Variant Phenotypes of Incomplete Achromatopsia in Two Cousins with GNAT2 Gene Mutations

Abstract

purpose. The present study was designed to elucidate the molecular genetic basis of a congenital stationary cone dysfunction characterized by congenital nystagmus, moderate visual impairment, and markedly disparate color vision deficiencies between two affected cousins.

methods. Ophthalmic examinations with emphasis on color vision and electrophysiology. Molecular genetic analysis of the X-linked cone opsin genes, mutation screening of the CNGA3, CNGB3, and GNAT2 genes, and heterologous splicing experiments.

results. Whereas the proband was found to carry a homozygous frameshift mutation (Tyr95fs) in GNAT2, her cousin was compound heterozygous for the Tyr95fs and a new intronic mutation c.461+24G→A. Heterologous expression in COS7 cells showed that the latter causes a splicing defect that results in early translation termination. Yet, this mutation is leaky, giving rise to small amounts of correctly spliced transcripts and offer an explanation for the diverging clinical findings in the cousins, one best described as incomplete achromatopsia and the other with oligocone trichromacy.

conclusions. The cases presented broaden the phenotypic spectrum of GNAT2 mutations and underline the increasing importance of molecular genetics in the clinical diagnosis of atypical ophthalmic phenotypes.

Congenital achromatopsia is a diagnostic term applied to rare hereditary conditions characterized by nonprogressive low vision from early infancy, severe color vision defects, reduced or absent photopic ERG responses, and essentially normal scotopic responses. Nystagmus and photophobia are frequently present and the foveal ring reflex is often indistinct or missing. However, no characteristic fundus changes are present. In the absence of any detectable cone activity, achromatopsia is designated complete and is transmitted as an autosomal recessive trait. Some subjects, however, show residual color vision, enabling them to recognize colors in daily life and to perform some color-matching tasks, depending on the fraction, type, and topographical distribution of the residual cone population. This condition is termed incomplete achromatopsia. A seemingly paradoxical situation occurs in some patients with achromatopsia who perform (nearly) normal color matches, whereas their photopic ERG responses are virtually undetectable. The term oligocone trichromacy has been introduced for this condition, expressing the notion that a small number of all three cone classes, sometimes even located preferentially in the foveal region may be able to provide color vision capabilities, although their combined electrical activity defies ERG measurement. ^{1 2}  

The molecular basis of incomplete achromatopsia is only partially elucidated. Known examples include blue cone monochromacy, an X-linked subtype of incomplete achromatopsia with preserved blue cone function caused by mutations in the red-green opsin gene cluster on chromosome X at q28. ^{3 4}  

Throughout recent years, considerable progress has been made in the genetics of complete achromatopsia (rod monochromacy). Mutations in three genes, CNGA3, CNGB3, and GNAT2, have been identified that account for ∼70% of achromatopsia cases in the central European population. ^{5 6 7 8 9} However, thorough clinical evaluations of such patients are still sparse, and it remains to be determined whether and to what extent rod monochromacy and incomplete achromatopsia are allelic or genetically distinct conditions. A recent study indicates that at least some mutations in the CNGA3 gene may be associated with a phenotype of incomplete achromatopsia. ¹⁰  

Herein, we report the clinical findings in two cousins with variant phenotypes of incomplete achromatopsia associated with mutations in GNAT2. Although a homozygous frameshift mutation in the proband has recently been identified, ⁸ we have now extended the molecular analysis to include the less severely affected cousin. He was found to carry an intronic mutation that activates a cryptic splice donor site and eventually introduces an early stop codon. Leakiness of this mutation provides an explanation for the milder phenotype compared with that of the proband. 

Materials and Methods

Subjects and Clinical Examination

Two cousins with congenital nystagmus, low vision, and electrophysiological evidence of a cone dysfunction differed radically with respect to color vision performance. No obvious clinical findings accounted for this discrepancy. Based on family history and data from the National Civil Register, a pedigree chart was constructed (Fig. 1) . 

Genealogical identification of the ancestors back to approximately 1800 was achieved through parish registers. Four persons (II:8, II:10, III:2, and III:4) underwent a standard ophthalmic examination, including evaluation of eye movements, visual acuity (VA), refraction, and slit lamp and fundus examinations, with fundus photography. Photopic visual field measurements were performed with a Goldmann perimeter (Haag-Streit AG, Bern, Switzerland) with objects I/4e and IV/4e. 

