The Multifunctional Peptidylglycine cY=Amidating Monooxygenase Gene: Exon/lntron Organization of Catalytic, Processing, and Routing Domains

L’Houcine Ouafik, Doris A. Stoffers, Tracey A. Campbell, Richard C. Johnson, Brian T. Bloomquist, Richard E. Mains, Betty A. Eipper

and

Department of Neuroscience The Johns Hopkins University School of Medicine Baltimore Maryland 21205 Laboratoire de Neuroendocrinologie Experimentale (L’H.0) INSERM U297 13326 Marseilles Cedex 15, France

codes the trans-membrane domain of PAM; alternative splicing at this site produces integral membrane or soluble PAM proteins. The COOH-terminal domain of PAM is comprised of a short exon subject to alternative splicing and a long exon encoding the final 68 amino acids present in all bifunctional PAM proteins along with the entire 3’-untranslated region. Analysis of hybrid cell panels indicates that the human PAM gene is situated on the long arm of chromosome 5. (Molecular Endocrinology 6: 15711584,1992)

Peptidylglycine Lu-amidating monooxygenase (PAM; EC 1.14.17.3) is a multifunctional protein containing two enzymes that act sequentially to catalyze the (Yamidation of neuroendocrine peptides. Peptidylglytine a-hydroxylating monooxygenase (PHM) catalyzes the first step of the reaction and is dependent on copper, ascorbate, and molecular oxygen. Peptidyl-cr-hydroxyglycine a-amidating lyase (PAL) catalyzes the second step of the reaction. Previous studies demonstrated that alternative splicing results in the production of bifunctional PAM proteins that are integral membrane or soluble proteins as well as soluble monofunctional PHM proteins. Rat PAM is encoded by a complex single copy gene that consists of 27 exons and encompasses more than 160 kilobases (kb) of genomic DNA. The 12 exons comprising PHM are distributed over at least 76 kb genomic DNA and range in size from 49-165 base pairs; four of the introns within the PHM domain are over 10 kb in length. Alternative splicing in the PHM region can result in a truncated, inactive PHM protein (rPAM-5), or a soluble, monofunctional PHM protein (rPAM-4) instead of a bifunctional protein. The eight exons comprising PAL are distributed over at least 19 kb genomic DNA. The exons encoding PAL range in size from 54-209 base pairs and have not been found to undergo alternative splicing. The PHM and PAL domains are separated by a single alternatively spliced exon surrounded by lengthy introns; inclusion of this exon results in the production of a form of PAM (rPAM-1) in which endoproteolytic cleavage at a paired basic site can separate the two catalytic domains. The exon following the PAL domain en0888-8809/92/1571-1584503.00/0 Molecular Endocmology CopyrIght 0 1992 by The Endocrme

INTRODUCTION

Many biologically active peptides are a-amidated at their COOH-terminus, a structural feature that is often essential for their biological activity. These peptides are produced from larger inactive precursors which are cleaved to form peptides having a glycine residue at their COOH-terminus (l-4). Conversion of a peptidylglycine substrate into an a-amidated product is a twostep reaction involving the copper, ascorbate, and molecular oxygen-dependent production of a peptidyl-ahydroxyglycine intermediate; at physiological pH a second enzymatic activity catalyzes the subsequent formation of the a-amidated product (4-10). Both enzymes are derived from the bifunctional peptidylglycine a-amidating monooxygenase precursor (PAM; EC 1.14.17.3) (Fig. 1; Refs. 8, 11, 12). PAM has broad substrate specificity and is found in many tissues. Amidation is often the rate-limiting step in peptide biosynthesis, and PAM expression is regulated in a tissuespecific fashion in response to endocrine manipulations (4).

Society

1571 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1572

ENDO.

1992

cDNA clones encoding 1 IO-kilodalton (kDa) bifunctional integral membrane protein forms of PAM have been isolated from bovine and rat pituitary, frog skin, rat heart atrium, and human and rat thyroid carcinoma cDNA libraries (13-l 9). These PAM precursor proteins are predicted to contain an NH,-terminal signal sequence followed by a short propeptide. The enzyme catalyzing the first step, peptidylglycine a-hydroxylating monooxygenase (PHM), constitutes the NH*-terminal third of the PAM precursor (Fig. 1). The enzyme catalyzing the second step, peptidyl-cu-hydroxyglycine LYamidating lyase (PAL), constitutes the mid-region of the PAM precursor. The lyase domain is followed by a hydrophobic putative trans-membrane domain and a hydrophilic COOH-terminal putative cytosolic domain. Endoproteolytic digestion of this bifunctional integral membrane protein releases soluble PHM and PAL catalytic domains that are very resistant to further endoproteolytic degradation (22). Several types of rat, bovine, and human PAM cDNA have been characterized; all appear to arise via alternative splicing (13-15, 17, 18, 20) (Fig. 1). Deletion of a 315nucleotide (nt) segment (referred to as exon A) between the PHM and PAL domains distinguishes rat PAM-l (rPAM-1) from -2 and human PAM-A (hPAM-A) from -B (this segment is 321 nt in hPAM). Deletion of a 258-nt segment (exon B) containing the putative transmembrane domain generates a soluble, intragranular bifunctional PAM protein. Frog skin amidating enzymeI and rPAM-4 encode soluble proteins that include only the monooxygenase domain; while rPAM-4 appears to arise via alternative splicing, amidating enzyme-l represents the product of a separate gene present in the tetraploid frog, Xenopus laevis (21). The various PAM proteins appear to differ in functionally significant ways. Expression of the individual forms of PAM cDNA in the neuroendocrine AtT-20 cell line indicates that the different PAM proteins exhibit distinct patterns of endoproteolytic processing and subcellular routing (23). In addition, the catalytic properties of PHM are altered upon limited endoproteolytic digestion of bifunctional PAM proteins (22). The nucleotide sequences of the known rat PAM cDNAs, as well as Southern blot analysis data, suggested that PAM was a single copy gene in the rat. As a first step toward understanding the mechanisms regulating PAM gene expression, we undertook elucidation of the structure of the gene encoding rat PAM. The genes encoding several multifunctional enzymes have been characterized; in general, they appear to have been constructed by assembling genes encoding the individual catalytic units (24-28). In this report we describe the isolation and characterization of genomic clones containing the entire coding sequence of PAM, analysis of the exon/intron structure of the gene and localization of the gene for human PAM to the long arm of chromosome 5.

Comparison

I

Copy PAM Gene

Preliminary Southern blot analysis using the bovine PAM (bPAM) cDNA probes shown in Fig. 1 indicated

of PAM mRNAs

PHM

i

Identified

PAL

TM .

in Various

Species

I i

Fig. 1. Forms of PAM mRNA Identified in Various Species The rPAM-1 protein is drawn to scale with the signal sequence and putative trans-membrane (TM) domain indicated by filled boxes; the PHM and PAL catalytic domains are indicated. The forms of PAM mRNA identified in SpragueDawley rat atrium and pituitary are identified to the leff (14, 17, 20). The forms of PAM mRNA identified in bovine intermediate pituitary (13) human thyroid carcinoma (18) rat medullary thyroid carcinoma (19) Wistar rat pituitary (15) and frog skin (16, 21) are identified to the right. The bovine PAM cDNA probes used in the initial library screen and the rat PAM cDNA probe used for Southern blot analysis are identified at the

bottom.

that the gene encoding rPAM was complex. Further Southern blot analysis was carried out using different probes to determine whether the rat genome contains a single copy of the PAM gene (Fig. 2). Since the 3’untranslated region of many mRNAs occurs in a single large exon, a cDNA probe from the 3’-untranslated region of rPAM was used. A single restriction fragment was visualized by this probe for each enzyme, indicating that the rat genome contains a single copy of the PAM gene (Fig. 2) (15). Subsequent analyses of Southern blots with several oligonucleotide probes contained within single exons confirmed this conclusion. Primer extension analysis was carried out to determine the length of the 5’-untranslated region (Fig. 3). A single major product of 167 nt was identified in atrium, hypothalamus, pituitary, and submaxillary gland; additional minor products were observed in anterior pituitary and submaxillary gland. The size of the major product suggests that rPAM transcripts in each of these tissues contain an additional 104 nt upstream of the nucleotides identified in rPAM-1 (17). The sequence reported for a PAM cDNA obtained from rat medullary thyroid carcinoma (19) includes 26 nt of sequence upstream of the sequence identified in rPAM-1. Exon/lntron

RESULTS The Single

Vo16No.10

Organization

of the Rat PAM Gene

When this project was initiated, the only PAM cDNA probes available were those for bPAM; therefore, our rat genomic library was screened with three bPAM

