Cholera toxin ADP-ribosylates the receptor-coupled form of pertussis toxin-sensitive G-proteins

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Vol. 165, No. 2, 1989

Pages 554-550

December 15, 1989

CHOLERA TOXIN ADP-RIBOSYLATES THE RECEPTOR-COUPLED FORM OF PERTUSSIS TOXIN-SENSITIVE G-PROTEINS Franz-Josef Klinz* and Tommaso Costa + Department of Neuropharmacology, Max-Planck-Institut fiir Psychiatric Am Klopferspitz 18a, D-8033 Planegg-Martinsried, F.R.G. Received October 9, 1989

SUMMARY: Cholera toxin catalyzes the ADP-ribosylation of 40 kDa pertussis toxin substrates in membranes from NG108-15 cells, which is increased in the presence of the opioid agonist DADLE. The basal ADP-ribosylation can be abolished by the opioid antagonist ICI 174864, suggesting that unoccupied opioid receptors interact spontaneously with the pertussis toxin substrates Gi/Go in the membrane. Treatment of NG 108-15 cells with the opioid agonist DADLE leads to a reduction of agonist-stimulated and basal ADP-ribosylation of 40 kDa substrates catalyzed by cholera toxin. This indicates that the spontaneous interaction between opioid receptors and G-proteins is decreased in membranes of cells in which the receptor was desensitized by prolonged exposure to the agonist. ©19~9Ao~demioP..... ~no. It is known that bacterial exotoxins from Vibrio cholerae.and Bordetella pertussis can catalyze the mono-ADP-ribosylation of guanine nucleotide binding regulatory proteins (G-proteins) in a remarkably specific fashion (1). Cholera toxin preferentially modifies the t~-subunits of the Gsgroup of G-proteins, whereas the Gi/Go-type ~-subunits are selective substrates for the transfer of ADP-ribose catalyzed by pertussis toxin. The site of modification by pertussis toxin is a cysteine residue present in a carboxy-terminal nonapeptide (2), whose sequence is highly conserved in Gisubtypes, Go, and transducins, but absent in Gs (3). Instead the ADP-ribose accepter site of cholera toxin is an arginine residue which is included in a typical consensus sequence present not only in Gs, but also in transducin, Go, and Gi-subtypes. The ~t-subunit of transducin is ADPribosylated by cholera toxin in retinal rod membranes (4) or in purified preparations of transducin reconstituted in rod outer segment membranes (5). Evidences for cholera toxin-catalyzed ADPribosylation of pertussis toxin substrates have been obtained, so far, only in membranes (6-10). This reaction occurs more efficiently in the absence of GTP (7,9,10), and, at least in some

* To whom correspondence should be addressed. Present address: Institut ffir Pharmakologie, Freie Universith't Berlin, Thielallee 69-73, D-1000 Berlin 33, F.R.G. + Present address: Laboratory of Theoretical and Physical Biology, NICHHD, NIH, Bethesda, Md 20892. ABBREVIATIONS: C-'/~, cholera toxin; DADLE, [D-Ala2,D-Leu5]enkephalin; G-protein, guanine nucleotide binding regulatory protein; Gi, inhibitory G-protein; Go, G-protein abundant in brain tissue; Gs, stimulatory G-protein; ICI 174864, [N,N'-diallyl-Tyrl,Aib2,3]I~u-enkephalin; kDa, kilodalton; NLX, naloxone; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. 0006-291X/89 $1.50 Copyright © 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

554

Vol. 165, No. 2, 1989

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

membranes, it is stimulated b y agonist-occupied receptors (9,10). Since the extent of ADP-ribose incorporation into pertussis toxin substrates catalyzed by cholera toxin under basal conditions is only a fraction of that induced by pertussis toxin, we considered that this subpopulation of cholera toxin-sensitive pertussis toxin substrates may represent the fraction of G-proteins that are spontaneously coupled to receptors in the absence of ligand. There is evidence that in membranes of NG108-15 (neuroblastoma x glioma hybrid) cells unoccupied opioid receptors are tightly precoupled to G-proteins (11). These precoupled forms of receptors are not inactive, but can elicit a basal GTPase activity in the absence of ligand, as suggested by the finding that some types of competitive antagonists can inhibit basal GTPase activity in a stereospecific fashion (negative antagonists) (12). We show in this paper that the basal level of cholera toxin-catalyzed transfer of ADP-ribose into 40 kDa pertussis toxin substrates can be virtually abolished in the presence of the negative antagonist ICI 174864, and that this effect can be prevented in a stereospecific fashion by naloxone enantiomers. We further show that in membranes obtained from NG108-15 cells, that had been exposed to an opioid agonist to desensitize and down-regulate opioid receptors, both agonist-stimulated and basal cholera toxin-catalyzed ADP-ribosylation of pertussis toxin substrates were diminished. Collectively, these data support the idea that cholera toxin selectively ADPribosylates receptor-coupled forms of Gi/Go in membranes.

