Embryonic expression profile of chicken CHD7, the ortholog of the causative gene for CHARGE syndrome

Authors


Abstract

BACKGROUND: CHARGE syndrome represents a constellation of malformations: C, coloboma of the iris or retina; H, heart defects; A, atresia of the choanae; R, retardation of growth and/or development; G, genital anomalies; and E, ear abnormalities. Recently, the Chromodomain helicase DNA-binding protein-7 (CHD7) at chromosome 8q12.1 was identified as a causative gene for CHARGE syndrome. Because CHD7 was identified as a causative gene using a positional cloning approach, the role of CHD7 in early embryogenesis needs to be further investigated. METHODS: Fertilized chick eggs were incubated to Hamburger and Hamilton stages 4–20 and were studied using whole mount in situ hybridization. Chicken EST clones corresponding to the chicken CHD7 (cChd7) sequence were identified using the chicken EST sequence database. From the EST clones, a digoxigenin-labeled RNA probe was synthesized and hybridized in situ to the embryonic specimens. RESULTS: The expression of cChd7 was pan-neuronal at stages 8–20. It was expressed throughout the rostral neural ectoderm and along the rostrocaudal axis but was absent from the more lateral, non-neuronal ectoderm. Adjacent to the neural tube, cChd7 transcripts were detected at the optic and otic placodes. At stage 20, cChd7 expression was observed in the branchial arches and olfactory placodes in addition to brain and optic and otic placodes. CONCLUSIONS:cChd7 was expressed in the neural epithelium, the otic placodes, the optic placodes, the branchial arches, and the olfactory placodes, which are the primordial tissues that give rise to organs affected in human CHARGE syndrome patients. Birth Defects Research (Part A), 2007. © 2006 Wiley-Liss, Inc.

INTRODUCTION

Hall and Hittner et al. independently documented that choanal atresia and ocular coloboma can be associated with a specific pattern of malformations (Hall, 1979; Hittner et al., 1979). Pagon et al. (1981) coined the term “CHARGE” to represent a constellation of nonrandomly associated malformations: C, coloboma of the iris or retina; H, heart defects; A, atresia of the choanae; R, retardation of growth and/or development; G, genital anomalies; and E, ear abnormalities. Since then, several hundred patients with CHARGE syndrome have been described, and the prevalence of the condition is estimated to be as high as 1/10,000 (Blake et al., 1998; Graham, 2001).

Vissers et al. (2004) identified the gene Chromodomain helicase DNA-binding protein-7 (CHD7) at chromosome 8q12.1 as a causative gene for CHARGE syndrome: 10 out of 17 (59%) of their patients had heterozygous mutations in CHD7. They documented that human CHD7 is ubiquitously expressed in several fetal and adult tissues (Vissers et al., 2004). Because CHD7 was identified as a causative gene using a positional cloning approach, the role of CHD7 in early embryogenesis needs to be further investigated.

We recently performed a genotype-phenotype correlation study among 17 CHARGE syndrome patients with CHD7 mutations and found that ocular coloboma (iridis, retina, or optic nerve) was almost universal (88%) and that hearing loss and developmental delays were universal (100%) (Aramaki et al., 2006). Examining whether the primordia of organs universally affected in CHARGE syndrome patients (i.e., eyes, ears, and brain) express CHD7 in normal embryos would provide important information.

In the present study, we investigated the expression pattern of the CHD7 ortholog in the early embryos of chicken, a model organism that lends itself to experimental manipulation and is rich in genome information (Burt, 2004; International Chicken Genome Sequencing Consortium, 2004).

MATERIALS AND METHODS

Identification of the Chicken Chd7 cDNA Sequence and Corresponding Clone

The putative chicken Chd7 cDNA sequence was obtained by comparing the human CHD7 protein sequence and the chicken genome sequence using the TBLASTN program of the BLAST software package (http://www.ncbi.nlm.nih.gov/BLAST/) (Altschul et al., 1990). The computationally predicted chicken Chd7 cDNA sequence (XM_419222) was queried against the chicken EST database at the Biotechnology and Biological Sciences Research Council (http://www.chick.umist.ac.uk/) (Boardman et al., 2002) in order to identify the expressed sequence tag (EST) clones corresponding to XM_419222.

