Molecular Vision 1998; 4:16 <>
Received 14 July 1998 | Accepted 12 September 1998 | Published 17 September 1998

Exon screening of the genes encoding the ß- and [gamma]-subunits of cone transducin in patients with inherited retinal disease

Yong Qing Gao,1 Michael Danciger,1,3 Novrouz B. Akhmedov,1 Dan Yun Zhao,1 John R. Heckenlively,1 Gerald A. Fishman,4 Richard G. Weleber,5 Samuel G. Jacobson,6 Debora B. Farber1,2

1Jules Stein Eye Institute, UCLA School of Medicine, Los Angeles, CA; 2Molecular Biology Institute, UCLA School of Medicine, Los Angeles, CA; 3Loyola Marymount University, Los Angeles, CA; 4University of Illinois at Chicago, Chicago, IL; 5Casey Eye Institute, Oregon Health Sciences University, Portland, OR; 6Scheie Eye Institute, University of Pennsylvania, Philadelphia, PA

Correspondence to: Michael Danciger, Jules Stein Eye Institute, UCLA School of Medicine, Los Angeles, CA, 90095-7008; FAX: (310) 794-2144; email:


Purpose: To screen the exons of the genes encoding the ß3-subunit (GNB3) and [gamma]c-subunit (GNGT2) of cone transducin for mutations in a large number of unrelated patients with various forms of inherited retinal disease including cone dystrophy, cone-rod dystrophy and macular dystrophy.

Methods: Exons of the two genes were screened for mutations by denaturing gradient gel electrophoresis (DGGE) and/or single strand conformation polymorphism electrophoresis (SSCP); any variants were sequenced directly.

Results: Although many sequence variants were found in both genes, none could be associated with disease. Additionally, the gene structure and sequence of the coding exons of GNB3 were determined and compared with those of the dog homolog. Both human and canine GNB3 have nine coding exons and their two predicted amino acid sequences have 97% identity.

Conclusions: The results indicate that GNB3 and GNGT2 are unlikely sites of mutations responsible for inherited retinal degenerations that predominantly effect cone-mediated function (cone and cone-rod dystrophies) or have a predilection for disease in the macula (macular dystrophies).


A rare autosomal recessive canine disorder selectively affecting cones (cd ) was originally identified in the Alaskan malamute dog [1]. Immunocytochemical studies comparing the proteins of pre-degenerate cd and normal dog retinas showed initially, the absence of cone-specific ß3-transducin subunit. At the same time, there were no differences between pre-degenerate and normal retinas in [alpha]-transducin, phosducin, the [alpha]-, ß- and [gamma]-subunits of cGMP-phosphodiesterase, opsin, S-antigen, IRBP and the middle and short wavelength sensitive cone markers cos-1, and cos-2 [2]. More recently, the cone specific [gamma]c-transducin was also shown to be absent from the pre-degenerate cd retina [3]. Although, both ß3- and [gamma]c-transducin are missing from the pre-degenerate cd retina, the mRNAs encoding each of the proteins are expressed and show no sequence differences from normal [3,4] suggesting that the primary defect is in another gene.

Cone and rod transducins are both heterotrimeric G proteins consisting of single [alpha]-, ß- and [gamma]-subunits. Although each of the subunits in cones is similar to its counterpart in rods they are nonetheless distinct. Functionally, however, the two transducins have the same role in the visual response of their corresponding photoreceptors. In both cases the transducin protein is separated into an [alpha]-subunit and a ß/[gamma]-subunit complex after interaction with light activated opsin. The[alpha]-subunit goes on to activate cGMP-phosphodiesterase causing the breakdown of cGMP and the closing of cGMP-gated cation channels on the plasma membrane. This leads to an interruption of the dark current of the photoreceptor and results in vision. In recovery, the ß/[gamma]-subunit complex binds back to the [alpha]-subunit, shutting down the cGMP-phosphodiesterase and its concomitant breakdown of cGMP. When cGMP is replenished by guanylate cyclase the dark current is resumed and the visual signal to the brain ends.

