ORIGINAL RESEARCH ARTICLE. G Sjøholt 1, AK Gulbrandsen 1, R Løvlie 1, JØ Berle 2, A Molven 1 and VM Steen 1

Transcription

1 (2000) 5, Macmillan Publishers Ltd All rights reserved /00 $ ORIGINAL RESEARCH ARTICLE A human myo-inositol monophosphatase gene (IMPA2) localized in a putative susceptibility region for bipolar disorder on chromosome 18p11.2: genomic structure and polymorphism screening in manic-depressive patients G Sjøholt 1, AK Gulbrandsen 1, R Løvlie 1, JØ Berle 2, A Molven 1 and VM Steen 1 1 Dr Einar Martens Research Group for Biological Psychiatry, Center for Molecular Medicine; 2 Institute of Psychiatry, University of Bergen, Haukeland University Hospital, Bergen, Norway For several decades, lithium has been the drug of choice in the long-term treatment of manicdepressive illness, but the molecular mechanism(s) mediating its therapeutic effects remain to be determined. The enzyme myo-inositol monophosphatase (IMPase) in the phospholipase C signaling system is inhibited by lithium at therapeutically relevant concentrations, and is a candidate target of lithium s mood-stabilizing action. Two genes encoding human IMPases have so far been isolated, namely IMPA1 on chromosome 8q and IMPA2 on chromosome 18p11.2. Interestingly, several studies have indicated the presence of a susceptibility locus for bipolar disorder on chromosome 18p11.2. IMPA2 is therefore a candidate for genetic studies on both etiology and lithium treatment of manic-depressive illness. Here we report that the genomic structure of IMPA2 is composed of eight exons, ranging in size from 46 bp to 535 bp. The promoter region contains several Sp1 elements and lacks a TATA-box, features typical for housekeeping genes. By a preliminary polymorphism screening of exons 2 8 in a sample of 23 Norwegian bipolar patients, we have identified nine single nucleotide polymorphisms (SNPs). Seven of the polymorphisms were located in the introns, one was a silent transition in exon 2 (159T C) and one was a transition in exon 5 (443G A) resulting in a predicted amino acid substitution (R148Q). Our data show that even in a small sample of bipolar patients, several variants of the IMPA2 gene can be identified. IMPA2 is therefore an intriguing candidate gene for future association studies of manic-depressive illness. Molecular Psychiatry (2000) 5, Keywords: bipolar disorder; inositol monophosphatase; IMPA2; inositol signaling; chromosome 18; lithium; DNA polymorphism Introduction Lithium may reduce the frequency, duration and severity of recurrent episodes of both mania and depression during maintenance therapy of bipolar disorder. 1 It has, however, been increasingly evident that lithium is far from the perfect mood-stabilizing drug, since only about a third of bipolar patients remain well and free of relapse after 5 years on lithium monotherapy. 2 4 Despite intensive research for several decades, the mechanisms of lithium action are still elusive and we are not able to explain the reasons for the marked variation in lithium response. Several possible effects of lithium on neuronal sig- Correspondence: Professor VM Steen, Dr Einar Martens Research Group for Biological Psychiatry, Center for Molecular Medicine, Haukeland University Hospital, N-5021 Bergen, Norway. Vidar.Steen molmed.uib.no Received 23 April 1999; revised and accepted 3 September 1999 naling have been suggested. 5,6 Among these, an interference with the recycling of inositol in the phospholipase C (PLC) signaling pathway is an attractive hypothesis. 7 Two dephosphorylating enzymes, myoinositol monophosphatase (IMPase) and inositol 1- polyphosphatase (IPPase) are both uncompetitively blocked by lithium at therapeutically relevant concentrations. Inhibition of these enzymes decreases the concentration of free myo-inositol in the cell, 8,9 but studies on the possible functional consequences of such a reduced level of inositol upon neuronal cell signaling have yielded conflicting results We have earlier postulated that allelic variation in the IMPA1 and INPP1 genes, encoding the IMPase and IPPase enzymes, respectively, may influence the varying lithium response among bipolar patients. 13,14 Recently, we identified several polymorphisms in the coding region of INPP1, including a single nucleotide substitution that may be associated with a favorable lithium response in Norwegian bipolar patients. 13 Also

