Constitutively Active Mutant of Rod A-Transducin in Autosomal Dominant

HUMAN MUTATION Mutation in Brief #970 (2007) Online

MUTATION IN BRIEF p.Gln200Glu, a Putative Constitutively Active Mutant of Rod α-Transducin (GNAT1) in Autosomal Dominant Congenital Stationary Night Blindness Viktoria Szabo1,2†, Hans-Jürgen Kreienkamp1†, Thomas Rosenberg3, and Andreas Gal1*

1Institut für Humangenetik, Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany; 3Gordon Norrie Centre for Genetic Eye Diseases, The National Eye Clinic for the Visually Impaired, Hellerup, Denmark; 2Permanent address: Department of Ophthalmology, Semmelweis University, Budapest, Hungary

*Correspondence to: A. Gal, Institut für Humangenetik, Universitätsklinikum Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany; Tel.: 49-40-42803-2120; Fax: 49-40-42803-5138; E-mail: [email protected]

Grant sponsor: Recognition Award of the Alcon Research Institute (Fort Worth, TX, to A.G.) and the German Academic Exchange Service (to V.S.).

†Viktoria Szabo and Hans-Jürgen Kreienkamp contributed equally to this work.

Communicated by Mark H. Paalman

Congenital stationary night blindness (CSNB) is a non-progressive Mendelian condition resulting from a functional defect in rod photoreceptors. A small number of unique missense mutations in the genes encoding various members of the rod phototransduction cascade, e.g. rhodopsin (RHO), cGMP phosphodiesterase β-subunit (PDE6B), and transducin α-subunit (GNAT1) have been reported to cause autosomal dominant (ad) CSNB. While the RHO and PDE6B mutations result in constitutively active proteins, the only known adCSNB-associated GNAT1 change (p.Gly38Asp) produces an α-transducin that is unable to activate its downstream effector molecule in vitro. In a multigeneration Danish family with adCSNB, we identified a novel heterozygous C to G transversion (c.598C>G) in exon 6 of GNAT1 that should result in a p.Gln200Glu substitution in the evolutionarily highly conserved Switch 2 region of α-transducin, a domain that has an important role in binding and hydrolyzing GTP. Computer modeling based on the known crystal structure of transducin suggests that the p.Gln200Glu mutant exhibits impaired GTPase activity, and thereby leads to constitutive activation of phototransduction. This assumption is in line with our results of trypsin protection assays as well as previously published biochemical data on mutants of this glutamine in the GTPase active site of α-transducin following in vitro expression, and observations that inappropriately activating mutants of various members of the rod phototransduction cascade represent one of the major molecular causes of adCSNB. © 2007 Wiley-Liss, Inc.

KEY WORDS: night blindness; CSNB; transducin; GNAT1; constitutive activation; phototransduction

INTRODUCTION Night blindness, reduced or absent dark adaptation, is a typical and early sign of various forms of retinal dystrophies. While night blindness as disease symptom is usually progressive and parallels the disintegration of rod photoreceptors, congenital stationary night blindness (CSNB; e.g. MIM# 163500) is a rare, non-progressive Mendelian condition due to a functional disorder of rod photoreceptors. CSNB is heterogeneous as to the mode of

Received 7 July 2006; accepted revised manuscript 12 April 2007.

