Cell, Vol. 62, 709-717,

August

24, 1990, Copyright

0 1990 by Cell Press

A Simple Structure Encodes G Protein-Activating Function of the IGF-WMannose 6-Phosphate Receptor Takashi Okamoto: Toshiaki Katada,? Yoshitake Murayama: Michio Ui,* Etsuro Ogata,’ and lkuo Nishimoto’5 * Life Science Laboratory Fourth Department of internal Medicine University of Tokyo School of Medicine Tokyo 112, Japan t Department of Life Science Faculty of Science Tokyo institute of Technology Yokohama 227, Japan t Department of Pharmaceutical Science Faculty of Pharmacology University of Tokyo Tokyo 113, Japan 5 Department of Medicine/Clinical Pharmacology Stanford University School of Medicine Stanford, California 94305

Summary The insulin-like growth factor-IVmannose Gphosphate receptor (IGF-IllmanGPR) can directly interact with and activate GIm2, a GTP binding protein (G protein). We found that the segment of residues 2410-2423 in the human IGF-ll/manGPR activates GI-~ in a manner simllar to G-coupled receptors. We observed a hierarchy of the segment action when tested on various G proteins, with an order of G,-* > G,., = GIm3> G,. The segment had no effect on G, or low molecular weight G ptoteins. The segment action depended on its primary structure and was potentiated when the segment was connected with a part of the receptor transmembrane region. Finally, the Gi.p-actlvating function of the human IGF-IllmanGPR could be blocked by an antibody against the segment, indicating a critical role for this small region of the receptor. Introduction GTP binding proteins (G proteins) are a family of membrane-associated regulatory proteins that transduce receptor-generated signals to effecters inside the plasma membrane (Gilman, 1987). The mechanism by which cell surface receptors regulate G proteins has recently been under intensive investigation. Extensive studies suggested that those receptors thus far demonstrated to be coupled to G proteins, referred to as G-coupled receptors, have not only similar primary sequences but a common membrane-spanning conformation with seven transmembrane regions (Sibley et al., 1987). This suggests that such a configuration may be essential for some mechanism involved in receptor-G protein communication. In contrast, several peptide ligands whose known receptors have a single transmembrane region can activate G protein-mediated pathways in intact cells and broken

ceil membranes (Goren et al., 1985; Luttrel et al., 1988; Church and Buick, 1988; Chedid et al., 1989; Nishimoto et al., 1987a, 1987b, 1989; Rothenberg and Kahn, 1988). In addition, there is convincing evidence that some cytoplasmic domains of multi-spanning receptors may play key roles in the interaction with G proteins (Kobilka et al., 1988; Rubenstein et al., 1987; Weiss et al., 1988). Furthermore, mastoparan (INLKALAALAKKIL) directly activates G proteins in a mode of action similar to G-coupled receptors (Higashijima et al., 1988). This raised the possibility that receptor-containing short segments of 10-20 residues long are involved in the receptor-induced G protein activation, although the sequence relationship between mastoparan and the G-coupled receptors remains undetermined. The linkage mechanism of receptors with G proteins will thus be further clarified by investigating at least two areas: whether multiple membrane-spanning structures are indispensable for the G protein-coupling ability of receptors, and whether any cytoplasmic segment of native receptors can directly interact with and activate G proteins. Recently, it has been reported that upon insulin-like growth factor-11 (IGF-Ii) stimulation, the IGF-illmannose g-phosphate receptor (IGF-IllmanGPR) functionally interacts with a Gr-like protein (Nishimoto et al., 1987a) and activates a calcium-permeable channel, augmenting calcium influx in mouse BALB/c 3T3 fibrobiasts (Kojima et al., 1988; Matsunaga et al., 1988). We further demonstrated the direct coupling of the receptor with Gr-2, a Gr-like protein possessing a 40 kd a subunit, in phosphoiipid vesicles as well as several kinds of ceil membranes (Nishimoto et al., 1989). The IGF-IllmanGPR has a single transmembrane domain, lacks intrinsic protein kinase activity, and has extracellular regions that bind two structurally unrelated ligands. The present study was conducted to delineate the mechanism that links the IGF-IllmanGPR with Gr-2. A comparison of the similarity of the IGF-IllmanGPR sequence with the sequence of mastoparan directed attention to a portion of the cytoplasmic domain of the receptor that has a capacity to activate Gr-2 directly. We will discuss the receptor-like mode of action of the segment, structure-function relationships, and functional involvement of the active segment in the IGF-IllmanGPR-Gr.2 interactions. Results