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional PAM Gene

Southern

Blot Analysis

1573

of Rat Genomic

DNA

Primer

Extension

Analysis of PAM Transcripts

bp 23,130 9,416 6,557 4,361 -

2,322 2,077 -

458-0 434,-

267

-

234

-

213

-

192 184

+

564 -

Fig. 2. Southern Blot Analysis Rat liver DNA (10 fig) digested with the enzymes indicated was fractionated on an 0.8% agarose gel, transferred to Hybond N, immobilized, and visualized with a cDNA probe contained within the 3’-untranslated region of rat PAM. The &cl fragment (1.3 kb) visualized on the Southern blot is the 1.3 kb Sac1 fragment containing the most 3’-exon of the PAM gene. cDNA probes spanning the sequence of bPAM (Fig. 1). A set of eight partially overlapping phage containing PAM exons was identified in this way. A different amplified rat genomic library was subsequently screened in order to obtain phage containing the missing regions of the PAM gene. A set of five additional phage was identified by screening with oligonucleotide probes, screening with polymerase chain reaction (PCR) fragments spanning missing exons, and screening with riboprobes specific to the end of an insert abutting a gap. Upon analysis of these 13 phage, it was clear that phage containing at least three exons (exons 1, 7, and 17) were still missing. The existence and approximate size of exon 1 can be inferred from the length of the primer extension product and a comparison of the sequence contained in exon 2 and the most 5’-region of the published sequences for rPAM (17,19). Repeated screening of a rat genomic library with an oligonucleo-

1%4

-

104

-

Fig. 3. Primer Extension Analysis of the 5’-Untranslated Region An antisense oligonucleotide [rPAM(63-46)] was used to prime the reverse transcription of poly(A)+ RNA isolated from atrium (1.5 pg) or total RNA isolated from atrium (21 pg), hypothalamus (23 fig), anterior pituitary (30 fig), and submaxillary gland (30 pg). A major product of 167 nt (arrow) was observed in each tissue. The streak at the top of the lane marked hypothalamus extends from the sample well and was not seen when analyzing poly(A)+ RNA.

tide probe corresponding to the most 5’-region reported by Bertelson et al. (19) failed to identify phage containing the V-end of the PAM gene; use of the same sequence for PCR amplification across intron A was similarly unsuccessful. Although assigned to a single exon in Table 1, the 5’-end of the PAM gene may consist of more than one exon. Exon 7 was identified by PCR amplification across the intron separating exons 6 and 7. A similar approach failed to amplify a fragment between exons 1 and 2, exons 7 and 8, or exon 17 and either exon 16 or 18. Therefore a PCRbased methodology [rapid amplification of genomic

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1574

ENDO.

Table

1. Exonnntron Exon (size)

1992

Vo16No.10

Junctions

in the rPAM

3’-End of donor exon

AGG

Gene lntron (approximate

size)

ydxag

-5 GGACCAGnd . . . . A(nd) . . . . . . ..atttttgcagGTTCtGA 401 TTAAGAGgtgcgtgctc....B(0.8kb).....tcttcttaagGTTTAAA 522 TAAAGAGgtgcgtggaa....C(>Zlkb).....tttttaacagTCTGACA 580 TTCGTGAgtaagtactc....D(>4.8kb)....cccctgacagTTGACTT 668 GTTACTGgtaaggactg.... E(ll.lkb)....tttctttcag 754 CCGAAAGgtttgtttga....F(l.lkb).....gacaacttagGTGTTGG 6 (86) 838 TTTCGAGgtgagtttag.. ..G(>13kb).....tttttctcagATAATCA 7 (84) 887 GTGTGCCgtaagtgcct....H(1.5 kb).....taatttttagCCAGCCT 8 (49) 955 GAGAAAGgtgagtagta....I(1.3kb).....ttccccacagTAGTGAA 9 (68) 10 (81) 1036 CATTTAGgtaagaactg....J(0.40kb)....cttcatcaagGTAAGGT 11 (77) 1113 GCCACAGgtgtgtagcg.. ..K(1.3 kb).....tttggggcagGCTTTCT 1217 12 (104) ACATCGGgtaagattct.. ..L(O.52 kb)....gttgacctagATTCAAA PAM5 [13] (1125) alt 3'-UTR....M(16kb)......ttccctacagCGGCACT 14 (185) 1402 CACAAAGgtgagctgga....N(1.8kb).....tcttttacagAAGCAGA 15 (72) 1474 GAGCAGGgtaagtgtcc....0(11.6kb)....cccctcgcagGTGATTT 1789 TCAGGAGgtgcatccag....P(>23 kb).....tctcttttagATTTCCA [I61 (315) PAM4 alt3'-UTR(retainedpartof intronP) [PI (1196) 17 (130) 1919 ATGGAAAgtaagtaatt.. ..4(>4.1kb)....ttttgtttagCTCTTTT 18 (117) 2036 AGAACCTgtgagtaaaa....R(2.4 kb).....tgtcttgcagGTTTTAT 2109 19 (73) CCACCAGgtgagcttcc.... S(l.l kb).....tttttttaagGTGTTCA 20 (211) 2320 GGAGAAGgtaccatttc....T(0.23kb)....tgtcctgtagAGTCCTC 21 (201) 2521 ATACCAGgtatttaact....U(2.8 kb).....ctcgtcgcagGTTTCCT 22 (116) 2637 ACGCAAGgtacctacat.. ..V(3,4kb).....gctttgacagCACTTCG 2737 23 (100) ACTGAAAgtaggttggg....W(3.9kb).....gccgttacagAAATGGA 2791 24 (54) ATCAAAGgttggtacgg....X(6.8 kb).....ttccctgaagAAGCCGA 2995 TTTGGAGgcaagtaaaa....Y(2.7kb).....tttaaagcagATCATGA [251(204) 3049 TTCCGAGgtatgcctgt....Z(0.80kb)....ttccccccagGAAAGGG WI (54) 3892 27 (843) Alternate 3’-untranslated regions: exon27 PAM-1,2,3 TGTAATAAAGTGTTTTCAGAGCATTAgcaagtcagtgtattttgtgattttttt exon13 PAM-5 GTGTTTTGTACAATAGCAGGTGTAGTCTCCacaataccataagtacccagtgaa retainedpartof intronP PAM-4 CCTATTAAAGAGCAAGTGAACTGGTGTCTGGTGATGGAgtgacagatgtgcagt 1 2 3 4 5

(-100) (405) (121) (58) (88)

G

GTTTTGT

5’-End 01 acceptor exon

Exon no

-4 402 523 581 669 755 839 888 956 1037 1114 PAM5 1218 1403 1475 1790

2 3 4 5 6 7 8 9 10 11

1920 2037 2110 2321 2522 2638 2738 2792 2996 3050

,::, 14 $1 17 18 19 20 21 22 23 [Z]

P61 27

Nucleotide sequences for the exon/intron junctions are compiled; exon numbers are indicated at both sides of the table and exons wrap-around. Bases are numbered as in Ref. 17 and modified in Ref. 20; additional sequence at the 5’-end is from Ref. 19. lntrons are identified alphabetically, and their approximate sizes are indicated; where there are gaps in the gene, intron sizes are indicated as more than the amount of intron present in the genomic clones isolated. The existence of exon 1 is inferred from the length of the primer extension product and the fact that nucleotide sequences at the 5’-end of PAM cDNAs isolated from atrial (17) and medullary thyroid carcinoma (19) cDNA libraries are not contained in exon 2. Consensus splice donor and acceptor sequences are from Ref. 41. Upper and lower case letters in the 3’-untranslated regions (UTR) indicate sequence before and after the poly(A) tail, respectively; poly(A) addition signals are shown in bold. For the 3’-untranslated region of rPAM-5, lower case letters indicate sequence not contained within the cloned cDNA (14). The asterisk indicates the first nucleotide (+l) of the published sequence (17). Numbers in brackets indicate alternatively spliced exons.

ends or (RAGE)] was developed in order to identify the genomic regions surrounding these exons. The physical map of the rPAM gene is shown in Fig. 4, and sequences at exon/intron junctions are presented in Table 1. The nucleotide sequences identified in all of the rPAM cDNAs (Fig. 1) are contained in these 27 exons. The 27 exons are distributed across at least 160 kilobases (kb) of genomic DNA; gaps remain at six positions within introns, so the PAM gene may be substantially larger than this estimate. The three alternate 3’-untranslated regions identified in various PAM cDNAs are each contained within single exons (exon 13, the retained part of intron P immediately following

exon 16 and exon 27) that are the largest in the PAM gene [1125, 1522, and 837 base pairs (bp), respectively]. The signal sequence and propeptide are included in exon 2. The monooxygenase domain is composed of 12 exons (exons 3-15, skipping exon 13) that are exclusively protein-coding and range in size from 49185 bp, with an average size of 90 bp. The exons encoding PHM are distributed over at least 76 kb genomic DNA; since four of the remaining gaps in the PAM gene lie within introns in the PHM domain, this region may be substantially larger. The introns within the PHM domain range in size from 400 bp to over 20

