Eur. J. Biochem. 195, 535-541 (1992) 0 FEBS 1991 0014295691000635

Structural requirements for the binding of the pituitary adenylate-cyclase-activating peptide to receptors and adenylate-cyclase activation in pancreatic and neuronal membranes Philippe GOURLET, Marie-Claire WOUSSEN-COLLE, Patrick ROBBERECHT, Philippe DE NEEF, Annick CAUVIN, Marie-Claire VANDERMEERS-PIRET, Andre VANDERMEERS and Jean CHRISTOPHE Department of Biochemistry and Nutrition, Medical School, Universite Libre de Bruxelles, Belgium (Received July 16, 1990) - EJB 90 0852

PACAP (pituitary adenylate-cyclase-activating peptide)-binding receptors were investigated in membranes from the rat pancreatic acinar cell line, AR 4-2J, the rat hippocampus and the human neuroblastoma cell line NB-OK, by '251-PACAP(1-27) (amino acid residues 1-27 of N-terminal amidated PACAP) binding and adenylate cyclase activation. The relative binding of "'I-PACAP(1- 27) to the receptor, and ability to activate adenylate cyclase were PACAP 3 PACAP(1- 27) > PACAP(2 - 38) > PACAP(1- 9)-VIP(10- 28)(PACAPVIP) > PACAP(2 - 27) > [Ser9, Tyrl31VIP > [Tyrl3]VIP 3 [Ser9]VIP 3 VIP(1- 23)-PACAP(24 - 27)(VIPPACAP) > VIP (vasoactive intestinal peptide). The N-terminal moiety of PACAP(1-27) was more important than the three amino acids at the C-terminus for '251-PACAP(1 -27)-binding site recognition. For rat pancreatic 12'I-VIP-binding sites tested with '2'I-VIP, the order of binding affinity was PACAP = PACAP(1- 27) 3 VIP = [Ser9]VIP = [Tyrl3]VIP = [Ser9,Tyrl3]VIP 3 PACAP-VIP 2 VIP-PACAP > PACAP(2-38) = PACAP(227). Pancreatic ' 'I-VIP-binding sites , when compared to '''I-PACAP(l - 27)-binding sites, showed little specificity and only weak coupling, so that PACAP and VIP-PACAP acted only as partial VIP agonists on adenylate cyclase.

PACAP (pituitary adenylate-cyclase-activating polypeptide) was isolated by Arimura et al. [l] from ovine hypothalamus on the basis of its ability to stimulate adenylate cyclase in rat anterior-pituitary cell cultures. The molecular cloning and characterization of ovine and human cDNA for PACAP are now established [2], indicating a similar amino acid sequence in both animal species. PACAP(1-27) (amino acid residues 1 - 27 of N-terminal amidated PACAP) which might reflect proteolytic processing of the triplet Gly28-Lys29-Arg30 and amidation of Gly28 in PACAP, is also present in the ovine hypothalamus [2]. PACAP(1-27) exhibits 68% similarity to vasoactive intestinal peptide (VIP) (Table l), but cannot be considered as a mere VIP variant for the following reasons: (a) PACAP is at least 1000-fold more potent than VIP in stimulating adenylate cyclase in cultured rat pituitary cells [l, 31; (b) PACAP increases CAMP levels in cultured rat astrocytes, whereas VIP is inefficient in this respect [4]; (c) PACAP cDNA is related to growth-hormone-releasing-factor cDNA and not to VIP cDNA [2]; (d) highly specific '251-PACAP(1 - 27)-binding Correspondence to J. Christophe, Department of Biochemistry and Nutrition, Medical School, Universite Libre de Bruxelles, 115 Bd. de Waterloo, B-1000 Brussels, Belgium Abbreviations. VIP, vasoactive intestinal peptide; PACAP, pituitary adenylate cyclase activating peptide; PACAP(1- 27), amino acid residues 1-27 of N-terminal amidated PACAP; PACAP-VIP, PACAP( 1 - 9)-VIP(10- 28); VIP-PACAP, VIP(1- 23)-PACAP(2427); K,,,, concentration exerting half-maximal stimulation of adenylate cyclase. Enzymes. Adenylate cyclase (EC 4.6.1.I); pyruvate kinase (EC 2.7.1.40).

