HB

E1/ (H

BB

+H

BG

+H

BE

1) % HBE1 (%)

0.03

0.04

0.01

Hba-x

F/F F/F

Cre+

F/F F/F

Cre+

Hba-a1/2

F/F F/F

Cre+

F/F F/F

Cre+

*

*p < 10-100

*

*

*p< 10-100

*

p= 0.27

p= 0.47

p= 0.51

p= 0.33

p= 0.35

p= 0.52

p= 0.7 p= 0.4

3

4

D

LCR

HBE1 HBG1/2 HBD HBB

3’HS1

Fig. S3. LRF inactivation induces γ-globin production in HSPC-derived erythroid cells. (A) Human CD34+ cells were induced into an erythroid lineage with EDM-based medium (see Methods). EDM-I: EDM supplemented with hydrocortisone, SCF and IL3; EDM-II: EDM with SCF; and EDM-III: EDM with no supplement. Lentivirus (pLKO vector) infection was performed on day 7 and puromycin selection started on day 9. (B) Protein samples were obtained from day 15 erythroblasts. LRF knockdown efficiency was assessed by Western blot using the anti-LRF antibody. GAPDH: loading control. (C) Relative copy number of indicated β-globin transcripts was determined by q-PCR. A plasmid containing the corresponding PCR amplicon was generated for each gene and used to calculate copy number. Copy numbers are reported relative to corresponding RPS18 levels in each sample. (D) Representative γ-globin FACS profiles of untreated, scrambled shRNA-treated, and LRF shRNA-treated human erythroblasts. Each sample was first incubated with either non-specific mouse IgG1κ (control) or anti-HBG1/2 antibody, followed by staining with fluorochrome-conjugated anti-mouse IgG1 antibody. γ-globin positivity was determined based on background signal levels (IgG control). (E) Representative HPLC profiles on day 15. %HbF was defined as a proportion of HbF relative to HbA0.

14

D

HB

G1

/2Ig

G c

on

tro

l

1.0%

28%

1.1%

37%

1.3%

50%

Untreated

E

1.3%

69%

LRF

GAPDH

Pa

ren

tal

Scra

mb

led

ZBTB

7A #

2

ZBTB

7A #

4

Day 0 7 9

Puromycin

EDM-I EDM-II EDM-III

11 15

q-PCR

Hb FACS

HPLC

ShRNA

infection

Thaw

hCD34+

Human CD34+-derived erythroid cell culture

B

Parental Scrambled ZBTB7A #2 ZBTB7A #4

A

ShRNA

ShRNA

HbF: 13.2% HbF: 25.7% HbF: 49.2% HbF: 71.0%

Scrambled ZBTB7A #2 ZBTB7A #4ShRNA

CH

bF

/(H

bF

+H

bA

0)

(%)

0

500

1000

1500

Re

lative

co

py#

HBE1 HBG1/2 HBB

Parental

Scrambled

ZBTB7A #2ZBTB7A #4

0.0

07

0.0

02

0.0

03

0.0

02

FSC

Fig. S4. Effects of LRF inactivation on HSPC-derived erythroid differentiation. (A) ShRNA-mediated LRF knockdown promotes a delay in erythroid differentiation. Representative FACS profiles of Day 10 and 14 erythroblasts are shown. Fewer CD235+CD71+ cells were evident on day 10 in LRF knockdown relative to control cells. (B) Representative images of Wright-Giemsa staining of cytospins prepared on Day14.

15

CD235aFSC CD235aFSC

Day 10 Day 14

Parental

ShRNA Scrambled

ShRNA ZBTB7A #2

ShRNA ZBTB7A #4

SSC

CD

71

SSC

CD

71

60%

42%

31%

30%

84%

96%

91%

98%

A

B

Parental

ShRNA Scrambled

ShRNA ZBTB7A #2

ShRNA ZBTB7A #4

Fig. S5. LRF inactivation in HUDEP-2 cells. (A) HUDEP-2 cells were maintained in SFEM medium in the presence of doxycycline (Dox) as described (15). To induce hemoglobin production, cells were cultured with EDM-II with Dox for 5 days and then cultured without Dox for two more days prior to analysis. (B) LRF KO in HUDEP-2 cells was validated by Western blot. BCL11A levels were also examined using BCL11A knockdown HUDEP-2 cells as a control. Tubulin: loading control. (C) Representative images of Wright-Giemsa staining of cytospins prepared on indicated days. (D) Immunophenotyping of control and LRF KO (ZBTB7AΔ/Δ) HUDEP-2 cells. FACS profiles of indicated surface markers were indistinguishable between genotypes.

16

SFEM Dox+ (Day 0) EDM-II Dox+ (Day 5) EDM-II Dox- (Day 11)

ZBTB7AΔ/Δ

ZBTB7A+/+

Switch to

EDM-II (Dox+)

Day0 Day5

Dox off

BSFEM

(Dox+)

Human immortalized erythroid line (HUDEP-2)

RNA-Seq

q-PCR

Hb FACS

HPLC

C

A

D ZBTB7AΔ/Δ ZBTB7A+/+

SFEM Dox+

EDM-II

Dox+ (Day5)

CD235a

CD

71

CD

36

α4-In

tegrin

FSC

SS

C

SFEM Dox+

EDM-II

Dox+ (Day5)

CD235a

CD

71

CD

36

α4-In

tegrin

FSC

SS

C

ZBTB

7A+/

+

BCL11A

LRF

Tubulin

ZBTB

7AΔ/

Δ

ZBTB

7A+/

+

+ShR

NA B

CL11

A

Differentiation

Maintenance

Fig. S6. HUDEP-2 cells exhibit gene expression patterns similar to those seen in human basophilic erythroblasts. Principal component analysis was performed using the RNA-Seq data (log10 read counts) of HUDEP-2 cells (this study) and HSPC-derived erythroid cells (57, 58). First (PC1) and third (PC3) components are shown, since the second (PC2) component captured separation of the HUDEP cells relative to all other datasets.

