POGZ Is Required for Silencing Mouse Embryonic

Article

POGZ Is Required for Silencing Mouse Embryonic b-like Hemoglobin and Human Fetal Hemoglobin Expression Graphical Abstract

Authors Bjorg Gudmundsdottir, Kristbjorn O. Gudmundsson, Kimberly D. Klarmann, ..., Thorunn Rafnar, John F. Tisdale, Jonathan R. Keller

Correspondence [email protected]

In Brief Gudmundsdottir et al. show that POGZ represses embryonic globin gene expression in mouse and human erythroid cells, in part by regulating Bcl11a expression in vitro and in vivo. The molecular pathways regulated by POGZ may represent potential therapeutic targets to increase fetal globin expression in patients with sickle cell disease and b-thalassemia.

Highlights d

d

d

d

Pogz is highly expressed in mouse megakaryocyte erythroid progenitors POGZ is required to repress murine embryonic b-like globin during erythropoiesis Pogz+/ mice develop normally and show elevated embryonic b-like globin expression POGZ knockdown decreases BCL11A and increases fetal globin expression in human cells

Gudmundsdottir et al., 2018, Cell Reports 23, 3236–3248 June 12, 2018 ª 2018 The Authors. https://doi.org/10.1016/j.celrep.2018.05.043

Data and Software Availability GSE113503

Cell Reports

Article POGZ Is Required for Silencing Mouse Embryonic b-like Hemoglobin and Human Fetal Hemoglobin Expression Bjorg Gudmundsdottir,1,9,10 Kristbjorn O. Gudmundsson,1,9,11 Kimberly D. Klarmann,1,2 Satyendra K. Singh,1 Lei Sun,1 Shweta Singh,1 Yang Du,3 Vincenzo Coppola,4 Luke Stockwin,5 Nhu Nguyen,3 Lino Tessarollo,1 Leifur Thorsteinsson,6 Olafur E. Sigurjonsson,6 Sveinn Gudmundsson,6 Thorunn Rafnar,7 John F. Tisdale,8 and Jonathan R. Keller1,2,12,* 1Mouse Cancer Genetics Program, Center for Cancer Research, National Cancer Institute at Frederick, Bldg. 560/12-70, 1050 Boyles Street, Frederick, MD 21702, USA 2Basic Research Program, Leidos Biomedical Research Inc., Frederick National Laboratory for Cancer Research, Bldg. 560/32-31D, 1050 Boyles Street, Frederick, MD 21702, USA 3Department of Pediatrics, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814, USA 4Wexner Medical Center, Ohio State University, 460 West 12thAvenue, Columbus, OH 43210, USA 5Drug Mechanisms Group, Developmental Therapeutics Program, Leidos Biomedical Research, Inc., National Cancer Institute at Frederick, Frederick, MD 21702, USA 6The Blood Bank, Landspitali University Hospital, Snorrabraut 60, 105 Reykjavik, Iceland 7Iceland Genomics Corporation, Snorrabraut 60, 105 Reykjavik, Iceland 8Molecular and Clinical Hematology Branch, NHLBI/NIDDK, NIH, Bethesda, MD 20814, USA 9These authors contributed equally 10Present address: Molecular and Clinical Hematology Branch, NHLBI, NIH, Building 10, Room 9N112, Bethesda, MD 20814, USA 11Present address: Department of Pediatrics, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, MD 20814, USA 12Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2018.05.043

SUMMARY

INTRODUCTION

Fetal globin genes are transcriptionally silenced during embryogenesis through hemoglobin switching. Strategies to derepress fetal globin expression in the adult could alleviate symptoms in sickle cell disease and b-thalassemia. We identified a zinc-finger protein, pogo transposable element with zinc-finger domain (POGZ), expressed in hematopoietic progenitor cells. Targeted deletion of Pogz in adult hematopoietic cells in vivo results in persistence of embryonic b-like globin expression without affecting erythroid development. POGZ binds to the Bcl11a promoter and erythroid-specific intragenic regulatory regions. Pogz+/ mice show elevated embryonic b-like globin expression, suggesting that partial reduction of Pogz expression results in persistence of embryonic b-like globin expression. Knockdown of POGZ in primary human CD34+ progenitor cellderived erythroblasts reduces BCL11A expression, a known repressor of embryonic b-like globin expression, and increases fetal hemoglobin expression. These findings are significant, since new therapeutic targets and strategies are needed to treat b-globin disorders.

During mouse embryonic development, three distinct populations of erythroid cells are generated (Baron et al., 2013). The first are primitive erythroid cells, which arise from the yolk sac and mainly express embryonic b-like globins (Hbb-bh1 and Hbb-y) and low levels of adult-type globins (Hbb-b1 and Hbb-b2) (Kingsley et al., 2006; Palis, 2014). The second are definitive erythroid cells from the yolk sac that seed the fetal liver (FL). Initially, they express Hbb-bh1 and Hbb-y and then switch to Hbb-b1 and Hbb-b2 expression (McGrath et al., 2011). The third population is hematopoietic stem cells (HSCs) that arise from intra-embryonic sites, including the aorta-gonad mesonephros region, that initially seed the FL and then home to the bone marrow and give rise to definitive erythroid cells. These erythroid cells express the adult Hbb-b1 and Hbb-b2 globins. Sickle cell disease (SCD) and b-thalassemia are inherited human hemoglobin disorders, which result from globin gene mutations and represent a significant global health issue. Natural variations in fetal hemoglobin expression have been linked to the severity of disease outcome, such that individuals with higher fetal hemoglobin levels have less severe symptoms in SCD and b-thalassemia (Sankaran et al., 2010a). Genome-wide association studies identified three loci associated with increased fetal hemoglobin levels (Galarneau et al., 2010; Lettre et al., 2008; Uda et al., 2008), including BCL11A, which was subsequently shown to function as a transcriptional repressor of fetal hemoglobin (Sankaran et al., 2008). Conditional loss of Bcl11a

3236 Cell Reports 23, 3236–3248, June 12, 2018 ª 2018 The Authors. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Figure 1. The Zinc-Finger Protein POGZ Is a Nuclear Protein Expressed in Human and Mouse Hematopoietic Cells (A) Real-time qPCR analysis reveals that POGZ expression is downregulated in KG1 cells after treatment with phorbol 12-myristate 13-acetate and tumor necrosis factor a (TNF-a). Experiments were performed in triplicate, and data are presented as mean ± SD (B) Western blot analysis shows reduced POGZ protein levels upon differentiation of KG1 cells. (C) Real-time qPCR analysis of Pogz expression in purified mouse hematopoietic stem cells (HSCs), common lymphoid progenitors (CLPs), common myeloid progenitors (CMPs), granulocyte macrophage progenitors (GMPs), and megakaryocyte erythroid progenitors (MEPs). Gene expression was normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; **p < 0.01. (D) Western blot analysis of POGZ in mouse tissue lysates indicates that POGZ is expressed in hematopoietic cells. (BL, blood; BM, bone marrow; LI, liver; SP, splenocytes; TH, thymocytes). Reduced exposure time is shown for thymus and spleen due to strong POGZ expression. (E) Immunofluorescence analysis of mouse erythroleukemia (MEL) cells. The cells were fixed, permeabilized, and stained with antibodies that recognize POGZ (red) and the nucleolar protein FIBRILLARIN (green). DNA was stained with DAPI (blue). POGZ is detected in the nucleus, but not the nucleolus.

in erythroid cells leads to increased embryonic b-like globin expression without affecting normal erythroid development, suggesting that BCL11A is a relevant therapeutic target (Xu et al., 2011). Recent experimental evidence confirmed that loss of Bcl11a expression in a preclinical model of SCD reversed sickling and end organ damage (Xu et al., 2011). Embryonic b-like globin is maintained in a repressed state by a multi-protein co-repressor complex, including BCL11A, GATA1, SOX6, and chromatin remodeling proteins, including Mi2b, HDAC1/2, LSD1/CoREST, and DMMT1 (Xu et al., 2010, 2013). These and other targets represent therapeutic opportunities to reactivate fetal globin to treat SCD and b-thalassemia (Bauer et al., 2012; Sankaran et al., 2008; Xu et al., 2013). We identified a previously uncharacterized transcriptional regulator of hematopoiesis, POGZ (KIAA0461, ZNF280E), in a screen of a human hematopoietic progenitor cell line model (KG1) and its more differentiated progeny (Gudmundsson et al., 2007). POGZ is a zinc-finger containing protein, which binds to SP1, LEDGF, and heterochromatin proteins (Bartholomeeusen et al., 2009; Gunther et al., 2000; Nozawa et al., 2010), suggesting POGZ may have an important role in gene regulation; however, its function in hematopoiesis is currently unknown (Gudmundsson et al., 2007; Ishikawa et al., 1997; Nomura et al., 1994; Okazaki et al., 2003). Domain structure predictions by SMART analysis (Letunic et al., 2009; Schultz et al., 1998) indicate that POGZ has at least 8 C2H2 zinc fingers, suggesting it can bind DNA (Figure S1A). We show here that Pogz is expressed in normal mouse hematopoietic stem and progenitor cells (HSPCs), with the highest levels of expression in megakaryocyte erythroid progenitors (MEPs). We discovered that POGZ is essential for normal murine embryonic development, and uncovered a function of POGZ in the regulation of embryonic b-like globin expression in vitro and in vivo. Using mouse models that conditionally delete Pogz in adult mice, we demonstrate that Pogz is intrinsically required for normal globin switching, in part, by regulating Bcl11a expression. Furthermore, we show that knockdown of POGZ expression in human erythroid cells derepresses fetal globin expression. Our data provide evidence that Pogz is a regulator of mouse embryonic b-like globin expression and human fetal hemoglobin expression. RESULTS Pogz Is Expressed in Normal Murine Hematopoietic Stem and Progenitor Cells We identified POGZ, a zinc-finger-containing protein, in a screen of the human hematopoietic progenitor cell line KG1, whose potential transcription factor activity and function in hematopoiesis were unknown (Gudmundsson et al., 2007). POGZ is expressed in KG1 cells, and POGZ RNA and protein levels are decreased during differentiation, suggesting that POGZ may function in hematopoietic cells (Figures 1A and 1B). POGZ is highly conserved from zebrafish to humans, with 90% homology in coding sequence between human and mouse and 94% homology in amino acid sequence (Figure S1A). We performed a detailed analysis of Pogz expression in purified mouse HSPC populations and confirmed that Pogz is expressed in normal HSCs and common lymphoid progenitors

Cell Reports 23, 3236–3248, June 12, 2018 3237

Figure 2. Pogz / Embryos Are Runted, with Reduced Numbers of Fetal Liver Cells (A) Pogz / mice do not survive beyond birth. Number and ratio of Pogz+/+, Pogz+/ and Pogz / progeny from intercrosses of mixed background (C57BL/6J 3 1291/SVImJ mice) Pogz+/ mice. The percentages within the parenthesis indicate observed progeny, and the expected outcome for these crosses is +/+ 25%, +/ 50%, and / 25%. (B) Gross morphology of E15.5 Pogz+/+ and Pogz / embryos. (C) Total fetal liver cellularity in E16.5 Pogz+/+ (n = 2), Pogz+/ (n = 2) and Pogz / (n = 4) embryos. Data are presented as mean ± SD. **p < 0.01. These data are representative of 5 separate litters.

