Effects of GTP analogs and metal ions on the binding of neurotensin to porcine brain membranes

Peptides,Vol. 14, pp. 37-45, 1993

0196-9781/93 $6.00 + .00 Copyright © 1993 Pergamon Press Ltd.

Printed in the USA.

Effects of GTP Analogs and Metal Ions on the Binding of Neurotensin to Porcine Brain Membranes R. E. C A R R A W A Y , l S. P. M I T R A

AND

T. W. H O N E Y M A N

Department o f Physiology, University o f Massachusetts Medical Center, 55 L a k e Avenue North, Worcester, M A 01655 R e c e i v e d 26 M a y 1992 CARRAWAY, R. E., S. P. M1TRA AND T. W. HONEYMAN. Effects oJ"GTP analogs and metal ions on the binding of neurotensin to porcine brain membranes. PEPTIDES 14(1) 37-45, 1993.--Using ~25I-labeled neurotensin (NT), porcine brain membranes were found to contain two types of high-affinity receptors, one class ( ~ J/3 of total) with an apparent Kd of 0.12 nM and another with an apparent Ko of 1.4 riM. Nonhydrolyzable analogs of GTP inhibited NT binding in a dose-dependent manner. In the presence of 60 #M guanosine 5'-(3-thio) 5'-(/3,3~-imino)triphosphate, NT binding was decreased by 35% with an associated decrease in the number of binding sites and little change in the Ko. Cross-linking of ~251-1abeled NT to brain membranes using disuccinimidyl suberate was found to specifically label two substances of ~ 120 kDa and ~ 160 kDa, which could represent different binding proteins or complexes. For a series of NT analogs, there was close agreement between the IC50 in the binding assay and the EDso in a bioassay based on ability to contract the guinea pig ileum. In addition, metal ions inhibited NT binding and the contractile action of NT with the same order of potency (Hg ++ > Zn ++ > Cu ++ > Mn ++ > Mg++ > Li++). There was a linear relationship between the standard reduction potential for these ions and the logarithm of the IC5o in the binding assay. The results suggest that porcine brain contains high-affinity, G-protein-linked receptors" for NT, the functioning of which depends upon group(s), perhaps sulfhydryl(s), which can interact strongly with certain heavy metal ions. Neurotensin

GTP analogs

Metal ions

Porcine brain membranes

N E U R O T E N S 1 N (NT) is a 13 amino acid peptide found within neurons of the central and peripheral nervous systems, as well as in endocrine cells of the pituitary, adrenal, and small intestine [for review (8,16,26)]. Receptors specific for N T have been demonstrated in CNS tissue from a number of mammals, including rat (15), guinea pig (23), mouse (19), cow (21 ), and human (14). Other studies have investigated the properties of N T receptors on gastrointestinal smooth muscle from rat (17), pig (25), guinea pig ( 11 ), and dog ( 1). Although there is disagreement concerning the molecular weight of the receptor, which varies from 52-1 l0 kDa (19-21,22), the overall concensus of these studies indicates that both high-affinity (Kd, 0.05-0.26 riM) and low-affinity (Kd, 2.3-20 nM) N T receptors exist [for review (l 6)]. Recent cloning of the rat N T receptor has shown that it is a 424 amino acid protein with seven transmembrane segments, as is typical of G-protein-linked receptors (28). Although there are several reports suggesting that N T stimulates phospholipid hydrolysis, inhibits cyclic A M P formation, and elevates cyclic G M P levels in a neuroblastoma cell line (3,4,26) and an adenocarcinoma cell line (2,5), there is little evidence for the specific involvement of G-protein(s), especially regarding the binding of

N T to membranes from normal tissues. In fact, one report has indicated no significant influence of guanyl nucleotides or divalent cations on the binding of N T to rat brain membranes (12). In addition, pretreatment of rat brain membranes with pertussis toxin, under conditions shown to cause ADP-ribosylation of G-proteins, did not attenuate the ability of N T to modulate dopamine D2 agonist binding (29). On the other hand, the suppressive effect of NT, given intracerebroventricularly to rats, on the baroreceptor reflex was antagonized by pretreatments with pertussis toxin and N-ethylmaleimide, both of which disrupt G-protein/receptor interactions (10). Thus, the question as to whether N T receptors in brain are coupled to G-proteins is still an open issue. In order to gather additional data for species comparisons, we have studied the binding of N T to porcine brain membranes and here we report the presence of two classes of N T binding sites that are differentially affected by G T P analogs. In addition, we present the finding that certain metal ions inhibit both the binding and biologic activity of N T in a m a n n e r suggesting important functional role(s) for electron-donating group(s) within the N T receptor(s) or associated protein(s).

Requests for reprints should be addressed to Robert E. Carraway.