17

−20 0 20 40 60

−15

−10

−50

510

15

PC1

PC

3

Polychromatophilic

Late basophilic

HUDEP-2 after differentiation (EDM-II Dox+ day5)

HUDEP-2 before differentiation (SFEM)

Early basophilic

Proerythroblast

Orthochromatophilic

Fig. S7. LRF inactivation induces γ-globin expression in HUDEP-2 cells. (A) RNA-Seq analysis of control and LRF KO HUDEP-2 cells before (SFEM, Dox+) and after (EDM day 5, Dox+) differentiation. Bar graphs show mRNA levels (mean FPKM values) of α- and β-globins in control (ZBTB7A+/+) and LRF KO (ZBTB7AΔ/Δ) HUDEP-2 cells. Values in parentheses indicate mean FKPM values of less abundant transcripts. P values: unadjusted P values determined using the DESeq Bioconductor package (34). (B) Bar graphs show proportions of γ-globin mRNA to total β-globin mRNA in control and LRF KO HUDEP-2 cells. Control: HUDEP-2_Cas9. (C) γ-globin protein levels were determined by Western blot using protein lysates from control, ZBTB7AΔ/Δ and BCL11AΔ/Δ HUDEP-2 cells. HBG protein was detected only in ZBTB7AΔ/Δ or BCL11AΔ/Δ cells. GAPDH: loading control. (D) Representative FACS profiles of γ-globin protein expression. (E) LRF protein was not detected in ZBTB7AΔ/Δ clones. HSP90: loading control. Asterisk denotes band corresponding to a residual LRF signal from a primary blot. (F) Bar graphs show transcript levels of known γ-globin repressors in control and ZBTB7AΔ/Δ HUDEP-2 cells. Mean FPKM values of two independent samples per genotype (before differentiation) are shown.

18

C

FSC

HBG

1/2

IgG

con

trol 1.1%

3.9%

1.2%

63%

ZBTB7A+/+ ZBTB7AΔ/ΔD

A

0

Mea

n FP

KM (x

104 )

5

10

15

20 Embryonic/Fetal β-globin Embryonic α-globinAdult β-globin

Adult α-globin

0

Mea

n FP

KM (x

106 )

1

2

3

4

0

Mea

n FP

KM (x

104 )

2

4

6

8

0

Mea

n FP

KM (x

104 )

5

10

15

SFEM Dox+

EDM-II Dox+(Day5)

ZBTB7A+/+

ZBTB7AΔ/Δ

ZBTB7A+/+

ZBTB7AΔ/Δ

HBE1 HBG2 HBG1 HBB

HBM HBA2 HBA1

LRF

HSP90

ZBTB

7A+/+

2-6

2-19

2-23

ZBTB7AΔ/Δclones

*

E

GAPDH

HBG1/2

Cont

rol

ZBTB

7AΔ/Δ

BCL1

1AΔ/Δ

Cont

rol

ZBTB

7AΔ/Δ

BCL1

1AΔ/Δ

1x105 cells 0.3x105 cells

+/+ Δ/Δ +/+

Δ/Δ+/+

Δ/Δ +/+

Δ/Δ +/+

Δ/Δ +/+

Δ/Δ+/+

Δ/Δ +/+

Δ/Δ+/+ Δ/Δ +/+

Δ/Δ +/+

Δ/Δ +/+

Δ/Δ+/+

Δ/Δ +/+

Δ/Δ

(10) (29) (31) (90) (235)(660) (734)(3015) (504)

B

+/+ Δ/Δ

HBG

/ (HB

B+HB

G) (

%) 100

80

60

40

20

0

HBBHBG1/2

HBZ

+/+ Δ/Δ +/+

Δ/Δ

% HBG

F

BCL1

1AKL

F1

GATA

1

ZFPM

1MB

D2CH

D4

NR2C

1

NR2C

2SO

X6

KDM1

A

EHMT

2

Mea

n FP

KM (x

103 )

0

2

4

6 ZBTB7A+/+

ZBTB7AΔ/Δ

p= 0.12

**p< 10-5

*

*

p= 0.2

p< 0.01

p= 0.8p= 0.9

p= 0.7p= 0.9

p= 0.8

p= 0.9

* p< 0.001

p= 0.1 p= 0.1

Fig. S8. LRF ChIP-Seq in human erythroid cells. (A) LRF-ChIP-Seq was performed using HSPC-derived erythroblasts (Day 8) in duplicate and HUDEP-2 cells in triplicates. Distribution of sample correlation coefficients among replicates was examined. Pearson's correlation coefficient values are shown. (B) Analysis revealed 5684 and 10385 LRF binding sites in human erythroblasts and HUDEP-2 cells, respectively. Venn diagram chart shows number of target genes shared between both cell types. (C) Heat map showing correlation between LRF (this study) and BCL11A and SOX6 (17) binding sites. LRF and BCL11A occupancy sites are mutually exclusive.

19

Replicate #1 Replicate #1 Replicate #1 Replicate #2

Repl

icate

#2

Repl

icate

#2

Repl

icate

#3

Repl

icate

#3r = 0.742 r = 0.805 r = 0.855 r = 0.866

AHUDEP-2

hCD34+-derived erythroblasts

B hCD34+-derived erythroblasts HUDEP-2

57421041

C

LRF

BCL1

1A

SOX6

SOX6

BCL11A

LRF 5684

0

2

0

108

35

3

34

372

% overlapping binding sites

0 20 40 60 80 1004643

Fig. S9. LRF binding motifs identified in HSPC-derived erythroid cells. LRF binding motifs were identified using motif discovery programs (Tomtom (59), MEME (60), DREAME (61), CentriMo (62), TRANSFAC (63) and JASPAR (64)). Motifs identified with high statistical significance (E-value (59) < 5x10^-3) are listed. Similar known motifs are also shown. Motif distribution curve (62) was generated based on the probability of the best match to a given motif occurring at a given position in input sequences. Grey vertical line indicates center of input sequence.

