Proc. Nati. Acad. Sci. USAVol. 88, pp. 1416-1420, February 1991Neurobiology

Human catechol-O-methyltransferase: Cloning and expression of themembrane-associated form

(nucleotide sequence/primary structure/enzymatic activity/inhibition)

BARBARA BERTOCCI*, VINCENZO MIGGIANOt, Most DA PRADA*, ZLATKO DEMBICt, HANS-WERNER LAHMt,AND PARICHEHR MALHERBE*t*Pharma Research Central Nervous System and tDepartment of Biology, Central Research Units, F. Hoffmann-La Roche Ltd, CH4002 Basel, Switzerland

Communicated by Erminio Costa, November 6, 1990

ABSTRACT A cDNA clone for human catechol-O-methyltransferase (hCOMT; S-adenosyl-L-methionine:cate-chol O-methyltransferase; EC 2.1.1.6) was isolated from ahuman hepatoma cell line (Hep G2) cDNA library by hybrid-ization screening with a porcine cDNA probe. The cDNA clonewas sequenced and found to have an insert of 1226 nucleotides.The deduced primary structure of hCOMT is composed of 271amino acid residues with the predicted molecular mass of 30kDa. At its N terminus it has a hydrophobic segment of 21amino acid residues that may be responsible for insertion ofhCOMT into the endoplasmic reticulum membrane. The pri-mary structure of hCOMT exhibits high homology to theporcine partial cDNA sequence (93%). The deduced amino acidsequence contains two tryptic peptide sequences (T-22, T-33)found in porcine liver catechol-O-methyltransferase (COMT).The coding region of hCOMT cDNA was placed under thecontrol of the cytomegalovirus promoter to tranfect humankidney 293 cells. The endogenous COMT activity, which wasw9.98 units per mg of protein in the untransfected cells,increased to 206 units per mg of protein upon transfection witha plasmid containing the COMT cDNA. The COMT activity ofrecombinant protein was inhibited competitively (IC50 = 700nM) by the selective COMT inhibitor Ro 40-7592. An anti-COMT monoclonal antibody recognized, on immunoblots, amajor polypeptide with apparent molecular mass of 29 kDa, inreasonable agreement with the predicted molecular mass. Therecombinant hCOMT was shown by immunoblot analysis to bemainly associated with the membrane fraction. RNA blotanalysis revealed one COMT mRNA transcript of 1.4 kilobasesin Hep G2 poly(A)+ RNA.

Catechol-O-methyltransferase (COMT; S-adenosyl-L-methi-onine; catechol O-methyltransferase; EC 2.1.1.6) is an en-zyme that catalyzes the transfer of a methyl group fromS-adenosyl-L-methionine to the m-hydroxy group of cate-cholamine neurotransmitters (dopamine, noradrenaline,adrenaline), their metabolites, and L-dopa, thereby inactivat-ing them. The enzyme has a broad substrate specificityaccepting as substrate also catechol steroids, a-methyldopa,and apomorphine (1). It is widely distributed in variouscerebral and extracerebral tissues of all mammalian species(2), including erythrocytes (3). The occurrence of at least twodistinct isoforms ofCOMT has been demonstrated, of whichone is soluble (S-COMT) and the other membrane-bound(MB-COMT) and whose relative abundance differs in varioustissues and species. The biochemical characteristics of MB-COMT are less extensively investigated. This isoenzyme islocalized in the microsomal fraction of different tissues, suchas liver (4), brain (5), erythrocytes (6), and appears to be anintegral membrane protein structurally distinct from the

cytosolic S-COMT (7). Although both forms of COMT cat-alyze the 0-methylation of catecholamines, MB-COMT hasa much higher affinity for its substrates (6, 8) and might,therefore, contribute markedly to the metabolism of lowconcentrations of endogenous catecholamines in central ner-vous system and peripheral tissues.

Studies of COMT activity in erythrocytes suggest thatindividual variations are genetically determined. Data fromsegregation analysis provided evidence for the control ofCOMT activity by a major autosomal locus with two alleles(9). Variations in COMT activity are found in patients withmajor depression, recurrent and bipolar disorders (10), aswell as in children with Down syndrome (11). Differences inthe enzymatic activity in human individuals may be involvedin the pathogenesis of various psychiatric and neurologicaldiseases (12). To study the molecular genetics of thesedisorders, it is necessary to isolate the gene or genes codingfor cytoplasmic and membrane-bound COMT.