(CLPs) and is reduced in common myeloid progenitors (CMPs) and granulocyte macrophage progenitors (GMPs), while Pogz expression is significantly higher in MEPs (Figure 1C). Inquiry of Pogz RNA expression in the BioGPS microarray database (Su et al., 2004) confirmed that Pogz is highly expressed in murine HSPCs and increased in MEPs and that Pogz is more broadly expressed in other tissues, including neural and eye tissue (Figure S1B). Finally, we compared POGZ protein expression in a limited tissue array and found that POGZ protein is highly expressed in adult mouse thymocytes and splenocytes, with lower levels of expression in peripheral blood cells (PBCs), bone marrow cells (BMCs), and liver cells (Figure 1D). Since POGZ is expressed in erythroid lineage cells, we used mouse erythroid leukemia (MEL) cells to examine the expression and subcellular localization of POGZ by immunofluorescence and determined that POGZ is mainly localized in the nucleus and is not present in the nucleolus (Figure 1E). Collectively, these results confirm that POGZ is expressed in normal mouse HSPCs, MEPs, and MEL cells, suggesting a potential role for POGZ in megakaryopoiesis and erythropoiesis. Reduced Output of Hematopoietic Cells and Deregulation of Genes Required for Erythropoiesis and Hemoglobin Switching in Pogz / FL Cells To uncover the physiological function of POGZ in hematopoietic development, we generated a mouse model to inactivate Pogz gene expression in vivo (Figures S1C–S1F) (Liu et al., 2003).

3238 Cell Reports 23, 3236–3248, June 12, 2018

We did not detect Pogz / pups at weaning in crosses of Pogz+/ mice, suggesting that the Pogz / mice died during embryonic development or shortly after birth. Further analysis showed that Pogz / embryos rarely survived beyond embryonic day 16.5 (E16.5) when backcrossed 10 generations onto C57BL/6J background mice. Timed-pregnancy studies showed that some Pogz / embryos were absorbed as early as E10.5, but we observed a consistent drop in animal survival around E15.5 (Figure 2A). The Pogz / embryos were generally smaller and appeared anemic compared to their wild-type littermates; however, the precise cause of death is currently unknown (Figure 2B). Since the FL is the major site of hematopoiesis in the embryo and Pogz / embryos survive until E15.5–E16.5, we harvested FL cells from Pogz+/+ and Pogz / embryos to examine lineage development by flow cytometry and performed differential gene expression analysis to identify potential pathways and target genes affected by the loss of Pogz. As expected, Pogz / FL was significantly smaller, with fewer cells (Figure 2C). We found that the frequency of myeloid (macrophages [Mac1+Gr1 ] and neutrophils [Mac1+Gr1+]) and B cells (CD19+) in Pogz / FL was similar to Pogz+/+ FL; however, their total numbers were reduced due to the overall reduction in FL cellularity (Figures S2A–S2C). We observed increased frequencies of more primitive erythroid cells (S0–S2 cells) in Pogz / FL (Figure S2D); however, the total number of the more mature erythroid cells was decreased at all stages of development (S3–S5) in Pogz / FL (Figure S2E). Thus, erythroid, B, and myeloid cells are present in Pogz / FL, but in greatly reduced numbers. Concomitantly, we performed microarray analysis of RNA expressed in Pogz+/+ and Pogz / E14.5 FL cells to identify potential target genes and pathways affected by the loss of Pogz. We found that 1,062 genes were differentially expressed in Pogz / versus Pogz+/+ FLs using >1.5-fold change in gene expression as a cutoff (Table S1). Ingenuity Pathway Analysis (IPA) of differentially expressed genes identified the ‘‘hematological system and development and function’’ as a top physiological system

Figure 3. Decreased Expression of Bcl11a and Increased Expression of Embryonic b-like Globins in Pogz / Fetal Liver Cells (A) Heatmap of differentially expressed genes in Pogz+/+ and Pogz / FL cells, indicating that genes expressed during erythroid lineage development are affected by loss of Pogz expression. RNA was isolated from E14.5 Pogz+/+ and Pogz / FL cells (n = 3 for each genotype), and gene expression was analyzed on Affymetrix Mouse 430 2.0 oligonucleotide arrays (blue indicates low expression, and red indicates high expression). (B) Bcl11a and Klf2 expression is downregulated in Pogz / FL cells. RNA was purified from E16.5 Pogz+/+ and Pogz / FLs, and Bcl11a, Klf2, Nfe2, Klf1, Gata1, and Fog1 expression was analyzed by real-time qPCR. Gene expression was normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001. (C) Reduction in BCL11A protein levels in Pogz / FL cells. Western blot analysis was performed on whole-cell lysates generated from E15.5 Pogz+/+ and Pogz / FL cells. (D) Upregulation of embryonic b-like globins in Pogz / FL cells. RNA was purified from E16.5 Pogz+/+ and Pogz / FLs and Hbb-y, Hbb-bh1, Hba-x, and Hbb-b1/2 expression analyzed by realtime qPCR. Gene expression was normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

affected, and ‘‘hematological disease’’ as a top disease and disorder affected (Figure S2F). Differentially expressed genes were linked to erythrocytosis and hereditary persistence of fetal hemoglobin (HPFH) and included Jak2, c-kit, and c-myb (all upregulated) and Bcl11a and Tfrc (both downregulated) in Pogz / FL cells compared to Pogz+/+ FL cells (Figure 3A). Taken together, the data from the microarray analysis suggests that loss of Pogz may lead to deregulation of erythropoiesis and globin gene expression. Loss of Pogz Expression Leads to Downregulation of Bcl11a and Increased Embryonic b-like Globin Expression Since Pogz is highly expressed in MEPs, and loss of Pogz expression affects the expression of genes and pathways required for erythroid development and globin gene expression (Figure 3A), we determined the expression of known transcriptional regulators of erythroid and globin gene expression, including Klf1, Klf2, Nfe2, Gata1, Fog1, and Bcl11a, at E16.5 in

Pogz+/+ and Pogz / FL by real-time qPCR. We found that the expression of Bcl11a and Klf2 was significantly reduced in RNA obtained from E16.5 Pogz / FL cells compared to Pogz+/+ FL cells, whereas the expression of Klf1, Gata1 and Fog1 was not significantly different (Figure 3B). We also confirmed a decrease of Bcl11a expression at the protein level in E15.5 Pogz / FL by western blot analysis (Figure 3C). Since BCL11A is a critical regulator of the switch between fetal and adult globin expression in definitive erythroid cells (Sankaran et al., 2008, 2009, 2010b), we analyzed the expression of the embryonic a- and b-like globins Hbb-y, Hbb-bh1, and Hba-x in E16.5 Pogz / FL RNA. We found that the expression of the embryonic globins was significantly upregulated in the Pogz / FL compared to Pogz+/+ FL (Figure 3D). Thus, expression of Bcl11a is decreased and embryonic globin expression is increased in Pogz / FL cells, suggesting that Pogz may function to regulate embryonic globin gene expression. Persistence of Embryonic Globin Expression Is Intrinsic to Pogz / Hematopoietic Cells To determine if the increased expression of embryonic globin observed in Pogz / FL cells was intrinsic to hematopoietic cells and not an indirect effect due to loss of Pogz function in the microenvironment, we transplanted 1 3 106 E15.5 Cell Reports 23, 3236–3248, June 12, 2018 3239

Figure 4. Persistence of Embryonic b-like Globin Expression Is Intrinsic to Pogz / Erythroid Cells (A) Summary of FL transplantation experiments. Hbb-y RNA expression was increased and Bcl11a decreased in PBCs obtained from mice 5 weeks after transplantation of Pogz+/+ or Pogz / FL cells. Gene expression was determined by realtime qPCR and normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. (B) Representative flow cytometry analysis of CD71 and Ter119 expression in mice transplanted with Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre BMCs 12 weeks post-pIpC treatment (n = 5 for each genotype). Gates were set around subsets of differentiating donor (CD45.2+ expression) bone marrow (BM) erythroid cells (S0–S5) as previously described (Koulnis et al., 2011). No differences in the percentages of donor-derived erythroid subsets were observed. (C and D) Pogz mRNA is not expressed in PBCs from mice transplanted with Pogzf/f; Mx1-cre BMCs and treated with pIpC compared to mice transplanted with control Pogz+/+; Mx1-cre BMCs (C), while Hbb-y mRNA expression is significantly increased in PBCs from mice transplanted with Pogzf/f; Mx1-cre BMCs (D). Experiments were performed in triplicate. Gene expression was normalized to b-actin expression.

Pogz+/+ and Pogz / FL cells into lethally irradiated recipient mice (Figure 4A). We found no difference in the number of differentiating erythroid cells at all stages of development (S0–S5) in the bone marrow of mice transplanted with Pogz / and Pogz+/+ FL cells (Figure S3A). Since CD45 is expressed on more primitive S1, S2, and S3 cells and is decreased during differentiation (S4–S5 cells), we gated on donor-derived CD45.2+ cells in mice transplanted with FL cells. Mice transplanted with Pogz / FL showed reduced numbers of S1/S2 cells and no difference in the number of S3, S4, or S5 cells compared to mice transplanted with Pogz / FL cells, suggesting relatively normal red cell output after transplantation (Figure S3B). We obtained RNA from PBCs 5 weeks after transplantation (short-term reconstitution) and analyzed Hbb-y and Bcl11a


expression by real-time qPCR in recipient mice reconstituted with Pogz+/+ or Pogz / hematopoietic cells. Hbb-y expression was significantly upregulated in all recipient mice that received Pogz / FL cells, while Hbb-y expression was silenced in recipient mice transplanted with Pogz+/+ FL cells (Figure 4A, left panel). The expression of Bcl11a was downregulated in mice transplanted with Pogz / FL cells in comparison to mice transplanted with Pogz+/+ FL cells (Figure 4A, right panel). Furthermore, analysis of Hbb-y expression 3 months post transplantation (long-term reconstitution) confirmed that Hbb-y expression remained elevated in mice transplanted with Pogz / FL cells, indicating that Hbb-y expression is not silenced in erythroid cells arising from long-term reconstituting HSPCs (Figure S4A). Collectively, these data suggest that loss of Pogz leads to an intrinsic derepression of embryonic b-like globin expression. While transplantation of Pogz / FL cells provides evidence for an intrinsic role of Pogz in regulating embryonic globin gene expression, we sought to confirm this in a model where we could delete Pogz in hematopoietic lineage cells in adult mice and limit the potential of non-cell-autonomous effects. First, we bred Pogz conditional mice (Pogzf/f) to EpoRcre transgenic mice (Heinrich et al., 2004) (Pogzf/f; EpoR-cre); however, Pogz was variably deleted in this model (data not

Figure 5. Pogz Regulates the Expression of Bcl11a and Hbb-y in MEL Cells and Fetal Liver Cells (A) Hbb-y expression is increased in PBC from Pogz+/ compared to Pogz+/+ mice. Gene expression was normalized to b-actin expression (n = 5 per group and data are presented as mean ± SD, *p < 0.05). (B) Lentiviral-mediated knockdown of Pogz in MEL cells represses Bcl11a expression and induces Hbb-y expression. RNA was harvested from MEL cells 72 and 96 hr after transduction with lentiviral vector expressing Pogz shRNA or a control shRNA vector and expression of Pogz, Bcl11a, and Hbb-y analyzed by real-time qPCR. Gene expression was normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; ***p < 0.001. (C) Western blot analysis of POGZ, BCL11A, and ACTIN expression following lentiviral-mediated knockdown of Pogz in MEL cells. Knockdown was performed with shRNA targeting Pogz or a control shRNA vector, and cell lysates were harvested 72 hr post-transduction. (D) Photomicrographs of cytocentrifuge preparations of MEL cells 72 hr after lentiviral-mediated knockdown indicating no effect on cell morphology.