2 Szabo et al. inheritance (autosomal dominant [ad], autosomal recessive, and X-linked), pattern of electroretinogram (Riggs or Schubert-Bornschein types), refractive error, and fundus appearance in the probands (for a recent review see Dryja 2000). Several large families with adCSNB have been documented (Carr 1974, al-Jandal et al. 1999), including the French Nougaret genealogy first described 1838 (Cunier 1838) and reinvestigated in 1907 (Nettleship 1907), and an extended Danish pedigree first published 1909 by Rambusch (Rambusch 1909) and rediscovered 1991 (Rosenberg et al. 1991). Affected family members in these two pedigrees present with identical electrophysiolo- gical and psychophysical findings. At the molecular level, two basic mechanisms of opposite nature have been suggested to explain the dominant phenotype; constitutive activation of various members of the rod phototransduction cascade, or inability to activate the cascade. Abnormally prolonged activation of photosignaling may mimic a weak background light, desensitizing the visual system. Dryja et al. (1993) suggested that a heterozygous p.Ala292Glu change of rod opsin resulted in an inappropriate continuous activity of the mutant protein, even without chromophore, in a patient with clinical features compatible with CSNB. Two further missense mutations, p.Thr94Ile of the gene (RHO) encoding rhodopsin (al-Jandal et al. 1999), and p.His258Asn of the rod cGMP phosphodiesterase β-subunit gene (PDE6B; Gal et al. 1994) were suggested to result in constitutive activation of the mutant protein and continuous phototransduction in two families with adCSNB. In contrast, a heterozygous missense mutation (p.Gly38Asp) in GNAT1 (MIM# 139330) (Dryja et al. 1996), identified in the Nougaret family, was shown to result in inability of the mutant α-transducin to activate its downstream effector, rod cGMP-specific phosphodiesterase (PDE). In the light of these findings, the question arises whether adCSNB-associated transducin mutations differ principally in their mode of action from those of other genes. Here we report a novel heterozygous missense mutation of the α- transducin gene that is likely to result in constitutive activation of phototransduction.

MATERIALS AND METHODS The Danish pedigree studied here consists of 9 family members presenting with typical symptoms of CSNB in three generations suggesting an autosomal dominant trait (Fig. 1). In addition to standard ophthalmologic examination, dynamic visual field measurement with a Goldmann apparatus (object size I/4e and IV/4e), colour vision testing (Ishihara, AOHRR, Farnsworth D-15, and Nagel anomaloscope), dark adaptometry a.m. Goldmann- Weekers with a diffuse light stimulus (integral technique), and full-field ERG following the ISCERG recommen- dations were performed. Informed consent was obtained from all human subjects included in this study. For linkage analysis, fluorescence-tagged microsatellite markers were analysed on an ABI PRISM 310 automated DNA sequencer (Applied Biosystems, Forster City, USA). For GNAT1 mutation numbering GenBank reference cDNA sequence NM_144499.1 was used with +1 corresponding to the A of the ATG translation initiation codon, which is codon 1. Protein expression. A His6-tagged expression clone for the transducin/Gai-chimera (chimera 8; Skiba et al., 1996) was kindly provided by Dr. H. Hamm (Vanderbilt University, USA). The p.Q200E mutation was introduced by site-directed mutagenesis, using appropriate PCR primers. For expression in bacteria, plasmids were transformed into BL21 cells; cells were cultured in terrific broth media including phosphate salts. After induction with IPTG overnight at 16°C, cells were lysed and recombinant proteins were purified essentially as described by Skiba et al. (1996). After elution from the Ni-NTA agarose with 100 mM imidazole, purified protein was dialyzed against 50 mM Tris-HCl; pH 8.0; 50 mM NaCl; 5 mM MgCl2 containing 50 μM GDP and stored at -80°C until further use. Trypsin protection assay. Purified protein was diluted to a final concentration of 0.1 mg/ml in 20 mM - Tris/HCl (pH 8.0) and 50 μM GDP. The AlF4 complex was generated by adding 1 mM NaF and 50 μM AlCl3. After incubation at room temperature for 60 min, trypsin was added to a final concentration of 0.01 mg/ml. Samples were incubated for 10 min at room temperature, followed by boiling in SDS sample buffer and SDS- polyacrylamide electrophoresis. Protein bands were visualized by staining with Coomassie Brilliant Blue. Molecular modeling. Molecular modeling was performed using the SWISS-Model server (Guex and Peitsch, 1997) at http://swissmodel.expasy.org. Modeling was based on the transducin/Gαi-chimera in complex with PDE- - - γ, RGS9 and GDP.AlF4 (Slep et al., 2001; template file: 1fqjA.pdb), and also on the isolated transducin/GDP.AlF4 -complex (template file 1tadA.pdb; Sondek et al., 1994). Structures were visualized using the PDBviewer software (Guex and Peitsch, 1997). p.Gln200Glu in Rod α-Transducin 3