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional

PAM

Gene

1575

Fig. 4. Structure of the Rat PAM Gene A scale marker is shown at the top; interruptions in the line indicate positions where the genomic clones isolated do not overlap. A, The physical map of the PAM gene is depicted. Regions corresponding to PHM, exon A, PAL, the trans-membrane domain (TMD) and the COOH-terminal domain (CD) are indicated. Exons are represented by numbered vertical bars; more than a single exon may precede exon 2 at the 5’-end of the PAM gene. lntrons are drawn to scale; gaps in the genomic sequence are indicated by interruptions in the line. Overlaps between A2 and Al and between G13 and G13a were established using riboprobes, and the exact length of each overlap was not determined. B, The phage used to characterize the PAM gene are diagramed. The Sad and Smal restriction sites used in mapping the genomic phage are indicated by upward and downard tic marks respectively. C, The exon-containing subcloned restriction fragments are indicated below each phage.

kb. While exons 8-l 3 are clustered within a 6-kb region, large introns separate many of the remaining individual exons or pairs of exons. The 315bp region separating the monooxygenase and lyase domains in rPAM-1 (referred to previously as optional exon A) is encoded by exon 16. The novel sequence that follows exon A in rPAM-4 represents a retained region of intron P (Fig. 1). Exon 16 is separated from the preceding final exon of the monooxygenase domain by an 11.6-kb intron and from the subsequent first exon of the lyase domain by intron P, which is over 23 kb in length. The lyase domain is composed of eight exons (exons 17-24) ranging in size from 54-209 bp, with an average size of 125 bp. The exons encoding PAL are distributed over at least 19 kb genomic DNA, with the introns ranging in size from 230 bp to over 4.1 kb. With the genomic clones identified, a single gap remains within one intron in the lyase domain, which is more compact than the monooxygenase domain. Exon 25 encodes the Vans-membrane domain (formerly referred to as exon B,), and exon 26 encodes the 54-bp domain formerly referred to as exon B,, (14). The remainder of the COOH-terminal domain and the entire 3’-untranslated region are encoded by exon 27. The final three exons of the PAM gene span only 4 kb genomic DNA. Comparison of the Genes Encoding Dopamine ,B-Monooxygenase

PAM and

PHM and dopamine p-monooxygenase, a rate-limiting enzyme in catecholamine biosynthesis, both use copper, ascorbate, and molecular oxygen (4, 29). Analysis of their amino acid sequences revealed a structural similarity extending over 270 residues of PHM (30). Comparing hPHM (18) and human dopamine ,Gmonooxygenase, 29% of the amino acid residues were identical over this 30-kDa region. Further evidence for an evolutionary relationship of PHM and dopamine p-

monooxygenase was sought by comparing the location of exon/intron junctions in the two genes (Fig. 5). Since the PAM gene has been characterized only in rat, the locations of exon/intron junctions in the human PAM gene were inferred by homology (18). The homologous region of PHM is encoded by 11 exons, whereas the homologous region of dopamine p-monooxygenase is encoded by eight exons (31). In four cases, two PHM exons appear to form a single dopamine p-monooxygenase exon; in one case, a single PHM exon appears to form two dopamine p-monooxygenase exons. lntron lengths are much shorter in dopamine p-monooxygenase than in PHM. Of the five exon/intron junctions that can be compared between PHM and dopamine p-monooxygenase, four were displaced by 2 bp or less. This conservation of exon/intron junctions lends further support to the existence of an evolutionary relationship between PHM and dopamine /I-monooxygenase, secretory granule-associated enzymes required for the production of neuropeptides and classical neurotransmitters. Correlation Structural

of Exon/lntron Organization Features of the PAM Protein

with

The protein coding exons of the PAM gene are drawn to scale in Fig. 6. The shape at the beginning and end of each exon indicates whether the exon/intron junction is type 0, type 1, or type 2; matching shapes can thus be spliced together without loss of reading frame. Exon 2, which includes much of the 5’-untranslated region, encodes the signal sequence and propeptide, terminating before the final nucleotide encoding the Lys-Arg sequence marking the end of the propeptide; the peptide encoded by exon 2 is not in the mature PHM protein. A maximum of 12 exons form the catalytic core of the PHM domain. Histidine-rich sequences thought to be involved in the interaction of the monooxygenase with copper are found in exons 5, 10, and 14. The 24

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1576

Vo16No.10

END0.1992

COMPARJSON

OF FliM AND DOPAMINE

BETA-MONOOXYGENASE

EXON/INTRON

STRUCTURE

PHM doman m

region of homology to DEM TYFC

\ \

I I ’ TYWC 1

HHM

HTHH

I I IHHM

3

f / ’

4

II II I

I

I I I

I 5

I I

HGHH

I I I I IHTHH 1

6

I I I 8

7

\

I I I

\ \ 9

I I I 10

Dopamlne beta-monooxygenase 0

regm of homology to PHI.4

Fig. 5. Comparison

of Exon/lntron Organization of PHM and Dopamine P-Monooxygenase The sequences of human PHM(50-320) and human dopamine P-monooxygenase(lSl-484) were aligned as suggested by Southan and Kruse (30) for the corresponding bovine proteins. The exon boundaries for human dopamine P-monooxygenase were taken from Kobayashi et al. (31); the exon boundaries for human PAM were predicted by analogy to rat PAM. Exons are drawn as rectangles and are numbered; introns are depicted by lines with the approximate length indicated above the line. The percentage

of amino acid identity is indicated below each PHM exon. Amino acid sequences thought to be important in catalysis are indicated using the single letter code.

PAM PROTEIN: EXON BOUNDARIES

A;; I

I monooxygenase

17

18

1s

20

I lyase

21

I domain

,‘+-

1 domain

22

(PHM)

23

TMD .ExO” BP

24

25

I (PAL)

‘e ‘Em” sb

,

processing information

2,

3’4TR

c!3 :=I

4 information

Location TYP 0

iYk

of exonhtron TYpe

1

53

junctions: TYP 2 Sk

Fig. 6. The PAM Protein-Correlation of Exons with Functional Domains The protein-coding exons of the PAM gene are drawn to scale; the number of the first amino acid residue in each exon indicated above the exon. Exons encoding 5’- or 3’-untranslated regions of PAM are not drawn to scale. Structural features

is

present in each exon are indicated. The shape of the exon indicates whether it is a type 0, type 1, or type 2 junction.

amino acids encoded by exon 15 comprise one of the most species-specific regions of PAM; for example, the paired basic site in bovine PAM corresponds to the sequence Lys383-Gln384 in rat PAM. Based on the sizes of the stable, catalytically active PHM proteins produced by digesting atrial membranes with exogenous endoproteases, this region appears to be especially protease-sensitive (22).

A maximum of eight exons form the catalytic core of the lyase domain (Fig. 6). The first exon of the lyase domain (exon 17) encodes the sequence -Arg-Gly-Asp(RGD). For a number of extracellular matrix proteins, their RGD sequences are important in their interactions with cell surface receptors of the integrin family (32, 33). Interestingly, an RGD sequence is found in three other enzymes involved in the posttranslational proc-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional PAM Gene

1577

essing of peptide hormones [furin and prohormone convertases 1 and 2 (PC1 and PC2)] (34-37). Exon 22 includes a potential N-glycosylation site and terminates immediately after the codons for a paired basic potential endoproteolytic cleavage site (Arg77g-Lys780). The remaining paired basic residues preceding the trans-membrane domain are included in exon 24. The remaining exons encode protein domains that are not essential for either catalytic activity of PAM. Exon 16 encodes the single paired basic amino acid site situated between the PHM and PAL catalytic domains of rat PAM; in the bovine neurointermediate pituitary, endoproteolytic cleavage at this site separates PHM-A from PAL (38). Expression of rPAM-1 and -2 (Fig.1) in AtT-20 cells indicates that exon 16 must be present for the efficient separation of PHM from PAL (23). The species-specific segment of PAM encoded by exon 15 is not cleaved by the endogenous endoproteases to which PAM is exposed in AtT-20 cells. The final three exons of the PAM gene may constitute a noncatalytic domain responsible for the routing of PAM proteins. Exons 25 and 26 correspond to the regions identified as optional exons B, and Bbr respectively, in our earlier analyses of forms of PAM mRNA (14). Exon 2.5 encodes the putative traans-membrane domain, terminating several amino acids after the ArgTrp-Lys-Lysag4 putative stop-transfer signal. Exon 26 is very species specific and includes an Arg-Lys sequence that would become available to intragranular endoproteases when expressed in the intragranular milieu. The remainder of the COOH-terminal domain is encoded by exon 27. Exon 27 includes two Arg-Lys sequences, several potential phosphorylation sites, a putative PEST sequence [rPAM(954-970) PEST score 20.8 using the PESTFIND program of Rogers et al. (39)], and the entire 3’-untranslated region. Studies on AtT-20 cells expressing PAM mRNAs truncated after the trans-membrane domain indicate that the presence of this COOHterminal domain is essential for retention of integral membrane forms of PAM in the perinuclear region (40). Alternative Splicing PAM mRNA