(legend continued on next page)


included), which precluded using these mice for further studies. Therefore, we bred Pogzf/f mice to Mx1-cre mice and transplanted BMCs from Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre into irradiated recipients to generate chimeric mice €hn et al., 1995). Six weeks after bone marrow transplanta(Ku tion (BMT), we treated mice with polyinosinic:polycytidylic acid (pIpC) to delete Pogz in hematopoietic cells. Twelve weeks after Pogz deletion we analyzed (1) BMCs to confirm that Pogz was deleted, (2) PBCs for complete blood cell (CBC) analysis, (3) BMCs for MEP and erythroid development, and (4) PBCs for expression of Pogz and Hbb-y globin. Pogz was efficiently deleted in the chimeric Pogzf/f; Mx1-cre transplanted mice (Figure S4B), and results of CBC analysis of mice 12 weeks after deletion of Pogz were normal, suggesting that Pogz is not required for normal red cell development in this model (Figure S4C). No differences in donor myeloid and B cell reconstitution were observed in mice transplanted with BMCs that lack Pogz (Figure S5A). No difference in the frequency or number of differentiating erythroid cells (S1–S5) was observed in BMCs from Pogzf/f; Mx1-cre transplanted mice compared to control transplanted mice (Figure S5B). Further, no significant differences in donorderived erythroid cell reconstitution were observed in mice transplanted with Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre BMC when gated on donor-derived CD45.2+ cells, providing additional evidence that loss of Pogz does not affect normal adult erythroid cell development (Figure 4B). Since MEPs are restricted progenitors for erythroid cells and platelets and express significant levels of Pogz, we examined if loss of Pogz affects their development. We found no differences in the number of donor-derived MEPs in BMCs of mice that lack Pogz (Figure S5C). In addition, there were no differences in the number of megakaryocyte progenitors (MkPs) and premagakaryocyte/erythroid progenitors in mice transplanted with Pogz / and Pogz+/+ BMCs (Figure S5D), suggesting that Pogz is not required for the development of erythroid or megakaryocyte progenitors. Finally, we analyzed the expression of Pogz, Hba-x, Hbb-bh1, and Hbb-y in PBCs of transplanted mice. Our data confirmed that Pogz is not expressed (Figure 4C), while Hbb-y is expressed in all mice transplanted with Pogzf/f; Mx1-cre BMCs (Figure 4D). We also found an upregulation of Hba-x and Hbb-bh1 expression in some of the animals (Figure S6A–S6D). Collectively, these data demonstrate that Pogz is intrinsically required to repress embryonic b-like globin Hbb-y in adult red blood cells. Interestingly, we found that the levels of Hbb-y expression were significantly increased in PBCs from adult Pogz+/ mice in comparison to Pogz+/+ mice, suggesting that Pogz-mediated repression of Hbb-y expression is dependent on the levels of Pogz expression in adult mice (Figure 5A). Furthermore, since Pogz+/ mice are viable, reproduce, and show no overt phenotype, the data suggest that reducing Pogz levels could result in

persistence of embryonic globin expression without significantly altering erythroid maturation. Pogz Negatively Regulates Hbb-y Expression, in Part through Bcl11a Our data suggest the possibility that POGZ represses Hbb-y expression, in part by regulating Bcl11a expression. To determine if Pogz and Bcl11a are coexpressed and developmentally regulated in a similar fashion during erythroid differentiation, we analyzed the expression profile of Pogz and Bcl11a in sorted erythroid cells from the bone marrow of normal C57BL/6J mice (Figure S6E). Our analysis shows that Pogz and Bcl11a are expressed at similar levels in all erythroid populations (Figures S6F and S6G). As a comparison, Gata1 and Klf1 are highly expressed in CD71HITer119+ cells (Figures S6H and S6I). To investigate if Pogz regulates the expression levels of Bcl11a and Hbb-y, we knocked down Pogz expression in MEL cells using lentiviral-mediated delivery of Pogz-specific small hairpin RNA (shRNA) and analyzed gene expression 72–96 hr post-transduction. The knockdown resulted in significant reduction in Pogz mRNA transcripts (Figure 5B) and loss of POGZ protein expression (Figure 5C), which did not affect MEL cell differentiation (Figure 5D) compared to control-treated cells. Loss of Pogz expression in MEL cells resulted in decreased Bcl11a expression and increased Hbb-y expression levels (Figures 5B and 5C). In addition, we overexpressed Pogz in E16.5 Pogz / FL by retroviral transduction and examined Bclla and Hbb-y expression (Figure 5E). Enforced expression of Pogz resulted in upregulation of Bcl11a expression and repression of Hbb-y expression. Taken together, these data suggest that Pogz positively regulates Bcl11a and represses Hbb-y expression. Interestingly, overexpression of Bcl11a did not reduce Hbb-y expression in Pogz / E16.5 FL cells, which suggests that Pogz may be required for Bcl11a-mediated repression and that there are additional mechanisms by which Pogz represses mouse embryonic b-like globin (Figure S6J). To further examine the requirement for BCL11A in Pogzmediated regulation of Hbb-y expression, we performed double knockdown experiments in MEL cells. We found that reducing either Pogz or Bcl11a expression increases Hbb-y expression with Bcl11a knockdown showing more efficient derepression of Hbb-y expression (Figure 6A). Knocking down both Pogz and Bcl11a did not increase Hbb-y expression above knocking down Bcl11a alone (Figure 6A). To examine if Bcl11a could rescue the upregulation of Hbb-y following Pogz knockdown, we overexpressed Bcl11a in MEL cells transduced with lentiviral vectors that express shPogz. We found that enforced Bcl11a expression partially reduces the increase in Hbb-y expression mediated by Pogz knockdown in MEL cells (Figure 6B), indicating that the Hbb-y derepression upon loss of Pogz is mediated, in part through BCL11A, although additional mechanisms are likely involved in this model.

(E) Re-expression of Pogz in Pogz / FL cells induces Bcl11a and reduces Hbb-y RNA expression. Pogz / FL cells were harvested at E16.5 and transduced with a control retrovirus or retroviral vector expressing the Pogz transgene. RNA was harvested 60 hr post-transduction, and expression of Hbb-y, Bcl11a, and Pogz were analyzed by real-time qPCR. Gene expression was normalized to b-actin expression. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; ***p < 0.001.


Figure 6. POGZ Regulates Hbb-y Expression, in Part by Regulating Bcl11a Expression (A) Western blot analysis of POGZ, BCL11A, and ACTIN protein levels, and real-time qPCR analysis of Hbb-y expression following lentiviral-mediated knockdown of Pogz, Bcl11a, or both Pogz and Bcl11a in MEL cells. MEL cells were transduced with shRNA targeting Pogz, Bcl11a, or a control shRNA vector and cell lysates harvested 72 hr posttransduction. Gene expression was normalized to b-actin expression. Experiments were performed in triplicates and data are presented as mean ± SD. PU, puromycin. BL, blasticidin. (B) Western blot analysis of POGZ, BL11A and ACTIN protein levels and real-time qPCR analysis of Hbb-y expression following lentiviral mediated knockdown of Pogz and overexpression of Bcl11a. Gene expression was normalized to b-actin expression. Experiments were performed in triplicates and data are presented as mean ± SD. *p < 0.05. (C) ChIP-qPCR analysis demonstrating that POGZ binds to the Bcl11a promoter ( 972) and enhancer site (+58). POGZ does not bind to a negative control region on chromosome 17. Sheared chromatin was prepared from MEL cells following lentiviralmediated knockdown of Pogz. Knockdown was performed with shRNA targeting Pogz or a control shRNA vector, and cell lysates were harvested 72 hr post-transduction. Chromatin was immunoprecipitated with an anti-POGZ antibody and a control antibody. The Bcl11a gene indicating sites relative to the transcription start site (TSS) that were examined by ChIP is shown below. Experiments were performed in triplicate, and data are presented as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

To test whether POGZ can repress Hbb-y through direct regulation of Bcl11a, we performed chromatin immunoprecipitation (ChIP) assays using MEL cells transduced with control or Pogz shRNA lentiviral vectors. POGZ binds to the Bcll1a promoter in

MEL cells at 972 and is greatly reduced at 3,469, and POGZ binding to these sites is significantly reduced in cells treated with Pogz shRNA (Figure 6C). POGZ did not show significant binding to a negative control region on mouse chromosome 17. We also examined a recently described enhancer element within intron 2 of the human BCL11A gene, which contains three DNase-I-hypersensitive sites at +55, +58, and +62 kb from the transcription start site (Bauer et al., 2013; Canver et al., 2015). The mouse Bcl11a gene has orthologous sequences within intron 2 (Bauer et al., 2013). We determined that POGZ binds the orthologous +58 sequences within the enhancer region using ChIP assays (Figure 6C) and that POGZ binding to these sites is reduced when Pogz levels are lowered by shRNA. Taken together, our data suggest that POGZ may be a regulator of Hbb-y expression by directly or indirectly interacting with the Bcl11a promoter and the orthologous Bcl11a enhancer elements.