20

2.0e-9

1.8e-6

3.8e-5

8.3e-4

1.0e-3

Known or

Similar MotifsE-value

CTCF (MA0139.1)

SP2 (MA0516.1)

INSM1 (MA0155.1)

EGD1 (MA0162.2)

SP1 (MA0079.3)

KLF5 (MA0599.1)

ELK4 (MA0076.2)

ELF1 (MA0473.1)

GABPA (MA0062.2)

ZNF263 (MA0528.1)

EHF (MA0598.1)

Erg (MA0474.1)

TAL1::GATA1 (MA0140.2)

GATA2 (MA0036.2)

Gata4 (MA0482.1)

Human erythroblasts

5.6e-20

LRF (Maeda et. al

Nature 2005)

Discovery/Enrichment

ProgramMotif Found

DREAME

DREAME

DREAME

DREAME

DREAME

DREAME

3.7e-4 CentriMoCTCF (MA0139.1)

Distribution

Not centrally

enriched

Not centrally

enriched

Not centrally

enriched

Not centrally

enriched

Fig. S10. LRF binding motifs identified in HUDEP-2 cells. Motif discovery was performed using HUDEP-2 LRF-ChIP-Seq data as described.

21

Zfp281_secondary

(UP00021_2)8.8e-36

1.4e-11

8.9e-7

4.6e-3

Known or

Similar MotifsE-value

Glis2_primary

(UP00024_1)

Sp4 (UP00002_1)

SP2 (MA0516.1)

ISM1 (MA0155.1)

Zbtb7b_secondary

(UP00047_2)

GATA3 (MA0037.2)

Gata1 (MA0035.3)

Gata4 (MA0482.1)

HUDEP-2 cells

4.1e-71

Zbtb7b_primary

(UP00047_1)

Zbtb7b_secondary

(UP00047_2)

Zfp281_secondary

(UP00021_2)

Discovery/Enrichment

ProgramMotif Found

MEME

DREAME

CentriMo

DREAME

DREAME

DREAME2.7e-3Bcl6b (UP00043_2)

Klf7 (UP00093_1)

Plagl1(UP00088_1)

Distribution

Not centrally

enriched

3.4e-6

7.2e-4

Zfp740_primary

(UP00022_1)

Zic1_primary

(UP00102_1)

Zfp281_primary

(UP00021_1)

CentriMo

CentriMo

CentriMo

DREAME

6.2e-5

1.3e-6

Not centrally

enriched

Fig. S11. LRF-interacting proteins identified by Y2H. (A) A Y2H screen identified components of the NuRD complex and chromatin remodelers as direct LRF binding proteins. LRF-BTB dimer structure based on a 2NN2 PDB file is shown. (B) Proteins listed were identified as direct LRF binding partners with high confidence. Protein domain architectures were obtained from SMRT (http://smart.embl.de/). The LRF binding region in each protein is depicted with a purple oval.

22

http://smart.embl.dehttp://smart.embl.de

Fig. S12. Interaction between LRF and NuRD components. (A) Bar graphs show abundance of indicated transcripts in mouse erythroblasts (left) and HUDEP-2 cells (right). FKPM values were obtained from RNA-Seq experiments. (B) Validation of Y2H results by IP. Exogenously-expressed Flag-tagged target proteins interact with endogenous LRF in 293T cells. Flag-tagged target proteins (CHD8, CHD3, BAZ1A or GATAD2B) were expressed in 293T cells, lysates were prepared 48 h later and IP was performed using standard methodology with anti-Flag antibody, followed by Western blot with anti-LRF antibody. Protein samples from empty vector-transfected cells served as negative controls. (C) LRF/GATAD2B interaction was validated in mouse erythroleukemia (MEL) cells. We generated a series of expression vectors encoding N- or C-terminally-tagged mouse LRF (mLRF) and performed IP in MEL cells. We could pull-down NuRD components and endogenous LRF protein only when we expressed the LRF tagged at the far C-terminus. Anti-HA antibody could pull down both endogenous and exogenous LRF plus NuRD components (far right lane), while antibody against Flag, which was placed 15 aa upstream of the HA tag, pulled down only exogenous LRF (2nd lane from the left). Although we tested 5 different (2 original and 3 commercially-available) anti-LRF antibodies for IP in mouse cells, none could pull down endogenous mouse LRF protein. Native LRF in the large NuRD complex may be inaccessible to anti-LRF antibodies in mouse cells (a model shown on right). Note: BCL11A image is from a different membrane. (D) Knockout of LRF or BCL11A was confirmed by Western blot using anti-LRF or -BCL11A antibodies. GAPDH: loading control. (E) Relative copy number of indicated β-globin transcripts was determined by q-PCR, as described in Fig. S3C. (F) Bar graphs show proportions of γ-globin mRNA to total β-globin mRNA in control and LRF KO HUDEP-2 cells. Control: HUDEP-2_Cas9.

23

B

CH

D8

Em

pty

CH

D8

Em

pty

Input IP

CH

D3

Em

pty

CH

D3

Em

pty

Input IPB

AZ

1A

Em

pty

BA

Z1A

Em

pty

Input IP

GA

TA

D2B

Em

pty

Em

pty

Input IP

WB: α-LRF

WB: α-Flag

IP: α-Flag

A

Zbtb

5

Zbtb

22

Zbtb

45

Chd8

Chd3

Gata

d2b

Src

ap

Baz1

a

Zbtb

7a

0

2

4

6

0

5

10

20

15M

ean F

PK

M (

x10

2)

Mean F

PK

M (

x10

2)

n/a

ZBTB5

ZBTB22

ZBTB45

CHD8

CHD3

GATAD2B

SRCAP

BAZ1A

ZBTB7A

Mouse splenic erythroblasts (FSChighCD44high) HUDEP-2 (before differentiation)

GA

TA

D2B

endo

exo

Sup

Bound

Sup

Bound

IP: α-FLAG α-HA

LRF

Gatad2b

Hdac1

Mta2

Bcl11a

C

FHA

BTB

D

LRF

BCL11A

Contr

ol

ZB

TB

7A

Δ/Δ

BC

L11A

Δ/Δ

GAPDH

DK

O #

5

DK

O #

18

E

0

500

1000

1500

Rela

tive c

opy#

HBE1 HBG1/2 HBB

Parental

ZBTB7AΔ/Δ

BCL11AΔ/Δ

DKO #18

N.D

.

