Molecular Cloning of cDNA and Chromosomal Assignment of the Gene for Human Phenylethanolamine N-Methyltransferase, the Enzyme for Epinephrine Biosynthesis*

Phenylethanolamine N-methyltransferase (PNMT; catalyzes the synthesis of epinephrine from norepinephrine, the last step of catecholamine biosynthesis. To isolate a cDNA clone for human PNMT, we first isolated a cDNA clone for bovine adrenal medulla PNMT using mixed oligodeosyribonu- cleotide probes whose synthesis was based on the partial amino acid sequence of tryptic peptides from the bovine enzyme. By screening a bovine adrenal medulla cDNA library, a cDNA clone with an insert of about 200 base pairs (bp) was isolated. This clone consisted of 84 bp of carboxyl-terminal coding region, which contained amino acid sequences corresponding to two tryptic peptides, and about 100 bp of 3”untranslated region. Using this cDNA fragment as the probe, we screened a human pheochromocytoma cDNA library and isolated

EC 2.1.1.28) catalyzes the synthesis of epinephrine from norepinephrine, the last step of catecholamine biosynthesis. To isolate a cDNA clone for human PNMT, we first isolated a cDNA clone for bovine adrenal medulla PNMT using mixed oligodeosyribonucleotide probes whose synthesis was based on the partial amino acid sequence of tryptic peptides from the bovine enzyme. By screening a bovine adrenal medulla cDNA library, a cDNA clone with an insert of about 200 base pairs (bp) was isolated. This clone consisted of 84 bp of carboxyl-terminal coding region, which contained amino acid sequences corresponding to two tryptic peptides, and about 100 bp of 3"untranslated region. Using this cDNA fragment as the probe, we screened a human pheochromocytoma cDNA library and isolated a cDNA clone with an insert of about 1.0 kilobase pairs, which contained the complete coding region of the enzyme. Northern blot analysis of human pheochromocytoma poly(A)+ RNA using this cDNA insert as the probe showed a single RNA species of about 1,000 nucleotides, suggesting that this clone is a fulllength cDNA. Determination of the nucleotide sequence revealed that human PNMT consists of 282amino acid residues with a predicted molecular weight of 30,853, including initial methionine. The amino acid sequence of the human PNMT was highly homologous (88%) to that of the bovine enzyme. Chromosomal assignment of the gene for human PNMT was carried out using mouse-human somatic cell hybrids. The PNMT gene was assigned to chromosome 17. on Priority Areas, Ministry of Education, Science, and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "uduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 503727.
** To whom correspondence should be addressed.
The abbreviations used are: PNMT, phenylethanolamine Nmethyltransferase; SDS, sodium dodecyl sulfate; HPLC, high performance liquid chromatography; bp, base pairs; kb, kilobase pairs. to form epinephrine, using S-adenosyl-L-methionine as the methyl donor (1, 2). The enzyme is found predominantly in chromaffin cells of adrenal medulla, in which epinephrine is synthesized as a hormone. It is also distributed in adrenergic neurons of the medulla oblongata and hypothalamus in the brain (3-5), where epinephrine functions as a neurotransmitter. Although the details of epinephrine neuron function in the brain are not yet clearly understood, these neurons may be involved in important neurophysiological functions such as cardiovascular and neuroendocrine regulation of the central nervous system (6). Several lines of evidence for the involvement of brain epinephrine neurons in blood pressure regulation have been presented in studies using genetically hypertensive rats, in which both PNMT activity and epinephrine level were elevated in discrete regions of the medulla oblongata and hypothalamus (7-10). Development of the enzyme activity is regulated by glucocorticoids in embryonic adrenal gland (11,12), and the enzyme activity in cultured chromaffin cells and in superior cervical ganglia is induced by nerve growth factor (13,14). The enzyme has been purified to homogeneity from bovine and rabbit adrenal medulla (15-18). It is a monomeric protein with a molecular weight of 30,000-31,000 (16, 18), and the enzymatically active charge isozymes have been noted in bovine and rabbit adrenal medulla (15,18,19).
Molecular cloning of the enzyme's cDNA should facilitate further studies on hormonal and developmental regulation of PNMT gene expression and genetic analysis of the isozymes as well as the structural basis of the enzyme reaction. As a first step, we describe here the complete nucleotide sequence of human PNMT cDNA and the deduced amino acid sequence of the enzyme. We also describe for the first time the chromosomal assignment of the gene for this enzyme.