Color vision was evaluated with the following tests: Ishihara 38 plates 1986 edition, Farnsworth-Munsell D-15 (F-M D-15) standard and desaturated, Farnsworth-Munsell 100-Hue test (F-M 100), Lanthony’s tritan album (LTA), American Optical Hand-Hardy-Rittler (AOHRR), Berson’s blue cone monocromacy test, ¹¹ and Nagel’s anomaloscope (Schmidt-Haensch, Berlin, Germany). The test battery also included a standardized color sheet (DS735) with 17 colors presented as separate circular plates 25 mm in diameter. 

Dark adaptation was measured with a Goldmann-Weekers instrument (Haag-Streit AG) according to the instruction manual for the integral examination of the whole retina. Normal values for comparison were obtained in 25 healthy individuals. 

Full-field ERG recordings according to the standard protocol of the International Society for Clinical Electrophysiology of Vision (ISCEV) were made from one eye only. ¹² The pupils were dilated with cyclopentolate 1% followed by dark adaptation for 30 minutes. The recordings were performed with a Burian-Allen monopolar electrode in topical anesthesia with oxybuprocaine 1%. A reference silver ear-clip electrode was placed on the opposite earlobe and a ground electroencephalogram (EEG) silver cup electrode on the forehead. 

Stimulations were produced with a xenon flash tube with adjustable intensity (0.16–2.2 J/flash) mounted in a Ganzfeld globe (Nicolet Biomedical Instruments, Madison, WI). The white background luminosity inside the sphere was adjusted to 17-cd/m² photopic measurements. A yellow background was obtained by inserting a glass filter (model OG 530; Schott; Mainz, Germany) in front of the illuminating bulb and adjusting the illumination lamp’s intensity to 100 cd/m². The flash intensity was 0.5 U, equivalent to 12 cd-s/m² at the surface of the bowl. Attenuation of the flash intensity was obtained by inserting neutral-density filters in front of the flash bulb. The duration of the flash was 20 to 30 μs. Amplification, filtering, and averaging of the recordings were accomplished with a workstation (Viking IV; Nicolet Biomedical Instruments). Band-pass filters were 0.2 to 1000 Hz, except for the recording of oscillatory potentials (OPs; band-pass 100–1000 Hz) and blue cone recordings (band-pass 5–1000 Hz). A unit log improvement in signal-to-noise ratio was obtained with a digital IIR high-Q band-pass filter with second-order slopes (center frequency 30.0 Hz; bandwidth 1.0 Hz). Multifocal ERG (mf)ERG was recorded with a visual evoked potential imaging system (VERIS ver. 4.1; EDI, San Mateo, CA). The examination was performed in tandem with the ERG recordings, using the same electrode. Stimulation was performed with 103 hexagons with full contrast and scaled with eccentricity, by using a camera attachment allowing for visual control of the fixation. The overall stimulation pattern subtended a visual angle of 20° to 30° on either side of fixation. Data processing was accomplished with the standard clinic software program. The study adhered to the tenets of the Helsinki Declaration, and written informed consent was obtained from all examined family members. 

Blood samples were taken from the proband (III:4), the affected cousin (III:2), his healthy brother and sister, and both parents of the two patients (Fig. 1) . 

Molecular Genetic Analysis

DNA was isolated from peripheral blood according to standard procedures. Analysis of the structure of the red and green pigment genes was performed as previously described. ¹³ The locus for the red and green pigment genes maps to the very distal portion of the long arm of the X chromosome (q28/151M). Four microsatellite markers spanning this region were used for genotyping: DXS984 (q27.1/138M), DXS292 (q27.3/140M), DXS548 (q28/145M), and DXS1113 (q27.3-q28). 

Mutation screening of the CNGA3, CNGB3, and GNAT2 genes was performed as described previously. ^{6 8 10} In addition, all introns and the putative promoter region up to −1200 upstream of the annotated translation start codon ¹⁴ of the GNAT2 gene were analyzed in subject III:2 by direct sequencing of PCR products. Segregation analysis of the mutations within the family was performed by sequencing of appropriate PCR products. Screening for the c.461+24G→A mutation in the control subjects was performed by PCR/restriction fragment length polymorphism (RFLP) analysis (gain of an XcmI site after amplification with a mismatch-containing sense primer). Sequences of the primers and PCR conditions are available from one of the authors (BW) on request. 