In the present communication, we describe the molecularcloning of a cDNA encoding the human liver COMT. Se-quence analysis and expression of the clone revealed that wehave isolated the cDNA coding for the membrane-boundenzyme.§

METHODSConstruction and Screening of cDNA Libraries. The Agt1

expression library from porcine liver was purchased fromClontech. By immunizing a rabbit with the highly purifiedsoluble form of porcine liver COMT we obtained an antise-rum that recognized (in ELISA and in immunoblot) theporcine enzyme and inhibited its activity. Immunoscreeningwith this antiserum was done according to Young and Davis(13). The Agtll library from Hep G2 cells was constructedand screened following described procedures (14).RNA Blot Analysis. Total RNA was extracted from a

human hepatoma cell line (Hep G2, ATCC HB 8065) by theguanidinium isothiocyanate/cesium chloride method (15).Poly(A)+ RNA was isolated by oligo(dT)-cellulose chroma-tography (16).

Poly(A)+ RNA (-5 gg) from Hep G2 cells was denaturedwith glyoxal, electrophoresed through 1.1% agarose contain-ing 25 mM NaH2PO4 (pH 7.0) (17), transferred to an Amer-sham Hybond-N filter (RPN 1520N), and hybridized with a32P-labeled nick-translated cDNA insert of AhCOMT F3.

Construction of Expression Plasmid Carrying HumanCOMT (hCOMT) cDNA. A 1230-base pair (bp) EcoRI frag-ment (hCOMT F3) cDNA was subcloned by blunt-ended

Abbreviations: COMT, catechol-O-methyltransferase; hCOMT, hu-man COMT; S-COMT, soluble COMT; MB-COMT, membrane-bound COMT; CMV, cytomegalovirus; mAb, monoclonal antibody.tTo whom reprint requests should be addressed.§The sequence reported in this paper has been deposited in theGenBank data base (accession no. M58525).

1416

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

6, 2

020

Proc. Natl. Acad. Sci. USA 88 (1991) 1417

ligation into the BamHI site of expression vector pBC12/CMV (constructed by B. R. Cullen, Duke University Med-ical Center). The pBC12/CMV vector is composed of asimian virus 40 origin connected to human cytomegalovirus(CMV) enhancer-promoter sequences with a unique BamHIsite for cloning. The rat preproinsulin II (rl2) intron andsimian virus 40 polyadenylylation signal region were placeddownstream to the BamHI site.

Transfection of 293 Cells and Assay of hCOMT. Humanembryonic kidney cells (293 cell line, ATCC no. CRL 1573)were maintained in minimal essential medium/8% fetal calfserum. Cells (5 x 105) were seeded in 100-mm dishes 24 hrbefore addition of plasmid DNA (20 ,ug per dish). Transfec-tion was performed using the modified calcium phosphateprecipitation method (18). Forty-eight hours after transfec-tion, cells were collected and homogenized in 10 mM phos-phate buffer, pH 7.6/10 mM benzamidine (_107 cells per ml).COMT activity was then measured in the homogenate by aradioenzymatic method with 2.5 mM pyrocatechol as sub-strate and S-adenosyl-L-[methyl-3H]methionine as methyldonor (19). One unit of COMT activity was defined as thequantity of enzyme producing 1 nmol of guaiacol per hr at37°C. For protein quantification the Pierce assay was used.

Isolation ofMicrosomes. After homogenization ofthe trans-fected 293 cells, a separation of the soluble from the mem-brane fraction was performed by differential centrifugation.The homogenate was centrifuged at 600 x g for 20 min, andthe resulting supernatant was recentrifuged at 100,000 x g for60 min. The pellet containing mitochondria and microsomeswas washed twice and finally resuspended in 10 mM phos-phate buffer, pH 7.6. The presence of recombinant hCOMT,both in the supernatant and in the pellet, was tested byimmunoblotting (20).

Soluble and membrane fractions from Hep G2 were ob-tained essentially as above.

