Pii: s0304-3940(99)00921-0

A cluster of single nucleotide polymorphisms in the 5H-leader of the human dopamine D3 receptor gene (DRD3) and its Sinthuja Sivagnanasundarama,1, Alex G. Morrisa,1, Emma J. Gaitondeb, Peter J. McKennac, John D. Mollonb, David M. Hunta,* aDepartment of Molecular Genetics, Institute of Ophthalmology, University College London, Bath Street, London, EC1V 9EL, UK bDepartment of Experimental Psychology, University of Cambridge, Downing Street, Cambridge, CB2 3EB, UK Received 26 August 1999; received in revised form 19 November 1999; accepted 19 November 1999 The association between schizophrenia and the Ser9Gly variant of the dopamine D3 receptor gene (DRD3) has been the subject of numerous studies. Under meta-analysis this site, or one or more in linkage disequilibrium with it, appears to contribute a small increase to the relative risk of schizophrenia. In this study, 768 bp of the 5H-leader region of DRD3 mRNA was screened for polymorphisms to assess their contribution to the association of DRD3 with schizophrenia. A cluster of three single nucleotide polymorphisms (SNPs) was identi®ed in tight linkage disequilibrium with each other and with the Ser9Gly polymorphism. One of the 5H-leader SNPs encodes a Lys9Glu variant within a 36 amino acid residue stretch of an upstream open reading frame (uORF). Two common haplotypes are found in the population examined; one is linked to the Ser9 coding variant and the other to the Gly9 variant. A panel of 73 schizophrenic patients and 56 matched controls recruited from the East Anglia region of the United Kingdom was screened for disease association at these sites.
Since the 5H-leader and coding sites are in tight disequilibrium, the combined genotype of all 4 sites was scored for each patient. A signi®cant association was seen between disease and the frequency distribution of these genotypes (x2 ˆ 13:19, d:f: ˆ 3, P ˆ 0:0042; Cochran method for sparse cells applied). A 20% excess of one of the heterozygous genotypes, in which the sequences differ at three of the four SNPs, including Ser9/Gly9 in the receptor and Lys9/Glu9 in the uORF, was found in the patient group. An absence of association of disease with the Ser9Gly polymorphism had previously been reported for this panel. This suggests that these SNPs and the corresponding coding changes may exert a combined or synergistic effect on susceptibility to schizophrenia. q 2000 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: 5H-leader; Dopamine receptor; Schizophrenia; Single nucleotide polymorphism; Association; Common complex disease The dopamine hypothesis of schizophrenia has been the residue 9 whereas the G substitution (allele 2) generates an subject of numerous studies. In particular, the suggestion MscI (or BalI) restriction enzyme site and encodes glycine that the dopamine D3 receptor gene (DRD3) is a candidate at residue 9. An excess frequency of homozygotes for both for increased susceptibility to the disease comes from the alleles was originally reported in schizophrenic patients [2].
high af®nity of D3 receptors for neuroleptic drugs [12,13].
This ®nding has been followed up by numerous replication Higher levels of DRD3 binding have been found in the studies, some of which have con®rmed this conclusion.
mesolimbic system of schizophrenics [8]. A common Others however have not supported it, including our study variant of a single nucleotide polymorphism (SNP) of A/G [3] which employed the same panel of cases and controls at position 25 of the DRD3 coding sequence has been iden- that has been used in this study. A meta-analysis [16] in ti®ed [7]. The A substitution (allele 1) encodes a serine at which over 30 of these case-control studies have been exam- ined, concluded that a relationship between schizophrenia * Corresponding author. Tel.: 144-171-608-6820; fax: 144-0- and homozygosity for the Ser9Gly polymorphism is present.
An MspI polymorphism located 40 kb downstream of the E-mail address: d.hunt@ucl.ac.uk (D.M. Hunt) 1 These two authors contributed equally to this study.
0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved.