Purification and Partial Amino Acid Sequences of Bovine
Adrenal PNMT-PNMT was purified to homogeneity from the 100,000 x g supernatant of bovine adrenal medulla by means of DEAE-Sepharose chromatography followed by ammonium sulfate fractionation and gel filtration on a Sephadex Portions of this paper (including "Materials and Methods," Tables 1-111, and Figs. 1, 2, 4, and 5 ) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press. G-100 column. The purified preparation showed a single band on SDS-polyacrylamide gel electrophoresis and had a specific activity of 740 nmol/h/mg protein. This is comparable to values reported for purified enzymes (15,18). About 100 g of adrenal medulla yielded about 30 mg of the pure enzyme. The pure enzyme preparation was reduced and carboxymethylated for trypsin digestion. The tryptic peptides were separated on a reverse-phase HPLC. Table I (see Miniprint) shows the amino acid sequences of four tryptic fragments. An attempt to sequence the amino-terminal region of the bovine enzyme was unsuccessful due to blocking of the amino terminus.
Isolutwn and Identification of cDNA Clone Encoding Bovine Adrenal PNMT-An amino acid sequence of peptide T4 seemed to be the most suitable for synthesizing oligodeoxyribonucleotide probes, because only 16 kinds of nucleotides could represent all possible nucleotide sequences corresponding to this amino acid sequence. The following mixed tetradecamer was chemically synthesized Bovine adrenal medulla cDNA library was first screened by colony hybridization with these synthetic probes. We isolated two hybridization-positive clones from about 5000 transformants. The inserts of these two clones had the same size (about 200 base pairs (bp)) and showed the same restriction map, suggesting that they are identical. Therefore, only one clone, named pbPNMT 14, was subjected to nucleotide sequence analysis. As shown in Fig. 1 (Miniprint), pbPNMT 14-insert cDNA contained nucleotide sequences corresponding to tryptic peptides T3 and T4 and poly(A)+ tail, although the NH2terminal amino acid residue of T3, Thr, was absent from the cDNA clone. In addition, the tryptophanyl residue in T4, which had been determined by amino acid sequencer, was found from the nucleotide sequence to actually be arginyl residue. Based on the coding frame of these two peptides, translational stop codon TGA was found at nucleotide position 85-87. These facts indicated that this clone represented a partial cDNA fragment for bovine PNMT consisting of 84 bp of carboxyl-terminal coding region and about 100 bp of complete 3"untranslated region.
Isolution and Churacterizatwn of cDNA Clone Encoding Human PNMT-The finding that the clone pbPNMT 14 contains a partial cDNA fragment encoding bovine PNMT allowed us to screen human pheochromocytoma cDNA library using this cDNA insert as the probe to isolate a human cDNA clone. After screening about 6 X lo5 recombinant phages, 12 hybridization-positive clones were isolated. Of these, four clones with the -1.0-kilobase pair (kb) insert were subcloned into Bluescript M13(+) and (-) vectors. Analysis of restriction mapping and determination of nucleotide sequence at 5'terminal region of these cDNA inserts suggested that two clones with identical restriction maps contained inserts of cDNA encoding the enzyme's entire coding region. One of these clones, designated phPNMT 901, was further analyzed for complete nucleotide sequence. This clone contained a cDNA insert of about 1.0 kb, including about 100 bp of poly(A)+ tail.