RESULTS AND DISCUSSION Of the nine night blind family members, five underwent detailed clinical investigations. All but one (II-2, see later) had experienced non-progressive night blindness from early infancy. The latter four family members (II-4, III-2, III-4, and III-7, aged, respectively, 53, 37, 31, and 25 years at the time of first clinical assessment in 1991) had normal visual acuity and colour vision, visual fields were unconstricted, and the fundi unremarkable. Refrac- tive values were –8.50 and –7.00 (spherical equivalent) in subject III-2, and –2.00 and –3.50 in subject III-4. Dark adaptometry confirmed that the three of them were completely night blind with an appr. 100-fold reduction in rod sensitivity. In the youngest subject, however, an initial cone phase of 2.0 log units was succeeded by a small rod deflection and additional 0.6 log increase in sensitivity. Scotopic electroretinograms (ERG) showed an absent rod b-wave in response to dim flashes of light and a cone-like response only to bright flashes. In light-adapted state, cone responses of normal-to-moderately decreased amplitudes and normal implicit time were found. This ERG pattern (Riggs-type ERG) suggests a presynaptic defect in rod phototransduction. Re-examination of two of the night-blind individuals (III-2 and III-4) 20 years after the initial examination (i.e. at age of 57 and 51 years) confirmed that the condition was non-progressive, except for mild changes induced by high myopia in one of them. One family member (II-2) had constricted visual fields, peripheral hyperpigmentation of the retina, and an extinguished ERG, typical symptoms of a degenerative retinopathy collectively termed retinitis pigmentosa (RP).

Figure 1. Pedigree of a Danish family (top part) in which the phenotype of an autosomal dominant congenital stationary night blindness is co-segregating with the heterozygous p.Gln200Glu mutation of GNAT1. The c.598C>G mutation can be detected by the restriction enzyme TseI (below the pedigree). The wild-type amplicon (394 bp) contains two TseI sites giving rise to fragments of 155, 128, and 111 bp. Due to the loss of one recognition site, an additional 283 bp fragment appears for persons carrying the c.598C>G mutation in heterozygous state. Lane 1: undigested product; lanes 2 and 3: Probands II-2 and III-4, respectively; lane 4: III-3; M: DNA molecular weight marker.

Using microsatellite markers corresponding to the loci D3S1263, D4S412, and D3S1289, that map less than 5 Mb apart from the three known adCSNB loci, RHO, PDE6B, and GNAT1, respectively, we genotyped a total of 16 family members. For the linkage relationships disease locus vs RHO or vs PDE6B, two-point analyses (MLINK program) yielded negative lod scores at all recombination fraction values (data not shown), making it unlikely that either of these genes is implicated in this trait. However, D3S1289, the locus close to GNAT1, that was informative only in one branch of the family, gave positive lod scores of Zmax=1.81 at θ= 0.00. Direct sequencing of all PCR- 4 Szabo et al. amplified exons and flanking intronic regions of GNAT1 detected a heterozygous C to G transversion of the first nucleotide (c.598C>G) in codon 200 in exon 6 in the DNA samples of probands II-2 and II-4 (results not shown), predicting a p.Gln200Glu (CAG>GAG) change in the protein. The mutation results in loss of a TseI site (Fig. 1) and restriction enzyme analysis of the family showed that the c.598C>G mutation was present in all affected family members in heterozygous state (data not shown). Two-point linkage analysis for disease locus vs C>G produced a lod score of Zmax=3.01 at θ= 0.00 assuming complete penetrance and a frequency of 0.0001 for the mutant allele. The c.598C>G change was absent from 104 alleles of unaffected and unrelated controls. Person II-2 is affected by retinitis pigmentosa, a degenerative retinopathy. He is heterozygous for the family- specific p.Gln200Glu GNAT1 mutation suggesting that the early onset night blindness was due to the GNAT1 mutation. No other change was detected by direct sequencing of the entire coding region of the gene. It is likely that the simultaneous occurrence of RP and CSNB is coincidental. This conclusion is supported by the fact that the p.Gln200Glu mutant allele of GNAT1 of person II-2 was inherited by his daughter who had typical signs of CSNB but no symptoms of RP at the age of 57 years. Rod transducin is a heterotrimeric (Gαβγ) G-protein complex that mediates signalling from the upstream receptor rhodopsin to the downstream effector rod cGMP-specific phosphodiesterase (PDE) in the phototransduction pathway. Light-activated rhodopsin binds to transducin α-subunit and catalyses the exchange of GDP for GTP, which enables Gα to dissociate from the Gβγ-subunits and rhodopsin, and to remove the inhibitory γ-subunit of cGMP-PDE. Activated PDEαβ hydrolyses cGMP whereas decreased intracellular cGMP levels lead to closure of cGMP-gated cation channels, membrane hyperpolarization, and photoreceptor signalling. The reaction is turned off as a result of GTP hydrolysis and reformation of the transducin α-GDP/βγ complex. In addition to the intrinsic GTPase activity of the activated α-transducin subunit itself, interaction of α–transducin with the rod-specific RGS9 protein and the γ-subunit of PDE enhances GTP hydrolysis and thereby accelerates recovery of rod photoreceptor active state (Slepak et al. 1995, Cowan et al. 2001, Slep et al. 2001).

