- Email: [email protected]
Missense mutants inactivate guanosine triphosphate cyclohydrolase I in hereditary progressive dystonia
Brain & Development 22 (2000) S111±S114
www.elsevier.com/locate/braindev
Original article
Missense mutants inactivate guanosine triphosphate cyclohydrolase I in hereditary progressive dystonia Satoshi Ueno*, Makito Hirano Department of Medical Genetics, Nara Medical University, Shijo-cho 840, Kashihara, Nara 634-8521, Japan
Abstract Hereditary progressive dystonia (HPD) with marked diurnal ¯uctuation is caused by mutant guanosine triphosphate (GTP) cyclohydrolase I (GCH). The clinical presentation of dominant HPD varies considerably. We proposed the hypothesis that a relative increase of mutant GCH capable of inhibiting normal GCH is responsible for heterogeneous phenotypic manifestations. In a Japanese family with a novel G90V mutation, an affected heterozygote had a higher mutant/normal mRNA ratio than an unaffected heterozygote. Co-expression analysis showed that mutant enzyme (GCH-G90V) inactivated the normal enzyme in the COS cells. Similarly, GCH-G203R showed the dominant negative effects. These results supported our proposed hypothesis. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Hereditary progressive dystonia; Guanosine triphosphate cyclohydrolase I; Missense mutation; Dominant negative effect
1. Introduction
2. Materials and methods
Hereditary progressive dystonia (HPD) with marked diurnal ¯uctuation/Dopa-responsive dystonia (DRD) is a disease characterized by childhood-onset dystonia that remarkably responds to small doses of levodopa [1,2]. Recently, several workers have demonstrated that HPD in some cases is caused by mutations in the gene encoding guanosine triphosphate (GTP) cyclohydrolase I (GCH) which is the ®rst enzyme in the pathway converting GTP to tetrahydrobiopterin, an essential cofactor of tyrosine hydroxylase [3±7]. HPD is a dominantly inherited disorder with the considerable diversity of clinical phenotypic expression [1,2]. We have previously proposed that a relative increase in the mutant product capable of inhibiting the wild-type gene product is responsible for the clinical phenotypic diversity [4,7]. This prediction was in part supported by the results of our recent study of truncated GCH mutants [8]. To expand this hypothesis to include missense mutations, we studied a novel missense mutation of the GCH gene in a Japanese family with HPD. Another GCH mutant with a substitution of arginine for glycine at position 203 (G203R), which was identi®ed originally in a British case [6] and recently in a Japanese case, was subjected to the present study.
2.1. Subjects
* Corresponding author. Tel./Fax: 181-744-22-9684. E-mail address: [email protected] (S. Ueno).
A 25-year-old woman began to notice tremor at rest in the upper extremities at age 11. She developed a gait disturbance due to dystonia in all limbs and wry neck at age 12. These symptoms worsened toward the evening, and were alleviated after sleep. Small doses of levodopa produced marked bene®cial effects on the patient, however, dyskinesia appeared in the neck with the administration of levodopa. The dystonia was relieved by trihexyphenidyl hydrochloride without dyskinesia. The blood samples were collected from the patient and her unaffected parents and healthy controls for this study after obtaining informed consent. 2.2. Sequencing of the GCH gene All exons and exon±intron junctions of the GCH gene were ampli®ed by PCR and sequenced as described previously [4,7,8]. 2.3. Quanti®cation of lymphocyte GCH mRNA Total RNA from peripheral blood lymphocytes was reverse-transcribed, and synthesized cDNA encoding GCH were quantitatively ampli®ed by PCR [4,7,8].