Generates

Different

Forms of

Each of the forms of Sprague Dawley rat PAM mRNA identified to date (Fig. 1) arises via alternative splicing (Fig. 7). Despite the complexity of the PAM gene, alternative splicing has been found to occur at only a few sites. Although primer extension studies (Fig. 3) identified a single major product in several tissues, further investigation will be required to determine whether alternative splicing occurs within the 5’-untranslated region of the PAM gene. Within the PHM domain, the only product of alternative splicing identified is rPAM-5, an apparently nonfunctional derivative of PHM (14). In rPAM-5, exon 13 encodes five amino acids before reaching a stop codon and includes an AT-rich region with four potential poly(A) addition signals; the only rPAM-5 cDNA isolated did not include a poly(A) tail. Exon 16 encodes the 315 bp optional exon A and is

followed immediately by the unique sequence of rPAM4. Generation of monofunctional rPAM-4 mRNA thus appears to involve retention of the 5’-end of intron P and use of an alternative poly(A) addition signal. Generation of rPAM-1 mRNA requires use of a cryptic 5’splice donor site included in the protein-coding sequence of rPAM-4. PAM mRNAs lacking exon A (rPAM2, -3, -3a, and -3b) are produced when exon 15 is spliced to exon 17, skipping exon 16 entirely. rPAM-4 mRNA is found only in tissues where rPAM-1 is prevalent (atrium and central nervous system) and is not a major form of PAM mRNA in any of the tissues examined (20). Exons 25 and 26 form a group of cassette exons; since the exon boundaries on both sides of these two exons are type 1, they can be spliced together in any combination and maintain the reading frame. The functional consequences of deleting exon 25 are obviously profound, with an integral membrane protein (e.g. rPAM-1, -2, and -3b) converted into a soluble, intragranular protein (rPAM-3a or -3). Splicing in this region is also tissue specific (20). In general the sequences flanking the exon/intron junctions of the rPAM gene (Table 1) are very similar to the exon/intron junction sequences tabulated for mammalian genes by Shapiro and Senepathy (41). Examination of the subset of exon/intron junctions that are subject to alternative splicing reveals some differences. lntron Y, which follows exon 25 (BJ (Fig. 7) contains a highly unusual AGgcaagt sequence; of the 1893 5’splice sites examined by Shapiro and Senepathy (41) only five lacked a gt sequence, and all contained instead a gc sequence. Included in this small group is human superoxide dismutase, another copper-dependent enzyme. Although less than 2% of the 3’-splice sites examined by Shapiro and Senepathy (41) contained an ag within IO bases upstream of the true 3’-splice site, intron Y terminates agcagAT. When compared to rPAM1 or -2, the type B rat PAM cDNA (Fig. 1) isolated from medullary thyroid carcinoma contains an extra Ala residue at the junction of exons 25 and 26, suggesting that under some circumstances the upstream ag serves as the 3’-end of intron Y. A similar situation may occur at the 3’-end of intron X, between exons 24 and 25 (B,). The type B rat PAM cDNA isolated from medullary thyroid carcinoma lacks a Glu at the junction of exons 24 and 25; inclusion of this triplet (gaa) in intron X would place an ag sequence in exactly the same position observed in intron Y. Chromosomal

Localization

of hPAM Gene

Two human somatic cell hybrid panels were used to determine on which human chromosome the PAM gene is located (Fig. 8). Species-specific pairs of sense and antisense oligonucleotides situated within the most 3’exon of the PAM gene were designed and used to determine which somatic cell hybrids contained the hPAM gene. The primers specific for rPAM efficiently amplify a fragment of the same size (294 bp) from mouse genomic DNA (not shown) but fail to amplify a

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1578

ENDO.

Vo16No.10

1992

ALTERNATIVE

Start

SPLICING

transcription A

MetB

~. PAM-l C

D

E

OF THE RAT PAM GENE

,2,3,3a,3b,4

PAM-l

PAM-l

FGHIJK

\\\ 1

2 (Exon

5’.UTR

A)

(rPAM-4)

PAM-2,3,3a,3b

PAY -5

Q

R

T

s

17

18

19

20



u A

A 21

AAAAAA

A 22

23

PAM-3

Fig. 7. Alternative Splicing Patterns The exons that must be spliced together to yield the forms of PAM mRNA identified to date are indicated. lntrons are labeled alphabetical order. Exon 1 has not yet been isolated, and its existence is inferred from the primer extension data; the 5’-untranslated region preceding exon 2 may be encoded by more than one exon.

in

A

-

406 294

bp bp

Fig. 8. Chromosomal Localization of the Human PAM Gene Pairs of sense and antisense oligonucleotide primers specific for human or rat PAM were used to amplify DNA prepared from each cell line or control sample as indicated. The primers specific for human PAM yield a 406-bp fragment, while the primers specific for rat PAM yield a 294-bp fragment. The rat PAM primers efficiently amplify a fragment of the expected size from mouse genomic DNA but not from CHO cell or human DNA. A, NIGMS human/rodent somatic cell hybrid panel 1. The identity of the 540bp fragment present in all of the mouse/human hybrid cell lines is not known: hybrid NA10611 is a CHO/human hybrid and lacks this band. B, BIOS Corporation somatic cell panel. Amplification of CHO cell DNA yielded a diffuse signal smaller than the 294-bp band expected from PAM.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional

PAM

Gene

1579

fragment of this size from Chinese hamster ovary (CHO) or human DNA. The primers specific for human PAM amplify a fragment of 406 bp (Fig. 8). Analysis of two somatic cell hybrid panels indicated 0% discordancy of the PAM gene with chromosome 5 (Table 2). Since the BIOS Corporation panel (Fig. 86) contains cells with partial deletions of chromosome 5, it was possible to localize the PAM gene to the long arm of chromosome 5 (5q). The gene for human PC1 , a subtilisin-like endoprotease involved in prohormone processing, is also situated on the long arm of chromosome 5 (42).

cDNAs derived from rat medullary thyroid carcinoma (19). The two PAM cDNAs isolated from a medullary thyroid carcinoma exhibited conservative C/T changes at two additional positions (19); T was present at both of these positions in the genomic DNA characterized. The Type B PAM cDNA identified in medullary thyroid carcinoma diverges from rat PAM-l within exon 26, and none of the cDNAs that we have isolated from pituitary or atrial libraries contained a sequence similar to the unique 3’-end of type B rat PAM cDNA (19). Peptide biosynthesis involves the participation of many enzymes including furin, PC1 , PC2, carboxypeptidase E (CPE), and angiotensin-converting enzyme. In situ localization places the genes for PAM (Ouafik, L’H., M.G. Mattei, P. Giraud, B.A. Eipper, and R.E. Mains, manuscript in preparation) and PC1 within the same region on the long arm of chromosome 5 (42). Colocalization of the genes encoding PC1 and PAM to the long arm of chromosome 5 may be of functional significance. The genes encoding PC2 and furin are located on human chromosomes 20 and 15, respectively (42). Gene structures have been elucidated for only a few of these enzymes, and PAM is one of the most complex. The gene encoding human furin encompasses only 5.4 kb genomic DNA and consists of eight exons (45). The gene encoding rCPE encompasses 50 kb genomic DNA and consists of nine protein-coding exons; the smallest exon contains 100 bp, and introns range in size from 1.6-12.5 kb (46); alternative splicing is not used as a mechanism of generating diversity in CPE. The gene encoding human angiotensin-converting enzyme is also complex, comprising 26 exons spanning 21 kb; tissuespecific use of two alternate promoters generates two angiotensin-converting enzyme mRNAs (47). Significant regions of the PHM domain and dopamine P-monooxygenase are clearly related in evolution. In addition to the 29% amino acid sequence identity between hPHM and dopamine p-monooxygenase over the region pointed out by Southan and Kruse (30), we show here that many of the exon/intron junctions in the homologous region are conserved (Fig. 7). PAM and dopamine P-monooxygenase may have evolved from an ancestral gene composed of more exons than either current gene, with different introns lost in the evolution