MATERIALS AND METHODS Culturing of NG108-15 (neuroblastoma x glioma hybrid) cells, membrane preparation and GTPase assay were performed as described previously (13). For cholera toxin-catalyzed ADPribosylation of membrane proteins, cholera toxin (List Biological Laboratories, Campbell, CA, USA) at 2 mg/ml was activated by incubation with an equal volume of 40 mM dithiotreitol for 10 min at 30°C. 50 Ixg membrane proteins were ADP-ribosylated in a total volume of 50 lal containing 100 mM potassium phosphate pH 7.0, 2.5 mM MgC12, 1 mM ATP, 10 mM thymidine, 10 mM arginine, 0.01% BSA, 200 ~tg/ml activated cholera toxin and 2 I~Ci [ct-32p]NAD (800-1000 Ci/mmol; NEN, Dreieich, FRG). Opioid agonists and antagonists were added to the reaction at concentrations indicated in the figure legends. The reaction was allowed to proceed for 120 min at 25oc and stopped by the addition of 500 i11 ice-cold 50 mM Tris-HC1 pH 7.5. The membranes were centrifuged for 10 min at 25000 x g and 4°C. The sediments were dissolved in sample buffer, heated for 5 min at 95°C and separated on 10% SDS-polyacrylamide gels prepared by the method of Laemmli (14). Molecular mass standards were [14C]-methylated proteins (Amersham Buchler, Braunschweig, FRG). After electrophoresis the gel was dried and exposed to Hyperfilm-~max (Amersham Buchler). Analysis of the incorporation of [32p]ADP-ribose into 40 kDa substrates was performed by scanning the autoradiograms with a LKB laser densitometer. Pertussis toxincatalyzed ADP-ribosylation of membrane proteins was performed as reported previously (15).

RESULTS Fig. 1 shows the cholera toxin-catalyzed incorporation of [32p]ADP-ribose into proteins of NG108-15 cell membranes. Two of them (43 and 47 kDa) correspond to the o~-subunits of Gs, of which the higher molecular weight form is predominant. The third band (40 kDa) displays an identical Rf value to the protein labeled by pertussis toxin in the same membrane (Fig. 1, lane 9). In agreement with a previous report (10), the labeling of this band was enhanced in a concentrationdependent fashion when the reaction was performed in the presence of the opioid agonist DADLE (Fig. 1, upper gel). In contrast, the peptidergic opioid antagonist ICI 174864 produced a 555

Vol. 165, No. 2, 1989

kDa 69~

1

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS 3

2

/+

5

6

7

9

8

L,6~

DADLE

30~ 69 IE117t~86L,

30 0

10 -9

10 -8

i

10 -7 10 -6

10 -5 I

concentration of [igand in reaction (M) Fig, 1: Effect of the opioid agonist DADLE (upper gel) and the antagonist ICI 174864 (lower gel) on cholera toxin-catalyzed ADP-ribosylation of pertussis toxin substrates in membranes from NG108-15 cells. 40 ~tg of membrane proteins ADP-ribosylated in the absence (lane 1) or presence of cholera toxin (lane 3-8) and 1 ~tg in the presence of pertussis toxin (lane 9) were analysed by SDS-PAGE. The concentrations of DADLE and ICI 174864 in the cholera toxin-catalyzed ADP-ribosylations are shown. Values on the left indicate the position of molecular mass marker proteins (lane 2) in kDa.

