SUPPLEMENTARY INFORMATION

Transcription

1 DOI:.3/ncb7 Hematopoiesis Expression (Protein) Competition PRC integration Polycomb-mediated gene repression Cell fate HSC PRC SELF-RENEWAL Cbx Cbx4 PRC progenitor genes PROG Cbx PRC Cbx4 DIFFERENTIATION MATURE CELL Cbx Cbx4 HSC genes Figure S PRC-Cbx control of HSC fate decisions. Model summarizing the findings in this paper. Polycomb Cbx orthologs compete for PRC integration and balance HSC self-renewal and differentiation. -containing PRC complexes induce self-renewal of HSCs by repressing the expression of progenitor-specific genes. Other Cbx proteins can outcompete from PRC, resulting in Cbx-, Cbx4-, or -containing PRC complexes that target HSC-specific genes and thereby induce entrance into the differentiation pathway. 3 Macmillan Publishers Limited. All rights reserved.

2 + + + Gr B CD3ε α-cbx α- α- α-actin Figure S Protein abundance of Cbx orthologs in differentiated hematopoietic. Western blot analysis of protein abundance of Cbx, and in granulocytes (Gr+), B (B+) and T- (CD3e+). Actin was used as loading control. Also see Figure S Macmillan Publishers Limited. All rights reserved.

3 A B C G/G: 7.5±4.4 S: 3.7±.5 G/M: 3.± AAD 3 counts G/G: 3.5±.5 S:.5±. G/M:.± G/G:.±.7 S:.±.5 G/M: 7.± ANNEXIN-V DAPI Figure S3_de Haan Figure S3 Effect of and overexpression on the cell-cycle. Post 5FU treated bone marrow were transduced with or Cbx retroviral vectors. GFP+ HSPCs were plated in cytokinesupplemented media and at day were used for apoptosis assays, cell-cycle analysis and cytospins. A) Apoptotic frequency of control, and bone marrow using FACS analysis of annexin-v and 7AAD positive. One representative FACS plots is shown from 3 independent experiments. B) Cell cycle analysis by measurement of DNA content in control, and HSPCs (n= independent experiments, mean±s.d.). C) May-Grünwald Giemsa staining of control, and. One representative image is shown. Scale bar μm. Also see Table S7 for raw data Macmillan Publishers Limited. All rights reserved.

4 A B Cbx #F9A.fcs B+ 3 3 CD4 3 CD GFP GFP C 5K CD3 5K CD3 K K # GFP+ x x^/ml PB other Gr/Mac+ CD3e+ B+ SSC SSC 5K K 5K K K 5K K CD/3 SSC SSC 5K K 5K CD3 5K K 5K K D Cbx Cbx4 5K 5K CD7 CD7 peripheral blood bone marrow spleen erythroid T-lymhoid immature μm Figure S4_de Haan Figure S4 -induced leukemia subtypes. A) Analysis of lineage contribution of transplanted HSPCs transduced with, Cbx, Cbx4 and retrovirus, weeks post-transplantation. n=7-9 individual mice, mean±s.e.m. B) FACS plot example of spleen from a mouse (mouse9 suppl table ) with (CD3+, CD4+, CD7+) T-cell leukemia. Cells within CD45.+, GFP+ fraction are shown. C) Staining of spleen of mouse 4 (suppl table ) with an immature leukemia type with CD3, CD7, CD7 and CD/3. Cells within CD45.+, GFP+ fraction are shown. D) May-Grunwald-Giemsa staining of in peripheral blood, bone marrow, and spleen of control mice and mice with indicated leukemia subtypes. One representative image is shown for every condition. Scale bar μm Macmillan Publishers Limited. All rights reserved.