37

38

CARRAWAY, MITRA AND HONEYMAN METHOD

TABLE 1

lodination of NT

PARAMETERS FOR NT BINDING TO PORCINE BRAIN MEMBRANES

Iodination of NT was performed using chloramine-T as previously described (9) and the resulting equimolar mixture of JzsI-[Tyr3]NT and ~25I-[Tyrll]NT was purified by HPLC on #Bondapak C-18 to a specific activity of 1000-2000 cpm/fmol as determined by radioimmunoassay. The column (3.9 X 300 mm) was eluted at 1.5 ml/min using a linear gradient (60 min) from 0.1% trifluoroacetic acid to this buffer made 50% in CH3CN. Neurotensin and 125I-NT eluted at ~ 3 4 and 40 min, respectively.

Preparation of Porcine Brain Membrane Fresh porcine cerebral cortex was obtained from 3-monthold Yorkshire pigs killed with an overdose of Nembutal (200 mg/kg). Membranes were prepared using a modification of the method described by Mills et al. (21) for bovine brain. The following buffers were used: buffer A (10 m M Tris-HC1, 20 uM bacitracin, 1 mh//benzamidine-HC1, 1 m M tosylphenyl-chloromethyl-ketone, I m M phenylmethylsulfonyl fluoride (PMSF), 1 m M ethylenediamine tetraacetic acid (EDTA), 1 m M 1,10phenanthroline, 0.01% soybean trypsin inhibitor, pH 7.5); buffer B (buffer A with 0.3 M sucrose). Fresh tissue (25 g) was minced with scissors, suspended in 75 ml buffer B, and then homogenized on ice using a Polytron at high speed for 2-3 min. Fresh PMSF (0.5 M in DMSO) was added during homogenization to 1 m M final concentration. Another 75 ml buffer B was added, the suspension was filtered through cheese cloth, and the filtrate was centrifuged at 1500 × g, 10 min, 4°C using a Beckman centrifuge

Site

Kd* (phi)

Bm,xt (fmol/mgprotein)

High affinity Low affinity

119 _+ 13 1412 _+284

139 _+22 441 _+68

* Data represent the mean _+ SEM for seven displacement experiments involving concentrations of NT in duplicate. t Data represent the mean _+SEM for membranes obtained from four different animals.

with J..17 rotor. The supernatants were removed and stored on ice while the pelleted material was rehomogenized and centrifuged as above. The pooled supernatants were then centrifuged at 30,000 × g, 40 rain, 4°C, and the pelleted material was resuspended in 150 ml of ice-cold buffer A. After recentrifugation at the same speed, the pelleted material was resuspended in 150 ml of ice-cold distilled water and stirred in the cold room for 30 min using a magnetic stirrer. This hypotonic shock served to rupture NT-containing vesicles, thus facilitating removal of endogenous NT from the preparation. The membranes were again centrifuged at 30,000 × g, 20 min, 4°C, and washed twice with buffer A. The pelleted material was resuspended in 100 ml buffer B and 10-ml aliquots were frozen using liquid nitrogen and stored at -80°C. Protein was determined using a modified Folin-Lowry assay (13).

Binding Assay A 100

2 0 ° C "~ e - - - - - - - - ~ e

.__~_ ~ - e

.q

s E

50

0

z_ r~ z

' _u

e'o

[3

'

60

T I M E (mln)

(.1

ILl a. U)

lOO

> ..I u.I n-

50

/ , e-'7"°

, °

, " e "-,7-,-

,

pH

FIG. 1. Time course (A) and pH dependence (B) of the binding of ]251NT to porcine brain membranes. See the Method section for details. Data are means of duplicate determinations differingby < 10%.

The following buffers were used: buffer C (10 m M Tris-HC1, pH 7.5); buffer D (10 m M Tris-HC1, 20 ~zM bacitracin, 1 mM 1,10-phenanthroline, 1 m M benzamidine-HCl, and 1 mM PMSF, added just before use). The frozen membranes (10 ml) were thawed rapidly with frequent shaking, washed once with 40 ml buffer C, and then suspended in 100 ml buffer C using a magnetic stirrer for 30 min at 4°C. The membranes were collected by centrifugation at 30,000 × g, 15 min, 4°C, then washed three times using 40 ml buffer C and finally suspended in 20 ml buffer D on ice. The binding assay was performed as described in previous studies with rat brain (6). In brief, ~251-NT (100,000 cpm) was incubated while shaking for 1 h at 22°C with an aliquot of the membrane preparation (200 ul, ~0.1 mg protein) and varying amounts of unlabeled NT in a total volume of 1.0 ml buffer D. Nonspecific binding was determined in the presence of 0.5 uM NT. The mixtures were then filtered through Whatman GF/C filter strips (presoaked in a solution of 0.3% polyethyleneimine and 0.1% BSA) using a Brandell Cell Harvester and washed three times with 5 ml of ice-cold 10 m M Tris-acetate (pH 7.4). The filter strip was then punched out and counted using a 16-well 3,-counter (Nuclear Enterprises, Scotland). When the effects of metal ions were tested 1,10-phenanthroline was omitted from buffer D. When testing the effects of GTP analogs, binding was performed in buffer D 6ontaining 0.6 m M MgCI2. Binding data were analyzed by nonlinear regression using a commercial program (NFIT, Island Products, Galveston, TX). Estimates for the number and apparent affinity of binding sites were calculated for ligand binding to a single and multiple classes of independent binding sites. The goodness of fit to these models was evaluated by performing an F test, and a p value of < 0.05