concentration-dependent diminution in the labeling of the 40 kDa substrate (Fig. 1, lower gel). Neither the agonist, nor the antagonist affected the labeling o f G s u-subunits by cholera toxin. Fig. 2 shows that the inhibition of cholera toxin-catalyzed incorporation of [32p]ADP-ribose into the 40 kDa substrate is receptor-mediated. The antagonist (-)naloxone (10 ~tM) did not affect the labeling of Gi/Go. However, when present in the reaction simultaneously with ICI 174864 (1 ~tM), the inhibitory effect of the negative antagonist was completely prevented. The inactive enantiomer (+)naloxone used at the same concentration did not prevent the inhibitory effect of ICI 174864. Thus, ICI 174864 affects the cholera toxin-mediated transfer of ADP-ribose into Gi/Go via a stereospecific interaction with the opioid receptor, and not by a nonspecific influence on the enzymatic reaction catalyzed by the toxin or exerted on the G-protein itself. We showed previously that exposure of intact NG108-15 cells to the opioid agonist DADLE leads to a time-dependent loss of receptor-mediated stimulation of GTPase activity (13). To examine the effect of agonist-induced loss of receptor responsiveness on cholera toxin-catalyzed ADP-ribosylation of Gi/Go, we studied membranes prepared from cells that had been exposed to DADLE (100 nM) for different times. As shown in Fig. 3, the incorporation of ADP=ribose into the 40 kDa band was significantly reduced as the time of agonist exposure increased. This diminution of cholera toxin-mediated ADP-rihosylation was apparent both in the presence and absence of DADLE in the reaction, indicating that both the basal and the agonist-stimulated labeling of Gi/Go 556

Vol. 165, No. 2, 1989

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

'2

1

kDa

3

14

: 6

5 :

7~

8

9

69-" 46--

30-" MW No ICt (-)NLX (+)NIX ICI ICI Stds Ligand 1/AVl lO/Ak,1 IO/AVl (-)NIX ( + )NIX I

I

No

Cholera Toxin

Pertussis Toxin

Toxins

Fig. 2: Effect of opioid antagonists on cholera toxin-catalyzed ADP-ribosylation of pertussis toxin substrates in membranes from NG108-15 cells. 40 ~tg of membrane proteins ADP-ribosylated in the absence (lane 1) or presence of cholera toxin (lane 3-8) and 0.5 rtg in the presence of pertussis toxin (lane 9) were analysed by SDS-PAGE. Lane 3: no ligand; lane 4:1 taM ICI 174864; lane 5:10 vM (-)naloxone; lane 6:10 paM(+)naloxone; lane 7:1 pM ICI 174864 + 10 pM (-)naloxone; lane 8:1 laM ICI 174864 + 10 pM (+)naloxone. Values on the left indicate the position of molecular mass marker proteins (lane 2) in kDa.

by the toxin are receptor-dependent. Fig. 4 shows a comparison of the rate of disappearance of cholera toxin-catalyzed incorporation of ADP-ribose into the 40 kDa band (quantified by densitometry of the autoradiogram shown in Fig. 3) and the rate of low-Kin GTPase activity, both measured in the same membrane preparation and in the presence of DADLE. It is clear that both

1

2

3

4

5

6

7

8

9

10

Ciso= ~ 13i~/13o~

Basal CTX-catalyzed ADP- ribosytation

fis= - - I ~ GidGo=--

DADLE-stimulated CTX-catatyzed ADP-ribosylation

0

4

15

60

120

2/,0

/,80

I

0 I

time of DADLE p r e f r e a f m e n f

(min)

Fig. 3; Effect of pretreatment of cultured NG108-15 cells with DADLE on basal (upper gel) and DADLE-stimulated cholera toxin-catalyzed ADP-ribosylation (lower gel) in membranes. 40 rtg of membrane proteins ADP-ribosylated in the absence (lane 1) or presence of cholera toxin (lane 2-9) and 0.5 I~g in the presence of pertussis toxin (lane 10) were analysed by SDS-PAGE. Shown is the time of pretreatment with 100 nM DADLE. Arrows on the left show the position of ADP-ribosylated Gsct- and GiJGoct-species. 557

Vol. 165, No. 2, 1989

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ,~,

/*.7.

i

[] ADP-ribosy[afion • GTPase

t~.3

3.9

3.5

i~o

2~o

3ao

4~o

s~o

time of DADLE prefreafmenf(min)

