cholecystokinin-like immunoreactive peptides in the Dungeness crab, Cancer magister (Dana): immunochemical and biological characterization

Regulatory Peptides, 7 (1983) 155-170 Elsevier

155

RPT 00234

Gastrin/cholecystokinin-like immunoreactive peptides in the Dungeness crab, Cancer magister (Dana)" immunochemical and biological characterization B r e t t A. L a r s o n * a n d S t e v e n R. V i g n a ** Department of Biology, University of Oregon, Eugene, OR 97403, U.S.A. (Received 31 May 1983; accepted for publication 2 August 1983)

Summary The purpose of this investigation was to characterize a gastrin/cholecystokinin-like immunoreactant (G/CCK-LI) extractable from the crab, Cancer magister. G / C C K LI was extracted best in boiling water and was found mainly in the stomach, hemolymph and carapace. A relatively large immunoreactive peptide in the stomach and apparently smaller forms in the hemolymph and carapace were separated by Sephadex G-50 fractionation. Anion-exchange chromatography further fractionated the stomach form into three major peaks. The crab material cross-reacted with three antisera specific for the common C-terminus of gastrin/CCK, but cross-reacted much less with three antisera directed against other portions of the gastrin molecule. Partially purified crab stomach G/CCK-LI inhibited the binding of labeled CCK to mouse brain G / C C K receptors but not to rat pancreatic CCK receptors. The crab peptide did not stimulate rat gastric acid or rat pancreatic amylase secretion. These results indicate that the crab peptides are structurally similar to, but distinguishable from, the bioactive C-terminal amino acid sequence common to gastrins and CCKs. gastrin; cholecystokinin; CCK receptors; neuropeptide; invertebrate gastrin/CCK; gastrointestinal hormone

* Present address: Department of Zoology and Cancer Research Laboratory, University of California, Berkeley, CA 94720, U.S.A. ** To whom correspondence should be addressed. Tel.: 503-686-4513. 016%0115/83/$03.00 © 1983 Elsevier Science Publishers B.V.

156 Introduction

Gastrin and cholecystokinin (CCK) are members of a family of well characterized regulatory peptides found in the mammalian digestive tract and nervous system (for reviews see refs. 1-3). Gastrins and CCKs share biological properties because of the structural similarity of their carboxyl (C)-termini [4,5]. Gut gastrin and cholecystokinin function as hormonal regulators of various digestive processes [6]. In the nervous system they probably function as neurotransmitters or neuromodulators [7]. Gastrin and/or cholecystokinin-like immunoreactants (G/CCK-LIs) recently have been reported in representatives of several invertebrate phyla [8,9]. All of these reports have been based on immunoreactivity in tissue extracts or specific immunostaining using C-terminal-specific G / C C K antisera. There is little information available on the structure, chemistry, biological properties or physiological role of any G/CCK-LI in the invertebrates. In addition to increasing our knowledge of invertebrate regulatory peptides, such information should provide insight into the basic properties and evolution of this important peptide family. In a recent study, we found G/CCK-LI to be widespread in invertebrate species and tissues, and identified the stomach of the Dungeness crab, Cancer magister, as a rich source of G/CCK-LI [8]. In this report we describe our initial studies of the tissue distribution, immunochemical characteristics and biological properties of Cancer magister G/CCK-LI. Portions of this work have been published previously in abstract form [10,11].

Materials and Methods

Materials The antisera used were kind gifts from Dr. G. Rosenquist, University of California, Davis, CA (No. 5135), Dr. J. Walsh, University of California, Los Angeles, CA (Nos. 1296, 161I, and 1295), and Dr. G. Dockray, University of Liverpool, U.K. (L48 and Ll12). CCK8 was a gift from the Squibb Institute for Medical Research (Princeton, N J), and 125I-labeled Bolton-Hunter CCK33 (125I-BH-CCK) was a gift from Dr. J. Williams, University of California, San Francisco, CA. 15-Leucine synthetic human gastrin-17I (G17I) was purchased from Research Plus (Bayonne, NJ), pentagastrin was purchased from Calbiochem (La Jolla, CA), and 125I (100 mCi/ml) was obtained from Amersham (Arlington Heights, IL). The following were purchased from Sigma (St. Louis, MO): phenylmethylsulfonyl fluoride (PMSF), iodoacetamide (IAA), protease (Pronase, Type V), and trypsin (DPCC treated, Type XI). Sephadex G-50 superfine was obtained from Pharmacia (Piscataway, N J) and diethylaminoethyl (DEAE) cellulose (DE-53) was purchased from Whatman (Clifton, N J). Chloramine T was obtained from Eastman Kodak (Rochester, NY). Tissue extractions Tissues were dissected from adult male crabs at 4°C, rinsed and stored frozen at -20°C. For aqueous extractions, various amounts of distilled water, acetic acid

157 (3%), ammonium acetate (0.1 M), or ammonium bicarbonate (0.25 M) were added to the thawed and minced tissues which were then boiled for 15 min. After cooling, the extracts were clarified by centrifugation (45 min, 20000 × g, 4°C), and the supernatants were stored at - 2 0 ° C until assayed. The protease inhibitors, PMSF and IAA, were dissolved (30 mg/ml) in 100% ethanol and used at 0.3 mg/ml in two distilled water extractions, both before and after boiling. Organic extractions including acetone and chloroform/methanol (2:1) were also tested. The hemolymph samples were either allowed to coagulate for 30 min on ice or heated in a boiling water bath for 10 min, centrifuged and stored at -20°C. In the tissue distribution experiments, 0.2 g/ml distilled water extractions were used whenever possible. If less than 0.2 g of tissue were available, enough water to give a final volume of 1 ml was used, resulting in increased radioimmunoassay (RIA) detection limits (pmol/g) as indicated in Table I. All other extractions were done in 0:1 g/ml boiling water.