Heterologous Splicing Assay

A 2903-bp genomic segment covering exons 4 to 5 and flanking sequences of introns 3 and 5 of GNAT2 were amplified from genomic DNA, by applying Pfu Turbo polymerase (Stratagene, Heidelberg, Germany) and were blunt-end cloned into the PCR-Script Amp Sk⁺ vector (pCR-Script Cloning Kit; Stratagene). The insert was excised by digestion with BamHI and NotI and cloned into BamHI/NotI–digested pSPL3_2096 (a derivative of pSPL3 [Invitrogen-Life Technologies, Karlsruhe, Germany], with a stuffer fragment cloned into the original NotI site). Allelic constructs were partially sequenced, including vector–insert boundaries and exon–intron borders. 

COS7 cells were cultured in DMEM supplemented with 5% (vol/vol) nonessential amino acids and 10% fetal calf serum. Cells were transfected with pSPL3-GNAT2 constructs with transfection reagent (Lipofectamine; Invitrogen). Thirty-six hours after transfection, total RNA was prepared (TRIzol reagent; Invitrogen) and reverse transcribed with either oligo-dT or SA2 primer (a reverse primer located on the 3′ tat exon of the pSPL3 vector) and avian myeloblastosis virus (AMV) reverse transcriptase (RT-PCR Kit; Takara, Kyoto, Japan). The cDNA was PCR amplified with pSPL3 exon primers. 

The RT-PCR products were either directly sequenced or sequenced after an intermitting cloning step applying dye terminator chemistry (BigDye Terminator Kit; Applied Biosystems, Inc. [ABI], Darmstadt, Germany). All sequences were run on an ABI3100 DNA sequencer manually checked, and analyzed on computer (SeqMan software; Lasergene, Madison, WI). 

To evaluate the presence of minute amounts of correctly spliced transcripts in RNA isolated from COS7 cells transfected with the mutant pSPL3-GNAT2 construct, diluted primary RT-PCR products were reamplified with a sense primer (GNAT_45B: 5′-ATG ACT CCG CAT CTT ACT ACC-3′) that locates at the boundary of exons 4 and 5. The specificity of this PCR reaction was verified on cloned wild-type and mutant cDNA templates. 

Results

Clinical Findings

The proband (Fig. 1 ; III:4), a female born in 1973, was referred to the National Eye Clinic (NEC) for the Visually Impaired at the age of 3 years due to congenital nystagmus and low vision. She was the single child of unrelated parents and was born at term after an uncomplicated pregnancy. Her father was reported to be red-green color-blind with normal vision, and a paternal cousin (Fig. 1 ; III:2) had congenital nystagmus. 

At first examination in 1977, alternating esotropia and pendular, coarse, horizontal nystagmus were present. Cycloplegic retinoscopy of the right eye was +6.0 D and of the left eye was +8.0 D. VA was 0.2. The fundus appeared normal, except for absent foveal reflexes and a slight grayish discoloration in the perifoveal region. 

The patient was re-examined in 1989 at the age of 16 years. She reported moderate photophobia and some difficulties with color discrimination. Apart from the visual difficulties she was a healthy young woman. 

VA was 0.1 for the right eye and <0.1 in the left eye. She had become emmetropic. In the fundi, a golden foveal reflex was noted. She identified none of the Ishihara plates, except for the first, but she was able to name 17 standard colors (DS735) correctly. She identified six plates of the AOHRR screening series, and the diagnostic series disclosed a mild R-G defect. F-M D-15 was performed with diagonal errors in axes between deutan and scotopic. She failed to identify any colors in Berson’s blue cone monochromacy test. Nagel anomaloscopy revealed extreme protanopia (Fig. 2A) . A full-field ERG demonstrated slightly subnormal scotopic recordings of rod-generated signals in contrast to a complete absence of the cone-elicited flicker. White single flashes on the light-adapting background generated a rod-like response (Fig. 3) . The blue cone signal was completely absent (not shown). Narrow band-pass filtering, however, demonstrated a rudimentary cone flicker response of 0.5 μV (Fig 3) . mfERG showed no recognizable cone pattern. 

The remaining eye examination, including Goldmann perimetry, was unremarkable. At re-examination in 2003, the patient’s refraction had changed to RE −6.0 D, LE −7.0 D. VA was 0.05 and 0.1 with correction, but otherwise her status was unchanged. 