0.0

07

0.0

6

0.0

09

1.0

6.5

F 10080

60

40

20

0

HB

G/ (H

BB

+H

BG

) (%

)

Pare

nta

lZB

TB

7A

Δ/Δ

BC

L11A

Δ/Δ

DKO

#18

HBB

HBG1/2

Fig. S13. Proposed model for γ-globin silencing mediated by the LRF-NuRD complex.

Fig. S14. Gene Set Enrichment Analysis (GSEA) of gene expression differences following ZBTB7A knockout. (A) GSEA was performed based on log2 fold-differences in expression magnitude between control and Zbtb7a KO splenic erythroblasts (Fig. 1A). Listed are gene Ontology (GO) categories showing statistically significant association after correction for multiple hypothesis testing (Q-value). A positive Edge value indicates upregulated gene expression in LRF KO cells. (B) Shown are significantly associated GO categories for expression differences between control and ZBTB7AΔ/Δ HUDEP-2 cells (EDM-II Dox+ Day 5).

24

GO:0005886 1.00E-05 4.84E-04 1.43 0.76 plasma membrane (CC)

GO:0045211 4.00E-05 1.46E-03 1.19 0.72 postsynaptic membrane (CC)

GO:0030054 8.00E-05 2.56E-03 1.16 1.07 cell junction (CC)

GO:0035115 8.00E-05 2.56E-03 1.12 1.11 embryonic forelimb morphogenesis (BP)

GO:0001501 1.00E-04 3.08E-03 1.13 1.84 skeletal system development (BP)

GO:0030424 1.30E-04 3.81E-03 1.11 0.43 axon (CC)

GO:0007165 1.60E-04 4.52E-03 1.11 1.23 signal transduction (BP)

GO:0005833 2.30E-04 5.93E-03 1.05 4.62 hemoglobin complex (CC)

GO:0005344 2.60E-04 6.42E-03 1.05 4.62 oxygen transporter activity (MF)

GO:0030049 2.60E-04 6.42E-03 1.07 1.13 muscle filament sliding (BP)

GO:0015671 2.80E-04 6.64E-03 1.05 4.62 oxygen transport (BP)

GO:0033077 4.20E-04 9.39E-03 -1.05 -3.20 T cell differentiation in thymus (BP)

GO:0072562 1.00E-05 2.35E-03 1.19 2.15 blood microparticle (CC)

GO:0005615 3.00E-05 4.30E-03 1.27 2.07 extracellular space (CC)

GO:0005576 4.00E-05 5.43E-03 1.19 2.07 extracellular region (CC)

GO:0032040 5.00E-05 6.14E-03 1.23 0.33 small-subunit processome (CC)

GO:0009897 6.00E-05 6.19E-03 1.19 1.48 external side of plasma membrane (CC)

GO:0019825 6.00E-05 6.19E-03 1.03 9.52 oxygen binding (MF)

GO:0030154 1.00E-04 9.56E-03 1.16 1.75 cell differentiation (BP)

B

A GO ID P value ScoreQ value Edge value Description

GO ID P value ScoreQ value Edge value Description

LCR HBBHBDHBG1HBG2HBE1

GATAD2BCHD3

MTA2 HDAC

LRF

NuRDBCL11A

SOX6

NuRDMTA2

HDAC

References

1. G. Stamatoyannopoulos, Control of globin gene expression during development and erythroid

differentiation. Exp. Hematol. 33, 259–271 (2005). Medline

doi:10.1016/j.exphem.2004.11.007

2. S. Pleasants, Epidemiology: A moving target. Nature 515, S2 (2014). doi:10.1038/515S2a

3. F. B. Piel, S. I. Hay, S. Gupta, D. J. Weatherall, T. N. Williams, Global burden of sickle cell

anaemia in children under five, 2010-2050: Modelling based on demographics, excess

mortality, and interventions. PLOS Med. 10, e1001484 (2013). Medline

doi:10.1371/journal.pmed.1001484

4. D. E. Bauer, S. C. Kamran, S. H. Orkin, Reawakening fetal hemoglobin: Prospects for new

therapies for the β-globin disorders. Blood 120, 2945–2953 (2012). Medline

doi:10.1182/blood-2012-06-292078

5. S. U. Lee, T. Maeda, POK/ZBTB proteins: An emerging family of proteins that regulate

lymphoid development and function. Immunol. Rev. 247, 107–119 (2012). Medline

doi:10.1111/j.1600-065X.2012.01116.x

6. S. U. Lee, M. Maeda, Y. Ishikawa, S. M. Li, A. Wilson, A. M. Jubb, N. Sakurai, L. Weng, E.

Fiorini, F. Radtke, M. Yan, H. R. Macdonald, C. C. Chen, T. Maeda, LRF-mediated Dll4

repression in erythroblasts is necessary for hematopoietic stem cell maintenance. Blood

121, 918–929 (2013). Medline doi:10.1182/blood-2012-03-418103

7. See supplementary materials on Science Online.

8. P. D. Kingsley, J. Malik, K. A. Fantauzzo, J. Palis, Yolk sac-derived primitive erythroblasts

enucleate during mammalian embryogenesis. Blood 104, 19–25 (2004). Medline

doi:10.1182/blood-2003-12-4162

9. M. H. Baron, J. Isern, S. T. Fraser, The embryonic origins of erythropoiesis in mammals.

Blood 119, 4828-4837 (2012). doi:10.1182/blood-2012-01-153486

10. K. Chada, J. Magram, F. Costantini, An embryonic pattern of expression of a human fetal

globin gene in transgenic mice. Nature 319, 685–689 (1986). Medline

doi:10.1038/319685a0

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15730849&dopt=Abstracthttp://dx.doi.org/10.1016/j.exphem.2004.11.007http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23874164&dopt=Abstracthttp://dx.doi.org/10.1371/journal.pmed.1001484http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22904296&dopt=Abstracthttp://dx.doi.org/10.1182/blood-2012-06-292078http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22500835&dopt=Abstracthttp://dx.doi.org/10.1111/j.1600-065X.2012.01116.xhttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23134786&dopt=Abstracthttp://dx.doi.org/10.1182/blood-2012-03-418103http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15031208&dopt=Abstracthttp://dx.doi.org/10.1182/blood-2003-12-4162http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=3951540&dopt=Abstracthttp://dx.doi.org/10.1038/319685a0

11. P. D. Kingsley, J. Malik, R. L. Emerson, T. P. Bushnell, K. E. McGrath, L. A. Bloedorn, M.

Bulger, J. Palis, “Maturational” globin switching in primary primitive erythroid cells.