sites are present in the rat pancreatic acinar cell line, AR 42J, [5] and the human neuroblastoma cell line, NB-OK, [6] and also in various rat brain areas (Cauvin et al., unpublished data). However, PACAP binds with high affinity to '251-VIPbinding sites in pancreas, lung and liver membranes from rat ([6] and unpublished results). The purpose of the present work was to identify which part of the PACAP molecule is responsible for the recognition of specific '2'I-PACAP(1- 27)-binding sites. To achieve this, we synthesized PACAP fragments and also VIP analogues in which one or two amino acids were replaced by their counterparts in PACAP, and two chimeric molecules combining the C-terminal part of VIP and the N-terminal part of PACAP or vice versa. All peptides were tested for their ability to occupy receptors and to stimulate adenylate cyclase in three membrane preparations rich in "'I-PACAP(1 - 27)-binding sites (the rat pancreatic carcinoma acinar cell line AR 4-2J, rat hippocampus and the human neuroblastoma cell line NBOK). Rat pancreatic membranes rich in '2'I-VIP binding sites served as a reference. MATERIALS AND METHODS Cell culture and crude membrane preparation

The human neuroblastoma cell line, NB-OK, was grown in RPMI medium enriched with 10% (by vol.) foetal calf serum (from Gibco), 20 mM glutamine and antibiotics (100 1U/ml penicillin/100 pg/ml streptomycin). The pancreatic acinar cell line, AR 4-2J, was grown in Dulbecco's modified Eagle's medium supplemented with 10% (by vol.) foetal

536 Table 1. Amino ucid sequences qf PACAP, P A C A P ( 1 - 2 7 ) , VIP, VIP analogues, peptide histidine isoleucinamide ( P H I ) and helodermin Similarities with PACAP(1-27) are underlined. The column on the right indicates the number of amino acid residues identical to those in PACAP(1- 27) All peptides except helodermin are amidated Peptide

Amino acid sequence

PACAP PACAP( 1 - 27) VIP [Ser9]VIP [Tyrl3]VIP [Ser9,Tyrl3]VIP PACAP-VIP VIP-PACAP PHI Helodermin

H - S - n - c - I - F - r - D - s - Y - ~ - ~ - ~ - R - K - ~ - ~ - ~ - ~ - K - K - G - K - R - Y - K - Q -27 R-V-K-N-K

1

5

Identical amino acids 10

15

calf serum, 20 mM sodium pyruvate, 20 mM glutamine and antibiotics (100 IU/ml penicillin, 100 pg/ml streptomycin and 2.5 pg/ml amphotericin B). At confluence, the cells were detached with a rubber policeman, centrifuged at 100 x g for 5 min, rinsed with the culture medium and lyzed in hypotonic 1 mM NaHC03 (pH 7.0), then quickly frozen in liquid nitrogen. For crude membrane preparations, the lyzate was defrosted and centrifuged at 100 x g for 10 min at 4' C. The resulting supernatant was centrifuged at 20000 x g for 10 min. The pellet was homogenized in 1 mM NaHC03 and immediately assayed. Membranes from rat hippocampus were prepared as follows: hippocampi were dissected out on ice and immediately homogenized in 10 vol. of ice-cold 0.25 M sucrose. After filtration through medical gauze, the homogenate was centrifuged at 800 x g for 10 min. The supernatant was centrifuged at 28 000 x g for 15 min and pelleted membranes, homogenized in 1 mM NaHC03,were used immediately. Rat pancreatic membranes were prepared as previously described [7] but 2-mercaptoethanol was omitted from all buffers [8]. They were stored in liquid nitrogen before use. Radioiodination and purification oftracers

VIP was radioiodinated and '251-VIP was purified as previously described [9, 101. 1251-PACAP(1-27) was also radioiodinated by the chloramine T method as described in [5], and further purified by HPLC. The radioiodination reaction was stopped by dilution, and the mixture immediately injected onto a reverse-phase 0.39 cm x 30 cm pBondapak C18 column (Waters Associates, Milford, MA, USA) equilibrated in solvent A (0.1/n' trifluoroacetic acid/5% CH3CN.) Elution was performed at a flow rate of 1 ml/min with a 0 - 35% linear gradient of solvent B (0.1 Yn trifluoroacetic acid/80% CH3CN) over 5 min, followed by a 35 - 75% linear gradient of solvent B over 40 min. Free '''1 and unlabelled PACAP were separated from three or four peaks of trichloroacetic-acid-precipitable 12'I. The major radioactive peak, eluting with 54% solvent B, showed the highest capacity to bind to AR 4-2J and neuroblastoma cell membranes. The biologically active fractions were pooled, CH3CN was evaporated, and the residue diluted in 10 mM sodium phosphate (pH 7.4) and 150 mM NaCl, containing 0.02% Tween 20 to avoid peptide adsorption to the tubes. The tracer was stored at -20"C, under which conditions high-affinity and low non-specific binding proper-

