a, Optimization of pH conditions for Oct4-GFP induction. Five days after CD45-positive cells were exposed to acidic solution treatment at different pH, Oct4-GFP expression was analysed by FACS (n = 3, average ± s.d.). b, Gating strategy for Oct4-GFP⁺ cell sorting. Top: representative results 7 days after the stress treatment. Bottom: non-treated control. P3 populations were sorted and counted as Oct4-GFP⁺ cells for all experiments. c, Controls for FACS analysis. In Oct4-GFP⁺ cell analysis, the grey and white histograms indicate the negative control (non-stress-treated Oct4-gfp haematopoietic cells) and the positive control (Oct4-gfp ES cells), respectively. Also, the green histograms indicate non-treated cells (left) and stress-treated cells at day 7 (right). In CD45⁺ cell analysis, the grey and white histograms indicate the negative (isotype) and positive controls, respectively. The red histograms indicate non-stress-treated cells (left) and stress-treated cells at day 7 (right). d, Oct4-GFP⁺ cell generation from various subpopulations of CD45⁺ cells. Seven days after the stress treatment, Oct4-GFP expression was analysed by FACS (n = 3, average ± s.d.). Among total CD45⁺ fraction and its subfractions of CD19⁺, CD90⁺, CD34⁺ and CD34⁻ cells, the efficacy of CD34⁺ cells was significantly lower than the others. P < 0.05 by the Newman–Keuls test and P < 0.01 by one-way ANOVA. e, Comparison of culture conditions for low-pH-induced conversion. Stress-treated cells were cultured in various media. The number of Oct4-GFP-expressing clusters was counted at day 14 (n = 3, average ± s.d.). ***P < 0.001 (B27+LIF versus all other groups); Tukey’s test. In the case of 3i medium, although the clusters appeared at a moderate efficiency, they appeared late and grew slowly. ACTH, ACTH-containing ES medium; ES+LIF+FBS, 20% FBS+LIF-containing ES culture medium; B27, DMEM/F12 medium containing 2% B27; B27+LIF, DMEM/F12 medium containing 2% B27+LIF; EpiSC, EpiSC culture medium containing Fgf2+activin. f, Signalling factor dependency of STAP cell generation. Growth factors that are conventionally used for pluripotent cell culture such as LIF, activin, Bmp4 and Fgf2 were added to basal culture medium (B27-supplemented DMEM/F12) in different culture phases (days 0–7, 2–7 and 4–7), and Oct4-GFP expression was analysed by FACS at day 7 (n = 3, average ± s.d.). g, h, Time course of apoptosis after the low-pH exposure. Stress-treated cells and non-stress-treated control cells were stained with CD45, annexin-V and propidium iodide at day 0 (immediately after stress treatment), day 3 and day 7. g, Blue bars, GFP⁺CD45⁻; orange bars, GFP⁻CD45⁺. Percentages in total cells included propidium-iodide-positive cells. h, Annexin-V-positive cells in these cell populations were analysed by FACS.

a, Oct4 protein expression in STAP cells was detected by immunostaining at day 2 (left) and day 7 (right). b, Live cell imaging of STAP conversion (grey, CD45 antibody; green, Oct4-GFP). See Methods for experimental details to monitor live CD45 immunostaining. c, Immunostaining of a parental CD45⁺ cell (left) and an Oct4-GFP⁺ cell (right). Scale bar, 10 μm. d, EdU uptake assay (n = 3, average ± s.d.). e, Schematic of Tcrb gene rearrangement. f, T-cell-derived STAP cells. Scale bar, 100 μm. g, Genomic PCR analysis of (D)J recombination at the Tcrb gene of T-cell-derived STAP cells. G.L. is the size of the non-rearranged germline type, whereas the smaller ladders correspond to the alternative rearrangements of J exons (confirmed by sequencing). Negative controls (ES cells), positive controls (lymphocytes) and T-cell-derived STAP (two independent preparations on d7) are indicated.

a, Expression of pluripotency marker genes in STAP cells derived from T cells (n = 3, average ± s.d.). b, Expression of pluripotency marker genes in STAP cells. In this experiment, Oct4-GFP⁺ cells seen in live cell imaging (Extended Data Fig. 2b) were analysed to confirm their conversion into STAP cells (n = 3, average ± s.d.). c, Haematopoietic marker expression during STAP conversion from CD45⁺ cells (n = 3, average ± s.d.). d, Formation of visceral endoderm-like surface epithelium in differentiating STAP cluster on day 2 (left) and day 8 (right). Scale bars, 50 μm.

a–c, Teratomas formed from STAP cell clusters included neuroepithelium (a), striated muscle (b) and pancreas (c; right, high-magnification view showing a typical acinar morphology and ductal structures). Scale bars, 100 μm. d, Teratoma-forming ability of Oct4-GFP⁺ and Oct4-GFP-dim cells (isolated by FACS, top). Oct4-GFP⁺ cells, but not Oct4-GFP-dim cells, efficiently formed teratomas (table at the bottom). However, because STAP cells were dissociation-intolerant, the teratoma-forming efficiency of dissociated Oct4-GFP⁺ cells was lower than that of non-dissociated STAP cell clusters. e, Gene expression of Oct4-GFP⁺ and Oct4-GFP-dim cells (n = 3, the average ± s.d.). Haematopoietic marker gene expression (left) and early lineage marker gene expression (right) are shown.