2 the IMPA1 gene is an interesting candidate for pharmacogenetic prediction of lithium response in bipolar disorder. The genomic structure of IMPA1, localized to chromosome 8q, has been determined, and one polymorphism and two short sequences reminiscent of inositol/choline responsive elements (ICRE) were detected in the 3 -UTR of the gene. 15 Although the DNA variation and/or the ICREs could be involved in inositol-regulated expression of IMPA1, there is so far no evidence for a role for this gene in bipolar disorder. During the characterization of IMPA1, we discovered a transcript from chromosome 18p that encodes a protein with high similarity to the IMPase enzyme. 15 This IMPA1-like gene, named IMPA2, was independently reported by Yoshikawa et al who localized it to chromosome 18p The functions of the IMPA2-encoded protein and its possible sensitivity to lithium have not yet been studied, but the predicted amino acid sequence contains regions with high similarity to the protein family of metal-dependent, lithium-inhibited phosphomonoesterases. 17 It is therefore likely that IMPA2 also encodes a lithium-inhibited myo-inositol monophosphatase participating in the inositol phospholipid signaling system. Intensive efforts have been made to identify the chromosomal loci and genetic factors involved in bipolar disorder, and among many candidate regions, a locus near the centromere on chromosome 18p has been pinpointed by several linkage studies Indeed, a recent study by Schwab et al on families with schizophrenia suggested the existence of a potential susceptibility locus for functional psychoses in general on chromosome 18p. 22 In line with these results, the enzyme activity of IMPase in transformed lymphoblastoid cell lines was reported to be lower in cells from bipolar patients than in cells from control individuals. 23 When the bipolar patients were grouped according to their clinical response to lithium therapy, the lithium responders exhibited significantly lower IMPase activity than the patients with poor lithium response. Since the IMPA2 gene maps to a putative susceptibility region for bipolar disorder and possibly encodes a lithium-inhibited IMPase, it should become the target for genetic studies of both lithium action and the pathophysiology of manic-depressive illness. We here report the genomic structure including the promoter region of IMPA2 along with data from a polymorphism screening in a sample of bipolar patients. In parallel with our work, characterization of IMPA2 has also been performed by Yoshikawa et al. 16 These results establish a necessary basis for further studies on whether or not the IMPA2 gene plays a role in the etiology of bipolar disorder and lithium therapy. Materials and methods Confirmation of the cdna sequence of IMPA2 The following EST clones were purchased from Research Genetics (Huntsville, AL, USA) and sequenced to confirm the IMPA2 cdna sequence: 39740, and (Clone ID). To obtain cdna clones that have full-length 5 -ends, 5 -RACE PCR was performed using the Human Brain, Marathon Ready cdna kit (Clontech, Palo Alto, CA, USA). This kit contains cdna ligated with adapters and primers complementary to the adapter sequence, and the reactions were carried out according to the manufacturer s recommendations. In the primary PCR, the gene-specific primer imp-b12 (see Table 1) was used together with the AP1 primer supplied in the kit. In the nested reaction, the gene-specific primer imp-b10 (see Table 1) was used together with the AP2 primer supplied in the kit. The PCR conditions were as follows: an initial denaturing step at 96 C for 1 min, then 35 (for the primary reaction) or 25 (for the nested reaction) cycles at 96 C for 30 s, 60 C for 30 s and 68 C for 4 min. Both programs ended with a final extension at 72 C for 10 min. Long-PCR cloning of human IMPA2 A set of oligonucleotide primer pairs (Table 1, long- PCR ) was designed on the basis of the IMPA2 cdna sequence and used to amplify the IMPA2 gene from genomic DNA (isolated manually from leukocytes by chloroform extraction of EDTA-anticoagulated blood). The GeneAmp XL PCR kit (Perkin-Elmer, Palo Alto, CA, USA) was used, according