RESULTS AND DISCUSSIONProtein Sequencing of Porcine Liver COMT. To obtain a

partial amino acid sequence of COMT, we purified thesoluble protein from porcine liver by using an immunoaffinitychromatography column (unpublished data). The highly pu-rified protein was then subjected to tryptic digestion followedby purification of the tryptic peptides by HPLC. The se-quences of two tryptic peptides, T-22: ILQYVLQPAVAGD-PQSVLDTIG and T-33: LLTIELNPDNAAIAQQVVDFA-GLQDWVTVVVGA, were determined by gas-phase micro-sequencing (22).

Cloning of Porcine Liver COMT cDNA. A polyclonal anti-body directed against highly purified porcine liver S-COMTwas used to screen a Agtll cDNA expression library fromporcine liver. In an initial immunoscreening, seven positiveclones with inserts ranging between 1100 and 1300 base pairs(bp) were detected. The inserts of five of them crosshybrid-ized to each other. The DNA sequences of the two largestcrosshybridizing clones (1 and 7) contained an open readingframe coding for a protein of 186 amino acids with a predictedmolecular mass of 20 kDa. The predicted amino acid se-quences contained a stretch of 33 amino acids that matchesperfectly with the porcine COMT tryptic peptide T-33 ob-tained by protein sequencing. This result strongly indicatedthat the isolated cDNA codes for a partial sequence of theporcine COMT.

Isolation and Sequence ofhCOMT cDNA. Using the porcineCOMT cDNA as hybridization probe, we isolated its humancounterpart by screening a human cDNA library. Because S-and MB-COMT had been demonstrated in the Hep G2 cells,its poly(A)+ RNA was used to construct a Agtll cDNAlibrary. An initial screening of 1.2 x 106 recombinants atmoderate stringency with a probe containing a 370-bp long

fragment of the porcine liver cDNA clone (pCOMT F7, Fig.la) revealed 10 positive clones with inserts ranging between1226 and 500 bp. Two largest clones that gave the strongestsignal, AhCOMT F3 and AhCOMT F8, contained 1226-bp and1150-bp EcoRI inserts, respectively. DNA sequence analysisrevealed that both clones were identical, except for a larger5'-untranslated region in AhCOMT F3.The nucleotide and deduced amino acid sequences ofclone

AhCOMT F3 are shown in Fig. lb. The nucleotide sequenceof clone ACOMT F3 (-204 to 1000) contains the largest openreading frame starting at nucleotide residues 1-3 and termi-nating at nucleotide residues 814-816 with a stop codon(TGA). This 813-bp open reading frame codes for a polypep-tide of 271 amino acid residues with a predicted molecularmass of 30 kDa. The first initiation codon (nucleotide positiona

H0

> O

2001 1 1

H Aa. tL <

rCL

600 800 1000 1200. I1

mRNA5' |I , 'I I

AUG pCOMTF7 term

b5' -GAATTCCCGCCATTGCCGCCATCGTCGTGGGGCTTCTGGGGCAGCTAGGGCTGGCCGCGC

TGCCTGCGCCGGACCGGGGCGGGTCCAGTCCCGGGCGGGCCGTCGCGGGAGAGAAATAAC

AGCCTGCTGAGAGGTGCTTTGAAGATGCCGGAGGCCCGCCTCTGCTGTTGGCAGCTGTGM P E A P P L L L A A V1

Membrane seg.TTGCTGGGCCTGGTGCTGCTGGTGGTGCTGCTGCTGCTTCT GGCACTGGGGCTGGGGCL L G L V L L V VL L L L L H W G W G

30CTGTCGCTTATCGGCTGGAACGAGTTCATCCTGCAGCCCATCCACAACCTGCTCATGGGTL S L I G W N E F I L Q P I H N L L M G

....-T-22 . . ...... .......

GACACCAAGGAGCAGCGCATCCTGAACCATGTGCTGCAGCATGCGGAGCCCGGGAACGCAD T K E Q R I L N H V L Q H A E P. G N A

......-...I 70CAGAGCGTGCTGGAGGCCATTGACACCTACTGCGAGCAGAAGGAGTGGGCCATGAACGTG

Q S V L E A I D T Y W E Q K E W A M N V80 90

GGCGACAAGAAAGGCAAGATCGTGGACGCCGTGATTCAGGAGCACCAGCCCTCCGTGCTGG D K K G K I V D A V I Q E H Q P S V L

100 110CTGGAGCTGGGGGCCTACTGTGGCTACTCAGCTGTGCGCATGGCCCGCCTGCTGTCACCA

L E L G A Y i G Y S A V R M A R L L S P... ~~...-T-33-... ... .......