Radioimmunoassays Gastrin/CCK RIAs were performed as described by Rosenquist and Walsh [12] using 15-1eucine synthetic human gastrin-17I as standard and label and anion-exchange resin separation. Monoiodogastrin was prepared using chloramine T by the method of Stadil and Rehfeld [13]. Potency estimates for crab G / C C K - L I are given in G17I equivalents. Antiserum No. 5135 was the only antiserum used for all experiments except the multiple antisera analyses. This antiserum is specific for the C-terminal pentapeptide region identical in mammalian gastrins and CCKs and thus binds both peptides with nearly equal affinity (see Table II). The C-terminal region is the bioactive center of these peptides and appears to be highly conserved in evolution [14]. The characteristics of the antisera used in the multiple antisera analyses are described in Table II.

Column chromatography Sephadex G-50 superfine columns (1 x 120 cm) were equilibrated and eluted at 4°C with 0.025 M veronal buffer (pH 8.6) containing 0.02% sodium azide. Fractions of 1.5 ml were collected at a flow rate of 6 ml. cm-2. h-1. Blue dextran or bovine serum albumin and Nal2SI were added to the samples to provide markers for the void and salt volumes, respectively. Individual samples were eluted from a Sephadex G-10 desalting column (0.8 × 25 cm) with distilled water at room temperature in 0.3 ml fractions. G / C C K - L I and standard peptides in the column fractions were detected by RIA with antiserum No. 5135. Batch adsorption and anion exchange chromatography were performed using DEAE cellulose (DE-53) equilibrated in 0.05 M ammonium bicarbonate (pH 8.2). In the batch procedure, crab G / C C K - L I was eluted at room temperature with 1.0 M ammonium bicarbonate in a large Biichner funnel. For the anion-exchange column (2.5 x 42 cm), linear gradients of 0.05-0.2 and 0.2-1.0 M ammonium bicarbonate were used for sample elution at 4°C. Fractions (5.4 ml) were collected at an average flow rate of 26 ml. cm- 2. h - 1.

158

Receptor assays Mouse cerebral cortical and rat pancreatic membrane preparations were prepared and receptor binding was performed as described by Saito et al. [15] and Steigerwalt and Williams [16]. 125I-labeled Bolton-Hunter porcine CCK33 label was used at a final concentration of 50 pM. Specific binding was expressed as percent of total binding minus nonspecific binding in excess cold CCK. The ability of CCK8 standard and crab G/CCK-LI to displace label from the brain and pancreatic CCK receptors was evaluated by incubating various concentrations of the peptides with fixed concentrations of membranes and label. Gastrin bioassay Gastric acid secretion in response to peptides was measured by a method modified from that described by Halter et al. [17]. Urethane-anesthetized rats were prepared for continuous perfusion of the stomach with 5~ glucose. The perfusate was recirculated via a flow-through conductivity cell. Conductivity was measured as an index of gastric acid secretion. Standards and extracts were administered i.v. into the external jugular vein. CCK bioassay Pancreatic amylase secretion in response to peptides was measured in vitro using the method of Scheele and Palade [18]. Individual lobules were dissected from the rat pancreas and separately incubated with CCK8 standard or crab G/CCK-LI in buffer for 2 h at 37°C. Amylase activity was measured in each lobule homogenate and suspending solution using the colorimetric method of Bernfeld [19]. Amylase release was expressed as the percent of total amylase activity present.

Results Extractability Boiling stomach tissue in distilled water or ammonium acetate gave the highest yields of G/CCK-LI. Lower yields were obtained with acetic acid, armnonium bicarbonate, acetone and chloroform:methanol extractions. Boiling the centrifugation pellets in water, acid or base a second time extracted little additional G/CCK-LI ( < 10~ of ori$inal). The protease inhibitors (PMSF and IAA) also did not increase the yield. Therefore, b ~ in distilled water (0.1 g/ml) was used for most subsequent experiments. Tissue distribution The distribution of G/CCK-LI in various Cancer magister tissues is shown in Table I. The RIA detection limits per gram wet tissue w @ t varied because of the small amounts of some tissues available, The esophagus, stomach, carapace (external) and hemolymph contained measurable immunoreactive material. Of the stomach regions, the anterior region of the cardiac stomach had the highest concentration of G / C C K - L I (13.0 pmol/g), followed by the posterior cardiac stomach or gastric mill

159 TABLE I Tissue distribution of Cancermagister gastrin/CCK-like immunoreactivity Tissue

Esophagus Anterior cardiac stomach Gastric mill (posterior cardiac stomach) Pyloric stomach Midgut Anterior midgut caeca Posterior midgut caecum Hindgut Hepatopancreas Bladder Propodus muscle of first walking leg Brain Commissural ganglia Esophageal ganglia Stomatogastric ganglia Pericardial organs Ventral thoracic ganglia Eyestalks (excluding carapace) Hypodermis Carapace Hemolymph