a, Immunostaining for Ki67 and BrdU. STAP cell clusters (top) and ES cell colonies (bottom) are shown. For BrdU uptake, BrdU was added into each culture medium (10 μM) for 12 h until fixation. Scale bar, 100 μm. b, Transformation assay by soft agar culture. Neither Oct4-GFP⁺ nor Oct4-GFP-dim cells showed colony formation in soft agar, whereas ES cells and STAP stem cells showed anchorage-independent growth in the same LIF-B27 medium. Scale bar, 100 µm. Proliferated cells were lysed and the amount of DNA in each well was estimated by chemical luminescence (graph). n = 3 , average ± s.d. c, No substantial change in chromosome number was seen with STAP cells in the CGH array. Genomic DNA derived from CD45⁺ cells (male) was used as reference DNA. The spikes (for example, those seen in the X chromosome) were nonspecific and also found in the data of these parental CD45⁺ cells when the manufacturer’s control DNA was used as a reference. d, qPCR analysis for pluripotency markers that highly express in ES cells, but not in EpiSCs. Average ± s.d. e, Immunostaining of markers for mouse EpiSC and ES cells. Scale bar, 100 μm. f, g, H3K27me3⁺ foci in female cells, which are indicative of X-chromosomal inactivation. These foci were not observed in male cells. Scale bar, 10 μm. In the case of female STAP cells, ~40% of cells retained H3K27me3⁺ foci (g). **P < 0.001; Tukey’s test. n = 3, average ± s.d. Although nuclear staining looked to be higher in STAP cells with H3K27me3⁺ foci (f), this appeared to be caused by some optical artefacts scattering from the strong focal signal. h, qPCR analysis for the tight junction markers Zo-1 and claudin 7, which were highly expressed in EpiSCs, but not in ES cells or STAP cells. **P < 0.01; ns, not significant; Tukey's test; n = 3, average ± s.d.

a, Alkaline phosphatase expression of STAP cells derived from adipose-derived mesenchymal cells. Scale bar, 100 μm. b, E-cadherin expression of STAP cells derived from adipose-derived mesenchymal cells. Scale bar, 50 μm. c, FACS sorting of dissociated neonatal cardiac muscle cells by removing CD45⁺ cells. d, Cardiomyocyte marker gene expression during STAP conversion from cardiomyocytes (n = 3, average ± s.d.).

a, 2N chimaeras generated with STAP cells derived from Oct4-gfp C57BL/6 mice (left) and 129/Sv × C57BL/6 F₁ mice (right). b, Generation of chimaeric mice from STAP cells by cluster injection. STAP cells used in the experiments above were generated from CD45⁺ lymphocytes of multiple neonatal spleens (male and female tissues were mixed). *All fetuses were collected at 13.5 d.p.c. to 15.5 d.p.c. and the contribution rate of STAP cells into each organ was examined by FACS. **The contribution of STAP cells into each chimaera was scored as high (>50% of the coat colour of GFP expression). ***B6GFP: C57BL/6 mouse carrying cag-gfp. c, Production of offspring from STAP cells via germline transmission. Chimaeras generated with 129/Sv × B6GFP STAP cells (obtained from the experiments shown in b) were used for germline transmission study. d, 4N embryos at E9.5 generated with STAP cells derived from F₁ GFP mice (B6GFP and DBA/2 or 129/Sv). B6GFP, C57BL/6 mouse carrying cag-gfp.

a, Compatibility of 2i conditions with STAP stem-cell derivation from STAP cells and STAP stem-cell maintenance. STAP stem cells could not be established directly from STAP cells in 2i + LIF medium (top). However, once established in ACTH medium, STAP stem cells were able to survive and expand in 2i + LIF medium. Scale bar, 100 μm. b, Q-band analysis (n = 4; all cell lines showed the normal karyotype). c, Multicolour FISH analysis (n = 8; all cell lines showed the normal karyotype) of STAP stem cells. d, Methylation status of the Oct4 and Nanog promoters. e, Electron microscope analysis of STAP stem cells. Scale bar, 1 μm. f, g, Beating cardiac muscle (mesoderm; 38%, n = 8). Red line indicates an analysed region for kymograph (g). h, Clonability of STAP stem cells. Clonal expansion from single STAP stem cells was performed. Pluripotency of clonal cell lines was confirmed by teratoma formation assay, showing the formation of neuroectoderm (left), muscle tissue (middle) and bronchial-like epithelium (right). Scale bar, 100 μm. i, Production of chimaeric mice from STAP stem-cell lines using diploid embryos. *These STAP stem-cell lines were generated from independent STAP cell clusters. j, Production of mouse chimaeras from STAP stem-cell lines by the tetraploid complementation method. *These STAP stem-cell lines were generated from independent STAP cell clusters. k, No H3K27me3-dense foci are seen in female STAP stem cells (n = 50; the CD45⁺ cell is a positive control). Scale bar, 10 μm.

a, Percentages of Oct4-GFP-expressing cells 7 days after stress treatment. Somatic cells were isolated from various tissues and exposed to different stressors. Oct4-GFP expression was analysed by FACS. b, Oct4 and Oct4-GFP expression induced in the reflux oesophagitis mouse model as an in vivo acid exposure model (top, experimental procedure). Oct4, but not Nanog, expression was observed in the oesophageal epithelium exposed to gastric acid (75% of 12 operated mice).