|Home | About | Journals | Submit | Contact Us | Français|
Genetic variation in the transcription factor Interferon Regulatory Factor 6 (IRF6) causes and contributes risk for oral clefting disorders. We hypothesized that genes regulated by IRF6 are also involved in oral clefting disorders. We used five criteria to identify potential IRF6 target genes; differential gene expression in skin taken from wild type and Irf6-deficient murine embryos, localization to the Van der Woude syndrome 2 (VWS2) locus at 1p36–1p32, overlapping expression with Irf6, presence of a conserved predicted binding site in the promoter region, and a mutant murine phenotype that was similar to the Irf6 mutant mouse. Previously, we observed altered expression for 573 genes; 13 were located in the murine region syntenic to the VWS2 locus. Two of these genes, Wdr65 and Stratifin, met four of five criteria. Wdr65 was a novel gene that encoded a predicted protein of 1250 amino acids with two WD domains. As potential targets for Irf6 regulation, we hypothesized that disease-causing mutations will be found in WDR65 and Stratifin in individuals with VWS or VWS-like syndromes. We identified a potentially etiologic missense mutation in WDR65 in a person with VWS who does not have an exonic mutation in IRF6. The expression and mutation data were consistent with the hypothesis that WDR65 was a novel gene involved in oral clefting.
Interferon regulatory factor 6 (IRF6) is a member of the Interferon regulatory factor family of transcription factors. IRF6 shares the highly-conserved helix-turn-helix penta-tryptophan DNA binding domain and the interferon associated protein interaction domain, suggesting it functions as a transcription factor [Kondo et al., 2002]. Mutations in the IRF6 gene are responsible for two autosomal dominant syndromic forms of cleft lip with or without cleft palate, Van der Woude and popliteal pterygium syndromes [Kondo et al., 2002]. Additionally, a study performed on 11 geographically distinct populations showed an association between the V274I polymorphism in IRF6 and isolated cleft lip and palate [Zucchero et al., 2004], and subsequent work identified a SNP in the binding site for the TFAP2A transcription factor that altered binding [Rahimov et al., 2008]. This body of work demonstrates that IRF6 is an integral component for development of the lip and palate. Other genes and loci have been shown to play a role in human clefting using candidate gene [Jugessur et al., 2009] and genome wide strategies [Birnbaum et al., 2009]. We hypothesize that by studying potential regulatory targets of IRF6 we will identify new genes involved in development of the lip and palate. In this study, we used a genomic approach to find genes that might be regulated by IRF6. Based on our observations, we hypothesize that WDR65 and Stratifin (SFN or 14-3-3 sigma) are necessary for human craniofacial development, and mutations in these genes may be found in families affected with VWS and VWS-related disorders.
The VWS2 locus was mapped between D1S2697 and D1S230 [Koillinen et al., 2001]. We used the UCSC genome browser (build hg18) to determine the probable region of synteny in the murine genome, and used the Ensembl genome browser to obtain gene content and sequence. We used Matinspector to identify potential IRF binding sites in the 1kb region upstream of the transcription start site of the VWS2 candidate genes. To enrich for functional IRF binding sites, only putative sites conserved in rat, mouse, and human were used (VISTA Browser).
Microarray expression data (GDS2359) [Ingraham et al., 2006] was verified with SYBR Green real-time PCR analysis using primer pairs designed with PrimerExpress (Applied Biosystems, Foster City, CA) and amplified using Power SYBR Green PCR Master Mix (Applied Biosystems) with 1ng of cDNA derived from 17.5 dpc wild type and Irf6 mutant murine skin RNA [Ingraham et al., 2006]. Standard curves were performed as recommended, and a Rpl4 primer set was used as the internal control. Sequence for all primers used in this study are listed in Supplemental Table I.
Wdr65 was PCR amplified from a testis cDNA library (BD Biosciences) using the forward primer MF2TOPO and the reverse primer MR23 and cloned into pENTR/D-TOPO cloning vector (Invitrogen, Carlsbad, CA). 5’ RACE was performed as recommended (Ambion, Foster City, CA). These products were sequenced on both the forward and the reverse strands at the University of Iowa DNA Core Facility. The human ortholog of Wdr65 was found utilizing tBLASTn and tBLASTx programs. Only one protein was highly similar over the entire sequence. The human and murine protein sequences were aligned using Tcoffee and Boxshade. Exons of murine Wdr65 were aligned with human genomic sequence to determine the exon boundaries of WDR65 using Sequencher 4.2 (GeneCodes, Ann Arbor, MI).