The absence and apparent involvement of ß3- and [gamma]c-transducin in a canine cone disease (even though the primary lesion is in another gene) and their central role in visual transduction makes the human genes encoding these proteins (GNB3 and GNGT2) prime candidates for study in human inherited retinal disease. This is supported by the fact that several other visual transduction proteins have been associated with inherited retinal disease, including opsin [5], the [alpha]-subunit of rod transducin [6], the [alpha]- and ß-subunits of rod cGMP-phosphodiesterase [7-10], arrestin [11], and others [12]. Since single base changes are the most common form of mutation, we adopted the strategy of exon screening of the GNB3 and GNGT2 genes.


Exon/intron boundaries of the ß3- and [gamma]c-transducin genes

Using the published sequences of human transducin-ß3 cDNA [13] and sequences of the canine GNB3 gene determined in our laboratory [4], we designed exon to exon primers for polymerase chain reaction (PCR) amplification. For the 5' and 3' UTRs (untranslated regions) we designed primers using sequences from canine exons 1 and 10, respectively. Direct sequencing of PCR products was performed with the Fmol DNA sequencing system (Promega, Madison,WI) with [[gamma]-32P] dATP. For exon 1, and introns 1, 8, and 9, it was necessary to first subclone PCR amplified fragments into a plasmid. Plasmid inserts were then sequenced using M13 primers and the Sequenase Kit (USB, Cleveland, OH) with [[alpha]-35S] dATP. We compared our results with results taken from a recent 222,930-bp sequencing project spanning chromosome 12p13 and including the GNB3 gene [14]. Our results confirmed the exon/intron sequence boundaries described. The sequences of the GNGT2 gene exons and their intron boundaries have recently been published [15].

Sequence comparison of the dog and human GNB3 coding regions shows that both contain 1020 bp predicting 340 amino acids with 97% identity. Both homologs have 10 exons and 9 introns with exons 2-10 containing the coding region. Except for introns 2 and 8 that are larger in the human, the exons and introns of the dog and human genes are similar in size.


Included in this study were 186 unrelated patients with retinal degenerative disease and 51 unrelated control subjects. Controls self reported normal vision; their ages and ethnicities are not known. Table 1 lists the numbers of patients studied by modes of inheritance and clinical diagnoses. For the GNB3 gene, 164 were screened by single strand conformation polymorphism electrophoresis (SSCP) with two different concentrations of glycerol; for the GNGT2 gene, 148 patients (126 of which were part of the GNB3 screening) were screened by SSCP with only one concentration of glycerol and by denaturing gradient gel electrophoresis (DGGE). Exons in which potential mutant sequence variants were present were screened in 30 and 51 controls for GNB3 and GNGT2, respectively. A number of patient and control DNAs were lost after the GNGT2 screening, and some new patient DNAs were acquired. All participants were fully informed of the nature of the investigations, and the research was performed in accordance to institutional guidelines and the Declaration of Helsinki.

Patients were evaluated clinically and with visual function tests including electroretinography (ERG) and perimetry. The retinal degenerative diseases were grouped into three categories: cone-rod dystrophy (CRD), cone dystrophy (CD) and macular dystrophy (MD). CRD patients had relatively early reduction in visual acuity, later impairment of peripheral vision, and subnormal cone and rod ERGs. CD patients also had progressive visual acuity loss and subnormal cone ERGs but rod ERGs that were normal or near normal. MD patients, although varying in clinical presentation, all had central retinal abnormalities and normal ERGs. In this category were patients with atrophic or hemorrhagic macular degeneration of uncertain type as well as patients with a diagnosis of Stargardt disease/fundus flavimaculatus, defined as an early-onset progressive maculopathy with atrophy and fleck-like lesions [16]. Patients with age-related macular degeneration were not included nor were any patients suspected of having toxic retinopathy.