To confirm the validity of the computationally predicted chicken Chd7 cDNA sequence XM_419222, we performed RT-PCR experiments using PCR primers located within the EST sequences. The primer sequences and PCR conditions are available upon request. Total RNA was extracted from 10 Hamburger and Hamilton (HH) stage 10 whole embryos (Hamburger, 1951). cDNA was directly synthesized using the SuperScript III amplification kit (Invitrogen, Carlsbad, CA) with an oligo-dT primer. The resulting RT-PCR product was directly sequenced from both directions using an ABI3100 autosequencer (ABI, Foster City, CA). The experimentally confirmed cDNA sequence was deposited in the GenBank database (http://www.ncbi.nlm.nih.gov/entrez/) as DQ978381 (cDNA sequence) and ABI96999 (corresponding protein sequence).

Then, a data set consisting of known chromodomain helicase DNA-binding proteins among the human, mouse, and chicken species and the ABI96999 protein sequence were used to construct an evolutionary tree. The GenBank accession numbers of those proteins are as follows: NP_001261 human CHD1, AAI15823 mouse Chd1, NP_990272 chicken Chd1, NP_001262 human CHD2, NP_001005273 human CHD3, NP_001264 human CHD4, NP_666091 mouse Chd4, NP_056372 human CHD5, NP_115597 human CHD6, NP_775544 mouse Chd6, NP_060250 human CHD7, XP_910521 mouse Chd7, NP_065971 human CHD8, NP_796198 mouse Chd9, and NP_079410 human CHD9. The sequence of the sucrose-nonfermenting 2 (SNF2) family N-terminal domain was excised from each protein and multiple alignments were created with ClustalW (Thompson et al., 1994). Sequences were analyzed by the Molecular Evolutionary Genetics Analysis (MEGA) version 3 (Kumar et al., 2004) with a neighbor-joining (N-J) algorithm and p-distance method.

The degree of identity and similarity between the domains and interdomain regions of the human CHD7 protein and the chicken ABI96999 protein were calculated using the “stretcher” function of the EMBOSS software package (Rice et al., 2000).

The chicken EST clones corresponding to DQ978381 cDNA sequence were obtained from the Medical Research Council's Geneservice (http://www.hgmp.mrc.ac.uk/geneservice/reagents/products/descriptions/chickenEST.shtml). To confirm probe specificity, we performed RT-PCR experiments using PCR primers located at the terminus of the in situ probe using cDNA of HH stage 10 whole embryos as synthesized above. RT-PCR analysis was performed with the primer pairs ChEST757_AntisenseF (5′-GCATCTGTCACCGAGCTACA-3′) and ChEST757_AntisenseR (5′- TTGACGGTTGAAGGTTACCA -3′, ChEST757_SenseF (5′-TCGTAGACAGTGGCTCGATG-3′) and ChEST757_SenseR (5′-AACTGCCAACTTCCAATGGT-3′), GAPDH_F (5′-CAGGTGCTGAGTATGTTGTGGAGTC-3′) and GAPDH_R (5′-TCTTCTGTGTGGCTGTGATGGC-3′). The resulting RT-PCR product was directly sequenced from both directions using an ABI3100 autosequencer (ABI).

Preparation of the RNA Probe

A DIG-labeled RNA probe for in situ hybridization was synthesized from a PCR-amplified template containing an RNA promoter according to the one-tube preparation method described by Ogasawara et al. (2001). The EST clones from MRC Geneservice had been directionally inserted into the EcoRI site of a pBluescript II KS+ plasmid vector (Stratagene, La Jolla, CA). The antisense RNA probe was synthesized from a T3 promoter. The sense probe was synthesized from a T7 promoter. To obtain the linearized template DNA containing the whole cDNA sequence and promoter for T3 or T7 RNA polymerase, we used a set of PCR primers: pBKS R-primer (ACAGCTATGACCATGATTAC), located at a downstream multicloning site, and pBKS F-primer (GCGTAATACGACTCACTATA), located on the T7 promoter flanking the poly(A) tail of the cDNA. Ten nanograms of plasmid DNA for each clone was amplified in 10 μL of PCR mixture under three-step PCR conditions (15 cycles at 94°C for 1 min, 50°C for 2 min, and 72°C for 3 min). Subsequently, 7.5 μL of additional DIG mixture containing the DIG RNA labeling mixture (Roche, Basel, Switzerland), RNase inhibitor, and T3 RNA polymerase (Invitrogen) or T7 RNA polymerase (Roche) was added and the mixture was incubated for 4 h at 37°C.