DISCUSSION

A single complex gene encodes the two enzymes that must act sequentially to bring about the cu-amidation of peptides in endocrine tissues, the nervous system, the gastrointestinal system, and other tissues producing LYamidated peptides. The 27 exons that account for all of the PAM cDNAs characterized to date are distributed over more than 160 kb genomic DNA. The two catalytic domains (PHM and PAL) account for much of the multifunctional PAM protein. The PHM domain encompasses a maximum of 12 exons; the PAL domain encompasses a maximum of eight exons. As observed for other multifunctional enzymes, expression of truncated cDNAs encoding a single catalytic domain leads to the production of active enzyme; thus synthesis of the bifunctional PAM protein is not required for proper folding (23; see Ref. 45). The protein-coding exons average 106 bp in length, typical of vertebrate exons, but the internal introns, ranging from 200 to over 20,000 bp in length, are much longer than typically seen in vertebrate genes (43, 44). A single difference (nt 1212 in exon 12) was noted when comparing our previous cDNA sequences to the genomic sequence; His305 was encoded by CAC in the genomic sequence and by CAT in all of the PAM cDNAs characterized previously in this laboratory (14, 17, 20). A CAC codon was observed at the corresponding position in PAM cDNAs derived from Wistar rat pituitary (15), and both sequences were observed in the PAM

Table

2. Analvsis

Human

Chromosome

A. NIGMS human/rodent Concordant hybrids (+/+) or C-/H Discordant hybrids

(+/-) or C/+)

Discordancy (%) B. BIOS Corporation Concordant hybrids (+/+) or C-/H Discordant hybrids

(+I-) or G/+)

Discordancy Cell lines discordant

(%)

of Human 1

Somatic 2

somatic 9 11 9

Cell 4

cell hybrid 8 9 10

9

Hvbrid 5

Panels

6

7

mapping panel 18 11 9

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

X

Y

17

14

8

12

1 8

8

7

IO

15

10

11

IO

9

12

11

9

11

8

9

6

7

9

7

1

4

IO

6

50

33

39

50

39

6

22

56

33

0

7

9

IO

10

11

8

3

8

7

39 56 cell hybrid 5 6

50 0 panel 6 19

39

50

56

56

61

44

17

44

39

8

6

7

7

7

7

8

10

11

8

6

4

4

11

7

11

8

7

6

18

20

19

19

0

17

19

18

18

18

18

17

15

14

17

19

21

21

14

18

14

17

18

19

72

80

76

76

0

68

76

72

72

72

72

68

60

56

68

76

84

84

56

72

56

68

72

76

were

positive.

50 somatic 7

7

3

were scored as positive for a given human hybrids for each chromosome is indicated

44

chromosome if 10% or more of the cells analyzed along with the percent of discordant hybrids.

The number

of concordant

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

and

MOL 1580

ENDO.

of PAM and dopamine p-monooxygenase (48, 49). By comparison, the sequence identity of CPE and the pancreatic carboxypeptidases CPA and CPB is only 17-21%, and the location of only one exon/intron junction is conserved (46). As observed for many other genes (48), alternative splicing is used to create functional diversity from the single copy PAM gene. Two sites within the PAM gene account for most of the alternatively spliced forms of PAM mRNA. Alternative splicing determines whether PAM proteins will be integral membrane proteins or soluble proteins and whether the two catalytic domains will be separated from each other by endogenous endoproteases cleaving at a paired basic amino acid site (Fig. 7). Exon 16 is present in PAM RNAs in heart and central nervous system but is commonly spliced out of PAM mRNAs in pituitary or submaxillary gland (20). The presence of exon A in a bifunctional PAM protein requires the use of a cryptic 5’-splice site within the protein-coding region of the sequence specific to rPAM4 and not the retention of an unspliced intron as postulated by Bertelson et a/. (19). The presence or absence of the exon encoding the trans-membrane domain (exon 25) determines whether the COOH-terminus of PAM will be exposed to the cytosol or sequestered within the lumen of the secretory pathway. Three alternate 3’-untranslated regions have been identified in the PAM gene (exon 13, the retained part of intron P, and exon 27); all contain consensus poly(A) addition signals (Table 1) (50) although the presence of a poly(A) tail in rPAM-5 mRNAs has not been documented. Use of alternate poly(A) addition sites determines whether monofunctional rPAM-4 or nonfunctional rPAM-5 protein will be made; neither rPAM-4 nor rPAM5 is a major transcript in any of the tissues examined. Although the utility of producing rPAM-5 is unclear, expression of an inactive form of glutamic acid decarboxylase, which is also governed by alternative splicing, is closely regulated during development (51). The interdomain regions of the multifunctional proteins that have been characterized tolerate higher rates of mutation than the catalytic domains and are sites at which exogenous endoproteases can cleave to release active catalytic domains (28, 52). Comparison of the sequences of frog, bovine, rat, and human PAM identified exon 15, the part of exon 25 preceding the transmembrane domain and exon 26 as poorly conserved regions (4). The sizes of the active PHM and PAL fragments generated by limited endoproteolytic digestion of atrial membranes are consistent with the identification of exons 15 and 25 as protease-sensitive interdomain regions (22). The fact that exon 27 is highly conserved suggests that it is functionally important. When precautions are taken to prevent endoproteolytic degradation during purification, most of the multifunctional proteins that have been studied [e.g. CAD, arom, and fatty acid synthase (FAS)] are isolated as intact polypeptides containing all of the catalytic domains (24,26,28). In contrast, endoproteolytic cleavage of PAM into its component parts occurs in a tissue-

Vo16No.10

specific fashion and may be governed in part by the presence or absence of exon 16 (optional exon A). Expression of rPAM-1 and rPAM-2 in AtT-20 cells indicates that the presence of this region is essential if the resident enzymes are to separate the PHM and PAL catalytic domains by endoproteolytic cleavage (23). Exon 16, while clearly not essential for expression of either catalytic activity, is highly conserved and thus not typical of an interdomain region and may serve as a domain permitting controlled endoproteolytic processing. Expression of exons 25-27 is not required for production of catalytic activity, and these three exons may function as a routing domain. PAM transcripts that include exon 25 (rPAM-1, -2, and -3b) have a transmembrane domain and are integral membrane proteins. Expression of mutant forms of PAM-l truncated immediately after the trans-membrane domain and lacking all of the peptide encoded by exons 26 and 27 indicates that this part of the protein plays a role in the proper routing of the protein in AtT-20 cells (40). Multifunctional enzymes are especially common in the biosynthetic pathways of lower eukaryotes, where they catalyze two or more consecutive reactions (52). The pentafunctional arom proteins of fungi and yeast, which catalyze steps two to six in prechorismate polyaromatic amino acid biosynthesis, are thought to have arisen by fusion of individual ancestral genes (26, 52). The trifunctional CAD protein, which catalyzes the first three steps in mammalian de nova pyrimidine biosynthesis (53), and the multifunctional FAS protein, which catalyzes all of the steps from acetyl-coenzyme A and malonyl-coenzyme A to palmitate (28, 54) are also thought to have arisen by fusion of ancestral genes. The expression of PAM, like the expression of many of these other multifunctional enzymes, is tissue specific and developmentally and hormonally regulated. Despite the complexity of the PAM gene, transcription appears to be initiated at the same start site in several different tissues. By linking expression of PHM and PAL in the bifunctional PAM protein, coordinate regulation is ensured. It will be of great interest to elucidate the structure of the two amidating enzyme genes in Xenopus laevis. Further investigations will focus on the 5’-flanking region of the PAM gene and identification of the sequences responsible for regulation of its expression.

