(
A) Heatmap showing row-normalized
z-scores of EN marker genes and
T-REX17 expression across various embryonic and adult tissues as measured by RNA-seq. The tree represents hierarchical clustering (based on Euclidean distance). RNA-seq expression values are calculated as normalized row-read-coverage. Note the high specificity of
T-REX17 expression as compared to a much broader expression of endodermal transcription factors, including
SOX17. A list of the curated dataset from the Roadmap Epigenome project used in this heatmap is provided in
Supplementary file 1. (
B) Heatmap showing column-normalized
z-scores of EN marker genes and
T-REX17 expression during in vitro pancreatic lineage differentiated cell types including primary isolated α- and β-cells as measured by RNA-seq (
Alvarez-Dominguez et al., 2020) (upper panel). The tree represents hierarchical clustering (based on Euclidean distance). RNA-seq expression values are calculated as normalized row-read-coverage.
T-REX17 and
SOX17 expression profiles plotted as row-read-coverage normalized
z-scores of the indicated in vitro generated cell type during pancreatic differentiation (lower panel). Note the transient and restricted expression of
T-REX17 specifically in definitive endoderm. (
C) Scatter plot showing the expression of a set of endoderm lncRNAs (
DEANR1, LHX1-DT, GATA6AS1, NKX2-1AS1, GATA3AS1, T-REX17) and the corresponding TFs (
FOXA2, LHX1, GATA6, NKX2-1, GATA3, SOX17) in the same set of EN tissues of
Figure 2B. A linear model excluding the expression in EN was fit for each lncRNA-TF couple. Pearson correlation coefficients as well as corresponding p-values are displayed in the bar-plot (bottom panel). Note that the
T-REX17-SOX17 couple has the lowest degree of tissue co-expression. (
D) Genome browser tracks displaying RNA levels at the
T-REX17 locus in PSCs and the three germ layers. Note
T-REX17 expression specificity as compared to
MRPL15. (
E) Uniform Manifold Approximation and Projection for Dimension Reduction (UMAPs) showing cell states (upper left panel) and
T-REX17 expression (upper right panel) in cells derived from a human gastrulating embryo (
Tyser et al., 2021). ScRNA-seq track from cells belonging to the endoderm cluster showing reads mapping to the
T-REX17 locus (bottom panel). (
F) Cell fractionation of endodermal cells followed by RT-PCR, agarose gel purification and band intensity quantification. Right panel indicates PCR products of respective target gene RNA of the respective cell fraction. Left panel indicates relative band intensity quantification (provided in
Supplementary file 1) of respective cell fraction (dark grey=chromatin fraction, light grey=nucleoplasmatic fraction, white = cytoplasmatic fraction) from RT-PCR-products in the left panel. Bar heights of each fraction represent mean values, error bars indicate SD (n=2). (
G) MinION-seq reads track showing
T-REX17 coverage and structure in endodermal cells. Sequencing read distribution histogram (top) and individual reads sorted by their start location (bottom) are displayed. Exon 1, 2, and 3 are highlighted by shading boxes. Sequence mismatches and matches are color coded as described. Split reads and deletions are shown as thin horizontal lines. (
H) Sanger sequencing of 3’/5’ RACE PCR products. Amplicon-specific sequencing results are shown below the query sequence (hg19). Sequencing mismatches are highlighted in red. Primer pairs relative positions used for the PCRs are shown for each product. Sanger sequencing chromatogram color code is used to show the raw reads data.