The proband’s father (Fig. 1 ; II:10) was examined at the age of 44 years. His VA was 1.0 in both eyes without glasses. Color vision testing including Ishihara, F-M D-15, LTA, and Nagel’s anomaloscope. He had a very wide range of Raleigh matches between 0/16 and 50/20 as extremes. These settings determine his color vision defect as either extreme deuteranomaly or deuteranopia. Slit lamp examination, ophthalmoscopy, and ERG yielded normal results. 

The paternal cousin (Fig. 1 ; III:2) was a male born in 1960. He was reported to the National Register for Visually Impaired Children at the NEC at the age of 6 years because of low vision and congenital nystagmus. When examined at the age of 30 years, he denied photophobia and any color discrimination difficulties. 

VA in the right eye was 0.2, −4.50 sphere, and VA in the left eye was 0.3, −2.0 sphere, −2.0 cyl × 30°. Horizontal nystagmus, with a rotator component and variable amplitude and frequency, was present. The Ishihara plates were correctly identified except for plate 12 (87 instead of 97), plate 14 (6 instead of 5), and plate 73 (13 instead of 73). These mistakes may be due to reduced VA as much as to any color vision deficiency. Four of six of the R-G plates in the AOHRR screening series were identified, and he read all plates in the diagnostic series without mistakes. The F-M D-15 tests showed a few insignificant inversions (Fig. 2B) , LTA and Berson’s blue cone monochromacy tests were read correctly, and he had an error score of 107 with the F-M 100 (Fig. 2C) . In contrast, examination with the Nagel anomaloscope revealed a protanomalous color vision defect (Fig. 2A) . 

ERG recordings demonstrated slightly subnormal rod responses and totally absent cone responses to flicker stimulation. However, as in the proband, a small rodlike signal appeared with single white flash stimulation on a light-adapting background. No blue cone response was present. With narrow band-pass filtering, a flicker response near the noise level of 0.2 μV was measured (Fig. 3) . mfERG revealed no recognizable cone pattern. The results of the remaining eye examinations were normal, including biphasic dark adaptometry and Goldmann perimetry. 

The cousin’s mother (Fig. 1 ; II:8) was examined at the age of 48 years. She had no visual symptoms. The routine examination demonstrated normal findings, and she read the color tests Ishihara, F-M D-15, and LTA correctly. 

Molecular Genetics

Figure 1 shows the composition of the red and green pigment gene tandem array as determined for the proband III:4; her father II:10; her paternal cousins III:1, III:2, and III:3; and the cousins’ parents II:7 and II:8, of which II:8 is the proband’s paternal aunt. 

The proband’s father II:10 carried an abnormal pigment gene array that contained a 5′ green-red 3′ fusion gene besides one red and two green pigment genes. This genotype causes a deutan defect only if the 5′ green-red 3′ fusion gene occupies the second position in the pigment gene array. ¹⁵ The father has a deutan defect, namely deuteranopia, and therefore the fusion gene must occupy the second position of the array, as indicated in Figure 1 . The proband’s genotype is composed of this abnormal pigment gene array on her paternal X-chromosome and a normal gene array on the her maternal X chromosome consisting of one red and three green pigment genes. This genotype does not explain the proband’s visual dysfunction. The proband’s paternal aunt (II:8) had a normal gene array consisting of one red and two normal green pigment genes on both her X chromosomes, and a similar gene array was found on her husband’s (II:7) X chromosome. As expected, all the X chromosomes of their children (III:1, III:2, and III:3) carried one normal red and two normal green pigment genes. 

II:8, III:1, and III:2 were genotyped with respect to four microsatellite markers spanning the q27-q28 region on the X chromosome. II:8 was heterozygous for markers DXS984 and DXS292, homozygous for marker DXS548, and heterozygous for marker DXS1113 (data not shown). The haplotypes of III:1 and III:2 were identical, except for the proximal marker DXS984, indicating that a crossover event had taken place between the markers DXS984 and DXS292. Because the red-green color vision locus is distal to DXS548, it is almost certain that the red and green pigment gene array transmitted to III:1 and III:2 originate from the same maternal X chromosome. Therefore, any phenotype produced by the red-green color vision locus would have to be shared by the brothers III:1 and III:2. These data clearly indicate that the cone dysfunction in subject III:2 is not caused by a defect in the X-linked cone opsin gene cluster. 

