Molecular cloning and chromosomal localization of a human gene encoding D-amino-acid oxidase.

Genomic clones covering the entire sequence of the gene encoding human D-amino-acid oxidase (DAO) (EC 1.4.3.3), one of the principal and characteristic flavoenzymes of peroxisomes, were isolated from human placental genomic libraries with the aid of a previously cloned cDNA for human DAO as a probe. Nucleotide sequence analysis revealed that the gene, present as a single copy in the human genome, comprises 11 exons and spans 20 kilobase pairs. The protein sequences containing the catalytically important residues, Tyr-228 and His-307, are coded for by separate exons. Heterologous transcription initiation sites were identified by primer extension analysis, and the sequence of the 5'-flanking region of the DAO gene was found to show some features common to other mammalian genes, such as those of glucocorticoid and the cAMP-responsive element. An additional noteworthy feature is the presence of promoter-like sequences in the first intron of the gene. In addition, two sequences of alternating pyrimidine and purine nucleotides, (CA)20 and (CA)17, are also present in the first intron. Such sequences may play some role in the expression of the DAO gene in human tissues. With the use of genomic DNAs prepared from human and Chinese hamster somatic hybrid cells as templates for the polymerase chain reaction, the gene for DAO was localized to human chromosome 12.

* This work was supported in part by Research Grant 2A-1 for Cardiovascular Diseases, Research Grant 9141202 for Aging and Health from the Ministry of Health and Welfare of Japan, and a grant from the Research Foundation for Cancer and Cardiovascular Diseases. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) Dl 1370.
To whom correspondence should be addressed. 'The abbreviations used are: DAO, D-amino-acid oxidase; kb, kilobase pair(s); bp, base pair(s); PCR, polymerase chain reaction; CRE, cyclic AMP responsive element; PIPES, piperazine-N,N'-bis(2ethanesulfonic acid). isomes (5), and DAO is regarded as a characteristic marker enzyme of the peroxisomes in porcine kidney. The enzyme is also known to be significantly decreased in patients with a subtype of Zellweger syndrome, a disorder of peroxisome formation (6). It has been assumed that the physiological role of DAO is related to the function of proximal tubules in the kidney and metabolism in peroxisomes, although D-amino acids do not appear to be involved in normal mammalian metabolism. Several approaches have been taken to obtain clues as to the physiological significance of this enzyme (7, 8); however, the biological function of DAO has remained undetermined.
Previously, we determined the primary structures of the porcine (9), human (lo), rabbit (ll), and mouse (12) DAO mRNAs by molecular cloning and sequence analysis of the respective cloned cDNAs. Chemical modification studies on DAO revealed several candidate amino acid residues for the active center of this enzyme (reviewed in Ref. 13). Moreover, D-propargylglycine, a suicide substrate, inactivates DAO on dynamic affinity labeling (14, 15), resulting in a great loss of activity and covalent modification at tyrosine 228 and histidine 307 of the primary structure of the porcine enzyme (16).
The availability of recombinant wild-type and mutant enzymes synthesized in vitro (17-19) and also produced in Escherichia coli (13,20) allowed us to re-examine the proposed active site of DAO, and we have been able to identify tyrosine 228 and histidine 307 as the most important residues for the catalytic function of porcine kidney DAO. Despite extensive studies on the enzymatic characteristics of DAO, little is known about the regulation of DAO expression, especially the mechanism underlying its induction or suppression at the molecular level, which appears to provide the molecular basis of the physiological function of DAO. In a functionally differentiated cell line, LLC-PK1, derived from porcine kidney proximal tubules, DAO activity was induced when a confluent monolayer was formed, and active biosynthesis of DAO was observed (21). In addition, in the ddY mouse strain, DAO expression was suggested to be controlled at the mRNA transcriptional level (22). In this study, to examine the structural organization of the human DAO gene and the regulation of its expression in a search for some clues as to the physiological function of DAO, we have isolated the human DAO gene from genomic libraries using cDNA as a probe. We have sequenced all of the coding and flanking regions of the human DAO gene and determined the transcription initiation site by primer extension analysis. In addition, we demonstrate that the gene for DAO is on human chromosome 12 through polymerase chain reaction analysis of human and Chinese hamster somatic hybrid cell lines.