20

25

30

35

19 20 20 21 22 22 13 14

ties were maintained for two months. Specific radioactivity was 0.5 mCi/nmol for 12sI-VIP and 2 mCi/nmol for 1251PACAP(1- 27).

'

Binding of 251-VIP to rat pancreatic membranes and of'251-PACAP(I -27) to rut A R 4-25, rut neuroblastoma and rat hippocumpus membranes

Binding was performed in 50 mM Tris/maleate (pH 7.4) containing 5 mg MgC12,0.5 mg/ml bacitracin, 1O/O (massivol.) bovine serum albumin, 20 -- 30 pM 12sI-VIP or 5 - 10 pM '251-PACAP(1 - 27), increasing concentrations of unlabelled peptide and approximately 2,3,5 and 30 pg membrane protein for hippocampus, AR 4-25, NB-OK and pancreatic membranes, respectively, in a total volume of 120 pl. Incubations were conducted at 37°C for 30 min for 1251-VIP,or 20 min for 1251-PACAP(1-27), to allow equilibration of binding. At the end of the incubation, the samples were diluted in 2 ml ice cold 0.1 M sodium phosphate buffer (pH 7.4) and immediately filtered through glass-fiber filters (GjFC, Whatman, Maidstone, Kent, UK) soaked in 0.1% poly(ethy1eneimine) for at least 24 h. The filters were rinsed three times with 2 ml of the same phosphate buffer, and their radioactivity was measured in a multichannel LKB Gammacounter. Nonspecific binding of 12'I-PACAP(1 - 27) was determined in the presence of 0.1 pM PACAP(1-27), or that of I2'I-VIP in the presence of 1 pM VIP. It represented about 5% of the total radioactivity added, and was essentially due to tracer binding to the filter. The ratio of total/non-specific binding was approximately three in the absence of non-radioactively labelled peptide. For each 251-tracerand meinbrane preparation, we verified that the tracer concentration did not affect the concentration of unlabelled peptide necessary for half-maximal inhibition of ''1 binding, so that the median inhibitory concentration values of unlabelled peptides were close to Kd values. Adenylate cyclase assa-v

This assay was performed according to the procedure of Salomon et al. [ll]. Membrane protein, used at the same concentrations as in the binding assays, was incubated in a total volume of 60 p1 containing 0.5 mM [IX-~~PIATP, 10 pM GTP, 5 mM MgC12, 0.5 mM EGTA, 1 mM CAMP, 1 mM theophylline, 10 mM phosphoenolpyruvate, 30 pg/ml pyruvate kinase and 30 mM Tris/HCl, at a final pH of 7.5. The

537 AR 4 - 2 J CELL MEMBRANES

9

100-

PACAP (1-27)

4

b-

0 3

m

rcu a

VIP

d I

,?i

4

PACAP (2-27)

50-

a H I

N Lo 4

L L 0

a-e

0-

z

B

F 100 PACAP 11-27] \

.'

[Ser9,Tyri31 VIP

PACAP

l o g [Pept ide] /M Fig. 1. Effects of increasing concentrations ofpeptides on '251-PACAP(I-27)-binding ( A ) and adenylate cyclase stimulation ( B ) in A R 4-2J cell membranes. In A and B the following symbols were used : ( 0 )PACAP(1- 27), (0)PACAP(2 - 27), (A)PACAP, and ( A ) PACAP(2 38). In C and D the following symbols were used: ( 0 )VIP, ( A ) [Ser9]VIP, (V)(Tyrl3]VIP, (0)[SerY,Tyrl3]VIP, (0)PACAP-VIP, and (B) VIP-PACAP. The results were the means of at least three experiments each performed in duplicate. Binding data were expressed as a percentage of '251-PACAP specifically bound in the absence of unlabelled peptide. Adenylate cyclase activity was expressed as a percentage of the increase in CAMPproduced in the presence of 0.1 pM PACAP(1-27). The basal activity was 134 8 pmol cAMP produced . min-' . mg protein-' and the activity stimulated by 0.1 pM PACAP was 393 f 25 pmol cAMP . min-' . mg protein-' (means f SEM, n = 4). Experiments were performed in duplicate