Figure 7. Lentiviral-Mediated Knockdown of POGZ Expression in Human Proerythroblasts Decreases BCL11A Expression and Increases Fetal Hemoglobin Expression (A) Western blot analysis of POGZ, BCL11A, and ACTIN proteins in cell lysates on days 3–14 of erythroid cultures. Numbers below POGZ and BCL11A bands indicate normalized protein levels in relation to ACTIN. (B) Western blot analysis of POGZ, HBG, and ACTIN protein levels in CD34+ cell-derived erythroblasts following lentiviral-mediated knockdown using two separate POGZ shRNA constructs. Knockdown was performed at day 2 of expansion culture with control shRNA or shRNAs targeting POGZ, and cell lysates were harvested 10 days post-transduction. (C) Flow cytometry analysis of CD71 and CD235a expression of erythroid cell cultures in vitro following lentiviral-mediated knockdown using two separate POGZ shRNA constructs. Knockdown was performed at day 2 of expansion culture with control shRNA or shRNAs targeting POGZ, and cells were analyzed 10 days post-transduction. (D) Western blot analysis of POGZ, BCL11A, and ACTIN protein levels in erythroid cell cultures in vitro following lentiviral-mediated knockdown of POGZ or BCL11A. Knockdown was performed at day 2 of expansion culture with control shRNA or shRNA targeting POGZ or BCL11A, and cell lysates were harvested 9 days post-transduction. (E) Representative bar graphs showing fetal hemoglobin (HBG1/2) as a percentage total b-globin (HBB+HBE+HBG1/2) expression following lentiviral-mediated shRNA knockdown of POGZ. Erythroblasts were harvested on day 9 posttransduction (day 11 of culture), and the quantities of HBB, HBE, and HBG proteins were determined by HPLC.

Reduction of POGZ Expression in Human Proerythroblasts Leads to Increased Expression of Fetal Hemoglobin To investigate whether POGZ regulates fetal globin expression in human cells, we examined erythroid cells differentiated from adult human CD34+ HSPCs in a modified two-phase in vitro erythroid culture system as previously described (Migliaccio et al., 2002). We confirmed that the purified CD34+ progenitors undergo erythroid differentiation in vitro by flow cytometry using CD34, CD45, CD71, and CD235a antibodies (Figure S7A). After 2 days, the cells in these cultures are predominantly primitive CFU-E and proerythroblasts that are CD34+CD45+ CD235a CD71low. Most cells undergo further differentiation to more mature CD34 CD45 CD71+CD235a+ polychromatic and orthochromatic erythroblasts after 11 days in culture (Figure S7A). In addition, compared to undifferentiated control


cells (Figure S7B, left panel), the cell pellets became increasingly red, an indication of increased hemoglobinization and erythroid differentiation (Figure S7B, right panel). We found that POGZ and BCL11A proteins are highly expressed in cells from day 3 cultures (proerythroblasts) and that their expression declines during erythroid differentiation, with little expression after 12–14 days in culture (Figure 7A). To investigate whether POGZ regulates fetal hemoglobin (HBG1/2) expression in human cells, we transduced primary human CD34+ cells on day 2 of expansion culture with a control lentiviral shRNA vector or lentiviral shRNA vectors targeting different regions of human POGZ coding sequence and measured fetal hemoglobin expression 10 days after transduction by real-time qPCR and western blot. Real-time qPCR analysis showed that both POGZ shRNA constructs significantly reduced POGZ transcript levels compared to control shRNA (Figure S7C). HBG1/2 expression was significantly upregulated and HBB expression was significantly downregulated, whereas HBE and HBA expression were marginally affected (Figure S7C).

Western blot analysis showed efficient reduction in POGZ protein expression in cultures transduced with POGZ shRNA constructs (Figure 7B), with robust increase in HBG1/2 protein levels, while POGZ or HBG1/2 protein expression was not affected in cultures transduced with control shRNA (Figure 7B). Knockdown of POGZ expression did not affect erythroid differentiation in these cultures, since there was no difference in the extent of erythroid differentiation (CD71+CD235a+ cells) in cultures treated with control shRNA and POGZ shRNA lentiviral vectors after 12 days (Figure 7C). In agreement with the mouse data above (Figure 5), BCL11A protein levels were significantly reduced upon POGZ lentiviral-vector-mediated knockdown, suggesting that POGZ may also regulate BCL11A expression in human erythroid cells (Figure 7D). In comparison, POGZ protein levels were minimally affected following lentiviral-vector-mediated BCL11A knockdown (Figure 7D). As expected, lentiviral-vector-mediated BCL11A knockdown signficantly increased HBG1/2 expression (Figure S7D). HBE expression was also significantly upregulated and HBB and HBA expression significantly downregulated (Figure S7D). Finally, knockdown of POGZ results in HBG1/2 protein levels representing roughly 25% of total b-globin, as assessed by high-performance liquid chromatography (HPLC), which is therapeutically relevant (Figure 7E). Taken together, the results suggest that POGZ is a repressor of fetal hemoglobin expression in humans. DISCUSSION In this report, we identified a previously uncharacterized zincfinger-containing protein, POGZ, which is expressed in mouse and human HSPCs and required to repress embryonic hemoglobin gene expression during normal hematopoietic development. Elevated embryonic globin expression correlated with reduced expression of Bcl11a, a known repressor of embryonic b-like globin expression, in Pogz / FL cells. We demonstrate, in two different animal models, that red cells develop normally in the absence of Pogz in vivo but that the red blood cells (RBCs) show increased embryonic globin expression. Thus, deregulation of embryonic globin expression is intrinsic to Pogz / hematopoietic cells, and embryonic globin expression can persist in adult mice after transplantation. Finally, we show that POGZ knockdown increases fetal globin expression in primary human erythroblasts, indicating that POGZ also regulates human fetal globin expression, which is the focus of our future studies. These findings are significant since improved therapeutic strategies are needed to treat hereditary globin disorders (Weatherall, 2010). Individuals affected by these diseases have moderate to severe anemia and other serious health issues; however, natural variations that result in HPFH expression are linked to lessening the severity of disease (Sankaran et al., 2010a). Therefore, POGZ may represent a potential therapeutic target to increase fetal globin expression in patients with SCD and b-thalassemia (Bauer et al., 2012). We found that Bcl11a expression is reduced in Pogz / FL cells, MEL cells treated with Pogz shRNA, and human CD34+ progenitors treated with POGZ shRNA, suggesting that Pogz positively regulates Bcl11a expression in mouse and human

cells. POGZ binding to the Bcl11a promoter and a recently identified intron 2 enhancer (Bauer et al., 2013; Canver et al., 2015) suggests that POGZ is directly regulating Bcl11a transcription. Future experiments, including ChIP sequencing and electrophoretic mobility shift assays (EMSAs), will determine if this regulation is direct or indirect via interaction with other DNA binding proteins. We also show that BCL11A does not repress embryonic b-like globin Hbb-y when Bcl11a is overexpressed in Pogz / FL erythroblasts, suggesting that POGZ may regulate embryonic globin expression by mechanisms other than regulation of Bcl11a expression. Since b-actin-Cre-mediated deletion of Pogz occurs early in development, the absence of Pogz could affect expression of genes other than Bcl11a, which could permanently affect BCL11A’s ability to properly function in these cells. Alternatively, BCL11A may require POGZ expression to repress embryonic globin gene expression. In support of this hypothesis, overexpression of Bcl11a in MEL cells transduced with shPogz RNA, where Pogz expression has been knocked down to 10%–15% of control Pogz expression levels, leads to partial repression of Hbb-y expression. BCL11A has been the focus of numerous studies to find unique therapeutic targets in SCD and b-thalassemia (Sankaran et al., 2008, 2009, 2010b). It is hypothesized that reduction of BCL11A expression in patients with SCD and b-thalassemia could lead to derepression of fetal hemoglobin, thereby alleviating the symptoms of these disorders (Bauer et al., 2012). The decrease in Bcl11a expression and loss of repression of b-like embryonic globin expression upon loss of Pogz indicates that POGZ may have the same therapeutic potential. Mice reconstituted with Pogzf/f; Mx-1-cre BMCs survive and show normal development of donor-derived MEP and erythroid lineage cells (S1–S5), as well as lymphoid and myeloid cells, suggesting that inhibiting POGZ function in adults would not have deleterious effects on the host hematopoietic system. However, additional studies are needed to determine whether loss of POGZ function in adults can affect HSC and multipotent progenitor development and function or affect other systems. Interestingly, we found that Pogz+/ mice, which develop normally and show no overt phenotypes, show increased embryonic globin expression levels in PBCs. Importantly, partial reduction of POGZ in human erythroblasts also derepressed fetal hemoglobin expression to levels reaching over 25% of total b-like globin. These results suggest that complete ablation of POGZ in vivo may not be required to obtain therapeutic benefits. Further in vitro and in vivo studies are needed to determine if this is feasible. KLF1 is a master regulator of erythroid development and b-globin expression (Perkins et al., 1995). Klf1 knockout (KO) mice die in utero around E15 due to defects in the differentiation of erythroid cells at the pro-erythroblast stage (Nuez et al., 1995; Perkins et al., 1995; Pilon et al., 2008). KLF1 represses the expression of embryonic globins by upregulating the expression of Bcl11a and promotes adult b-globin expression in definitive erythroid cells (Tallack and Perkins, 2013; Zhou et al., 2010). Klf1 expression was not affected by loss of Pogz expression in FL cells. However, global gene expression analysis of Klf1 / erythroid progenitors demonstrates that Pogz is among the significantly downregulated genes (Pilon et al., 2008). Furthermore,


analysis of submitted ChIP-sequencing data suggests that KLF1 binds the POGZ promoter in human primary erythroid cells, indicating that POGZ may be a direct KLF1 target (Su et al., 2013). Recent evidence suggests that POGZ may have physiological functions in other systems. Specifically, studies analyzing the genetic basis of autism spectrum disorders (ASDs) and intellectual disability have detected inactivating mutations in POGZ in some of these patients (Iossifov et al., 2012; Stessman et al., 2016; Tan et al., 2016; White et al., 2016). Our preliminary studies also suggest a function for Pogz in the mammalian neural system, since loss of Pogz affects the proliferation of mouse neural progenitor cells in fetal and adult brain (K.O.G., unpublished data). Interestingly, potential disrupting mutations in the BCL11A gene have been found in ASDs, and BCL11A has been implicated in neuronal morphogenesis (Iossifov et al., 2012; John et al., 2012). In addition, it was shown in two separate studies that individuals presenting with ASD and developmental delay had common microdeletions of BCL11A rendering them haploinsufficient for the gene. Interestingly, these individuals have elevated expression of fetal hemoglobin (Basak et al., 2015; Funnell et al., 2015). These data suggest that BCL11A and POGZ could function within the same regulatory networks in the neural system. In summary, our data show that POGZ is essential for normal embryonic development and that loss of the gene leads to deregulation of embryonic globin expression, in part through Bcl11a. Reduction of POGZ expression in erythroid cells could have therapeutic implications in SCD and b-thalassemia. EXPERIMENTAL PROCEDURES Mice Conventional Pogz / mice and conditional Pogzf/f, and Pogzf/f; Mx1-cre mice were generated as described in Supplemental Experimental Procedures. Female mice aged 8–12 weeks were used as recipients for all transplantation experiments. Mice were housed, fed, and handled in accordance with the National Institutes of Health guidelines for animal care and use and the Guide for the Care and Use of Laboratory Animals, 8th Edition. All mouse experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the National Cancer Institute at Frederick, which is accredited by Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International. Real-Time qPCR Analysis of POGZ expression in KG1 cells by real-time qPCR was performed as described previously (Gudmundsson et al., 2007). Globin expression was analyzed in CD34+ HSPC-derived human erythroblasts using Taqman assays. For mouse FL cells, PBC and BMC RNA was isolated and real-time qPCR analysis performed in triplicate using Power SYBR Green PCR Master Mix (Life Technologies) and a 7500 Real-Time PCR System (Life Technologies) as previously described (Oakley et al., 2012). The DCt method was used to calculate relative changes in gene expression. Primer sequences are presented in Supplemental Experimental Procedures.