H u m a n D-Amino-acid Oxidase Gene

EXPERIMENTAL PROCEDURES
Materials-Reagents were obtained as follows: [Y-~'P]ATP (specific activity, 111 TBq/mmol) from Du Pont/New England Nuclear; [a-"PIdCTP, specific activity, -111 TBq/mmol) from Amersham International; restriction endonucleases from Toyobo Co. and Takara Shuzou; T4 polynucleotide kinase, a random primer DNA labeling kit, a DNA ligation kit, and E. coli JM109 competent cells from Takara Shuzou; avian myeloblastosis virus reverse transcriptase from Midwest Bio-Products; an in uitro packaging kit from Stratagene; human kidney total RNA and a XEMBL-3 SP6/T7 human genomic library containing Sau3AI partial digests of human placental DNA from Clontech Laboratories; nitrocellulose filters (pore size, 0.45 pm) from Advantec and Schleicher & Schuell; agarose (type I) from Sigma; pGEM3Zf(+) and pGEM7Zf(+) plasmid DNAs from Promega Biotec; XgtWESXB phage DNA EcoRI arms from Bethesda Research Laboratories; a double-stranded nested deletion kit including exonuclease I11 from Pharmacia LKB Biotechnology Inc.; a somatic cell hybrid panel from Bios Co.; a PCR reagent kit and Taq DNA polymerase from Perkin-Elmer Cetus Instruments; a Taq dye primer cycle sequencing kit and a Taq dye-labeled dideoxynucleotide terminator cycle sequencing kit from Applied Biosystems, Inc.; and an imaging plate for an image analyzer from Fuji Photo Film. Oligonucleotides were synthesized with an automated DNA synthesizer (Applied Biosystems, Inc., Model 380A). All other chemicals used were of analytical grade and purchased from Sigma, Nakarai Tesque, or Wako Pure Chemical Co.
Southern Blot Hybridization Analysis of Human DAO Gene-High molecular weight DNA from human placenta was digested with EcoRI or BamHI, electrophoresed on a 0.7% agarose gel (3 to 10 pg/lane), and then transferred to a nitrocellulose filter according to Southern (23). The filter was hybridized with the ApaLI-PuuII fragment (1.4 kb) of human DAO cDNA (10) and then radiolabeled by the random hexamer primer method of Feinberg and Vogelstein (24). The hybridization conditions were as described previously (25), and several washing conditions were used to change the stringency: 0.1 X SSC (1 at 65 "C, 2 X SSC at 65 "C, and 2 X SSC at 55 "C. X SSC = 0.15 M sodium chloride, 0.015 M trisodium citrate, pH 7.0) Construction and Screening of Phage Genomic Libraries-Complete EcoRI digests of human placental DNA were size-fractionated by 0.5% agarose gel electrophoresis; DNA from the fractions containing -12-kb fragments was ligated into the EcoRI site of bacteriophage XgtWESXB; and then the resulting recombinant phage DNA was packaged in uitro. Unamplified libraries were screened by the plaque hybridization method of Benton and Davis (26) using 32P-labeled human DAO cDNA (10) containing the entire coding sequence (ApaLI-PuuII fragment, 1.4 kb) as a probe. Another XEMBL-3 SP6/ T7 human phage genomic library containing Sau3AI partial digests of human placental DNA was also screened by the same procedure as described above, except for the probe used. In this case, we used the 5"portion of a cDNA fragment (EcoRI-HincII, 0.59 kb) as a probe. The positive clones isolated from each library, designated as XHDAO-1 and XHDAO-5, respectively, were plaque-purified and amplified, and then the genomic inserts were subcloned for sequence analysis.