NEUROTENSIN RECEPTOR BINDING 300

39 //

A

30 Broom

,.,\

250 ~o 20

200

lO

o

150

o

SITE I 90

SITE 2 92

, 200

250

",

x

E

(fMoles/mg)

i";:. 50

.... -,-100

150

300

Z O tn

100

50

0 0

I

I

I

I

I

100

200

300

400

500

/f/l'w

I

600

5050

NT CONCENTRATION (pM)

70

ff

/4"

B 60 A

50

E

"~ +o 20 Q

z

8mox ( f M o l e l / ~ ) i

30

Kd ( ~ ) * 134

15

0

10

20

$

10 0 10 20 30 40 50 60 B~UNO ( f M o l e s / m 9 ) 0

I

I

I

100

200

300

70 I

400

500

600

1500

NT CONCENTRATION (pM)

FIG. 2. (A) Dose dependence of the displacement of 12~I-NTbinding to brain membranes by unlabeled NT. Each value is the mean of duplicate determinations from a representative experiment that was replicated seven times. Inset: Scatchard plot of the same data. (B) Binding of ~25I-NT to brain membranes as a function of ~25I-[Tyr3]NT concentration. Each value is the mean of duplicate determinations from a representative experiment that was replicated three times. Inset: Scatchard plot of the same data.

was accepted as indicating the multiple-site model fit the data significantly better than a single-site model.

Peptide Synthesis Solid phase peptide synthesis was performed using a manual R A M P S system (Dupont, Boston, MA) and F m o c amino acids

as described previously (7). After cleavage from the resin, each peptide was purified to homogeneity by reverse-phase H P L C using a Delta Pak C-18 column (19 × 300 m m ) as described (7). A m i n o acid analysis verified the presence of integral molar ratios of the expected amino acids for each peptide. Neurotensin, NMN, and XP were obtained from Peninsula (Belmont, CA).

40

CARRAWAY, MITRA AND HONEYMAN

i

I A-DOSE RESPONSE

100

~"~'s

•

T5

o ;T

GppNHp GTPI~S

50 25

IR"

•

0

;,

I

I

0.5 1.0

i

I

I

10,000

/

B-DISPLACEMENT CURVE

i

I

10 100 1,000 CONCENTRATION (pM) 251 ~

q CONTROL Kd-104 pM ,, slte$=118 fmol/mg

GppNHp Kd=120 pM sltes-80 fmol/mg

Z~ 75 O '~"

•

50

X~~~UND

Q. ~~ILI U,I

25

(pM)

0

1.0

10 100 1,000 NT CONCENTRATION (pM)

10,000

FIG. 3. Effect of GTP analogs on ~25I-NT binding to brain membranes. (A) Dose dependence of the inhibition of ~2Sl-NT binding by Gpp(NH)p (circles) and GTP3,S (triangles). Data are means of duplicate determinations differing by < 10% and representative of five separate experiments. (B) Displacement curves for '251-NT binding in the presence (open circles) and absence (closed circles) of 60 uM Gpp(NH)p. Inset: Scatchard plots of the same data.

Bioassay The contractility of freshly dissected guinea pig ileum was examined as previously described (7) using a 5-mljacketed organ bath with Tyrode's solution. Tissue was obtained from halothaneanesthetized guinea pigs and 2-cm sections were mounted under a tension of 1 g. Both ends of the gut were ligated so that mucosal secretions did not enter the bathing fluid. Samples of 10-200/zl were added to the bath at 10-min intervals with two washes in between to avoid desensitization. Histamine (0.1-1 # M ) and carbachol (0. l-1 ~tM) were used as standards. While dose-response curves for these standards were similar throughout the 3-4 h of testing, the sensitivity of the tissue to N T often increased

more than twentyfold over the first 90 min, perhaps due to receptor desensitization. Therefore, the results reported here were obtained after the mounted tissue had been exposed to carbachol (0.2/~M) and washed at 15-min intervals for 100 min. For experiments in which the effects of metal ions were tested, the following protocol was used. Copper sulfate, a l u m i n u m sulfate, and chloride salts of the other metals, purchased from Fisher Scientific (Boston, MA), were dissolved in water and administered to the bath in 10-200-#1 aliquots. Testings were performed until reproducible responses to 0.4 n M N T and 0.1 # M histamine were obtained. The metal ion was added to the bath at the lowest concentration to be tested, and after 5 min, 0.4 n M N T was tested and the tissue washed twice with buffer. The metal ion

NEUROTENSIN RECEPTOR BINDING

41

A

1o0 8 "5

so

o z

6o

z m

S~

o W m.

