Figure 1.
Validation of selected binding sites.
Real-time PCR showing relative enrichment values for all 3 biological replicates after amplification. 5′proximal promoter regions were selected for primer sites. SOX2-Exon, ACTB-promoter and HBB-promoter were used as negative controls.
Figure 2.
Influence of peak finding algorithms on binding sites.
Venn diagrams, illustrating the overlaps between different peak analysis programs. A–C: sorted replicate-wise. D: Real-time PCR validation, showing relative enrichment values for 2 non-amplified biological replicates. Ten randomly chosen peak regions, identified by our peak analysis were chosen for this analysis. ZNF398, POMGNT1, ZNF532 and MAGED2 were identified by all three algorithms. FIGN and LPHN2 were only detected by brute-force. PIPOX and H2AFY were only detected by MA2C. TMEM139 and ZIC4 were only detected by TAMALPAIS. E: Venn diagram, showing the overlap between different cell lines- NCCIT, this study, H9 [3] and NTERA2 [13].
Figure 3.
Correlation between octamer motifs and peak score values.
Boxplot, showing the distribution of the quality of octamer motifs in relation to our defined peak score. For a peak-score of 0.5, half of the motifs will have a motif-score of 7.3 and above. The average motif score will decrease slightly for a peak score of 0.33 and a significant drop in the motif score can be perceived for a peak score of 0.11.
Figure 4.
The NANOG promoter harbours an evolutionary conserved binding site.
The conserved binding site is shown for OCT4 (red) and SOX2 (bold). A: Bandshift showing a supershift with OCT4 antibody, using NCCIT-derived nuclear extracts and a Cy5 labelled probe in the 5′region of the NANOG promoter bearing the OCT4-SOX2 motif. Binding specificity was tested using oligonucleotide competitors. 1) 20-fold excess of unlabelled competitor. 2) Supershift with OCT4 (sc9081) antibody. 3) Nuclear extract with Cy5-labelled probe. B: Alignment of the OCT4-SOX2 binding sequence in multiple species. C: Bitscore model of the re-constructed OCT4 PWM. Note that the OCT4 PWM sequence is presented in the opposite strand with respect to the sequence shown in (B) above.
Figure 5.
A gene regulatory network based on the 31 genes common in OCT4 ChIP-on-Chip targets derived from NCCIT, NTERA2 and H9 cells.
GADD45G was also included in this analysis. The network was generated using the web-based program STRINGS [27]). Pink lines: connectivity based on experimental evidence. Green lines: connectivity based on text mining.
Table 1.
Examples of Homeodomain containing genes bound by OCT4 in NCCIT and H9 cells [3].
Figure 6.
Six distinct OCT4 binding modules.
Shown are the peak scores, relative to the overlap between MAC2, TAMALPAIS, the in-house developed algorithm - brute-force [19] and the biological replicates. Peak profiles could be screened for the octamer and SOX2 motifs.
Figure 7.
The USP44 promoter harbours the evolutionary conserved OCT4 binding site but lacks the SOX2 motif.
A: Sequence containing the conserved POU site as displayed by the UCSC genome browser. B: Real time PCR confirmation of the presence of the OCT4 binding site. Position 0 indicates the conserved region seen in (A). C: Multiple alignments showing evolutionary conservation of the OCT4-bound region. The sequences depicted in blue and green are uncharacterised with respect to transcription factor recognition and binding. D: Western blot analysis of proteins bound to biotinylated oligos representing the promoter fragment shown in (A). The OCT4 antibody shows higher binding intensity to the USP44- specific probe compared to the corresponding scrambled oligo.
Figure 8.
The GADD45G promoter harbours the evolutionary conserved OCT4 binding sites.
A: Bandshifts showing supershifts with OCT4 antibody using NCCIT cells derived- nuclear extracts using two probes in the 5′region of the GADD45G promoter containing an OCT4 motif at positions 9–15 (lane 1–3) and 17–23 (lane 4–6) of 31 nucleotides. Lane 3,6: Nuclear extract plus labelled probe. Lane 2,5: same as lanes 3 and 6 but with the addition of OCT4 antibody (sc-9081). Lane 1,4: same as lanes 3 and 6 but with the addition of a 20-fold increase in unlabelled competitor oligo. B: Multi-species alignment of the selected region chosen for the bandshift assay, the conserved OCT4 binding site is highlighted in red. C: Real time PCR confirmation of the presence of the OCT4 binding site. Position 0 indicates the position shown in the alignment in panel 2B.
Figure 9.
Over-expressing GADD45G in NCCIT cells.
A: Presence of GFP expression 48 h post-transfection (left) compared to the phase-contrast image of the cells. The map of the vector used is presented below. B: Scatter plot comparing the transcriptomes of GADD45G transfected cells against cells transfected with the wild-type vector. GADD45G-mediated induction of transcription factors such as HAND1 (purple), GATA4 (green), and ID2 (brown) depicted in boxes. C: Table listing the most significant GO:biological processes related to the up-regulated (>2-fold) genes. D: Real time PCR validation of a selection target genes (NANOG, SOX2 and BMP4 were below detection score 0.01).
Figure 10.
Potential new interaction partners of OCT4.
De novo motif discovery for genes, identified as OCT4 indirect targets and differentially regulated (2-fold and above) in NCCIT cells but lacking the OCT4 and SOX2 motif within the promoter region analysed. The 4 most significant motifs identified and the potential transcription factor binding sites related to these motifs are displayed. In addition, putative regulated genes harbouring these motifs in their promoter regions shown. Red depicts up-regulated and green down-regulated in response to the ablation of OCT4 activity in ES and EC cells.
Figure 11.
Sequence alignments of selected OCT4-regulated genes under the distinct modules.
The OCT4 binding motif is represented in red and that of SOX2 in green.
Figure 12.
Hypothetical model based on module 4 of how OCT4 could be involved in regulating its target genes via non-direct DNA binding.
OCT4 might be recruited by a mediator complex (X), which has additional affinity for the discussed transcription factors (A – H). Alternatively, there might be a direct interaction between OCT4 and the transcription factor(s) (indicated by ?), which might then potentially bind to the identified in silico cis elements. Arrows: Red- induction and green- repression of transcription of the respective target genes.