DNA Sequence Determination-DNA restriction fragments containing exons and flanking regions of the DAO gene were cloned into the pGEM7Zf(+) and pGEM3Zf(+) vectors. The 11.4-kb BanHI fragment of clone XHDAO-5, which was hybridized with the human DAO cDNA probe by Southern blot hybridization, was isolated and digested with EcoRI. The 9.4-and 2-kb fragments thus produced were then subcloned into the pGEM7Zf(+) vector. A series of unidirectional deletion mutant clones spanning both the 9.4-and 2-kb EcoRI-BamHI fragments was generated by digestion with exonuclease 111 (27). Doubled-stranded DNA templates prepared from mutant plasmids were sequenced by the dideoxynucleotide chain termination method (28) using fluorescently tagged M13 universal or reverse sequencingprimers. Enzymatic extension reactions for DNA sequencing were performed with Taq DNA polymerase and a DNA thermal cycler (Perkin-Elmer Cetus Instruments, Model PJ 2000) according to the manufacturer's recommendations. This thermal cycling reaction, based on the linear polymerase chain reaction, consists of steps programed in the step-cycle mode, with 15 cycles at 95 "C for 30 s, 55 "C for 30 s, and 70 "C for 1 min and with another 15 cycles at 95 "C for 30 s and 70 "C for 1 min, using -1.2-1.5 pg of doublestranded plasmid DNA as the template. The sequencing reaction products were analyzed automatically with a DNA sequenator (Applied Biosystems, Inc., Model 373A). Sequencing was performed with overlapping clones throughout, and synthetic oligonucleotides based on the previously determined sequences were synthesized and used as primers to facilitate sequencing. The 12-kb EcoRI fragment of clone XHDAO-1 was also isolated and digested with SalI. The 7-and 5-kb fragments thus produced were then subcloned into the pGEM3Zf(+) vector. Synthetic oligonucleotides specific for several of the exons were used as primers for the thermal cycling sequencing reactions with dye-labeled dideoxynucleotide terminators, and the reaction products were analyzed as described above. The nucleotide sequences of the exons were determined on both strands. The assembly of DNA sequences and homology searches of the GenBank" (release 69, September 1991) and EMBL (release 28, September 1991) data bases were performed with a computer program developed by Software Development Co. (Version 7.0).
Primer Extension Analysis-Synthetic oligonucleotides of 30 bp complementary to the 3'-end of exon U (5"CGGAAGTACCAA-ACTGCTGGAGAGACCAGA-3') were labeled at the 5'-ends with [Y-~'P]ATP and T4 polynucleotide kinase and used as primer. The labeled primer was then hybridized to 50 pg of total RNA isolated from human kidney for 12 h at 45 "C in 20 pl of a solution containing 50% formamide, 40 mM PIPES, pH 6.4,0.4 M NaCl, and 1 mM EDTA. The RNA/DNA hybrids were precipitated with ethanol and then subjected to the primer extension reaction with avian myeloblastosis virus reverse transcriptase. The extended cDNAs were electrophoresed on a 6% polyacrylamide gel containing 7 M urea along with dideoxy-DNA sequence samples prepared by the modified linear polymerase chain reaction method with the use of Taq DNA polymerase and a radiolabeled primer identical to that used for the primer extension described above. The gel was dried under reduced pressure, exposed to a photostimulatable phosphor imaging plate, and then analyzed with an image analyzer ( The samples were overlaid with 100 p1 of mineral oil to prevent condensation; placed in a preheated DNA thermal cycler at 94 "C; and then subjected to 30 cycles consisting of a 1-min denaturation period at 94 "C, a 2-min annealing period at 55 "C, and a 3-min period at 72 "C for extension of the annealed primers. Following amplification, the reaction products were electrophoresed on a 1.5% agarose gel and then analyzed by Southern blot hybridization as described above. Synthetic oligonucleotides (21-mer) corresponding to the middle portion of the target sequence were labeled at the 5'-ends with T4 polynucleotide kinase and used as probe.