4O 2o

0.000,

0.001

0.01 '

0:1

1:0

1()

' 100

1000

SALT CONCENTRATION (raM) +1.0 B

• Cu++

~

-e

_ __ r-u.uu

• Fo+++

• ~ Cd++

o

nZ

Pb+÷

• Co++

-1.0

• Zn++

e

~

+

AI+++

-2.0 Z (a

-3.0 ,.**,

,.,,

,.,

,.o

1,

1,0

IC60(mM)

FIG. 4. Dose response (A) for the inhibition of 1251-NTbinding to brain membranes by varioussaltsand correlation(B) betweenthe IC5oobservedand the standard reduction potential of the metal ion involved. Data represent mean of duplicate determinations differing by < 10%.

was added at a higher concentration and testings continued until the IC50 was obtained. At this time, the responsiveness of the tissue to 0.1 u M histamine was determined.

Cross-Linking 125I-Labeled N T to Receptor(s) Chemical cross-linking using disuccinimidyl suberate (DSS) was performed as follows. Membranes (0.5 mg protein/ml in 10 ml of 10 m M Hepes buffer, pH 7.5) were incubated with J2sIlabeled NT (106 cpm/ml) in the presence and absence of 1 ~M NT for 1 h at 22°C. Then the membranes were washed three times in ice-cold buffer (10 ml) and the resuspended membranes were then reacted for 1 min on ice with 1 m M DSS (diluted from 0.1 M stock in DMSO). The reaction was stopped by addition of 0.1 M ammonium acetate. After washing three times,

the incorporation of radioactivity was seen to be 20-30%. The membranes were then reduced with 5% mercaptoethanol and analyzed by SDS-polyacrylamide electrophoresis (7.5% acrylamide slab gels). The radiolabeled bands were detected by autoradiography of the dried gel. RESULTS

Characteristics of the Binding Assay Under the conditions of the assay (50-100 pM 125I-NT and 30-100 #g membrane protein/ml), equilibrium was achieved after ~ 3 0 min at 20°C and after ~ 1 2 0 min at 4°C, giving 36% of the added counts in the specifically bound fraction (Fig. 1A). Nonspecific binding, determined in the presence of 0.5 #M NT, was consistently 5-15% of total binding. As assessed by

42

CARRAWAY, MITRA AND HONEYMAN TABLE 2

HPLC, the 125I-NTwas not degraded significantly during a 60rain incubation with membranes at 22°C (recovery: control, 78 +_ 3%; with membranes, 77 +_ 5%; n = 3). Binding was pH dependent, exhibiting an optimum near pH 7 and falling offmore steeply on the alkaline side (Fig. 1B).

COMPARISON OF IC~o VALUES IN THE BINDING ASSAY TO ED~o VALUES IN THE BIOASSAY FOR NT AND RELATED PEPTIDES

Scatchard Analysis

Compound #

Peptide

l 2 3 4 5 6 7 8 9 t0 11 12 13 14 15 16 17

ZLYENKPRRPYIL (NT) KIPYIL (NMN) ZGKRPWIL (XP) FHPKRPWIL (XP-1) HPKRPWIL (XP-2) XPRRPYIL (NT6-13) YKKRRPY1L YKKKRRPYIL dKYPRRPYIL RRPYIL RRIYIL RRVYIL RRPDIL RRPYEL RRPYYL RRPYIY RRPYYY RRPYIL-amide

Addition of increasing amounts of unlabeled NT to a constant amount of ~25I-NT and membrane generated the binding saturation curve shown in Fig. 2A. Scatchard analysis (Fig. 2A, inset) indicated that the data could be best fit by assuming the presence of two classes of binding sites, one with relatively high affinity (K~, ~0.2 nM) and another with lower affinity (Ko, ~ 1.4 nM). Similar data obtained from seven experiments are summarized in Table 1. The high-affinity binding was examined more closely by addition of increasing amounts of ]25I-NT to membranes (Fig. 2B and inset). The Kd for this experiment, 0.13 nM, was near to that obtained in Fig. 2A, indicating that the affinity of iodinated NT for the high-affinity sites was similar to that for NT. Owing to the large amount of labeled material required, we could not examine the lower-affinity binding in this manner.