to the manufacturer s recommendations. When amplifying the IMPA2 promoter and exon 1, GC-Melt reagent (Advantage-GC PCR kit, Clontech) was added. The amplification conditions were as follows: an initial denaturing step at 96 C for 1 min, then cycles of 96 C for 30 s, C for s and 68 C for 1 12 min. All programs ended with a final extension at 72 C for 10 min. Genomic DNA walking To amplify the 3 -end of intron 1, the 5 -end and the 3 -end of intron 5 and the 3 -end of the gene, genomic DNA walking was performed using the Human Promoter Finder Walking Kit (Clontech). The volume of the initial PCR was 50 l, containing 1 Tth reaction buffer (Clontech), 2 U rtth/vent R DNA polymerase (Perkin-Elmer) and 1 l genomic library DNA (supplied in the kit). In the initial reaction, IMPA2-specific primers (Table 1, primary walking ) were used in combination with the AP1 primer supplied in the kit. The volumes of the nested PCRs were l, containing 1 XL PCR buffer (from the GeneAmp XL PCR kit, Perkin-Elmer), 2 4 U rtth/vent R DNA polymerase and l of the initial PCR product. Nested amplification was performed using another set of genespecific primers (Table 1, nested walking ) in combination with the AP2 primer supplied in the kit. The primary PCR was performed by an initial denaturing step at 96 C for 1 min, followed by 7 cycles at 96 C for 30 s and 72 C for 4 min, then 37 cycles at 96 C for 30 s, 60 C for 15 s and 68 C for 4 min. The nested PCR started with an initial denaturing step at 96 C for 30 s, then 25 cycles of 96 C for 30 s, 60 C for 15 s and 68 C for 4 min. All programs ended with a final extension at 72 C for 10 min. 173

3 174 Table 1 Oligonucleotide primers for PCR amplification of overlapping IMPA2 fragments from human genomic DNA Primer name Primer sequence Direction Type of PCR PCR product size (kb) imp-b10 a 5 -GAAACCTCTCTCGCAACTCAGAAATAAT-3 Reverse primary walking imp-b8 a 5 -TCTTCCACAAGGTGATCTGTTTCTGTC-3 Reverse nested walking 0.3 imp-b5 5 -GCTGCAGATCTTGTGACAGAAA-3 Forward imp-b12 5 -GCTGTGGGTGAGCACACACT-3 Reverse Long-PCR 10.0 imp-b7 5 -GCTTCTGGGGCCAAGTGTG-3 Forward imp-b28 5 -TCCAATGCTAACCGCCACAGT-3 Reverse Long-PCR 2.4 imp-b1 5 -GCGGTTAGCATTGGATTT-3 Forward imp-b14 5 -TGTACAGCCGCTCCTCTGTG-3 Reverse Long-PCR 2.2 imp-b31 a 5 -GCGCAGCCTTCATCTTTGAAT-3 Forward primary walking imp-b3 a 5 -CCAGCGGCTCCGGGTCTC-3 Forward nested walking 0.5 imp-b40 a 5 -CAGCTCCCAGCTGACCTAGCTACT-3 Reverse primary walking imp-b16 a 5 -GCAGCCGCTCCATGTTACTC-3 Reverse nested walking 0.6 imp-b9 5 -AATTGGCCCCAAACGTGAC-3 Forward imp-b30 5 -CAGTGCAGGCCAAACTGGTAATA-3 Reverse Long-PCR 0.9 imp-b11 5 -CGAGTGATTGGAAGCTCCACATT Forward imp-b22 5 -ACGATCCGCTTTATCAGCAGTTCTA-3 Reverse Long-PCR 1.9 imp-b23 a 5 -CTCAGGCCTTACAGACGATTAACTA-3 Forward primary walking imp-b25 a 5 -GGGAGTTGTCACGCTACAGTGAGT-3 Forward nested walking 3.5 a For the fragments amplified by PCR-based, genomic walking, only the gene-specific primers are shown for the primary and nested PCR. Screening of a human genomic lambda library To isolate genomic clones containing IMPA2, an IMPA2-specific probe was amplified from lymphocyte cdna using the primer pair imp-b27 (forward; 5 -GCGGCGGACTAGGCACAGA-3 ) and imp-b22 (reverse; Table 1). This probe was radioactively labeled and used to isolate positive IMPA2 clones from a human genomic Lambda FIX II library (Stratagene, La Jolla, CA, USA) by standard methods. 24 IMPA2- containing clones were further screened for exon 1 sequence, by performing PCR with the following primer set: imp-b27 (forward; see above) and imp-b38 (reverse; 5 -GGCCGCCTGGAAGCACTC-3 ). DNA sequencing The PCR products were purified using the QIAquick PCR Purification kit (Qiagen, San Francisco, CA, USA) or Jetsorb Gel Extraction kit (Genomed, Research Triangle Park, NC, USA). EST-containing plasmids were purified using the Qiagen Plasmid Midi kit (Qiagen). DNA was sequenced using the ABI Prism drhodamine Terminator Cycle Sequencing kit or ABI Prism Big Dye Terminator Cycle Sequencing kit (PE Applied Biosystems, Foster City, CA, USA) and the reactions were analyzed on an ABI 377 DNA Sequencer (PE Applied Biosystems). Sequencing primers were either the same as those used in PCR or they were designed specifically. Screening of IMPA2 exons 2 8 for polymorphisms DNA samples were