and Discussion

Sequence Investigations The IGF-IllmanGPR lacks both protein kinase activity and a multiple membrane-spanning structure (Nissley et al., 1985). Nevertheless, it is capable of coupling to Gr-2 iocated inside the plasma membrane (Nishimoto et al., 1989). We therefore reasoned that the receptor might contain unknown key sequences in its cytopiasmic domain that are capable of directly activating Gr-2. The primary sequence of the human IGF-IllmanGPR (Morgan et al.,

Cell

710

rat

receptor

1*q5.1q25:

***t******NR**N**f***Gt**f*t*S*

human bovine

receptor receptor

2288-2318: 1299-1329:

KKERRETVISKLTTCCRRSSNVSYKYSKVNK ****t**p,j*MSR********~********l . . . . 00 0'. 0.0

rat human bovine

rat human bovine

rat human bovine

1926-1964 2319 -2357 1330-1368

1965-2001 2358-2392 1369-1407

2002 -2039 2393-2430 1408-1444

*EA.*T*******f*t****t*******T-**T.*****V*~* A- -LSSLHGDDQDSEDEVLTIPEVKVHS-GRGAEAESS *ADT*SA*'**********'*L***RPP**Rpp**ApG**GG 0 0 0'0 00 00 0.0' o-00 2

* 0. 1

.

IGF-

The same residue as the human receptor is indicated by (‘). Both the hydrophobic residues (0) and basic residues (0) that are common among these receptors are indicated (Kyte and Dcoliile, 1982). Representative segments where at least 5 conserved hydrophobic residues and 2 conserved basic residues are assembled within 17 residues are underlined. The number at the left represents the sequence number.

0

Q*L**P*RKV’K*.**GE*M*‘:*******R**FRPG*R*

HPVRNAQSNALQE-REDDRVGLVRGEKARKGKSSSAQQK 0

00

-

.0000~0

3 2040-2060: 2431-2451: 1445-1460:

.O

t*****ft**t*f***V*A**L***D*********~T**

EEETDENETEWLMEEIQLPPPRQGKEGQENGHITTKSVK ***Attf*****t****PPt+Rp*+t****t*VAA*t*R 000 00000 0'0 0'0

00. rat human bovine

Figure 1. Cytoplasmic Domains of the II/ma&P Receptors of Three Species

0. .o.

0

4

PTTPAI**t*****f****t* TVSSTKLVSFHDDSDEDLLHI

..TpI...T*t********** 0.

00’0

1987) was compared with the sequences of demonstrated Gi-Coupled receptors, subtypes of the muscarinic acetylcholine receptors (Kubo et al., 1988; Banner et al., 1987) and the a2-adrenergic receptor (Kobilka et al., 1987). We also compared the sequence of the IGF-IllmanGPR with that of mastoparan, which is known to activate G, and Gr proteins (Higashijima et al., 1988). No stringent sequence homologies were found between the IGF-Ill man6PR and those molecules. However, we found that several regions in the cytoplasmic domain of the IGF-III man6PR shared the structural characteristics of mastoparan, in which several basic residues are interspersed among hydrophobic residues (Figure 1). In addition, we searched for sequences that were conserved among human, rat (MacDonald et al., 1988) and bovine (Lobe1 et al., 1987) IGF-IllmanGPR, since the rat and human receptors can equipotently couple to porcine and bovine Gi-2 (Nishimoto et al., 1989; Okamoto et al., 1990). Among these three species, residues 2410-2423 of the human receptor are conserved, and this sequence shares the hydropathicity characteristics of mastoparan. Of interest, the 14 residue sequence RVGLVRGEKARKGK had a structural similarity to the terminal portion of the third cytoplasmic loops of most G-coupled receptors and, in addition, the second inner loops and C-terminal regions of some Gr-coupled receptors (see below). Effect of Synthetic Segments on GTPyS Binding of GIe2 We chemically synthesized four peptides corresponding to the four regions of homology shown in Figure 1. Each was tested in solution for its effect on GTPyS binding of Gi.2 (Figure 2). The 14 residue peptide corresponding to position 2410-2423 (referred to as peptide 14) facilitated the GTPyS binding rate of Gi-2 w3-fold. This rate, stimulated by 100 FM peptide 14, was equal to or more than the