Gastrin/CCK-like immunoreactivity a (G17I equiv.; pmol/g wet weight) 0.13+0.07 (5) b 13.04 + 3.72 (6) 4.62 + 1.29 (6) 2.29 + 0.62 (6) < 0.05 (9) < 0.08 (9) < 0.05 (9) < 0.05 (9) < 0.05 (9) < 0.05 (9) < 0.05 (9) < 0.14 (pooled sample) c < 0.25 (pooled sample) < 3.79 (pooled sample) < 3.55 (pooled sample) < 0.88 (pooled sample) < 0.07 (9) < 0.10 (9) < 0.05 (1) 2.65 + 0.50 (7) 109.13 5:9.90 pmol/1 (8)

a Using 125I-G17Ias label, G17I as standard, and antiserum No. 5135. b The mean + S.E.M. and the sample size in parentheses are given. c Samples were pooled from nine individuals because of the small individual tissue weights. region (4.6 p m o l / g ) , and the most posterior region, the pyloric stomach (2.3 p m o l / g ) . G / C C K - L I concentrations in extracts of various regions of the carapace including sections f r o m the dorsal and ventral surfaces, the merus of the first periopod, the a b d o m e n and the eyestalk ranged from 1 to 4 p m o l / g . Other tissues, including various neural ganglia, neurohemal organs and hypodermis underlying the carapace, did not have G / C C K - L I detectable b y antiserum No. 5135. The lumen of the crab stomach is lined with a chitinous cuticle similar to the external carapace. We dissected and separately extracted the various tissue layers associated with the stomach. The reticular connective tissue and external muscles and nerves surrounding the stomach did not have detectable immunoreactive material. Extracts of mucosal tissue, including the cuticle, and serosal tissue, including the epithelium, from the cardiac stomach b o t h contained G / C C K - L I , the latter having over twice the tissue concentration of the former. Immunochemical characterization Evidence for the peptidic nature of the G / C C K - L I was obtained b y incubating a

160 crude boiling water extract of crab stomach with the nonspecific protease, Pronase, at an enzyme concentration of I m g / m l . Pronase eliminated detectable immunoreactivity ( < 50 f m o l / m l ) after incubation for 10 min at room temperature. The immunoreactivity in control samples incubated with heat-inactivated Pronase was not reduced. Fractionation of a crude cardiac stomach extract on Sephadex G-50 superfine revealed a major peak (approx. 55% of initial) of G / C C K - L I eluting (K~v = 0.57) just prior to a 17 residue gastrin standard (Fig. 1). Other immunoreactive peaks (each approx. 20~ of initial) eluted in (K~v = 0.0) and just after (Kay = 0.25) the

O i to) o ou

3

8

|[

1

1

3600-

2400-

J 1200'

~' r.

0

o '7m

30.

o

0E Q.

20

>-

t0-

I-L) < n~ o Z

0

. i~..e.e_ i ....

I

Hemolymph

ID.--O-0----O ..... -O--------

11

120-

1

A

- ........

;

Coropoce 80W

-Ii

v

0

40-

.................................. -.; ~ .i .~ .~ .:~ .~ .;~ .'~ .~ .~ ,'.o ,'., ;.2 Kay, BSA to ; ~ I

Fig. 1. Separation of crude crab stomach (top), hemolymph (middle) and carapace (bottom) extracts on Sephadex G-50 columns. Elution volume is expressed as the elution constant (Kav= the fractional elution volume between bovine serum albumin (BSA) and Na12~I).The columns were separately cal~rated with G34, CCK33, G17 and caerulein-10 (see arrows).

161

void volume. When crab hemolymph is fractionated under the same conditions a major peak (approx. 86% of initial) is again found in the void volume, but a significant amount (approx. 34% of initial) elutes (Kay = 0.87) just prior to the elution volume of a 10 residue calibration marker (Fig. 1). All of the G / C C K - L I in carapace extracts (approx. 92% of initial) elutes (Kay -- 0.85) at this same 10 residue position (Fig. 1). None of this apparently small-sized component is present in cardiac stomach extracts. In light of the ability of trypsin to cleave many larger precursor proteins to smaller forms (including mammalian gastrin and cholecystokinin and even a molluscan gastrin-like peptide [20]), we incubated a sample of the major G-50 purified stomach component with trypsin (1 mg/ml) for 1 h at 37°C and refractionated the sample on G-50. The resulting elution profile was not significantly changed from that of a control sample treated with heat-inactivated trypsin (data not shown). The native trypsin was able to convert G34 to G17. To rule out the possibility that the stomach might contain a smaller form of G / C C K - L I nonspecificaUy bound to larger proteins, we incubated a crude cardiac stomach extract in 8 M urea. Fractionation on Sephadex G-50 in 8 M urea did not change the elution profile from that of a control sample (data not shown).

! c

35 g C- 1 2 0 0 .

3O

c.

I 1ooo.

i o.

800-

5

_> 6oo

8 0 400 Z D :E

1o

08 0,6

200

5

g 0.4 2 0.2

_:_:-f-""'-"-''"'~-.7..:.-.:. 20

60

1OO

440

"':".......... "'"" 180 220 260 FRACTION NUMBER

300

"-"-"-" ....:...= 340

380

00

420

Fig. 2. DEAE cellulose anion-exchange chromatography of Sephadex G-50 purified crab stomach G / C C K - L I . G/CCK-LI in the crude stomach extract had previously been batchwise adsorbed to and eluted from DE-53 and then fractionated on Sephadex G-50 (see text for details). The major G-50 peak material (Kay ffi 0.57) was loaded onto a DE-53 anion-exchange column previously equilibrated with 0.05 M ammonium bicarbonate. The elution profiles are shown for absorbance at 280 n m ( . . . . . . ), conductivity ( ) and G / C C K - I ~ e immunoreactivity measured with antiserum No. 5135 (O Q).