Figure 2. Structure of the GTPase catalytic center of the transducin/Gαi chimera used to analyze the Gα/RGS9/PDEγ - heterotrimeric complex by Slep et al. (2001). A: The transition state analog GDP.AlF4 is shown to make close contact to the - catalytic Q200 residue; AlF4 mimics the planar conformation of the γ-phosphate group during hydrolysis, which is stabilized by interacting with the amide of Q200 (arrow). By homology modeling of the mutant transducin/Gαi, the E200 and L200 residues (panels B and C, respectively) were incorporated into this structure. Note that both E200 and L200 move away from the catalytic site, most likely due to electrostatic repulsion in case of E200, explaining GTPase deficiency and constitutive - 2+ activation of the mutant transducin. Blue, nitrogen; red, oxygen; yellow, phosphorus. AlF4 and the Mg ion associated with the - GDP.AlF4 complex are shown in different shades of gray.

There are three related families of GTP-binding proteins. Although members of each of these families perform different cellular functions, the function of each depends critically on an intrinsic GTP-driven conformational switch, combined with an intrinsic GTPase activity. The so-called Switch II domain of G-proteins plays an important role in binding and hydrolyzing GTP. In particular, the glutamine-200 residue that is evolutionarily highly conserved in the various G-protein α-subunits, has been consistently implicated as being important for this p.Gln200Glu in Rod α-Transducin 5 reaction. Some indication of the functional consequences of the Gln to Glu change might be gleaned from the analysis of the (distantly) related small GTPase ras oncogen, the crystal structure of which was defined. Seventeen different point mutations were introduced in the human HRAS gene by site-directed mutagenesis to replace the glutamine-200 residue (in the corresponding position, Gln-61), including glutamic acid (Der et al. 1986). Each mutant protein hydrolysed GTP at approximately one-eighth the rate of the normal HRAS protein, suggesting that this glutamine represents a strong conformation-determining region of the protein with a major effect on its ability to hydrolyse GTP. Molecular modeling of the p.Q200E mutation was performed within the context of the known three dimensional - structure of a transducin/Gαi chimera in complex with the phosphodiesterase γ−subunit, RGS9 and GDP.AlF4 , a transition state analog for GTP hydrolysis (Fig. 2) (Slep et al., 2001). For comparison, we included another Gln- 200 mutant, p. Q200L, first described and analysed by Srinivasan et al. (1999). The p.Q200L mutant has recently been extensively studied by Majumdar et al. (2006). These authors have shown that this mutant exhibited severely compromised rates of GTP hydrolysis and, consequently, was capable of stimulating constitutively PDE activity. The structure predicted by the SWISS-MODEL server indicates that the p.Q200E mutation does not induce gross perturbations in the transducin structure; the major effect of the mutation is a relative move of the glutamate-200 side chain out of the GTPase active site into a position where contact with the tetrafluoride complex is impossible, most likely due to electrostatic repulsion. A similar movement of the Glu-200 side chain was observed when - modeling was performed within the context of the isolated transducin α-subunit in complex with GDP.AlF4 (structure solved by Sondek et al., 1994; not shown). The effect of the p.Q200E mutation is most similar to that of the p.Q200L mutation (Fig. 2). This latter mutant