0387-7604/00/$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0387-760 4(00)00135-2
S112
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
2.4. Preparation of GCH-G203R cDNA Genomic DNA of another Japanese patient with typical HPD was available for the GCH gene analysis. The determined sequence contained a G to A change in exon 5, resulting in the G203R mutation. This mutation was ®rst reported in a British DRD case [6]. GCH-G203R cDNA was synthesized by PCR-based mutagenesis [9], and inserted into an expression vector pcDNA3 for the present study. 2.5. Expression of GCH in transfected COS cells The COS-7 cells were transfected with the cDNA for wild-type GCH (GCH-WT), the newly recognized mutant GCH, or GCH-G203R in the expression vector pcDNA3. In cotransfection experiments, the cells were transfected with the cloned wild-type and mutant GCHcDNA mixed at various ratios using Lipofectamine (GIBCO laboratories, Grand Island, NY). The cells were analyzed for the expression of GCH mRNA, enzyme protein, and enzyme activity according to the same method as that described in our previous paper [8]. 3. Results 3.1. Mutation and expression of the GCH gene in lymphocytes The GCH gene from the patient contained a novel G to T transversion in exon 1 (Fig. 1A), leading to a substitution of valine for glycine at position 90 (G90V). The BsaAI-restriction site analysis showed that the patient and her mother were heterozygous for this mutation, while her father and 100 control subjects had only the normal sequence (not shown).
The ratio of mutant/normal GCH mRNA in the patient's lymphocytes was 1:54 ^ 0:04 (mean ^ SD), signi®cantly higher than that in the mother's, 1:00 ^ 0:06 (Fig. 1B). 3.2. GCH expression in transfected COS cells GCH mRNA was not detected in non-transfected COS-7 cells but was present in cells transfected to express wildtype and mutant GCH (Fig. 2). The enzyme activity for GCH-G90V in the COS cells is a ,12% of control enzyme activity, and for GCH-G203R, ,6%. GCH protein was not detected in the control COS-7 cell extracts, whereas in the transfected COS cells with GCH-WT, GCH-G90V or GCHG203R cDNA, the protein was detected with an apparent molecular weight (28 kD) for the monomeric form of the normal enzyme. The expression level of GCH-G203R was lower than GCH-WT and GCH-G90V. 3.3. Inhibition of normal GCH enzyme activity by mutant GCH The co-expression of the mutant with wild-type GCH in COS cells showed that the two mutant GCHs inactivated the normal enzyme in a dose-dependent manner. GCH activity was determined with COS-7 cells cotransfected with GCHWT cDNA and mutant GCH cDNA in various ratios. When COS cells were cotransfected with the expression plasmid for GCH-WT and the plasmid for GCH-G90V or GCHG203R in the 1:1 ratio, the enzyme activity decreased to 70 and 62% of control, respectively (Fig. 3). This effect was clearly speci®c, since the co-expression of GCH with unrelated proteins including Cu/Zn superoxide dismutase or chloramphenicol acetyl transferase cDNA had no effect on the enzyme activity (data not shown).
Fig. 1. (A) Direct sequence analysis of ampli®ed genomic DNA from a control and the patient with dopa-responsive dystonia. The patient is heterozygous for a G to T transversion. (B) Quantitative reverse-transcribed polymerase chain reaction analysis of normal and mutant GCH mRNA was done with primers R1: 5 0 CTGGGCGAGAACCCCCAGCGGCA-3 0 , and E1mF: 5 0 -CTGGGCGAGAACCCCCAGCGGCACG-3 0 carrying a substitution of C for A. This substitution creates a new BsuAI restriction site in the mutant allele, thereby the mutant fragment can be separated from normal ones. The expressed ratio of mutant/normal GCH mRNA is higher in the patient than in her asymptomatic mother. Quanti®cation was carried out with the normal and mutant fragments ampli®ed by ®ve separate RT-PCR ampli®cations. (*P , 0:05, repeated analysis of variance). Autoradiographs (inset) show normal (upper band) and mutant (lower band) GCH mRNA.