Fig. 4: Comparison of DADLE-stimulated cholera toxin-catalyzed ADP-ribosylation and GTPase in membranes from DADLE-pretreated NG 108-15 cells. The autoradiogram of the lower gel from Fig. 3 was scanned by densitometer and the reduction in labeling of 40 kDa substrates calculated as percent of control. Shown is the decrease in ADPribosylation 03) and GTPase (0).

processes disappear with very similar rates. From these data the half-time of loss of responsiveness for the fast component of desensitization is 11.1 (+/- 3) min. We found previously that pretreatment of intact NG108-15 cells for 24 h with high concentrations (10 ~tg/ml) of cholera toxin virtually abolishes toxin-catalyzed labeling of 40 kDa substrates in membranes (Klinz and Costa, submitted). This suggests that cholera toxin-catalyzed ADP-ribosylation of Gi/Go can occur in intact cells, despite the high concentrations of GTP present in the physiological intracellular environment. We studied the consequence of cholera toxin treatment of intact cells on the relative abilities o f the opioid agonist D A D L E and antagonist ICI 174864 to stimulate and inhibit low-Km GTPase activity, respectively. As shown in Fig. 5, cholera toxin treatment can diminish both agonist-mediated stimulation and antagonist-mediated inhibition

A

,,

.~

• DADLE

~.~ 20' Q K . ~

16"

I---

0'1 i Chotera Toxin (pg/mt)

10

Fig. 5: Effect of cholera toxin treatment of NG108-15 cells on ICI 174864-mediated inhibition of GTPase in membranes. Cultured NG108-15 cells were treated for 24 h with cholera toxin at concentrations ranging from 0.1 to 10 ~tg/ml. GTPase was measurd in membranes under basal conditions (O) or in the presence of 10 ~ DADLE (0) and 10 p2VlICI 174864 (zx). 558

Vol. 165, No. 2, 1989

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

of GTPase activity, similarly to the effect of pertussis toxin (12). Unlike pertussis toxin, however, cholera toxin can suppress the effect of the antagonist at lower concentrations than the effect of the agonist. Since the effect of the antagonist on basal GTPase is a measure of the proportion of precoupled complexes between opioid receptors and G-proteins, these data indicate that even in vivo the preferential pertussis toxin-sensitive substrate of cholera toxin is a form of Gi/Go complexed to the receptor.

DISCUSSION The data presented in this study provide support to the idea that pertussis toxin-sensitive Gproteins serve as substrates of cholera toxin in their receptor-coupled form, and that a substantial proportion of spontaneously-formed complexes between opioid receptors and G-proteins exist in the membranes of NG108-15 cells. The finding that G-proteins others than Gs can be substrate of cholera toxin was first reported by Abood et al. (4), who documented toxin-catalyzed transfer of ADP-ribose into the a-subunit of transducin. Subsequently, Graves et al. (6) demonstrated GTPsensitive labeling of a 41 kDa band by cholera toxin in adipocyte membranes, and further evidences (7) indicated that these lower molecular weight bands are the 0~-subunits of the pertussis toxin substrates. More recently, Gierschik and Jakobs (9) demonstrated that cholera toxin-catalyzed ADP-ribosylation of Gi can be enhanced by agonist-occupied receptor in membranes from HL-60 cells, suggesting that Gi is best ADP-ribosylated by this toxin when existing in an active, receptorbound form. Similar observations were later extended to NG108-15 cells (10). We presented evidence that opioid receptors in intact membranes from NG 108-15 cells behave as a tight complex with G-proteins (11), and that a substantial portion of the basal pertussis toxin-sensitive GTPase activity in these membranes reflects an agonist-independent spontaneous interaction between opioid receptors and Gi/Go (12). The data presented here provide three additional evidences that unoccupied receptor can interact with, and activate Gi/Go, and that this activated complex between opioid receptor and G-protein is the target of the ADP-ribose transfer catalyzed by cholera toxin. First, the basal labeling of the 40 kDa substrate was almost abolished in a stereospecific fashion by the competitive peptidergic opioid antagonist ICI 174864, at concentrations similar to those required to inhibit basal pertussis toxin-sensitive GTPase activity (12). We showed previously for GTPase that the effect of ICI 174864 could not be attributed to contamination of membranes with endogeneous agonists which may tonically activate the opioid receptor. Such an explanation is also unlikely for ICI 174864-mediated inhibition of ADP-ribosylation, since naloxone, used at saturating concentrations, was inactive, indicating that the effect of the antagonists depends on their degree of negative intrinsic activity and not on their ability to compete with agonist binding on the receptor site. It is interesting, however, that the extent of negative intrinsic activity of these antagonists measured from their effects on GTPase and on ADPribosylation is not in perfect agreement. In fact, naloxone exhibits a partial negative effect on GTPase, but is absolutely inactive in ADP-ribosylation. It remains to be clarified whether this difference results from the fact that receptor-Gi and receptor-Go complexes contribute to GTPase and ADP-ribosylation in a different proportion, or whether it merely reflects different sensitivities of the two assay systems. 559

