Transcript

Eur. J. Biochem. 174,219-224 (1988) 0 FEBS 1988

Characterization of covalently cross-linked somatostatin receptors in hamster beta cell insulinoma Patrizia COTRONEO Jean-Claude MARIE' and Gabriel ROSSELIN

Instituto di Clinica Medica, Universita Cattolica, Roma Unit6 de Recherches sur les Peptides Neurodigestifs et le Diabete, Unite 55, Institut National de la Santk et de la Recherche Medicale, Paris

(Received November 10, 1987/January 22,1988) - EJB 87 1284

The selective binding of somatostatin-28 (SS-28) to beta cells of hamster insulinoma was characterized using HPLC-purified '251-[Le~8,~-Trp22,Tyr25]SS-28 or lZ5I-SS-28. A single class of high-affinity sites (Kd = 53 f 5 pM) was observed with a binding capacity of 2.85 pmol/mg membrane protein. A large number of relatively low-affinity sites was found also. The order of potency of different peptides to inhibit '251-SS-28 binding is SS- 28 > SS-14 > SMS-201-995 and the respective half-maximal inhibitory doses are 0.16 nM, 10 nM and 1000 nM. CCK8 and other active pancreatic peptides (glucagon, insulin, gastric inhibitory peptide, vasoactive intestinal peptide, oxyntomodulin) do not inhibit the SS-28 receptor binding. '251-SS-28-labeled beta membranes were successfully cross-linked using either the cleavable cross-linker dithiobis(succinimidy1propionate) (1 mM) alone or with a heterobifunctional agent, N-hydroxysuccinimidyl-4-azidobenzoate (HSAB). In both cases five molecular components were revealed, after polyacrylamide gel electrophoresis of the membrane proteins and autoradiography, with the following molecular mass: 196-kDa, 132 kDa, 69 kDa, 45 kDa and 28 kDa. The labeling of 196-kDa, 132-kDa and 45-kDa species was specific in that they could be inhibited by unlabeled SS-28. The major labeled species corresponds to the 132-kDa band and no change in the mobility of this HSAB covalently bound SS-28 receptor was found after addition of dithiothreitol, suggesting that this specific receptor does not contain interchain disulphide bonds. The molecular mass of SS-28 receptors differs markedly from that of guinea- pig pancreatic acinar membranes, where a single 93-kDa protein is identified as a '251-SS-28 receptor site in comparative experiments. Both the binding kinetics and structural differences sustain the selective action of SS- 28 in the endocrine pancreas.

Somatostatin (SS), or somatotropic release inhibiting factor, was initially isolated from the ovine hypothalamus [l]. Somatostatin-28 [2] is mostly abundant in gut epithelium whereas somatostatin-14 [3] is mainly present in D cells of islets of Langerhans [3] and appears early in the developmental mammals [4]. Both inhibit, at different concentrations, the release of insulin and glucagon in vitro and in vivo [5, 61. The mechanism by which the somatostatin initiates the cellular response is the binding of somatostatin to specific receptors as shown in different tissues (reviewed in [7]). The somatostatin receptors have also been found in isolated rat islets of Langerhans [XI and in membrane of hamster insulinoma [9]. This insulinoma represents an invaluable model for somatostatin receptor studies, since it is composed of mainly beta cells [lo] and an array of other regulatory receptors (glucagon, oxyntomodulin, gastric inhibitory peptide) related to the beta cell function [ll - 131 (reviewed in [14]). Further- more, in this tumor the glucose-induced insulin release is inhibited by SS-14 [15] suggesting the presence of functional somatostatin receptors in these beta cells. Whereas the under- standing of the mechanism of the somatostatin action requires the characterization of authentic binding components of the

Correspondence to J.-C. Marie, Unitit 55 de I'INSERM, 184 Rue du Faubourg-Saint-Antoine, F-75573 Paris Cedex 12, France

Abbreviations. SS, somatostatin; DSP, dithiobis(succinimidy1- propionate); HSAB, N-hydroxysuccinimidyl-4-azidobenzoate; VIP, vasoactive intestinal peptide; GIP, gastric inhibitory pepide; IC50, half-maximal inhibition dose.