Blood 107, 1665–1672 (2006). Medline doi:10.1182/blood-2005-08-3097

12. K. R. Peterson, C. H. Clegg, C. Huxley, B. M. Josephson, H. S. Haugen, T. Furukawa, G.

Stamatoyannopoulos, Transgenic mice containing a 248-kb yeast artificial chromosome

carrying the human β-globin locus display proper developmental control of human globin

genes. Proc. Natl. Acad. Sci. U.S.A. 90, 7593–7597 (1993). Medline

doi:10.1073/pnas.90.16.7593

13. J. Xu, C. Peng, V. G. Sankaran, Z. Shao, E. B. Esrick, B. G. Chong, G. C. Ippolito, Y.

Fujiwara, B. L. Ebert, P. W. Tucker, S. H. Orkin, Correction of sickle cell disease in adult

mice by interference with fetal hemoglobin silencing. Science 334, 993–996 (2011).

Medline doi:10.1126/science.1211053

14. T. Maeda, K. Ito, T. Merghoub, L. Poliseno, R. M. Hobbs, G. Wang, L. Dong, M. Maeda, L.

C. Dore, A. Zelent, L. Luzzatto, J. Teruya-Feldstein, M. J. Weiss, P. P. Pandolfi, LRF is

an essential downstream target of GATA1 in erythroid development and regulates BIM-

dependent apoptosis. Dev. Cell 17, 527–540 (2009). Medline

doi:10.1016/j.devcel.2009.09.005

15. R. Kurita, N. Suda, K. Sudo, K. Miharada, T. Hiroyama, H. Miyoshi, K. Tani, Y. Nakamura,

Establishment of immortalized human erythroid progenitor cell lines able to produce

enucleated red blood cells. PLOS ONE 8, e59890 (2013). Medline

doi:10.1371/journal.pone.0059890

16. T. Maeda, R. M. Hobbs, T. Merghoub, I. Guernah, A. Zelent, C. Cordon-Cardo, J. Teruya-

Feldstein, P. P. Pandolfi, Role of the proto-oncogene Pokemon in cellular transformation

and ARF repression. Nature 433, 278–285 (2005). Medline doi:10.1038/nature03203

17. J. Xu, V. G. Sankaran, M. Ni, T. F. Menne, R. V. Puram, W. Kim, S. H. Orkin,

Transcriptional silencing of gamma-globin by BCL11A involves long-range interactions

and cooperation with SOX6. Genes Dev. 24, 783–798 (2010). Medline

doi:10.1101/gad.1897310

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16263786&dopt=Abstracthttp://dx.doi.org/10.1182/blood-2005-08-3097http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=8356061&dopt=Abstracthttp://dx.doi.org/10.1073/pnas.90.16.7593http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21998251&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21998251&dopt=Abstracthttp://dx.doi.org/10.1126/science.1211053http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19853566&dopt=Abstracthttp://dx.doi.org/10.1016/j.devcel.2009.09.005http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23533656&dopt=Abstracthttp://dx.doi.org/10.1371/journal.pone.0059890http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15662416&dopt=Abstracthttp://dx.doi.org/10.1038/nature03203http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20395365&dopt=Abstracthttp://dx.doi.org/10.1101/gad.1897310

18. J. D. Buenrostro, P. G. Giresi, L. C. Zaba, H. Y. Chang, W. J. Greenleaf, Nat. Methods 10,

1213–1218 (2013). doi:10.1038/nmeth.2688

19. J. W. Rupon, S. Z. Wang, K. Gaensler, J. Lloyd, G. D. Ginder, Methyl binding domain

protein 2 mediates γ-globin gene silencing in adult human βYAC transgenic mice. Proc.

Natl. Acad. Sci. U.S.A. 103, 6617–6622 (2006). Medline doi:10.1073/pnas.0509322103

20. M. Amaya, M. Desai, M. N. Gnanapragasam, S. Z. Zhu, D. C. Williams Jr., G. D. Ginder,

Mi2β-mediated silencing of the fetal γ-globin gene in adult erythroid cells. Blood 121,

3493–3501 (2013). doi:10.1182/blood-2012-11-466227

21. V. G. Sankaran, T. F. Menne, J. Xu, T. E. Akie, G. Lettre, B. Van Handel, H. K. Mikkola, J.

N. Hirschhorn, A. B. Cantor, S. H. Orkin, Human fetal hemoglobin expression is

regulated by the developmental stage-specific repressor BCL11A. Science 322, 1839–

1842 (2008). Medline

22. J. Xu, D. E. Bauer, M. A. Kerenyi, T. D. Vo, S. Hou, Y. J. Hsu, H. Yao, J. J. Trowbridge, G.

Mandel, S. H. Orkin, Corepressor-dependent silencing of fetal hemoglobin expression by

BCL11A. Proc. Natl. Acad. Sci. U.S.A. 110, 6518–6523 (2013). Medline

doi:10.1073/pnas.1303976110

23. V. G. Sankaran, J. Xu, T. Ragoczy, G. C. Ippolito, C. R. Walkley, S. D. Maika, Y. Fujiwara,

M. Ito, M. Groudine, M. A. Bender, P. W. Tucker, S. H. Orkin, Developmental and

species-divergent globin switching are driven by BCL11A. Nature 460, 1093–1097

(2009). Medline doi:10.1038/nature08243

24. I. T. Chan, J. L. Kutok, I. R. Williams, S. Cohen, L. Kelly, H. Shigematsu, L. Johnson, K.

Akashi, D. A. Tuveson, T. Jacks, D. G. Gilliland, Conditional expression of oncogenic K-

ras from its endogenous promoter induces a myeloproliferative disease. J. Clin. Invest.

113, 528–538 (2004). Medline doi:10.1172/JCI20476

25. M. Y. Su, L. A. Steiner, H. Bogardus, T. Mishra, V. P. Schulz, R. C. Hardison, P. G.

Gallagher, Identification of biologically relevant enhancers in human erythroid cells. J.