MATERIALS Construction

AND METHODS and Screening

of Rat Genomic

Libraries

High molecular weight genomic DNA prepared from rat liver was digested with a concentration of Sau3Al selected to yield 15 to 20-kb fragments. Digested DNA fragments (0.5 fig) were filled in with dATP and dGTP using Klenow DNA polymerase, ligated to the partially filled-in Xhol sites of XFIX vector arms (1 pg; Stratagene, La Jolla, CA) using Tq DNA ligase, and packaged with a Gigapack Gold packaging extract. The resultant library had a titer of 1.36 x 10” plaque forming units (pfu)/pg when plated on LE392 cells and a titer of 1.22 x 1 O6 when plated on P,392 cells. Triplicate lifts of 400,000 pfu

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional

PAM

Gene

1581

plated on LE392 cells were screened with three cDNA probes that span the sequence of bPAM-1 (Fig.1): 0.8 kb (1-781); 0.7 kb (782-1503); and 2.2 kb (1504-3724). Eight phage were plaque purified: G3, G5, and G7 hvbridized with the most 5’probe; G7, G13, and G16 hybridized with the mid-region probe; G17, G18, and G19 hybridized with the most 3’-probe. Overlaps between phage were established by comparing restriction digests as well as using short riboprobes prepared from the ends of neighboring phage to look for cross-hybridization. When the exons contained in these eight phage were identified, it was clear that they did not account for the entire sequence of rPAM-1 and that many of the clones isolated did not overlap. Sense and antisense oligonucleotide primers flanking the missing regions of rPAM-1 were prepared and used to synthesize fragments specific to the missing regions by PCR amplification; these fragments were radiolabeled by random priming and used as probes in further library screens. Five additional genomic clones (G2, Al, A2, G13a, and G19a) were isolated from an amplified commercial rat genomic library (Stratagene; Sprague-Dawley rat testis library) in the XDASHII vector. Phage Al and A2 were identified upon screening 270,000 pfu of the commercial genomic library with a probe spanning rPAM(1060-1397); A2 was also recoqnized by a probe spanning rPAM(709-978). No phage wereidentified by a probe corresponding to rPAM(l789-2034). When another 500,000 pfu from the same commercial library were screened with the rPAM(l789-2034) probe, no positive phage could be identified. The commercial librarv (600,000 pfu) was also screened using an end-labeled ojigonucleotidk probe corresponding to rPAM(523-553) and yielded phage G2. In order to close the gaps between nonoverlapping phage, riboprobes were prepared from the ends of genomic inserts adjacent to gaps. Phage DNA from G13 and G19 was cleaved with Rsal, and riboprobes were synthesized using the appropriate polymerase and L~-[~*P]CTP (800 &i/mmol). These riboprobes were used to screen 500,000 pfu from the commercial library; phage G13a (G13 3’-riboprobe) and G19a (G19 5’-riboprobe) were identified in this manner. Riboprobes prepared from G2 and G3 failed to identify additional sequences in the PAM gene. The 13 partially overlapping genomic clones span a total of approximately 160 kb genomic DNA, leaving six gaps (Fig. 4). Southern

Blot Analysis

High molecular weight Sprague Dawley rat liver DNA was digested to completion with several restriction endonucleases. Digested DNA was fractionated by electrophoresis on an 0.8% agarose gel at 0.7 V/cm for 14 h (55). DNA was transferred to a Hybond-N membrane (Amersham Corp., Arlington Heights, IL) using an LKB Vacu-Gene vacuum blotting unit (LKB, Piscataway, NJ). Prehybridization, hybridization, and washing were carried out as described (56). The 660-bp Sacl/EcoRI fragment of rPAM-1 [Fig.1 ; rPAM(3232-3886)] was labeled by random priming (Boehringer Mannheim, Indianapolis, IN) with @P]dCTP to a specific activity of 1 x 1 O9 cpm/pg. Characterization Fragments

of Phage-

and Exon-Containing

Phage DNA was purified from LE392 cells grown in NZCYM medium and infected at a multiplicity of infection of 0.005 using a polyethylene glycol procedure (55). Complete restriction maps were obtained using Sac1 and Smal, and additional restriction enzyme digests were analyzed when necessary. Maps were confirmed by hybridization with a set of [32P]CTPlabeled riboprobes of varying lengths. Riboprobes were synthesized from either end of the insert using T, or TS RNA polymerase for varying periods of time. Restriction fragments containing exons were identified by hybridization with rPAM cDNA probes or end-labeled oligonucleotide probes and subcloned (Fig. 4C). Fragments were subcloned in pUCl8 or pBSll SK-. Ligated material was used to transform Escherichia co/i NM522 cells, and restriction

digests of recombinant clones were analyzed on agarose gels. Sequencing of double-stranded DNA was performed using the dideoxy chain termination method (57) with the T, DNA polymerase sequencing kit (Pharmacia LKB) using vector primers, primers contained within rPAM-1, rPAM-4, or rPAM-5, and intron-specific primers. Oligonucleotides were synthesized on a MilliGen Cyclone Nucleic Acid Synthesizer (Millipore, Bedford, MA) using phosphoramidite chemistry. Sequence data were obtained for all exons, and exon-intron junctions were identified by comparing the genomic sequence to the cDNA sequences (14, 17, 20)‘. The single exonic sequence differing from published cDNA sequences was verified by sequencing both strands. lntronic sequences differing from the consensus 5’- and 3’-splice donor and acceptor sites and any ambiguous sequences were verified by sequencing both strands. Restriction mapping and PCR were used to localize exons within subcloned fragments or within the phage and to determine distances between exons. Sense and antisense oligonucleotide primers situated within the exon(s) in question were used in combination or individually with primers situated within the appropriate vector arms. PCR reactions were carried out in 50 ~1 vol in 10 mM Tris HCI, pH 8.3, at 25 C, 50 mM KCI, 1.5 mM MgCI,, 0.01% gelatin, 200 WM each of four dNTPs, and 1 WM each primer (58). Reactions typically contained l10 ng phage or plasmid DNA or 200-500 ng genomic DNA and 1.25 units AmoliTaa DNA oolvmerase (Perkin Elmer Cetus. Norwalk, CT). Samples’were b&layered with light mineral oii and subjected to 25-30 cycles of amplification in an MJ Research Thermal cycler (MJ Research Inc., Watertown, MA). The initial denaturation step was performed at 94 C for 5-7 min. Reaction conditions were generally as follows: denaturation at 94 C for 1 min, annealing at 50 C for 1 min, extension at 70 or 72 C for 2-4 min; the last extension time was lengthened to 10 min. Amplified DNA fragments were analyzed on agarose gels and visualized with ethidium bromide. Identification

of Exons

Not Found

in Genomic

Clones

Exons 7 and 17 were not contained within any of the genomic phage identified. Exon 7 was subsequently identified by amplification of genomic DNA with an antisense primer contained within exon 7 and a sense primer contained within exon 6. The amplified fragment was subcloned, and the exon/intron junctions were sequenced. Exon 17 was not recovered in any of the 13 genomic clones, and PCR amplification from either exon 16 or exon 18 into exon 17 failed to yield a product. Therefore, the methodology developed to extend cDNA sequences [rapid amplification of cDNA ends (RACE)] (59) was modified for use in identifying introns surrounding a known exon (biotin-RAGE) (60). The introns preceding and following exon 17 and the intron following exon 7 were identified in this way. In brief, rat liver genomic DNA was cut with a restriction enzyme recognizing a four-bp sequence not present in the exon of interest (in this case, Mbol). A poly(A) tail was added to the cut genomic DNA using terminal deoxynucleotidyl transferase as described in the RACE protocol (59). Amplification was then carried out using an exon-specific primer and the RAGE-hybrid primer, which consists of three consecutive restriction sites (EcoRI, Sacl. and BsolO6) followed bv an oligo(dT) tail (5’-GAAiTCGAGCTCATCGAT,,-3’). Souihern blot analysis was performed on an aliquot of the amplified DNA to test for the presence of sequences homologous to the exon in question. An aliquot of this reaction mix was aqain amplified using a biotinyjated cDNA-specific primer and-the RAGE-hvbrid primer. The oroducts of the second PCR reaction were bound to streptavidin-linked magnetic beads (Dynal Inc., Great Neck, NY), and nonbiotinylated contaminants were washed away. DNA bound to the magnetic beads was again subjected to PCR amplification using the same biotinylated ’ Sequence data across all exon/intron junctions have been submitted to GenBank, accession numbers LO1 668-L01693.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1582

ENDO.

1992

Vo16No.10

primer (or a nested, exon-specific primer) and the RAGE hybrid primer. The products of this second PCR amplification were fractionated by agarose gel electrophoresis, excised from the gel, subcloned, and sequenced. To control for errors introduced by PCR, at least two subcloned fragments were sequenced. In addition, intronic primers flanking exon 17 were used to amplify a genomic fragment including exon 17; the sequence of this subcloned fragment was identical to those obtained using the RAGE-hybrid primer (60). Primer

Extension

We would like to thank Dr. Ethylin W. Jabs (Department of Pediatrics and Medicine, The Johns Hopkins Hospital) for kindly providing DNA from hybrid cell line panels and advice on determining chromosomal localization (supported by NIH Grant HG-00373) Dr. Martin Rechsteiner (University of Utah) for providing his PESTFIND program, and Zina Garrett for help in preparing the manuscript.