Nucleotide and Deduced Amino Acid Sequence of Human PNMT cDNA- Fig. 2 (Miniprint) shows partial restriction map and sequence strategy of the cloned phPNMT 901 insert. The sequence was determined on both strands of the cDNA, crossing restriction enzyme sites. The complete nucleotide sequence and deduced amino acid sequence are shown in ? k c ser C l y n l a ASP A r g Ser P r o A s n A l a G l y A l a A l a Pro ASP Ser A l a Pro G i y G l n 10 ccc LCC GTG GCT TCG GCC TAC CAG CGC T I C GAG CCG CGC GCC TAC CTC CGC M C AAC TAC A l a la V a l A l a Ser A l a T y r Gln A r g P h e Glu Pro A r g A l a T y r L e u A r g ASn ASn 3. This clone contains 846 bp of a single long open reading frame, starting with an ATG at position 1-3 and ending with a TGA stop codon at position 847-849. The nucleotide sequence corresponding to the COOH-terminal region of the protein at position 766-849 including the stop codon was found to be highly homologous to that of the bovine PNMT cDNA, pbPNMT 14, indicating that this is a cDNA clone encoding human PNMT. The assignment of this initiation site is supported by the fact that the nucleotide sequence preceding the ATG codon, GCAGC, agrees well with the consensus sequence CCACC, in which adenine at position -3 is highly conserved, found in many eukaryotic mRNA (33,34). This open reading frame codes for a protein with 282amino acid residues and a predicted molecular weight of 30,853, including the initial methionine. The clone had only 6-bp 5'-untranslated and 84-bp 3"untranslated sequences. A consensus polyadenylation signal AATAAA (35) is located 11 nucleotides upstream of the poly(A)+ addition site.
Comparison of Amino Acid Sequence and Amino Acid Composition of Human and Bovine PNMT-The nucleotide sequence of cDNA for bovine adrenal PNMT and the deduced amino acid sequence have recently been reported (36). However, we found this sequence to be incorrect in the COOHterminal portion of the protein due to possible misreading of the nucleotide sequence (see "Discussion"). Fig. 4 (Miniprint) shows the amino acid sequence of human PNMT together with that of the bovine enzyme, of which the latter was corrected by assuming sequence homology. For maximum homology, only one gap was inserted at the carboxyl-terminal region in human PNMT. The sequences differ only at 33 positions, and about 88% homology was observed between the two species. The amino acid sequence was completely conserved at positions 25-47,52-83, and 171-229, while a certain divergency was observed at positions 4-24,111-145, and 272-283. The areas of conservation would be important for the enzymatic function.
The amino acid compositions of human and bovine PNMT predicted from the nucleotide sequences were compared (Table 11). The human and bovine enzymes were very similar, reflecting a high degree of sequence homology (88% homology, Fig. 4). The amino acid composition of the purified enzyme almost completely agreed with that deduced from the cDNA sequence of the bovine enzyme modified from the original sequence (see "Discussion"). Relatively high contents of arginine, glutamic acid, and leucine were observed.
RNA Blot Analysis-In order to examine the length of PNMT mRNA, poly(A)+ RNA extracted from human pheochromocytoma was subjected to Northern blot analysis using the 32P-labeled phPNMT 901 insert as the probe. As shown in Fig. 5 (Miniprint), a single band corresponding to about 1.0 kb was observed, almost the name size as the cDNA insert of phPNMT 901, suggesting that the cDNA insert described above is a full-length cDNA.
Chromosomal Assignment of Human PNMT Gene-To carry out chromosomal assignment of PNMT gene, 15 DNAs from mouse-human somatic cell hybrids with selected human chromosomes were subjected to Southern blot analysis using the phPNMT 901 cDNA insert as a probe. EcoRI-digested human genomic DNA gave a single band at 21 kb, while mouse DNA also gave a signal at 25 kb due to cross-hybridization. The presence of this human-specific band in these hybrid cells is listed in Table 111. From the karyotype analysis data for the hybrid cells, the human PNMT gene was assigned to chromosome 17.

DISCUSSION
This paper describes the isolation and nucleotide sequence of a full-length cDNA clone encoding human PNMT. This cDNA clone was isolated by cross-hybridization with a partial bovine cDNA fragment in a clone, pbPNMT 14, which was also isolated in the present work. The identity of the latter clone was established by the presence of amino acid sequences of two tryptic peptides determined by direct analysis (see Fig.  1). The complete primary structure of human PNMT was deduced from the nucleotide sequence of the cDNA. The enzyme consisted of 282 amino acid residues with a molecular weight of 30,853, including initial methionine, which is in good agreement with the value (30,000-31,000) determined by SDS-polyacrylamide gel electrophoresis for the purified enzyme (16, 18).