GGGGCGAGGCTCATCACCATCGAGATCAACCCCGACTGTGCCGCCATCACCCAGCGGATG

G A R L I T I E I N P D © A A I T Q R M...... 1.. .150

GTGGATTTCGCTGGCATGAAGGACAAGGTCACCCTTGTGGTTGGAGCGTCCCAGGACATCV D F A G M K D K V T L V V G A S Q D I

160 170ATCCCCCAGCTGAAGAAGAAGTATGATGTGGACACACTGGACATGGTCTTCCTCGACCACI P Q L K K K Y D V D T L D M V F L D H

180 190TGGAAGGACCGGTACCTGCCGGACACGCTTCTCTTGGAGGAATGTGGCCTGCTGCGGAAG

W K D R Y L P D T L L L E E i G L L R K200 210

GGGACAGTGCTACTGGCTGACAACGTGATCTGCCCAGGTGCGCCAGACTTCCTAGCACAC

G T V L L A D N V I W P G A P D F L A H220 230

GTGCGCGGGAGCAGCTGCTTTGAGTGCACACACTACCAATCGTTCCTGGAATACAGGGAG

V R G S S W F E i T H Y Q S F L E Y R E240 250

GTGGTGGACGGCCTGGAGAAGGCCATCTACAAGGGCCCAGGCAGCGAAGCAGGGCCCTGAV V D G L E K A I Y K G P G S E A G P AM*

260 270CTGCCCCCCCCGGCCCCCCTCTCGGGCTCTCTCACCCAGCCTGGTACTGAAGGTGCCAGACGTGCTCCTGCTGACCTTCTGCGGCTCCGGGCTGTGTCCTAAATGCAAAGCACACCTCGCCGAGCCTGCGCCCTGACATGCTAACCTCTCTGAACTGCAACACTGGATTGTTCTTTTTTAAGACTCAAAAAAAAAAAAAAAAAAAA-3'

3'

-145-85-25

36

96

156

216

276

336

396

456

516

576

636

696

756

816

876936996

FIG. 1. hCOMT. (a) Restriction map of the cDNA clone hCOMTF3 (bases 1-1226). The thick black bar indicates the hCOMT-codingsequence. The porcine fragment (pCOMT F7, 370 bp), which wasused for isolation ofthe human clone, is indicated below the solid bar.(b) Nucleotide and predicted amino acid sequences of clonedhCOMT. Nucleotides are numbered in the 5' -) 3' direction, andnumbers are shown at right of the sequence. Nucleotide residue 1 isthe adenine of the initiating methionine codon ATG, and the nucle-otides on the 5' side of residue 1 are indicated by negative numbers.The putative membrane segment is indicated by the box. The sevencysteines are circled. Tryptic peptides (T-22, T-33) that were isolatedand sequenced from porcine liver COMT are overlined; residuesdiffering from those predicted in the human sequence are indicatedby a dotted line. *, Stop codon.

Neurobiology: Bertocci et al.

Dow

nloa

ded

by g

uest

on

May

6, 2

020

1418 Neurobiology: Bertocci et al.

kb

4.974 -

0.947 -

0.832 -

0.564 ---

FIG. 2. Northern blot hybrid-ization of poly(A)+ RNA fromHep G2 cells. The hybridizationprobe was the 32P-labeled nick-translated cDNA insert from theAhCOMT F3 clone. Positions ofmolecular size standards are indi-cated at left.

1-3) is embedded in the sequence TGAAGATG, which doesnot perfectly match with the consensus sequence CCAC-CATG frequently found for the eukaryotic translation initi-ation (23). In the 5'-untranslated region there is an in-framestop codon at the position (-159 to -157). AhCOMT F3contains 204 and 207 nucleotides in the 5' and 3' untranslatedregions, respectively. The homologues of the porcine trypticpeptide sequences, T-22 and T-33, were also found in thepredicted protein sequence of hCOMT (Fig. lb).To confirm the assigned initiator codon, a cRNA was

synthesized in vitro by using the pSPI19 vector into whichthe hCOMT cDNA was subcloned. When added to a rabbitreticulocyte lysate containing [35S]methionine, the cRNAdirected the translation of a single labeled polypeptide withan estimated molecular mass of 29 kDa (data not shown),implying that the imposed initiator methionine residue can berecognized by the in vitro translation system.No potential N-linked glycosylation site (Asn, Xaa, Ser/