stimulation by intact rat IGF-IllmanGPR observed for Gr.2 in phospholipid vesicles (Nishimoto et al., 1989). Because muscarinic receptors, a class of physiological Gi-coupled receptors, increase the GTPyS binding rate 2-fold in a similar setting (Kurose et al., 1986) a 3-fold increase in the peptide 14-stimulated rate may reflect physiologically significant activation of G,+ Peptide 14 (100 PM)-induced stimulation of GTPyS binding displayed kinetics quite similar to those of muscarinic receptor-induced stimulation (Kurose et al., 1986), that is, an increased maximal value of kapp (the apparent first-order rate constant for GTPyS binding) without significant change in its affinity for GTPyS (data not shown). The effect of peptide 14 on initial rates of binding was dose dependent (Figure 28). Effects were observed with as little as 1 uM, a concentration ml00 times higher than that of Gr-2, and 100 uM peptide 14 prompted 1 mol of G,.* to bind ~1 mol of GTPyS, which implies full activation of Gr.2 within 5 min. The GTPyS binding rate appeared to reach saturation by 100 FM peptide 14 when assayed at 5 min. However, at the 1 min time point, 1 uM peptide 14 produced a greater rate of GTPyS binding than 100 uM peptide 14 (data not shown). The other synthetic peptides did not induce accelerated GTPyS binding (Figure 2). Because all of these inactive peptides consist of several basic residues interspersed among hydrophobic amino acid residues, more strict conditions of primary structure appeared necessary for successful activation of Gr.2 by peptides. Effect of Peptide 14 on GDP Release and GTPase Activities of GIe2 Peptide 14 stimulated the release of GDP from and the GTP hydrolysis activity of Gr.2. The ratio of 100 uM peptide 14-induced stimulation of GDP release to basal GDP release was comparable to the ratio of 100 PM peptide

G Protein-Activating 711

Segment

of IGF-IllmanGP

Receptor

1.0

Figure 2. Effect of Peptide 14 and Other Segments of the Human IGF-IllmanGPR on G,.2 0.9 0.8 0.7 0.6 0.5

b

0.2

0.4 0.0

0.0 0

2

4 Time

6 (min)

8

10

0

0.1

1

10

[peptide] (PM)

14-stimulated GTPyS binding to the basal binding (data not shown). The turnover number of GTP hydrolysis was augmented by 100 PM peptide 14 from 0.02 per min to 0.04 per min. Such an increment of the GTPase activity induced by peptide 14 was again comparable to those induced by muscarinic (Kurose et al., 1988) and a,-adrenergic (Cerione et al., 1986) receptors. Taken together, these results show that peptide 14 activates Gi-2 in every aspect of G protein activation, notably an increase in the GTPyS binding, GDP release, and GTP hydrolysis activities. Effect of Pertussis Toxin on Peptide 14 Action The effect of 100 PM peptide 14 was about 70% attenuated in the case of Gi-2 pretreated with pertussis toxin in the presence of NAD as compared with control Gi-2 pretreated with the toxin in the absence of NAD. The toxin completely inhibited the stimulation of Gi-2 induced by peptide 14 at less than 10 FM, with incomplete inhibition at higher peptide concentrations. Gi.2 was assumed to be fully ADP-ribosylated, since no additional [a”P]NAD was incorporated into Gi.2 after incubation with the toxin in the presence of unlabeled NAD (data not shown). Therefore, the incomplete inhibition may be due to incomplete sensitivity of ADP-ribosylated Gi.2 to the action of peptide 14. Effect of Magnesium Ion on Peptide 14 Action G proteins require millimolar concentrations of Mg*+ to bind GTP effectively, whereas G-coupled receptors stimulate GTP binding of G proteins by reducing the Mg*+ requirement to micromolar concentrations (Gilman, 1987). We tested effects of peptide 14 and Mg*+ on Gi-2 activity (Figure 3). In the mid-range of Mg2+ concentrations (1 PM to 1 mM), 100 PM peptide 14 increased the GTPyS binding rate by -2-to 3-fold, while at very low (lO mM) Mg*+ concentrations, the peptide had no additional effect. Thus, the mode of action of peptide 14 is consistent with that of G-coupled receptors, even with respect to Mg*+ concentration requirement.