FL and BMC Transplantations FL cells were harvested from E14.5–E16.5 Pogz+/+ and Pogz / embryos, and BMCs were isolated from adult Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre mice and then transplanted as described in Supplemental Experimental Procedures using standard methodologies (Gudmundsson et al., 2012, 2014). Statistical Analysis Statistical analysis was performed using GraphPad Prism (GraphPad Software). An unpaired Student’s t test was used to calculate statistical significance. Results were considered significant if p < 0.05. Results are presented as the mean ± SD. DATA AND SOFTWARE AVAILABILITY The accession number for the microarray data reported in this paper is GEO: GSE113503. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, seven figures, and two tables and can be found with this article online at https://doi.org/10.1016/j.celrep.2018.05.043. ACKNOWLEDGMENTS We wish to thank Mr. Steven Stull and Ms. Terri Stull for excellent animal technical support; Kathleen Noer, Roberta Matthai, and Guity Mohammadi at the NCI-Frederick Flow Cytometry Core for flow cytometric analysis; Ms. Bobbi Smith for CBC analysis; and Dr. Miriam Anver at the NCI-Frederick Histotechnology/Pathology Laboratory for tissue sectioning and staining and pathology analysis. We thank Dr. Matthew Hsieh for information related to Institutional Review Board (IRB) protocols and Dr. Naoya Uchida for technical support regarding human CD34+ cell cultures. This project was funded in part by federal funds from the Frederick National Laboratory for Cancer Research, NIH (contract HHSN261200800001E) and by the intramural research program of the NHLBI and NIDDK, NIH (HL006009-09) and USUHS (R086414217). The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsements by the US government. AUTHOR CONTRIBUTIONS B.G., K.O.G., and J.R.K. designed and conducted experiments, analyzed results, and wrote the manuscript. J.F.T. designed and conducted experiments and analyzed results. K.D.K., S.K.S., L.S., S.S., Y.D., V.C., L.S., N.N., L. Tessarollo, L. Thorsteinsson, O.E.S., S.G., and T.R. designed and conducted experiments. DECLARATION OF INTERESTS The authors declare no competing interests. Received: October 3, 2017 Revised: March 27, 2018 Accepted: May 14, 2018 Published: June 12, 2018 REFERENCES

Flow Cytometry Single-cell suspensions were prepared from Pogz+/+ or Pogz / FLs or from BMCs and PBCs from animals transplanted with Pogz+/+ or Pogz / FL cells or Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMCs. Cells were incubated with the antibodies described in Supplemental Experimental Procedures and then analyzed by FACS-CantoII (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).


Baron, M.H., Vacaru, A., and Nieves, J. (2013). Erythroid development in the mammalian embryo. Blood Cells Mol. Dis. 51, 213–219. Bartholomeeusen, K., Christ, F., Hendrix, J., Rain, J.C., Emiliani, S., Benarous, R., Debyser, Z., Gijsbers, R., and De Rijck, J. (2009). Lens epithelium-derived growth factor/p75 interacts with the transposase-derived DDE domain of PogZ. J. Biol. Chem. 284, 11467–11477.

Basak, A., Hancarova, M., Ulirsch, J.C., Balci, T.B., Trkova, M., Pelisek, M., Vlckova, M., Muzikova, K., Cermak, J., Trka, J., et al. (2015). BCL11A deletions result in fetal hemoglobin persistence and neurodevelopmental alterations. J. Clin. Invest. 125, 2363–2368. Bauer, D.E., Kamran, S.C., and Orkin, S.H. (2012). Reawakening fetal hemoglobin: prospects for new therapies for the b-globin disorders. Blood 120, 2945–2953. Bauer, D.E., Kamran, S.C., Lessard, S., Xu, J., Fujiwara, Y., Lin, C., Shao, Z., Canver, M.C., Smith, E.C., Pinello, L., et al. (2013). An erythroid enhancer of BCL11A subject to genetic variation determines fetal hemoglobin level. Science 342, 253–257. Canver, M.C., Smith, E.C., Sher, F., Pinello, L., Sanjana, N.E., Shalem, O., Chen, D.D., Schupp, P.G., Vinjamur, D.S., Garcia, S.P., et al. (2015). BCL11A enhancer dissection by Cas9-mediated in situ saturating mutagenesis. Nature 527, 192–197. Funnell, A.P., Prontera, P., Ottaviani, V., Piccione, M., Giambona, A., Maggio, A., Ciaffoni, F., Stehling-Sun, S., Marra, M., Masiello, F., et al. (2015). 2p15p16.1 microdeletions encompassing and proximal to BCL11A are associated with elevated HbF in addition to neurologic impairment. Blood 126, 89–93. Galarneau, G., Palmer, C.D., Sankaran, V.G., Orkin, S.H., Hirschhorn, J.N., and Lettre, G. (2010). Fine-mapping at three loci known to affect fetal hemoglobin levels explains additional genetic variation. Nat. Genet. 42, 1049–1051. Gudmundsson, K.O., Thorsteinsson, L., Sigurjonsson, O.E., Keller, J.R., Olafsson, K., Egeland, T., Gudmundsson, S., and Rafnar, T. (2007). Gene expression analysis of hematopoietic progenitor cells identifies Dlg7 as a potential stem cell gene. Stem Cells 25, 1498–1506. Gudmundsson, K.O., Stull, S.W., and Keller, J.R. (2012). Transplantation of mouse fetal liver cells for analyzing the function of hematopoietic stem and progenitor cells. Methods Mol. Biol. 879, 123–133. Gudmundsson, K.O., Oakley, K., Han, Y., and Du, Y. (2014). Analyzing gene function in adult long-term hematopoietic stem cells using the interferon inducible Mx1-Cre mouse system. Methods Mol. Biol. 1194, 313–325. Gunther, M., Laithier, M., and Brison, O. (2000). A set of proteins interacting with transcription factor Sp1 identified in a two-hybrid screening. Mol. Cell. Biochem. 210, 131–142. €ller, U. (2004). A mouse model for Heinrich, A.C., Pelanda, R., and Klingmu visualization and conditional mutations in the erythroid lineage. Blood 104, 659–666.

Letunic, I., Doerks, T., and Bork, P. (2009). SMART 6: recent updates and new developments. Nucleic Acids Res. 37, D229–D232. Liu, P., Jenkins, N.A., and Copeland, N.G. (2003). A highly efficient recombineering-based method for generating conditional knockout mutations. Genome Res. 13, 476–484. McGrath, K.E., Frame, J.M., Fromm, G.J., Koniski, A.D., Kingsley, P.D., Little, J., Bulger, M., and Palis, J. (2011). A transient definitive erythroid lineage with unique regulation of the b-globin locus in the mammalian embryo. Blood 117, 4600–4608. Migliaccio, G., Di Pietro, R., di Giacomo, V., Di Baldassarre, A., Migliaccio, A.R., Maccioni, L., Galanello, R., and Papayannopoulou, T. (2002). In vitro mass production of human erythroid cells from the blood of normal donors and of thalassemic patients. Blood Cells Mol. Dis. 28, 169–180. Nomura, N., Miyajima, N., Sazuka, T., Tanaka, A., Kawarabayasi, Y., Sato, S., Nagase, T., Seki, N., Ishikawa, K., and Tabata, S. (1994). Prediction of the coding sequences of unidentified human genes. I. The coding sequences of 40 new genes (KIAA0001-KIAA0040) deduced by analysis of randomly sampled cDNA clones from human immature myeloid cell line KG-1. DNA Res 1, 27–35. Nozawa, R.S., Nagao, K., Masuda, H.T., Iwasaki, O., Hirota, T., Nozaki, N., Kimura, H., and Obuse, C. (2010). Human POGZ modulates dissociation of HP1alpha from mitotic chromosome arms through Aurora B activation. Nat. Cell Biol. 12, 719–727. Nuez, B., Michalovich, D., Bygrave, A., Ploemacher, R., and Grosveld, F. (1995). Defective haematopoiesis in fetal liver resulting from inactivation of the EKLF gene. Nature 375, 316–318. Oakley, K., Han, Y., Vishwakarma, B.A., Chu, S., Bhatia, R., Gudmundsson, K.O., Keller, J., Chen, X., Vasko, V., Jenkins, N.A., et al. (2012). Setbp1 promotes the self-renewal of murine myeloid progenitors via activation of Hoxa9 and Hoxa10. Blood 119, 6099–6108. Okazaki, N., Kikuno, R., Ohara, R., Inamoto, S., Aizawa, H., Yuasa, S., Nakajima, D., Nagase, T., Ohara, O., and Koga, H. (2003). Prediction of the coding sequences of mouse homologues of KIAA gene: II. The complete nucleotide sequences of 400 mouse KIAA-homologous cDNAs identified by screening of terminal sequences of cDNA clones randomly sampled from size-fractionated libraries. DNA Res. 10, 35–48. Palis, J. (2014). Primitive and definitive erythropoiesis in mammals. Front. Physiol. 5, 3.

Iossifov, I., Ronemus, M., Levy, D., Wang, Z., Hakker, I., Rosenbaum, J., Yamrom, B., Lee, Y.H., Narzisi, G., Leotta, A., et al. (2012). De novo gene disruptions in children on the autistic spectrum. Neuron 74, 285–299.

Perkins, A.C., Sharpe, A.H., and Orkin, S.H. (1995). Lethal beta-thalassaemia in mice lacking the erythroid CACCC-transcription factor EKLF. Nature 375, 318–322.

Ishikawa, K., Nagase, T., Nakajima, D., Seki, N., Ohira, M., Miyajima, N., Tanaka, A., Kotani, H., Nomura, N., and Ohara, O. (1997). Prediction of the coding sequences of unidentified human genes. VIII. 78 new cDNA clones from brain which code for large proteins in vitro. DNA Res 4, 307–313.

Pilon, A.M., Arcasoy, M.O., Dressman, H.K., Vayda, S.E., Maksimova, Y.D., Sangerman, J.I., Gallagher, P.G., and Bodine, D.M. (2008). Failure of terminal erythroid differentiation in EKLF-deficient mice is associated with cell cycle perturbation and reduced expression of E2F2. Mol. Cell. Biol. 28, 7394–7401.

