Experiments were carried out as described in Fig. 2. (A) The relative percentage of EBs with beating foci were determined from 3 distinct EB populations (group 1: 100% Ainv; group 2: 50% AinV / 50% iGATA4ES (mixed); group 3: 100% iGATA4ES) cultured with or without induction. Gata4 induction at day 2 in mixed EBs results in a significant increase in the number of beating EBs at day 10 (asterisk, p<0.01). These data represent the combined results of 4 independent assays. Error bars represent the standard error of the mean. (B) Shown are results from a single but reproducible representative experiment in which mixed EBs were either left uninduced (−Dox) or induced at different days as indicated (d1 = day 1, etc.). EBs were cultured until day 6, plated for cardiomyocyte assays and beating foci counted at day 10. In a reproducible manner, maximal foci are seen with induction at day 2–3. (C) Efficient induction of cardiogenesis occurs over a range of iGATA4 ESC ratios. EBs were derived from starting ratios of parental AinV and iGATA4 ESCs (Ainv: GATA4, respectively), as indicated. At day 6, EBs were replated on gelatinized petri dishes in media supporting cardiomyocyte differentiation. The resulting EB clusters were counted at day 10. Shown is the combined result of three independent experiments. * p<0.006 and ** p<0.033 when comparing day 2 uninduced (−, grey bars) to the induced (+, green bars) group using Student's T-test. Even at the 1:10 ratio there is a clear trend toward induction, although it is more variable and did not show statistical significance. Error bars represent the standard error of the mean. (D) Shown is a representative example of flow cytometry using a 1:1 ratio of iGATA4ES:Ainv EBs that were either left uninduced (−dox) or induced at day 2. The EBs were plated at day 6, and at day 10 the EBs were harvested, dissociated, and evaluated concurrently for GFP expression (fluorescence) and cTnT expression (using intracellular antibody staining). The cTnT-positive cell population is indicated by the block grey arrow in the +dox panel. The induced EBs show typically 10% cardiac cells (compared to 1% in control samples), although this is likely an under-estimate, since the beating foci containing cardiomyocytes are difficult to dissociate, and only single cells were counted. The GFP+ cells fall outside the cTnT gate, and the percentage of double-positive cells in the induced samples is never above background. (E) Quantitative real time PCR analysis for cardiac markers, as indicated, in day 6 induced mixed EBs normalized to levels in day 6 uninduced controls. Shown are results from three independent experiments, with relative expression data log2 transformed. Error bars indicate the standard error of the mean with asterisks denoting statistical significance (p<0.05). In other cases, a single sample gave a relatively high difference. While this limited statistical significance, the trend for all cardiac markers is nevertheless clear.

Shown are accumulated qPCR results from three independent experiments measuring transcripts at day 3 for cells from EBs induced to express GATA4 protein at day 2. Levels from induced EBs were compared with uninduced controls and the relative differences are plotted with the data log2 transformed. Early mesoderm markers are repressed, while Sox17 transcript levels are increased. Error bars indicate the standard error of the mean with asterisks denoting statistical significance (p<0.05).

Shown are results from three independent experiments using three different iGATA4ES subclones. Cells were harvested at day 6 and qPCR was performed to measure relative fold change comparing the values from each subclone to its uninduced control, shown in the graphs with the data log2 transformed. Top panel: A subset of defined visceral endoderm markers are increased in levels (Apoc2, Cubn), while others (Emp2, Pem1) are not, Middle panel: Expression of pan-endoderm markers (FoxA2, FoxA3, Tspan7) are increased whereas by day 6 Bry and Flk1 transcript levels are unchanged or downregulated. Bottom panel: definitive endoderm marker levels are generally increased (Cxcr4, Dlx5, FoxG1, Idb4, Pax9, Pyy). Error bars indicate the standard error of the mean with asterisks denoting statistical significance (p<0.05). In some other cases, a single sample gave a relatively high difference. While this limited statistical significance, the trend is nevertheless clear.

EBs were generated from the iGATA4ES line and either left uninduced or induced at day 2 to express Gata4. EBs were harvested at day 2.5, 3, 3.5, or 4 as indicated, dissociated, and cells evaluated by flow cytometry for cell death (7-AAD) and apoptosis (Annexin V). These data are from a representative experiment that was reproducible with equivalent results.

(A) Cre-mediated recombination between lox sites facilitates integration of the Gata4 cDNA into the targeting site of parental AinV cells. The FLAG-Gata4ORF-iresEGFP transgene is flanked by a lox site (left black triangle) and a phosphoglycerokinase promoter (PGK-ATG). The target site in AinV ES cell contains a tetracycline response element (tetOP) followed by a lox site and a neomycin resistance gene lacking its start codon (NEO-pA). Upon site-specific integration, the transgene is placed under control of the tetracycline response element. Note that PGK-ATG lies in frame with NEO-pA to establish neomycin resistance. (B–E) Representative results using a clonal iGATA4ES cell line. Essentially identical results were obtained using pooled clones or several independent lines derived by serial dilution cloning. (B, C) Merged 10x phase and fluorescence images of uninduced (B) and induced (C) iGATA4ES cells. GFP expression only in induced cells indicates successful induction and non-leakiness of transgene expression. (D) Left panel: Western blot analysis using total cell extracts from uninduced (−) or induced (+) iGATA4ES-derived embryoid bodies. Protein detection was performed using a monoclonal anti-FLAG antibody or an anti-GATA4 antibody, as indicated. Right panel: Coomassie Blue staining of the western blotted gel. (E) Gel mobility shift analysis using nuclear extracts from parental cells (AINV), or iGATA4ES cells, under uninduced (−) or induced (+) conditions. The specific competitor (sp. u.c.) is a 20x excess of unlabeled probe. The non-specific competitor (unsp. u.c.) is a 20x excess of unlabeled probe containing mutations in the GATA-binding site.