S. Sivagnanasundaram et al. / Neuroscience Letters 279 (2000) 13±16 MscI site has also been identi®ed, but was not associated with the MscI site or with disease [4].
In all the above studies, it is the frequency of the two variant types of receptor gene that have been assessed in cases and controls, whereas the important factor may be the level of receptor protein production. This is determined by upstream regulatory elements, both in the promoter and in the 5H-leader. To address this problem, we have sequenced 768 bp of the 5H-upstream region of human DRD3. Reverse transcription PCR (RT PCR) of this upstream region, using polyA1 mRNA from human adult brain, demonstrated that the 768 bp stretch is situated in the 5H-leader region of the mRNA. This region was then screened for polymorphisms using both single stranded conformation polymorphism Fig. 1. Sequence of the 5H-leader of DRD3 mRNA together with (SSCP) and sequencing analysis. Three novel SNPs were 174 bp of the coding sequence. The arrows above and below the identi®ed and their frequency, and that of the MscI poly- sequence indicate the forward (F) and reverse (R) primers used.
morphism, were scored in unrelated schizophrenic patients The position from the translation start site of the 5H-end of each primer is given with the PCR fragment number. The base substi- As previously reported [3], a panel of unrelated patients tution for each SNP is shown, including the 125 change in the DRD3 coding sequence that gives rise to Ser9Gly. The 36 amino of European descent with chronic schizophrenia and age, acids encoded by the uORF are shown with the coding change sex, and ethnically matched controls was recruited from East Anglia, United Kingdom. Informed consent was obtained, subjects were interviewed using the Present region of DRD3 in DNA samples obtained from patients and State Examination [17] and diagnoses were made using a controls. Overlapping PCR products were generated from structured interview with DSM-IIIR criteria to assess recent four different PCR reactions. The primer pairs used are or chronic psychopathology. Data were also gathered on shown in Fig. 1. The nature of the base changes responsible personal and family psychiatric and medical history; 22% for the band shifts was determined by sequencing. For each had a ®rst- or second-degree relative with schizophrenia or different fragment detected, the sequence of the correspond- schizoaffective disorder. Controls were given a semi-struc- ing PCR product was con®rmed with a minimum of 20 tured interview and subjects with any history of major individuals comprising at least 10 patients and 10 controls.
illness, either in themselves or in a ®rst degree relative, Statistical comparisons of the patients and controls in 2 £ m were excluded. Symptoms were rated on an ordinal scale contingency tables as well as the estimates of relative risk for response to clozapine. DNA was extracted from blood were carried out using standard methods [10]. The SPSSw using the Nucleon II kit (Scotlab Inc., Shelton, CT).
program was used for the Mann±Whitney±Wilcoxon Rank Seven high-density gridded membranes of a human PAC Sum test for association of clozapine response ratings with library (RPC11) from the UK HGMP resource centre were screened with a 219 bp fragment generated by PCR from the The sequence to 2768 was found to correspond to 5H end of the coding region of the DRD3 gene, using primer mRNA. Unfortunately, all attempts to identify precisely pair exon1F and exon1R (Fig. 1). A positive signal was the transcription start site with 5H-RACE were unsuccessful.