receptors, there is as yet no data on the molecular characterization of beta cell somatostatin receptors. In order to substantiate further the first step of somatostatin action and its paracrine effect in the endocrine pancreas it is imperative to characterize the SS-28 receptor carefully and to study the structure of the receptor proteins. Therefore, we have for the first time: (a) characterized the SS-28 receptor in insulinoma cells by binding experiments using HPLC-purified ' 251-SS-28; (b) assessed the selective binding of different somatostatin analogs to the membrane of hamster insulinoma and its specifity in presence of unrelated peptides ; (c) characterized the molecular components of somatostatin receptors by affin- ity labeling with 1251-SS-28 utilizing the heterobifunctional and homobifunctional cross-linking agents: dithiobis(suc- cinimidylpropionate) (DSP) and N-hydroxysuccinimidy1-4- azidobenzoate (HSAB) and by SDS gel electrophoresis analy- sis; (d) compared the labeling by '251-SS-28 of plasma mem- branes from beta cells and pancreatic acinar cells.

MATERIALS A N D METHODS Chemicals

Somatostatin (SS-28) and [Leu8,~-TrpZ2,Tyrz5]SS-28 were purchased from Peninsula Laboratories Inc. (Belmont, Calif., USA) SS-14 was provided by Dr J. Diaz (Sanofi Recherche, Montpellier, France) and SMS-201- 995 (an octapeptide analog of somatostatin-14) from Sandoz (Easel, Switzerland),

220

oxyntomodulin was generously given by D. Bataille (CCIPE, Montpellier, France). Bifunctional cross-linking reagents were obtained from Pierce Eurochimie (Rotterdam, Holland) acrylamide, bisacrylamide from Serva (Heidelberg, FRG), SDS from Sigma (St Louis, Mo., USA) and molecular mass markers from Bio-Rad (Richmond, Calif., USA). Both CF,COOH and CH3CN, purchased from Merck (Darmstadt, FRG), were of spectroscopic grade. All the other chemicals were of the highest available grade.

Preparation of pancreatic membranes

The insulin-secreting hamster pancreatic insulinoma was serially subcutaneously transplanted in Syrian hamsters. The animals were killed 60 days after the grafting and the enriched membrane fraction was prepared as described [13]. Briefly, the tumors were dissected, homogenized, using a Waring blender, in a freshly prepared 20 mM Hepes buffer (pH 7.5) containing 0.25 M sucrose, 0.01 M triethanolamine, 5 mM EDTA, 0.1 M phenylmethylsulfonyl fluoride and bacitracin (1 mg/ml) and centrifuged at 2000 x g for 20 min, after which the supernatant was centrifuged at 20000 x g for 20 min and the pellet resuspended and washed twice in 20mM Hepes buffer (pH 7.5). The membranes were collected and stored at -80°C until use. The membrane protein content was mea- sured by the method of Bradford [16]. Pancreatic acini mem- branes from male Hartley guinea-pig pancreas were gener- ously provided by C. Susini (INSERM U 151, Toulouse, France) [17].

Iodination andpurijication of labeled peptides

Iodination of pure somatostatin analog [LeU8,D-Trp22,- T ~ r ~ ~ l S S - 2 8 was performed by the chloramine T method [18] : 5-10 pg SS-28 analog was iodinated with 1 mCi NalZ5I and 5 pl chloramine T (1 mgiml). The reaction was stopped after 30 s by adding 120 pl tyrosine (2 mg/ml) [19]. We have purified the '251-SS-28 as follows: the iodinated material was diluted with 1 ml 10% acetonitrile in 0.1% trifluoroacetic acid and immediately applied to a SEP-PAK C18 cartridge packed with octadecylsilica support (Waters, Milford, Mass., USA), which was previously equilibrated with 10% acetonitrile in 0.1 YO CF3COOH [20]. The cartridge was washed in the same buffer to remove the free iodine and iodinated tyrosine. The labeled peptide was eluted with 60% acetonitrile in 0.1 YO CF,COOH and purified by reverse-phase HPLC using a p-Bondapak C1 8 column packed with a 10-pm-particle-size octadecylsilica support in a stainless-steel column (0.39 x 30 cm) from Waters (USA) as shown in Results.