100

(A) Time course of GTPyS binding to G,.2 after treatment with 100 NM peptides. In the presence of 60 nM [%]GTPvS. 10 nM G,.2 was incubated for the indicated periods with various synthetic peptides (no. 1 [O], no. 2 [ 0 1, no. 3 [ 01, no. 4 [peptide 14][ 01, or water [0]), corresponding to the regions of the human IGF-III man6PR indicated in Figure 1. The fraction bound to GTPyS of total G,.2 is indicated as a function of incubation time. Values represent the mean +SE of three independent experiments. (8) Dose response relationship for these seg ments on the GTPvS binding rate of G,.z. G,.z was incubated for 5 min with each peptide in the same settina as in (A). The GTPvS binding rate per 5 min (kol pe; &ol of G,.* per 5 minj is indicated as a function of the concentrations of each peptide examined. Values represent the mean of three experiments, while SE was less than 5% of each value.

Effect of Phospholipids on Peptide 14 Action The rat IGF-IllmanGPR can stoichiometrically interact with Gi.2 in phospholipid vesicles (Nishimoto et al., 1989). We found that phospholipids also enhanced the action of peptide 14. In the presence of 0.003%-0.02% soybean phosphatidylcholine, 30 PM peptide 14 stimulated the GTPyS binding rate of Gi.2 from the basal binding rate of 0.3-0.4 mol per mol of Gi.2 per 5 min to 0.8-0.9 mol per mol of Gi-2 per 5 min. However, when phosphatidylcholine was present at less than 0.003% or above 0.02%, it only increased the binding rate from similar basal rate to 0.5-0.6 mol per mol of Gi-2 per 5 min. Concentrations of phosphatidylcholine, 0.003%-0.02%, were estimated to weigh 40-240 times more than Gi.2, a ratio known to be appropriate for reconstitution of proteins into phospholipid vesicles (Zimniak and Racker, 1978). Phospholipids may therefore enhance the peptide 14-Gi-2 interaction through the reconstitution of Gi.2. In accordance, the dose response curve of the peptide 14 action on Gi-2 was shifted left -3-fold in reconstituted vesicles with Gi.2 (compare Figure 4 with Figure 26), although the potentiating effect of phospholipids on the peptide 14 action might be less than that on the action of mastoparan (Higashijima et al., 1988). Effect of Transmembrane Domain-Connected Peptide 14 on GiS2 Even in the presence of phospholipids, peptide 14 concentrations 100-1000 times higher than that of Gi-2 were required to stimulate G,-*. We noted that phospholipids had a lesser effect on peptide 14-Gi.2 interaction than on mastoparan-G, interaction, and peptide 14 has a lower content of hydrophobic residues (Kyte and Doolittle, 1982) than mastoparan or transmembrane domains of receptors, both of which can directly interact with plasma membrane phospholipids. The small effect of phospholipids on the activity of peptide 14 suggested that this region plays no role in the receptor interaction with plasma membranes. We therefore tested whether the peptide 14 segment became more effective on Gi.2 in phospholipid vesi-

Cell 712

x l.0 z 0.9

z gs 0.9 2 c z 08.

z

5 0.7 E 5 0.6

2 0.6 al z 0.4 f q

0.8

;

0.2

0.5

g 0.4

-log [Mg2’l Figure

3. Effect

of Mg2+ and Peptide

0

14 on GTP7.S Binding

[Peptide

141

of Peptide

14 on GTPyS

Binding

Mel per Mol of G Protein

to Various

10 [peptidel

Figure 4. Effect on G,~2

of Transmembrane

io’

lo3

lo4

(nM)

Domain-Connected

Peptide

14

G,.s reconstituted in phospholipid vesrcles was incubated with transmembrane-plus peptide 14 (LTACLLTLLLYRVGLVRGEKARKGK) and its related peptides for 5 min, and its GTPyS binding was measured. (0): transmembrane-plus peptide 14, (0): transmembrane domain (LTACLLTLLLY) alone, ( 0 ): peptide 14. (m): peptide 14 in the presence of 10 mM LTACLLTLLLY. Values represent the mean +SE of three experiments.