Thr) (24) was found in the predicted amino acid sequence ofhCOMT, which is in good agreement with the observationthat when a highly purified porcine liver COMT was incu-bated with a mixture of endoglycosidases, the molecular sizeof the COMT (tested by SDS/PAGE) was not reduced (datanot shown).Northern (RNA) blot analysis of poly(A)+ RNA from Hep

G2 cell revealed a single band of 1.4 kilobases (kb) when

Amino acids

FIG. 3. Hydropathy plot of the amino acid sequence of thehCOMT computed according to Kyte and Doolittle (25) at the spansetting of 17 amino acids. Solid bar indicates the putative membranesegment. Positive values represent hydrophobic regions, and nega-tive values represent hydrophilic regions.

either the partial porcine cDNA or AhCOMT F3 cDNA wasused as hybridization probe (Fig. 2).Hydropathy and Secondary Structure of the Predicted

Amino Acid Sequence of hCOMT. The results of the hydrop-athy index evaluation of hCOMT protein by the method ofKyte and Doolittle (25) are illustrated in Fig. 3. According tothese authors, any protein sequence corresponding to amembrane-spanning domain is characterized by a stretch ofat least 18 amino acids exhibiting an average hydropathyindex >1.6. In the hCOMT sequence, the N-terminal region(amino acids 6-26) meets these criteria, indicating that thehCOMT we cloned corresponds to the membrane-associatedenzyme. Furthermore, the nature of the N-terminal regionwas investigated by using the criteria ofVon Heijne (26). Theamino acids 1-28 of hCOMT contain a short negativelycharged N-terminal region (amino acids 1-5), a 21-residue-long highly hydrophobic core (amino acids 6-26), and apositively charged C-terminal region. Such features are notcharacteristic of known eukaryotic signal peptides (27) butseem rather to belong to a class of N-terminal anchorsegments, resulting in an N-terminally anchored protein withmost of its mass exposed on the cytosolic face of theendoplasmic reticulum membrane.Sequence Comparison. Comparison of the partial COMT

sequence deduced from the porcine cDNA (amino acids1-186) with the deduced amino acid sequence of hCOMT(amino acid 86-271) (Fig. 4) shows high homology. The

pCOMT: [Y]R A 1H 1V R [7K G1Q[IT T VIY ]Q R P S V L L E L G A Y©g)G Y S A V R M A R L L L SjA

hCOMT: W A ]M VJ D K K G K I D jAVjI[ H Q P S V L L E L G A Y©G Y S A V R M A R L L S P G:

porcine T-33: L L T I E L N P D N A A I A Q Q V V D F A G L Q D|W|V T V V V G A

pCOMT: R L L T I E L N P D N A A I A Q Q V V D F A G L Q D R V T V V V G A S Q D I I P Q L K K K Y D V D T0000 0

hCOMT: R L K V T V V G A S Q D I I P Q L K K K Y D V D T

pCOMT: L D M V F L D H W K D R Y L P D T L L L E E®G L L R K G T V L L A D N V I®P G A P D F L A H V R

hCOMT: L D M V F L D H W K D R Y L P D T L L L E E®G L L R K G T V L L A D N V i©P G A P D F L A H V R

pCOMT: GF)G RF E T H|F s1 Y S Q M|V D G L E K A|V|Y K G P G S PA QW *

hCOMT: S S F EOT H|Y QHSF L E Y|R E V|V D G L E K A|I|Y K G P G SEAGP *

49

134

99

184

149234

186271

FIG. 4. Comparison of the hCOMT primary structure with that of the partial porcine sequence. The predicted amino acids 1-186 of porcineCOMT have been aligned with amino acids 86-271 of hCOMT by using the GAP computer program (28). Amino acid sequence obtained fromthe porcine liver COMT peptide (T-33) is shown over the predicted porcine sequence. Identical residues are boxed. The cysteine residues arecircled. A * between residues represents chemically homologous residues.