obtained from 23 Norwegian lithium-treated bipolar patients 13 and two healthy Norwegian control subjects. Primers were designed from the sequences flanking the exons of IMPA2 (Table 2). Exons 2 8 were amplified by PCR. The reaction volumes were l, containing 10 mm Tris-HCl (ph 9.0), 1.5 mm MgCl 2, 50 mm KCl, 0.01% gelatin, 0.1% Triton X-100, mm of each dntp, 0.6 M of each primer, U SuperTaq DNA polymerase (HT Biotechnology, Cambridge, UK) and ng purified genomic DNA. The amplification conditions were as follows: an initial denaturing step at 96 C for 1 min, then cycles of 96 C for 30 s, C for s and C for 1 3 min. All programs ended with a final extension at 72 C for 10 min. One strand of the amplified products was sequenced as described above using the PCR primers shown in Table 2. When DNA polymorphisms were suspected, they were confirmed by sequencing the opposite strand. Computer analysis Comparisons of gene sequences were performed using the BLAST network service at the National Center for Biotechnology Information (NCBI), USA ( ncbi.nlm.nih.gov/blast/). The search for ICREs was performed using the Findpatterns program of the GCG program package at the Biotechnology Centre, Univer-

4 Table 2 Oligonucleotide primers used for PCR amplification of the IMPA2 exons 175 Primer name Primer sequence Direction Exon No. Exon size PCR product (bp) size (bp) imp-b ACGGTCCGACCCAGACAGTA-3 Forward imp-b ACCCCTGGAGAGATTTTTACATTCT-3 Reverse imp-b CCCAATGGCAGGCTGAATA-3 Forward imp-b TCCAGAAATAACAAATGGTTGAGTACA-3 Reverse imp-b CCCAGAGAAACCATGATGACA-3 Forward imp-b AAACAAGTGCAGATTTCCTAACATA-3 Reverse imp-b AGTTAGTGAAACTGTGGGAAGTAGT-3 Forward imp-b CCAAAGCCAGTCTGGACTTGT-3 Reverse imp-b GGAAAGGTAATTTCTGTCACAGTGTA-3 Forward imp-b GGCCACGTGAACAAAAACACTAA-3 Reverse imp-b GCTCACGTTGGAAGCCTGTACTC-3 Forward imp-b GGGCTTCCCTGAGTCTGATGTT-3 Reverse imp-b TGGCTCCAGGCCCTATATTCAG-3 Forward imp-b GGAAATTCAGTTGGTGGGGACTT-3 Reverse imp-b TAAGGCAGTGGCAGCTTCAGAG-3 Forward imp-b GCAAAGCTTCCTGAGAGCATAGAT-3 Reverse sity of Oslo, Norway. Promoter analyses were performed at the TFMATRIX transcription factor binding site profile database ( db/tfsearch.html) with a threshold value of 85.0% significance for vertebrate transcription factors. Transcription start site predictions were performed at the promoter predicting database for eukaryote organisms at Lawrence Berkeley National Laboratory ( Nucleotide sequences Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession numbers AF AF Results and discussion The IMPA2 gene is localized to a putative susceptibility region for bipolar disorder on chromosome 18p11.2 and seems to encode a lithium-inhibited IMPase enzyme. We have determined the genomic structure of IMPA2 and used the sequence information to perform a polymorphism screening in a sample of manic depressive patients, with the aim of exploring the possible role of this gene in the pathogenesis and drug response of lithium-treated bipolar disorder. Isolation of genomic fragments covering the IMPA2 gene To determine the genomic organization of IMPA2, we first considered screening a human BAC library. Unfortunately, a screening service ordered from Research Genetics failed to identify any positive BAC clones, despite the use of two different primer pairs (from exon 2 and the 3 -UTR, respectively) which both amplified IMPA2-specific PCR products from genomic DNA. We then performed long-pcr-based amplification of the IMPA2 gene directly from genomic DNA. Based upon the cdna sequence, 16 several PCR primers were designed, of which five of the primer combinations (imp-b5/imp-b12, imp-b7/imp-b28, imp-b1/imp-b14, imp-b9/imp-b30 and imp-b11/imp-b22) yielded specific products from genomic DNA (Figure 1). The lengths of these PCR products (approximately 10 kb, 2.4 kb, 2.2 kb, 0.9 kb and 1.9 kb, respectively) were compared with the lengths of the PCR products amplified from cdna using the same primer-pairs (138 bp, 111 bp, 79 bp, 168 bp and 630 bp, respectively). The difference in size indicated the presence and extent of one or more introns. Fragments from the 3 -end of intron 1, the 5 -end and 3 -end of intron 5 and the region flanking the 3 -end of IMPA2 (0.3, 0.5, 0.6 and 3.5 kb, respectively; see Figure 1 and Table 1) were amplified by PCR-based, genomic