€ttner, R., Crenshaw, John, A., Brylka, H., Wiegreffe, C., Simon, R., Liu, P., Ju E.B., 3rd, Luyten, F.P., Jenkins, N.A., Copeland, N.G., et al. (2012). Bcl11a is required for neuronal morphogenesis and sensory circuit formation in dorsal spinal cord development. Development 139, 1831–1841. Kingsley, P.D., Malik, J., Emerson, R.L., Bushnell, T.P., McGrath, K.E., Bloedorn, L.A., Bulger, M., and Palis, J. (2006). ‘‘Maturational’’ globin switching in primary primitive erythroid cells. Blood 107, 1665–1672. Koulnis, M., Pop, R., Porpiglia, E., Shearstone, J.R., Hidalgo, D., and Socolovsky, M. (2011). Identification and analysis of mouse erythroid progenitors using the CD71/TER119 flow-cytometric assay. J. Vis. Exp 54, 2809. €hn, R., Schwenk, F., Aguet, M., and Rajewsky, K. (1995). Inducible gene tarKu geting in mice. Science 269, 1427–1429. Lettre, G., Sankaran, V.G., Bezerra, M.A., Arau´jo, A.S., Uda, M., Sanna, S., Cao, A., Schlessinger, D., Costa, F.F., Hirschhorn, J.N., and Orkin, S.H. (2008). DNA polymorphisms at the BCL11A, HBS1L-MYB, and beta-globin loci associate with fetal hemoglobin levels and pain crises in sickle cell disease. Proc. Natl. Acad. Sci. USA 105, 11869–11874.

Sankaran, V.G., Menne, T.F., Xu, J., Akie, T.E., Lettre, G., Van Handel, B., Mikkola, H.K., Hirschhorn, J.N., Cantor, A.B., and Orkin, S.H. (2008). Human fetal hemoglobin expression is regulated by the developmental stage-specific repressor BCL11A. Science 322, 1839–1842. Sankaran, V.G., Xu, J., Ragoczy, T., Ippolito, G.C., Walkley, C.R., Maika, S.D., Fujiwara, Y., Ito, M., Groudine, M., Bender, M.A., et al. (2009). Developmental and species-divergent globin switching are driven by BCL11A. Nature 460, 1093–1097. Sankaran, V.G., Lettre, G., Orkin, S.H., and Hirschhorn, J.N. (2010a). Modifier genes in Mendelian disorders: the example of hemoglobin disorders. Ann. N Y Acad. Sci. 1214, 47–56. Sankaran, V.G., Xu, J., and Orkin, S.H. (2010b). Transcriptional silencing of fetal hemoglobin by BCL11A. Ann. N Y Acad. Sci. 1202, 64–68. Schultz, J., Milpetz, F., Bork, P., and Ponting, C.P. (1998). SMART, a simple modular architecture research tool: identification of signaling domains. Proc. Natl. Acad. Sci. USA 95, 5857–5864.


Stessman, H.A.F., Willemsen, M.H., Fenckova, M., Penn, O., Hoischen, A., Xiong, B., Wang, T., Hoekzema, K., Vives, L., Vogel, I., et al. (2016). Disruption of POGZ is associated with intellectual disability and autism spectrum disorders. Am. J. Hum. Genet. 98, 541–552. Su, A.I., Wiltshire, T., Batalov, S., Lapp, H., Ching, K.A., Block, D., Zhang, J., Soden, R., Hayakawa, M., Kreiman, G., et al. (2004). A gene atlas of the mouse and human protein-encoding transcriptomes. Proc. Natl. Acad. Sci. USA 101, 6062–6067. Su, M.Y., Steiner, L.A., Bogardus, H., Mishra, T., Schulz, V.P., Hardison, R.C., and Gallagher, P.G. (2013). Identification of biologically relevant enhancers in human erythroid cells. J. Biol. Chem. 288, 8433–8444. €ppelTallack, M.R., and Perkins, A.C. (2013). Three fingers on the switch: Kru like factor 1 regulation of g-globin to b-globin gene switching. Curr. Opin. Hematol. 20, 193–200. Tan, B., Zou, Y., Zhang, Y., Zhang, R., Ou, J., Shen, Y., Zhao, J., Luo, X., Guo, J., Zeng, L., et al. (2016). A novel de novo POGZ mutation in a patient with intellectual disability. J. Hum. Genet. 61, 357–359. Uda, M., Galanello, R., Sanna, S., Lettre, G., Sankaran, V.G., Chen, W., Usala, G., Busonero, F., Maschio, A., Albai, G., et al. (2008). Genome-wide association study shows BCL11A associated with persistent fetal hemoglobin and amelioration of the phenotype of beta-thalassemia. Proc. Natl. Acad. Sci. USA 105, 1620–1625.


Weatherall, D.J. (2010). The inherited diseases of hemoglobin are an emerging global health burden. Blood 115, 4331–4336. White, J., Beck, C.R., Harel, T., Posey, J.E., Jhangiani, S.N., Tang, S., Farwell, K.D., Powis, Z., Mendelsohn, N.J., Baker, J.A., et al. (2016). POGZ truncating alleles cause syndromic intellectual disability. Genome Med. 8, 3. Xu, J., Sankaran, V.G., Ni, M., Menne, T.F., Puram, R.V., Kim, W., and Orkin, S.H. (2010). Transcriptional silencing of gamma-globin by BCL11A involves long-range interactions and cooperation with SOX6. Genes Dev. 24, 783–798. Xu, J., Peng, C., Sankaran, V.G., Shao, Z., Esrick, E.B., Chong, B.G., Ippolito, G.C., Fujiwara, Y., Ebert, B.L., Tucker, P.W., and Orkin, S.H. (2011). Correction of sickle cell disease in adult mice by interference with fetal hemoglobin silencing. Science 334, 993–996. Xu, J., Bauer, D.E., Kerenyi, M.A., Vo, T.D., Hou, S., Hsu, Y.J., Yao, H., Trowbridge, J.J., Mandel, G., and Orkin, S.H. (2013). Corepressor-dependent silencing of fetal hemoglobin expression by BCL11A. Proc. Natl. Acad. Sci. USA 110, 6518–6523. Zhou, D., Liu, K., Sun, C.W., Pawlik, K.M., and Townes, T.M. (2010). KLF1 regulates BCL11A expression and gamma- to beta-globin gene switching. Nat. Genet. 42, 742–744.

Cell Reports, Volume 23

Supplemental Information

POGZ Is Required for Silencing Mouse Embryonic b-like Hemoglobin and Human Fetal Hemoglobin Expression Bjorg Gudmundsdottir, Kristbjorn O. Gudmundsson, Kimberly D. Klarmann, Satyendra K. Singh, Lei Sun, Shweta Singh, Yang Du, Vincenzo Coppola, Luke Stockwin, Nhu Nguyen, Lino Tessarollo, Leifur Thorsteinsson, Olafur E. Sigurjonsson, Sveinn Gudmundsson, Thorunn Rafnar, John F. Tisdale, and Jonathan R. Keller

Figure S1. Pogz domain structure, expression profile, and generation of mice with targeted Pogz gene deletion, Related to Figure 2. (A) Summary of human and mouse POGZ proteins using SMART (Simple Modular Architecture Research Tool) analysis (http://smart.emblheidelberg.de/) demonstrating identical domain structure with the presence of multiple C2H2 zinc fingers, a CENPB-type domain and a DDE domain. Regions of low complexity are represented in magenta. (B) Pogz expression profile in mouse primary cells, tissues, and cell lines. Probeset ID mouse: 1455046_a_at. Data were obtained from the BioGPS gene expression database (http://biogps.org). (C) Schematic overview of the Pogz targeting strategy. The exons are depicted as blue boxes, loxp sites as black arrowheads, and Frt sites as green arrowheads. The Pogz gene spans approximately 26.4 kb and contains 19 exons. To delete as many functional protein domains as possible, we flanked a 4.13 kb region on the 3´end of the Pogz gene with loxp sites. This region contains exons 13-19 encoding three zinc fingers, the CENPB domain, the DDE domain and a part of the 3’ untranslated region. The conditional knockout vector for Pogz was constructed using the recombineering method and is described in the supplemental methods (Liu et al., 2003). For generating homozygous Pogz−⁄− embryos, heterozygous Pogz+/- mice were intercrossed and pregnant mice harvested at different time points. (D) Genotyping of E15.5 Pogz embryos using primer pairs specifically designed to generate wild type or knockout PCR products. (E) RT-PCR analysis of the full length Pogz coding sequence using RNA isolated from the brains of E14.5 Pogz+/+, Pogz+/- and Pogz-/- embryos. (F) Western blot analysis of whole cell lysates generated from E15.5 Pogz+/+, Pogz+/- and Pogz-/- fetal brains.

Figure S2. Reduced output of hematopoietic cells in Pogz-/- fetal livers, Related to Figure 2 and Figure 3. Total number of (A) macrophages (Mac1+Gr1-), (B) neutrophils (Mac1+Gr1+) and (C) B cells (CD19+) in E15.5 Pogz+/+ and Pogz-/- fetal livers. N=3 for each genotype. Data are presented as mean ± SD. * P < 0.05. *** P < 0.001 and are representative of three separate experiments. (D) Schematic overview of differentiating red blood cells and corresponding flow cytometry gates for FL cells. Representative flow cytometry analysis of CD71 and Ter119 expression in E15.5 Pogz+/+ and Pogz-/- fetal liver (FL) cells (N=4 for each genotype). Gates were set around subsets of differentiating FL erythroid cells (S0-S5) as previously described (Koulnis et al., 2011). The frequencies for all cell populations are shown (N= 4 mice/group), and are representative of three separate experiments. Pogz-/- FLs show increased percentages of S0, S1 and S2 cells. (E) Pogz-/- FLs show a reduction in the total number of more differentiated erythroid cells (S3-S5). Data are presented as mean ± SD. * P < 0.05; ** P < 0.01, and are representative of three separate experiments. (F) Loss of Pogz affects expression of genes linked to hematological disease. Genes denoted with green color are downregulated in Pogz-/- fetal livers. Genes denoted with red color are upregulated in Pogz-/- fetal livers.

Figure S3. Erythroid development in mice transplanted with Pogz-/- fetal liver cells, Related to Figure 4. Representative flow cytometry analysis of CD71 and Ter119 expression in recipients of E15.5 Pogz+/+ (N=5) and Pogz-/- FL cells (N=4) 4 months after transplantation. Gates were set around subsets of differentiating BM erythroid cells (S0-S5) as previously described (Koulnis et al., 2011). (A) Flow cytometry analysis of total BMCs from recipient mice and (B) flow cytometry analysis of CD45.2+ donor-derived BMCs. Data are presented as mean ± SD. ** P < 0.01, and are representative of two separate experiments.