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
Fig. 2. The activity of COS-7 cells transfected with GCH-WT cDNA was 0:67 ^ 0:03 nmol/h per mg protein (mean ^ SD), while the native COS-7 cells showed no enzyme activity. The activities in COS-7 cells with mutant GCH expression were 0:075 ^ 0:002 nmol/h per mg protein for GCHG90V and 0:032 ^ 0:002 nmol/h per mg protein for GCH-G203R. (*P , 0:05, repeated analysis of variance). Assays were done in quadruplicate. In the middle panels, reverse-transcribed polymerase chain reaction shows that GCH mRNA was not detected in non-transfected COS-7 cells, but was present in the cells transfected with wild-type GCH (GCH-WT) cDNA and mutant GCH cDNAs. Immunoblotting demonstrates that GCH subunit was expressed in the COS-7 cells transfected with the expression vector for wild-type and mutant GCH. The expression level of GCH-G203R was lower than those of other two transfected cells.
4. Discussion We have characterized the mutant GCH gene in a new Japanese family with HPD, where a G to T transversion in exon 1 caused a G90V substitution (Fig. 1A). Both our patient with HPD and her unaffected mother are heterozygous for this mutation. The relative expression of mutant GCH mRNA was higher in the patient than in her mother. This observed ®nding within the family, which is similar to those in other HPD families previously reported by us, is in agreement with our proposed hypothesis that the mutant/ normal GCH mRNA ratio may contribute to the phenotypic variation of HPD [4,7,8]. At the present time, we can not exclude the possibility that heteroduplex formation between wild-type and mutant GCH cDNAs during the PCR may affect the quantitation with the possibility that the mutant cDNA is underestimated. It was interesting to note that this patient showed resting tremor in the upper limbs and dopainduced dyskinesia, which were previously unreported
S113
symptoms in genetically proven HPD. This suggested that the clinical phenotype of HPD was far more variable than previously thought [1,2]. The in vitro transfection experiments showed that the wild-type and mutant GCH proteins were expressed in the COS cells. The levels of GCH enzyme activity for GCHG90V and GCH-G203R were ,12% and ,6% of control activity, respectively. The two mutants were examined for their ability to inactivate the native enzyme using the cotransfection assay. It was demonstrated that GCH-G90V and GCH-G203R mutants inactivated the normal enzyme, and GCH-G203R was more effective in reducing the enzyme activity than GCH-G90V. This fact indicated that the dominant negative effect was mutant-dependent. Similar ®ndings were obtained in the study of the ClC-1 channel gene; different missense mutations showed different magnitudes of dominant negative effects on the normal channel [10]. Immunoblotting experiments with the transfected COS cell extracts showed that, when GCH-G90V and GCHG203R were expressed individually, each displayed a molecular mass (28 kD) for a normal GCH subunit. However, the expression of GCH-G203R was reduced compared with GCH-WT and GCH-G90V. This reduction may be in part due to the fact that the G203R mutation, a non-conservative amino acid change, causes the conformational changes of the GCH molecule, leading to aberrant processing, or to premature degradation of the GCH. The results of the in vitro experimentation may in part account for clinical phenotypic variation within the same family members. We have shown that, in the HPD family associated with GCH-G90V, the mutant/normal GCH mRNA ratio is 1.54 for the affected carrier, and 1.00, for the unaffected carrier. Fig. 3A showed that this difference corresponded to a ,10% decrease in the enzyme activity. Although it is not certain whether this decrease may be suf®cient enough to produce phenotypic differences in these two carriers, the possibility exits that such a change increases the risk for reaching the threshold of the GCH activity below which dystonia occurs. This prediction is in agreement with the conclusion of the study of the sodium channel gene by Cannon and colleagues [11]. In the previous paper, we provided support, which had been obtained from our study of the truncated mutant GCH, for the hypothesis that the dominant negative effect of the truncated mutant GCH on the normal enzyme may be one of the molecular mechanisms determining the clinical phenotypic heterogeneity of HPD [8]. In the present study, we expanded this hypothesis to include missense mutations of GCH. Acknowledgements This study was supported by Nakajima Memorial Research Grant (to S.U.) and Nara Medical University Research Grant, and by Grant-in-Aid for Scienti®c Research
S114
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
Fig. 3. GTP cyclohydrolase I (GCH) activity of COS-7 cells co-expressed with wild-type GCH (GCH-WT) and (A) GCH-G90V or (B) GCH-G203R. For each set, 100% GCH-WT mRNA is expected to have 100% activity, and other points are expressed as a percentage relative to this control. The dashed line indicates the GCH activity when the wild-type GCH and mutant GCH with a slight residual activity presumably acted independently. Reverse-transcribed polymerase chain reaction (RT-PCR) was performed for the separation of mutant PCR fragments. The separation for GCH-G90V was done as described in the legend of Fig. 1. The separation for GCH-G203R was done with the forward primer carrying a substitution of C for G to create a new DdeI restriction site in the mutant fragments (E5mF: 5 0 -ATCACGGAAGCCTTGCGGCCTGCTGGACTC-3 0 ) and the reverse primer (R3: GACAGACAATGCTACTGGCAGT). Autoradiographs (inset) show normal (upper band) and mutant (lower band) GCH cDNA fragment.
