Introduction

Congenital microphthalmia is an ocular malformation characterized by a small eye (ocular anteroposterior axial length < 20 mm), with a reported prevalence at birth of 1. 5 per 10 000.1 Severity of the disease often ranges from mild to extreme (anophthalmia), even within a single family. Three transcription factor genes have been involved in isolated microphthalmia: CHX10, SOX2 and RAX.2, 3, 4, 5 Recently, a 24 bp deletion in the Sonic Hedgehog gene was found in a three-generation family.6 In addition, four other isolated microphthalmia genes have been mapped.7, 8, 9, 10 In a 5-generation family, we previously mapped the gene responsible for microphthalmia to a 13.8 cM region that was delimited by two intervals in close vicinity: a centromeric D15S1002–D15S1048 and a telomeric D15S1043–D15S1040 intervals.9 Here, we report a physical and partial transcript map of the 15q12–q15 chromosomal region and genetic refinement of the critical interval.

Methods

Physical and transcript map was constructed using BLAST homology search. We compared various sequences from the critical 15q12–q15 region, namely known microsatellite loci, STSs, gene sequences and transcripts, to genomic clones and contigs GenBank: http://www.ncbi.nlm.nih.gov/Genbank. In an attempt to identify candidate genes, we performed a BLASTx comparison between the sequences of the three contigs encompassing the critical interval and proteins including FGFs and various transcription factors containing either a homeodomain, a paired domain, a bHLH domain, a zinc finger or a leucine zipper domain. Each contig sequence was also compared to dbEST and TIGR databases with the aim of finding candidate transcripts dbEST: http://www.ncbi.nlm.nih.gov/dbEST; TIGR: http://www.tigr.org/tdb/.

In order to identify novel potentially polymorphic microsatellite DNA markers, we performed a BLASTn search using poly(AC or GT) sequences against the three overlapping contigs (Figure 1a). Family members were genotyped using the novel microsatellite markers (Figure 1a; Table 1). PCR products were size-fractionated by capillary electrophoresis (MegaBace, Amersham Pharmacia Biotech, France). Allele sizes were determined using the Genetic Profiler software.

Figure 1
figure 1

Genetic, physical and partial transcript map of the autosomal dominant colobomatous microphthalmia critical region on chromosome 15q12–q15. (a) Contigs and polymorphic markers. D15S markers are from the Genethon map; m: novel polymorphic poly(AC) markers; underlined loci represent boundaries of the critical intervals. (b) Genes. Candidate genes are indicated in relief and underlined. Potential genes encoding proteins with unknown function are represented by an asterisk. Trinucleotide repeats are indicated. (c) Retinal ESTs. Fetal retinal ESTs are underlined.

Table 1 Primers for PCR amplification of 20 novel informative polymorphic markers (m), 3 novel trinucleotide repeats and coding sequences of the CKTSF1B1, CX36 and KLF13 genes

Exons and flanking intronic DNA sequences of three candidate genes, namely CKTSF1B1, KLF13 and CX36, were PCR-amplified (Table 1) and sequenced in two affected individuals.We identified two CAG and one CGG repeats, mapping within the refined 2.5 Mb interval (Figure 1b). Analysis of these three repeats in two severely affected individuals from generation IV and in their parents and grandparents was performed using PCR-amplification (Table 1) followed by both electrophoresis on a 3% agarose gel and capillary electrophoresis.

Results

BLAST analysis and sequence annotation databases (Map View: http://www.ncbi.nlm.nih.gov/mapview/build34; UCSC Browser: http://genome.ucsc.edu/) made it possible to focus on three contigs encompassing the linked region (Figure 1). We analysed 34 novel potential polymorphic microsatellite markers, 20 of them were informative in the family studied (Figure 1a). On the basis of haplotype analysis (Figure 2), we refined the critical region, which was previously delimited by two intervals in close vicinity, namely a centromeric D15S1002–D15S1048 interval and a telomeric D15S1043–D15S1040 region. A new proximal limit for the centromeric interval was defined by a recombination event which occurred with marker m3 in individuals III:11 and IV:17 (Figure 2), whereas another recombination event in individual V:3 allowed us to place the distal limit at the marker m17 locus. The new centromeric interval m3–m17 contains large repeated regions, either duplicated or triplicated, thus preventing from using the polymorphic poly(AC or GT) markers mapping within this interval for accurate genotyping. Regarding the telomeric interval, individual V:3 is recombinant with marker m76. A recombination event between loci D15S1007 and m24 occurred in individual IV:6. The new telomeric interval is flanked by m76 and m24 markers. Thus, the refined critical region is composed of a 1.9 Mb centromeric m3–m17 interval and a 2.5 Mb telomeric m76m24 interval.