Preparation of the Chick Embryos

Fertilized chick (Gallus gallus) eggs were incubated at 37.8°C and 50% humidity to HH stages 4–20. Chick embryos were fixed in 4% paraformaldehyde in PBS at 4°C overnight. Fixed embryos were rehydrated in a graded series of ethanol and substituted in PBS with 0.1% Triton X-100 (PBST). The embryos were treated with proteinase K (2–10 μg/mL) at room temperature for 15–30 min and were postfixed in 4% paraformaldehyde and 0.2% glutaraldehyde in PBST at room temperature for 30 min.

Whole Mount In Situ Hybridization of Chick Embryo

The specimens were hybridized in situ using digoxigenin-labeled antisense probes. After washing with PBST three times, the specimens were rinsed in hybridization buffer (50% formamide, 5 × 3 times, 2% blocking reagent, 0.1% Triton X-100, 0.1% CHAPS, 1 mg/mL yeast torula RNA, 5 mM EDTA, 50 μg/mL heparin) at room temperature for 10 min, then in hybridization buffer at 65°C for 1 h as a prehybridization step. Hybridization was carried out in a hybridization mix containing 1 μg/mL of digoxigenin-labeled probe incubated at 70°C for 16 h. Nonspecific binding was blocked using 1.5% blocking reagent (Roche) in KTBT (10 mM KCl, 50 mM Tris-HCl, 150 mM NaCl, 0.1% Triton–X-100), and the embryos were incubated overnight with an alkaline phosphatase-conjugated antidigoxigenin antibody (Roche). Embryos were thoroughly washed in six changes of KTBT at room temperature. Following two washes in NTMT (0.1 M NaCl, 0.1 M Tris-HCl, 50 mM MgCl2, 0.1% Triton X-100), the color was developed using BM purple (Roche). BM purple is a chromogenic substrate for alkaline phosphatase designed for precipitating enzyme immunoassays. It develops a permanent, dark purple spot at the alkaline phosphatase binding site. At least three embryos were examined at each developmental stage. Images of stained embryos were acquired with a Fuji Digital Camera HC-2500 (Fujifilm, Tokyo, Japan) mounted on a Leica MZ9.5 high-performance stereomicroscope (Leica Microsystems, Wetzlar, Germany). Digitized images were adjusted for levels and color balance to reflect their true color using Adobe Photoshop software (Adobe, San Jose, CA). Whole images were treated uniformly so as to maintain the original data.

For the section analysis, stained embryos were immersed in a graded series of sucrose (10–30%) in PBS and were embedded in OCT compound (Tissue-Tek; Sakura Finetek, Torrance, CA). Serial frozen sections (12 μm) were cut and mounted on glass slides. Images of the sections were acquired using a DXM1200 digital camera (Nikon, Tokyo, Japan) mounted on a BX50 microscope (Olympus, Tokyo, Japan).