from the Ministry of Education, Science and Culture of Japan (to S.U.). References [1] Segawa M, Hosaka A, Miyagawa F, Nomura Y, Imai H. In: Eldridge R, Fahn S, editors. Hereditary progressive dystonia with marked diurnal ¯uctuation, Advances in Neurology, vol. 14. New York: Raven Press; 1976. pp. 215±233. [2] Narabayashi H, Nagatsu T, Yanagisawa N, Mizuno Y, editors. Doparesponsive dystonia, Delineation of the clinical syndrome and clues to pathogenesis Advances in Neurology, vol. 60. New York: Raven Press; 1993. pp. 577±585. [3] Ichinose H, Ohye T, Takahashi E, Seki N, Hori T, Segawa M, et al. Hereditary progressive dystonia with marked diurnal ¯uctuation caused by mutations in the GTP cyclohydrolase I gene. Nature Genet 1994;8:236±242. [4] Hirano M, Tamaru Y, Nagai Y, Ito H, Imai T, Ueno S. Exon skipping caused by a base substitution at a splice site in the GTP cyclohydrolase I gene in a Japanese family with hereditary progressive dystonia/ dopa responsive dystonia. Biochem Biophys Res Commun 1995;213:645±651.
[5] Furukawa Y, Shimadzu M, Rajput AH, Shimizu Y, Tagawa T, et al. GTP-cyclohydrolase I gene mutations in hereditary progressive and dopa-responsive dystonia. Ann Neurol 1996;39:609±617. [6] Bandmann O, Nygaard TG, Surtees R, Marsden CD, Wood NW, Harding AE. Dopa-responsive dystonia in British patients: new mutations of the GTP-cyclohydrolase I gene and evidence for genetic heterogeneity. Hum Mol Genet 1996;5:403±406. [7] Hirano M, Tamaru Y, Ito H, Matsumoto S, Imai T, Ueno S. Mutant GTP cyclohydrolase I mRNA levels contribute to dopa-responsive dystonia onset. Ann Neurol 1996;40:796±798. [8] Hirano M, Yanagihara T, Ueno S. Dominant negative effect of GTP cyclohydrolase I mutations in dopa-responsive hereditary progressive dystonia. Ann Neurol 1998;44:363±371. [9] Higuchi R, Krummel B, Saiki RK. A general method of in vitro preparation and speci®c mutagenesis of DNA fragments; study of protein and DNA interactions. Nucleic Acids Res 1998;16:7351± 7367. [10] Steinmeyer K, Lorenz C, Pusch M, Koch MC, Jentsch TJ. Multimeric structure of ClC-1 chloride channel related by mutations in dominant myotonia congenita (Thomsen). EMBO J 1994;13:737±743. [11] Cannon SC, Corey DP. Loss of Na 1 channel inactivation by anemone toxin (ATXII) mimics the myotonic state in hyperkalemic periodic paralysis. J Physiol 1993;466:501±520.