Proc. Natl. Acad. Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

May

6, 2

020

Proc. Natl. Acad. Sci. USA 88 (1991) 1419

extent of sequence identity of hCOMT with porcine is 83%,but when the homology is calculated considering conserva-tive amino acid substitutions, the figure is 93%. Nucleotidesequence homology between human and porcine COMT wasfound to be 83%. Seven cysteines are contained in the humansequence (Fig. lb), of which at least four are conservedbetween the human and porcine enzyme (Fig. 4).Based on substantial evidence obtained by affinity-labeling

studies, Borchardt (29) has suggested that two nucleophilicresidues, probably sulfhydryl, exist at the active site ofCOMT and that both are necessary for enzymatic activity.

a c,, T 5203

200 -

.S(

o

OO 0-

0- c>1EC

*_100 -

b

We attempted to assign the active site in the predictedsequence of COMT. As indicated in Fig. 4, a stretch of 71amino acids (residues 79-150 of porcine and 164-235 ofhuman sequences, respectively) is identical in the predictedsequences of both species. In particular, the two cysteines,separated by 15 amino acids, in the middle of the conservedstretch might be important for the catalytic activity ofCOMT.Further studies, such as site-directed mutagenesis, areneeded to substantiate this assumption.

Transient Expression of hCOMT cDNA in 293 Cells. Toconfirm that the isolated cDNA indeed encodes hCOMT, thehCOMT F3 cDNA was subcloned into the eukaryotic expres-sion vector pBC12/CMV (Fig. Sa). 293 cells were transfectedwith the resultant construct, and COMT activity was mea-sured in the cell homogenate. 293 cells contained very littleendogenous enzyme activity, and this baseline level did notchange after transfection with a plasmid lacking the hCOMTsequence. In contrast, 293 cells transfected with plasmidpBC12/CMV-hCOMT containing the entire hCOMT se-quence showed a 20-fold increase in COMT activity (Fig. 5b).The activity increased linearly as a function of the concen-tration of the homogenate and of the incubation time (15-90min).

Inhibition of hCOMT Activity by Ro 40-7592. The effect ofthe two COMT inhibitors, Ro 40-7592 (3,4-dihydroxy-4'-methyl-5-nitrobenzophenone) and tropolone (2-hydroxy-2,4,6-cycloheptatrien-1-one) on recombinant hCOMT wasinvestigated. Ro 40-7592 is a highly potent, competitive, andselective inhibitor and is presently under clinical trial for thetherapy of Parkinson disease. This compound, used in com-bination with L-dopa and a peripheral decarboxylase inhibi-tor, is designed to prolong the duration of L-dopa action bypreventing its conversion to 3-O-methyldopa (21).The inhibitory effect of Ro 40-7592 on the recombinant

COMT activity was analyzed in the homogenate of cellstransfected with the pBC12/CMV-hCOMT (Fig. 6). Therecombinant enzyme was inhibited in a dose-dependent man-ner by Ro 40-7592 with an IC50 (-700 nM) found similar tothat measured for the native enzyme in various animal tissues(21). With tropolone, a relatively weak inhibitor of COMT, a50% inhibition of recombinant COMT was observed at aconcentration of '-4 mM, again in good agreement with theresults obtained with the native enzyme (21).Immunoblot Analysis of Expressed hCOMT in 293 Cells. To

characterize the recombinant hCOMT expressed in 293 cells,the homogenates from untransfected cells as well as fromcells transfected with plasmid pBC12/CMV-hCOMT weresubjected to Western (immunological) blotting using mono-

100-1,...................I.... .....

>

F-

2 OC,

m :CL

C-L

FIG. 5. Structure of the plasmid carrying cDNA for hCOMT andexpression of hCOMT in transfected 293 cells. (a) pBC12/CMV-hCOMT contains the entire coding region of the human COMT(hCOMT F3) ligated into the unique BamHI site of the vector(Klenow filled) as described. (b) Expression of hCOMT in trans-fected 293 cells. After transfection with the indicated plasmid, a cellhomogenate was prepared for the assay of COMT activity. Basalactivity was measured in untransfected cells (indicated as none).Results, expressed as COMT activity per protein concentration, arethe means ± SEMs of three independent experiments done induplicate. SV40, simian virus 40.

0F-0

0

u

:,E

0

80-

60-

40-

20*

.1.