RESULTS

Identification of the Chicken Chd7 cDNA Clone

The computationally predicted chicken Chd7 cDNA sequence (XM_419222) that was deduced from NW_060290 matched with the 13 chicken EST sequences (ChEST931i1, ChEST437p13, ChEST771p24, ChEST738p21, ChEST295o7, ChEST938j11, ChEST901b1, ChEST908m6, ChEST287a1, ChEST675j20, ChEST322d21, ChEST936m8, ChEST757h23). We determined the complete coding sequence of XM_419222 by RT-PCR experiments using PCR primers located within the EST sequences and deposited it as DQ978381 (cDNA sequence) and ABI96999 (corresponding protein sequence) in the GenBank database (http://www.ncbi.nlm.nih.gov/entrez/). The protein ABI96999 was considered to be the ortholog of human CHD7 based on the following rationale: first, the flanking genes in human CHD7 gene and the putative chicken homologue {CYP7A1-RAB2-[CHD7]-CRH-EYA1 and Cyp7a1-Rab2-[ABI96999]-Crh-Eya1} are orthologous to each other, and their order on the chromosome is evolutionarily conserved between these two species (Fig. 1). Second, the ABI96999 protein sequence was located closest to human CHD7 and mouse Chd7 in the evolutionary tree consisting of all the known Chd homologue protein sequences among the human, mouse, and chicken species including ABI96999 (Fig. 2). Third, human CHD7 protein and ABI96999 protein had a similar domain structure and the sequence identity and similarity figures between the two protein sequences of each domain and interdomain region were consistently high (Fig. 3). Based on chromosomal synteny between the human CHD7 and ABI96999, closeness of the two proteins within the evolutionary tree of chromodomain proteins, and the high degree of sequence similarity and identity along the entire protein, we concluded that ABI96999 represents the ortholog of the human CHD7 protein. Hereafter, we refer to DQ978381 as the chicken Chd7 cDNA (cChd7) and ABI96999 as the chicken Chd7 protein (cChd7).

Figure 1.

Conserved order of genes flanking CHD7 on human chromosome 8 and ABI96999 on chicken chromosome 2. CYP7A1: cholesterol 7alpha-hydroxylase; RAB2: RAS-associated protein 2; CHD7: chromodomain helicase DNA-binding protein 7; CRH: corticotropin-releasing hormone; EYA1: eyes absent 1. ABI96999 is the putative chicken Chd7 gene. Note conserved order of orthologous genes between human and chicken genomes.

Figure 2.

Evolutionary tree among known human, mouse, and chicken CHD family proteins and ABI96999, the putative chicken Chd7 protein. The tree was constructed by the neighbor-joining method. The scale bar represents the number of substitutions per site.

Figure 3.

Sequence identity and similarity figures between the human CHD7 protein and ABI96999 (the putative chicken Chd7 protein) sequences of each domain and interdomain region. chromo: chromatin organization modifier (chromo) domain; SNF2: sucrose-nonfermenting 2 (SNF2) family N-terminal domain; helicase: helicase c-terminal domain; BRK: BRK domain. Note conserved domain structure between human CHD7 protein (top) and putative chicken Chd7 protein (bottom). The degree of identity and similarity between the domains and interdomain regions of human CHD7 protein and the putative chicken Chd7 protein are presented as percentages between the two proteins. The bar indicates the regions of cChd7 spanned by the antisense in situ probe.

EST clone ChEST757h23 (GenBank accession number BU233774.1), the sequence of which completely matched that of the amino-terminal portion of the putative cChd7 cDNA sequence, was obtained from the Medical Research Council's Geneservice and was used to generate the in situ probe (Fig. 3). The 3′-untranslated region, which is usually used as the in situ probe because of its high sequence specificity, was unavailable at the time of this study and we chose the amino-terminal portion instead. The sequence of ChEST757h23 was unique among the currently available Chd family mRNA sequences.

Specificity of the In Situ Probe

To confirm the specificity of the probe for in situ hybridization, we performed RT-PCR experiments using PCR primers located at the terminus of the in situ probe (Fig. 3) and template mRNA derived from HH stage 10 whole embryos. Sequencing of the RT-PCR product, which was of the expected size, revealed that the products represented part of cChd7 cDNA, demonstrating the specificity of the PCR primers (Fig. 4).

Figure 4.

Specificity of the in situ probe. RT-PCR experiments using PCR primers located at the terminus of the in situ probe. Total RNA was extracted from Hamburger and Hamilton stage 10 whole embryos. Lane 1: RT-PCR product amplified with ChEST757_antisense primers; lane 2: RT-PCR product with ChEST757_Sense primers; lane 3: RT-PCR product with primers to amplify a ubiquitously expressed gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH); lane 4: negative control. Note strong amplification of the specific product in lane 2.