162 To characterize further the stomach G / C C K - L I and to obtain a concentrated source of the peptide for biological activity studies, we partially purified G / C C K - L I from 50 crab stomachs (170 g). As a first step, we incubated the crude extract with D E A E cellulose (DE:53) equilibrated in 0.05 M a m m o n i u m bicarbonate. After washing, the adsorbed G / C C K - L I was eluted with 1.0 M a m m o n i u m bicarbonate, tyophilized, redissolved and fractionated on a Sephadex G-50 column (3 x 90 cm) equilibrated with 0.1 M a m m o n i u m bicarbonate. The resulting elution profile was virtually identical to that of the crude stomach extract (data not shown). The major peak fractions eluting just prior to the G17 marker were pooled, diluted with an equal volume of distilled water, and applied to a DE-53 anion-exchange column. Three major immunoreactive peaks were resolved: Peaks II, III and IV in order of increasing adsorption affinity (Fig. 2). Separate refractionation of a sample from each of these three peaks on Sephadex G-50 columns (1 x 120 cm) revealed that each contained G / C C K - L I eluting in the same volume as the major stomach form (Kay = 0.57) (data not shown). Portions of DE-53 peaks II, III and IV were pooled separately, lyophilized, redissolved in distilled water, and desalted on a Sephadex G-10 column. The void volume fractions containing the least salt (conductivity) were pooled, lyophilized, and redissolved in distilled water for bioactivity studies. Final yields of peaks II, III and IV were 15, 160 and 80 immunoreactive pmol, respectively, of the original 3 nmol in the crude extract.

TABLE II Characteristics of antisera used in radioimmunoassays Antiserum 5135

RIA titer 1 : 600000

iD5o a 3

L48

1 : 100000

Ll12

1:8000

1296

1:225000

4

1611

1 : 400000

3

1295

1:60000

5

5

65

Specificity b C-terminus of G/CCK (reacts nearly equally with G17 and CCKS, but weakly with 04) C-terminus of G/CCK (reacts equally with 017 and CCK8, but weakly with 04) C-terminus of G/CCK (reacts equally with G17, CCK8 and 04) C-terminus of gastrin (reacts weaklywith CCK8 and 04) C-terminus of gastrin (reacts very weakly with CCK8 and 04) N-terminus of 017 (does not react with CCK8 or 04)

Reference

12

36

37 12, 38 12 12, 38

a Concentration (pM) of standard 0171 needed for half-maximalinhibition of label bindln~b For full details see text and references.

163

Multiple antisera analyses To characterize the crab G/CCK-LI further and to rule out the possibility of nonspecific effects of apparent crab G/CCK-LI with the single antiserum used for most of these studies (No. 5135), we measured the cross-reactivities of various samples with five different gastrin/CCK antisera. Potency estimates were obtained at those sample dilutions resulting in inhibition of label binding closest to half-maximal (IDs0) and relative to a G17I standard (Table III). Potency estimates using antiserum No. 1295 are not given because of the unusual sample dilution curves obtained with this antiserum (Fig. 3) indicating nonspecific effects in the RIA. The hypodermis, eyestalk, and brain extracts exhibited some cross-reactivity with antisera L48 and/or Ll12, but not with No. 5135 at the doses tested. G/CCK-LI in

10 7 t

10 6 i

to 4

lo 3

10 B 1

10 4

I ;

I

107 I

106 I

104

tO3

I

I

105 I

I02

104

107 I

J

1oI

10 4

I

10 s I

10 3 t

105 I

Reoprocol Odution (DE-g)

104 I

10 z

101

I

I

ReciprocoI Dilution (other s o m p l e s )

,85' 1.4-

.61

12 i

~ .5" i .3-

1,0-

.4-

,8.6-

5135

u. £3 Z

G171 Stomoch Pot tiolly Purified

•

Stomoch (DE n)

•

Coropoce Hemolymph

•

.2-

.4.

.2'

x 0

L48 r ----

10 4

103

1~

10t

I

I

i

10 3 i

10 2 i

1

1o

lOO

1

lO

10 4

101 i

104

103

102

101

,

,

,

t

tO

Reciprocoi Dilution (oil somples)

0 m

20.

1.6. 1.2. .8 .4'

t00

100 G17I (pM)

Fig. 3. Radioimmunoassay dilution curves for the inhibition of 125I-G17I binding to antisera No. 5135, L48, Ll12, Nos. 1296, 1611, and 1295 produced by addition of graded dilutions of Cancer extracts and standard G17I. The samples were G17I ( x x ) , crude extracts of Cancer stomach ( O ©), carapace (11 II) and hemolymph (~, zx), and partially purified DE-53 peak II G/CCK-LI from the crab stomach (DE-II) ($ e). Binding is expressed as the ratio of bound to free radiolabeled peptide corrected for nonspecific binding in the absence of antiserum. The concentration of G17I (pM) and the reciprocal dilution of the crab samples are plotted on the abscissa.

164

the crude and partially purified stomach extracts and in the carapace extract cross-reacted most with the C-terminal, tetrapeptide-specific antiserum, Ll12. These three samples cross-reacted relatively well with the other antisera (No. 5135 and L48) specific for the common C-terminal amino acids shared by gastrin and CCK, but had much less cross-reactivity with the more gastrin-specific antisera (Nos. 1296 and 1611). The hemolymph sample, on the other hand, cross-reacted with all of the antisera except Ll12. Representative sample dilution curves with various antisera are shown in Fig. 3. Both parallelism and nonparallelism to the G17I standard curve are exhibited..

