Editor—Retinitis pigmentosa (RP, MIM 268000) is the most frequent form of retinal dystrophy world wide. The clinical findings are night blindness and narrowing of the visual field. Examination of the fundus of the eye in RP patients usually shows bone spicula pigmentation of the retina, waxy pallor of the optic disc, attenuation of the retinal blood vessels, and no results detectable by electroretinogram.1

RP shows notable allelic and non-allelic heterogeneity2(RET-GEN-NET htp://www.sph.uth.tm.edu/Retnet/home.htm). By using classical linkage strategies and the direct and indirect candidate gene approach, the number of RP loci identified has grown increasingly since 1989 and to date more than 30 autosomal RP loci have been identified, including syndromic and non-syndromic forms of the disease. Autosomal recessive RP (ARRP) is the commonest form of RP and to date at least 13 independent ARRP loci have been identified.3-15

Our group proposed the hypothesis that the alteration of functions related to neurotransmission in the external plexiform layer of the retina could be related to RP.14 In order to test this model, we used homozygosity mapping to analyse different genes involved in retinal neurotransmission. Using this indirect candidate gene approach, we identified the locus RP25 in an important subgroup of ARRP patients from our cohort. In fact, around 14% of the ARRP families from southern Spain showed linkage toRP25. 14 RP25 is an ARRP locus located on the long arm of chromosome 6 between markers D6S257 and D6S1644 (MIM 602772). This chromosomal region contains the GABRR1and GABRR2 genes, both being expressed in the retina. These genes encode the rho1 and rho2 subunits of the C type receptor for γ-aminobutyric acid (GABAc receptor).16 17The GABAc receptor is expressed in the horizontal and bipolar cells of the retina.18 19 For this reason, we considered both genes to be attractive candidates for mutation analysis.

In order to identify the intron-exon boundaries ofGABRR1, we selected the gene that encodes the β1 subunit of the GABAa receptor whose complete genomic structure is known (M59212). Comparing the cDNA of theGABRR1 gene (M62323) and the cDNA of theGABRB1 gene (X14767),20 we obtained regions of high homology, approximately 59%, in the fragments corresponding to exons 6, 7, 8, and 9. However, the homology observed in the fragments corresponding to exons 1, 2, 3, 4, and 5 was less than 42%.

Afterwards, we localised by homology the different putative exons of the cDNA of the GABRR1 gene, which permitted the design of exonic primers to amplify exon-exon fragments containing all the introns. All the primers had the universal M13 primers attached 5′ (see table 1 for more details). The large PCR products were purified and then sequenced by the biochemical method of Sanger using dideoxynucleotides as terminators (fmol^®DNA Sequencing System Promega, Madison, WI). Electrophoresis was carried out in the automatic sequencer Alf-Express (Amersham-Pharmacia Biotech) at 1500 V and 50°C using Long Ranger Singel^TM (FMC) matrix.

The sequences obtained were analysed with the Alf-Manager^TMprogram and were aligned afterwards with the cDNA sequence of the GABAc receptor (M62323) using the command Bestfit for GCG or the Multalin programs (Multiple Alignment with Hierarchical Clustering) or both.21

Using this approach, we identified the four fragments corresponding to the last four introns of the GABRR1 gene (table 1).The information regarding introns 1, 2, 3, 4, and 5 of theGABRR1 gene has been published elsewhere.22

In order to perform mutation screening ofGABRR1 andGABRR2, the index patients of the ARRP families that showed linkage to RP25, RP5.II.1, RP73.II.1, RP167.II.8, and RP214.II.5, were selected.14 The DNA samples were PCR amplified using intronic primer pairs (tables 2 and 3). The products obtained were analysed by direct sequencing and fluorescent single strand conformational polymorphism analysis (SSCP) in the Alf-Express automatic sequencer (Amersham-Pharmacia Biotech) at 15 W. The migration patterns of each of the fragments were analysed using the Fragment Manager^TM program. The DNA fragments corresponding to exons 4 and 8 of the GABRR1 gene were digested with the restriction endonucleases MspI (Roche Diagnostic) and EagI (Amersham-Pharmacia Biotech) respectively before SSCP analysis.

The sequences obtained were aligned with the previously published cDNA sequence (M62323), the sequence we obtained after the analysis of the intron-exon boundaries, and the sequence provided by Hackamet al. 22 A total of 12 variants were found, 10 in GABRR1 (table 4), four of which are described in this work, and two inGABRR2 (table 5). However, none of them appear to be disease causative since they were found in the controls.

All the polymorphisms detected (tables 4 and 5) were confirmed by restriction analysis following the manufacturer's instructions. The 5′UTR-RsaI and IVS2+45C→G polymorphisms were genotyped by PCR digestion using a RsaI (Roche Diagnostic) site and MaeIII (Amersham-Pharmacia Biotech) site, respectively, introduced into the PCR primer next to the nucleotide change (table 4).

The 12 polymorphisms identified in GABRR1and GABRR2 were genotyped in all families previously linked to RP25. In the analysis of M20V of GABRR1 and V84V ofGABRR2 in the consanguineous family RP5, patients RP5.II.1 and RP5.II.3 were observed to have homozygous M20V and V84V changes, while a third patient, RP5.II.2, was heterozygous for these variants. The analysis of the other changes, namely IVS2+45C→G, IVS6-33C→T, and A389A ofGABRR1 and V84V ofGABRR2, in the consanguineous family RP167, showed that patient RP167.II.8 was homozygous for the normal alleles, while patient RP167.II.3 was homozygous for the mutated ones (fig 1). These results exclude the GABRR1 andGABRR2 genes as the cause of RP in both consanguineous families (RP5 and RP167). Since theRP25 locus was identified by homozygosity mapping, these data argue against the involvement of these genes inRP25.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Segregation of the polymorphisms of the GABRR1 and GABRR2 genes in two families linked to RP25, RP5 and RP167.

The RP25 locus is the third gene involved in RP and the seventh one related to retinal degeneration localised on chromosome 6. According to the data from the human transcription map,24 the initial RP25critical region colocalises with two loci involved in retinal degeneration, an autosomal dominant Stargardt-like locus (STGD3)25 and an autosomal dominant cone-rod dystrophy locus (CORD7),26 sharing a region of 4.8 cM. These disorders are different, but it cannot be excluded that the same gene could be responsible forSTGD3, CORD7, andRP25. This allelic heterogeneity has already been reported for the peripherin/RDS gene, the ABCR gene, and theCRX gene.27-30 Recently, a kindred with autosomal dominant cone-rod dystrophy with features of Stargardt-like disease where genetic analysis has shown linkage toCORD7 and STGD3on chromosome 6q14 has been identified.31 On the other hand, linkage analysis in one family of Pakistani origin has refined the RP25 critical region from 16.1 cM14 to 2.4 cM between D6S1053 and D6S430.32However, according to the physical and genetic maps available, the data provided by Khaliq et al 32would not be consistent with the overlap ofRP25 andCORD7/STGD3.

On the whole, the data reported argue against the involvement of theGABRR1 and GABRR2genes in RP25. However, the exclusion of both genes does not rule out other genes involved in neurotransmission within the critical region. In order to address the search for additional candidate genes for RP25, our current efforts include building a physical map across the current critical region to localise the STSs, ESTs, and polymorphic markers to the critical region.