walking. We were not able to amplify the upstream region of IMPA2, containing the promoter and the 5 -end of intron 1 by PCR-based techniques. A human genomic lambda library was therefore screened with an IMPA2- specific cdna-probe. One clone (No. 13) was found to contain the complete IMPA2 exon 1 with flanking regions (Figure 1). Characterization of the genomic structure of IMPA2 The various PCR fragments were partially sequenced to determine the exact organization of exons and introns in IMPA2. As shown in Figure 1 and Table 2, IMPA2 consists of eight exons ranging in size from 46 bp (exon 4) to 535 bp (exon 8). Introns 1 and 5 are of unknown size, whereas the remaining introns span from 0.7 kb (intron 6) to approximately 10 kb (intron

5 176 Figure 1 The genomic organization of the human myo-inositol monophosphatase gene 2 (IMPA2). Boxes represent the different exons in which the coding and untranslated sequences are shown as filled and open boxes, respectively. The approximate intron sizes (kb) are based upon agarose gel electrophoresis of restriction enzyme-digested PCR products; the sizes of intron 1 and intron 5 are not known. The localization and nature of nine IMPA2 polymorphisms are indicated. Arrowheads connected with lines show the sizes and localization of the overlapping PCR products used to characterize IMPA2. The lines with only one arrowhead indicate the PCR walking products of IMPA2. Each arrowhead with its B-number corresponds to an IMPA2- specific primer listed in Table 1. The position of lambda clone No. 13, containing the IMPA2 promoter region is also shown. 2) (Figure 1). All intronic splice sites were in accordance with the GT- (5 -end) and AG- (3 -end) consensus rule (Figure 2). The ATG start codon is located in the first exon. A TGA stop codon and a poly-a signal (AATAAA) are present in exon 8, and the poly-a tail of the transcript starts 14 bp downstream of the poly- A signal. By sequencing different EST clones, we confirmed that the IMPA2 transcript 16 contains an open reading frame of 867 bp and a 3 -UTR of 420 bp. Isolation and sequencing of 5 -RACE products obtained from human brain cdna made it possible to read cdna sequence 214 bp upstream of the ATG start codon. This result is compatible with a promoter predicting computer analysis (TFSEARCH), which indicates possible transcription start sites in positions 214 and 202 (Figure 3). It is also in agreement with the 1.5-kb total size of the IMPA2 transcript. 16 However, it should be noted that GC-rich, TATA-less promoters may have alternative transcription start sites. 25 We have also amplified all exons with flanking regions (Table 2) to verify both their size and sequence. It should be mentioned that during the DNA sequencing of 25 DNA samples, obtained from 23 bipolar patients and two normal controls, we found that all subjects were homozygous for a G in position 837, instead of a C as reported by Yoshikawa et al. 16 This difference does not alter the predicted amino acid in codon 279. Sequence analysis of the IMPA2 promoter region Figure 3 shows the sequence of the putative promoter region, exon 1 and the 5 -end of intron 1. The overall GC-content is 76%, and the region immediately upstream of exon 1 contains more than ten possible Sp1 elements and lacks a TATA-box. These features are typical for so-called housekeeping genes. 25 However, it should be noted that the data obtained by Yoshikawa et al 16 indicate that the expression of IMPA2 may vary widely between different tissues. It is therefore interesting to note that the IMPA2 promoter region and exon 1 contain several other possible consensus sites for transcription factors, such as GATA-1, MZF1, p300 and USF (Figure 3). In IMPA1, two tandemly organized sequences similar to the inositol/choline responsive element (ICRE) in Saccharomyces cerevisiae, are present in the 3 -UTR of the gene. 15 Such elements may serve as binding sites for the heterodimeric ino2p/ino4p transcription factor, which functions as a positive regulator of the expression of several phospholipid biosynthetic genes, in response to the cellular level of inositol and choline. Using the 5 -WYTTCAYRTGS consensus sequence of ICRE motifs, 26 we detected two nearby sequences in the promoter region, each having one