DE- 111'(Rec~procol O,luhon) 1000 100~ 10 100 -

80

60

E

40

x

20

a5 i

i

100 1g3

Cortex

80

6O

40

20

.&

;

.~ CCK 8

Ib

(nM)

Fig. 4. Inhibition of specific t25I-BH-CCK binding to rat pancreatic (top) and mouse cerebral cortical (bottom) membrane preparations by unlabeled CCK8 (O O) and partially purified DE-53 peak III G/CCK-LI from the crab stomach (DE-III)(@ @). The membrane preparations were incubated with 50,pM 125I-BH-CCK for 120 min at 22°C and pH 6.5. Varying concentrations of CCK8 and peak III G/CCK-LI were.added as shown. Nonspecific binding in the presenc~ of 500 nM unlabeled CCK8 was subtracted from total binding, and specific binding was expressed as percent of binding in the absence of unlabeled peptides.

165

Biological characterization G / C C K - L I from the partially purified DE-53 peak III component (cf. Fig. 2) competed with 125I-BH-CCK for binding to mouse cerebral cortical G / C C K receptors (Fig. 4). The highest concentration of peak III G/CCK-LI tested (25 nM as determined by RIA with antiserum No. 5135) inhibited the binding of a2SI-BH-CCK by 39%. Equivalent inhibition is obtained with 0.15 nM porcine CCK8. Construction of a complete dose-response curve was not possible because of the limited amount of sample available. Peak II and peak IV material at the highest concentrations tested (3 and 10 nM, respectively) did not significantly inhibit a25I-BH-CCK binding to the cortical receptors, whereas these same immunoreactive concentrations of peak III inhibited binding by 23% and 29%, respectively. At the same concentrations used in the mouse radioreceptor assay, none of the partially purified G/CCK-LI components significantly affected the binding of ~25I-BH-CCK to rat pancreatic CCK receptors (Fig. 4). Peak II G/CCK-LI, prior to desalting on G-10, and peaks III and IV G/CCK-LI did not stimulate rat gastric acid secretion in vivo at any dose (up to 23, 20 and 10 immunoreactive pmol, respectively). In the same bioassay 5 pmol of pentagastrin elicited a clear response. In the in vitro rat pancreatic lobule bioassay all three components were unable to stimulate the secretion of amylase using final immunoreactive concentrations of 1.2 nM, 25 nM, and 10 nM for peaks II, III and IV, respectively. The concentration of porcine CCK8 producing a half-maximal response in this bioassay was 30 pM.

Discussion

Crab G / C C K - L I occurs in several molecular forms and is found mainly in three tissues. Gastrointestinal G/CCK-LI is most concentrated in the anterior cardiac stomach and is best extracted by boiling water (an optimal extraction method for some forms of G / C C K in other species as well [9,21]). Because G / C C K - L I was found in the external carapace, it was possible that the extractable gastricG/CCK-LI was limited only to the chitinous cuticle lining the gastric lumen. Evidence that the cuticle is not the only source of the stomach G/CCK-LI was obtained by extraction and RIA of the different stomach-associated tissue layers. In agreement with our earlier work, relatively little G / C C K - L I was found in crab neural tissues, unlike other arthropods and most other invertebrates that have been investigated [8]. G / C C K peptides have been found in the skin of certain frog species (caerulein; [22]) and the body wall of the earthworm [8,23]. The present demonstration of G / C C K - L I in the crab carapace is possibly another example of G / C C K peptides localized in the integument. The cellular source of this carapace G/CCK-LI is not clear for many reasons including the lack of detectable antiserum No. 5135 cross-reactive material in the hypodermis (Table III) which elaborates the carapace. However, G / C C K - L I found in the hypodermis with antisera L48 and Ll12 may be a precursor form of the carapace G/CCK-LI with different antigenic properties. The sensitivity of crab G/CCK-LI to pronase digestion, resistance to boiling and

166 T A B L E III G / C C K - L I concentrations (pM) in Cancer magister tissue extracts measured by radioimmunoassay using

various antisera a,b Sample

Antiserum c 5135

Crude stomach Partially purified stomach (DE-53 Peak II) Carapace Hemolymph Hypodermis Eyestalk (excluding carapace) Brain

320 d 510000 50.0 13.0 < 10 < 10 < 10

L48

Ll12 360

1200000 78.0 22.0 85.0 44.0 85.0

110000 22000000 2 700 < 150 830 290 < 600

1296

1611

< 10

< 10

2000 31.0 57.0 < 10 <:20 < 100

980 < 10 16.0 < 10 < 20 < 100

a Detection limits varied because of the small a m o u n t s of some tissues available and antisera differences in least detectable doses. b Extracts of ventral thoracic, commissural, esophageal and stomatogastric ganglia and propodus muscle of the first periopod did not have detectable cross-reactivity with any of the antisera. c For antisera specificities see Table II. a Potency estimates were determined with all antisera using 1251-G17I as label, G17I as standard, and those sample dilutions producing label binding inhibition closest to half-maximal for each antiserum.