www.elsevier.com/locate/braindev
Original article
Missense mutants inactivate guanosine triphosphate cyclohydrolase I in hereditary progressive dystonia Satoshi Ueno*, Makito Hirano Department of Medical Genetics, Nara Medical University, Shijo-cho 840, Kashihara, Nara 634-8521, Japan
Abstract Hereditary progressive dystonia (HPD) with marked diurnal ¯uctuation is caused by mutant guanosine triphosphate (GTP) cyclohydrolase I (GCH). The clinical presentation of dominant HPD varies considerably. We proposed the hypothesis that a relative increase of mutant GCH capable of inhibiting normal GCH is responsible for heterogeneous phenotypic manifestations. In a Japanese family with a novel G90V mutation, an affected heterozygote had a higher mutant/normal mRNA ratio than an unaffected heterozygote. Co-expression analysis showed that mutant enzyme (GCH-G90V) inactivated the normal enzyme in the COS cells. Similarly, GCH-G203R showed the dominant negative effects. These results supported our proposed hypothesis. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Hereditary progressive dystonia; Guanosine triphosphate cyclohydrolase I; Missense mutation; Dominant negative effect
1. Introduction
2. Materials and methods
Hereditary progressive dystonia (HPD) with marked diurnal ¯uctuation/Dopa-responsive dystonia (DRD) is a disease characterized by childhood-onset dystonia that remarkably responds to small doses of levodopa [1,2]. Recently, several workers have demonstrated that HPD in some cases is caused by mutations in the gene encoding guanosine triphosphate (GTP) cyclohydrolase I (GCH) which is the ®rst enzyme in the pathway converting GTP to tetrahydrobiopterin, an essential cofactor of tyrosine hydroxylase [3±7]. HPD is a dominantly inherited disorder with the considerable diversity of clinical phenotypic expression [1,2]. We have previously proposed that a relative increase in the mutant product capable of inhibiting the wild-type gene product is responsible for the clinical phenotypic diversity [4,7]. This prediction was in part supported by the results of our recent study of truncated GCH mutants [8]. To expand this hypothesis to include missense mutations, we studied a novel missense mutation of the GCH gene in a Japanese family with HPD. Another GCH mutant with a substitution of arginine for glycine at position 203 (G203R), which was identi®ed originally in a British case [6] and recently in a Japanese case, was subjected to the present study.
2.1. Subjects
* Corresponding author. Tel./Fax: 181-744-22-9684. E-mail address: [email protected] (S. Ueno).
A 25-year-old woman began to notice tremor at rest in the upper extremities at age 11. She developed a gait disturbance due to dystonia in all limbs and wry neck at age 12. These symptoms worsened toward the evening, and were alleviated after sleep. Small doses of levodopa produced marked bene®cial effects on the patient, however, dyskinesia appeared in the neck with the administration of levodopa. The dystonia was relieved by trihexyphenidyl hydrochloride without dyskinesia. The blood samples were collected from the patient and her unaffected parents and healthy controls for this study after obtaining informed consent. 2.2. Sequencing of the GCH gene All exons and exon±intron junctions of the GCH gene were ampli®ed by PCR and sequenced as described previously [4,7,8]. 2.3. Quanti®cation of lymphocyte GCH mRNA Total RNA from peripheral blood lymphocytes was reverse-transcribed, and synthesized cDNA encoding GCH were quantitatively ampli®ed by PCR [4,7,8].