mismatch (underlined; see below) as compared to the 11-bp consensus sequence. The two motifs, 5 -ATTTTACGTGG (sense direction) and 5 -ATTTCATAAGC (antisense direction) are separated by 34 bp, and are positioned approximately 1.3 kb upstream of the proposed transcription start site (see GenBank accession number AF for complete sequence of the promoter region). It should be noted, however, that our proposal of regulatory elements and binding sites for transcription factors are based on computer analysis only, and functional studies are necessary to identify those elements that are responsible for the regulation of IMPA2 expression. The evolutionary relationship between IMPA1 and IMPA2 By alignment of the open reading frames, an identity of 54% at the nucleotide level was found between IMPA1 and IMPA2. At the amino acid level the identity is 52%. The two genes differ markedly in their 5 -ends,

6 177 Figure 2 Alignment and comparison of the exon intron junction sites of IMPA1 and IMPA2. The exons are surrounded by boxes with their sizes given in parentheses. Vertical bars between the letters represent identical nucleotides. When known, sizes of the introns are given between the exons. where IMPA1 contains an untranslated exon and an extra intron, as compared to IMPA2 (Figure 2). The ATG codon of IMPA2 is localized 33 bp further upstream, relative to the IMPA1 start codon, resulting in a predicted protein with 11 additional residues in the amino terminal. Except for the first exon 1/intron 1 in IMPA1, the two genes exhibit an identical exon intron organization. Consequently, the corresponding exons 3 8 of IMPA1 and exons 2 7 of IMPA2 have exactly the same sizes (Figure 2), as is also the case for the coding parts of exon 9 of IMPA1 and exon 8 of IMPA2. In contrast, the untranslated parts of the exons and the introns of the two genes exhibit no similarity. Our findings strongly suggest that the two genes have evolved by duplication of a common ancestral gene. Three different enzymes with inositol dephosphorylating activity have earlier been isolated from the slime mold Dictyostelium discoideum. 27 Moreover, the tomato plant Lycopersicon esculentum has been reported to possess three different IMPA-like genes. 28 In humans an additional transcript almost identical to the IMPA1 mrna originates from chromosome 8q, but this transcript contains an early stop codon and seems to represent a pseudogene of IMPA1. 15 It is plausible that the ancestral IMPA gene duplicated early in evolution and as a result, species from different phyla have more than one IMPA gene encoding homologous enzymes with IMPase like functions. The possible difference in function, expression and lithium inhibition of the two human IMPase enzymes remains to be studied. Screening for polymorphisms in the IMPA2 gene Based on the assumption that IMPA2 may play a role in the etiology of bipolar disorder and in the therapeutic response to lithium-treatment, we searched for polymorphisms in the IMPA2 gene. DNA samples obtained from 23 well-characterized Norwegian manic-depressive patients 13 and two normal controls were employed in the screening study. Each of the exons 2 8, with at least 50 bp of flanking intronic sequence (Table 2), were amplified by PCR from genomic DNA followed by direct DNA sequencing of the PCR products. The promoter region and exon 1 remain to be screened due to technical difficulties in amplifying this very GC-rich region by PCR. As indicated in Figure 1, we detected nine different single nucleotide variations in the examined parts of IMPA2. In the two normal controls there were no variations, and in the 23 patients we found seven substitutions, one deletion and one insertion. The polymor-