reaction was initiated by adding membranes, and was terminated after a 12-min incubation at 37" C by adding 0.5 m10.5 M SDS containing 0.5 mM ATP, 0.5 mM cAMP and 20000 cpm [8-3H]cAMP (to determine cAMP recovery). cAMP was separated from ATP by two successive chromatographies on Dowex 50 W-X8 and neutral alumina. Under all conditions tested, cAMP production was linear during the incubation period, and proportional to the amount of protein added.

degradation in a 477 A sequencer (Applied Biosystems, Foster City, CA, USA). Other chemicals All reagents for peptide synthesis were purchased from Novabiochem (Laufelfingen, Switzerland). The origins of the reagents used for peptide radioiodination, adenylate cyclase assay and binding assay is detailed in [lo].

Pep t ides PACAP was a gift from Dr. A. Arimura and Dr. D. H. Coy (Tulane University, New Orleans, LA, USA), VIP and PACAP(2 - 38) were from Dr. J. P. Durieux (Novabiochem, Laufelfingen, Switzerland). PACAP(1- 27), PACAP(2 - 27), [Ser9]VIP, [Tyrl3]VIP, [Ser9,Tyrl3]VIP, PACAP-VIP, and VIP-PACAP (see Table 1) were synthesized with a Biotronic Synostat P automatic solid-phase peptide synthesizer (Braun, Melsungen, F.R.G.) using a 4-(2',4'-dimethoxyphenylfluoren9-y1methoxycarbonylaminomethyl)phenoxy resin and fluoren-9-ylmethoxycarbonyl-labelled amino acids activated with N-hydroxybenzotriazol and benzotriazol-I-yloxytris(dimethy1amino)phosphonium hexafluorophosphate [12]. The peptides were purified by HPLC and their conformity established by both total amino acid composition and Edman

RESULTS Interaction of PACAP(1- 27), PACAP, PACAP fragments, V I P and V I P analogues with specijk '251-PACAP(I-27)binding sites in membranes from rat pancreatic carcinoma acinar cell line A R 4-2J, human neuroblastoma cell line NB-OK and rat hippocampus The present data (Figs. 1 and 2, Table 2) confirmed the presence of specific and selective high-affinity 1251PACAP(1-27)-binding sites in rat pancreatic AR 4-25 membranes [5] and human neuroblastoma NB-OK membranes [6]. We also showed similar receptors to be present in rat brain hippocampus membranes (Fig. 3, Table 2). Like AR 4-25 and NB-OK 1251-PACAP(1- 27)-binding sites, rat-brain 1251-

538 Table 2. K d of ten PACAP(l-27)-related and VIP-related peptides required f o r 50% binding inhibition of '251-PACAP in rat A R 4-2J, human NB-OK and rat hippocampus membranes and '''I-VIP in rat pancreatic membranes SEM ( n 3). Experiments were performed in duplicate Values are the means Peptide

with '251-PACAP(1 -27)-binding sites on

Kd with '251-VIP-binding sites on

Kd

~-

human NB-OK

rat AR 4-25

rat hippocampus

rat pancreas

nM PACAP(1- 27) PACAP PACAP(2 - 27) PACAP(2 - 38) VIP [Ser9]VIP [Tyrl3]VIP [Ser9,Tyrl3]VTP VIP-PACAP PACAP-VIP

+ +

0.5 f 0.3 0.3 f 0.3 40 +_ 10 4 i 2 500 1000 100 200 300 & 100 100 2 20 300 50 20 5

1.5k 0.3 0.9k 0.1 400 k 20 10 f 2 % 3000 =z 1000 =z 2000 500 f 100 % 2000 150 f 30