Exon screening by SSCP and/or DGGE

DNA was extracted from blood samples by standard methods. For SSCP, exons of the GNB3 and GNGT2 genes were amplified directly from genomic DNAs in the presence of [[alpha]-32P]dCTP with primer pairs that flanked each exon (Table 2). The PCR buffer was at pH 9.0 and contained 1.5 mM MgCl2. The PCR protocol was as follows: 94 °C for 2 min followed by 30 cycles of (94 °C, 1 min, 60 °C, 30 s, 72 °C, 45 s) followed by 5 min at 72 °C. The amplicons were denatured with NaOH and heat before being electrophoresed in 6% non-denaturing acrylamide gels buffered with 1x TBE (90 mM Tris base, 90 mM boric acid, 2 mM EDTA, pH 8.3). The GNB3 gene was screened by SSCP alone, but each GNB3 exon was run in two separate gels; one with 10% glycerol, the other with no glycerol. The GNGT2 gene was screened by DGGE and SSCP: each GNGT2 exon was run only in one SSCP gel with 5% glycerol. For DGGE each of the three exons of GNGT2 was amplified by the same primer pairs as for SSCP, but with the exception that one of the pairs had a "GC clamp" associated with it. The 36-bp "GC clamp" was attached to the primer with the highest GC content; its sequence (Table 2) was kindly provided by Drs. John Nakamoto and Neil vanDop (Department of Endocrinology, UCLA). Each GC-clamped amplicon was then run in two 14% acrylamide denaturing gradient gels (10% to 50% denaturing gradient and 40% to 80% denaturing gradient) in a 60 °C bath as previously described [17].

DGGE is a costly, labor-intensive procedure. Consequently, after the screening of the small (3 exons) GNGT2 gene (by DGGE and SSCP), and before screening the larger (9 coding exons) GNB3 gene, we compared the DGGE/SSCP screening method with the one involving only SSCP (described above for GNB3). Summing this study of GNGT2 and a previous study we found a total of 12 variants in 25 exons of two genes. In our studies of 77 exons of five genes screened by the SSCP/two gel method we found 35 variants. All genes were screened in a large number of patients. Since the number of variants was about one for every two exons with either method, we screened the GNB3 gene by the easier SSCP/two gel method.

Direct Sequencing of variant exons

The same primers used to amplify exons for SSCP were used to amplify variant exons for sequencing. The amplicons were purified and sequenced directly by the standard dideoxy method using [[gamma]-32P]dATP and the Fmol sequencing kit (Promega).


Exon screening - GNB3 and GNGT2

A total of four sequence variants were found in the GNB3 gene, and four in the GNGT2 gene. In GNB3 they were: (1) an insert of a single G in a cluster of G's starting after the ninth base pair downstream of exon 4 (in intron 4) which did not produce any change in the 14-bp splice site; (2) a GtoA transition 38 bp downstream from exon 4 (and 57 bp upstream from exon 5) which did not create a new or alter an existing branchpoint; (3) a GtoA transition predicting Gly272Ser in exon 9; and (4) a CtoT transition (C825T) predicting Ser275Ser in exon 9. In GNGT2 they were: (1) a GtoT transversion 31 bp upstream of the transcription start site which did not appear to be in any known regulatory sites; (2) an AtoG transition predicting Gln17Arg in exon 2; (3) a GtoT transversion predicting Leu11Phe in exon 2; and (4) a C deletion at the seventeenth base pair after the stop codon in exon 3. For each of these variants Table 3 shows the number of patients, their diagnoses, the number of controls and whether or not an additional variant is present in the same patient or control.

One AR-MD patient carried the transversion upstream of the transcription start site and the Gln17Arg missense in GNGT2 and the GtoA transition in intron 4 of GNB3. However, these variants did not segregate with disease in the patient's family. Another patient (AR-CRD) carried the transversion upstream of the transcription start site and the C deletion downstream of the stop codon in GNGT2 and the GNB3 intron 4 insert. DNAs from this patient's family were not available for testing, but none of his variants appear to predict any significant change to the protein encoded by this gene. The Ser275Ser variant of GNB3 was present in an approximately equal frequency in patients (53/164 = 32%) and controls (10/30 = 33%). The Gly272Ser of the same gene was also present in approximately equal frequency in patients (12/164 = 7%) and controls (2/30 = 7%). The two variants together were present in five patients and one control. All of the variants found in GNGT2 were present only in simplex or AR patients. Gln17Arg was present in 9/148 patients (one patient that carried other variants is referred to above) and 4/51 controls. These are approximately equal frequencies of 6% and 8%, respectively. The Leu11Phe was found in a simplex patient and a control. Neither parent of this proband had disease nor was any second variant found in her GNB3 or GNGT2 exons. The GtoT transversion upstream of the transcription start site was present in 8/148 patients (5%) and 1/51 controls (2%). No second variant was found in six of the eight patients; the other two patients are discussed above. Likewise, two of the three patients carrying the C deletion downstream of the stop codon revealed no additional sequence variants, and the other patient is discussed above.