7 178 Figure 3 The nucleotide sequence of the 5 -end of IMPA2. Capital letters indicate the sequence of exon 1. Putative regulatory elements as revealed by computer analysis (TFSEARCH) are underlined and labeled. Nucleotide positions are given relative to the first nucleotide in the ATG start codon. phisms were all verified by bi-directional sequencing and named according to the nomenclature system of the Nomenclature Working Group on Human Mutations. 29 In the coding regions, we identified a substitution of G to A in exon 5 (443G A), predicting an amino acid change from a basic arginine to a polar glutamine residue in codon 148 (R148Q). This amino acid substitution is apparently non-conservative, but the arginine residue in position 148 is not part of the highly conserved motifs which are suggested to be involved in the metal binding and the catalytic activity of human inositol phosphatases. 17 We also discovered a silent T to C transition (159T C) in exon 2. In the intronic sequence, there were G to A substitutions in the 3 -end of intron 1 (97 15G A), in the 5 -end of intron 2 ( G A) and in the 3 -end of intron 4 (382 44G A). The 97 15G A and G A substitutions occurred together in four patients, and three of these subjects did also have the G A variation. All patients were heterozygous for the G A substitutions. These data indicate that the three G A substitutions are in linkage disequilibrium. In the 5 -end of intron 5, we found an insertion of a single A-nucleotide ( A). We also detected three variations in the screened parts of the intron 6 DNA sequence, namely a single nucleotide deletion (599+75delT) and two different G to A substitutions (599+97G A and G A). The allele and genotype distributions of the nine polymorphisms observed among the 23 bipolar patients are given in Table 3. For each polymorphism there was one predominating allele with frequency ranging from 0.65 to The degree of variation in the coding region of IMPA2 appears somewhat greater

8 Table 3 The allele and genotype distributions of nine different IMPA2 polymorphisms in 23 bipolar patients Allele distribution (frequency) Genotype distribution 97 15G A G A GG GA AA 40 (0.87) 6 (0.13) T C T C TT TC CC 44 (0.96) 2 (0.04) G A G A GG GA AA 40 (0.87) 6 (0.13) G A G A GG GA AA 42 (0.91) 4 (0.09) G A G A GG GA AA 45 (0.98) 1 (0.02) insA A +A (0.98) 1 (0.02) delT +T T (0.93) 3 (0.07) G A G A GG GA AA 35 (0.76) 11 (0.24) G A G A GG GA AA 30 (0.65) 16 (0.35) than what we previously reported for the IMPA1 gene. 14 In the latter, we only detected one silent polymorphism in one subject after performing a cdna screening of 21 patients and 20 controls. However, both IMPA genes seem more resistant to DNA change than the third known gene that encodes a lithiuminhibitable, dephosphorylating enzyme in the phospholipase C signaling pathway, namely INPP1. Three silent and one missense polymorphism were found in the INPP1 coding region when examining the same 23 patients who have been used in the present study. 13 In conclusion, we have determined the genomic structure of the human IMPA2 gene and identified nine polymorphisms, including one that leads to an amino acid substitution. The genomic information presented here will be employed to screen for additional DNA variations in the promoter and exon 1 regions. To further explore the possible role of IMPA2 in the pathophysiology and drug response of lithium-treated manic-depressive illness, the polymorphisms should now be included in association studies of both family and case/control-based materials. Acknowledgements This study has been supported by Dr Einar Martens Research Fund, the Research Council of Norway (Mental Health Program) and in part by Eva Torhild s Memorial Fund for Medical Research. References 1 Goodwin FK, Jamison KR. Manic-depressive Illness. Oxford University Press: New York, Peselow ED, Fieve RR, Difiglia C, Sanfilipo MP. Lithium prophylaxis of bipolar illness. The value of combination treatment. Br J Psychiatry 1994; 164: Maj M, Pirozzi R, Magliano L. Late non-response to lithium prophylaxis in bipolar patients: prevalence and predictors. J Affect Disord 1996; 39: Vestergaard P, Wentzer LR, Brodersen A, Rasmussen NA, Christensen H, Arngrim T et al. Outcome of lithium prophylaxis: a prospective follow-up of affective disorder patients assigned to high and low serum lithium levels. Acta Psychiatr Scand 1998; 98: Manji HK, Potter WZ, Lenox RH. Signal transduction pathways. Molecular targets for lithium s actions. Arch Gen Psychiatry 1995; 52: Lenox RH, McNamara RK, Papke RL, Manji HK. Neurobiology of lithium: an update. J Clin Psychiatry 1998; 59: Berridge MJ, Downes CP, Hanley MR. Neural and developmental actions of lithium: a unifying hypothesis. Cell 1989; 59: Allison JH, Blisner ME, Holland WH, Hipps PP, Sherman WR. Increased brain myo-inositol 