5 5. c-myc relative expression (ddct) 5 5 relative expression (ddct).5..5 T-cell leukemic mice. T-cell leukemic mice. Lmo Evi relative expression (ddct).5..5 relative expression (ddct) 4. T-cell leukemic mice T-cell leukemic mice Figure S5_de Haan Figure S5 Oncogene expression in -induced lymphoid leukemias. Relative gene expression of, c-myc, Evi-, Lmo in spleen from individual (n=3) and T-cell leukemic mice (n=5). Gapdh was used for normalization Macmillan Publishers Limited. All rights reserved.

6 A 5 B population doublings 5 WT Pc AA α-ringb α-bmi flag-ip WT AA Pc WT AA Pc Week C.. α FLAG Ink4a Gapdh 4 α H3K7me3 Ink4a Gapdh 4 Ink4a Gapdh %...4 % 4 % 4. WT AA Pc WT AA Pc WT AA Pc Figure S -induced HSC self-renewal is Polycomb dependent. A) Population doublings of HSPCs WT, flag- AA, and flag- ΔPc. Bars represent mean from independent experiments. B) Flagimmunoprecipitation of Ringb and Bmi with flag- WT, flag- AA, and flag- ΔPc in 3D. One representative experiment is shown from two independent experiments. Also see Figure S9. C) Binding of WT, flag- AA, and flag- ΔPc to the transcription start site of Ink4a. HSPCs were transduced with indicated vectors. Chromatin-immunoprecipitation was then performed using Flag-M agarose beads and H3K7me3 antibodies coupled to sepharose beads. was used as negative control. Pulled down DNA was purified and binding to Ink4a was assessed using qpcr. Bars represent mean from independent experiments. Also see Table S7 for raw data. 3 Macmillan Publishers Limited. All rights reserved.

7 A.75 αcbx flag- flag- B 3.5 αcbx7 flag- flag- %.5 % C E relative expression (ddct) % GAPDH α H3K7me3 flag- flag GAPDH D % GAPDH flag- flag GAPDH p5 p p9 Figure S7 and bind the senescence locus. A-D) Post-5FU treated bone marrow transduced with or Cbx- retroviral vectors were sorted based on GFP expression and cultured for 7 days in cytokine-supplemented media. Next, ChIP was performed in control, and bone marrow using indicated antibodies. Numbers correspond to the Ink4b-Ink4a-Arf primer pairs used for qpcr, as shown in Figure 4a. Bars represent mean from independent experiments. E) Relative expression of p5, p and p9 after and overexpression in bone marrow compared to control, as determined by rt-pcr. Expression was calculated according to ddct of Gapdh. Bars represent mean from independent experiments. Also see Table S7 for raw data Macmillan Publishers Limited. All rights reserved.

8 5 5 _ 5 _ Setdb Sri 3M4Rik Ramp Ankrd Exocb Srpk ENSMUSG Tmem3b D5Ertd579e Hagh Fmnl Tob 75N9Rik Snx Slca Cyb5r Atg5 Ndrg Per Rapgds SUPPLEMENTARY INFORMATION A 5 kb kb chr Npm chr Il7ra B Npm Sfpq Sp3 Ptprc Actin Flag Flag Flag Flag Npm Sfpq Sp3 Ptprc Actin overexpression flag- flag- overexpression flag- flag- overexpression flag- flag- overexpression flag- flag- overexpression flag- flag- C D Gfib Pnpla7 Srgap C35ORik Soat Slc4a Mcfl Sqrdl Ctdsp Ethe Ncf Chst Cd4 Pilra Mgam Alox5ap Fosl Ppp3ca module Slc5a4 Chd7 Anxa3 Ube3c Myl Pxn Aacs Hdac4 Lrrfip module 3 Pppr9a 7PRik Prkg Fnbpl Extl Sfpq Slc3a Pag Blm Lrrcd Mapkapk Notch Ef Arfip Tmd Igfr Mrpl33 Lass Cdk Acpl Lrrfip Gpi Fahd norm. expression (log) norm. expression (log) norm. expression (log) norm. expression (log) LSK LSK Ter9 GR Sfpq Ube3c Flag LSK LSK Ter9 GR ChIP Figure S_de Haan Figure S and targets and gene network. A) Representative examples of ChIP-seq results in, -overexpressed and - overexpressed bone marrow. The y-axis of the binding profiles indicates tag counts. B) Validation of ChIP-seq targets by flag-chip, -ChIP and -ChIP followed by qpcr. Flag-ChIP and Cbx-ChIPs represent independent experiments. Npm and Sfpq are targets (either not bound by, or stronger bound by than by ), Sp3 and Ptprc are specific targets. Actin was used as a negative control locus. Also see Supplementary Table 7 for raw data. C) Module represents a myeloid specific sub-network within the network. D) Module 3 represents an erythroid specific sub-network. Right panels of C) and D): Genes (represented by nodes) are clustered on basis of similarities in gene expression profiles and plotted in two graphs that show an inverse expression pattern along hematopoietic differentiation (L-S+K+ HSCs, L-S-K+ progenitors, erythroid (Ter9+), granulocytes (Gr+)). Green lines represent expression changes of targets, grey dotted lines of targets. Raw expression data can be found in Table S Macmillan Publishers Limited. All rights reserved.

