(a) Scheme showing activation of the canonical BMP signalling pathway through phospo-Smad1/5 binding of the double BMP responsive element (BRE) and the resulting transcription of GFP. (b) Three-dimensional whole-mount image of an E10.5 immunostained BRE–GFP mouse embryo (37 somite pairs). Anti-CD31 (magenta), cKit (red) and GFP (green) antibody staining shows the predominantly ventral distribution of GFP⁺ cells in various cell types within the aorta and underlying mesenchymal cells. (c) High-magnification image of a transverse section of an E11 BRE–GFP dorsal aorta stained with anti-GFP-antibody and 4,6-diamidino-2-phenylindole (DAPI). Intra-aortic hematopoietic clusters contain GFP⁺ and GFP⁻ cells. Some endothelial cells are GFP⁺. (d) Representative FACS plot with side scatter on the y axis showing percentage of E11 AGM GFP⁺ cells (see for control). (e) Percentage donor cell chimerism in the peripheral blood of adult irradiated transplant recipients at 4 months after injection of E11 AGM GFP⁺ and GFP⁻ cells (2–4.5 embryo equivalents (ee) of AGM cells injected per mouse; n=3 independent transplantation experiments). *P=0.05 by z-Test for proportions.

Representative FACS plots showing percentage of GFP⁺ cells with side scatter shown on the y axis in the (a) E14 FL and (c) adult bone marrow (BM) of BRE–GFP mice (see for controls). Transplantation results of adult irradiated recipients injected with GFP⁺ and GFP⁻ cells sorted from (b) E12, E13 or E14 FL (n=3) or (d) adult BM (n=4). Secondary irradiated adult recipients received BM cells, unsorted (3 × 10⁶) or GFP⁺ (1.5 × 10⁵) and GFP⁻ (0.8, 1.8 and 11.8 × 10⁶) sorted cells from repopulated primary recipients. Data from all secondary recipients were pooled. Peripheral blood (PB) donor chimerism is shown for all recipients. Representative FACS plots showing the sorting strategy for LSK-SLAM GFP⁺ and GFP⁻ cells from (e) E14 FL and (g) adult BM. Gates are shown for the Lin- and LSK fractions. Histogram analysis shows percentages of CD150⁺ cells in the LSK fractions and the percentages of GFP⁺ cells in the LSK-SLAM fractions of FL and BM. Grey curves show control non-transgenic LSK SLAM cells. Percentages of LSK SLAM cells that are GFP⁺ (black) or GFP⁻ (grey) as found in (f) E14 FL (n=3) and (h) adult BM (n=3). **P<0.001 T-test.

(a) FPKM values for BMP receptor and Smad genes expressed by GFP⁺ and GFP⁻ LSK SLAM sorted HSCs from E14 FL and adult BM. ND, not detected. (b–d) Gene list enrichment analysis on genes with more than two fold difference in expression level between GFP⁺ and GFP⁻ LSK SLAM sorted cells from FL and BM using Enrichr web-application (threshold false discovery rate (FDR)=0.01). GFP⁺ high genes are enriched in (b) BMPR2 gene targets (FDR<1.0e⁻¹³; Gene Expression Omnibus (GEO) Kinase perturbation gene sets). Gene lists available on request. (c) Also, overrepresented in GFP⁺ high genes are genes involved in homoeostasis (FDR=9.8e⁻⁴; Mouse Genome Informatics (MGI) mammalian phenotype term MP0001764), metabolism (FDR=5.4e⁻⁴; MGI mammalian phenotype term MP0005266), MYC (FDR=2.7e⁻⁴; TRANSFAC and JASPAR Position Weight Matrix (PWM) genesets), and STAT5B (FDR=2.7e⁻⁴; TRANSFAC and JASPAR PWM gene sets); (d) GFP⁻ high genes are enriched for NFE2 (FDR=9.6e⁻³; TRANSFAC and JASPAR PWM gene sets), NFKB1 (FDR=2.7e⁻³; TRANSFAC and JASPAR PWM gene sets), SP1 (FDR=2.7e⁻³; TRANSFAC and JASPAR PWM gene sets), and haematopoietic system (FDR=5.2e⁻⁷; MGI mammalian phenotype term MP0002396)-related genes.

Limiting dilution transplantation experiments show the frequencies of HSCs in GFP⁺ and GFP⁻ cell fractions from (a) E14 FL and (c) adult BM. See for detailed results. Bar graphs showing percentages of long-term repopulating (LTR) HSCs that are GFP⁺ (black) or GFP⁻ (grey) in (b) E14 FL (n=5) and (d) adult BM (n=5) as calculated from limiting dilution transplantations. Summary of lineage output analysis of repopulated recipient mice injected with BM (e) GFP⁺ and (f) GFP⁻ cells at the clonal HSC level (n=3). Sorted GFP⁺ and GFP⁻ adult BM cells from 14–32-week-old BRE–GFP (Ly5.2) mice were injected intravenously into Ly5.1 irradiated (9 Gy) recipients and donor chimerism measured at 4 months post transplantation. Peripheral blood was analysed for donor Ly5.2 cell marker in granulocyte–macrophage (GM) lineages (Gr-1 and CD11b) and B- and T-cell lineages (B220 and CD3) by flow cytometry. Per cent donor contribution and donor percentage of GM and B+T were calculated according to Cho et al., to reveal HSC type in each recipient. P values were calculated by Fisher exact test between the GFP⁺ and the GFP⁻ fractions (P=0.007) and by two sample test for proportion (χ²) between balanced (*P=0.005), My (**P=0.031) and Ly-HSCs (P=0.727) of the two groups, GFP⁺ and GFP. (g) Fragments per kilobase of transcript per million mapped reads (FPKM) values for selected lymphoid genes expressed by GFP⁺ and GFP⁻ LSK SLAM sorted HSCs from adult BM.