1-phosphate in lithium-treated rats. Biochem Biophys Res Commun 1976; 71: Berridge MJ, Downes CP, Hanley MR. Lithium amplifies agonistdependent phosphatidylinositol responses in brain and salivary glands. Biochem J 1982; 206: Dixon JF, Lee CH, Los GV, Hokin LE. Lithium enhances accumulation of [3H]inositol radioactivity and mass of second messenger inositol 1,4,5-trisphosphate in monkey cerebral cortex slices. J Neurochem 1992; 59: Kennedy ED, Challiss RA, Ragan CI, Nahorski SR. Reduced inositol polyphosphate accumulation and inositol supply induced by lithium in stimulated cerebral cortex slices. Biochem J 1990; 267: del Rio E, Shinomura T, van der Kaay J, Nicholls DG, Downes CP. Disruption by lithium of phosphoinositide signalling in cerebellar granule cells in primary culture. J Neurochem 1998; 70: Steen VM, Løvlie R, Osher Y, Belmaker RH, Berle J, Gulbrandsen AK. The polymorphic inositol polyphosphate 1-phosphatase gene as a candidate for pharmacogenetic prediction of lithium-responsive manic-depressive illness. Pharmacogenetics 1998; 8: Steen VM, Gulbrandsen AK, Eiken HG, Berle J. Lack of genetic variation in the coding region of the myo-inositol monophosphatase gene in lithium-treated patients with manic depressive illness. Pharmacogenetics 1996; 6: Sjøholt G, Molven A, Løvlie R, Wilcox A, Sikela JM, Steen VM. Genomic structure and chromosomal localization of a human myoinositol monophosphatase gene (IMPA). Genomics 1997; 45: Yoshikawa T, Turner G, Esterling LE, Sanders AR, Detera-Wadleigh SD. A novel human myo-inositol monophosphatase gene, IMP.18p, maps to a susceptibility region for bipolar disorder. Mol Psychiatry 1997; 2: York JD, Ponder JW, Majerus PW. Definition of a metaldependent/li(+)-inhibited phosphomonoesterase protein family based upon a conserved three-dimensional core structure. Proc Natl Acad Sci USA 1995; 92: Berrettini WH, Ferraro TN, Goldin LR, Weeks DE, Detera-Wadleigh S, Nurnberger JI et al. Chromosome 18 DNA markers and manicdepressive illness: evidence for a susceptibility gene. Proc Natl Acad Sci USA 1994; 91: Stine OC, Xu J, Koskela R, McMahon FJ, Gschwend M, Friddle C et al. Evidence for linkage of bipolar disorder to chromosome 18 with a parent-of-origin effect. Am J Hum Genet 1995; 57: Nöthen MM, Cichon S, Rohleder H, Hemmer S, Franzek E, Fritze J et al. Evaluation of linkage of bipolar affective disorder to chromosome 18 in a sample of 57 German families. Mol Psychiatry 1999; 4: Detera-Wadleigh SD, Badner JA, Berrettini WH, Yoshikawa T, Goldin LR, Turner G et al. A high-density genome scan detects evidence for a bipolar-disorder susceptibility locus on 13q32 and 179

9 180 other potential loci on 1q32 and 18p11.2. Proc Natl Acad Sci USA 1999; 96: Schwab SG, Hallmayer J, Lerer B, Albus M, Borrmann M, Honig S et al. Support for a chromosome 18p locus conferring susceptibility to functional psychoses in families with schizophrenia, by association and linkage analysis. Am J Hum Genet 1998; 63: Shamir A, Ebstein RP, Nemanov L, Zohar A, Belmaker RH, Agam G. Inositol monophosphatase in immortalized lymphoblastoid cell lines indicates susceptibility to bipolar disorder and response to lithium therapy. Mol Psychiatry 1998; 3: Ausubel FM, Brent R, Kingston RE, Moore D, Seidman JG, Smith JA et al. Current Protocols in Molecular Biology. John Wiley and Sons: New York, Azizkhan JC, Jensen DE, Pierce AJ, Wade M. Transcription from TATA-less promoters: dihydrofolate reductase as a model. Crit Rev Eukaryot Gene Expr 1993; 3: Schuller HJ, Richter K, Hoffmann B, Ebbert R, Schweizer E. DNA binding site of the yeast heteromeric Ino2p/Ino4p basic helix-loophelix transcription factor: structural requirements as defined by saturation mutagenesis. FEBS Lett 1995; 370: Van Dijken P, Bergsma JC, Hiemstra HS, De Vries B, Van Der Kaay J, Van Haastert PJ. Dictyostelium discoideum contains three inositol monophosphatase activities with different substrate specificities and sensitivities to lithium. Biochem J 1996; 314: Gillaspy GE, Keddie JS, Oda K, Gruissem W. Plant inositol monophosphatase is a lithium-sensitive enzyme encoded by a multigene family. Plant Cell 1995; 7: Antonarakis SE. Recommendations for a nomenclature system for human gene mutations. Nomenclature Working Group. Hum Mutat 1998; 11: 1 3.