Sephadex G-50 elution profile indicate its peptidic nature. Sephadex G-50 fractionation of the stomach G/CCK-LI revealed at least three well-resolved peaks. The large amount of protein eluting in the void volume, which can have nonspecific effects in the RIA, may indicate that part or all of the apparent immunoreactivity in the void peak is artifactual, as has been shown previously for many 'big big gastrins' [24]. The intermediate peak represented only a small portion (22~) of the total stomach G/CCK-LI. The major stomach peak representing 58~ of the total immunoreactivity eluted just prior to the volume of standard human G17. Because of its relatively high concentration this form of stomach G / C C K - L I was characterized further. Its resistance to trypsin cleavage indicated that this peptide, like mammalian G17, probably does not have any internal lysine or arginine residues. Anion-exchange fractionation on DEAE cellulose and refractionation on Sephadex G-50 revealed that the major stomach component consisted of at least three forms (peaks II, III and IV; Fig. 2) separable by charge. Their affinity for DEAE suggests that these peptides are slightly acidic at pH 8.2. Both hemolymph and carapace have major G/CCK-LI components eluting from Sephadex G-50 just prior to the elution volume of eaerulein, a 10-residue standard peptide. Material of this size does not appear to be present in the stomach. The apparently different M r variants in the stomach and other tissues may be biosynthetically related, but no evidence for this has yet been found. The major stomach component (Kay - 0.57) was resistant to trypsin cleavage, but it is possible that other forms are cleavable by trypsin-like enzymes or that endogenous enzymes are capable of specific cleavage. More work is needed to clarify the relationship between the stomach, hemolymph and carapace G/CCK-LI.

167

The Sephadex G-50 elution position of the major stomach G / C C K - L I corresponds well with that of G / C C K - L I found after gel filtration of nervous tissue extracts in two other arthropods, the insects Manduca [25] and Calliphora [9]. Aplysia gut extracts and hemolymph and Lumbricus intestinal extracts also appear to have G / C C K - L I components of about this size [20,23]. Only the Aplysia form is apparently cleavable by trypsin. Detectable hemolymph G / C C K - L I has been reported in two other invertebrate species, the molluscs Aplysia and Otala, where its apparent size and concentration are comparable to that found in crab hemolymph [20]. Along with the crab carapace, G / C C K - L I of this apparent small size has also been reported in extracts of Aplysia and Otala gut, Calliphoraand Helix neural tissue, and Lumbricus intestine [9,20,23,26]. As has been found in other studies of invertebrate neuronal G / C C K - L I employing multiple anti-G/CCK sera [9,27,28], the Cancer stomach and carapace G / C C K LIs have antigenic determinants in common with the C-terminus of mammalian G / C C K . Cross-reactivity was found with the three antisera specific for the C-terminal sequence shared by gastrin and CCK (No. 5135, L48 and Ll12). Of the antisera tested, this G / C C K - L I cross-reacted most avidly with the tetrapeptide-specific antiserum Ll12, a property shared by the Calliphora brain peptide [9]. Much lower cross-reactivity was observed with antisera more specific for the C-terminus of gastrin (Nos. 1296 and 1611), but since no CCK-specific antiserum was used, we do not know yet whether the crab C-terminal immunoreactivity is more gastrin- or CCK-like. Apparent cross-reactivity seen in the crude stomach extract with an antiserum (No. 1295) specific for the N-terminus of G17 may indicate some components with N-terminal homology. However, the unusual sample dilution curves with this antiserum seem to indicate nonspecific effects in the RIA (Fig. 3). The sample dilution curves with the six different antisera shown in Fig. 3 exhibit various degrees of nonparallelism to the G171 standard curve. The nonparallelism exhibited suggests that the antigenic binding site in crab G / C C K - L I is not identical to the C-terminus of mammalian G / C C K [3]. However, other factors may be involved [3,29]. Because sample and standard curves are not parallel, potency estimates do not allow 'absolute' quantitation. However, such estimates are valid for comparing relative amounts of immunoreactive material, especially when the samples are assayed at comparable dilutions in the same RIA as we have done in these studies. The concentrations we have reported here are likely to be underestimates. All reports so far of non-protochordate invertebrate G / C C K - L I have been based on immunochemical cross-reactivity only. Bevis and Thorndyke [31] have demonstrated both stimulation of gastric enzyme secretion by porcine CCK in the protochordate, Styela, and stimulation in mammalian bioassays by the Styela peptide [32]. We attempted to supplement our immunochemical data for crab G / C C K - L I by determining its ability to bind to mammalian G / C C K receptor preparations and elicit biological responses in mammalian G / C C K target organs. Partially purified crab stomach G / C C K - L I (peak III) was capable of inhibiting the binding of porcine CCK33 label (125I-BH-CCK) from mouse brain, but not rat pancreatic receptor preparations at the doses tested. Peak II and peak IV material did not inhibit 125I-BH-CCK binding to either receptor preparation at the doses tested.

