This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Figure S1. Expression of Key Genes During In Vitro PGC Specification, and Generation and Analysis of EGFP-Blimp1 Knock-in Mice, Related to Figure 1 and Figure 7(A) The expression patterns of genes for germ cell specification/development, pluripotency, an early mesodermal program, and epiblast development during in vitro PGC specification are shown with standard deviations (SDs) of two biological replicates. (B) Schematic representation of the wild-type Blimp1 locus, the targeting vector, and the EGFP-Blimp1 knock-in allele. White boxes: non-coding exons; black boxes: coding exons; black triangles: loxP sequences; LA: long arm; SA: short arm; folded line: MC1-DTA-pA cassette for negative selection. Key restriction enzyme sites and primers for genotyping are shown. (C) Histological sections of the testes and ovaries of the wild-type and EGFP-Blimp1 homozygous knock-in mice stained by hematoxylin and eosin. EGFP-Blimp1 homozygous knock-in mice are healthy and fertile, with their testes and ovaries showing normal spermatogenesis and oogenesis, respectively, indicating that the EGFP-BLIMP1 fusion protein functions appropriately in vivo. Bar, 50 µm. (D) Immunofluorescence analysis of MVH (red) and EGFP-BLIMP1 (green) expression counterstained by DAPI in the embryonic gonads (top, male; bottom, female) at E12.5 of EGFP-Blimp1 heterozygous knock-in mice. Note that EGFP-BLIMP1 localizes specifically in the nuclei of MVH-positive PGCs. Bar, 20 µm. (E) Western blot analysis of EGFP-BLIMP1 and BLIMP1 expression in PGCs at E12.5 in EGFP-Blimp1 homozygous knock-in and wild-type mice, respectively. αTUBLIN was used as a control. (F) Contour graphs for scatter plot comparisons of EGFP-BLIMP1 peaks in the two biological replicates (Rep.1 and Rep. 2) of d2 and d6 PGCLCs (top left and right, respectively), and of averaged EGFP-BLIMP1 peaks in d2 and d6 PGCLCs (bottom).
Figure S2. Distribution of Histone H3, and Normalization and Comparison of ChIP-seq Data for H3K3me2 and H3K27me3, Related to Figure 3 and Figure 5(A) ChIP-Q-PCR analysis of histone H3 on the LINE L1 ORF2, SINEB1, and IAP for ESCs, EpiLCs, d2 and d6 PGCLCs (color codes as indicated). (B) The log2 IP/input-frequency plots of histone H3 for the genome (single-copy regions, 2 Kb sliding windows with 1 Kb overlaps, red) and around TSSs (within 1Kb) of the HCP, ICP, and LCP genes (color codes as indicated) during in vitro PGC specification. (C) ChIP-seq values of H3K9me2 on the repetitive elements normalized by the ChIP-Q-PCR values of H3K9me2 on the LINE L1 ORF2 for ESCs, EpiLCs, and d2 and d6 PGCLCs (color codes as indicated). (D) ChIP-seq values of H3K9me2 on SINE B1 and IAP normalized by the ChIP-Q-PCR values of H3K9me2 on the LINE L1 ORF2 plotted against ChIP-Q-PCR values for SINE B1 and IAP for ESCs, EpiLCs, and d2 and d6 PGCLCs (SDs, two biological replicates). (E) ChIP-seq values of H3K27me3 on the repetitive elements normalized by the ChIP-Q-PCR values of H3K27me3 on the LINE L1 ORF2 for ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs (color codes as indicated). (F) ChIP-seq values of H3K27me3 on SINE B1 and IAP normalized by the ChIP-Q-PCR values of H3K27me3 on the LINE L1 ORF2 plotted against ChIP-Q-PCR values for SINE B1 and IAP for ESCs, EpiLCs, and d2 and d6 PGCLCs (SDs, two biological replicates). (G) Contour graphs for scatter plot comparisons of H3K9me2 around the TSSs (within 1 Kb) in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, or in d2 or d6 PGCLCs (right). (H) Venn diagrams of genes with H3K9me2 detected within 1 Kb from TSSs in the two biological replicates of ESCs, EpiLCs, and d2 and d6 PGCLCs. (legend continued on next page)
15546 964297
EpiLC
16926 4134
d2 PGCLC
16581 116275
d6 PGCLCESC
135121175 128
K15840 105483
EpiLC
16611 24345
d2PGCLC
16622 212151
d6PGCLCd2 LIF Ag
16954 4216
ESC
16036427 454
A B0
1
2
3
4
5
ESCEpiLC
d2 PGCLCd6 PGCLC
LINE1ORF2
SINE B1 IAP
∆Ct (
ChI
P-In
put)
H3 IP/Input (log2)-3 -2 -1 0 1 2
Freq
uenc
y
ESC
EpiLC
d2 PGCLC
d6 PGCLC
Genome
TSS
HCPICPLCP
D
Kurimoto et al., Supplemental Figure S3
OCT4, ESC OnlyTop50%
OC
T4
ES
C
Epi
LC
d2 d6
H3K27acOCT4, ESC&EpiLCTop50%
OC
T4
ES
C
Epi
LC
d2 d6
H3K27acOTX2, EpiLC OnlyTop50%
OTX
2
ES
C
Epi
LC
d2 d6
H3K27ac
±10 Kb0 0 55