In-situ hybridization was performed on coronal sections of murine embryonic heads. The processing of the sections and probes were performed as previously described [Goudy et al., 2010]. The probes for Wdr65 were made from the Riken clone 1110020C03 (MGC Image#1478343) that contained 630 bases of exons 20–23. For the antisense probe, the plasmid was linearized with SacI and labeled with T3 polymerase. For the sense probe, the plasmid was linearized with SalI, and labeled with T7 polymerase. For semi-quantitative PCR experiments, cDNA was obtained from the following sources; multi-tissue panels (BD Bioscience), human (Stratagene) and murine skin (Origene, Rockville, MD), C57Bl/6 embryonic stem cells, and from murine palate obtained by Laser Capture Microdissection (Pixel II system, Molecular Devices, Sunnyvale, CA). For microdissection, embryos at different developmental stages were embedded in OTC and frozen in liquid nitrogen. Tissue sections (7 µm) were dehydrated using serial ethanol concentrations, followed by xylene. The desired cells were captured, and RNA was extracted using a TRIzol solution, followed by RNeasy and DNase I cleanup, and cDNA was made as recommended (Ambion).
We obtained written informed consent from all subjects and approval for all protocols from the Institutional Review Boards at the University of Iowa, the University of São Paulo State and CONEP/Brazil, and the Karolinska Institutet. Subjects were examined by a clinical geneticist or by a genetic counselor, and diagnoses were made as described previously for VWS [Kondo et al., 2002], VWS2 [Koillinen et al., 2001] and BPS [Shanske et al., 2004]. Sample collection and processing and DNA sequence analysis were performed as previously described [Kondo et al., 2002].
Previous linkage studies of VWS pedigrees suggest little evidence for genetic heterogeneity for VWS. However, one VWS-like pedigree from Finland was found not to be linked to the IRF6 region. Rather, linkage was observed at the interval 1p36–1p32, and the authors named this locus VWS2 [Koillinen et al., 2001]. As one criterion for a regulatory target for IRF6, we hypothesized that the VWS2 locus contains such a gene. The VWS2 locus was mapped between the markers D1S2697 and D1S230 [Koillinen et al., 2001], near the genes CLCNKB and INADL, respectively. The murine orthologs of these two genes were located at 140,676,436 and at 97,887,844 on murine chromosome 4. In this region, the 694 genes and their order were conserved, confirming a syntenic relationship (Fig. 1a).
As a second criterion for genes regulated by Irf6, we hypothesized that their expression would be altered in tissues that lack Irf6. We took advantage of our previous microarray experiments that identified 573 differentially expressed genes between wild type and Irf6-deficient murine embryos at 17.5 days post conception (dpc) [Ingraham et al., 2006]. Of these genes, 13 were located in the syntenic region of the VWS2 locus (Table I). The microarray data (Fig. 1b) were verified by real-time PCR for Irf6, Wdr65 and Sfn (Fig. 1c). Two of the 13 genes were omitted because the probes mapped to an intron of one gene (Hdac1) or because the amount of mRNA in normal skin was deemed insignificant in a second gene (Sh2d5). Thus, we limited further studies to the remaining 11 candidate genes.
We hypothesized that target genes of Irf6 will show spatial and temporal expression patterns similar to Irf6. Irf6 showed tissue-specific expression, and of the eleven candidate genes, only Wdr65 had a distinct tissue-specific pattern (Fig. 2). We also compared temporal expression patterns in whole murine embryos. Irf6 expression was first seen in embryonic stem cells, disappeared, reappeared at 7 dpc embryos and peaks in 17 dpc embryos [Kondo et al., 2002]. Only two candidate genes, Wdr65 and Paqr7, showed similar changes in temporal expression during embryonic development (Fig 2). As a final comparison, we measured gene expression during development of the palate. For this analysis, cDNA was generated from whole palates before fusion (12.5–13.5 dpc), at fusion (14.5 dpc), after fusion (15.5 dpc), and from the medial edge epithelium (MEE). As expected from previous studies [Ingraham et al., 2006; Richardson et al., 2006], Irf6 was present in these tissues, except in the MEE isolated from Irf6-deficient embryos (Fig. 2). Most candidate genes showed constitutive expression in all palatal tissues, even in the absence of Irf6. Significantly, Wdr65 was present in the 14.5 dpc MEE, but was absent in MEE from Irf6-deficient embryos. In sum, these expression studies were most consistent with the hypothesis that Wdr65 is a target for regulation by Irf6.
IRF6 contains the highly conserved DNA binding domain of the IRF family, and a previous study showed that this domain binds to the consensus IRF binding site [Little et al., 2009]. In an effort to determine which of the eleven candidate genes could be direct targets of Irf6, we searched for putative IRF binding sites near each gene. A single conserved putative binding site was found for Wdr65 and Sfn (Table I). However, the binding site identified for Wdr65 was found in an exon of Ebna1bp2, a gene found 63 bp upstream of Wdr65 and transcribed in the opposite direction.