0387-7604/00/$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S 0387-760 4(00)00135-2
S112
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
2.4. Preparation of GCH-G203R cDNA Genomic DNA of another Japanese patient with typical HPD was available for the GCH gene analysis. The determined sequence contained a G to A change in exon 5, resulting in the G203R mutation. This mutation was ®rst reported in a British DRD case [6]. GCH-G203R cDNA was synthesized by PCR-based mutagenesis [9], and inserted into an expression vector pcDNA3 for the present study. 2.5. Expression of GCH in transfected COS cells The COS-7 cells were transfected with the cDNA for wild-type GCH (GCH-WT), the newly recognized mutant GCH, or GCH-G203R in the expression vector pcDNA3. In cotransfection experiments, the cells were transfected with the cloned wild-type and mutant GCHcDNA mixed at various ratios using Lipofectamine (GIBCO laboratories, Grand Island, NY). The cells were analyzed for the expression of GCH mRNA, enzyme protein, and enzyme activity according to the same method as that described in our previous paper [8]. 3. Results 3.1. Mutation and expression of the GCH gene in lymphocytes The GCH gene from the patient contained a novel G to T transversion in exon 1 (Fig. 1A), leading to a substitution of valine for glycine at position 90 (G90V). The BsaAI-restriction site analysis showed that the patient and her mother were heterozygous for this mutation, while her father and 100 control subjects had only the normal sequence (not shown).
The ratio of mutant/normal GCH mRNA in the patient's lymphocytes was 1:54 ^ 0:04 (mean ^ SD), signi®cantly higher than that in the mother's, 1:00 ^ 0:06 (Fig. 1B). 3.2. GCH expression in transfected COS cells GCH mRNA was not detected in non-transfected COS-7 cells but was present in cells transfected to express wildtype and mutant GCH (Fig. 2). The enzyme activity for GCH-G90V in the COS cells is a ,12% of control enzyme activity, and for GCH-G203R, ,6%. GCH protein was not detected in the control COS-7 cell extracts, whereas in the transfected COS cells with GCH-WT, GCH-G90V or GCHG203R cDNA, the protein was detected with an apparent molecular weight (28 kD) for the monomeric form of the normal enzyme. The expression level of GCH-G203R was lower than GCH-WT and GCH-G90V. 3.3. Inhibition of normal GCH enzyme activity by mutant GCH The co-expression of the mutant with wild-type GCH in COS cells showed that the two mutant GCHs inactivated the normal enzyme in a dose-dependent manner. GCH activity was determined with COS-7 cells cotransfected with GCHWT cDNA and mutant GCH cDNA in various ratios. When COS cells were cotransfected with the expression plasmid for GCH-WT and the plasmid for GCH-G90V or GCHG203R in the 1:1 ratio, the enzyme activity decreased to 70 and 62% of control, respectively (Fig. 3). This effect was clearly speci®c, since the co-expression of GCH with unrelated proteins including Cu/Zn superoxide dismutase or chloramphenicol acetyl transferase cDNA had no effect on the enzyme activity (data not shown).
Fig. 1. (A) Direct sequence analysis of ampli®ed genomic DNA from a control and the patient with dopa-responsive dystonia. The patient is heterozygous for a G to T transversion. (B) Quantitative reverse-transcribed polymerase chain reaction analysis of normal and mutant GCH mRNA was done with primers R1: 5 0 CTGGGCGAGAACCCCCAGCGGCA-3 0 , and E1mF: 5 0 -CTGGGCGAGAACCCCCAGCGGCACG-3 0 carrying a substitution of C for A. This substitution creates a new BsuAI restriction site in the mutant allele, thereby the mutant fragment can be separated from normal ones. The expressed ratio of mutant/normal GCH mRNA is higher in the patient than in her asymptomatic mother. Quanti®cation was carried out with the normal and mutant fragments ampli®ed by ®ve separate RT-PCR ampli®cations. (*P , 0:05, repeated analysis of variance). Autoradiographs (inset) show normal (upper band) and mutant (lower band) GCH mRNA.