Biol. Chem. 288, 8433–8444 (2013). Medline doi:10.1074/jbc.M112.413260

26. T. Maeda, T. Merghoub, R. M. Hobbs, L. Dong, M. Maeda, J. Zakrzewski, M. R. van den

Brink, A. Zelent, H. Shigematsu, K. Akashi, J. Teruya-Feldstein, G. Cattoretti, P. P.

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16608912&dopt=Abstracthttp://dx.doi.org/10.1073/pnas.0509322103http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19056937&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23576758&dopt=Abstracthttp://dx.doi.org/10.1073/pnas.1303976110http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19657335&dopt=Abstracthttp://dx.doi.org/10.1038/nature08243http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=14966562&dopt=Abstracthttp://dx.doi.org/10.1172/JCI20476http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23341446&dopt=Abstracthttp://dx.doi.org/10.1074/jbc.M112.413260

Pandolfi, Regulation of B versus T lymphoid lineage fate decision by the proto-oncogene

LRF. Science 316, 860–866 (2007). Medline doi:10.1126/science.1140881

27. Y. Ishikawa, M. Maeda, M. Pasham, F. Aguet, S. K. Tacheva-Grigorova, T. Masuda, H. Yi,

S. U. Lee, J. Xu, J. Teruya-Feldstein, M. Ericsson, A. Mullally, J. Heuser, T.

Kirchhausen, T. Maeda, Role of the clathrin adaptor PICALM in normal hematopoiesis

and polycythemia vera pathophysiology. Haematologica 100, 439–451 (2015). Medline

doi:10.3324/haematol.2014.119537

28. L. Cong, F. A. Ran, D. Cox, S. Lin, R. Barretto, N. Habib, P. D. Hsu, X. Wu, W. Jiang, L. A.

Marraffini, F. Zhang, Multiplex genome engineering using CRISPR/Cas systems. Science

339, 819–823 (2013). Medline doi:10.1126/science.1231143

29. M. C. Canver, D. E. Bauer, A. Dass, Y. Y. Yien, J. Chung, T. Masuda, T. Maeda, B. H. Paw,

S. H. Orkin, Characterization of genomic deletion efficiency mediated by clustered

regularly interspaced palindromic repeats (CRISPR)/Cas9 nuclease system in mammalian

cells. J. Biol. Chem. 289, 21312–21324 (2014). doi:10.1074/jbc.M114.564625

30. J. Black, Isoelectric focusing in agarose gel for detection and identification of hemoglobin

variants. Hemoglobin 8, 117–127 (1984). doi:10.3109/03630268408991705

31. A. Shevchenko, H. Tomas, J. Havliš, J. V. Olsen, M. Mann, In-gel digestion for mass

spectrometric characterization of proteins and proteomes. Nat. Protoc. 1, 2856–2860

(2006). Medline doi:10.1038/nprot.2006.468

32. H. B. Kramer, K. J. Lavender, L. Qin, A. R. Stacey, M. K. Liu, K. di Gleria, A. Simmons, N.

Gasper-Smith, B. F. Haynes, A. J. McMichael, P. Borrow, B. M. Kessler, Elevation of

intact and proteolytic fragments of acute phase proteins constitutes the earliest systemic

antiviral response in HIV-1 infection. PLOS Pathog. 6, e1000893 (2010). Medline

doi:10.1371/journal.ppat.1000893

33. S. Anders, P. T. Pyl, W. Huber, HTSeq—a Python framework to work with high-throughput

sequencing data. Bioinformatics 31, 166–169 (2015). Medline

doi:10.1093/bioinformatics/btu638

34. S. Anders, W. Huber, Differential expression analysis for sequence count data. Genome Biol.

11, R106 (2010). Medline doi:10.1186/gb-2010-11-10-r106

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=17495164&dopt=Abstracthttp://dx.doi.org/10.1126/science.1140881http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25552701&dopt=Abstracthttp://dx.doi.org/10.3324/haematol.2014.119537http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=23287718&dopt=Abstracthttp://dx.doi.org/10.1126/science.1231143http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=17406544&dopt=Abstracthttp://dx.doi.org/10.1038/nprot.2006.468http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20463814&dopt=Abstracthttp://dx.doi.org/10.1371/journal.ppat.1000893http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25260700&dopt=Abstracthttp://dx.doi.org/10.1093/bioinformatics/btu638http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20979621&dopt=Abstracthttp://dx.doi.org/10.1186/gb-2010-11-10-r106

35. P. V. Kharchenko, M. Y. Tolstorukov, P. J. Park, Design and analysis of ChIP-seq

experiments for DNA-binding proteins. Nat. Biotechnol. 26, 1351–1359 (2008). Medline

doi:10.1038/nbt.1508

36. J. D. Buenrostro, B. Wu, H. Y. Chang, W. J. Greenleaf, Curr. Protoc. Mol. Biol. 109,

21.29.1-9 (2015).

37. A. B. Vojtek, S. M. Hollenberg, J. A. Cooper, Mammalian Ras interacts directly with the

serine/threonine kinase Raf. Cell 74, 205–214 (1993). Medline doi:10.1016/0092-

8674(93)90307-C

38. A. B. Stergachis, S. Neph, R. Sandstrom, E. Haugen, A. P. Reynolds, M. Zhang, R. Byron, T.

Canfield, S. Stelhing-Sun, K. Lee, R. E. Thurman, S. Vong, D. Bates, F. Neri, M. Diegel,

E. Giste, D. Dunn, J. Vierstra, R. S. Hansen, A. K. Johnson, P. J. Sabo, M. S. Wilken, T.

A. Reh, P. M. Treuting, R. Kaul, M. Groudine, M. A. Bender, E. Borenstein, J. A.

Stamatoyannopoulos, Conservation of trans-acting circuitry during mammalian

regulatory evolution. Nature 515, 365–370 (2014). Medline doi:10.1038/nature13972

39. J. C. Ulirsch, J. N. Lacy, X. An, N. Mohandas, T. S. Mikkelsen, V. G. Sankaran, Altered

chromatin occupancy of master regulators underlies evolutionary divergence in the

transcriptional landscape of erythroid differentiation. PLOS Genet. 10, e1004890 (2014).