cles when the segment was connected with the native transmembrane domain of the human IGF-IllmanGPR. We synthesized the 25 residue peptide that combined the 11 residue segment of the transmembrane domain with the peptide 14 sequence (LTACLLTLLLYRVGLVRGEKARKGK; the underlined portion is the C-terminal part of the transmembrane domain of the human IGF-lllman6PR). The peptide, referred to as transmembrane-plus peptide 14, exerted striking stimulation of Gi.2 reconstituted in phospholipid vesicles at concentrations comparable to that of Gi.2. The action of transmembrane-plus peptide 14 was estimated to be 300-1000 times more potent than peptide 14 and reached saturation at 100 nM (Figure 4). The half maximum concentration (EC& of the peptide was 30 nM, which is comparable to the concentration (10 nM) of Gi-2 present in each experiment. In contrast, the transmembrane segment itself (LTACLLTLLLY) was totally without effect on Gi.2, indicating that the peptide 14 segment has direct bearing on the Gi.2 activation induced by transmembrane-plus peptide 14. In addition, the effect of peptide 14 on Gi.2 in phospholipid vesicles was not altered by the presence or absence of the isolated transmem-

1. Effect

1

to G,.s

G,.s was incubated with (0) or without (0) 100 uM peptide 14 at various concentrations of Mg 2+ for 5 min. The GTP$S binding rate was measured. Each value represents the mean of three experiments.

Table

0.1

brane segment, suggesting that the transmembrane segment must be connected with peptide 14 to potentiate the action of pephde 14 on Gi.2. Transmembrane-plus peptide 14 also showed a stimulatory effect (ECsO = 1 PM) on G,.2 ~30 times more potent than peptide 14, even when tested in the absence of phospholipids (data not shown). Yet, given the fact that the action of transmembrane-plus peptide 14 was about 1000 times more potent than that of the isolated peptide 14 in the presence of phospholipids, it is reasonable to assume that the potentiating effect of the transmembrane segment is mainly mediated by its interaction with phospholipids. These results suggest that the peptide 14 segment in its native IGF-IIlmanGPR may be able to communicate effectively with Gi.2 with the aid of the receptor transmembrane domain and the plasma membrane.

G Proteins

per Unit Period

ilh

62

Gt.1

63

GO

GS

c-K&as

0 (a) 1 10 30 100 (b)

0.41 0.44 0.55 0.66 0.92

0.37 0.44 0.46 0.49 0.58

0.37 0.39 0.46 0.47 0.54

0.36 0.37 0.39 0.41 0.46

0.31 0.31 0.31 0.32 0.32

0.47 0.47 0.47 0.47 0.47

0.42 0.42 0.42 0.42 0.42

O-9 1 (a)

2.2

1.6

1.5

1.3

1.0

1.0

1 .o

p21

smg

p25A

GTPyS binding was measured per 5 min in Gi.a, per 4 min in G,., and Gr.3, per 2 min in G,, per 10 min in G,, and per 60 min in c-Ki-ras p21 and smg p25A in the conditions described in Experimental Procedures. The fraction bound to GTPyS of total G protein for these respective periods is presented. These periods are appropriate to approximate and compare the GTPyS binding rates, not only because the GTP$S binding to each G protein increases almost linearly within each period but also because each basal GTPyS binding is similar. Each G protein has a distinct GTPyS binding rate as described (Oinuma et al., 1967; Carty et al., 1990). The lowest line indicates the ratio of the stimulated GTPyS binding rate with 100 uM peptide 14 to unstimulated rate. Values represent the mean of three independent experiments for each concentration of peptide 14.