obtained for clone 221/M4. The 5H-¯anking region of The translation start site in exon 1 was previously deduced DRD3 was then ampli®ed from this clone using the method from an open reading frame in the sequence of a cDNA of single-primer walking PCR [11], initially with primer clone situated 69 bp downstream of an in-frame termination exon1R to extend upstream. The ampli®cation products codon [13]. Our extended sequence (Fig. 1) contains an were cloned into pTAg (R&D Systems) and sequenced upstream open reading frame (uORF) between 2228 and using exon1R as a sequencing primer and ABI PRISMe 2120 that would encode a putative 36 residue peptide; the dye terminator cycle sequencing. Sequencing was then presence of Kozak sequences [6] at the putative start site extended to the 768 bp stretch, as well as to 174 bp of the strengthens the case for the translation of this peptide. No 5H-coding sequence, using primers designed to the new other ORFs were found in this region of the mRNA. SSCP sequence. All sequencing was carried out on an ABI analysis of PCR products 2 and 3 each gave identical band- Model 373a sequencer. The extent of the 5H-leader upstream ing patterns across all individuals screened, indicating that of the translation start site was examined by generating no polymorphic sites are present. In contrast, fragments 1 DRD3 cDNA from adult human brain polyA1 mRNA and 4 showed a range of banding patterns. PCR product 1 (Clonteche) with a Promegae reverse transcription-poly- had a single polymorphic site at 2707 (C/G), whereas PCR merase chain reaction (RT PCR) kit and primers 1F and product 4 had two polymorphic sites at 2343 (A/G) and exon1R (Fig. 1). An isotopic PCR-based SSCP method [5] 2204 (A/G). Assuming random mating, the observed geno- was used to screen from 2768 to 2150 bp of the 5H-¯anking typic frequencies for each SNP are not signi®cantly different S. Sivagnanasundaram et al. / Neuroscience Letters 279 (2000) 13±16 from the expected frequencies, suggesting that each is in tested by using a 2 £ m x 2 test, for which all cells with Hardy±Weinberg equilibrium. The Ser9Gly polymorphism expected values ,5 were collapsed [11]. A signi®cant in this panel of patients and controls had previously been difference between cases and controls (x2 ˆ 13:19, found to be in Hardy±Weinberg equilibrium [3].
d:f: ˆ 3, P ˆ 0:0042) suggests that there may be an associa- The identities of all three upstream polymorphic sites tion between these four SNPs and schizophrenia. The rela- were scored for the 73 patient and 56 control DNA samples.
tive risk of the two common genotypes was calculated using The remaining 11 patient and 21 control samples [3] proved the odds ratio. The relative risk for individuals that are refractory to PCR ampli®cation for at least one of the frag- homozygous for the GGAA haplotype is 1.15 (C.I. 95% ˆ ments. The level of linkage disequilibium between all four 0.54±2.5), which suggests that this genotype does not affect SNP sites, including the Ser9Gly coding A/G variant at the susceptibility to schizophrenia. The relative risk of the 125, was analyzed by the EH program [15], which indi- C/G G/G G/A A/G genotype, which corresponds to hetero- cated that the four SNPs are in tight linkage disequilibrium zygotes for the two most common haplotypes, is 3.13 (C.I.
(P , 10227), as expected for sites that lie within an 732 bp 95% ˆ 1.28±7.6). From inspection, this class is 20% over- stretch of DNA. The combined genotype of all four sites represented in patients. Since these two genotypes were was therefore scored for each patient; the corresponding selected post-hoc for the estimation of relative risk, the frequencies of each of the 19 combined genotypes found hypothesis that the heterozygosity for these two haplotypes in the panel are shown in Table 1. Inspection of the mean may increase the risk to schizophrenia remains to be tested genotype frequencies of the pooled cases and controls, in a separate sample. This avoids applying an overly conser- reveals that two of the genotypes are much more common vative multiple test correction for the 19 genotypes found.
than the 17 remaining genotypes. Since one of the common The presence of multiple haplotypes in the population genotypes is homozygous at all three upstream sites and for examined indicates that these SNPs are suf®ciently ancient Ser9, this identi®es one common haplotype as GGAA.
for recombination to have taken place between sites that are Given the presence of linkage disequilibrium, the other common Gly9 coding haplotype can be identi®ed by The association of clozapine response with the common subtraction as CGGG. These haplotypes differ therefore at C/G G/G A/G A/G genotype was assessed by comparing the two sites (2707 and 2204): one is linked to the Ser9 coding response ratings of patients on clozapine (n ˆ 49) with this variant and the other to the Gly9 variant of the D3 receptor.
