Figure 1.
GBX2 overexpression and ChIP in human PC-3 cells.
(A) Schematic representation of the HA-GBX2 and HA-GBX2ΔHD recombinant proteins containing the proline-rich region (PR), DNA-binding homeodomain (HD), and the HA epitope tag located at the amino terminus. Immunoflouresence of transiently transfected human PC-3 cells with HA-GBX2 (C, D), and, HA-Gbx2ΔHD (F, G). Blue channel identifies DAPI staining in the nucleus (B, E). Green channel identifies GFP-GBX2 fusion proteins. (D, G) Merge displays nuclear localization of GFP-GBX2 fusion proteins. Western blots of total lysates (H) and HA-immunoprecipitated samples (I) from mock, HA-Gbx2, and HA-Gbx2ΔHD transfected PC-3 cells.
Figure 2.
Bioinformatic analysis of the top 286 GBX2 target genes.
(A) Tissue expression analysis for the top 286 genes targeted by GBX2 determined by DAVID. Of the top 286 genes targeted by GBX2, 51% are expressed in the nervous system: 38% brain, 1% brain stem, 2% brain cortex, 3% hippocampus, 7% fetal brain. (B) The top two GBX2 DNA-binding consensus motifs bioinformatically determined by Motif Sampler analysis of GBX2 ChIP-Seq target sequence fragments.
Table 1.
ChIP-Seq identified GBX2 targets.
Figure 3.
Confirmation of murine GBX2 by Western blot and mass spectral analysis.
(A) Western analysis of recombinant GBX2 proteins. The amino acid sequences for GBX2 (B) and GBX2ΔHD (B') recombinant proteins. Bold type indicates matched peptides identified by mass spectrometry. (C) MS/MS fragmentation data table includes: precursor mass (peptide chosen for MS/MS), approximate weight of the band analyzed by mass spectrometry, peptide sequence, location of the peptide in the protein sequence, the Mascot ion score, and the mass error for each peptide sequence. The low mass error score in parts per million (ppm = {[observed mass – theoretical mass]/theoretical mass}×106) suggests that the observed mass matches the theoretical mass.
Figure 4.
Binding to ROBO1 and regulation of NC cell patterning by GBX2.
(A) Gel-shift analysis for identified GBX2 target ROBO1. A reduction in the mobility of [ÿ -32P] ATP labeled ROBO1 100-mer probe is observed with the addition of GBX2 (black arrows), whereas no shift is observed with the addition of GBX2ΔHD (compare lane 2 to lane 4). A supershift is observed in lane 3 with the addition of anti-GBX2. Addition of identical ROBO1 100-mer unlabeled specific competitor probe at 100x, 300x, and 500x molar concentrations in lanes 5–7. Addition of ROBO1 45-mer unlabeled non-specific competitor probe, omitting the GBX2 DNA-binding sequence in lanes 8–10. (B–I) Whole mount in situ hybridization for Robo1 (B–E) and the migrating neural crest marker, Sox10 (F–I) at gestational stage E9.5, demonstrates abnormal expression in Gbx2−/− mutants. Image analysis of embryos in a right lateral view (B,C) and dorsal view (D,E) reveals a reduction in expression of Robo1 in rhombomere 4 (compare white arrows in B and D to C and E) and disorganized expression in the rhombomere 1 domain (compare black arrows in B and D to C and E) in Gbx2–/– mutants compared to the WT control. Lateral and dorsal views reveal a reduction in Sox10 expression within the otic vesicle (compare F and H to G and I). Two distinct streams of NCCs into pharyngeal arch 1 and pharyngeal arch 2 are defined within control embryos (F, H) whereas in Gbx2−/− mutants (G,I) the NC streams appear disrupted. In the mutant, expression of Sox10 within the NC stream into presumptive pharyngeal arch 1 is significantly downregulated (compare black arrows in F and H to G and I), and the NCCs within the pharyngeal arch 2 appear to be compacted more posteriorly, truncating the stream (compare white arrows in F and H to G and I) when compared to the WT control. ov = otic vesicle.
Figure 5.
GBX2 directly targets multiple genes associated with Usher syndrome and inner ear development.
(A) Reverse transcription (RT)-PCR analysis of Gbx2 and identified targets, Pcdh15, Ush2a, and Notch2, in E13.5 wild-type mouse cochlear or vestibular inner ear tissues. Myo15 positive control expression is observed in cochlear and vestibular tissues. (B,C,D) Gel-shift analysis for identified GBX2 targets USH2A, PCDH15, and NOTCH2. A reduction in the mobility of [ÿ -32P] ATP labeled USH2A, PCDH15, and NOTCH2 100-mer probes is observed with the addition of GBX2 (black arrows), whereas no shift is observed with the addition of GBX2ΔHD (compare lane 2 to lane 4). A supershift is observed in lane 3 with the addition of anti-GBX2. Addition of identical USH2A, PCDH15, and NOTCH2 100-mer unlabeled specific competitor probes at 100x, 300x, and 500x molar concentrations in lanes 5–7. Addition of USH2A, PCDH15, and NOTCH2 45-mer unlabeled non-specific competitor probe, omitting the GBX2 DNA-binding sequence in lanes 8–10.
Figure 6.
GBX2 binds to and functionally interacts within the EEF1A1 core promoter.
(A) EEF1A1 locus depicting non-coding exons (white boxes), coding exons (black boxes), and the ChIP-Seq identified location of the GBX2 DNA-binding sequence (red bar). Alignment of the human and mouse ChIP-Seq identified EEF1A1 promoter region using sequences obtained from Ensembl [61], EEF1A1 TATA box (underlined sequence) and GBX2 DNA-binding sequence (red box). (B) Gel-shift analysis for identified GBX2 target EEF1A1. A reduction in the mobility of [ÿ -32P] ATP labeled EEF1A1 100-mer probe is observed with the addition of GBX2 (black arrows). A supershift is observed in lane 3 with the addition of anti-GBX2. Addition of identical EEF1A1 100-mer unlabeled specific competitor probes at 100x, 300x, and 500x molar concentrations in lanes 4–6. Addition of EEF1A1 45-mer unlabeled non-specific competitor probes, omitting the GBX2 DNA-binding sequence in lanes 7–9. (C) EEF1A1 promoter luciferase reporter assay. HEK 293 cells were transiently transfected with either the empty pGL4.10[luc2] vector (white bar), the pGL4.10[luc2] vector containing the functional EEF1A1 promoter sequence and the TATA box (TATATAA; black bars), the pGL4.10[luc2] vector containing the mutated EEF1A1 promoter sequence with a mutated TATA box (TATATAA changed to GCGCGCC; striped bars), and the pGL4.70[hRluc] Renilla vector. Substantial luciferase activity was observed in cells transfected with the pGL4.10[luc2] vector containing the functional EEF1A1 promoter compared to the empty pGL4.10[luc2] reporter construct (compare lane 3 to lane 1). Maximal luciferase activation is observed upon the addition of GBX2 in cells with the pGL4.10[luc2] vector containing the functional EEF1A1 promoter sequence (compare lane 4 to lane 3), and activation is reduced in GBX2ΔHD (lane 6) cells and cells expressing GBX2 and the pGL4.10[luc2] mutated EEF1A1 reporter construct (compare lane 4 to lane 5). Luciferase activities were normalized to Renilla luciferase activities. * P = 0.0047 (two-tailed P value).