The proband was then screened for mutations in CNGA3 and CNGB3, and no putatively pathogenic sequence alteration was found. However, analysis of the GNAT2 gene revealed the presence of a homozygous deletion–insertion mutation (c.285_291del7insCTGTAT; Fig. 4 ) which results in a frameshift and subsequent translation termination another 61 deviant codons downstream (p.Y95fsX61). ⁸ Both parents were heterozygous for this mutation. 

The proband’s affected cousin was also heterozygous for the c.285_291del7insCTGTAT mutation, as was his mother (Fig. 1) . Yet this single heterozygous GNAT2 mutation did not explain the phenotype, because other heterozygotes (e.g., the proband’s mother II:11 and the cousin’s mother II:8) had no visual symptoms. However, analysis of the remaining GNAT2 exons did not reveal any other putative disease-associated mutation. We therefore sequenced the complete GNAT2 gene, including the promoter region and all introns in this subject. This analysis led to the identification of an unknown A-to-G substitution at position +24 in intron 4 (Fig. 4 , c.461+24G→A). 

Analysis of this substitution in 156 control subjects ruled out the possibility of a common polymorphism. We reasoned that this substitution might affect splicing of the GNAT2 pre-mRNA and therefore tested this substitution in a heterologous splicing assay. A genomic segment that included the complete intron 4, its flanking exons 4 and 5, and ∼200 bp of each of the adjacent introns 3 and 5 was amplified from patient III:2 and a control subject and was cloned into the exon trapping vector pSPL3. Plasmids with or without the c.461+24G→A substitution were transfected into COS7 cells, and RT-PCR products were amplified from RNA of transfected cell cultures, with primers specific for the flanking tat exons of the pSPL3 vector. RT-PCR products from cultures transfected with plasmid constructs with the c.461+24G→A substitution were clearly larger than products from cultures transfected with the control plasmid (Fig. 5A) . 

Subsequent sequence analysis revealed a 21-bp insertion between exons 4 and 5 (Fig. 5B) . This insertion results from the inclusion of adjacent intron 4 sequences, due to the usage of an alternate splice donor sequence activated by the c.461+24G→A substitution and gives rise to a premature stop codon, right at the genuine exon–intron boundary (Fig. 5C) . Although agarose electrophoresis and sequence analysis of RT-PCR products indicated a rather homogenous population of mutant cDNA, we tested whether minute amounts of correctly spliced transcripts might be produced in cells transfected with the mutant construct. By applying a sense primer situated at the boundary of exons 4 and 5 we were able to obtain appropriate PCR products from diluted primary RT-PCR amplifications (Fig. 5D) . These products are specific for correctly spliced cDNAs, as shown with cloned wild-type and mutant cDNAs as templates. 

Discussion

Congenital nystagmus (CN) is a common condition encountered in a pediatric low-vision clinic. As a purely symptomatic diagnosis, CN requires a diagnostic evaluation, including a family history. In infants, early ERG examination may preclude extensive pediatric assessment by disclosing one of many conditions in which a routine ophthalmic examination fails to demonstrate morphologic changes. These conditions, include, among others, Leber’s congenital amaurosis, congenital stationary night blindness (CSNB), Åland eye disease/incomplete CSNB, complete and incomplete achromatopsia, blue cone monochromacy, and some rare cone-dysfunction disorders such as oligocone trichromacy and cone dysfunction with supernormal rod response. Later in infancy, the assessment may be expanded to include various psychophysical examinations such as color vision testing. 

Despite pertinent differential diagnostic elucidation, some cases remain for which a precise taxonomic classification is unattainable. The presented cases belong to this category. 

The results of the color vision tests were confusing and did not support an unambiguous classification of the defect. The type of color vision deficiency in the proband is best described as incomplete achromatopsia or extreme protanopia, whereas the cousin had a mild protanomalous trichromacy (oligocone trichromacy), and the proband’s father displayed a dichromatic deutan defect. The latter is neatly explained by the presence of a green-red hybrid opsin gene typically associated with deutan color vision defects. The affected cousin, in contrast, had a normal X-linked opsin gene array. Furthermore, a comparison of microsatellite marker haplotypes between the cousin and his unaffected brother proved that they had inherited the same distal X chromosome region from their mother (Fig. 1) . Thus, the red-green opsin gene locus can be excluded as a cause of or contributor to the disease phenotype of the cousin. 