Medline doi:10.1371/journal.pgen.1004890

40. M. C. Canver, E. C. Smith, F. Sher, L. Pinello, N. E. Sanjana, O. Shalem, D. D. Chen, P. G.

Schupp, D. S. Vinjamur, S. P. Garcia, S. Luc, R. Kurita, Y. Nakamura, Y. Fujiwara, T.

Maeda, G.-C. Yuan, F. Zhang, S. H. Orkin, D. E. Bauer, BCL11A enhancer dissection by Cas9-mediated in situ saturating mutagenesis. Nature 527, 192–197 (2015).

doi:10.1038/nature15521

41. A. L. Jackson, P. S. Linsley, Recognizing and avoiding siRNA off-target effects for target

identification and therapeutic application. Nat. Rev. Drug Discov. 9, 57–67 (2010).

Medline doi:10.1038/nrd3010

42. D. B. Constam, A. R. Tobler, A. Rensing-Ehl, I. Kemler, L. B. Hersh, A. Fontana,

Puromycin-sensitive aminopeptidase. Sequence analysis, expression, and functional

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=19029915&dopt=Abstracthttp://dx.doi.org/10.1038/nbt.1508http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=8334704&dopt=Abstracthttp://dx.doi.org/10.1016/0092-8674(93)90307-Chttp://dx.doi.org/10.1016/0092-8674(93)90307-Chttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25409825&dopt=Abstracthttp://dx.doi.org/10.1038/nature13972http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25521328&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=25521328&dopt=Abstracthttp://dx.doi.org/10.1371/journal.pgen.1004890http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20043028&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20043028&dopt=Abstracthttp://dx.doi.org/10.1038/nrd3010

characterization. J. Biol. Chem. 270, 26931–26939 (1995). Medline

doi:10.1074/jbc.270.45.26931

43. L. Schomburg, H. Kollmus, S. Friedrichsen, K. Bauer, Molecular characterization of a

puromycin-insensitive leucyl-specific aminopeptidase, PILS-AP. Eur. J. Biochem. 267,

3198–3207 (2000). Medline doi:10.1046/j.1432-1327.2000.01348.x

44. B. Tolhuis, R. J. Palstra, E. Splinter, F. Grosveld, W. de Laat, Looping and interaction

between hypersensitive sites in the active β-globin locus. Mol. Cell 10, 1453–1465

(2002). Medline doi:10.1016/S1097-2765(02)00781-5

45. S. H. Song, C. Hou, A. Dean, A positive role for NLI/Ldb1 in long-range β-globin locus

control region function. Mol. Cell 28, 810–822 (2007). Medline

doi:10.1016/j.molcel.2007.09.025

46. W. Deng, J. Lee, H. Wang, J. Miller, A. Reik, P. D. Gregory, A. Dean, G. A. Blobel,

Controlling long-range genomic interactions at a native locus by targeted tethering of a

looping factor. Cell 149, 1233–1244 (2012). Medline doi:10.1016/j.cell.2012.03.051

47. D. M. Bodine, T. J. Ley, An enhancer element lies 3′ to the human A γ globin gene. EMBO J.

6, 2997–3004 (1987). Medline

48. M. Kellis, B. Wold, M. P. Snyder, B. E. Bernstein, A. Kundaje, G. K. Marinov, L. D. Ward,

E. Birney, G. E. Crawford, J. Dekker, I. Dunham, L. L. Elnitski, P. J. Farnham, E. A.

Feingold, M. Gerstein, M. C. Giddings, D. M. Gilbert, T. R. Gingeras, E. D. Green, R.

Guigo, T. Hubbard, J. Kent, J. D. Lieb, R. M. Myers, M. J. Pazin, B. Ren, J. A.

Stamatoyannopoulos, Z. Weng, K. P. White, R. C. Hardison, Defining functional DNA

elements in the human genome. Proc. Natl. Acad. Sci. U.S.A. 111, 6131–6138 (2014).

Medline doi:10.1073/pnas.1318948111

49. O. Yildirim, R. Li, J. H. Hung, P. B. Chen, X. Dong, L. S. Ee, Z. Weng, O. J. Rando, T. G.

Fazzio, Mbd3/NURD complex regulates expression of 5-hydroxymethylcytosine marked

genes in embryonic stem cells. Cell 147, 1498–1510 (2011). Medline

doi:10.1016/j.cell.2011.11.054

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7592939&dopt=Abstracthttp://dx.doi.org/10.1074/jbc.270.45.26931http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=10824104&dopt=Abstracthttp://dx.doi.org/10.1046/j.1432-1327.2000.01348.xhttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=12504019&dopt=Abstracthttp://dx.doi.org/10.1016/S1097-2765(02)00781-5http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=18082606&dopt=Abstracthttp://dx.doi.org/10.1016/j.molcel.2007.09.025http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22682246&dopt=Abstracthttp://dx.doi.org/10.1016/j.cell.2012.03.051http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=3691478&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24753594&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24753594&dopt=Abstracthttp://dx.doi.org/10.1073/pnas.1318948111http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22196727&dopt=Abstracthttp://dx.doi.org/10.1016/j.cell.2011.11.054

50. T. Shimbo, Y. Du, S. A. Grimm, A. Dhasarathy, D. Mav, R. R. Shah, H. Shi, P. A. Wade,

MBD3 localizes at promoters, gene bodies and enhancers of active genes. PLOS Genet.

9, e1004028 (2013). Medline

51. A. Miccio, G. A. Blobel, Role of the GATA-1/FOG-1/NuRD pathway in the expression of

human β-like globin genes. Mol. Cell. Biol. 30, 3460–3470 (2010). Medline

doi:10.1128/MCB.00001-10

52. N. Sakurai, M. Maeda, S. U. Lee, Y. Ishikawa, M. Li, J. C. Williams, L. Wang, L. Su, M.

Suzuki, T. I. Saito, S. Chiba, S. Casola, H. Yagita, J. Teruya-Feldstein, S. Tsuzuki, R.