G$otein-Activating

Segment

of IGF-IllmanGP

Receptor

Figure 5. Structure-Function Peptide 14 Sequence

0.7

$

0.6

z

0.5

-c g

0.0 0.4

% 0.9 c .F z 3

0.8

2

0.6

k

0.5

+

peptide 10

+

peptide 14R

-)

peptide 14K peptide 14H

of

G,.p was incubated with various residuesubstituted or -truncated peptides from the peptide 14 sequence for 5 min. and the GTPrS binding rate was assayed. Values represent the mean of four experiments. Each SE was less than 5% of the value. Effects of residue-truncated or residue-elongated peptides (A), peptides modified at the C-terminus with glycine (B), and basic residue-substituted peptides (C and D) are indicated. Inset: effect of the substitution of the second arginine residue of peptide 14 with proline (peptide 14Pa) or glutamine (peptide 14Qa).

0.8 5

Relationship

0.7

0.4 0.0

0

0.1

1

10

100

0

0.1

1

10

100

[peptidel (PM)

Effect of Peptide 14 on Other G Proteins To examine whether the action of peptide 14 is selective for Gr.2 or ubiquitous on many G proteins, we tested the effect of peptide 14 on various kinds of G proteins. As shown in Table 1, peptide 14 preferentially acted on Gr-2, less effectively on Gr.1 and Gr.3, and weakly on G,. Peptide 14 may be able to act on each of the Gr proteins because they have highly homologous structures (Itoh et al., 1988a). In contrast to these Gr-related G proteins, the effect of peptide 14 on G,, c-Ki-ras ~21, or smg p25A was minimal. The latter two proteins belong to another family of G proteins with low molecular masses of 20-30 kd. Thus, the stimulatory effect of peptide 14 showed preferences on Gr proteins. Experimental conditions such as Mg*+ concentrations (between 1O-7 M and lo-* M) or the presence of phospholipids (between 0.01% and 1% phosphatidylcholine) did not alter the preference order of the action of the peptide on Gi proteins and G,. The sensitivity to peptide 14 suggests that Gi.2 may be the most appropriate G protein that the peptide 14 segment activates. Structure-Function Relationship of Peptide 14 Sequence TIuncafed Peptides The action of the peptide 14 segment was further characterized by experiments using residue-truncated or residue-substituted peptides (Figure 5). About 30-fold higher concentrations of peptide 18 (REDDRVGLVRGEKARKGK) were required to stimulate GTPrS binding to Gr-2 to the same extent as peptide 14, while peptide 9 (RGEKARKGK) had no discernible effects on the stimulation of Gr-2

(Figure 5A). This suggests that the G protein-activating function of the peptide 14 segment depended on its primary structure, prompting us to investigate the detailed structure implicated in the stimulatory action of peptide 14. Next we examined peptides that were truncated in the C-terminal motif of peptide 14. As shown in Figure 5A, neither peptide 10 (RVGLVRGEKA) nor peptide 4 (RKGK) produced any significant stimulation of GTPyS binding to Gr-2 in a concentration range up to lO,OOO-fold of the G protein. Fifty-fold increased concentrations of peptide 12 (RVGLVRGEKARK) were required to stimulate Gi-2 to an extent similar to peptide 14. The peptide 12 potency was comparable to the potency of mastoparan, which has a primary sequence resembling peptide 12 at the C-terminus (data not shown). These results suggest an essential role of the C-terminal structure of peptide 14. Peptides Modified at C-Terminus We further tested the effects of peptides with modified C-terminal motifs, RKGGK, RKGGGK, and RKGGGGK (Figure 5B), since several segments structurally related to peptide 14 in Gi-coupled receptors contain a B-B-X-X-B motif at the C-termini of the segments (see below). Peptide 14G2 (RVGLVRGEKARKGGK) had a stimulatory effect that was regarded as equipotent to peptide 14; peptide 14G4 (RVGLVRGEKARKGGGGK) showed only a minimal effect similar to the effect of peptide 12; and an intermediate potency of peptide 14G3 (RVGLVRGEKARKGGGK) was observed in the stimulation of GTPyS binding. These results indicate that the C-terminal motif should be B-BX-B or B-B-X-X-B for peptide 14-related sequences to activate Gr-2 appropriately, and that the peptides containing