As a final criterion for potential Irf6 target genes, we compared the phenotypes of mutant murine strains to determine if the function of the candidate genes were consistent with craniofacial development. Of the eleven candidate genes, seven have been mutated in mice (Table 1). Of these, only Sfn mutant mice showed cleft palate and other craniofacial abnormalities [Herron et al., 2005].
Overall, Sfn and Wdr65 met four of the five criteria as candidates for regulation by IRF6 (Fig. 3). We then wished to screen these two genes for DNA variants that may cause VWS and VWS-like syndromes. However, WDR65 was a novel gene whose complete structure needed to be determined. When viewed in the ENSEMBL genomic database, we observed a predicted gene, ENSMUST00000081921 upstream from the Riken clone 1110020C03. We hypothesized that the predicted gene represented the 5’ end of the Riken clone. Using a PCR primer designed from each of these clones, MF2TOPO and MR23 (Fig. 4a), we amplified a 4 kb product from a murine testis cDNA library. Using 5’ RACE we amplified a 238 bp product. When sequenced, these products generated a complete cDNA whose length was 3958 bp. When aligned with the murine genomic sequence, Wdr65 has 23 exons.
To determine the human ortholog, we performed a BLAST search with Wdr65 and identified human cDNA clones, including NM_152489. This clone contained the first eleven exons, with 5’ and 3’ untranslated regions (UTRs) and a polyA tail (Fig. 4b). PCR experiments using human and murine cDNA, resulted in the identification of a shorter transcript, with only eleven exons, that we named WDR65b and Wdr65b, respectively. To clone the large human WDR65 transcript, we performed PCR experiments using primers designed to the most 5’ and 3’ human cDNA clones. We amplified a 4126 bp product that contains 24 exons from human testis cDNA. This larger human transcript is similar to the murine Wdr65 transcript, except it has an additional exon, exon 19, which adds 33 amino acids. In total, WDR65 human and murine genes produced two transcripts, WDR65b (11 exons), and a larger transcript, WDR65, with 24 exons in human and 23 exons in mouse. The long transcript was the major expressed isoform in both human and mouse (Supplemental Fig. 1). The human and murine WDR65 genes were predicted to encode proteins that are highly similar (Supplemental Fig. 2). And in each species, both the long and short isoforms were predicted to encode two WD-repeat functional domains (Supplemental Fig. 3). WD-repeat domains comprise a highly conserved motif defined by starting with GH (Gly-His) followed by 23–41 amino acid core and ending with WD (Trp-Asp).
To identify cell types that expressed Wdr65 in craniofacial structures, in situ experiments were performed on coronal sections from 14.5, and 15.5 dpc murine embryos. At 14.5 dpc the palates were juxtaposed and strong Wdr65 staining was seen in the nasal epithelium and detected in the medial edge epithelium (MEE) (Fig. 5a). In 15.5 dpc mice the expression was very strong in the nasal respiratory epithelium and epidermis, but was not seen in the oral epithelium (Fig. 5b). The expression of Wdr65 in the nasal epithelium was limited to the respiratory epithelium and not the olfactory epithelium. Significantly, we observed that the strong oral and nasal respiratory epithelia expression seen in the wild type (Fig. 5c) was markedly decreased or absent in the Irf6 mutant sections (Fig. 5d). These data are consistent with the hypothesis that Wdr65 is a candidate target of Irf6 and is strongly, though not completely, dependent on Irf6 for proper expression.
In the VWS2-linked pedigree, we sequenced the entire SFN gene and all exons for WDR65. SFN has a single exon, and the putative binding site for IRF6 was located in a multi-species conserved region located 900 bp upstream of the transcriptional start site. The only DNA variant observed in these two regions was rs3065004, a highly polymorphic microsatellite repeat located in the 3’UTR of SFN. For WDR65, we observed that the sequence was heterozygous for two DNA variants, rs663824 (non-synonymous, Asn241Asp) and rs603123 (synonymous). These DNA variants were not etiologic because they were observed in samples from unaffected individuals.
For this study, we sequenced samples from 48 individuals with VWS that lacked a mutation in any exon for the 5’UTR and coding region of IRF6 [Kondo et al., 2002; de Lima et al., 2009]. To test whether mutations in SFN and WDR65 contributed to VWS, we sequenced SFN as above, but limited our analysis of WDR65 to the exons coding for the WD repeat domain (exons 7 through 10). These exons were chosen because a deleterious mutation would most likely be located in a functional domain [Braun et al., 2006]. We found a missense mutation in exon 10 (c.1567G>T; p.Asp523Tyr) in one individual from Brazil with VWS (Fig. 6a). This residue is conserved through chicken (Fig. 6b). Since this mutation was not seen in controls and significantly changes the biochemical properties of a conserved residue, we conclude that it is potentially etiologic. To screen for other DNA variants, we performed DNA sequence analysis on all 24 exons in 32 DNA samples. We observed 21 additional DNA variants. Genotypes for all DNA variants were not significantly different from Hardy-Weinberg equilibrium, and all but one were detected previously in control populations (Supplemental Table II), suggesting that they are not etiologic. The one novel DNA variant was located in intron 5 and its function was unknown.