Effect of GTP Analogs The nonhydrolyzabte analogs of GTP, GTP3,S [guanosine 5'-(3-thio) triphosphate], and Gpp(NH)p [guanosine 5'-(/3,3'imino) triphosphate] decreased NT binding in a dose-dependent manner (Fig. 3A). In the presence of 0.6 m M Mg++, 60 uM Gpp(NH)p decreased NT binding by ~35% (Fig. 3B), and Scatchard analysis of the displacement curve (Fig, 3B, inset) indicated a 32% decrease in the number of high-affinity binding sites without a significant change in the Kd [control, 104 pM; Gpp(NH)p, 120 pM]. When assayed with the same receptor preparations, 3H-QNB (quinuclidinyl benzilate) binding to the muscarinic receptor was decreased ~ 10% by 20-100 u M concentrations of Gpp(NH)p or GTP3,S (data not shown).

18

IO

0_ o

80

~ 0 z_

6o

-~

4o

13 ILl a. m

2o

0.()1

[Val ]-NT \• ,, \,,

,

$-

IS -

IS

8 - 13

[Tyr -]-NT -+~<.

-/-

/

9

0.3 2.0 0.8 1.0 1.5 0.2 0.3 0.25 0.15 1.4 > 10,000 > 10,000 > 10,000 > 10,000 >10,000 > 10,000 > 10,000 40

directly related to the valency of the cation (Fig. 4A). Instead, for the series of cations investigated, we found a linear relationship (r = 0.86) between the standard reduction potential and the logarithm of the IC50 (Fig. 4B). Thus, metal ions with a stronger tendency to accept electrons were more effective inhib-

In general, increasing concentrations of salt decreased the binding of NT to membranes; however, ICs0 values were not

lOO . . . . . . . . . . . . .

0.2 1.6 0.5 0.6 1.0 0.1 0.1 0.15 0.15 0.2 250 100 900 1000 104) 40 1100 3

Bioassay ED~ (riM)

Given are mean values determined from at least two experiments involving duplicate testings at multiple concentrations. Peptide structures are given using the single letter code for amino acids: Z, pyroglutamic acid; L, leucine; I, isoleucine; Y, tyrosine; W, tryptophan; V, valine;F, phenylalanine;P, proline; R, arginine;K, lysine; H, histidine: G, glycine;E, glutamic acid; D, aspartic acid.

Effect of Salts

o

Binding IC~ (nM)

•

....

- - --_<,

---

"\~tTyrl-NX

~)

0.1

1.0

1

PEPTIDE

CONCENTRATION

, 100

1000

10,000

(nM)

FIG. 5. Log dose-response relationships for the inhibition of J25I-NTbinding to brain membranes by various analogs of NT. Data are means of duplicate determinations differing by <10%.

NEUROTENSIN RECEPTOR BINDING

A

B

C

43

D

200-* #2 ~ P '

9 2 -*

shown in Fig. 5. From these and similar data, IC50 values were determined and compared to ED~0 values obtained for the contractile effect of these peptides on the isolated guinea pig ileum (Table 2). Although potencies of the peptides varied over a ten thousandfold range, the values for the two assays were in fair agreement. For the naturally occurring analogs (compounds 15, Table 2), IC50 and EDs0 values were nearly equal. Peptides with increased positive charge near to position 6 in NT (compounds 6-9, Table 2) were slightly more potent than NT in both assays. Analogs of NT(8-13) (compound 10) with alterations in positions 10-13 (compounds 11-18, Table 2) were strikingly less potent in both assays. Also note that the EDs0 for NT in the bioassay (~0.3 nM, Table 2) was similar to the K~ obtained for the high-affinity component of the receptor binding (~0.1 nM, Table 1).

Molecular Weight of the Receptor

68-*

46-*

22..*

FIG. 6. Specific labeling of brain membranes with ~251-NTusing crosslinking reagent, DSS. Membranes were incubated with ~25I-NT_+ 1 uM unlabeled NT. Unbound peptide was removed by centrifugation and membranes were reacted with 1 mM DSS for 1 min as described in the Method section. Membraneswere then solubilizedand subjectedto SDSPAGE using 7.5% acrylamide under reducing conditions and labeled proteins were visualizedusingautoradiography. Lanes A and B, duplicate samples minus added NT: lanes C and D, duplicate samples with added NT.

itors of NT binding. Mercuric salts, which are known to readily react with sulflaydryl groups, were the most potent. Metal ions also inhibited the contractile action of NT on the isolated guinea pig ileum. The order of potency, Hg ++ > Zn ++ > Cu ++ > Mn ++ > Mg ++ > Li +, was the same as that shown for NT binding. The IC5o for each of the above ions (/~M total, not corrected for binding to buffer constituents) was 0.005, 0.03, 0.06, 0.2, 0.8, and 60, respectively. The contractile effect of histamine (0.1 #M), tested after 5-min exposure to these ion concentrations, was inhibited <15%, except in the case of Hg ++ (20%).