168

Receptors in the mouse brain bind gastrin and CCK with almost equal affinity, whereas rat pancreatic receptors are much more specific for the CCK-like structure [15,16]. Thus, our receptor binding results, in conjunction with the antisera cross-reactivity studies, support the concept that crab G / C C K - L I has structural similarity to the C-terminal pentapeptide sequence common to gastrin and CCK. Unlike cnidarian G/CCK-LI, which was reported to be more similar to CCK than gastrin [28], our results do not suggest any closer similarity to CCK than gastrin for crab G/CCK-LI. In mammalian gastrin and CCK bioassays, crab G / C C K - L I had no apparent G / C C K biological activity using doses at which mammalian gastrins and CCKs produce maximal responses. The inability of crab G / C C K - L I to stimulate rat pancreatic enzyme secretion correlates with its inability to bind to rat pancreatic receptors. However, the reported binding properties of gastrin receptors in the rat stomach are similar to those in the mouse brain [15,33]. Thus, the inability of crab G / C C K - L I to stimulate acid secretion from the rat stomach may indicate that it is capable of binding to gastrin-like receptors but is unable to elicit the normal biological response. Alternatively, it is possible that crab G / C C K - L I may be incapable of binding to the stomach receptors, may have been degraded faster than G17I in vivo, or the concentrations used may have been too low. Since there is no bioassay available mediated by brain G / C C K receptors, a direct test of this hypothesis was not possible. Large quantities of pure or synthetic crab G / C C K - L I would be needed to test thoroughly for competitive inhibition of gastrin-stimulated gastric acid secretion. No physiological function has been demonstrated for any invertebrate G / C C K - L I [3]. The concentration of crab G / C C K - L I in the anterior stomach and its presence in the hemolymph are consistent with a possible gastrointestinal hormonal role. In fact, Denuce [34] has reported that porcine CCK may affect hepatopancreatic enzyme release in crustaceans. However, we have not detected changes in the hemolymph levels of G / C C K - L I after feeding fasted crabs [10]. The relatively high concentrations of G / C C K - L I found in carapace extracts are probably not high enough to indicate a structural role, but may suggest possible involvement in the molting process. However, crab G / C C K - L I does not seem to be associated with all chitinous tissues, because we did not detect any cross-reactivity in the hindgut, which has a cuticular lining similar to the stomach. In summary, crab G / C C K - L I exhibits the four basic characteristics of the gastrin-cholecystokinin system proposed by Rehfeld [1]: (1) Homology. Immunological cross-reactivity and receptor binding data indicate structural similarity to the C-terminus of mammalian G / C C K . (2) Heterogeneity. Several molecular forms of G / C C K - L I are present in the crab. Variants in size, 'macroheterogeneity', and charge, 'microheterogeneity', are found in the stomach. (3) Ubiquity. The G / C C K - L I is located in at least three different tissues including the stomach, carapace and hemolymph. (4) Differential principality. Different molecular forms predominate in the stomach versus the carapace and hemolymph. Based on these criteria and the other evidence presented, crab G / C C K - L I appears to be an authentic member of the G / C C K regulatory peptide family. However, immunological cross-reactivity

169

may not reflect chemical identity (see, for example, ref. 35). Final verification will require determination of the structures of these peptides. Such studies are in progress in our laboratory.

Acknowledgements We thank James L.M. Morgan for advice and assistance, the Oregon Institute of Marine Biology for assistance in collecting animals, Drs. G. Rosenquist, J. Walsh, and G. Dockray for gifts of antisera, Dr. J. Williams for 125I-BH-CCK, and the Squibb Institute for Medical Research for CCK8. This work was supported by funds from the Lerner Fund for Marine Research, a Grant-in-Aid of Research from Sigma Xi, The Scientific Research Society, and USPHS Training Grant 5 T32 GMO 7413 to B.A.L., and grants from the Medical Research Foundation of Oregon and NSF PCM 81-04338 to S.R.V.

References 1 Rehfeld, J.F., Four basic characteristics of the gastrin-cholecystokinin system, Am. J. Physiol., 240 (1981) G255-G266. 2 Mutt, V., Chemistry of the gastrointestinal hormones and hormone-like peptides and a sketch of their physiology and pharmacology, Vitamins Hormones, 39 (1982) 231-427. 3 Larson, B.A., Gastrin/cholecystokinin-like immunoreactivd peptides in the Dungeness crab, Cancer magister (Dana): Immunochemical, biological, and structural characterization, Ph.D. dissertation, University of Oregon, Eugene, 1983, 168 pp. 4 Tracy, H.J. and Gregory, R.A., Physiological properties of a series of synthetic peptides structurally related to gastrin I, Nature (London), 204 (1964) 935-938. 50ndetti, M.A., Rubin, B., Engel, S.L., Pluscec, J. and Sheehan, J.T., Cholecystokinin-pancreozymin: recent developments, Am. J. Dig. Dis., 15 (1970) 149-156. 6 Grossman, M.I., Physiological effects of gastrointestinal hormones, Fed. Proc., 36 (1977) 1930-1932. 7 Williams, J.A., Cholecystokinin: ahormone and a neurotransmitter, Biomed. Res., 3 (1982) 107-121. 8 Larson, B.A. and Vigna, S.R., Species and tissue distribution of cholecystokinin/gastrin-like substances in some invertebrates, Gen. Comp. Endocrinol., 50 (1983) 469-475. 9 Dockray, G.J., Duve, H. and Thorpe, A., immunochemical characterization of gastrin/cholecystokinin-like peptides in the brain of the blowfly, Calliphora vomitoria, Gen. Comp. Endocrinol., 45 (1981) 491-496. 10 Larson, B.A. and Vigna, S.R., Cholecystokinin-like immunoreactivity in the Dungeness crab, Cancer magister, Am. Zool., 20 (1980) 830 (Abstract). 11 Larson, B.A., Scalise, F.W., Reeve, J.R., Jr. and Vigna, S.R., Immunochemical characterization and localization of gastrin/cholecystokinin-like immunoreactants in the Dungeness crab, Cancer magister, Am. Zool., 22 (1982) 949 (Abstract). 12 Rosenquist, G.L. and Walsh, J.H., Radioimmunoassay of gastrin. In G.B.J. Glass (Ed.), Gastrointestinal Hormones, Raven Press, New York, 1980, pp. 769-795. 13 Stadil, F. and Rehfeld, J.F., Preparation of 125I-labelled synthetic human gastrin I for radioimmunoanalysis, Scand. J. Clin. Lab. Invest., 30 (1972) 361-368. 14 Dockray, G.J., Evolutionary relationships of the gut hormones, Fed. Proc., 38 (1979) 2295-2301. 15 Saito, A., Goldfine, I.D. and Williams, J.A., Characterization of receptors for cholecystokinin and related peptides in mouse cerebral cortex, J. Neurocliem., 37 (1981) 483-490. 16 Steigerwalt, R.W. and Williams, J.A., Characterization of cholecystokinin receptors on rat pancreatic membranes, Endocrinology, 109 (1981) 1746-1753.