genotype versus all other genotypes. No signi®cant associa- The A-204G SNP encodes a non-conservative Lys9Glu tion was found (P ˆ 0:24); this genotype is not associated substitution in the putative 36 residue peptide encoded by therefore with clozapine response in this sample. We have formerly found that patients in this panel either taking or not The frequency distribution of the 19 combined genotypes taking clozapine did not differ signi®cantly in the allele or in the patient group compared to the control group was genotype frequencies of the Ser9Gly variant [3].
a Frequency of 5H-leader and coding SNPs in the DRD3 gene in patients and controls. The number and % frequency in parentheses of each genotype is shown. The Ser9Gly genotypes were determined by Gaitonde et al. (1996) [3]. x 2 test between 73 patients and 56 controls for all genotypes in the table: P ˆ 0:004. Relative risk was estimated by the odds ratio (OR) and 95% con®dence interval (CI) for the two most common genotypes, which are indicated by asterisks: *OR ˆ 1.15 (CI 95%, 0.54±2.5) **OR ˆ 3.13 (CI 95% ˆ 1.28±7.6).
S. Sivagnanasundaram et al. / Neuroscience Letters 279 (2000) 13±16 The absence of an association between schizophrenia and zygosity at the dopamine D3 receptor gene. J. Med. Genet., the Ser9Gly polymorphism in this panel of patients and controls has been reported previously [3], and this subset [3] Gaitonde, E.J., Morris, A., Sivagnanasundaram, S., McKenna, P.J., Hunt, D.M. and Mollon, J.D., Assessment similarly showed no association with this polymorphism of association of D3 dopamine receptor MscI polymorph- (P ˆ 0:13). However, when the SNPs in the coding and ism with schizophrenia: analysis of symptom ratings, 5H-leader are considered together, a signi®cant association family history, age at onset, and movement disorders.
with disease is seen. A similar combined `super-allele' was Am. J. Med. Genet., 67 (1996) 455±458.
found by examining the effect of SNPs in the alcohol dehy- [4] Griffon, N., Crocq, M.A., Pilon, C., Martres, M.P., Mayerova, A., Uyanik, G., Burgert, E., Duval, F., Macher, J.P., Javoy- drogenase (Adh) gene on the level of its expression in Agid, F., Tamminga, C.A., Schwartz, J.C. and Sokoloff, P., Drosophila melanogaster [14]. This quantitative trait was Dopamine D3 receptor gene: organization, transcript modulated by the combination of the individual effects of variants, and polymorphism associated with schizophre- SNPs located in the coding sequence, introns and 3H- nia. Am .J. Med. Genet., 67 (1996) 63±70.
untranslated region of Adh, that are non-randomly asso- [5] Hayashi, K., Laboratory Protocols for Mutation Detection, HUGO, Oxford University Press, Oxford, 1996, pp. 14±22.
ciated. This suggests that when the relationship between [6] Kozak, M., Comparison of initiation of protein synthesis in polymorphic sites and a complex disease is being assessed Procaryotes, Eucaryotes, and organelles. Microbiol. Rev., for association studies, examination of all of the closely- linked polymorphic sites in combination can be more [7] Lannfelt, L., Sokoloff, P., Martes, M., Pilon, C., Giros, B., revealing than when individual sites are considered in isola- Schwartz, J., Jonsson, E. and Sedvall, G., Amino acid substitution in the D3 dopamine receptor as a useful poly- tion. Meta-analysis of the large number of Ser9Gly studies morphism for investigating psychiatric disorders. Psychiat.
has shown that schizophrenia is associated with homozyg- osity at this site [16]. However, it is unknown at present [8] Murray, A.M., Ryoo, H.L., Gurevich, E. and Joyce, J.N., whether the Gly9Ser substitution alters the pharmacological Localization of dopamine D3 receptors to mesolimbic and properties of the DRD3 receptor. The association of schizo- D2 receptors to mesostriatal regions of human forebrain.