Binding of '25-Z-SS-28 analog Binding of 1251-SS-28 analog to pancreatic plasma mem-

brane was carried out in 10 mM Hepes buffer (pH 7.5) con- taining 1% bovine serum albumin and 1 mg/ml bacitracin (incubation buffer). Plasma membranes (50 pg protein/ml) were incubated with '251-[Le~8,~-Trp22,Tyr25]SS-28 analog (0.01 nM) in a total volume of 500 pl, in the presence or absence of unlabeled somatostatin-28. The incubation was routinely performed at 20°C for 60 min and stopped by adding ice-cold 10 mM Hepes buffer (pH 7.5). Membrane- bound labeled somatostatin was separated by centrifugation at 40000 x g for 30 min and washed twice with the incubation buffer and once with 10 mM Hepes 10% sucrose. The radioac- tivity of the pellet was determinated by a gamma counter.

In all the experiments the non-specific binding of labeled somatostatin was evaluated as the radioactivity that was not displaced by an excess of unlabeled SS-28 (100 nM). The specifity of the '251-SS-28 binding to pancreatic plasma mem- brane of hamster insulinoma was studied in the presence of somatostatin analogs and unrelated peptides, as indicated in the corresponding figure. The degradation of ' 251-SS-28 analog during the incubation was evaluated by measuring the ability of labeled somatostatin to rebind to specific receptors present in fresh plasma membrane obtained from the same insulinoma.

Cross-linking

The pancreatic plasma membranes (300 pg protein/ml) were incubated with 0.01 nM '251-SS-28 analog and centri- fuged as described above. The washed pellet was resuspended in 1 ml 60 mM Hepes buffer (pH 7.5) for cross-linking by DSP and in 1 ml50 mM phosphate buffer (pH 7.5) for HSAB cross-linking. The cross-linking reagent DSP was dissolved in dimethylsulfoxide and added immediately at a final concen- tration of 1 mM. After 20 rnin at 4°C the reaction was stopped by adding 500 pl 60 mM Hepes buffer (pH 7.5) containing 60 mM CH3COONH4. For the HSAB cross-linking 10 p1 1 mM photoreactive cross-linker (final concentration) in 2% dimethylsulfoxide was added to the membrane suspension. After exposure to a 275-W sun-lamp at a distance of 12 cm for 15 min in ice, the reaction was stopped by adding 500 p1 50 mM phosphate buffer, pH 7.5. In all cases the treated mem- branes were harvested by centrifugation at 40000 x g for 30 min. Pancreatic acinar membranes (100 pg protein/ml) were incubated in 10 mM Hepes buffer containing 0.1 % bo- vine serum albumin, 0.2 M CaC12, 0.02% soybean trypsin inhibitor, 1 mM benzamidine and bacitracin (0.5 mg/ml) at 20°C for 60min. The resulting membrane pellet was covalently labeled as described above.

SDSlpolyacry lamide gel electrophoresis

Affinity-labeled membrane was solubilized with 60 mM Tris/HCl buffer (pH 6.8), containing 0.001 YO (mass/vol.) bromophenol blue, 3% (mass/vol.) SDS, 10% (mass/vol.) glycerol in either the presence or absence of 100 mM dithio- threitol, and heated at 60°C for 30 min. Electrophoresis was performed according to Laemmli [21] using 1.5-mm slab gels of 7 - 15% acrylamide. After electrophoresis the gel was fixed and stained with Coomassie brillant blue. Autoradiograms were obtained from the dried gel after exposure to Kodak X-Omat AR film with enhancing screens at - 80 "C for 4 - 7 days.

RESULTS

Preparation and purification of iodinated SS-28

After reverse-phase HPLC purification of the iodination mixture, one major form of '251-SS-28 was isolated with a retention time of 27 min; it was well separated from unlabeled somatostatin-28, which has a retention time of 15 rnin (Fig. 1). We have previously shown that the iodination of VIP or GIP leads to a more hydrophobic compound than the native peptide [13,20]. This is in agreement with the observation that halides generally increase the hydrophobic nature of mol- ecules [22]. Free iodine and iodinated tyrosine have been elim- inated by the prepurification step with a SEP-PAKCI8 car-