Whole-Mount In Situ Hybridization Studies

Whole-mount chick embryos at HH stages 4–20 were analyzed for cChd7 expression using in situ hybridization with digoxigenin-labeled antisense RNA probes. Images obtained using the sense probe, which theoretically should give a negative result for mRNA, are presented together with images obtained using the antisense probe. No signals were obtained using the sense probes in any of the hybridization experiments.

Stage 4–7 embryos.

At HH stages 4–7, no cChd7 expression was detected (Fig. 5).

Figure 5.

Whole-mount in situ hybridization using antisense cChd7 digoxigenin RNA probe in Hamburger and Hamilton developmental stages (left: stage 4; middle: stage 7; right: stage 8). Images at the top were obtained with the antisense probes and those at the bottom were obtained with the sense-probes. Note initiation of cChd7 expression at stage 8.

Stage 8–11 embryos.

At HH stage 8, cChd7 expression was detected along the entire rostrocaudal axis of the neuroectoderm (Fig. 5). The expression was relatively weak in the most caudal regions and was absent from the more lateral, non-neuronal ectoderm.

Stage 12 and 13 embryos.

At stages 12 and 13, cChd7 expression was seen in the neural ectoderm and was uniformly expressed at high levels (Fig. 6A). Two paraxial crescent signals representing the dorsal halves of the otic placodes were identified at the hindbrain level (Fig. 6B). In transverse sections, cChd7 expression was observed in the neural ectoderm, optic vesicles, and otic pits (Fig. 7). In the rostral neural tube, the expression of cChd7 transcripts was identified in the dorsal roof plate (Fig. 7B,D), but in the caudal neural tube, expression was seen in the entire neural tube (Fig. 7F). Furthermore, higher magnification of the optic vesicles revealed that cChd7 transcripts were expressed in the neuroepithelium but not in the surface ectoderm (Fig. 7C). No signals were detected in the notochord or the heart primordium (Fig. 7D).

Figure 6.

Whole-mount in situ hybridization using cChd7 digoxigenin RNA probe in Hamburger and Hamilton developmental stage 12 (dorsal view). (A) Image obtained with the antisense probe; (B) a close-up view of (A). The arrows indicate the positive staining of the otic placode, appearing as paraxial crescent signals at the hindbrain level. (C) Image obtained with the sense probe.

Figure 7.

Transverse sections of a Hamburger and Hamilton stage 12 embryo. The stained embryo was sectioned at three levels (B, D, and F), indicated in (A) (whole-mount view). Magnified views of sections (B) and (D) are shown in (C) and (E), respectively. In (C), the neuroepithelium of the optic vesicle is shown on the right, and the surface ectoderm is shown on the left. fb: forebrain; op: optic vesicle; ne: neuroepithelium; se: surface ectoderm; hb: hindbrain; nt: neural tube; nc: notochord; ot: otic pit; ht: heart; s: somite.

Stage 14 and 15 embryos.

From stages 14 and 15, low levels of cChd7 expression were observed in the telencephalon, diencephalon, mesencephalon, metencephalon, and myelencephalon (Fig. 8). In addition, cChd7 was detected in the optic and otic primordium.

Figure 8.

Whole-mount in situ hybridization of Hamburger and Hamilton stage 14 embryos using antisense (A,C,E) and sense (B and D) probes. (E) Contralateral view of (C). (C) A close-up view of (A). (D) A close-up view of (B). The optic and the otic primordium are indicated by the arrow and arrowhead, respectively.

Stage 20 embryos.

Besides that in the brain and the optic placode including the lens vesicle (Fig. 9), cChd7 expression was also observed in the branchial arches and olfactory placodes. No signals were detected in the heart.

Figure 9.

Whole-mount in situ hybridization of Hamburger and Hamilton stage 20 embryos using antisense (A) and sense (B) probes (lateral view). The olfactory placode is indicated by the arrow.

DISCUSSION

In the present study, we investigated the expression pattern of the chicken CHD7 ortholog, cChd7, in early chick embryos at the gastrulation stage and the neurulation stage, hoping to correlate the expression pattern with the disease spectrum of CHARGE syndrome. Our expression study in normal chick embryos suggested that several, if not all, of the cardinal features of CHARGE syndrome could be accounted for by the reduced expression of CHD7.