RESULTS
Blotting Analysis of Chromosomal DNA-To determine the organization and structure of the human DAO gene, we first carried out Southern blot analysis of the chromosomal DNA from human placenta. As shown in Fig. lA, a simple  single positive band (12 kb) on blotting analysis, we purified the 12-kb DNA fragment by agarose gel electrophoresis and cloned it into the XgtWESXB phage. About 2.2 X loG plaques were screened by the in situ procedure with the "P-labeled cDNA fragment for human kidney DAO containing the whole coding region as a probe (ApaLI-PouII fragment, 1.4 kb), and eight positive independent phage clones were identified. Restriction endonuclease analysis of these clones revealed the presence of only one group of recombinant phages (designated a s XHDAO-1). Partial nucleotide sequence analysis and Southern blot hybridization analysis using an oligonucleotide probe corresponding to the 5"region of the coding sequence, however, indicated that this clone (XHDAO-1) contained only the 3"portion of the DAO gene and that the 12-kb positive band detected on Southern blot analysis (Fig. 1) represented two DNA fragments of identical size. Therefore, we screened another human placental genomic DNA library constructed with DNA partially digested with Sau3AI and cloned it into the XEMBL-3 SP6/T7 phage. About 2 X 10" plaques were screened by the same procedure as described above using the 5'-portion of the cDNA fragment (EcoRI-HincII fragment, 0.59 kb) as a probe, with two independent, but identical phage clones being identified (designated as XHDAO-5). Restriction endonuclease analysis revealed that XHDAO-1 and XHDAO-5 overlapped each other and contained the same 2-kb EcoRI-RamHI fragment, as shown in Fig. 2. The inserts were separately cloned into the pGEM3Zf(+) and pGEM7Zf(+) vectors for sequence analysis. Structure of Human DAO Gene-The human DAO cDNA sequence previously determined (10) was aligned with the chromosomal genomic sequence. The two sequences matched well, but the 5"untranslated region of the cDNA was found to be interrupted a t 9 bp upstream of the translation initiation codon. Therefore, we sequenced all the 5"upstream region and found an mRNA-coding segment corresponding only to the 5"untranslated region (designated as exon U) a t -5 kb upstream of the translation initiation site. The nucleotide sequences of the chromosomal genomic exons and the cDNA showed perfect agreement. Thus, the DAO gene, spanning -20 kb of the human genome, consists of 10 exons for the protein-coding region and the 3'-untranslated region and another one for the 5'-untranslated region. All 11 exons are in the two DNA inserts, and the gene contains 10 introns. The organization of the DAO gene and a partial restriction map are presented in Fig. 2.
The transcription initiation site of the DAO gene was mapped by primer extension of human kidney total RNA using reverse transcriptase. The noncoding sequence was extended from the oligonucleotides corresponding to the 3'-end of exon U. As shown in Fig. 3, the transcription initiation sites of human DAO transcripts were heterologous, and four sites were determined. The total length of exon U is 134 bp. assuming that transcription begins with t,he first T nucleotide revealed by the primer extension assay. In addition, alignment of the sequences of the porcine and mouse cDNAs with that of human genomic DNA allows assignment of putative start sites for the porcine and mouse mRNAs, both of which reside in exon U (Fig. 4). On the basis of these assignments, the  Histidine 307, which was suggested to play a subsidiary role in the catalytic reaction together with tyrosine 228, was also coded for by exon 10. The exon-intron junctions of the human DAO gene and the organization of each exon are summarized in Fig. 5. Analysis of Nucleotide Sequence of 5'-FEanking Region-To determine the nucleotide sequence common to the regulatory regions of eukaryotic genes, sequence determination for the 5"flanking region of the human DAO gene was extended up to nucleotide 247 upstream of the capping site determined by primer extension analysis as described above. As shown in Fig. 4, the 5'-flanking region of the human DAO gene was observed to have a high G + C content (61.94%). A TATA box sequence, which determines the specificity of mRNA synthesis initiation by RNA polymerase 11, was not found in the 5'-flanking region of the DAO gene. Instead, sequence 5'-TCCATA-3' was found at positions -23 