Vol. 165, No. 2, 1989

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Second, agonist-induced loss of receptor responsiveness in intact cells, not only leads to a reduction of agonist-stimulated ADP-ribosylation by cholera toxin in membranes that is temporally correlated with the reduction of agonist-stimulated GTPase activity, but also clearly depresses basal labeling by the toxin. Using both pertussis toxin-catalyzed ADP-ribosylation and immunoblot analysis it has been shown that DADLE treatment of NG108-15 cells does not alter the function or the levels of the pertussis toxin substrates Got~ and Gi2~, which both display apparent molecular masses of 40 kDa in membranes from NG108-15 cells (Lang and Costa, unpublished results). Thus, the agonist-induced diminution of basal ADP-ribosylation by cholera toxin is an indication that spontaneous complexes between receptors and G-proteins are less abundant in membranes obtained from cells in which the receptor is desensitized by prolonged exposure to the agonist. Third, we have shown here that exposure of intact ceils to high concentrations of cholera toxin reduces antagonist-mediated inhibition of GTPase more efficiently than agonist-mediated stimulation. This is consistent with the idea that the effect of the antagonist results from the disruption of spontaneously-formed complexes between opioid receptors and G-proteins, and that the toxin preferentially modifies Gi/Go in their receptor-complexed form even in membranes of intact cells. Thus, our data are in substantial agreement with those of Navon and Fung (5), who analysed the mechanism of cholera toxin-catalyzed ADP-ribosylation of the o~-subunitof transducin in a reconstituted system. Their data suggested that the preferred substrate of cholera toxin is the heterotrimeric transducin complexed to the light receptor rhodopsin.

ACKNOWLEDGMENTS We are grateful to Christine Gless for technical assistance and to Ursula B/iuerle for artwork. We thank Prof. Albert Herz for his support. This work was supported by the Deutsche Forschungsgemeinschaft, Bonn.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Althaus, F.R., and Richter, C. (1987) ADP-Ribosylation of Proteins, pp. 131-182, SpringerVerlag, Berlin. West, R.E., Jr., Moss, J., Vaughan, M., Liu, T., and Liu, T.-Y. (1985) J. Biol. Chem. 260, 14428-14430. Lochrie, M.A., and Simon, M.I. (1988) Biochemistry 27, 4957-4965. Abood, M.E., Hurley, J.B., Pappone, M.-C., Bourne, H.R., and Stryer, L. (1982) J. Biol. Chem. 257, 10540-10543. Navon, S.E., and Fung, B.K.-K. (1984) J. Biol. Chem. 259, 6686-6693. Graves, C.B., Klaven, N.B., and McDonald, J.M. (1983) Biochemistry 22, 6291-6296. Owens, J.R., Frame, L.T., Ui, M., and Cooper, D.M.F. (1985) J. Biol. Chem. 260, 1594615952. Aksamit, R.R., Backlund, P.S., Jr., and Cantoni, G.L. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7475-7479. Gierschik, P., and Jakobs, K.H. (1987) FEBS LetL 224, 219-223. Milligan, G., and McKenzie, F.R. (1988) Biochem. J. 252, 369-373. Costa, T., Klinz, F.-J., Vachon, L., and Herz, A. (1988) Mol. Pharmacol. 34, 744-754. Costa, T., and Herz, A. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, in press. Vachon, L., Costa, T., and Herz, A. (1987) Mol. Pharmacol. 31,159-168. Laemmli, U.K. (1970) Nature 227,680-685. Klinz, F.-J., Yu, V.C., Sadee, W., and Costa, T. (1987) FEBS Lett. 224, 43-48. 560