Proc. Natl. Acad. Sci. USA, 91 (1994) 11271±11275.
phrenia with heterozygosity in this study is dependent on the [9] Parola, A. and Kobilka, B., The peptide product of a 5H- inclusion of the three 5H-leader sites. This implies that leader cistron in the beta 2 adrenergic receptor mRNA inhi- disease susceptibility could be related to different levels of bits receptor synthesis. J. Biol. Chem., 269 (1994) 4497± expression of the two protein variants. The 5H-leader may modulate translational activity via the translational control [10] Rosner, B., Fundamentals of Biostatistics, Duxbury Press, elements as found in the 5H-leader of human GLUT-1 [11] Screaton, G.R., Bangham, C.R. and Bell, J.I., Direct sequen- mRNA [1]. However, it remains to be established whether cing of single primer PCR products: a rapid method to the two common 5H-leader haplotypes of DRD3 alter the achieve short chromosomal walks. Nucl. Acids Res., 21 level of protein synthesis. The uORF in this region of the mRNA has similarities to that present in the 5H leader of the [12] Sokoloff, P., Giros, B., Martres, M.P., Bouthenet, M.L. and Schwartz, J.C., Molecular cloning and characterization of a b2-adrenergic receptor mRNA [9]. A peptide rich in basic novel dopamine receptor (D3) as a target for neuroleptics.
residues is encoded by this uORF and has been shown to down-regulate expression. Mutations that replace basic for [13] Sokoloff, P., Andrieux, M., Besancon, R., Pilon, C., Martres, neutral residues in this peptide increased translation by M.P., Giros, B. and Schwartz, J.C., Pharmacology of human 50%. The presence and role of such a peptide in DRD3 dopamine D3 receptor expressed in a mammalian cell line: comparison with D2 receptor. Eur. J. Pharmacol., 225 (1992) expression should therefore be investigated. In this regard, it is interesting to note that the uORF adjacent to DRD3 [14] Stam, L. and Laurie, C., Molecular dissection of a major encodes three basic residues in the ®rst nine N-terminal gene effect on a quantitative trait: the level of alcohol dehy- residues, and that the SNP at 2204 would generate a drogenase expression in Drosophila melanogaster. Genet- Lys9Glu substitution in this putative peptide.
[15] Terwilliger, J. and Ott, J., Handbook of Human Genetic Linkage, Johns Hopkins University Press, Baltimore, MD, This work was supported by a project grant (037331) [16] Williams, J., Spurlock, G., Holmans, P., Mant, R., Murphy, from the Wellcome Trust. Sinthuja Sivagnanasundaram K., Jones, L., Cardno, A., Asherson, P., Blackwood, D., muir, was in receipt of a Wellcome Trust Prize Studentship.
W., Meszaros, K., Aschauer, H., Mallet, J., Laurent, C., Pekkarinen, P., Seppala, J., Stefanis, C.N., Papadimitriou, G.N., Macciardi, F., Verga, M., Pato, C., Azavedo, H., [1] Boado, R.J., Tsukamoto, H. and Pardridge, W.M., Evidence Crocq, M-A., Gurling, H., Kalsi, G., Curtis, D., McGuf®n, P.
for translational control elements within the 5'-untranslated and Owen, M.J., A meta-analysis and transmission disequi- region of GLUT1 Glucose Transporter mRNA. J. Neuro- librium study of association between the dopamine D3 receptor gene and schizophrenia. Mol. Psychiatry, 3 [2] Crocq, M.A., Mant, R., Asherson, P., Williams, J., Hode, Y., Mayerova, A., Collier, D., Lannfelt, L., Sokoloff, P., [17] Wing, J.K., Cooper, J.E. and Sartorius, N., The Measure- Schwartz, J.C., Gill, M., Macher, J.P., McGuf®n, P. and ment and Classi®cation of Psychiatric Symptoms, Owen, M.J., Association between schizophrenia and homo- Cambridge University Press, Cambridge, 1974, pp. 189±228.

Source: http://vision.psychol.cam.ac.uk/jdmollon/papers/DRD3_schizophrenia.pdf