to -18. Therefore, the heterologous transcription initiation observed on primer extension analysis may have resulted from the lack of a welldefined TATA box motif. In addition, there is a G + C-rich region between positions -93 and -82 that contains a GC box homologous sequence in its core (5-out of 6-base pair match with the consensus sequence). Moreover, a sequence resembling that of a canonical CAMP-responsive element, 5'-TGAgaTCT-3' (mismatches are shown as lower-case letters hereafter), is present between positions -145 and -138. When  we searched the 5"flanking region for hormone-responsive element consensus sequences, a sequence including 7 nucleotides in common with the 12-nucleotide consensus glucocorticoid responsive element (5'-GgTAcANNNTGTTcT-3') (32) was found between positions -220 and -206 (5'-aGacCtTCTTGTgCT-3'). This sequence contains the 4 nucleotides determined to be the most important in binding to the glucocorticoid receptor at identical positions (underlined). Another hormone-responsive element-like sequence was present between positions -60 and -53 (5'-GTGCaGTG-3'), which is homologous to the sterol-dependent repressor sequence present in the human low-density lipoprotein receptor and hydroxymethylglutaryl-CoA reductase genes. Furthermore, the 5"flanking region of the human DAO gene exhibited significant homology as a whole (51.4% over 243 bp) to the putative regulatory region (positions -311 to -568) of the human corticotropin @-lipoprotein precursor gene (33) when the two sequences were aligned with the introduction of gaps to maximize homology (data not shown). Identification of Internal Promoter-like Region in First Intron of DAO Gene-When we searched for promoter-like regions within the DAO gene, three copies of a TATA box sequence and one copy of a CAAT box sequence were found in the first intron downstream from the mapped transcription initiation site. As shown in Fig. 6, the first TATA box upstream of the translation initiation codon is at position 4997 and is preceded by a CAAT box sequence at position 4133. The second TATA box is at position 3790, and the third one at position 2770. In addition, an AP-1-binding site is at position 2036 (5"TGAGTCAG-3')) and two copies of an SP- 1- Chromosomal Localization of Human DAO Gene-To establish the chromosomal location of the human DAO gene, we examined somatic cell hybrids as to their human chromosome contents using the polymerase chain reaction to specifically amplify human sequences, and not their Chinese hamster counterparts. DNA samples prepared from human cells, Chinese hamster cells, and 25 somatic cell hybrids were used as templates for PCR analysis. The target sequence for amplification was selected from the exon 10 sequence since this exon is the largest in size and contains the 3"untranslated region. A pair of 21-mer oligonucleotides (Fig. 7B) spanning a 338-bp fragment were synthesized, one corresponding to the 5'-end of the exon 10 coding sequence and the other to the 3"untranslated region. Another 21-mer oligonucleotide (Fig.  7B) corresponding to the middle portion of the target sequence in exon 10 was also synthesized for use as the hybridization probe for the sensitive and specific detection of the amplified DNA fragment. Assignment of the desired PCR product to a cell line in the somatic cell panel was thus The terminator codon (TGA) is in shaded. The 5"primer and probe were synthesized in the sense strand, whereas the 3"primer was synthesized in the antisense strand.

CCTCTCCTAG IGTCATCCACAAC---. Exon lo----U G G T T C T G A A A A C T C C T G C
accomplished by simple visualization of bands on autoradiography. PCR was carried out under standard conditions (29) using 50 ng of hybrid DNA, in a total volume of 100 p1. The reaction products of the amplification were electrophoresed on agarose gel and then analyzed by Southern blotting using internal oligonucleotides as hybridization probe. As shown in Fig. 7A, the human DNA was specifically amplified and detected as a band of the expected size (338 bp). No band was observed for Chinese hamster parent cell lines, confirming the absence of cross-species homology for the selected primer set. The results of Southern blot analysis of the human DNA, hamster DNA, and DNAs from 25 somatic cell hybrids are summarized in Table I. Although human chromosome 5 is very frequently retained in this hybrid panel, the results suggest strongly overall that DAO is on human chromosome 12.