Effect of NT Analogs Dose-response curves for the inhibition of receptor binding by NT and various representative synthetic analogs of NT are

Chemical cross-linking of ~25I-labeled NT to porcine membranes in the presence and absence of 1 uM NT led to the identification of two specifically labeled proteins with apparant molecular weights of 120 kDa and 160 kDa (Fig. 6). These proteins could represent two types of NT receptor or the same receptor with and without an associated protein. DISCUSSION The time- and temperature-dependent binding of ~25I-labeled NT to porcine brain membranes displayed properties suggesting that these preparations contained high-affinity receptor(s) that were specific for NT. That these represented functional receptors was suggested by the fact that binding and biologic activity exhibited similar dependencies on specific groups within NT and were similarly inhibited in the presence of certain metal ions. Although there are other possibilities, the unusual sensitivity to mercuric ion suggested the involvement ofthiol group(s) in the binding process and this has been substantiated by additional experiments. [Mitra, S. P.; Carraway, R. E. Importance ofthiol group(s) in the binding of ~25I-labeledneurotensin to membranes from porcine brain. Submitted.] The binding of NT to porcine brain membranes suggested the existence of two classes of binding sites, one with high affinity (Kd, 0.12 nM) and one with lower affinity (Kd, 1.4 nM). The apparent Kd values are similar to those reported for the binding of ~25I-NT to brain membranes from mouse [Kds, 0.13 and 2.3 nM(19)], rat [KoS, 0.18 and 5.1 n M (15)], guinea pig [Kds, 0.16 and 6.3 n M (23)], and human (K~s, 0.26 and 4.3 nM). The use of tritiated NT has revealed the presence of only one class of binding sites in the brain membranes tested [bovine: Kd, 3.3 n M (21); human: Kd, 2.0 n M (14)], except in a recent study by Miyamoto-Lee and coworkers [Kds, 0.13 and 20 n M (22)]. Whether both or only one of these sites represent functional receptors is not yet clear because the available bioassays based on central effects of NT, e.g., hypothermia, analgesia, and increased dopamine turnover, do not permit one to accurately determine the effective NT concentrations at the receptor(s). However, an increase in inositol phospholipid metabolism and elevation in intracellular Ca +2 (26), a stimulation of cyclic GMP synthesis (3), and an inhibition of cyclic AMP synthesis (4) in response to NT have been measured in cultured murine neuroblastoma cells (clone N 1E-115). If one corrects the EDs0 values determined for these responses, taking into consideration the fact that some responses occur before equilibrium is achieved, then the data appear to be consistent with mediation by a receptor having a Ka of ~ 1-2 n M (4). These and other findings have lad

44

CARRAWAY, MITRA AND HONEYMAN

some investigators to suggest that the low-affinity NT binding sites mediate the biologic actions of this peptide (17). In the present study, we have shown that the EDs0 for the contractile effect of NT on guinea pig ileum is ~0.3 riM. Since this response occurs and fades within ~ 2 min, there is likely to be insufficient time for the added NT to equilibrate across the tissue. In addition, degradation within the tissue space will further decrease the concentration of NT at the receptors. Therefore, the actual EDso for this effect is likely to be considerably less than that measured. Preliminary experiments in our laboratory have shown that guinea pig ileal muscle contains high- and lowaffinity NT receptors similar to those described here for porcine brain (Kas, 0.09 n M and 2.4 nM, respectively). [Carraway, R. E.; Mitra, S. P. Implication of two receptor types, one releasing acetylcholine (Kd, 0.15 nM) and one substance P (Kd, 2.5 nM) which mediate the contractile effect of neurotensin in guinea pig ileum. Abstract to Symposium on Gastrointestinal Hormones, Leuven, Belgium, September 1-5, 1992.] Thus, it seems probable that, in this case, the high-affinity sites mediate the contractile effect of NT on the guinea pig ileum. As to the molecular nature of the receptor types observed here, it is clear that the high-affinity form (Ko, 0.12 nM) is sensitive to GTP analogs and, therefore, could be G-protein linked. However, while treatment with Gpp(NH)p decreased the number of high-affinity (Kd, 0.12 nM) sites, it did not lead to a concomitant increase in the low-affinity (Kd, 1.4 nM) component. Thus, it does not appear that the high- and low-affinity sites merely represent G-protein coupled and uncoupled forms of the same receptor. One possibility is that the G-protein uncoupled form had an even lower affinity and was not observed in these experiments. Interestingly, when we cross-linked ~2SI-NTto porcine brain membranes, two proteins (Mr, 120 and 160 kDa) were specifically labeled (Fig. 6). It is possible that the smaller protein is the NT receptor and the larger one represents the same protein cross-linked to a 40-45 kDa G-protein. Alternatively, the two forms could represent two different receptors or differential degradation of a single receptor. The possible existence of two different NT receptors is supported by the finding that levocabastine can distinguish between high- and low-affinity receptor forms in rat brain membranes (24). The effects of metal ions on the NT receptor and on the contractile action of NT in the isolated ileum could be very complex. The observed correlation between the potencies of the metal ions in inhibiting binding and their standard reduction