Figure S3. Normalization and Comparison of ChIP-seq Data for H3K4me3 and H3K27ac, and Relationships between H3K27ac Peaks and OCT4-binding Sites During In Vitro PGC Specification, Related to Figure 2 (A) Contour graphs for scatter plot comparisons of H3K4me3 in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, d2 LIF Ag, d2 or d6 PGCLCs (right). (B) Venn diagrams of genes with H3K4me3 peaks around TSSs (< 2Kb) in the two biological replicates of ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs. (C) Contour graphs for scatter plot comparisons of H3K27ac in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, d2 LIF Ag, d2 or d6 PGCLCs (right). (D) Heat map representation of the relationships of H3K27ac peaks during in vitro PGC specification with (left) OCT4-binding sites specific to ESCs (top 50%), (middle) OCT4-binding sites common to ESCs and EpiLCs (top 50%), and (right) OTX2-binding sites specific to EpiLCs (top 50%) (Buecker et al., 2014).
(Legend for Figure S2, continued)(I) Contour graphs for scatter plot comparisons of genome-wide H3K9me2 (2 Kb sliding windows with 1 Kb overlaps) in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, d2 or d6 PGCLCs (right). (J) Contour graphs for scatter plot comparisons of H3K27me3 around the TSSs (within 1 Kb) in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, d2 LIF Ag, d2 or d6 PGCLCs (right). (K) Venn diagrams of genes with H3K27me3 detected within 1 Kb from TSSs in the two biological replicates of ESCs, EpiLCs, d2 LIF Ag, and d2 and d6 PGCLCs. (L) Contour graphs for scatter plot comparisons of genome-wide H3K27me3 (2 Kb sliding windows with 1 Kb overlaps) in the two biological replicates (Rep.1 and Rep. 2) of ESCs, EpiLCs, and d2 and d6 PGCLCs (left), and between averaged values in ESCs and those in EpiLCs, d2 LIF Ag, d2 or d6 PGCLCs (right).
Figure S4. Analysis of the H3K27me3 Targets During In Vitro PGC Specification, Related to Figure 3(A) Relationships between the log2 gene expression levels and the log2 H3K27me3 (top), H3K9me2 (middle), and H3K4me3 (bottom) levels during in vitro PGC specification. (B) Fraction of the genomic region with fold H3K27me3 IP level changes > 2 plotted against the genomic loci in EpiLC to d2 PGCLC (top) and EpiLC to d6 PGCLC (bottom) comparisons. (C) GO analysis for the ESC, EpiLC, d2 and d6 PGCLC PRC2 targets. (D) Contour graphs for scatter plot comparisons of the log2 H3K27me3 levels between d2 PGCLCs and EpiLCs (top), d2 LIF Ag (middle), or d6 PGCLCs (bottom). (legend continued on next page)
Figure S5. Analysis of Bivalent Genes During In Vitro PGC Specification, Related to Figure 4(A) Relationships between the log2 H3K4me3 levels and the log2 H3K27me3 IP levels during in vitro PGC specification. (B) Transitions of the numbers of HCP, ICP, and LCP genes (all genes including those with log2 expression levels > 8) that bear H3K4me3 (blue), H3K27me3 (green), bivalent modifications (orange), and no modifications (pale blue) during in vitro PGC specification. (C) Sequential ChIP-Q-PCR analysis of bivalent genes in EpiLCs. The IP efficiency of the 1st ChIP using the anti-H3K4me3 antibody and of the second ChIPs using the anti-H3K27me3 antibody or normal IgG were indicated as ∆Ct values from the input (upper three panels). The fold enrichment in the second ChIP was indicated as the ∆Ct value between the two 2nd IPs (anti-H3K27me3 antibody - normal IgG) (bottom panel). ChIP-Q-PCR was performed for the H3K27me3-only region (Hoxa1 3’ UTR, blue), H3K4me3-only promoter (H3K4me3 peak around the TSS of Gapdh, green), and 12 bivalent promoters (H3K4me3 peaks around the TSSs of Hoxa1, Hoxa5, Pcsk9, Arg2, Mthfd2l, Cdk5r1, Gnaz, Nrp1, Adora2a, Rtn4rl1, Reln, Igfbp5, red). Note that the bivalent promoters were enriched in both the 1st and 2nd IPs (upper two panels). Bars indicate SDs of the two biological replicates. The primers used are given in Supplemental Table S1. (D) Venn diagram representation of the transitions of bivalent genes (log2 expression levels < 8, numbers indicated) and their GO enrichment during in vitro PGC specification. (E) Venn diagram representation of the bivalent genes (log2 expression levels < 8, numbers indicated) and their GO enrichment in comparison between d6 PGCLCs and PGCs at E11.5 (Sachs et al., 2013). (F) Transitions of the histone modifications of the germline genes (log2 expression levels < 8).