Figure S4. Persistence of fetal globin expression and analysis of PBCs in transplant recipient mice, Related to Figure 4. (A) Persistence of embryonic Hbb-y globin expression in mice transplanted with Pogz-/- FL cells. RNA was purified from PBCs isolated from mice 3 months after transplantation with Pogz+/+ and Pogz-/- FL cells, and Hbb-y expression analyzed by real-time RT-qPCR. Gene expression was normalized to β-actin expression. Experiments were performed in triplicate and data are presented as mean ± SD, and representative of two separate experiments. (B) PBCs were isolated from mice transplanted with Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre 12 weeks after treatment with pIpC. DNA was extracted from PBCs and individual recipient mice were genotyped using primers that specifically amplify Pogz+/+ and Pogz-/- DNA by PCR followed by gel electrophoresis. (C) CBC analysis of PBCs obtained from mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre or BMC 12 weeks post pIpC injections (N=5 for each genotype). No significant differences in red blood cell counts, hemoglobin concentration, hematocrit (%), platelets, mean platelet volume, and white blood cell counts were observed in mice transplanted with Pogz+/+; Mx1-cre BMC or Pogzf/f; Mx1-cre BMCs. Data are presented as mean ± SD. * P = 0.01, and are representative of three separate experiments.

Figure S5. Analysis of hematopoietic development in mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMC, Related to Figure 4. (A) No difference in neutrophil or B cell development was observed in mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMCs 12 weeks after pIpC treatment. The frequency and total number of donor neutrophils and mature B cells for individual mice is presented. (B) Representative flow cytometry analysis of CD71 and Ter119 expression in differentiating erythroid cells (S0-S5) in mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMCs 12 weeks after pIpC treatment. No difference in the frequency of differentiating erythroid cells in S1, combined S2/3/4 or S5 gates was observed. (C) BMCs were stained with lineage markers, c-Kit, Sca1, CD34 and Fcrγ to analyze the frequency of MEPs by flow cytometry. No difference in the frequency of donor MEPs was observed in mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMCs 12 weeks after pIpC treatment. (D) Frequency of donor Mk and preMegE progenitors in mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMC 12 weeks after pIpC injections. BMCs were stained with lineage markers, c-Kit, Sca1, CD41, CD150, and CD105 and analyzed by flow cytometry. The data presented in A-D are representative of at least two separate experiments with five mice analyzed per group.

Figure S6. Expression and regulation of hemoglobin and transcription factor expression in mice that lack Pogz expression, Related to Figure 4 and Figure 5. (A-D) Hba-x and Hbb-bh1 expression was elevated is some mice transplanted with Pogzf/f; Mx1-cre BMCs. BMCs were harvested from recipient mice transplanted with Pogz+/+; Mx1-cre or Pogzf/f; Mx1-cre BMC 12 weeks post pIpC treatment, and expression of Hba-x and Hbb-bh1 was analyzed by real-time qRT-PCR (N=5 per genotype) in two separate experiments. Gene expression was normalized to β-actin expression. (E-I) Pogz, Bcl11a, Gata1 and Klf1 RNA expression was determined in FACS sorted erythroid cells from the bone marrow of normal mice by qRT-PCR. Bone marrow erythroid cells from C57BL/6J mice were stained with Ter119 and CD71 antibodies and sorted into 4 populations (Ter119-CD71+ {S1-2}, Ter119+CD71HI {S3}, Ter119+CD71Med {S4}, and Ter119+CD71- {S5}) using the gates defined. (F) Pogz and (G) Bcl11a show similar RNA expression profiles throughout erythroid differentiation, whereas (H) Gata1 and (I) Klf1 RNA expression is predominantly in Ter119+CD71HI (S3) cells. (J) Overexpression of Bcl11a in E16.5 Pogz-/- FL cells does not repress Hbb-y expression. E16.5 Pogz+/+ and Pogz-/- FL erythroid progenitors (N=2 each genotype) were transduced with a control retroviral vector or a Bcl11a retroviral vector, and harvested 60 hrs post transduction for qRT-PCR analysis of Hbb-y, Pogz and Bcl11a expression. Gene expression was normalized to β-actin expression.

Figure S7. POGZ regulates fetal hemoglobin expression in normal human erythroid expansion and differentiation culture in vitro, Related to Figure 7. CD34+ cells were purified from mobilized peripheral blood cells, and then cultured in a modified 2 phase in vitro erythroid expansion and differentiation system. (A) Flow cytometry analysis of CD34, CD45, CD71 and CD235a expression on day 2 (blue histograms) and day 11 (red histograms) of erythroid cell cultures in vitro. (B) Photographs of undifferentiated control cell pellet at day 0 (left panel) and cell pellet of differentiated cells at day 11 demonstrating hemoglobinization of cells in culture (right panel). (C) Bar graphs showing expression of POGZ, HBG1/2, HBB, HBE and HBA following lentiviral mediated shRNA knockdown of POGZ using two POGZ shRNA constructs. Erythroblasts were harvested on day 10 post transduction, and expression analyzed by real time RT-qPCR. Gene expression was normalized to ACTIN expression. Experiments were performed in triplicates and data are presented as mean ± SD. * P < 0.05; ** P < 0.01; *** P < 0.001. (D) Bar graphs showing expression of BCL11A, HBG1/2, HBB, HBE and HBA following lentiviral mediated shRNA knockdown of BCL11A. Erythroblasts were harvested on day 10 post transduction, and expression analyzed by real time RT-qPCR. Gene expression was normalized to ACTIN expression. Experiments were performed in triplicates and data are presented as mean ± SD. *** P < 0.001.

Supplemental Experimental Procedures Cell Cultures: KG1 cells were maintained in IMDM media (Life Technologies, Gaithersburg, MD, USA) containing 20% FBS and 1% Penicillin/Streptomycin (Life Technologies), and differentiated into myeloid cells by adding 20 ng/ml PMA (Sigma-Aldrich, St. Louis, MO, USA) and 20 ng/ml TNF-α (R&D Systems, Minneapolis, MN, USA). Murine erythroleukemia (MEL) cells were maintained in RPMI-1640 media (Life Technologies) containing 10% FBS and 1% Penicillin/Streptomycin (Life Technologies). Human CD34+ cells were obtained from peripheral blood stem/progenitor cells mobilized by granulocyte colony-stimulating factor (G-CSF) under studies (02-H-0160 and 08-H-0156) that were approved by the institutional review boards (IRB) of the National Heart, Lung, and Blood Institute (NHLBI) and the National Institute of Diabetes, Digestive, and Kidney diseases (NIDDK). All patients gave written informed consent for the sample donation and consent documents are maintained in the medical records. The consent documents were approved by the IRB prior to study initiation and were reviewed and updated yearly. The CD34+ cells were thawed and cultured in a 2 stage culture system as previously described (Migliaccio et al., 2002). Briefly, the cells were cultured in an expansion media consisting of IMDM with 1% Penicillin/Streptomycin, 20% FBS, 10 ng/ml Stem Cell Factor (SCF), 2.0 U/ml Erythropoietin (EPO), 1 ng/ml IL-3, 1 µM Dexamethasone and 1 µM Estradiol, for 5 days, with fresh media added as needed to keep the cells at 0.1-1x106 cells/ml. On day 6 cells were transferred to differentiation media consisting of IMDM, 1% Penicillin/Streptomycin,

20% FBS, 2.0 U/ml EPO, 10 ng/ml Insulin, 0.5 mg/ml Holo Transferrin and 2% BSA Fraction V. The differentiation media was changed every other day until cells were harvested. Western Blot Analysis: Whole-cell lysates for western blot analysis were prepared from KG1 cells, MEL cells, wild type adult mouse tissues, Pogz+⁄+, Pogz+/- and Pogz−⁄− fetal brains, fetal livers and CD34 HSPC derived eryhtroblasts. The lysates were resolved on 4-12% Tris-Glycine pre-cast gels (Life Technologies), transferred to nitrocellulose membranes and probed with specific antibodies against Pogz (from Bethyl Laboratories, Montgomery, TX, USA), Bcl11a (Abcam, Cambridge, MA), fetal hemoglobin; HBG1/2 (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and Actin (Millipore, Billerica, MA, USA). Goat anti-mouse IgG-HRP or goat antirabbit IgG-HRP (Santa Cruz Biotechnology) were used as secondary antibodies, and ECL detection kit (GE Healthcare Biosciences, Pittsburgh, PA, USA) to develop blots. Immunofluorescence: MEL cells were fixed with 2% paraformaldehyde on poly-L-lysine coated slides followed by permeabilization with 0.05% Triton X-100. The cells were blocked with 5% normal goat serum and 3% BSA for 30 minutes at room temperature. Cells were then labeled with primary antibodies to Pogz (1:50; IHC-00712, Bethyl Laboratories) and fibrillarin (1:200; sc-166021, Santa Cruz Biotechnology) overnight at 4°C. After washing with PBS, cells were incubated with Texas Red goat anti-rabbit (1:500; Invitrogen) and Alexa-Fluor 488 goat anti-mouse (1:500; Invitrogen) at room temperature for 1 hour. Cells were mounted with ProLong Gold anti-fade reagent with DAPI to stain for the nucleus and visualized by a LSM 710 scanning confocal microscope (Zeiss). Mice: Mice were housed, fed and handled in accordance with the National Institutes of Health guidelines for animal care and use, and the Guide for the Care and Use of Laboratory Animals,