potentials suggests that the metal's avidity for electron-donating atoms is an important determinant of the interaction. In general, the order of binding ability to transition metals for amino acid side chains is Cys > His > Asp/Glu. The fact that there are an uneven number of Cys residues in the cDNA-predicted sequence of the NT receptor (28) and the extreme potency of mercuric ion, which is known to bind strongly to free sulfhydryls, suggests that a Cys residue is likely to be one site of metal interaction. The inhibitory effects of metal ions on the biologic response to NT were qualitatively in agreement with the results obtained with membrane receptor. However, it was difficult to perform a quantitative comparison due to differences in the buffers employed. We could not use the same medium for both assays since the physiolgic buffers required to study smooth muscle contraction contain sodium, potassium, calcium, and magnesium ions, which interfere in the binding assay. Furthermore, some of the metals under study are known to form strong complexes with other constituents, such as phosphate ion. In spite of these limitations, the results obtained by bioassay indicated that the tested metal ions inhibited the contractile effect of NT without greatly decreasing the response to histamine at that time and that the order of potency was similar to that obtained in the binding assay. The structure-function studies performed here indicated that the ICs0 values determined in the binding assay were in fair agreement with the EDs0 values obtained in the bioassay for quite a few analogs of NT (Table 2), except for compounds 1018. Compound 10 [NT(8-13)], for example, was equipotent with NT in the binding assay but only 25% as potent as NT in the bioassay. This discrepancy may be due to differential degradation of this peptide that, unlike NT, has a free N-terminus. Interestingly, analogs of NT(8-13) (compounds 11-18) also exhibited disparity in these two assays. In summary, our analyses indicate that porcine brain membranes contain high-affinity, G-protein-linked NT receptors, which display an affinity for NT, a structural specificity and a sensitivity to metal ions that resemble those shown for functional NT receptors in the guinea pig ileum. ACKNOWLEDGEMENTS This work was supported by National Institutes of Health Grant DK 28565. Excellent technical help was provided by Rebecca Salmonsen.

REFERENCES 1. Ahmad, S.; Berezin, I.; Vincent, J. P.; Daniel, E. E. Neurotensin receptors in canine intestinal smooth muscle: Preparation of plasma membranes and characterization of(Tyr3-t25I)labelled neurotensin binding. Biochem. Biophys. Acta 896:224-238; 1987. 2. Amar, S.; Kitabgi,P.; Vincent,J-P. Activationofphosphatidylinositol turnover by neurotensin receptors in the human adenocarcinoma cell line. FEBS Lett. 201:31-36; 1986. 3. Amar, S.; Kitabgi,P.; Vincent, J-P. Stimulation ofinositol phosphate production by neurotensin in neuroblastoma NI El 15 cells: Implication of GTP-bindingproteins and relationshipwith the cyclicGMP response. J. Neurochem. 49:999-1006; 1987. 4. Bozou, J-C.; Amar, S.; Vincent, J-P.; Kitabgi P. Neurotensin-mediated inhibition of cyclicAMP formation in neuroblastomaN 1E115 cells: Involvement of the inhibitory GTP-binding component ofadenyl cyclase. Mol. Pharmacol. 29:489-496; 1986. 5. Bozou, J-C.; Rochet, N.; Magnaldo, 1.; Vincent, J-P.; Kitabgi, P. Neurotensin stimulates inositol triphosphate-mediatedcalcium mobilizationbut not protein kinaseC activationin HT29 cells. Biochem. J. 264:871-878; 1989.

6. Carraway, R. E.; Mitra, S. P.; Cochrane, D. E. Structure of a biologically active neurotensin related peptide obtained from a pepsin treated albumin(s). J. Biol. Chem. 262:5968-5973; 1987. 7. Carraway, R. E.; Mitra, S. P.; Paradise, C. Characterization of large neuromedin-N using antisera towards regions of the neurotensin/ neuromedin-N precursor. Peptides 12:601-607; 1991. 8. Ferris, C. F. In: Makhlouf, G. M., ed. Handbook of physiology, the gastrointestinal system, vol. II. Bethesda, MD: American Physiological Society; 1989:559-586 9. Feurle, G. E.; Carraway, R. E.; Rix, E.; Knauf, W. Evidence for the presence ofxenopsin related peptides in gastric mucosa of mammals. J. Clin. Invest. 76:156-162; 1985. 10. Fu, M. J.; Lin, K. S.; Chan, J. Y. H.; Chan, S. H. H. Participation of pertussis toxin-sensitiveGTP-binding regulatory proteins in the suppression of baroreceptor reflex by neurotensin in the rat. Regul. Pept. 37:167-180; 1992. 11. Goedert, M.; Hunter, J. C.; Nincowic, M. Evidencefor neurotensin as a non-adrenergic, non-cholinergicneurotransmitter in guinea pig ileum. Nature 311:59-62: 1984.