I

-8 -7 -6 -5log [Ro 40-7592] (M)

-4

FIG. 6. Inhibition of recombinant hCOMT by the selective in-hibitor Ro 40-7592. The homogenates of cells transfected withplasmid pBC12/CMV-hCOMT were tested for COMT activity in thepresence of different concentrations of the COMT inhibitor Ro40-7592. Each value represents the means + SEMs of four differentmeasurements.

Neurobiology: Bertocci et al.

ZIT II _ _

H~nImI

Dow

nloa

ded

by g

uest

on

May

6, 2

020

1420 Neurobiology: Bertocci et al.

14.3 -

mAbkDa Co6O-I B7

46 --

30--

21.5 --

14.3 -

0

9;>0 I1i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

L Homogenates

FIG. 7. Immunoblot analysis ofrecombinant and native hCOMT.Homogenate (30 f.g) from untransfected cells (a), homogenate (30,ug), cytosolic (15 lsg), and membrane (15,tg) fractions from 293 cellstransfected with pBC/CMV-hCOMT vector (b), and cytosolic (270tug) and membrane (450,g) fractions from Hep G2 (c) were subjectedto SDS/PAGE (15% polyacrylamide gel) and immunoblotted withmAb Co 60-1B/7 at 10 lug/ml. Bound immunoglobulins were de-tected by incubation with iodinated sheep anti-mouse immunoglob-ulin. 14C-labeled protein standards are indicated by horizontal bars.Std, standards.

clonal antibody (mAb) Co 60-1B/7. This antibody was raisedagainst S-COMT from porcine liver and was shown tocrossreact with human and rat COMT (unpublished data). Apredominant immunoreactive band of 29-kDa apparent mo-lecular mass and a weaker band of 25.5 kDa were detected(Fig. 7a) in the homogenates from transfected cells but not inthe homogenates from untransfected cells.To verify that the cDNA clone encodes the membrane-

associated form of COMT, the homogenates from 293 cellstransfected with plasmid pBC12/CMV-hCOMT were sub-jected to differential centrifugation. The resulting membraneand cytosolic fractions were tested by immunoblotting withmAb Co 60-1B/7 (Fig. 7b). COMT immunoreactivity was

found mostly in the membrane fraction, which indicates thatthe expressed hCOMT is incorporated mainly into the endo-plasmic reticulum of 293 cells. The weaker immunoreactiveband of 25.5 kDa was found in the cytosolic fraction. Fig. 7cshows an immunoblot of the cytosolic and membrane frac-tions from Hep G2 cell for comparison. This cell containsS-COMT and MB-COMT forms of the enzyme, which bothreact with mAb Co 60-1B/7. The molecular sizes of 29 kDaand 25.5 kDa observed for the recombinant COMT expressedin 293 cell are in good agreement with MB-COMT andS-COMT from Hep G2 cell.These results suggest that the 25.5-kDa protein may arise

from the 29-kDa protein by posttranslational processing. Al-though a computer search for potential proteolytic cleavagesite in the region of interest (amino acids 27-46) predictedcleavage sites for trypsin, chymotrypsin, and proline endopep-tidase, we have no clue for the exact nature ofthe process. For293 cells we speculate that, because of lack of an efficientprocessing enzyme, most COMT remains membrane-bound.

CONCLUSIONOur results indicate that the cDNA clone that has beenisolated encodes hCOMT. This conclusion -is based on thefollowing lines of evidence: (i) the predicted amino acidsequence of the human clone shows a high homology to

c porcine COMT (93%). (ii) Two tryptic peptide sequences (22and 33 amino acids) of the purified porcine liver COMT are

Hep G2 found in the predicted amino acid sequence of hCOMT. (iii)-rev - ~The expression of the cDNA in transfected cells produced a

protein that possesses the full biological activity ofCOMT inmAb the enzymatic assay and is recognized by anti-COMT mAb in

kDa Co60-1B 7 the immunoblot. Moreover, immunoblot analysis showedthat the recombinant hCOMT was actually associated with

46 - - the membrane fraction.The successful cloning of hCOMT will enable us to pro-

duce the enzyme in quantities needed for further structural30 and functional analysis. Finally, we hope that the cDNA

clone will provide valuable tools for the investigation not only21 =, -- of the cellular expression of COMT in various regions of the4

brain but also of the genetic linkage of COMT alleles with14.3 certain diseases of the central nervous system.