Analysis

Total RNA was prepared from adult rat tissues using the acid guanidinium isothiocyanate-phenol-chloroform procedure (61). Poly(A)+ RNA was purified by oligo(dT)-cellulose affinity chromatography. An antisense oligonucleotide primer [rPAM(6346)] was radiolabeled using Tq polynucleotide kinase; after incubation at 70 C for 5 min, the labeled oligonucleotide was annealed to poly(A)+ RNA or total RNA in a-20 bl reaction in the presence of 83 mM Tris HCI, pH 8.3, 133 mM KCI, 1 and mM EDTA at 42 C for 2 h. The mixture was diluted to 40 ~1 with 20 ~1 transcription buffer consisting of 30 mM Tris HCI, pH 8.3, 16 mM MgCl*, 8 mM dithiothreitol, 16 pg/ml BSA, 1 mM each deoxynucleotide triphosphate, 40 U RNasin (Promega, Madison, WI), and 40 units avian myeloblastosis virus reverse transcriptase (Pharmacia LKB Biotechnology) (24). The samples were incubated at 42 C for 1 h. The products were analyzed by electrophoresis on a 6% polyacrylamide gel containing 8 M urea and detected by autoradiography. Chromosomal

Acknowledgments

Localization

Pairs of sense and antisense oligonucleotide primers capable of distinguishing rodent from human PAM were used to determine which members of two separate hybrid cell panels contained the gene for hPAM (62). The primer pairs were selected from within the lengthy exon containing the COOH-terminal end of the PAM protein and the 3’-untranslated region; since cDNAs for mouse and Chinese hamster PAM have not been characterized, primers were selected based on the sequence of rPAM and tested for species specificity. Both primers of the pair specific to rPAM [rPAM(3125-3141) and (3418-3402)] differ from the corresponding region of hPAM (18) at four of 17 sites. Both primers of the pair specific to hPAM [hPAM(2944-2963) and (3349-33291 differ from the corresponding region of rPAM [rPAM(3059-3078) and (34673447)] at five sites. The primer pairs are species specific when used to amplify 100 ng genomic DNA or 10 pg plasmid DNA; with larger amounts (10 ng) of human plasmid DNA, the rat PAM primer pair amplifies the hPAM plasmid. The rPAM primers amplify a band of the expected size when used with mouse DNA; the rPAM primers do not efficiently amplify a fragment of the expected size from CHO cell DNA. Two somatic cell hybrid cell panels (kindly provided by Dr. Ethylin W. Jabs, Department of Pediatrics and Medicine, The Johns Hopkins Hospital) were screened for the presence of hPAM by using the PCR: NIGMS rodent somatic cell hybrid panel 1 consists of 17 mouse/human hybrids and 1 CHO/ human hybrid; the BIOS Corporation (New Haven, CT) somatic cell hybrid panel consists of 25 CHO/human hybrids. PCR amplification was carried out as described with 0.1 O-O.25 pg genomic DNA except that dNTP concentrations were 1 mM and all four primer concentrations were 1 KM. An initial denaturation step was performed at 94 C for 4 min. Amplification (30 cycles) was carried out by denaturation at 94 C for 1 min, annealing at 50 C for 1 min and extension at 72 C for 2 min; in the final round of amplification the extension time was increased to 10 min. A 20-~1 aliquot of each sample was fractionated on a 1% agarose gel and visualized with ethidium bromide.

Received June 9, 1992. Revision received July 21, 1992. Accepted July 21, 1992. Address requests for reprints to: Dr. Betty Eipper, Department of Neuroscience, The Johns Hopkins University School of Medicine, 725 North Wolfe Street, Baltimore, Maryland 21205. This work was supported by NIH Grant DK-32949, National Institute on Drug Abuse Grants DA-00098 and DA-00097, and American Heart Association Grant 89831.

REFERENCES 1. Bradbury AF, Smyth DG 1991 Peptide Amidation. Trends Biochem Sci 16:112-l 15 2. Sossin WS, Fisher JM, Scheller RH 1989 Cellular and molecular biology of neuropeptide processing and packaging. Neuron 2:1407-l 417 3. Mains RE, Dickerson IM, May V, Stoffers DA, Perkins SN, Ouafik L’H, Husten EJ, Eipper BA 1990 Cellular and molecular aspects of peptide hormone biosynthesis. Front Neuroendocrinol 11:52-89 4. Eipper BA, Stoffers DA, Mains RE 1992 The biosynthesis of neuropeptides: peptide a-amidation. Annu Rev Neurosci 15:57-85 5. Bradbury AF, Smyth DG 1987 Enzyme-catalysed peptide amidation. Eur J Biochem 169:579-584 6. Young SD, Tamburini PP 1989 Enzymatic peptidyl CYamidation proceeds through formation of an a-hydroxyglycine intermediate. J Am Chem Sot 111 :1933-l 934 7. Katopodis AG, Ping D, May SW 1990 A novel enzyme from bovine neurointermediate pituitary catalyzes dealkylation of a-hydroxyglycine derivatives, thereby functioning sequentially with peptidylglycine a-amidating monooxygenase in peptide amidation. Biochemistry 29:61156120 8. Perkins SN, Husten EJ, Mains RE, Eipper BA 1990 pHDependent stimulation of peptidylglycine a-amidating monooxygenase activity by a granule-associated factor. Endocrinology 127:2771-2778 9. Tajima M, lida T, Yoshida S, Komatsu K, Namba R, Yanagi M, Noguchi M, Okamoto H 1990 The reaction product of peptidylglycine a-amidating enzyme is a hydroxyl derivative at a-carbon of the carboxyl-terminal glycine. J Biol Chem 265:9602-9605 10. Suzuki K, Shimoi H, Kawahara T, Matsuura Y, Nishikawa Y 1990 Elucidation of amidating reaction mechanism by frog amidating enzyme, peptidylglycine cu-hydroxylating monooxygenase, expressed in insect cell culture. EMBO J 9:4259-4265 11. Perkins SN, Husten EJ, Eipper BA 1990 The 108-kDa peptidylglycine a-amidating monooxygenase precursor contains two separable enzymatic activities involved in peptide amidation. Biochem Biophys Res Commun 171:926-932 12. Kato I, Yonekura H, Tajima M, Yanagi M, Yamamoto H, Okamoto H 1990 Two enzymes concerned in peptide hormone a-amidation are synthesized from a single mRNA. Biochem Biophys Res Commun 172:197-203 13. Eipper BA, Park LP, Dickerson IM, Keutmann HT, Thiele

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

Multifunctional

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26. 27.

28.

29.

30.

31.

32.

PAM

Gene

EA, Rodriguez H, Schofield PR, Mains RE 1987 Structure of the precursor to an enzyme mediating COOH-terminal amidation in peptide biosynthesis. Mol Endocrinol 1:777790 Eipper BA, Green CB-R, Campbell TA, Stoffers DA, Keutmann HT, Mains RE, Ouafik L’H 1992 Alternative splicing and endoproteolytic processing generate tissue-specific forms of pituitary peptidylglycine a-amidating monooxygenase (PAM). J Biol Chem 267:4008-4015 Kato I, Yonekura H, Yamamoto H, Okamoto H 1990 Isolation and functional expression of pituitary peptidylglytine a-amidating enzyme mRNA. FEBS Lett 269:319323 Ohsuye K, Kitano K, Wada Y, Fuchimura K, Tanaka S, Mizuno K, Matsuo H 1988 Cloning of cDNA encoding a new peptide C-terminal a-amidating enzyme having a putative membrane-spanning domain from Xenopus laevis skin. Biochem Biophys Res Commun 150:1275-l 281 Stoffers DA, Green CB-R, Eipper BA 1989 Alternative mRNA splicing generates multiple forms of peptidyl-glytine a-amidating monooxygenase in rat atrium. Proc Natl Acad Sci USA 86:735-739 Glauder J, Ragg H, Rauch J, Engels JW 1990 Human peptidylglycine a-amidating monooxygenase: cDNA, clon. -. ing and functional expression of a truncated form in COS cells. Biochem Biophvs Res Commun 169:551-558 Bertelsen AH, Be&dry GA, Galella EA, Jones BN, Ray ML, Mehta NM 1990 Cloning and characterization of two alternatively spliced rat cu-amidating enzyme cDNAs from rat medullary thyroid carcinoma. Arch Biochem Biophys 279:87-96 Stoffers DA, Ouafik L’H, Eipper BA 1991 Characterization of novel mRNAs encoding enzymes involved in peptide a-amidation. J Biol Chem 266:1701-l 707 Mizuno K, Ohsuye K, Wada Y, Fuchimura K, Tanaka S, Matsuo H 1987 Cloning and sequence of cDNA encodinq a peptide C-terminal &amidating enzyme from Xenopui laevis. Biochem Bioohvs Res Commun 148:546-552 Husten EJ, Eipper BA i991 The membrane-bound bifunctional peptidylglycine a-amidating monooxygenase protein. J Biol Chem 266:17004-l 7010 Milgram SL, Johnson RC, Mains RE 1992 Expression of individual forms of peptidylglycine cu-amidating monooxygenase in AtT-20 cells: endoproteolytic processing and routing to secretory granules. J Cell Biol 117:717-728 Bein K, Simmer JP, Evans DR 1991 Molecular cloning of a cDNA encoding the amino end of the mammalian multifunctional protein CAD and analysis of the 5’-flanking region of the CAD gene. J Biol Chem 266:3791-3799 Davidson JN, Rao GN, Niswander L, Andreano C, Tamer C, Chen K-C 1990 Organization and nucleotide sequence of the 3’ end of the human CAD gene. DNA Cell Biol 9:667-676 Hawkins AR 1987 The complex Arom locus of Aspergihs nidulans. Curr Genet 11:491-498 Kameda K, Goodridge AG 1991 Isolation and partial characterization of the gene for goose fatty acid synthase. J Biol Chem 266:419-426 Witkowski A, Rangan VS, Randhawa ZI, Amy CM, Smith S 1991 Structural organization of the multifunctional animal fatty-acid synthase. Eur J Biochem 198:571-579 Stewart LC, Klinman JP 1988 Dopamine beta-hydroxylase of adrenal chromaffin granules: structure and function. Annu Rev Biochem 57:551-592 Southan C, Kruse LI 1989 Sequence similarity between dopamine P-hydroxylase and peptide or-amidating enzyme: evidence for a conserved catalytic domain. FEBS Lett 255:116-l 20 Kobayashi K, Kurosawa Y, Fujita K, Nagatsu T 1989 Human dopamine p-hydroxylase gene: two mRNA types having different 3’-terminal regions are produced through alternative oolvadenvlation. Nucleic Acids Res 17:10891102 ’ ’ ’ Ruoslahti E, Pierschbacher MD 1987 New perspectives