0.3 0.1 0.2 f 0.1 20 5 2 f l 150 + I 0 50 f 1 0 40 + 1 5 15 2 100 2 2 0 8 + I

+ +

+

+

*

1 f 0.2 0.9 & 0.2 40 2 1 0 30 2 1 0 3 + I 3 * l 3 + I 3 f 0.5 10 2 6 k 1

+

HUMAN NEUROBLASTOMA CELL MEMBRANES

A

L

-Y

1

VIP

PACAP-VIP

~~~

Iserg,Tyri31 VIP

;sB PACAP-VIP

\

\ /

~ ~ eiyrl31 r ? VIP

[ ~ e r 9VI1P VID-PACAP ,---VIP

~~

-11 -10

-9

-e

-7

'

/ I

'

-10 -6 log [Peptidel/M

1

-'9 -E

8

8

-7

-6

Fig. 2. Effects of' increasing concentrations ofpeptides on '251-PACAP(I -27)-binding (A, C ) and adenylate cyclase stimulation l B , D ) in NBO K ce1lmembranr.s. The symbols and representation of results are as in Fig. 1. For adenylate cyclase activity (mean f SEM, n = 3), experiments were performed in duplicate: the basal value was 204 f 15 pmol CAMP . min-' . mg protein-' and the stimulated value in the presence of 0.1 pM PACAP was 594 f 22 pmol CAMP . min-' . mg protein-'

PACAP(1- 27)-binding sites recognized VIP with very low affinity, and did not interact with the parent peptides, peptide histidine isoleucinamide, secretin, growth hormone releasing factor and glucagon each at 1 pM (data not shown). The order of binding affinities of the peptides was similar for the three membrane preparations: PACAP 3 PACAP(1- 27) >

PACAP(2 - 38) > PACAP-VIP > PACAP(2 - 27) > [Ser9,Tyrl3]VIP > [Tyrl3]VIP 3 [Ser9]VIP > VIP-PACAP > VIP. However, these preparations differed in their Kd values: hippocampus membranes presented a threefold and fivefold higher Kd for PACAP(1-27) than AR 4-25 and NBOK membranes, respectively. NB-OK membranes showed

539 RAT HIPPOCAMPUS MEMBRANES

A

0

z

n

100

0 3

m

-

-

.rN

a I

2 U

50

a H

LD

4 cu

0 U

ae

0

PACAP-VIP

\

PACAP 12-27] -11 -10

-9 -8

-7

-6

-10

-9 -8

-7

-6

l o g [Peptide] /M

Fig. 3. Effects of increasing concentrations ofpeptides on 12'I-PACAP(I -27)-binding ( A and C ) and adenylate cyclase stimulation ( B and D ) in membranes from rat brain hippocampus. Same symbols and representation of results as in Figs 1 and 2. For adenylate cyclase activity (mean SEM, n = 4), experiments were performed in duplicate: the basal value was 72 f 7 pmol CAMP.min-' 'mg protein-' and the stimulated value in the presence of 0.1 pM PACAP was 137 f 10 pmol CAMP'rnin-' 'mg protein-'

higher affinity for VIP and VIP analogues than AR 4-25 and hippocampus membranes. All peptides stimulated adenylate cyclase and their order of affinity was in good agreement with binding data (Figs 1 to 3). The ability of PACAP(1-27) and PACAP to stimulate adenylate cyclase activity in AR 4-2J and NB-OK cell membranes (Kact)was 3 - 10-fold higher than their receptor binding affinity (Kd).This was also the case for PACAP(2-38) in NB-OK cell membranes. K,,, and Kd values were broadly comparable for all VIP analogues and for PACAP(2-27), although these parameters were not easily determined, due to the high peptide concentration required for total receptor binding and adenylate cyclase activation. K,,, and Kd values were almost identical for hippocampus membranes.

for adenylate cyclase activation, but dose/response curves were about 10-fold less sensitive than concentration/binding curves. PACAP and VIP-PACAP acted as partial agonist with an intrinsic activity of 0.75 and 0.70, respectively, considering the maximal response to VIP as 1.0. The intrinsic activity of PACAP(2 - 7) and PACAP(2 - 38) could not be determined with accuracy. DISCUSSION