Bhatia, T. Maeda, The LRF transcription factor regulates mature B cell development and

the germinal center response in mice. J. Clin. Invest. 121, 2583–2598 (2011). Medline

doi:10.1172/JCI45682

53. J. M. Polo, T. Dell’Oso, S. M. Ranuncolo, L. Cerchietti, D. Beck, G. F. Da Silva, G. G.

Prive, J. D. Licht, A. Melnick, Specific peptide interference reveals BCL6 transcriptional

and oncogenic mechanisms in B-cell lymphoma cells. Nat. Med. 10, 1329–1335 (2004).

Medline doi:10.1038/nm1134

54. A. F. Ghetu, C. M. Corcoran, L. Cerchietti, V. J. Bardwell, A. Melnick, G. G. Privé,

Structure of a BCOR corepressor peptide in complex with the BCL6 BTB domain dimer.

Mol. Cell 29, 384–391 (2008). Medline doi:10.1016/j.molcel.2007.12.026

55. L. C. Cerchietti, A. F. Ghetu, X. Zhu, G. F. Da Silva, S. Zhong, M. Matthews, K. L. Bunting,

J. M. Polo, C. Farès, C. H. Arrowsmith, S. N. Yang, M. Garcia, A. Coop, A. D.

Mackerell Jr., G. G. Privé, A. Melnick, A small-molecule inhibitor of BCL6 kills DLBCL

cells in vitro and in vivo. Cancer Cell 17, 400–411 (2010). Medline

doi:10.1016/j.ccr.2009.12.050

56. C. Duy, C. Hurtz, S. Shojaee, L. Cerchietti, H. Geng, S. Swaminathan, L. Klemm, S. M.

Kweon, R. Nahar, M. Braig, E. Park, Y. M. Kim, W. K. Hofmann, S. Herzog, H. Jumaa,

H. P. Koeffler, J. J. Yu, N. Heisterkamp, T. G. Graeber, H. Wu, B. H. Ye, A. Melnick, M.

Müschen, BCL6 enables Ph+ acute lymphoblastic leukaemia cells to survive BCR-ABL1

kinase inhibition. Nature 473, 384–388 (2011). Medline doi:10.1038/nature09883

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24385926&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20439494&dopt=Abstracthttp://dx.doi.org/10.1128/MCB.00001-10http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21646720&dopt=Abstracthttp://dx.doi.org/10.1172/JCI45682http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15531890&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=15531890&dopt=Abstracthttp://dx.doi.org/10.1038/nm1134http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=18280243&dopt=Abstracthttp://dx.doi.org/10.1016/j.molcel.2007.12.026http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=20385364&dopt=Abstracthttp://dx.doi.org/10.1016/j.ccr.2009.12.050http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21593872&dopt=Abstracthttp://dx.doi.org/10.1038/nature09883

57. X. An, V. P. Schulz, J. Li, K. Wu, J. Liu, F. Xue, J. Hu, N. Mohandas, P. G. Gallagher,

Global transcriptome analyses of human and murine terminal erythroid differentiation.

Blood 123, 3466–3477 (2014). Medline doi:10.1182/blood-2014-01-548305

58. J. Li, J. Hale, P. Bhagia, F. Xue, L. Chen, J. Jaffray, H. Yan, J. Lane, P. G. Gallagher, N.

Mohandas, J. Liu, X. An, Isolation and transcriptome analyses of human erythroid

progenitors: BFU-E and CFU-E. Blood 124, 3636–3645 (2014). doi:10.1182/blood-2014-

07-588806

59. S. Gupta, J. A. Stamatoyannopoulos, T. L. Bailey, W. S. Noble, Quantifying similarity

between motifs. Genome Biol. 8, R24 (2007). Medline doi:10.1186/gb-2007-8-2-r24

60. T. L. Bailey, C. Elkan, Fitting a mixture model by expectation maximization to discover

motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 2, 28–36 (1994). Medline

61. T. L. Bailey, DREME: Motif discovery in transcription factor ChIP-seq data. Bioinformatics

27, 1653–1659 (2011). Medline doi:10.1093/bioinformatics/btr261

62. T. L. Bailey, P. Machanick, Inferring direct DNA binding from ChIP-seq. Nucleic Acids Res.

40, e128 (2012). Medline doi:10.1093/nar/gks433

63. V. Matys, O. V. Kel-Margoulis, E. Fricke, I. Liebich, S. Land, A. Barre-Dirrie, I. Reuter, D.

Chekmenev, M. Krull, K. Hornischer, N. Voss, P. Stegmaier, B. Lewicki-Potapov, H.

Saxel, A. E. Kel, E. Wingender, TRANSFAC and its module TRANSCompel:

Transcriptional gene regulation in eukaryotes. Nucleic Acids Res. 34, D108–D110 (2006).

Medline doi:10.1093/nar/gkj143

64. A. Mathelier, X. Zhao, A. W. Zhang, F. Parcy, R. Worsley-Hunt, D. J. Arenillas, S.

Buchman, C. Y. Chen, A. Chou, H. Ienasescu, J. Lim, C. Shyr, G. Tan, M. Zhou, B.

Lenhard, A. Sandelin, W. W. Wasserman, JASPAR 2014: An extensively expanded and

updated open-access database of transcription factor binding profiles. Nucleic Acids Res.

42, D142–D147 (2014). Medline doi:10.1093/nar/gkt997

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24637361&dopt=Abstracthttp://dx.doi.org/10.1182/blood-2014-01-548305http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=17324271&dopt=Abstracthttp://dx.doi.org/10.1186/gb-2007-8-2-r24http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=7584402&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=21543442&dopt=Abstracthttp://dx.doi.org/10.1093/bioinformatics/btr261http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=22610855&dopt=Abstracthttp://dx.doi.org/10.1093/nar/gks433http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16381825&dopt=Abstracthttp://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=16381825&dopt=Abstracthttp://dx.doi.org/10.1093/nar/gkj143http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=24194598&dopt=Abstracthttp://dx.doi.org/10.1093/nar/gkt997

aad3312MasudaRefs.pdfReferences

SOM.page.1.EXPRESS.no.movies.pdfTranscription factors LRF and BCL11A independently repress expression of fetal hemoglobin