lacks any detectable GTPase activity, in line with the view that the carbonyl of Gln-200 (which is still present in Glu-200) is involved in positioning the hydrolytic water molecule. In addition, the amide nitrogen of Gln-200 (missing in Glu-200) is required to support the planar - transition state for GTP hydrolysis, which is mimicked by the GDP.AlF4 complex. On a biochemical level, the similarity between the p.Q200L and p.Q200E mutant transducin molecules becomes apparent in trypsin protection - assays. Transducin is readily degraded by trypsin; however, after addition of GDP.AlF4 , a conformational change occurs which interferes with tryptic cleavage. Both the p.Q200E mutant transducin (reported here) and the - p.Q200L mutant are insensitive to GDP.AlF4 (Fig. 3; and see Fig. 4 in Majumdar et al., 2006), suggesting that the - GDP.AlF4 complex cannot be formed properly. Recent studies have suggested that in addition to the lack of hydrolytic activity, the extended lifetime of the α-subunit/GDP.P complex is also likely to contribute to the (constitutive) activation of PDE (Majumdar et al. 2006). In addition, analysis of transgenic GtαQ200L mice showed that the return of the mutant transducin to the rod outer segments (from the inner segments and synaptic terminals) during dark adaptation is markedly slower than in wild-type animals (Kerov et al. 2005).

Figure 3. Analysis of proteolytic cleavage of α−transducin in the presence of AlF4. Wild-type and p.Gln200Glu mutant His6-transducin/Gαi-chimera were pretreated for 60 min with GDP and - with or without AlF4 , as indicated. After treatment with trypsin, samples were boiled in SDS sample buffer, separated by SDS-PAGE and stained with Coomassie brilliant Blue. Molecular weight markers (in kDa) are shown on the left. Note the appearance of a protected double band at approx. 32 kDa in the wild type (left panel), but not the p.Gln200Glu mutant (right panel) α−transducin.

Remarkably, the mutation p.G38D in transducin also causes adCSNB. Residue Gly-38 is also in the vicinity of - the active site, and introduction of aspartate at this position abolishes the AlF4 -effect on tryptic cleavage, similar to the p.Q200E mutation. However, the p.G38D mutant transducin exhibits only slightly reduced GTPase activity. 6 Szabo et al.

Instead, the interaction with the effector molecule PDEγ is interrupted, leading to attenuated visual signaling (Muradov and Artemyev, 2000). Modeling of the p.G38D mutant indicated a perturbation of the structure surrounding the PDEγ binding site (Moussaif et al., 2006). We did not observe such a perturbation; however, our modeling experiments cannot be compared directly to those of Moussaif et al. (2006), as their data were obtained on the basis of the homologous G44V/Gαi mutant. In conclusion, our modeling and biochemical data suggest that the mutation p.Q200E most likely interferes with the GTPase activity of the transducin α-subunit. Thus, unique mutations of genes encoding various members of the rod phototransduction cascade seem to be the major cause of adCSNB. The allelic and nonallelic genetic heterogeneity is largely simplified at the level of the molecular mechanism leading to adCSNB which turns out to be, in the majority of cases, a constitutive (inappropriate) activation of the visual cascade.

REFERENCES al-Jandal N, Farrar GJ, Kiang AS, Humphries MM, Bannon N, Findlay JB, Humphries P, Kenna PF. 1999. A novel mutation within the rhodopsin gene (Thr-94-Ile) causing autosomal dominant congenital stationary night blindness. Hum Mutat 13:75-81.