(Legend for Figure S4, continued)(E) GO analysis of genes bearing higher levels of H3K27me3 (> 2 fold) in d2 LIF Ag compared to d2 PGCLCs. (F) Expression of PRC2 components during in vitro PGC specification (SDs, two biological replicates). (G) Expression of H3K4 methyltransferases during in vitro PGC specification (SDs, two biological replicates).
Figure S6. Comparison of the Chromatin States between EpiLCs and EpiSCs, Related to Figure 2, Figure 3, Figure 4, and Figure 6(A) Selected GO terms enriched in non-house keeping genes (log2 expression levels < 8 either in EpiLCs or EpiSCs, or >2 fold difference between EpiLCs and EpiSCs) associated with strong H3K27ac peaks differential between EpiLCs and EpiSCs (IP/input > 16, and > 2 fold difference) (within gene bodies or < 15 Kb from TSSs). (B) ChIP-seq tracks of H3K27ac, H3K4me3, and H3K27me3 for selected genes in EpiLCs and EpiSCs. (C) Number of strong H3K27ac peaks differential between EpiLCs and EpiSCs (color codes as indicated) around the non-house keeping genes classified in the indicated GO terms (neuron differentiation, embryonic development, pattern specification) and the EpiLC genes. The classification of GO terms was defined using AmiGO 2 (Ashburner et al., 2000; Carbon et al., 2009). (D) Venn diagram representation of the bivalent genes (log2 expression levels < 8, numbers indicated) and their GO enrichment in comparison between EpiLCs and EpiSCs. (E) The log2 H3K27me3 IP level-frequency plots for the germline genes in EpiLCs (magenta) and EpiSCs (blue). Pale purple and cyan lines represent the plots for all genes as references.
positive regulation of transcription,DNA-dependent
mesoderm development
neuron differentiation
Fgf8, Msgn1, Tbx3, Hoxa1
Tbx6, Mesp1, Efna1, Vegfc
Hoxa1, Hoxa2, Hoxa4, Hoxa5
Gata6, Irf1, Atxn7, Ablm1, Foxo3
Mesp1, Tbx6, Tbx3, Ext1, Nanog
Ascl1, Numb, Neurog1, Epha1
1E-2 1E-9P value
ESC
EpiLCd2 PGCLC
d6 PGCLC
d2 PGCLC
d6 PGCLC
H3K
27m
e3E
GFP
-BLI
MP
1
Fgf4Fgf3Hoxb4Hoxb5 Hoxb2
100Kb 100Kb
BLIMP1 bound genes T bound genes
Figure S7. Analysis of the BLIMP1 and T Target Genes, Related to Figure 7(A) ChIP-seq track transitions for H3K27me3- (top, blue) and BLIMP1-binding (bottom,red) in the 100 Kb regions around Fgf3 and Fgf4 (left) and Hoxb cluster (right). (B) Distribution of BLIMP1 (in d2 and d6 PGCLCs) and T [in EpiLC aggregates stimulated by BMP4 for 36 hrs (Aramaki et al., 2013)] peaks (color codes as indicated) in the genome represented by peak numbers per base pair plotted against the distances from the TSSs (red dotted line). (C) The promoter classes for genes bound by BLIMP1 and T. (D) The sequence motifs for BLIMP1 binding in comparison to those identified previously (Doody et al., 2010; Kuo and Calame, 2004). (E) Motif probability graph showing the position of the consensus motifs in the BLIMP1-binding sites. (F) The sequence motifs for T binding. A motif sequence from the JASPAR CORZ database (TF00006 as PazarID) (Mathelier et al., 2014) is shown as a reference. (G) Motif probability graph showing the position of the consensus motifs in the T-binding sites. (H) Expression of representative targets of BLIMP1 (left) and T (right) during in vitro PGC specification. (I) Venn diagram showing the overlap between genes (< 15 Kb from the TSSs) bound by BLIMP1 in d2 PGCLCs and those bound by T in EpiLC aggregates stimulated by BMP4 for 36 hrs. (J) GO analysis of BLIMP1 targets (core enrichment genes identified in Figure 7E). (K) GO analysis of T targets (core enrichment genes identified in Figure 7F).