221

1, ,

I 20 40 6 T I M E ( m i d

0.1

i 1.05 2

rn a

0

Fig. 1. Purification ~ f ' ~ ~ I - [ L e u ' , ~ - T r p ~ ~ , T y r ~ ~ ] S S - 2 8 by HPLC. The iodination mixture eluted from the SEP-PAK C18 cartridge (Waters) (see Materials and Methods) was subjected to a p-Bondapak CIS column (Waters) and eluted with 23% acetonitrile in 0.1% trifluoroacetic acid at a flow rate of 1 ml/min. The radioactivity was continuously monitored (- - -) by a Berthold rate meter (model LB 2040) and 0.5-ml fractions were collected. The elution position of unlabeled [Le~~,~-Trp~~,Tyr~~]SS-28 (-) was determined by absorbance using a Waters spectrophotometer (model 440). The small first peak indicated the void volume

tridge. The mean specific activity of the 1251-SS-28 analog was calculated as described [23] and was 11.8 TBq/mol. This corresponds to a mean incorporation of 0.15 lZ5I atom/mol- ecule SS-28, which minimizes the production of diiodinated derivatives. Assuming that the major peak of lZ5I-SS-28 rep- resented monoiodinated somatostatin the specific activity was 74 PBq/mol. This pure iodinated somatostatin-28 was routinely used in the following experiments.

Specificity of SS-28 binding to beta cell hamster membrane

Optimal somatostatin binding was obtained after 1 h of incubation at 20°C, in agreement with data of Reubi et al. [9]. Total binding represented as much as 79 4% and non- specific binding was as low as 6 & 1% of total radioactivity in the presence of 100 nM unlabeled SS-28. In our incubating conditions the degradation of labeled somatostatin-28 did not account for more than 10% of the total radioactivity. The competitive inhibition of iodinated SS-28 binding to pancre- atic plasma membrane by native SS-28, SS-14 and SMS-201- 995 showed that SS-28 was more potent than SS-14 and SMS- 201 - 995 in displacing the labeled somatostatin (Fig. 2) with a half-maximal inhibition occurring at concentrations of 0.16 nM, 10 nM and 1000 nM respectively. Furthermore the Scatchard analysis showed a population of high-affinity bind- ing sites with a Kd of 53 +_ 5 pM and a binding capacity of 2.58 2 0.5 pmol SS-28/mg protein (Fig. 2, inset), and a class of binding sites with low affinity with a Kd of 18 f 5 nM and a binding capacity of 26 f 3.6 pmol/mg (Table 1). A variety of unrelated polypeptides (glucagon, VIP, insulin, gastrin, secretin, GIP, oxyntomodulin and CCK,) were tested at con- centrations as high as 1 pM for their ability to inhibit the specific binding of 1251-SS-28. All these eight peptides failed to interact with the membrane-bound labeled somatostatin (Fig. 2). It can be noted that cholecystokinin has been shown to reduce 1251-SS-14 binding to pancreatic acinar cell plasma membranes [24], which in our system we did not observe.

oL -L I iolBlpM ioo I i io , -\ 611 1 1 1 0 9 8 7 6 -LOG [MI

Fig. 2. Competitive displacement of '251-[Leu8, D-Trp2',Tyr2 5]SS-28 binding to beta membrane by peptides. Membranes were incubated with 1251-SS-28 (0.01 nM) in the presence of increasing concentrations of peptides: native SS-28 (O) , SS-14 (m), SMS-201-995 ( A ) and unrelated peptides (VIP, GIP, glucagon, oxyntomodulin, insulin, se- cretin, gastrin, CCKs) ( x ). Specific binding was measured and has been plotted against log (peptide concentration). A Scatchard analysis of the data is shown in the inset. Results show the means of four independent experiments with duplicate determinations of each (Kd = 53 & 5 pM, Number of sites = 2.85 0.5 pmol/mg). For clarity

standard errors are not indicated; they were below 15% of the mean values

Table 1 . Characteristics of binding sites of SS-28 in membranes from beta cells Values reported were obtained by Scatchard analysis and represent the mean SEM of four separate experiments

Binding sites Kd No. of sites

M pmol/mg membrane protein

High affinity 5.3 f 0.5 x lo-'' 2.85 k 0.5 Low affinity 1.8 f 0.5 x lo-' 26.5 k 3.6

Also, in the presence of 0.2 mM CaCl, the maximal amount lZ5I-SS-28 bound of to beta cell membranes was unchanged (total binding = 69 f 6% and non-specific binding 4 f 0.7%) whereas it was observed that CaClz modified the somatostatin binding in pancreatic acini [I71 and in brain [25] membranes.