Electroretinographically, the proband and her cousin presented the characteristics of rod monochromacy. The unusual presence of a rod-like response to single-flash stimulation in the light-adapted state may reflect an incomplete suppression of rod activity or, alternatively, the presence of an adaptive mechanism by which residual cone activity is transmitted through neural rod circuits incapable of transmitting responses to flicker stimulation. 

Despite barely detectable cone function in electroretinographic recordings, a certain residual cone activity is indicated by the partially preserved color vision and the normal peripheral cone functions, as measured by light-adapted visual fields. This could either be accomplished with subpopulations of functional cones or a subtotal suppression of cone activity. Whichever the mechanism, the relatively preserved ability of color discrimination with barely detectable ERG cone activity remains paradoxical. 

Recently, Michaelidis et al. ¹⁶ described the phenotypes of several affected subjects from a large consanguineous Pakistani family with a homozygous frameshift mutation in the GNAT2 gene. In accordance with our findings, these authors detected residual color vision in three of the five affected subjects. In addition, small photopic ERG responses elicited by short-wavelength stimuli were recorded. ¹⁶  

How can these phenotypes and in particular the residual cone function be explained at the cellular level? Besides those cases with a GNAT2 mutation, residual cone function has also been found in some patients with missense mutations in the CNGA3 gene. ¹⁰ In vitro studies have shown that at least one of the CNGA3 mutant polypeptides expressed in incomplete achromats retains channel function, albeit with altered biophysical properties. ¹⁷ It is therefore most likely that functionally altered cGMP-gated channels are responsible for the residual cone function in these subjects. 

Yet, the two subjects with GNAT2 mutations described herein carried mutations of a sort that may explain the phenotype of incomplete achromatopsia. The index patient carried a homozygous frameshift mutation that results in a largely shortened polypeptide composed of the first 95 genuine amino acid residues followed by 61 residues encoded by a different reading frame. The cousin is a compound heterozygote carrying the Y95fsX61 mutant allele, which he shares with the proband, and the newly identified splicing mutation c.461+24G→A that leads to the formation of a termination codon at codon position 154 (p.Y154X). 

The nature of the latter mutation may provide a clue for understanding his mild phenotype with considerable color discrimination capabilities. The c.461+24G→A mutation activates a cryptic splice donor remote from the genuine splice donor. A reason for the preferential usage of the cryptic splice site may lie in the weakness of the genuine splice donor. Although the 3′ terminal end of the GNAT2 exon 4, a TTA, is rather uncommon for splice donor sequences, the AGG trinucleotide at the end of the extended exon in the mutant fits more closely to the consensus sequence. Indeed, splice site predictions using the GeneSplicer program (www.tigr.org/genesplicer/) ¹⁸ provide a score of 4.19 for the cryptic splice donor (in the presence of the c.461+24G→A substitution) compared with a score of 3.83 for the regular splice donor. 

Nevertheless, the c.461+24G→A mutation leaves the genuine splice donor intact, which may in turn compete with the activated cryptic splice donor. Indeed, we were able to demonstrate the presence of low amounts of correctly spliced transcripts, at least in the heterologous expression system. Provided that the c.461+24G→A mutation is also leaky in vivo and that small amounts of correctly spliced GNAT2 transcripts are formed in cones, some low level of phototransduction may proceed in these cones and thus explain the cousin’s having better color discrimination capabilities than the proband. 

Yet, psychophysical analysis showed that even the proband had some color discrimination. A possible explanation for this unexpected finding comes from recent analysis of the homologous zebrafish nof mutant. This mutant carries an early nonsense mutation in the orthologous gnat2 gene that is associated with a preserved retinal morphology and normal levels of other phototransduction enzymes. ¹⁹ Intriguingly, nof cones respond to bright light by a transducin-independent mechanism that may involve the release of Ca²⁺ from sequestered stores into the photoreceptor cytoplasm. It has been argued that such a light-evoked increase in intracellular Ca²⁺ may cause changes of the membrane potential sufficient for subsequent neuronal signaling. Desensitization of cGMP-gated channels and decreased guanylate cyclase activity through inhibition of the guanylate cyclase activator proteins are well-known negative-feedback mechanisms of increased cytoplasmic Ca²⁺ levels ²⁰ that by themselves may cause a suppression of the dark current. If such a mechanism also exists in humans, it may well explain the residual psychophysical cone function in the proband. The presence of a deutan genotype on one of her X chromosomes may further contribute to her particular color vision phenotype. 