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
Fig. 2. The activity of COS-7 cells transfected with GCH-WT cDNA was 0:67 ^ 0:03 nmol/h per mg protein (mean ^ SD), while the native COS-7 cells showed no enzyme activity. The activities in COS-7 cells with mutant GCH expression were 0:075 ^ 0:002 nmol/h per mg protein for GCHG90V and 0:032 ^ 0:002 nmol/h per mg protein for GCH-G203R. (*P , 0:05, repeated analysis of variance). Assays were done in quadruplicate. In the middle panels, reverse-transcribed polymerase chain reaction shows that GCH mRNA was not detected in non-transfected COS-7 cells, but was present in the cells transfected with wild-type GCH (GCH-WT) cDNA and mutant GCH cDNAs. Immunoblotting demonstrates that GCH subunit was expressed in the COS-7 cells transfected with the expression vector for wild-type and mutant GCH. The expression level of GCH-G203R was lower than those of other two transfected cells.
4. Discussion We have characterized the mutant GCH gene in a new Japanese family with HPD, where a G to T transversion in exon 1 caused a G90V substitution (Fig. 1A). Both our patient with HPD and her unaffected mother are heterozygous for this mutation. The relative expression of mutant GCH mRNA was higher in the patient than in her mother. This observed ®nding within the family, which is similar to those in other HPD families previously reported by us, is in agreement with our proposed hypothesis that the mutant/ normal GCH mRNA ratio may contribute to the phenotypic variation of HPD [4,7,8]. At the present time, we can not exclude the possibility that heteroduplex formation between wild-type and mutant GCH cDNAs during the PCR may affect the quantitation with the possibility that the mutant cDNA is underestimated. It was interesting to note that this patient showed resting tremor in the upper limbs and dopainduced dyskinesia, which were previously unreported
S113
symptoms in genetically proven HPD. This suggested that the clinical phenotype of HPD was far more variable than previously thought [1,2]. The in vitro transfection experiments showed that the wild-type and mutant GCH proteins were expressed in the COS cells. The levels of GCH enzyme activity for GCHG90V and GCH-G203R were ,12% and ,6% of control activity, respectively. The two mutants were examined for their ability to inactivate the native enzyme using the cotransfection assay. It was demonstrated that GCH-G90V and GCH-G203R mutants inactivated the normal enzyme, and GCH-G203R was more effective in reducing the enzyme activity than GCH-G90V. This fact indicated that the dominant negative effect was mutant-dependent. Similar ®ndings were obtained in the study of the ClC-1 channel gene; different missense mutations showed different magnitudes of dominant negative effects on the normal channel [10]. Immunoblotting experiments with the transfected COS cell extracts showed that, when GCH-G90V and GCHG203R were expressed individually, each displayed a molecular mass (28 kD) for a normal GCH subunit. However, the expression of GCH-G203R was reduced compared with GCH-WT and GCH-G90V. This reduction may be in part due to the fact that the G203R mutation, a non-conservative amino acid change, causes the conformational changes of the GCH molecule, leading to aberrant processing, or to premature degradation of the GCH. The results of the in vitro experimentation may in part account for clinical phenotypic variation within the same family members. We have shown that, in the HPD family associated with GCH-G90V, the mutant/normal GCH mRNA ratio is 1.54 for the affected carrier, and 1.00, for the unaffected carrier. Fig. 3A showed that this difference corresponded to a ,10% decrease in the enzyme activity. Although it is not certain whether this decrease may be suf®cient enough to produce phenotypic differences in these two carriers, the possibility exits that such a change increases the risk for reaching the threshold of the GCH activity below which dystonia occurs. This prediction is in agreement with the conclusion of the study of the sodium channel gene by Cannon and colleagues [11]. In the previous paper, we provided support, which had been obtained from our study of the truncated mutant GCH, for the hypothesis that the dominant negative effect of the truncated mutant GCH on the normal enzyme may be one of the molecular mechanisms determining the clinical phenotypic heterogeneity of HPD [8]. In the present study, we expanded this hypothesis to include missense mutations of GCH. Acknowledgements This study was supported by Nakajima Memorial Research Grant (to S.U.) and Nara Medical University Research Grant, and by Grant-in-Aid for Scienti®c Research
S114
S. Ueno, M. Hirano / Brain & Development 22 (2000) S111±S114
Fig. 3. GTP cyclohydrolase I (GCH) activity of COS-7 cells co-expressed with wild-type GCH (GCH-WT) and (A) GCH-G90V or (B) GCH-G203R. For each set, 100% GCH-WT mRNA is expected to have 100% activity, and other points are expressed as a percentage relative to this control. The dashed line indicates the GCH activity when the wild-type GCH and mutant GCH with a slight residual activity presumably acted independently. Reverse-transcribed polymerase chain reaction (RT-PCR) was performed for the separation of mutant PCR fragments. The separation for GCH-G90V was done as described in the legend of Fig. 1. The separation for GCH-G203R was done with the forward primer carrying a substitution of C for G to create a new DdeI restriction site in the mutant fragments (E5mF: 5 0 -ATCACGGAAGCCTTGCGGCCTGCTGGACTC-3 0 ) and the reverse primer (R3: GACAGACAATGCTACTGGCAGT). Autoradiographs (inset) show normal (upper band) and mutant (lower band) GCH cDNA fragment.