Cross-linking When 1251-SS-28-labeled pancreatic plasma membrane of

hamster insulinoma was cross-linked with 1 mM DSP and analyzed by polyacrylamide gel electrophoresis, the auto- radiography revealed one major band of 132 * 0.7 kDa and four minor bands of 196+ 1.1 kDa; 6 9 2 0 . 6 kDa; 45 f 0.5 kDa and 29 & 0.8 kDa (Fig. 3, lane A). Also, two bands with molecular masses close to 45 kDa are present and could be due to post-translational modification. These values are a mean of eleven experiments and each value was obtained from linear regression plots between log (molecular mass) and the mobility of the following protein markers: myosin, P-galactosidase, phosphorylase b, bovine serum albumin,

222

Fig. 3. SDS-PAGE autoradiogram of lZ5I-SS-28 cross-linked to pancreatic beta cell membranes; effect of increasing concentrations of unlabeled SS-28. Beta cell membranes were incubated with 1251-SS-28 and increasing concentrations of SS-28, then treated with 1 mM DSP. After the reaction, unbound radiolabel was removed by centrifugation and the pelleted membranes were dissolved in SDS sample buffer prior to their separation by SDS-PAGE, as described in Materials and Methods. On the left, lane A shows the autoradiographic pattern of membranes incubated with lZ5I-SS-28 alone; lane B-H with 0.1 nM, 0.5 nM, 1 nM, 5 nM, 10 nM, 50 nM, 0.1 pM SS-28 respectively. Molecular mass values of the radioactive bands (28 - 196) are indicated in kDa. Quantitative measurement of the corresponding bands are given on the right by the histogram showing relative absorbance. The quantification was performed with a Biocom program and the relative absorbance (0-0.5) was calculated relative to the maximal absorbance obtained with the band of 132 kDa, lane B. The bands (196 kDa, 132 kDa and 45 kDa) diminish in a concentration-dependent rnanneI

ovalbumin, carbonic anhydrase, soybean trypsin inhibitor and lysozyme. The specifity of the '251-SS-28 labeling of somatostatin receptors was evaluated in the presence of differ- ent concentrations of unlabeled SS-28 (0.1 - 100 nM), (Fig. 3, lanes B - H). In these conditions the components of 196 kDa, 132 kDa and 45 kDa disappeared in a dose-dependent man- ner, indicating that these bands represent specific components of the receptor. However, the labeled components of 69 kDa and 28 kDa persisted in the presence of high concentrations of unlabeled peptide (100 nM) even though the intensity of the bound peptide (28 kDa) is much less intense in lanes F - H than in lanes A and B. These bands are likely to represent non-specific labeled proteins. These observations are further sustained by the quantitative measurement of the intensity of the corresponding autoradiogram bands (Fig. 3, right). The major labeled band (1 32 kDa) is highly sensitive to unlabeled somatostatin, It can also be noted that in the presence of 100 pM unlabeled somatostatin (lane B) the labeling is quite considerably stronger than that on lane A. This also occurs with the 196-kDa species but not with the other specifically labeled band (45 kDa). Thus, it is difficult to ascertain whether there is a cooperative effect.

We have further investigated whether the proteolytic degradation or protein aggregation during the experiment could account for the heterogeneity of three components which specifically bound the '251-SS-28. It is known that several proteases are not inhibited by using bacitracin, even at 1 mg/ml. The same labeling was observed when the binding was performed in the presence of phenylmethylsulfonyl fluo- ride, a potent inhibitor of proteolysis, which was added at the beginning, or when the solubilizdtion of labeled membranes in SDS buffer was perfomed at 100°C, 60°C and 37°C. Thus, protein aggregation or proteolytic degradation is unlikely to account for the three specific receptor components. Also, the affinity labeling with another cross-linker (HSAB) revealed the same five components (Fig. 4, lanes A-C) as with the DSP. We noticed that, whereas the 196-kDa band was regu- larly labeled when using DSP, it was not always observed with

Fig. 4. Covalent labeling of lZ5I-SS-28 beta pancreatic membranes by DSP or HSAB. Pancreatic beta membranes were incubated with '"1- SS-28 and the bound pellet was cross-linked either with DSP (lanes A, B) or HSAB (lanes C, D) as described in Materials and Methods. The arrows indicate the molecular mass of the three specific labeled somatostatin receptor components identified previously in Fig. 3. The same labeling is observed with either DSP (lane A) or HSAB (lane C) cross-linking experiments. When solubilization was performed in the presence of 0.1 M dithiothreitol (a) the disulfide bond present in the DSP cross-linker is readily reduced (lane B) and (b) the labeling with the non-reducible cross-linker HSAB shows that the labeled component (132 kDa) is unaffected by the active reductant (lane D)