Our study demonstrates the value of molecular genetic analysis for diagnostic and taxonomic classification of atypical congenital cone dysfunctions. To our knowledge, this is the first report of an oligocone trichromacy phenotype to be elucidated at the molecular level. The variant phenotypes of the two cousins with GNAT2 mutations also provide good examples of the phenotypic spectrum of incomplete achromatopsia that include subjects with almost complete monochromacy and others with mild color vision defects depending on the number and the distribution of the remaining photoreceptors. 

 Supported in part by Grants SFB430/A5 and Wi 1189/61 from the Deutsche Forschungsgemeinschaft and by the Danish Society for the Blind.

 Submitted for publication March 19, 2004; revised June 21, 2004; accepted July 29, 2004.

 Disclosure: T. Rosenberg, None; B. Baumann, None; S. Kohl, None; E. Zrenner, None; A.L. Jorgensen, None; B. Wissinger, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Segregation of two different GNAT2 mutations in a family with incomplete achromatopsia. The affected proband, III:4, was homozygous for a frameshift (fs) mutation (c.285_291del7insCTGTAT) in the GNAT2 gene. She inherited the mutation from each of her parents, II:10 and II:11, who were heterozygous carriers. The proband’s unaffected aunt, II:8, was also a heterozygous carrier of this frameshift mutation, and she passed this allele on to her affected son, III:2, who inherited a GNAT2 gene containing a splice mutation (c.461+24G→A) from his unaffected father, II:7. III:2 was therefore compound heterozygous with respect to GNAT2 mutations. The red-green color vision locus of the segregating X chromosomes of the family members carried either a normal array of a 5′ red pigment gene (red arrow) followed by two or three green pigment genes (green arrow) or an abnormal array of one red gene followed by a green-red fusion gene (green-red arrow) and two green genes, consistent with a deutan defect in subject II:10.

Figure 2.

 

Psychophysical color vision tests. (A) Nagel anomaloscope. (•) proband; (▪) affected cousin. (B) F-M 15-D results of the affected cousin: standard test (left) and desaturated test (right). (C) Farnsworth-Munsell 100-hue results of the affected cousin.

Figure 3.

 

Full-field ERG from two subjects with incomplete achromatopsia due to GNAT2 mutations with a normal subject for comparison (top traces). Traces 1 to 4 were recorded in the dark-adapted state and traces 5 to 7 with a light-adapting background: 1, rod-response; 2, mixed cone-rod response; 3, oscillatory potentials; 4, 30-Hz flicker cone response; 5, cone response; 6, 30-Hz flicker cone response; and 7, narrow-band filter cone response. Horizontal bar: 100 ms (traces 1 and 2 and 4–6), 40 ms (trace 3), and 20 ms (trace 7). Vertical bar: 200 μV (traces 1 and 2), 40 μV (trace 3), 100 μV (traces 4 and 6), 40 μV (trace 5), and 0.2 μV (trace 7).

Figure 4.

 

Compound heterozygous GNAT2 mutation in subject III:2. Electropherogram sections presenting the 285_291del7insCTGTAT mutation (bottom left) and the c.461+24G→A mutation (bottom right) in subject III:2 in comparison with the respective wild-type sequences (top).

Figure 5.

 

Heterologous splicing analysis of the c.461+24G→A mutation. (A) RT-PCR products obtained from RNA of COS7 cells transfected with either the mutant (lanes 2 and 4) or the control construct (lanes 1 and 3). Reverse transcription was primed with either oligo-dT (lanes 1 and 2) or SD6 primer (lanes 3 and 4). Lanes M: 100-bp size standards. (B) Electropherogram section of directly sequenced RT-PCR amplified from COS7 RNA transfected with either the mutant or the control constructs. (C) Schematic overview of the splicing of wild-type and mutant transscripts. The c.461+24G→A mutation activates a cryptic splice donor 21 bp downstream of the genuine splice site. Extension of exon 4 leads to the formation of a stop codon. (D) Detection of correctly spliced cDNA derived from mutant constructs. Applying a primer situated at the boundary of exons 4 and 5 (right), appropriate PCR products were obtained upon reamplification with 1:100 (lane 3) and 1:1000 (lane 4) dilutions of primary mutant RT-PCR products. The specificity of the assay was verified with ∼250 pg plasmid DNA of cloned control (lane 1) and mutant (lane 2) cDNAs. Lane M: pBR322/AluI size standard.