We are grateful to Prof. G. Garotta for his helpful advice through-°n U out this work. We thank Prof. W. Haefely, Drs. M. Steinmetz, P.

-0n Schoch, and J. G. Richards for their critical evaluation of themanuscript, Dr. T. Giller for subcloning of hCOMT cDNA into themammalian vector, M. Buhler, V. Gyoerffy, and K. McKune for

IL IS 7- rj_ _ Itechnical assistance, and M. Wdonwiciu tor secretarial work.1. Axelrod, J. (1966) Pharmacol. Rev. 18, 95-113.2. Kawaia, S. (1983) in Methods in Biogenic Amine Research, eds.

Parvez, S., Nagatsu, T., Nagatsu, I. & Parvez, H. (Elsevier,Amsterdam), pp. 417-439.

3. Axelrod, J. & Cohn, C. K. (1971) J. Pharmacol. Exp. Ther. 176,650-654.

4. Inscoe, J. K., Daly, J. & Axelrod, J. (1965) Biochem. Phar-macol. 14, 1257-1263.

5. Roth, J. A. (1980) Biochem. Pharmacol. 29, 3119-3122.6. Assicot, M. & Bohuon, C. (1971) Biochimie 53, 871-874.7. Jeffery, D. R. & Roth, J. A. (1984) J. Neurochem. 42, 826-832.8. Rivett, A. J., Eddy, B. J. & Roth, J. A. (1982) J. Neurochem.

39, 1009-1016.9. Weinshilboum, R. M. (1989) in Handbook of Experimental

Pharmacology, Catecholamines II, eds. Trendelenburg, V. &Weiner, N. (Springer, Berlin), Vol. 90/1, pp. 391-425.

10. Karege, F., Bovier, P., Gaillard, J.-M. & Tissot, R. (1987) ActaPsychiatr. Scand. 76, 303-308.

11. Gustavson, K. H., Wetterberg, L., Backstrom, M. & Ross,S. B. (1973) Clin. Genet. 4, 279-280.

12. Guldberg, H. C. & Marsden, C. A. (1975) Pharmacol. Rev. 27,135-206.

13. Young, R. A. & Davis, R. W. (1983) Proc. Natl. Acad. Sci.USA 80, 1194-1198.

14. Huynh, T. V., Young, R. A. & Davis, R. W. (1984) in PracticalApproaches in Biochemistry, ed. Glover, D. (IRL, Oxford), pp.49-77.

15. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter,W. J. (1979) Biochemistry 18, 5294-5299.

16. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) MolecularCloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY), pp. 7.26-7.29.

17. McMaster, G. K. & Carmichael, G. G. (1977) Proc. Natl.Acad. Sci. USA 74, 4835-4838.

18. Chen, C. & Okayama, H. (1987) Mol. Cell. Biol. 7, 2745-2752.19. Zurcher, G. & Da Prada, M. (1982) J. Neurochem. 38, 191-195.20. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl.

Acad. Sci. USA 76, 4350-4354.21. Zurcher, G., Keller, H. H., Kettler, R., Borgulya, J., Bonetti,

E. P., Eigenmann, R. & Da Prada, M. (1990) Adv. Neurol. 53,497-503.

22. Hewick, R. M., Hunkapiller, M. W., Hood, L. E. & Dreyer,W. J. (1981) J. Biol. Chem. 256, 7990-7997.

23. Kozak, M. (1984) Nucleic Acids Res. 12, 857-872.24. Marshall, R. D. (1972) Annu. Rev. Biochem. 41, 673-702.25. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132.26. Von Heijne, G. (1986) J. Mol. Biol. 189, 239-242.27. Von Heijne, G. (1986) Nucleic Acids Res. 14, 4683-4690.28. Needleman, S. B. & Wunsch, C. D. (1970) J. Mol. Biol. 48,

443-453.29. Borchardt, R. T. (1977) in Structure and Function of Mono-

amine Enzymes, eds. Usdin, E., Weiner, N. & Youdim,M. B. H. (Dekker, New York), pp. 707-726.

a bUntrans- pBCl 2/CMV-hCOMTfected 293 Transfected 293

m - m< C7

Q EB(D (00P 0 0

kDa co ) z a

46

-0

E0z

30 m

21.5 -

Proc. Natl. Acad. Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

May

6, 2

020