HSAB. This might be attributed to the different spacing of the chemically reactive groups of the cross-linker. The solubilization of cross-linked samples by either DSP or HSAB with or without dithiothreitol was performed in comparative experiments. While in the presence of dithiothreitol the ob-

223

Fig. 5. Comparative covalent labeling of '251-SS-28 to beta cells or acinar cellpancreatic membranes. Pancreatic beta cells (300 pg/ml) or acini membranes (100 Fg/ml) were incubated with lZ5I-SS-28 at 20°C for 60 min in the presence or absence of unlabeled SS-28 (100 nM). The total binding and the non-specific binding with acini membrane were 43 & 2% and 12 6%, respectively, and those observed with the beta membranes correspond to 79 4% and 6 & 1 YO. After cross- linking with DSP, SDS-PAGE analysis was performed as described in Materials and Methods. The autoradiogram shows a single band of 93 kDa (lane A) in the absence of unlabeled SS-28, while in the case of pancreatic beta cell membrane we obtained three bands of 196 kDa, 132 kDa and 45 kDa (lane B). Thus substantial differences are observed in the labeling of these two membranes

served DSP labeling was greatly reduced by cleavage of the disulfide bond of the cross-linker (Fig. 4, lane B), the HSAB- labeled component of 132 kDa was unaffected by this re- ductant (Fig. 4, lane D). However, the use of the reducing agent decreases the intensity of the 45-kDa band and clearly increases the intensities of the 28-kDa and the 69-kDa bands. This may be accounted for by the presence of dithiothreitol, which may alter the facility of certain labeled species to enter the gel. The possibility cannot be excluded that internal di- sulfide bonds, which are known to generate a more compact structure, are present and their reduction by dithiothreitol may lead to these increased diffuse bands in agreement with observation by others [26]. This shows that in the presence of an active reductant the specific labeling by DSP is confirmed, and with HSAB (a non-reducible cross-linker) the major labeled species (132 kDa) do not contain any interchain di- sulfide bonds.

After binding of 1251-SS-28 to pancreatic acinar guinea- pig membrane in the presence or absence of unlabeled SS-28 (100 nM), cross-linking with DSP was performed and fol- lowed by SDS-PAGE analysis. The autoradiogram showed only one specific band of 93 kDa (Fig. 5, lane A), which disappeared in the presence of 100 nM unlabeled 58-28, in agreement with Susini [27]. The migration of the beta cell SS-28 receptor components is clearly different from that ob- served with acinar membrane (Fig. 5, lane B).

DISCUSSION The present results yield essential information on the struc-

ture of somatostatin receptors in the pancreatic beta cell. We

have clearly demonstrated that these somatostatin receptors (a) bind more specifically to SS-28 than to SS-14 or SMS- 201 - 995 with the respective half-maximal inhibitory dose (Ic50) = 0.16 nM, 10 nM and 1000 nM; (b) do not display cross-reactivity with biologically active peptides in the pan- creas (VIP, GIP, glucagon, oxyntomodulin, CCKs, insulin); (c) can be identified as three specific somatostatin protein complexes of 196 kDa, 132 kDa and 45 kDa, with a major species of 132 kDa. This insulinoma expresses a population of high-affinity sites for somatostatin-28 in agreement with Reubi et al. [9]. The difference observed in Kd values and the number of sites may be attributed to the use of HPLC-purified '251-SS-28 or different experimental conditions. However, the I C ~ O value of SS-28 in inhibiting '251-SS-28 binding to the insulinoma (0.16 nM) compares well with the half-maximal inhibition of insulin release from purified beta cells by 0.1 nM SS-28 [28]. The property of these high-affinity sites for dis- criminating between the different analogs of somatostatin and unrelated peptides substantiates the selective action of the SS-28 on insulin release [5, 61. The concentrations of SS-28 necessary for demonstrating the low-affinity sites are unlikely to preclude any role of this site in the physiological action of somatostatin. The relevance of low-affinity sites remains unclear. Indeed, our results and the demonstration that SS-28 and SS-14 display different binding affinity with somatostatin receptors in pituitary [29], rat pancreatic acinar cells [30] and guinea-pig acinar pancreatic membranes [I 71 further docu- ment the need to understand the receptor interaction at the molecular level. The molecular analysis revealed three repro- ducible and specific '251-protein complexes with distinct mo- lecular masses. If we assume that one molecule of somato- statin is bound per receptor the apparent molecular masses of these complexes are 193 kDa, 129 kDa and 42 kDa. The identification of authentic specific somatostatin receptors is coherently demonstrated by the following data: (a) the dis- appearance of these labeled proteins in the presence of increas- ing concentrations of unlabeled somatostatin-28 (0.1 - 100 nM) fits well with the biochemical dose-response; (b) cross-linking with another agent (HSAB, a heterobifunctional reagent) resulted in the identification of the same specific complexes. The molecular heterogeneity of the 1251-SS-28 protein can be interpreted in different ways: either by the proteolytic degradation of receptors during the membrane preparation or experimental procedure or by the cross-linking of one binding component with other protein species within the plasma membrane. The first hypothesis cannot be com- pletely eliminated although the incubating medium of mem- branes contains bacitracin, a proteinase inhibitor. However, it is unlikely since the same results were obtained when the binding was performed in presence of phenylmethylsulfonyl fluoride. The second hypothesis may be considered, but the DSP cross-linking reaction is performed at a low temperature (4 "C) in order successfully to minimize any cross-reactions. Furthermore, the identification of the same specific labeled protein was obtained by a labeling procedure utilizing HSAB as the cross-linker, even though the chemically reactive groups and their spacing differ between DSP and HSAB. The cross- link obtained by using HSAB involves a substitution of an amino group and a substitution of amino acids through a radical mechanism, in contrast with the cross-link obtained by using DSP which involves two amino groups. On the other hand we observed a single '251-SS-28 protein in pancreatic acinar membrane (93 kDa), in agreement with the molecular mass of the acinar component, which has been reported after different covalent labeling with 1251-SS-14 [24,27]. The major