8th Edition. All mouse experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the National Cancer Institute at Frederick, which is accredited by AAALAC International. Targeted Disruption of the Pogz Gene: The conditional knockout vector for Pogz was constructed using the recombineering method as previously described (Liu et al., 2003). Briefly, using probes for the 5´and 3´ ends of the gene, three different 129/SV BAC clones (543F12, 60907, 529K9; all from Life Technologies) which contained the gene (or parts of the gene) were identified. BAC 60907 was found to contain the entire region of interest and was used for sequence retrieval. Primers used in the study were designed using MacVector (MacVector, Inc., Cary, NC, USA) and are listed in Table S2. All PCR reactions were performed using the ROCHE Expand High-Fidelity Kit (Roche Applied Science, Indianapolis, IN, USA). Exons 1319, which encode the CENPB and DDE domains and a portion of the zinc finger region of the Pogz protein, were targeted. These were flanked by a loxP site at the 5′ end and a FRT-NeoFRT-loxP cassette on the 3′ end. The Pogz targeting vector was linearized and introduced into CJ7 ES cells by a standard electroporation method. ES cell clones resistant for G418 and FIAU were selected and analyzed using Southern blot hybridization of EcoRI-digested genomic DNA. Two independently targeted ES cell clones were injected into C57BL/6 blastocysts to produce chimeras demonstrating germ-line transmission of the targeted gene. To delete the neo cassette from the line, Pogzneo/neo mice were crossed with β-actin-flp mice to produce Pogzf/+ mice. These were subsequently crossed to β-actin-cre mice to produce heterozygous Pogz+⁄− mice. To detect the full length Pogz coding sequence, cDNA from mouse fetal brain was amplified using the Advantage 2 PCR system (Clontech Laboratories, Inc. Mountain View, CA, USA) according to the manufacturer’s recommendations. Primer sequences are listed in Table S2. To

conditionally delete the Pogzf allele, an Mx1-cre+ transgene was introduced into Pogzf/f mice to generate Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre mice. Microarray analysis: Microarray analysis was performed at the NCI Frederick Laboratory of Molecular Technology. Briefly, RNA was isolated from E14.5 Pogz+/+ and Pogz-/- fetal livers (3 of each genotype) using the Ambion RNAqueous-4PCR Kit (Life Technologies). RNA quality was analyzed using Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA). Approximately 100 ng RNA was reverse transcribed from each sample, labeled with biotin and hybridized to Affymetrix GeneChip Mouse Genome 430 2.0 microarrays (Affymetrix, Santa Clara, CA, USA). Data was analyzed using the GeneSpring software (Agilent Technologies Inc.). The cutoff value was set at ±1.5 fold and the false discovery rate at P=0.05. Of the 1232 transcripts with an Id, 1185 could be mapped and 1062 were analysis ready. Pathway analysis was performed using Ingenuity Pathway Analysis (IPA) (Ingenuity Systems, Inc., Redwood City, CA, USA). Microarray data was deposited to the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/). Accession number GSE113503. Flow Cytometry: For flow cytometric analysis of mouse cells, single-cell suspensions were prepared from E14.5-E16.5 Pogz+/+ or Pogz-/- FL cells, or BMC and PBC from animals transplanted with Pogz+/+ or Pogz-/- FL cells, or Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre BMCs. Where applicable, the cells were incubated with ACK lysis buffer (Life Technologies) to lyse red cells followed by a wash in PBS/2% BSA buffer. FL cells were resuspended in PBS/2% BSA and treated with Fc receptor blocking antibodies (anti-mouse CD16/32), and then incubated with the following conjugated monoclonal antibodies for lineage analysis: Isotype controls: rIgG2a (eBR2a), rIgG2b (eB149/10H5), CD71 (R17217), Ter119 (TER-119), anti-Gr-1 (RB68C5), anti-Mac-1 (M1/70), CD19 (1D3). For HSPC analysis, BMCs were incubated with

biotinylated antibodies against Lin markers (Mac-1, Gr-1, B220, Ter119, CD4, CD8), PEconjugated c-Kit (2B8), APC-conjugated Sca-1 (E13-161.7), PE-Cy7-conjugated Flk2 (A2F10 ) FITC–conjugated CD34 (RAM34), PE-Cy5-conjugated IL-7r (A7R34), and streptavidineflour450. HSCs are Lin-negative (L), c-Kit+ (K) and Sca-1+ (S) (LSK) CD34-Flk2- cells, and common lymphoid progenitors (CLP) are LSK Flk2+IL7r+ cells. LK progenitors were isolated by replacing Flk2 and IL7r antibodies with PE-Cy7-conjugated FcγRII/III ( 2.4G2) including common myeloid progenitors (CMP) LK CD34+FcR-, granulocyte/macrophage progenitors (GMP) LK CD34+, FcR+, and megakaryocyte-erythrocyte progenitors (MEP) are LK CD34FcR-. Megakaryocyte progenitors, (MP) progenitors are LK CD150+CD41+ cells, and preMegE are LK CD41-CD150+CD105- and were identified using PE-conjugated CD41 (MWReg30), APC-conjugated CD150 (mShad150), biotin-conjugated CD-105 (MJ7/18), and streptavidineflour450. Throughout the staining process, the cells were kept at 4°C, and the antibodies were used at a concentration of 0.2-0.5 μg/1 × 106 cells. For flow cytometric analysis of human CD34+ cells differentiated towards the erythroid lineage, cells were resuspended in PBS/2% BSA and treated with Fc receptor blocking antibodies (anti-human CD16/32), and then incubated with the following conjugated monoclonal antibodies for lineage analysis: FITC and APCconjugated-mIgG1 (MOPC-21), APC-conjugated mIgG2a (G155-178), PE-conjugated mIgG2b (27-35), FITC-conjugated CD34 (581), APC-conjugated CD45 (HI30), APC-conjugated CD71 (M-A712), PE-conjugated CD235a (HIR2). All antibodies were purchased from BD Biosciences (San Jose, CA, USA) or eBiosciences (San Diego, CA, USA). The cells were analyzed by

FACS-CantoII or FACSCalibur (BD Biosciences) and data analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA). Fetal Liver and Bone Marrow Cell Transplantation: Fetal livers were harvested from E14.5E16.5 Pogz+/+ and Pogz-/- embryos and BMCs were isolated from adult Pogz+/+; Mx1-cre and Pogzf/f; Mx1-cre mice and transplanted as previously described (Gudmundsson et al., 2014; Gudmundsson et al., 2012). Briefly, FL cells were dissociated in PBS 4° C by mechanical disruption, and femurs were flushed using 5 ml syringe with a 29G × ½ needle and cells were passed through through 0.45 μm mesh filters (Millipore) to obtain single cell suspensions. The cells were resuspended in PBS and FL cells were transplanted at 0.25 × 106 cells/ 0.2 ml, and BMC at 1 x 106/ 0.2ml. The cell suspension was injected into the tail vein of lethally irradiated (10 Gy) recipient mice (B6.SJL-Ptprca Pep3b/BoyJ, CD45.1, Charles River Laboratories) to track donor reconstitution. Six weeks after transplantation of Pogzf/f; Mx1-Cre BMCs and controls, the mice were injected with 300µg of polyinosinic–polycytidylic acid solution (pIpC) dissolved in physiological water (NaCl 0.9%) (tlrl-pic, Invivogen or P1530, Sigma-Aldrich) 2 or 3 times every other day. The day of last injection was defined as day 0. Lentiviral mediated shRNA knockdown: Mouse and human Pogz and Bcl11a shRNA target sequences were derived from The RNAi Consortium (TRC) portal (http://www.broadinstitute.org/rnai/public/) for oligo generation (sequences listed in Table S2). Oligos containing the target sequences were purchased, annealed and cloned into the pLKO.1 puro vector according to TRC protocols (http://www.broadinstitute.org/rnai/public/resources/protocols). Infectious lentivirus was generated by transfecting the shRNA constructs and packaging plasmids (PMD2G and pCMV8.74) into 293T/17 cells using the LipoD293 transfection agent (SignaGen, Rockville,

MD, USA). Virus containing supernatants were collected 48 hours post transfection. To calculate viral titers, NIH3T3 cells were transduced with serial dilutions of viral supernatants. Media containing puromycin (2 μg/ml) or blasticidin (10 μg/ml) was added 48 hours post transduction and resistant colonies counted 3 days later. For Pogz or Bcl11a knockdown, MEL cells or human CD34+ HSPC derived erythroblasts were transduced with shRNA lentivirus by spinoculation. Briefly, lentivirus and cells (4:1 ratio) were spun at 2000 x g for 90 minutes at 37°C. At 24-48 hours post transduction, puromycin (2 μg/ml) or blasticidin (10 μg/ml) was added to the cultures. MEL cells were harvested at 72-96 hours post transduction and CD34+ HSPC derived erythroblasts 9-12 days post transduction for real-time qRT-PCR and western blotting. Retroviral transduction. MSCV-mPogz-IRES-eGFP retroviral construct was produced at the NCI-Frederick Protein Expression Laboratory and used to generate infectious viral particles in Ecotropic Plat-E packaging cells. Briefly, Plat-E cells were transiently transfected with a control retrovirus or MSCV-mPogz-IRES-eGFP retrovirus using Fugene 6 (Roche) and viral supernatant harvested 48 hours post-transfection. To calculate viral titers, NIH3T3 cells were transduced with serial dilutions of viral supernatants and GFP positive colonies counted 48 hours posttransduction. E16.5 Pogz-/- fetal liver cells were transduced in 1:4 ratio with control or MSCVmPogz-IRES-eGFP retrovirus on plates coated with Retronectin (Takara Bio Inc., Mountain View, CA, USA) in DMEM media containing 15% FBS, 1% P/S, SCF (100 ng/ml), IL-3 (6 ng/ml), IL-6 (10 ng/ml) and Polybrene (1/1000). Two consecutive transductions were carried out with 24 hour interval and cells harvested for RNA isolation 48-60 hours after the second transduction. MEL cells were transduced with MSCV-mBcl11a-IRES-eGFP retrovirus following lentiviral mediated Pogz knockdown. A second transduction was performed 24 hours later and

puromycin added to the cultures (2 µg/ml). Cells were harvested at 96 hours post transduction for real-time qRT-PCR and western blotting. Statistical Analysis: Statistical analysis was performed using Graphpad Prism (GraphPad Software, Inc.). Unpaired student’s t-test was used to calculate statistical significance. Results were considered significant if P < 0.05. Results are presented as mean ± standard deviation. Chromatin immunoprecipitation (ChIP) assay: ChIP assay was performed on chromatin isolated from MEL cells using the ChIP-IT Express Enzymatic kit according to the manufacturers recommendations (Active Motif, Carlsbad, CA, USA). Briefly, 1.5x107 MEL cells were crosslinked in media containing 1% formaldehyde and nuclei extracted by incubating the cells in lysis buffer for 30 min on ice followed by dounce homogenization and incubation in enzymatic digestion for 10 min to shear the chromatin. Sheared chromatin was immunoprecipitated overnight at 4°C in a solution containing protein-G magnetic beads and antiPogz or rabbit IgG (both from Bethyl). Immunoprecipitated chromatin was washed, eluted from the beads and the crosslinking reversed. Purified DNA was used for qPCR with the following primers for the Bcl11a enhancer region: +58 ChIP-F: 5’- AAA GGT GTT GGG TTC TGA GG3’; +58 ChIP-R: 5’- ATC AGC AGC GAG CTC TCA TA-3’. The following primers were used for the Bcl11a promoter region: mBcl11a (-3496)-F: 5’-AAG CCA TTT CTG GAG AGG TAA A-3’; mBcl11a (-3496)-R: 5’-GTA TTG TGG AGC TGG GTG AA-3’; mBcl11a (-972)-F: 5’CTT CTC GGT CTA TGT ATT CCA ATC T-3’; mBcl11a (-972)-R: 5’-GAT CTG AGC GAC

CCT ACA AAC-3’. A validated mouse negative control primer set from Active Motif (negative primer set 2) was used as a negative control. Reverse Phase - High Performance Liquid Chromatography (HPLC): Hemoglobin quantity (HBB, HBE, HBG1/2) in erythroblast cultures was determined by HPLC analysis using the Agilent 1100 HPLC series (Agilent Technologies, Santa Clara, CA) according to the manufacturers recommendations.