Figure 2
figure 2

Pedigree of the colobomatous microphthalmia family. Blackened symbols indicate affected individuals; hatched symbols represent individuals with uncertain status; the disease-associated haplotypes are boxed.

A total of 57 known genes were located within the three contigs overlapping the critical interval (Figure 1b). Of these, three genes, namely CKTSF1B1, CX36 and KLF13 encode proteins with a possible role in eye development. However, no change in the coding sequence of these genes could be identified in affected individuals. Morever, subsequent physical mapping data (Map View: http://www.ncbi.nlm.nih.gov/mapview/build34; UCSC Browser: http://genome.ucsc.edu/) indicated that the CX36 gene maps telomeric to the m76m24 interval.

Analyses of three trinucleotide repeats in affected individuals and healthy controls did not reveal any modification of the migration pattern, ruling out an abnormal expansion of these repeats as the molecular basis of the disease in this family.

Finally, BLASTn search for homologies using the three candidate contigs against dbEST and TIGR (dbEST: http://www.ncbi.nlm.nih.gov/dbEST; TIGR: http://www.tigr.org/tdb/) databases identified 9 fetal retinal ESTs and 28 retinal ESTs (Figure 1c).

Discussion

A linkage analysis in a large autosomal dominant colobomatous microphthalmia family using 20 novel polymorphic microsatellite DNA markers allowed us to refine the region 15q12–q15 including the disease-causing gene to two intervals in close vicinity, namely a centromeric interval, flanked by microsatellite DNA markers m3–m17, and a telomeric interval, m76m24, encompassing respectively 1.9 and 2.5 Mb. We set up a detailed physical and transcript map, which pointed out three candidate genes: CKTSF1B1, KLF13 and CX36, encoding proteins which may be involved in eye development.

First, CKTSF1B1 maps within the refined interval and its transcripts are expressed at a high level in adult and fetal brains.11 It encodes Gremlin, a protein which belongs to the family of bone morphogenetic protein (BMP) regulators. A targeted inactivation of murine Bmp7 leads to microphthalmia or anophthalmia.12 In chick embryos, disruption of the BMP signalling by intraocular overexpression of a BMP antagonist results in microphthalmia.13 In mouse Gli3−/− mutants, gremlin overexpression causes an optic cup reduction.14 Thus, gremlin interacts with GLI3, a member of the SHH pathway that has recently been shown to be essential for eye morphogenesis.6 Second, the KLF13 gene codes for a zinc-finger transcription factor, expressed in mouse fetal eye (http://odin.mdacc.tmc.edu/RetinalExpress). Finally, the CX36 gene, which encodes connexin36 belongs to the connexin family, including Cx50, a gene whose targeted invalidation in the mouse results in microphthalmia.15 However, complete sequencing of CKTSF1B1, KLF13 and CX36 exons and flanking intronic sequences led us to exclude these three candidate genes as the causative gene in this family.

To date, the 4.4 Mb refined interval includes 21 other genes encoding proteins that have no known role in eye development. So, they should be considered as candidate by their position only. A total of 11 retinal ESTs were detected within the critical interval: three correspond to hypothetical proteins and the others have no homology with known proteins. Noteworthy, no mouse mutant with microphthalmia has been mapped yet to the mouse chromosome 2 and 7 regions, which share homologies with human chromosome 15q12–q15.

In the five-generation family studied, the apparently increasing penetrance of the disease through successive generations may suggest anticipation, a feature known to result from trinucleotide repeat expansions. However, no abnormal expansion of the three trinucleotide repeats studied was found in this family. Thus, the apparent anticipation may be due to another yet unidentified trinucleotide repeat expansion or may simply be fortuitous.

Finally, the combined genetic and physical refinement of the disease locus to a 4.4 Mb region, the exclusion of three candidate genes (CKTSF1B1, KLF13 and CX36) and the identification of retinal ESTs, represent important steps in the attempt to identify the disease-causing gene itself.