DISCUSSION
We have determined the nucleotide sequence of the human DAO gene and analyzed its organization. Perfect agreement between the nucleotide sequences of the chromosomal gene exons and the cDNA sequence was observed. The human genome contains a single copy of the DAO gene, as was also -acid Oxidase Gene the case with the porcine genome (9). We could not find any cross-hybridized DNA segment, even under nonstringent hybridization conditions. Jacobs et al. (34) previously elucidated the partial nucleotide sequence of a porcine DAO genomic clone to determine the nucleotide sequence of the mRNA. Since they could not obtain full-length cDNA for DAO, no information on the exon segment in the porcine genome corresponding to the 5"untranslated region of the mRNA was reported; only the structures of exons 1-4 of the porcine DAO gene were described. The presence of a 5'-noncoding exon (exon U) at -5 kb upstream of the translation initiation codon was demonstrated in the human genome. In addition, a sequence homologous to a portion of the 5'-untranslated region of the porcine DAO mRNA was found in the first intron of the human DAO gene. A corresponding sequence in porcine mRNA was present in a 3.3-kb mRNA, but absent in a 2.7-kb mRNA, which suggests that alternative splicing of this DNA segment in the porcine genome might take place and contribute in part to the heterogeneity of mRNA species in porcine kidney. Elucidation of the structure of the porcine DAO gene would be necessary to confirm this possibility. The determination of exon-intron junctions is also important since there is convincing evidence that, in a number of proteins, structural or functional units are encoded by discrete exons with splice junctions coincident with the boundaries of the protein motif. In the human DAO gene, a putative hydrophobic FAD-binding region is encoded by exon 1. The 2 chemically modifiable amino acid residues, Tyr-228 and His-307, are encoded by exons 7 and 10, respectively. These 2 residues, which are completely conserved among the four species of DAO examined, are modified by D-propargylglycine, a suicide substrate, and thus are suggested to play important roles in the catalytic function of DAO. Exon 10 exhibited several interesting structural features indicating some role in the expression of the DAO gene, such as the presence of His-307, the carboxyl-terminal tripeptide sequence of Ser-His-Leu for the translocation into peroxisomes (35), the hexanucleotide sequence ATTAAA for the polyadenylation signal, and the relatively large size of 514 bp. The organization of the DAO gene may also provide some clues as to the phylogenetic relationship of the DAO gene with other genes.
Multiple initiation sites for human DAO transcription were observed on primer extension analysis. This is consistent with common structural properties in the 5"flanking region of the DAO gene, namely the absence of well-defined TATA and CAAT boxes. Similar heterogeneity of the transcription initiation sites was also reported for a mouse immunoglobulin Cp transcript (36). It would be interesting to determine whether or not any stimulation could induce predominant transcription from one of the four initiation sites. The 5'flanking region of the human DAO gene was compared with the nucleotide sequence data bases to see whether or not it shared common regulatory sequences. Homology between the human DAO and human corticotropin @-lipoprotein precursor genes was significant immediately upstream of the transcription initiation site. The production of corticotropin and related peptides in the pituitary is regulated negatively by glucocorticoids and positively by corticotropin-releasing factor. A sequence homologous to a glucocorticoid-responsive element was also found in the 5"flanking region. Moreover, there is another noteworthy DNA segment homologous to the sequence necessary for induction by CAMP. It is not known yet whether or not these regions are important in the transcriptional regulation of the DAO gene, but it is tempting to speculate that glucocorticoids and CAMP might be involved in the modulation of the gene expression of DAO. Identifica-