Bartsocas-Papas syndrome (BPS) is also referred to as lethal popliteal pterygium syndrome because the affected population has overlapping characteristics of PPS. But unlike PPS, BPS is autosomal recessive, and has a more severe phenotype that leads to early lethality. DNA from a patient with BPS was obtained [Shanske et al., 2004] and the whole gene for SFN and all the exons for WDR65 were sequenced. We observed no DNA variants in SFN. In WDR65, we observed eight rare DNA variants, seven were in introns and one was synonymous (Supplemental Table II). Two of these DNA variants were not found in other populations, but they were found in introns and their function was unknown. Overall, we cannot exclude WDR65 as the causative gene for this recessive disorder since no heterozygous DNA variants were observed in the patient who is from a consanguineous family.
The goal of this study was to identify new candidate genes for cleft lip and palate that are regulated by IRF6. IRF6, a gene critical for palatal development, encodes a putative transcription factor, and we hypothesized that gene targets would be potential cleft genes. From our genomic approach we identified two strong candidate targets for Irf6, Sfn, and the novel gene Wdr65. While mutation analyses at two syndromic clefting populations, VWS2 and BPS, failed to support a direct role for SFN and WDR65, a potentially etiologic missense mutation in WDR65 was discovered in a patient with VWS. Therefore, DNA variation in WDR65 may contribute to human oral clefting.
We identified and cloned WDR65, a gene that produced two transcripts of 11 and 24 exons. Both transcripts encoded two WD domains. Proteins containing WD-repeat domains are found most commonly in eukaryotic organisms. This is an extremely large family of proteins comprising over 100 human proteins with various functions [Smith et al., 1999]. Most WD-repeat proteins have 4–16 WD-repeats, and a minimum of four repeats were needed to form a proper tertiary structure, and seven produced the most stable structure [Li and Roberts, 2001]. Since WDR65 only has two predicted WD-repeats, it may not function in the traditional fashion of WD-repeat domain proteins.
The missense mutation found in a patient with VWS suggests that WDR65 plays a role in the fusion of lip and palate. The strongest expression of Wdr65 was seen in simple epithelium, specifically the ventral nasal respiratory epithelium, but not the nasal olfactory epithelium. Three other genes have a similar pattern of expression and have a role in palatal development, Gabrb3 [Hagiwara et al., 2003], Hoxa-2 [Nazarali et al., 2000] and Tbx1 [Goudy et al., 2010]. These genes all showed strong nasal epithelial expression during palatal fusion. The expression in the nasal epithelia suggests a role for this epithelium during the fusion of the palate to the nasal septum, which occurs at the same time as palatal fusion.
The increase in Wdr65 expression in whole palates and in 14.5 dpc MEE correlated with the development and dissolution of the epithelial seam. Previous studies suggest involvement in at least two potential pathways. First, several WD-repeat domain proteins function in programmed cell death, which could facilitate the dissolution of the medial edge seam in the MEE [Li and Roberts, 2001]. Second, WDR65 was identified recently as a candidate for a gene involved in cilial function [McClintock et al., 2008]. This hypothesis was based on the observation that expression of 1110020C03RIK was statistically overrepresented in tissues that contain highly ciliated cells, including olfactory epithelium, testis, vomeronasal organ, trachea and lung. However, immuno-staining shows direct staining of microvilli in bronchial epithelium, fallopian tube and epididymis mucosa (www.proteinatlas.org). It is interesting to speculate about a role for WDR65 in microvilli. During the adhesion and fusion of the primary palate, microvilli appeared, then disappeared and finally reappeared on the epithelial surfaces [Millicovsky and Johnston, 1981]. Further studies are needed to address the potential roles of Irf6 and WDR65 in apoptosis and changes in microvilli appearance during the development of the lip and palate.
The authors wish to acknowledge technical assistance from M. Malik, L. Henkle, K. Boyer, and R. Williamson. We would also like to thank Kevin Knudtson at the University of Iowa DNA Core Facility. This work was supported in part by the Predoctoral Training Program in Genetics 5 T32 GM08629 (N.K.R.), US National Institutes of Health grant DE16215 (J.C.M., B.C.S.), DE08559 (J.C.M.) and DE13513 (B.C.S., N.K.R.).
Ensembl genome browser: http://www.ensembl.org
UCSC Genome Browser: http://genome.ucsc.edu
VISTA browser: http://pipeline.lbl.gov