224

beta cell band, like the acinar receptors, was not affected by the reducing agent dithiothreitol, suggesting that this receptor does not have subunits connected by interchain disulfide bonds unlike the case of nerve growth factor [31] or insulin [32]. However, Srikant and Patel have reported components of different molecular mass (200 kDa, 80 kDa and 70 kDa) in the rat pancreatic acinar membranes [30] and in the rat adrenal cortex (200 kDa) [33]. Whether the differences ob- served in the structure of somatostatin receptors can be ac- counted for by tissue specificity or species differences is yet to be fully established. On the other hand, further investigations are also necessary to evaluate to what extent the characteristic structures are related to the different pharmacological speci- ficities of these receptors.

In conclusion, the present paper shows that specific mem- brane proteins are involved in the binding of somatostatin-28 on the pancreatic beta cell, which is different from that ob- served in the acini. This SS-28 receptor differs from those previously described by: (a) its greater affinity for the SS-28 than for SS-14, (b) the absence of any Ca2+ effect on the binding, (c) the presence of several receptor components after cross-linking experiments, (d) the presence of a main component of 132 kDa. We are now undertaking the solubilization and the purification of these receptors to characterize their structural properties fully and to determine their relationship with the inhibitory guanine nucleotide regu- latory protein. Our preliminary results suggest that the recep- tor is a glycoprotein. The distinctive characteristic of the beta cell SS-28 receptors fits well with the preponderant SS-28 action in inhibiting insulin release from the beta cells.

This work was supported by Institut National de la Sunti et de la Recherche Midicale (INSERM) and by the Fondation de la Recherche en Hormonologie (FRH), and a grant from the Diabetes Department of Istituto di Clinica Medica (Universita Cattolica, Roma) was awarded to Dr P. Cotroneo. We are grateful for the excellent cooper- ation of Prof. G. Ghirlanda and the valuable technical skill of R. de Chasseval. We also thank Dr C. Gespach for making available several peptides, P. Victor-Raphael and A. Le Ntdic for their technical assis- tance, C. Foulloy for her editorial assistance and Jean-Paul Wattiaux for his help in the quantitative density measurements.

REFERENCES 1.

2.

3.

4.

5.

Brazeau, P., Vale, W. L., Burgus, R., Ling, N., Butcher, M., Rivier, 1. &Guillemin, R. (1973) Science (Wash. DC) 179,77- 79.

Pradayrol, L., Jornvall, H., Mutt, V. & Ribet, A. (1980) FEBS Lett. 109, 55 - 58.

Dubois, M. P. (1975) Proc. Nut1 Acad. Sci. USA 72, 1340- 1343.

