Figure 1.
SELEX-Selection of specific oligonucleotides bound to GST-TBX1.
A: A pipeline illustrating the SELEX method is shown. The dsDNA was generated by PCR of the selected oligonucleotides at each found and incubated with GST-TBX1. A total of 6 rounds of selection was performed. B: EMSA was used to detect specific GST-TBX1 and [α-32P]dCTP labeled DNA complexes at 0, 2, 4 and 6 rounds of selection, with or without cold competitor (R6, cold PCR products from round 6; T, ds DNA harboring the published Brachyury half site). C: Sequence alignment shows that the optimal DNA binding motif for TBX1 is AGGTGT(G/T)(A/T) followed by two repeated similar motifs termed the Tandem Repeat (TR) and Half Site Partial Site as shown (½SPS). D: Distribution of sequences with different consensus binding motifs within the pool of oligonucleotides after 6 rounds of selection (total number = 60).
Figure 2.
Specificity of new binding sites tested via EMSA and luciferase assays.
A: The binding sites identified from the SELEX experiment were tested with GST-TBX1 separately by EMSA to determine if there is specific binding (Tandem Repeat, TR; Half Site Partial Site, ½SPS, half site; ½ Site). For comparison, the palindromic T-site was also tested but binding was very weak as compared to the newly identified binding sites and was only observed after extensive overexposure of the autoradiogram (not shown). B: Luciferase reporter constructs containing 6 copies (6x) of the TR and 6x of the ½SPS, respectively, were co-transfected with full length Tbx1-pCDNA3.1 and compared to the empty pCDNA3.1 transfection to determine if TBX1 could activate transcription of a reporter via these sites. A significant increase of luciferase activity was observed in the presence of full-length TBX1 for both the 6x TR and the 6x½SPS when compared to transfection of the empty pCDNA3.1 vector (TR: 29 fold; Students t test, *p<0.001; ½SPS: 5.6 fold; Students t-test, *p<0.02). The mutations analyzed were those previously tested in a half site where AGGTGTGA was mutated to AATTTTGA [31]. When these nucleotide changes were present in the TR, there was a dramatic decrease in activation by Tbx1-pCDNA3.1 (7.4 fold; Students t test, ♦p<0.001). The same mutation in the 6x½SPS construct did not show a significant change when compared to the normal ½SPS (n.s. not significant). All data are presented as means ±SD; n≥3.
Figure 3.
Surrounding nucleotides outside of the ½ Site are crucial for binding.
A: EMSA was performed on mutated sequences that were generated in the second ½ site to test if variation at position, P8 (A→T), P11 (G→C) and P13 (G→T) affects binding. When P8 was changed, the faster migrating binding conformation was lost. When P11 was changed most of the slower migrating binding conformation was lost. Almost all binding was completely lost when P13 was changed. B: Luciferase reporter assays were performed to test their effect on transcription. Constructs harboring 6 copies of the mutated binding site (mutated at either P8, 11 or 13, respectively) were co-transfected with the full length Tbx1 gene. Mutation of these nucleotides affected luciferase activation. Data are presented as means ± SD; n≥3. Student's t-test, *p<0.02.
Figure 4.
Human Mutations in Mouse Tbx1.