1583

33. 34.

35.

36.

37.

38.

39.

40.

41.

42.

43. 44. 45.

46.

47.

48.

49.

50.

51.

52.

in cell adhesion: RGD and integrins. Science 238:491497 Albelda SM, Buck CA 1990 lntegrins and other cell adhesion molecules. FASEB J 4:2868-2880 Hatsuzawa K, Hosaka M, Nakagawa T, Nagase M, Shoda A, Murakami K, Nakayama K 1990 Structure and expression of mouse furin, a yeast Kex2-related protease. J Biol Chem 265:22075-22078 Smeekens SP, Steiner DF 1990 Identification of a human insulinoma cDNA encoding a novel mammalian protein structurally related to the yeast dibasic processing protease KexP. J Biol Chem 265:2997-3000 Seidah NG, Marcinkiewicz M, Benjannet S, Gaspar L, Beaubien G, Mattei MG, Lazure C, Mbikay M, Chretien M 1991 Cloning and primary sequence of a mouse candidate prohormone convertase PC1 homologous to PC2, furin, and Kex2: distinct chromosomal localization and messenger RNA distribution in brain and pituitary compared to PC2. Mol Endocrinol 5:ll l-l 22 Bloomauist BT. EiDDer BA. Mains RE 1991 Prohormoneconverting enzymes: regulation and evaluation of function using antisense RNA. Mol Endocrinol 5:2014-2024 Eipper BA, Perkins SN, Husten EJ, Johnson RC, Keutmann HT, Mains RE 1991 Peptidyl-a-hydroxyglycine 01amidating lyase. J Biol Chem 266:7827-7833 Rogers S, Wells R, Rechsteiner M 1986 Amino acid sequences common to rapidly degraded protein: the PEST hypothesis. Science 234:364-368 Milgram SL, Tausk FA, Eipper BA, Mains RE, Routing of peptidylglycine a-amidating monooxygenase (PAM) in transfected endocrine and nonendocrine cell lines. Program of the 74th Annual Meeting of The Endocrine Society, San Antonio, TX, 1992 (Abstract) Shaoiro MB. Seneaathv P 1987 RNA solice iunctions of different classes of eukaryotes: sequence siatistics and functional implications in gene expression. Nucleic Acids Res 15:7155-7174 Seidah NG, Mattei MG, Gaspar L, Benjannet S, Mbikay M, Chretien M 1991 Chromosomal assignments of the genes for neuroendocrine convertase PC1 (NECl) to human 5ql5-21, neuroendocrine convertase PC2 (NEC2) to human 2Opll.l-11.2, and furin (mouse 7[Dl -E2] region). Genomics 11 :103-l 07 Gilbert W 1985 Genes-in-pieces revisited. Science 228:823-824 Hawkins JD 1988 A survey on intron and exon lengths. Nucleic Acids Res 16:9893-9908 Roebroeck AJM, Schalken JA, Leunissen JAM, Onnekink C, Bloemers HPJ, Van de Ven Wim JM 1986 Evolution conserved close linkage of the c-fex/fps proto-ocogene and genetic sequences encoding a receptor-like protein. EMBO J 5:2197-2202 Jung Y-K, Kunczt FJ, Pearson RK, Dixon JE, Fricker LD 1991 Structural characterization of the rat carboxypeptidase-E gene. Mol Endocrinol5:1257-1268 Hubert C, Houot A-M, Corvol P, Soubrier F 1991 Structure of the angiotensin l-converting enzyme gene. J Biol Chem 266:15377-l 5383 Smith CWJ, Patton JG, Nadal-Ginard B 1989 Alternative splicing in the control of gene expression. Annu Rev Genet 23:527-577 Marchionni M, Gilbert W 1986 The triosephosphate isomerase gene from maize: introns antedate the plantanimal divergence. Cell 46:133-l 41 Wickens M 1990 How the messenger got its tail: addition of poly(A) in the nucleus. Trends Biochem Sci 15:277281 Bond RW, Wyborski RJ, Gottlieb DI 1990 Developmentally regulated expression of an exon containing a stop codon in the gene for glutamic acid decarboxylase. Proc Natl Acad Sci USA 87:8771-8775 Duncan K, Edwards RM, Coggins JR 1987 The pentafunctional arom enzyme of Saccharomyces cerevisiae is

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

MOL 1584

53.

54.

55.

56.

57.

END0.1992

a mosaic of monofunctional domains. Biochem J 246:375-386 Faure M, Camonis JH, Jacquet M 1989 Molecular characterization of a Dictyostelium discoidem gene encoding a multifunctional enzyme of the pyrimidine pathway. Eur J Biochem 179:345-358 Amy CM, Williams-Ahlf B, Naggert J, Smith S 1992 Intronexon organization of the gene for the multifunctional animal fatty acid synthase. Proc Natl Acad Sci USA 89:11051108 Sambrook J, Fritsch EF, Maniatis T 1989 Molecular Cloning, ed 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor Ouafik L’H, May V, Keutmann HT, Eipper BA 1989 Developmental regulation of peptidylglycine a-amidating monooxygenase (PAM) in rat heart atrium and ventricle. J Biol Chem 264:5839-5845 Sanger F, Nicklen S, Coulson AR 1977 DNA sequencing

Vo16No.10

58.

59.

60. 61. 62.

with chain-terminating inhibitors. Proc Natl Acad Sci USA 7415463-5467 lnnis MA, Gelfand DH 1990 Optimization of PCR. In: lnnis MA, Gelfand DH, Sninsky JJ, White TJ (eds) PCR Protocols. A Guide to Methods and ADDlications. Academic Press, Inc., New York, pp 3-12 ’ ’ Frohman MA 1990 RACE: raoid amolification of cDNA ends. In: lnnis MA, Gelfand DH, S&sky JJ, White TJ (eds) PCR Protocols. A Guide to Methods and ADDliCaiions. Academic Press, Inc., New York, pp 28-38 ’ ’ Bloomquist BT, Johnson RC. Mains RE. Rapid isolation of flanking genomic DNA using biotin-RAGE. DNA Cell Biol, in press Chomczynski P, Sacchi N 1987 Single-step method of RNA isolation of acid guanidinium thiocyanate-phenolchloroform extraction. Anal Biochem 162:156-l 59 Dionne CA, Kaplan R, Seuanez H, O’Brien SJ, Jaye M 1990 Chromosome assignment by polymerase chain reaction techniques: assignment of the oncogene FGF-5 to human chromosome 4. BioTechniques 8:190-194

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 December 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved.

intron organization of catalytic, processing, and routing domains.

Peptidylglycine alpha-amidating monooxygenase (PAM; EC 1.14.17.3) is a multifunctional protein containing two enzymes that act sequentially to catalyz...
2MB Sizes 0 Downloads 0 Views