9 A B mflag-ip f- f- f- f- whole film with marker,, different exposure Cbx mouse anti-bmi mouse anti-bmi mflag-ip mflag-ip f- f- f- f- f- f- f- f- rabbit anti- rabbit anti- lower exposure mflag-ip mflag-ip f- f- f- f- f- f- f- f- Cbx Actin H3 rabbit anti-ringb goat anti-mel C anti-actin anti-cbx D m-flag-ip fc7-wt fc7-aa fc7-dpc fc7-wt fc7-aa fc7-dpc m m m r-anti-ringb anti-cbx anti-cbx7 m-flag-ip m m m-anti-bmi fc7-wt fc7-aa fc7-dpc fc7-wt fc7-aa fc7-dpc igg Figure S9 Uncropped western blots. Full scans of cropped westernblots of (A) Figure b, (B) Figure d, (C) Figure S, and (D) Figure Sb Macmillan Publishers Limited. All rights reserved.

10 Supplementary Tables Supplementary Table : -induced leukemic mice. This table show cell counts and FACS analysis of from different tissues from all individual -induced leukemic mice and control mice. FTP= femur, tibia and pelvic bones, WBC= white blood, RBC= red blood Supplementary Table : Differential and targets This table shows the final lists of differential and target genes. Supplementary Table 3: Gene-ontology analysis of differential and targets Classification of differential and targets in GO-categories. Supplementary Table 4: and targets show reciprocal expression patterns during HSC differentiation Summary of all genes presented in modules,, and 3 of the gene network. Connectivity represents the amount of connections within the network (edges). In addition, gene expression data in two genetically distinct mice,c57bl/ (B) and DBA (D), of these genes in different hematopoietic cell populations are shown in worksheet,3 and 4. Differential clustering between and targets was tested using binomial distribution. Stem= Stem (Lin -, Sca +, ckit + ), Pro= progenitors (Lin -, Sca +, ckit - ), Ter=erythroid (Ter9 + ), Gr=Granulocytes (Gr + ). E=equal, D=down, U=up. Supplementary Table 5: targets in HSCs overlap with targets in ESCs Overlapping and unique targets in ESCs (Morey et al., ) and HSCs. Supplementary Table : Gene-ontology analysis of common targets in ESCs and HSCs. Full list of significant overrepresented GO-categories of common ESC and HSC targets (sheet ). In addition, genes that were found that fall into these categories are listed in sheet. Supplementary Table 7: raw data Raw data for experiments with small n (n<5). Raw data for each figure panel are listed in separate worksheets. Supplementary Data : ChIP-seq analysis of and binding sites. ChIP-seq identified targets for and by peakcalling using empty vector as background (worksheet and ), and by cross background peakcalling (worksheet 3 and 4). In the final worksheet, overlapping and peak positions are listed. 3 Macmillan Publishers Limited. All rights reserved.