Ghirlanda, G., Bataille, D., Dubois, M. P. & Rosselin, G. (1978) Metabolism 27, S1,1167-1170.

Mandarino, L., Stenner, D., Blanchard, W., Nissen, S., Gerich, J., Ling, N., Brazeau, P., Bolhen, P., Esch, F. & Guillemin, R. (1981) Nature (Lond.) 291, 76-77.

6. Vaysse, N., Pradayrol, L., Chayvialle, A., Pignal, F., Esteve, J. P., Susini, C., Descos, F. & Ribet, A. (1981) Endocrinology 108,

7. Lewin, M. J. M. (1986) Scand. J. Gastroenterol. 21, suppl. 119,

8. Mehler, P. S., Sussman, A. L., Maman, A. & Sussman, K . E.

9. Reubi, J. C., Rivier, J., Perrin, M., Brown, M. & Vale, V. (1 982)

1843 - 1847.

42-46.

(1980) J. Clin. Invest. 66, 1334-1338.

Endocrinology 110, 1049- 1051. 10. Kirkman, H. A. (1962) StanfordMed. Bull. 20, 163. 11. Goldfine, I., Roth, J. & Birnbaumer, L. (1972) J . Biol. Chem. 247,

12. Bataille, D., Besson, J., Bastard, C., Laburthe, M. & Rosselin, G. (1977) in Hormonal receptors in digestive tract phy.c:ology (Bonfils, S . , Fromageot, P. & Rosselin, G., eds) pp. 113 - 123, Elsevier, Amsterdam.

13. Maletti, M., Portha, B., Carlquist, M., Kergoat, M., Laburthe, M., Marie, J. C. & Rosselin, G. (1984) Endocrinology 11.5,

1211-1218.

1324-1331. 14. Rosselin, G. (1987) Front. Horm. Res. 16, 51 - 8 2 . 15. Shapiro, S., Eto, S., Fleischer, N. & Baum, S. C. (1975) Endocrin-

16. Bradford, M. (1976) Anal. Biochem. 72, 248. 17. Zeggari, M., Viguerie, N., Susini, C., Esteve, J. P., Vaysse, N.,

Rivier, J., Wunsch, E. & Ribet, A. (1986) Peptides 7, 953- 959.

18. Greenwood, F. C., Hunter, W. M. & Glover, J. S . (1963) Biochem. J. 89, 114-123.

19. Antoniotti, H., Fagot-Revurat, P., Esteve, J. P., Fourny, D., Pradayrol, L. & Ribet, A. (1984) J . Chromatogr. 196, 181- 188.

20. Marie, J. C., Hui Bon Hoa, D., Jackson, R., Heiblum, G. & Rosselin, G. (1985) Regul. Pept. 12, 113-123.

21. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685. 22. Harris, M. J., Takery, H. & Rytting, H. J. (1973) J . Phys. Chem.

23. Marie, J. C., Boissard, C. & Rosselin, G. (1984) Peptides 5 , 179-

24. Sakamoto, C., Goldfine, I. D. & Williams, J. A. (1984) J. Biol.

25. Reubi, J. C. & Maurer, R. (1986) Regul. Pept. 14, 301 -31 1. 26. Kasuga, M., Van Obberghen, Nissley, S. P. & Rechler, M. M.

27. Susini, C., Bailey, A., Szecowka, J. & Williams, J. (1986) J . Biol.

28. Schmit, F., Pipeleers, D. (1986) in Diabetologiu, 22ndAnnu. Meet.

29. Srikant, C. B. & Patel, Y. C. (1981) Nature (Lond.) 294, 259-

30. Srikant, C. B. & Patel, Y. C. (1986) J. Biol. Chem. 261, 17, 7690. 31. Massague, J., Guillette, B. J., Czech, M. P., Morgan, C. &

Bradshaw, R. A. (1981) J. Biol. Chem. 256,9419-9424. 32. Kasuga, M., Hedo, J. A,, Yamada, K. M. & Kahn, C. R. (1982)

J . Biol. Chem. 257,10392-10399. 33. Srikant, C. B. & Patel, Y. C. (1986) Biochem. Biophys. Res.

Commun. 139, 757-762.

ology 97,442 - 447.

2694 - 2703.

182.

Chem. 259,9623 - 9627.

(1981) J. Biol. Chem. 256, 5305-5308.

Chem. 261,16738-16743.

EASD, 28,8,592 (A).

260.


Recommended