Three populations of EBs were initiated in parallel at day 0: the first population is derived only from AinV parental ES cells, a second is comprised of an equal mixture of 50% AinV and 50% iGATA4ES cells, and a third is comprised entirely of iGATA4ES cells. At day 2, EBs were induced (+ dox) while control samples were left uninduced (− dox). At day 3, Gata4 transgene expression is monitored by observation of GFP expression (top panels: 10x phase; bottom panels: corresponding views under fluorescence). At day 6, EBs were plated onto gelatin under permissive conditions that allow differentiation (beating) of the cardiac progenitors. At day 10, the number of EBs containing beating foci was counted for each experimental condition.

Erythroid colony assays were carried out using the three different EB populations (100% Ainv; 50% AinV / 50% iGATA4ES (mixed); 100% iGATA4ES) cultured with (+) or without (−) induction by doxycycline. Gata4 induction at day 2 of iGATA4ES or mixed EBs results in a significant decrease in the number of primitive erythroid colonies (p<0.01 and p<0.05 respectively). These data represent the combined results of 4 independent colony assays. Error bars represent the standard error of the mean.

Shown are representative qPCR results using cDNA derived from cells sorted out of mixed EBs (group 2, 50% iGATA4ES / 50% AinV). EBs were induced with doxycycline at day 2, and at day 3 the GATA4-expressing cells were sorted as GFP+ and compared to levels of transcripts (as indicated) in the GFP−. The normalized expression data is shown log2 transformed. The GATA4− expressing GFP+ cells express relatively high levels of endoderm markers Sox17 and Dkk1, and relatively low levels for the mesoderm markers. Multiple independent experiments gave similar results.

(A) Shown is the flow cytometry strategy, with live cells gated for epCAM (endoderm) and Dpp4 (visceral endoderm). (i) A representative experiment is shown of an uninduced iGATA4ES control and the same cells induced to express GATA4 at day 2 of EB development. Cells were harvested and analyzed at day 4 on a FACSCalibur cytometer. This experiment was repeated (n = 3) with similar results using independent subclones. (ii) Gating strategy used in the preparative sorting experiments with iGATA4 EBs induced at day 2 and dissociated at day 4 of development. Cells were gated as negative (N, epCAM-/Dpp4−), definitive endoderm (DE, epCAM+/Dpp4−), or visceral endoderm (VE, epCAM+/Dpp4+) and populations were sorted with a BD FACSVantage cell sorter. Secondary analyses showed the populations are greater than 90% pure (not shown). (B) Shown are representative qPCR analyses to validate the sorting strategy. The relative transcript levels are plotted with the data log2 transformed, compared to those in the control uninduced sample. Double negative and VE and DE endoderm samples are labeled as in panel Aii. Left panel: Known VE markers, Right panel: Known pan and DE markers. Both definitive and visceral endoderm sorted cells show expression of pan-endoderm markers (FoxA2, FoxA3, Tspan7) whereas the expression of definitive endoderm (Cxcr4, Gsc) or visceral endoderm (Apoc2, Cubn) marker levels are relatively enhanced in their respective sorted populations. As expected, the double-negative sample, containing mesoderm, is also positive for Cxcr4 and Gsc. Both endoderm subtypes demonstrate expression of the anterior endoderm marker Hex and decreased levels of visceral endoderm markers Emp2 and Pem1. Equivalent results were obtained in at least 3 independent experiments.

(A) Cell populations were sorted at day 4 as described in Fig. 8 and were then mixed with dissociated cells from day 2 AinV EBs in a 1:1 ratio at a concentration of 1.25 × 10⁵ cells/ml. Reaggregated EBs were plated on gelatinized petri dishes four days later in media supporting cardiomyocyte differentiation. The resulting EB clusters were counted at day 10 (relative to the responding AinV cells). Shown are the combined results of two independent experiments. The asterisk indicates p<0.011 when comparing AinV:VE to AinV using Student's T-test. (B) Differential expression of cardiac inducing molecules in the two endoderm populations. Shown are representative qPCR results using cDNA generated from sorted cell populations. Cells were derived and sorted as described in Fig. 8, and the levels relative to uninduced controls are plotted with the data log2 transformed. Both endoderm sub-types express enhanced levels of the Wnt antagonists Dkk1 and Sfrp5, whereas the visceral endoderm, compared to definitive endoderm, shows relatively increased levels of Bmp2 and Bmp6 transcripts. Equivalent results were obtained in at least three independent experiments.

The EBs derived entirely from iGATA4ES cells were induced to express Gata4 at day 2, plated at day 10 under conditions that normally support mesoderm development, and harvested at day 17. Shown are representative qPCR results analyzing cDNA isolated from induced EBs compared with levels of transcripts (as indicated) in the uninduced control cells. The relative levels are plotted with the data log2 transformed. The GATA4-expressing cells have relatively high levels of transcripts for liver markers Afp and Alb, and relatively unchanged levels for the pancreas markers Pdx1 and Ins. Multiple independent experiments gave equivalent results.