NEUROTENSIN RECEPTOR BINDING 12. Goedert, M.; Pittaway, K.; Williams, B. J.; Emson, P. Specific binding of tritiated neurotensin to rat brain membranes: Characterization and regional distribution. Brain Res. 304:71-78; 1984. 13. Hess, H. H.; Lees, M. B.; Derr, J. E. A linear Lowry-Folin assay for both water soluble and sodium dodecyl sulfate solubilized proteins. Anal. Biochem. 85:295-300; 1978. 14. Kanaba, K. S.; Kanaba, S.; Okazaki, H.; Richelson, E. Binding of [3H] neurotensin in human brain: Properties and distribution. J. Neurochem. 46:946-952; 1986. 15. Kitabgi, P.; Carraway, R. E.; Rietschoten, J. V.; Granier, C.; Morgat, J. P.; Mendez, A.; Leeman, S. E.; Freychet, P. Neurotensin: Specific binding to synaptic membranes from rat brain. Proc. Natl. Acad. Sci. USA 74:1846-1850; 1977. 16. Kitabgi, P.; Checler, F.; Mazella, J.; Vincent, J. P. Pharmacology and biochemistry of neurotensin receptors. Rev. Clin. Basic Pharmacol. 5:397-486; 1985. 17. Kitabgi, P.; Kwan, C. Y.; Fox, J. E. T.; Vincent, J. P. Characterization of neurotensin binding to rat gastric smooth muscle receptor sites. Peptides 5:917-923; 1984. 18. Lucchesi, P. A.; Romano, F. D.; Scheid, C. R.; Yamaguchi, H.; Honeyman, T. W. Interaction of agonists and selective antagonists with gastric smooth muscle muscarinic receptors. NaunynSchmiedebergs Arch. Pharmacol. 339:145-151; 1989. 19. Mazella, J.; Chabry, J.; Kitabgi, P.; Vincent, J. P. Solubilization and characterization of active neurotensin receptors from mouse brain. J. Biol. Chem. 263:144-149; 1988. 20. Mazella, J.; Kitabgi, P.; Vincent, J-P. Molecular properties of neurotensin receptors in rat brain: Identification ofsubunits by covalent labeling. J. Biol. Chem. 260:508-514; 1985. 21. Mills, A.; Demoliou-Mason, C. D.; Barnard, E. A. Purification of

45

22.

23.

24. 25. 26.

27. 28. 29.

the neurotensin receptor from bovine brain. J. Biol. Chem. 263:1316; 1988. Miyamoto-Lee, Y.; Shiosaka, S.; Tohyama, M. Purification and characterization of neurotensin receptor from rat brain with special reference to comparison between newborn and adult age rats. Peptides 12:1001-1006; 1991. Sadoul, J. L.; Mazella, J.; Amar, S.; Kitabgi, P.; Vincent, J. P. Preparation of neurotensin selectively iodinated on the tyrosine-3 residue: Biological and binding properties. Biochem. Biophys. Res. Comrnun. 120:812-819; 1984. Schotte, A.; Leysen, J. E.; Laduron, P. M. Evidence for a displaceable non-specific (3H)-neurotensin binding site in rat brain. Naunyn Schmiedebergs Arch. Pharmacol. 333:400-405: 1986. Seybold, V. S.; Treder, B. G.; Aanonsen, L. M.; Parsons, A.; Brown, D. R. Neurotensin binding sites in pordine jejunum: Biochemical characterization and intramural localization. Synapse 6:81-90; 1990. Snider, R. M.; Forray, C.; Pfenning, M.; Richelson, E. Neurotensin stimulates inositol phospholipid metabolism and calcium mobilization in murine neuroblastoma cone N1E115. J. Neurochem. 47: 1214-1218; 1986. St. Pierre, S. A.; Kerouac, R.; Quirion, R. In: Hearn, M. T., ed. Peptide and protein reviews, vol. 2. New York: Marcel Dekker; 1985: 83-171. Tanaka, K.; Masu, M.; Nakanishi, M. Structure and functional expression of the cloned rat neurotensin receptor. Neuron 4:847854; 1990. Von Euler, G.; Van der Ploeg, I.; Fredholm, B. B.; Fuxe, K. Neurotensin decreases the atfinity of dopamine D2 agonist binding by a G-protein-independent mechanism. J. Neurochem. 56:178-183; 1991.