170 17 Halter, F., Kohler, B., Leuenberger, J., Smith, G.M. and Varga, L., The development of a reliable and sensitive bioassay for gastrin in body fluids, Eur. J. Pharmacol., 35 (1976) 369-378. 18 Scheele, G.A. and Palade, G.E., Studies on the guinea pig pancreas. Parallel discharge of exocrine enzyme activities, J. Biol. Chem., 250 (1975) 2660-2670. 19 Bernfeld, P., Amylases, a and ft. In S.P. Colowick and N.O. Kaplan (Eds.), Methods of Enzymology, Academic Press, New York, 1 (1955) 149-158. 20 Straus, E., Yalow, R.S. and Gainer, H., Molluscan gastrin: concentration and molecular forms, Science, 190 (1975) 687-689. 21 Dockray, G.J., Immunoreactive component resembling cholecystokinin octapeptide in intestine, Nature (London), 270 (1977) 359-361. 22 Anastasi, A., Erspamer, V. and Endean, R., Isolation and amino acid sequence of caerulein, the active decapeptide of the skin of Hyla caerulea, Arch. Biochem. Biophys., 125 (1968) 57-68. 23 Rzasa, P., Kaloustian, K.V. and Prokop, E.K., Immunochemical evidence for a gastrin-like peptide in the intestinal tissues of the earthworm Lumbricus terrestris, Comp. Biochem. Physiol., 71A (1982) 631-634. 24 Rehfeld, J.F., Schwartz, T.W. and stadil, F., Immunochemical studies on macromolecular gastrins. Evidence that 'big big gastrins' are artifacts in blood and mucosa, but truly present in some large gastrinomas, Gastroenterology, 73 (1977) 469-477. 25 Kramer, K.J., Speirs, R.D. and Childs, C.N., Immunochemical evidence for a gastrin-like peptide in insect neuroendocrine system, Gen. Comp. Endocrinol., 32 (1977) 423-426. 26 Osborne, N.N., Cuello, A.C. and Dockray, G.J., Substance P and cholecystokinin-like peptides in Helix neurons and cholecystokinin and serotonin in a giant neuron, Science, 216 (1982) 409-411. 27 Duve, H. and Thorpe, A., Gastrin/cholecystokinin (CCK)-like immunoreactive neurones in the brain of the blowfly, Calliphora erythrocephala (Diptera), Gen. Comp. Endocrinol., 43 (1981) 381-391. 28 Grimmelikhuijzen, C.J.P., Sundler, F. and Rehfeld, J.F., Gastrin/CCK-like immunoreactivity in the nervous system of coelenterates, Histochemistry, 69 (1980) 61-68. 29 Yalow, R.S. and Berson, S.A., Problems of validation of radioimmunoassays. In W.D. Odell and W.H. Daughaday (Eds.), Principles of Competitive Protein-Binding Assays, J.B. Lippincott, Philadelphia, 1971, pp. 374-400. 30 Ross, G.T., Factors governing choice of a reference preparation for competitive binding assays. In W.D. Odell and W.H. Daughaday (Eds.), Principles of Competitive Protein-Binding Assays, J.B. Lippincott, Philadelphia, 1971, pp. 325-338. 31 Bevis, P.J.R. and Thorndyke, M.C., Stimulation of gastric enzyme secretion by porcine cholecystokinin in the ascidian Styela clara, Gen. Comp. Endocrinol., 45 (1981) 458-464. 32 Thorndyke, M.C. and Bevis, P.J.R., CCK-like and secretin/VIP-like activity in protochordate gut extracts. In Proceedings of the 4th Int. Symp. on Gastrointestinal Hormones, Stockholm, Sweden, June 1982, p. 58 (Abstract). 33 Takeuchi, K., Speir, G.R. and Johnson, L.R., Mucosal gastrin receptor. I. Assay standardization and fulfillment of receptor criteria, Am. J. Physiol., 237 (1979) E284-E294. 34 Denucb, J.M., Do crustaceans react to mammalian gastrointestinal hormones? Preliminary experiments with cholecystokinin-pancreozymin (CCK-PZ) and secretin, Arch. Int. Physiol. Biochim., 90 (1982) B20. 35 Julliard, J.H., Shibasaki, T., Ling, N. and Guillemin, R., High-molecular-weight immunoreactive fl-endorphin in extracts of human placenta is a fragment of immunoglobulin G, Science, 208 (1980) 183-185. 36 Dockray, G.J., Cholecystokinins in rat cerebral cortex: Identification, purification and characterization by immunochemical methods, Brain Res., 188 (1980) 155-165. 37 Dockray, G.J., Williams, R.G. and Zhu, W., Development of region-specific antisera for the C-terminal tetrapeptide of gastrin/cholecystokinin and their use in studies of immuno, eactive forms of cholecystokinin in rat brain, Neurochem. Int., 3 (1981) 281-288. 38 Dockray, G.J. and Walsh, J.H., Amino terminal gastrin fragment in serum of Zollinger-Ellison syndrome patients, Gastroenterology, 68 (1975) 222-230.