PACAP and PACAP(1- 27) are newly discovered members of the VIP family, originally extracted from ovine hypothalamus. These neuropeptides are most probably widely distributed [2]. Despite strong similarity to VIP (Table l), PACAP, in its two forms, is not a mere VIP variant. It interacts with specific receptors that display a 500 - 2000-fold greater Interaction of PACAP(I-27), PACAP, PACAP fragments, specificity for PACAP compared to VIP, the ligand showing VIP-binding sites VIP and VIP analogues with the next highest affinity for 1251-PACAP(1- 27)-binding sites, in rat pancreatic membranes among parent peptides. '251-PACAP(1- 27)-binding sites are All peptides tested were able to occupy rat pancreatic abundant in all areas of the rat central nervous system (Cauvin '251-VIP-bindingsites labelled with 1251-VIPand to stimulate et al., unpublished results), in the human neuroblastoma cell adenylate cyclase activity (Fig. 4 and Table 2). The order line NB-OK [6], in pituitary cells [l, 31, and in the acinar AR of affinity for 1251-VIP-binding-site occupancy was 4-25 cell line derived from rat pancreas [5]. There is also a PACAP = PACAP-27 3 VIP = [Ser9]VIP = [Tyrl3]VIP = small population of specific 1251-PACAP(1- 27)-binding sites [Ser9,Tyrl3]VIP 3 PACAP-VIP 3 VIP-PACAP(2 - 38) = in rat liver membranes, coexisting with a majority of highPACAP(2-27). These values for PACAP, VIP and their affinity and low-affinity VIP-binding receptors. In the latter analogues were relatively close, suggesting that 1251-VIP-bind- case, specific hepatic IZ5I-PACAP-binding sites are difficult ing sites were virtually unable to discriminate between VIP to characterize as PACAP also binds with relatively high and PACAP. The same order of binding affinity was observed affinity to both subclasses of VIP-binding receptors, with

540 RAT P A N C R E A T I C MEMBRANES

A

C

VIP, PACAP

\f

-11 -10 -9

-8

-7

-6 -10 -9 l o g [Pept idel /M

-8

-7

vu :o

-6

Fig. 4. Effects of incrcwsing concentrations of peptides on 1251-VIP-binding( A , C ) and adenylate cyclase stimulation ( B ,0 ) in rat pancreatic membranes. The symbols and representation are as in Figs 1 - 3, except that adenylate cyclase stimulation was expressed as a percentage of the value observed in the presence of 0.1 pM VIP. Basal and 1 FM-VIP stimulated adenylate cyclase activities were 36 f 2 pmol CAMP.min . mg protein-' and 548 i 20 pmol CAMP.rnin-l mg protein-' (mean 5 SEM, n = 3). Experiments were performed in duplicate

'

PACAP(1-27) being even more active than VIP itself for binding to '251-VIP-binding sites (Robberecht et al., unpublished data). In the present study, a similar interaction with 1251-VIP-bindingsites was also observed in rat pancreatic membranes. Considering that PACAP reproduces the vasodilation properties of VIP in vivo, it is tempting to conclude that both PACAP(1-27) and PACAP are likely to occupy all '2sI-VTP-binding sites with high affinity and, in addition, specific "'I-PACAP( 1 - 27)-binding sites capable of mediating adenylate cyclase activation [l, 51. The purpose of the present work was to define some of the PACAP amino acids that are of importance for optimal recognition of 12sI-PACAP(1- 27)-binding sites and binding specificity. We synthesized seven peptides: PACAP(2 - 27) and PACAP(2 - 38), to evaluate the importance of the N-terminal histidine ; two chimeric molecules, called VIPPACAP and PACAP-VIP (see Table 1) to estimate the relative contribution of the C-terminus and N-terminus of the ligand to receptor specificity; three VIP analogues substituted by the corresponding amino acids of PACAP in position 9 and/or 13. We thought that amino acids in these positions were probably more important for receptor recognition than those in position 4, 5 or I1 for two reasons. Glycine at position 4 is also present in peptide histidine isoleucinamide, and isoleucine in position 5 is also present in helodermin; these two peptides have no more affinity than VIP for the '2sI-PACAP(1-27)-binding