Carr RE. 1974. Congenital stationary nightblindness. Trans Am Ophthalmol Soc 72:448-487.

Cowan CW, He W, Wensel TG. 2001. RGS proteins: lessons from the RGS9 subfamily. Prog Nucleic Acid Res Mol Biol 65:341-359.

Cunier F. 1838. Histoire d'une héméralopie héréditaire depuis deux siècles dans une famille de la commune de Vendémian, près Montpellier. Ann Soc Médicin de Gand 4:385-395.

Der CJ, Finkel T, Cooper GM. 1986. Biological and biochemical properties of human rasH genes mutated at codon 61. Cell 44:167-176.

Dryja TP. 2000. Molecular genetics of Oguchi disease, fundus albipunctatus, and other forms of stationary night blindness: LVII Edward Jackson Memorial lecture. Am J Ophthalmol 130:547-563.

Dryja TP, Berson EL, Rao VR, Oprian DD. 1993. Heterozygous missense mutation in the rhodopsin gene as a cause of congenital stationary night blindness. Nat Genet 4:280-283.

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

Gal A, Orth U, Baehr W, Schwinger E, Rosenberg T. 1994. Heterozygous missense mutation in the rod cGMP phosphodiesterase beta-subunit gene in autosomal dominant stationary night blindness. Nat Genet 7:64-68.

Guex N, Peitsch MC (1997). SWISS-MODEL and the Swiss-Pdb viewer: An environment for comparative protein modelling. Electrophoresis 18:2714-2723.

Kerov V, Chen D, Moussaif M, Chen Y-J, Artemyev NO. 2005. Transducin activation state controls its light-dependent translocation in rod photoreceptors. J Biol Chem 280:41069-41076.

Majumdar S, Ramachandran S, Cerione RA. 2006. New insights into the role of conserved, essential residues in the GTP binding/GTP hydrolytic cycle of large G proteins. J Biol Chem 281:9219-9226.

Moussaif M, Rubin WW, Kerov V, Reh R, Chen D, Lem J, Chen C-K, Hurley JB, Burns ME, Artemyev NO. 2006. Phototransduction in a transgenic mouse model of Nougaret night blindness. J Neurosci 26:6863-6872.

Nettleship E. 1907. A history of congenital stationary nightblindness in nine consecutive generations. Trans Ophthalmol Soc UK 27:269-293.

Rambusch SHA. 1909. Den medfodte Natteblindheds Arvelighedsforhold. Oversigt over det Kgl. Danske Videnskabernes Selskabs Forhandlinger 3:337-347.

Rosenberg T, Haim M, Piczenik Y, Simonsen SE. 1991. Autosomal dominant stationary night-blindness. A large family rediscovered. Acta Ophthalmol (Copenh) 69:694-702. p.Gln200Glu in Rod α-Transducin 7

Skiba NP, Bae H, Hamm HE. 1996. Mapping of Effector Binding Sites of Transducin α-Subunit Using Gαt/Gαi1 Chimeras. J Biol Chem 271:413-424.

Slep KC, Kercher MA, He W, Cowan CW, Wensel TG, Sigler PB. 2001. Structural determinants for regulation of phosphodiesterase by a G protein at 2.0 A. Nature 409:1071-1077.

Slepak VZ, Artemyev NO, Zhu Y, Dumke CL, Sabacan L, Sondek J, Hamm HE, Bownds MD, Arshavsky VY. 1995. An effector site that stimulates G-protein GTPase in photoreceptors. J Biol Chem 270:14319-14324.

Sondek J, Lambright DG, Noel JP, Hamm HE, Sigler PB. 1994. GTPase mechanism of G-proteins from the 1.7-A crystal structure of transducin α-GDP-AlF-4. Nature 372:276-279.

Srinivasan J, Gundersen RE, Hadwiger JA. 1999. Activated Gα subunits can inhibit multiple signal transduction pathways during Dictyostelium development. Dev Biol 215:443-452.