A: The position of the three known mutations in TBX1 in human patients are shown with respect to its domain structure [19]–[21]. The three mutations lie within the region that was cloned to generate the GST-TBX1 fusion protein. B: After mutagenesis, the wild type (WT) and mutant proteins were induced, purified and used for EMSAs. The protein-DNA complexes are shown before and after IPTG induction. C, D: EMSAs of the WT and three mutated proteins using the radiolabeled TR (C) and ½SPS motifs (D). E: Luciferase assays were performed after co-transfecting the reporter construct (6x TR or 6x ½ SPS) with WT or each mutated full length Tbx1-pCDNA3.1 construct. The F148Y mutation led to a decrease in activation (TR- 17 fold decrease, ♦p<0.0005; ½SPS- 4.5 fold decrease, ♦p<0.0003) when compared to the WT transfection. The H194Q mutation did not lead to a statistically significant change in activation but the trend was in the direction of decreased activity (not significant- n.s). The G310S mutation led to a smaller but still significant decrease in activation (TR- 2.2 fold decrease, ♦p<0.01; ½SPS- 1.3 fold decrease, ♦p<0.05). Equal amounts of WT and mutated Tbx1 was co-transfected with the respective reporter constructs to determine if there was suppression of the mutant phenotype. There was a slight increase in activation when compared to the mutated F148Y alone transfection (TR-4 fold, •p<0.006; ½SPS- 1.9 fold, •p<0.006). Under these new conditions, the F148Y mutated TBX1 with WT protein still showed reduced activation when compared to WT TBX1 (TR- *p<0.0003; ½SPS- *p<0.006). The H194Q+WT combination did not show any significant change (TR- p<0.4; ½ SPS- p<0.1). The G310S+WT combination showed a significant increase in activation (TR- 2 fold, •p<0.001; ½SPS 1.5 fold, •p<0.02) when compared to G310S mutant alone. All data are presented as means ±SD; n≥3. p-values were determined using the Student's t-test. F: Immunoflourescence experiments were performed with antibodies to TBX1 on transfected Jeg3 cells to valdiate that the mutated constructs were localized to the nucleus (green). Nuclear localization was confirmed by observing expression in DAPI stained nuclei shown in blue.
Figure 5.
Genome wide search of T-sites.
A, B: Representative examples depicting the genome-wide search for ½ SPS and TR motifs. The first search was done for the ½ site in evolutionarily conserved blocks within 100 kb upstream and downstream of the TSS of genes. The second search for the TR was expanded since there were fewer sites (303), irrespective of distance to the TSS and conservation across mammals or vertebrates. C, D: Bar graphs depict the number of genes comprising top gene ontology categories using the Database for Annotation, Visualization and Integrated Discovery (DAVID v6.7- http://david.abcc.ncifcrf.gov).
Figure 6.
The Fgf8 locus has a ½SPS that is bound and activated by TBX1.
A: Snapshot from the UCSC Genome Browser showing the position of the ½SPS, located 4 kb downstream of the Fgf8 gene. The region is ultraconserved from humans to fugu. B: EMSA for the 40 bp element in the Fgf8 locus harboring the ½SPS motif with GST-TBX1. Lanes with unlabeled competitor is shown. C: The 400 bp element within Fgf8 locus was subjected to luciferase assays in cell culture and was activated in the presence of Tbx1-pCDNA3.1 (5-fold; Students t- test *p<0.002; data are presented as means ±SD; n≥3). D: Whole mount in situ hybridization of Fgf8 antisense mRNA in Tbx1+/+ (left) versus Tbx1-/- (right) mouse embryos at E10.5. Expression of Tbx1 is reduced in the distal pharyngeal apparatus (arrow), but remains in the rest of the embryo (first pharyngeal arch, head, limb buds, somites).
Figure 7.
The Bmper locus has an intronic TR site that is bound and activated by TBX1.
A: Snapshot of the UCSC browser showing the TR site in intron 13 in the Bmper locus and it is evolutionarily conserved in some rodent species. B: GST-TBX1 was able to bind to the TR motif in the Bmper locus. C: Luciferase assay results demonstrate a 1.8 fold increase (Students t-test *p<0.003; ±SD; n≥3) in activation when Tbx1pCDNA3.1 was cotransfected with the 400 bp element. D: Whole mount in situ hybridization of Bmper antisense mRNA in Tbx1+/+ (left) versus Tbx1-/- (right) embryos at E10.5. Expression of Tbx1 is reduced in the core mesoderm of the pharyngeal apparatus (long arrow) and otic vesicle (short arrow), but remains in the somites.
Figure 8.
TBX1 can bind to a TR site in the Otog-MyoD locus.
A: Snapshot from the UCSC genome browser showing the TR site in intron 13 of the Otog gene and 77 kb downstream from the TSS of MyoD. B: EMSA of GST-TBX1 and the TR motif in the Otog-MyoD locus with or without unlabeled TR double stranded oligonucleotide competitor. C: Luciferase reporter assays using the intronic element showed a 2.5 fold increase (Student's t-test, *p<0.05; means ±SD; n≥3) in activation when in the presence of TBX1. D: Expression of MyoD is lost in the 1st pharyngeal arch core mesoderm in Tbx1-/- embryos at E10.5 (arrow) but it remains in the somites. Forebrain expression in both embryos represents a staining artifact.