8
SUPPLEMENTAL TABLES (Table S1-Table S6), see separate Excel documents
Table S1. Primers Used in This Study, Related to All Figures
Table S2. Antibodies Used in This Study, Related to All Figures
Table S3. Outlines and Mapping Statistics of ChIP-seq experiments, Related to All
Figures
Table S4. Gene Expression and Chromatin States Analyzed in This Study, Related
to All Figures
Table S5. Cell-type specific H3K27ac peaks and TF binding motifs in Such Peaks,
Related to Figure 2
Table S6. BLIMP1 and T Binding Sites During In Vitro PGC Specification, Related
to Figure 7
9
SUPPLEMENTAL EXPERIMENTAL PROCEDURES
Analysis of the Gene Expression for In Vitro PGC Specification
The gene expression data on an Affymetrix Gene Chip microarray reported in (Hayashi
et al., 2011; Nakaki et al., 2013) were used for the transcriptome analysis of in vitro
PGC specification (GEO numbers: GSM744093 and GSM744094 for ESCs,
GSM744095 and GSM744096 for EpiLCs, GSM1070847 and GSM1070848 for d2
PGCLCs, and GSM744101 and GSM744102 for d6 PGCLCs).
A gene list was created using Ensemble Genes with genes that are included in the
microarray probes for Affymetrix Mouse430_2. In the case that multiple probes were
assigned to a single gene, the probe that gave the highest average expression values was
selected. The promoter classes (HCPs, ICPs, and LCPs) were defined as reported
previously (Borgel et al., 2010).
Significantly expressed genes were defined as those showing an averaged log2
expression level > 8 in two biological replicates. Differentially expressed genes
(Figure 1C) were defined by significant expression levels and fold changes > 2 in at
least one pair-wise comparison among ESCs, EpiLCs, and d2 and d6 PGCLCs, and such
genes were further classified by the highest expression levels among the four key cell
types. Gene ontology analysis was performed using the DAVID gene ontology
functional annotation tool (http://david.abcc.ncifcrf.gov/) (Huang da et al., 2009a, b).
Generation of EGFP-Blimp1 Knock-in Mice and ESCs
All the animal experiments were performed under the ethical guidelines of Kyoto
University and RIKEN CDB. Noon of the day when the copulation plugs of mated
females were identified was designated as embryonic day (E) 0.5.
The targeting vector for the EGFP-Blimp1 knock-in allele, in which EGFP cDNA and a
linker sequence (the BspEI-HindIII sequence in the pEGFP-C1 plasmid; Addgene) were
inserted into the first ATG of the Prdm1 genes, was constructed using SalI-NotI and
SwaI-XhoI sites of the DT-A-pA/loxP/PGK-Neo-pA/loxP vector
(http://www.cdb.riken.jp/arg/cassette.html). The targeting vector was linearized using
the SalI site, and electroporated into the TT2 ESC line (Yagi et al., 1993).
Homologous recombination was screened by PCR using the primer set NeoGt-1/F and
Prdm1_N3659/R (Supplemental Table S1), and confirmed by Southern blot analysis
with the 5’- and 3’- probes (PCR amplicons by primers described in Supplemental Table
S1), which detected AvrII and HindIII sites of the Prdm1 locus, respectively. Random
integration was ruled out by Southern blot analysis using a probe targeting the Neo
locus. The homologous recombinant ESCs were injected into eight-cell-stage embryos
of ICR mice to generate chimeric mice. Chimeras with a high ESC contribution were
judged by coat color and mated with C57BL/6 females to generate Prdm1+/EGFP-Blmp1-Neo