from the Ministry of Education, Science and Culture of Japan (to S.U.). References [1] Segawa M, Hosaka A, Miyagawa F, Nomura Y, Imai H. In: Eldridge R, Fahn S, editors. Hereditary progressive dystonia with marked diurnal ¯uctuation, Advances in Neurology, vol. 14. New York: Raven Press; 1976. pp. 215±233. [2] Narabayashi H, Nagatsu T, Yanagisawa N, Mizuno Y, editors. Doparesponsive dystonia, Delineation of the clinical syndrome and clues to pathogenesis Advances in Neurology, vol. 60. New York: Raven Press; 1993. pp. 577±585. [3] Ichinose H, Ohye T, Takahashi E, Seki N, Hori T, Segawa M, et al. Hereditary progressive dystonia with marked diurnal ¯uctuation caused by mutations in the GTP cyclohydrolase I gene. Nature Genet 1994;8:236±242. [4] Hirano M, Tamaru Y, Nagai Y, Ito H, Imai T, Ueno S. Exon skipping caused by a base substitution at a splice site in the GTP cyclohydrolase I gene in a Japanese family with hereditary progressive dystonia/ dopa responsive dystonia. Biochem Biophys Res Commun 1995;213:645±651.
[5] Furukawa Y, Shimadzu M, Rajput AH, Shimizu Y, Tagawa T, et al. GTP-cyclohydrolase I gene mutations in hereditary progressive and dopa-responsive dystonia. Ann Neurol 1996;39:609±617. [6] Bandmann O, Nygaard TG, Surtees R, Marsden CD, Wood NW, Harding AE. Dopa-responsive dystonia in British patients: new mutations of the GTP-cyclohydrolase I gene and evidence for genetic heterogeneity. Hum Mol Genet 1996;5:403±406. [7] Hirano M, Tamaru Y, Ito H, Matsumoto S, Imai T, Ueno S. Mutant GTP cyclohydrolase I mRNA levels contribute to dopa-responsive dystonia onset. Ann Neurol 1996;40:796±798. [8] Hirano M, Yanagihara T, Ueno S. Dominant negative effect of GTP cyclohydrolase I mutations in dopa-responsive hereditary progressive dystonia. Ann Neurol 1998;44:363±371. [9] Higuchi R, Krummel B, Saiki RK. A general method of in vitro preparation and speci®c mutagenesis of DNA fragments; study of protein and DNA interactions. Nucleic Acids Res 1998;16:7351± 7367. [10] Steinmeyer K, Lorenz C, Pusch M, Koch MC, Jentsch TJ. Multimeric structure of ClC-1 chloride channel related by mutations in dominant myotonia congenita (Thomsen). EMBO J 1994;13:737±743. [11] Cannon SC, Corey DP. Loss of Na 1 channel inactivation by anemone toxin (ATXII) mimics the myotonic state in hyperkalemic periodic paralysis. J Physiol 1993;466:501±520.