The map locations of the GNB3 gene at human chromosome 12p13 [18] and the GNGT2 gene at human chromosome 17q21 [15] do not correspond to any currently known loci of inherited retinal disease. However, because the cone specific ß3- and [gamma]c-subunits of transducin are absent in the pre-degenerate cd dog (even though the genes encoding these proteins appear not to be the primary site of the genetic lesion), and because they play significant roles in phototransduction, it was logical to perform mutation analysis of the GNB3 and GNGT2 genes in patients with retinal degenerations since there are loci for these diseases that have not yet been mapped.

We were not able to assign disease-causing potential to any of the four variants found in non-coding regions of the two genes we studied. The two intron 4 variants in GNB3 neither altered existing nor produced new splice sites or branch points. One AR patient carried the GtoA transition in GNB3 along with two GNGT2 variants, but these did not segregate with disease in the patient's family. Another patient carried the G insert in GNB3 along with the upstream (of the transcription start site) transversion and the downstream (of the stop codon) deletion in GNGT2, but DNAs from his family were not available for testing. Nevertheless, because the DNA sequences around the two GNGT2 variants just mentioned did not correspond to any known regulatory sequences we could not assign any deleterious effect to them. Likewise, we could not find disease-causing potential in any of the four variants found in the coding regions of the two genes. The Leu11Phe in GNGT2 was present in only one simplex patient. Neither of the parents of this patient had disease nor was a second variant found. Additionally, this variant was present in a control subject. The Gln17Arg appears to be a polymorphism being present in equal frequencies in patients (6%) and controls (8%). It was present only in simplex and AR patients, and did not segregate with disease in the family of the one patient that carried other variants. Its presence in four controls (two of the controls were over 50 years) rules out AD disease. Both the Ser275Ser and Gly272Ser were present in relatively equal frequencies in patients and controls suggesting that they too are polymorphisms. Furthermore, neither are in the functional N-terminal or C-terminal binding domains of the transducin-ß3 protein.

Therefore, in our screening for exonic mutations in GNB3 and GNGT2 in 164 and 148 patients, respectively, with a number of different forms of retinal disease, we found several sequence variants, but none that could be associated with disease. In another study, exons of the gene encoding the [alpha]-subunit of cone transducin (GNAT2) were screened in patients with Stargardt disease. As in our case, no sequence variants could be associated with disease [19].

It should be noted that neither GNB3 nor GNGT2 (nor any gene for which negative exon screening results have been obtained) can be definitively ruled out as the site of mutations responsible for disease. This is because: (1) the SSCP or DGGE techniques may not detect all sequence variants although this would be a very small percentage; (2) mutations may be present in 5', 3' or intronic sequences not screened; and (3) mutations in this gene may be rare and/or present only in specific diseases such that not enough patients of a particular disease type were screened to detect them. Nevertheless, based on our data, it is unlikely that either the GNB3 or GNGT2 genes are common sites of mutations responsible for the inherited retinal diseases of cone dystrophy, cone-rod dystrophy or the (non-age related) macular dystrophies that we tested.


This research was supported by grants from the Foundation Fighting Blindness (DBF, MD, RGW and SGJ), from the George Gund Foundation (DBF and MD), the NIH (DBF, EY08285 and EY00331; SGJ, EY05627), and a Research to Prevent Blindness Unrestricted Grant (RGW and DBF). DBF is the recipient of a Research to Prevent Blindness Senior Scientific Investigators Award.

We thank Dr. Olivia Ong for providing the genomic sequence of the GNGT2 gene before publication.


1. Aguirre GD, Rubin LF. Pathology of hemeralopia in the Alaskan malamute dog. Invest Ophthalmol 1974; 13:231-235.

2. Gropp KE, Szel A, Huang JC, Acland GM, Farber DB, Aguirre GD. Selective absence of cone outer segment beta 3-transducin immunoreactivity in hereditary cone degeneration (cd). Exp Eye Res 1996; 63:285-296.

3. Akhmedov NB, Piriev NI, Pearce-Kelling S, Acland GM, Aguirre GD, Farber DB. Canine cone transducin-gamma gene and cone degeneration in the cd dog. Invest Ophthalmol Vis Sci 1998; 39:1775-1781.