Recently, Baetge et al. (36) reported the nucleotide sequence of cDNA encoding bovine adrenal PNMT. The nucleotide sequence reported by Baetge et al. (36) differs somewhat from our sequence corresponding to the COOH-terminal portion of the bovine enzyme. Comparison of the amino acid sequence of bovine PNMT deduced by Baetge et al. (36) with our amino acid sequence of tryptic peptides obtained from the pure preparation of the bovine enzyme revealed that peptides T1 and T2 could be assigned to amino acid positions 30-33 and 154-171, respectively (Fig. 4). However, peptides T3 and T4 could not be identified in Baetge's sequence, although the carboxyl-terminal sequence of peptide T4, -Ala-Gln-Lys-Lys-OH, was assigned to their amino acid position 278-281. When our nucleotide sequence for the bovine cDNA clone, pbPNMT 14, is compared with their sequence in the region correspond-ing to our clone, our nucleotide sequence ACC = GCC at nucleotide position 58-66 in Fig. 1 was ACC TG GCC in their sequence. The former sequence should becorrect for the following reasons. First, our nucleotide sequence can completely code for the amino acid sequence of peptide T4. If their sequence were correct, translational frame-shift would occur and, therefore, peptide T4 could not be assigned to any other portion of the protein. Second, AuaI cleaved the cDNA at the nucleotide positions 58-66. Since the recognition sequence for this restriction endonuclease is CYCGRG, where the capitals Y and R denote a pyrimidine and a purine nucleotide, respectively, this finding is consistent with our sequence but not with theirs. There is another apparent frame-shift due to possible misreading of the nucleotide sequence at amino acid position 246-248 in Baetge's report (Fig.  1 of Ref. 36). The sequence AGC TAT seems likely to be AGC GGC TAT, because insertion of one guanine at this position allows a high degree of amino acid sequence homology between human and bovine PNMT at the COOH-terminal position 247-283, although we have no direct data on the nucleotide sequence of this region. The validity of the argument is also confirmed by the amino acid composition (Table  11). Amino acid composition predicted from the corrected amino acid sequence of the bovine enzyme as described above coincided almost completely with that obtained by direct analysis of the pure enyzme. In contrast, the amino acid sequence as previously described (36) should have resulted in much more Cys, His, and Leu and less Val and Phe than the observed values. Taking these facts into account, we aligned amino acid sequences of human and bovine PNMT deduced from cDNAs so as to perform the maximum matching (Fig.  4).
We searched the GenBank and NBRF-PIR data bases with pVAX computer using SEQF and SEQFP subprograms of IDEAS homology search program and were unable to identify any sequences homologous to the nucleotide or predicted amino acid sequence of human PNMT. Recently, cDNA clonings of &adrenergic receptor (37, 38) and hydroxyindole 0-methyltransferase (39) have been reported. Since the former binds norepinephrine as a ligand and the latter binds Sadenosylmethionine as a substrate, the primary structure of these macromolecules may have some regional homology to that of PNMT, which has binding sites for both compounds. However, we could not find significant homology between these molecules and human PNMT. The apparent lack of homology between bovine hydroxyindole 0-methyltransferase and bovine PNMT has also been reported (39). These facts suggest that the regional homology under statistical significance and/or three-dimensional configuration of peptide backbone will be responsible for constructing active binding sites. It has been suggested that the catecholamine pathway enzymes, tyrosine hydroxylase, dopamine 0-hydroxylase, and PNMT, might be structurally related and thus have evolved from a common ancestral gene (40). We compared the predicted amino acid sequences of human tyrosine hydroxylase (27,41,42) and human PNMT and found no significant homology between them.
Several studies have shown that PNMT from bovine and rabbit adrenal medulla exists as differently charged isozymes (15, 18, 19), and Wong et al. (18) suggested a difference in primary structure based on peptide mapping of the bovine isozymes. As for the human enzyme, it is not known whether it exists as isozymes or not. Although multiple forms of PNMT mRNA were not observed by Northern blot analysis (Fig. 5), we are now screening the cDNA library further to elucidate this problem.

7675
The present paper describes for the first time the chromosomal assignment of PNMT gene. The genes for the enzymes of the catecholamine pathway, tyrosine hydroxylase, dopamine @-hydroxylase, and PNMT, are located on different chromosomes: chromosome 11 (43), 9 (44), and 17, respectively. The cloned cDNA for human PNMT reported here provides a useful tool for studies on regulation of gene expression. Mouse-human somatic c e l l h y b r l d s w i t h selected human chromosomes