First, the present study indicated that cChd7 was expressed throughout the neural epithelium during neurulation. Documentation of cChd7 expression in the forebrain in normal chick embryos agrees with the clinical observation that global developmental delay is a universal feature of CHARGE syndrome in humans (Aramaki et al., 2006; Vissers et al., 2004).

cChd7 expression in the otic placode in normal chick embryos agrees with the clinical observation that sensorineural hearing loss is a universal feature and that semicircular canal defects are a specific feature of CHD7 mutation-positive CHARGE patients (Aramaki et al., 2006; Vissers et al., 2004). Likewise, the documentation of cChd7 expression in the optic placode in normal chick embryos may account for the clinical observation that coloboma (iris, retinal, or optic disc coloboma) is an almost universal feature of CHARGE syndrome (Aramaki et al., 2006; Vissers et al., 2004).

Among the six cardinal features of CHARGE syndrome, C (coloboma of the iris or retina), R (retardation of growth and/or development), and E (ear abnormalities) can be accounted for by the cChd7 expression patterns during gastrulation and neurulation (stages 4 and 15). Furthermore, cChd7 expression in the branchial arches and olfactory placodes of pharyngula embryos (stage 20) may account for heart defects (H) and genital anomalies (G), respectively. In terms of cardiac morphogenesis, cChd7 expression was not observed during the gastrulation and neurulation periods, during which the precardiac mesoderm is formed (stage 4 and later), the endocardial tubes start to fuse (stages 8–11), and the looping process begins (stage 9 and later) (Martinsen, 2005). However, during the pharyngula period, cChd7 was expressed in the branchial arch, to which neural crest–derived cells are a major contributor. The documentation of cChd7 expression in the branchial arches of normal chick embryos agrees with the clinical observation that many of the cardiac defects in CHARGE syndrome patients originate as forms of neural crestopathy (Clark, 1996).

Documentation of cChd7 in the olfactory placode may shed light on the pathogenesis of genital anomalies in CHARGE syndrome. Chalouhi et al. (2005) recently demonstrated that CHARGE syndrome patients fulfill the criteria for Kallmann syndrome, which combines hypogonadotropic hypogonadism with various degrees of deficiency in the sense of smell as a result of olfactory bulb aplasia or hypoplasia. Because hypogonadism in Kallmann syndrome results from the failed embryonic migration of GnRH-synthesizing neurons along the olfactory nerve pathway (Schwanzel-Fukuda et al., 1989), they suggested that hypogonadotropic hypogonadism in CHARGE syndrome may result from the same embryonic defect (Chalouhi et al., 2005). The identification of cChd7 in the olfactory placode in the present study supports Chalouhi's hypothesis.

Vissers et al. (2004) reported “ubiquitous” expression in adult and fetal tissues in a study using RT-PCR. RT-PCR is the most useful tool for detecting extremely low quantities of RNA. However, accurate quantification is not possible using normal PCR techniques. As the obtained PCR products do not necessarily reflect the amount of DNA originally present in a sample, it is impossible to quantify RNA expression by estimating the amount of template DNA in proportion to the amount of PCR products (Bustin, 2002). Vissers' data indicated that a low level of CHD7 expression was present in various tissues, but their data do not necessarily indicate that the CHD7 expression levels in various tissues are of equal magnitudes. Hence, we do not think that their data conflict with our data.

CHD7 mutations account for only ∼60–65% of CHARGE cases, and there may be other causative mutations that lead to the CHARGE syndrome phenotype. Indeed, one patient with CHARGE syndrome has been reported to have a mutation in SEMA3E (Lalani et al., 2004), a gene other than CHD7. A comparison of the expression patterns of other candidate genes and those of CHD7 could help to identify additional causative genes of CHARGE syndrome.

In conclusion, cChd7 was expressed at the neural epithelium, the otic placodes, and the optic placodes, which are the primordial tissues giving rise to organs affected in CHARGE syndrome patients.

Acknowledgements

We thank Mr. H. Suzuki for technical support and Ms. K. Shinohara for secretarial assistance.

Ancillary