Disruption of an AP-2alpha binding site in an IRF6 enhancer is associated with cleft lip

Abstract

Previously we have shown that nonsyndromic cleft lip with or without cleft palate (NSCL/P) is strongly associated with SNPs in IRF6 (interferon regulatory factor 6). Here, we use multispecies sequence comparisons to identify a common SNP (rs642961, G>A) in a newly identified IRF6 enhancer. The A allele is significantly overtransmitted (P = 1 x 10(-11)) in families with NSCL/P, in particular those with cleft lip but not cleft palate. Further, there is a dosage effect of the A allele, with a relative risk for cleft lip of 1.68 for the AG genotype and 2.40 for the AA genotype. EMSA and ChIP assays demonstrate that the risk allele disrupts the binding site of transcription factor AP-2alpha and expression analysis in the mouse localizes the enhancer activity to craniofacial and limb structures. Our findings place IRF6 and AP-2alpha in the same developmental pathway and identify a high-frequency variant in a regulatory element contributing substantially to a common, complex disorder.

Figure 1

Transcription factor AP-2α binds to MCS-9.7 and rs642961 disrupts its binding site. (a) Multispecies sequence alignment of the MCS-9.7 segment corresponding to nucleotides from –14470 to –14535, with respect to IRF6 translation initiation codon, that contains variants -14474A>G, rs642961 and -14523G>A and the predicted AP-2α binding site (highlighted in yellow). (b) The double-stranded IRDye-700 labeled oligonucleotide probes used in EMSA. The probes correspond to MCS-9.7 nucleotides –14495 to –14522 and differ only at the rs642961 SNP. (c) EMSA using probe alone (lane 1) and incubated with the human recombinant AP-2α protein, allele G (lane 2) and allele A (lane 3). Competition with increasing amounts of unlabeled probes with allele G (lanes 4–6) and allele A (lanes 7–9). Supershift is formed when AP-2α is pre-incubated with anti-AP-2α antibody (AP-2α-Ab) (lane 10). (d) Quantification of ChIP by real-time PCR. AP-2α-Ab and IgG immunoprecipitated chromatin fragments were amplified using primers flanking the MCS-9.7 segment with four putative AP-2α binding sites shown in Supplementary Figure 1 and a control region of a similar length devoid of predicted AP-2α binding sites. The curves represent the accumulation of PCR products given as log values (logΔR) of SYBR-Green fluorescence intensity plotted against the number of cycles. MCS-9.7 amplification plot of the immunoprecipitated chromatin from uninfected (Plot I) and ad-AP-2α-infected (Plot III) HaCaT cells. Amplification curves of the control region are shown on Plot II for the uninfected cells and on Plot IV for the infected cells.

Figure 2

MCS-9.7 shows IRF6 enhancer activity in transgenic mouse assay. (a-a”) Lateral views of three independent transgenic embryos at embryonic day 11.5 (E11.5) expressing LacZ directed by MCS-9.7. (b) Frontal view and (c) expanded view of the orofacial region of the embryo shown in panel a”. White arrow points to LacZ expression at the fusion sites between the lateral nasal (ln), medial nasal (mn) and maxillary (mx) prominences towards the end of upper lip formation at E11.5. (d) Sagittal virtual section through the orofacial region of the embryo shown in panel a” generated with optical projection tomography. Shown in red (white arrow) is LacZ expression in ectoderm covering fusing facial prominences (ln, mn and mx). See 3D views of embryos a-a″ in Supplementary videos online