4. Akhmedov NB, Piriev NI, Ray K, Acland GM, Aguirre GD, Farber DB. Structure and analysis of the transducin beta3-subunit gene, a candidate for inherited cone degeneration (cd) in the dog. Gene 1997; 194:47-56.

5. Gal A, Apfelstedt-Sylla E, Janecke AR, Zrenner E. Rhodopsin mutations in inherited retinal dystrophies and dysfunctions. In: Osborne NN, Chader GJ, editors. Progress in Retinal and Eye Research. Vol 16. Great Britain: Elsevier Science; 1997. p. 51-79.

6. Dryja TP, Hahn LB, Reboul T, Arnaud B. Missense mutation in the gene encoding the alpha subunit of rod transducin in the Nougaret form of congenital stationary night blindness. Nat Genet 1996; 13:358-360.

7. Huang SH, Pittler SJ, Huang X, Oliveira L, Berson EL, Dryja TP. Autosomal recessive retinitis pigmentosa caused by mutations in the alpha subunit of rod cGMP phosphodiesterase. Nat Genet 1995; 11:468-471.

8. McLaughlin ME, Ehrhart TL, Berson EL, Dryja TP. Mutation spectrum of the gene encoding the beta subunit of rodphosphodiesterase among patients with autosomal recessive retinitis pigmentosa. Proc Natl Acad Sci U S A 1995; 92:3249-3253.

9. Danciger M, Heilbron V, Gao YQ, Zhao DY, Jacobson SG, Farber DB. A homozygous PDE6B mutation in a family with autosomal recessive retinitis pigmentosa. Mol Vis 1996; 2:10 <>.

10. Bayes M, Giordano M, Balcells S, Grinberg D, Vilageliu L, Martinez I, Ayuso C, Benitez J, Ramos-Arroyo MA, Chivelet P, et al. Homozygous tandem duplication within the gene encoding the beta-subunit of rod phosphodiesterase as a cause for autosomal recessive retinitis pigmentosa. Hum Mutat 1995; 5:228-234.

11. Fuchs S, Nakazawa M, Maw M, Tamai M, Oguchi Y, Gal A. A homozygous 1-base pair deletion in the arrestin gene is a frequent cause of Oguchi disease in Japanese. Nat Genet 1995; 10:360-362.

12. Farber DB, Danciger M. Identification of genes causing photoreceptor degenerations leading to blindness. Curr Opin Neurobiol 1997; 7:666-673.

13. Levine MA, Smallwood PM, Moen PT Jr, Helman LJ, Ahn TG. Molecular cloning of beta 3 subunit, a third form of the G protein beta-subunit polypeptide. Proc Natl Acad Sci U S A 1990; 87:2329-2333.

14. Ansari-Lari MA, Shen Y, Muzny DM, Lee W, Gibbs RA. Large-scale sequencing in human chromosome 12p13: experimental and computational gene structure determination. Genome Res 1997; 7:268-280. 9074930

15. Ong C, Hu K, Rong H, Lee RH, Fung BK. Gene structure and chromosome localization of the G gamma c subunit of human cone G-protein (GNGT2). Genomics 1997; 44:101-109.

16. Weleber RG, Eisner A. Cone degeneration and color vision defects. In: Newsome DA, editor. Retinal Dystrophies and Degenerations. New York; Raven Press; 1988. p. 233-256.

17. Gao YQ, Danciger M, Zhao DY, Blaney J, Piriev NI, Shih J, Jacobson SG, Heckenlively JH, Farber DB. Screening of the PDE6B gene in patients with autosomal dominant retinitis pigmentosa. Exp Eye Res 1996; 62:149-154.

18. Levine MA, Modi WS, O'Brien SJ. Chromosomal localization of the genes encoding two forms of the G protein beta polypeptide, beta 1 and beta 3, in man. Genomics 1990; 8:380-386.

19. Magovcevic I, Weremowicz S, Morton CC, Fong SL, Berson EL, Dryja TP. Mapping of the human cone transducin alpha-subunit gene (GNAT2) to 1p13 and negative mutation analysis in patients with Stargardt disease. Genomics 1995; 25:288-290.

Gao, Mol Vis 1998; 4:16 <>
©1998 Molecular Vision <>
ISSN 1090-0535