site. Serine and threonine in position 11 were assumed to be almost equivalent. The absence of histidine in position 1 in PACAP(2-27) and PACAP(2-38) decreased the affinity for both "'1PACAP(1- 27) and 12sI-VIP-binding sites. PACAP(2 - 27) had 70 - 80 times less affinity than PACAP(1- 27) for NBOK and AR 4-23 cell membranes, and 270 times less affinity for rat hippocampus membranes, suggesting that human neuroblastoma and rat peripheral '2sI-PACAP(1 - 27)-binding sites could differ somewhat from central nervous 12'I-PACAP(1 - 27)-binding sites. In the three systems, PACAP(2-38) was only 10- 13 times less active than PACAP, suggesting that the amino-acid-27 - 38 extension of PACAP contributed to receptor recognition when ligand affinity was reduced, due to the absence of His 1. On rat pancreatic '251-VIP-binding sites PACAP(2 - 27) and PACAP(2-38) both had 40-fold less affinity than the corresponding complete ligands. This is in line with the fact that VIP(2 -28) has 80 and 70 times less affinity than VIP(128) for rat and in human intestinal membranes, respectively [13]. The necessity, for optimal secretin and glucagon receptor recognition, of a histidine residue in position 1 in secretin [14] and glucagon [15] (or its replacement by phenylalanine [16, 17]), has also been documented. On the basis of adenylate cyclase activation, PACAP(2 38) was a full agonist with AR 4-2J membranes and rat hippo-

541 campus membranes, and PACAP(2 - 27) a full agonist with AR 4-2J and NB-OK membranes. Furthermore, PACAP(2 27) and PACAP(2 - 38) had 800 - 1000-fold and 60 - 80-fold less affinity than PACAP(1- 27) and PACAP, respectively, for activating adenylate cyclase in the three systems considered, the K,,, of the two fragments being very similar to their Kd values in binding studies. Taken together, these results suggest that PACAP(2 - 27) and PACAP(2 - 38) acted as partial agonists capable of eliciting a maximal response in highly coupled systems, whereas enzyme activation through the full agonists PACAP(1- 27) and PACAP required only partial receptor binding. The three C-terminal amino acids of PACAP(1-27) were not a major requirement for '251-PACAP(1- 27)-binding. Indeed, the chimeric VIP-PACAP molecule we synthesized (Table 1) possessed 22 amino acids in common with PACAP and differed from PACAP in positions 4, 5,9,11 and 13. VIPPACAP had only 1.5 - 3 times greater affinity than VIP for 1251-PACAP(1- 27)-binding sites and therefore still had 300 - 600 times lower affinity than PACAP(1- 27). Conversely, VIP-PACAP had three times less affinity than VIP for the lZ5I-VIP-bindingsites. Affinity recognition of 251-PACAP(1 -27)-binding sites depended much more on the N-terminal moiety of the molecule, with a probable contribution of each amino acid in this area. Replacement of Asp9 or Leu13 in VIP by serine or tyrosine, respectively, increased 2 - 5-times the affinity of VIP for lZ5I-PACAP(1- 27)-binding sites. A simultaneous change at both positions increased 10-fold the affinity of the peptide but [Ser9,Tyrl3lVIP still had 40 - 200 times less affinity than PACAP. The chimeric molecule PACAP-VIP had 20 - 50times higher affinity than VIP but had 20- 100-times less activity than PACAP itself. The whole PACAP(1-27) molecule was thus necessary for optimal receptor binding. In brief, lZ5I-PACAP(1-27)-binding sites in three membrane preparations from rat hippocampus, human neuroblastoma NB-OK, and rat pancreatic acinar cell line AR 4-2J discriminate very efficiently between peptide analogues, contrasting with rat pancreatic '251-VIP-binding sites which accommodate, with relatively high affinity, ligands as diverse as VIP, PACAP, VIP analogues modified in positions 1-6 [lo, 16, 171, peptide histidine isoleucinamide [18, 191, helodermin [20, 211, growth-hormone-releasing factor derivatives [lo, 22, 231 and secretin analogues [23]. Aided by grant 5-ROI-DK 17010-13from the National Institutes of Health (Bethesda, MD, USA), grant 3.4504.85 from the Fund for Medical Scientific Research (Belgium) and a Concerted Research Action grant from the Ministry of Scientific Policy (Belgium).

REFERENCES 1. Miyata, A., Arimura, A , , Dahl, R. R., Minamino, N., Uehara, A,, Jiang, L., Culler, M. D. & Coy, D. H. (1989) Biochem. Biophys. Res. Commun. 164, 567 - 574.

2. Kimura, C., Ohkubo, S., Ogi, K., Hosoya, M., Itoh, Y., Onda, H., Miyata, A., Jiang, L., Dahl, R. R., Stibbs, H. H., Arimura, A. & Fujino, M. (1990) Biochem. Biophvs. Res. Commun. 166, 81 - 89. 3. Gottschall, P. E., Tatsuno, I., Miyata, A. & Arimura, A. (1990) Endocrinology 127,272-277. 4. Tatsuno, I., Gottschall, P. E., Koves, K. & Arimura, A. (1990) Biochem. Biophys. Res. Commun. 168, 1027- 1033. 5. Buscail, L., Gourlet, P., Cauvin, A,, De Neef, P., Gossen, D., Arimura, A., Miyata, A., Coy, D. H., Robberecht, P. & Christophe, J. (1990) FEBS Lett. 262, 77-81. 6. Cauvin, A,, Buscail, L., Gourlet, P., De Neef, P., Gossen, D., Arimura, A,, Miyata, A,, Coy, D. H., Robberecht, P. & Christophe, J. (1990) Peptides 11, 773 -777. 7. Svoboda, M., Robberecht, P., Camus, J., Deschodt-Lanckman, M. & Christophe, J. (1976) Eur. J. Biochem. 69, 185-193. 8. Robberecht, P., Waelbroeck, M., Camus, J. C., De Neef, P. & Christophe, J. (1984) Biochim. Biophys. Acta 773, 271 -278. 9. Laburthe, M., Bataille, D. & Rosselin, G. (1977) Acta Endocrinol. 84, 588 - 599. 10. Robberecht, P., Coy, D. H., De Neef, P., Camus, J., Cauvin, A., Waelbroeck, M. & Christophe, J. (1986) Eur. J . Biochem. 159, 45 - 49. 11. Salomon, Y., Londos, C. & Rodbell, M. (1974) Anal. Biochem. 58, 541 -548. 12. Fournier, A,, Danho, W. & Felix, A. M. (1989) Int. J . Pept. Protein Res. 33, 133- 142. 13. Couvineau, A,, Rouyer-Fessard, C., Fournier, A., St Pierre, S., Pipkorn, R. & Laburthe, A. (1984) Biochem. Biophys. Res. Commun. 121,493-498. 14. Robberecht, P., De Neef, P., Waelbroeck, M., Camus, J. C., Scemama, J.-L., Fourmy, D., Pradayrol, L., Vaysse, N. & Christophe, J. (1988) Pancreas 3, 529-535. 15. Lin, M. C., Whright, D. E., Hruby, V. J. & Rodbell, M. (1975) Biochemistry 14, 1559- 1563. 16. Robberecht, P., Waelbroeck, M., Camus, J. C., De Neef, P., Coy, D. H. & Christophe, J. (1984) Peptides5, 877-881. 17. Robberecht, P., De Neef, P., Gourlet, P., Cauvin, A., Coy, D. H. & Christophe, J. (1989) Regul. Pept. 26, 117-126. 18. Bataille, D., Gespach, C., Laburthe, M., Amiranoff, B., Tatemoto, K., Vauclin, N., Mutt, V. & Rosselin, G. (1980) FEBS Lett. 114,240- 242. 19. Robberecht, P., Tatemoto, K., Chatelain, P., Waelbroeck, M., Delhaye, M., Taton, G., De Neef, P., Camus, J. C., Heuse, D. & Christophe, J. (1982) Regul. Pept. 4 , 241 -250. 20. Robberecht, P., Waelbroeck, M., De Neef, P., Camus, J. C., Vandermeers, A,, Vandermeers-Piret, M.-C. & Christophe, J. (1984) FEBS Lett. 172, 55-58. 21. Zhou, Z., Gardner, J. D. & Jensen, R. T. (1989) Am. J. Physiol. 256, G283 - G290. 22. Laburthe, M., Amiranoff, G., Boige, N., Rouyer-Fessard, C., Tatemoto, K. & Moroder, L. (1983) FEBS Lett. 159, 89-92. 23. Robberecht, P., Waelbroeck, M., Coy, D. H., De Neef, P., Camus, J.-C. & Christophe, J. (1986) Peptides 7, Suppl. 1, 53-59. 24. Robberecht, P., Waelbroeck, M., Noyer, M., Chatelain, P., De Neef, P., Konig, W. & Christophe, J. (1982) Digestion 23,201 210.