Cdl 714

these motifs are similarly active. Taken together with the results above, the two kinds of C-terminal motifs can play a key role in the Gc2-activating function of peptide 14. Residue-Substituted Peptides Another characteristic structure of the peptide 14 sequence was interspersed basic residues at its N-terminal portion as well as at its C-terminus. Three sets of residue-substituted peptides elucidated the role of these basic residues. First, we tested the effects of peptide 14R (RVGLVRGERARRGR), 14K (KVGLVKGEKAKJGKJ, and 14H (HVGLVHGEHAHHGH), in which all of the basic residues of peptide 14 were substituted with arginine, lysine, or histidine residues, respectively (Figure 5C). Peptide 14R stimulated the initial GTPyS binding rate of Gr.2 approximately two times as potently as peptide 14. In contrast, peptide 14H elicited only minimal stimulation from Gr.2, whereas peptide 14K displayed an intermediate potency. Thus, the kinds of basic residues located in the segment could have direct bearing on the G protein-activating capacity. Second, we examined peptides each of whose basic residues in the N-terminal side was substituted with an alanine residue (Figure 5D). Peptides 14A, (AVGLVRGEKARKGK) and 14A3 (RVGLVRGEAARKGK) indicated actions 10 and 3 times less potent than peptide 14, respectively. Peptide 14A2 (RVGLVAGEKARKGK) elicited little stimulation from Gr.2. Since peptide 14A2 was almost inactive, we further substituted the N-terminal second arginine residue (R2) of peptide 14 with residues other than alanine. As indicated in the inset of Figure 5D, Rp-substituted peptides acted on Gr-2 to various degrees depending on the substituted residues. These results suggest that R2 plays a major role in the Gra-activating function of peptide 14 in addition to the N-terminal first arginine residue, while the third lysine residue is less involved in the action of peptide 14. Similar results were obtained by measuring GDP release from Gr.2 (data not shown). While those experiments are still limited in scope, they suggest an outline of the structure-function relationship of the IGF-IllmanGPR segment. Obviously, the G protein-activating function of the segment depends on its primary structure, especially the two basic residues in the N-terminal side and the C-terminal motif, B-B-X-B or B-B-X-X-B. In the IGF-IllmanGPR, no segments other than the peptide 14 sequence retain such structures. Further attention should be directed to the cytoplasmic segments of other general receptors that possess those structural characteristics. For example, several cytoplasmic regions of MI-MS muscarinic acetylcholine receptors and up-adrenergic receptors retain such features. If the sequence length is limited to that within 20 residues, those are: first, the second cytoplasmic loops at no. 123-140 (porcine and rat MI acetylcholine receptors), 121-136 (porcine and human M2 receptors), 131-146 (human M3 receptor), 165-162 (rat M3 receptor), 130-147 (human M., receptor), 129-146 (rat M4 receptor), 127-144 (rat M5 receptor), 131-146 (human az-adrenergic receptor); second, the third cytoplasmic loops at no. 316-334 and 346-365 (porcine and rat MI receptors), 370-387 (por-

cine and human M2 receptors), 333-350 and 362-400 (human M3 receptor), 252-262 and 287-304 (rat M3 receptor), 223-240 and 240-256 (human Mq receptor), 222-239 (rat M,, receptor), 361-399 (human and rat M4 receptors), 359-371 and 361-400 (rat Ms receptor), and 294-313 or 301-313 (human a*-adrenergic receptor); and third, the C-terminal region at no. 430-447 (porcine and rat MI receptors). While the search for general receptors for such sequences will be described elsewhere, information from these sequences may provide a helpful suggestion concerning the investigation of transmembrane signaling functions of segment-containing receptors. Involvement of Peptide 14 Segment in Receptor Function The direct linkage of the rat IGF-IllmanGPR with Gr-2 is quite similar to the typical coupling observed for conventional G-coupled receptors (Nishimoto et al., 1969; Okamoto et al., 1990). In the present study, the segment that activates Gi.2 in a receptor-like manner could be located in the cytoplasmic domain of the human IGF-IllmanGPR. Therefore, the IGF-IllmanGPR might activate Gr-2 through the peptide 14 segment. Involvement of the peptide 14 segment would account for the pertussis toxin-sensitivity of the IGF-IllmanGPR-Gr.2 linkage (Nishimoto et al., 1989) and the selective action of peptide 14 on G, proteins over G, may explain the finding that low concentrations of IGFII attenuate PGE,-stimulated CAMP production in intact cells (Richman et al., 1980). To assess the involvement of the peptide 14 region in receptor signaling, we raised a polyclonal antibody against peptide 14 (RP14AB) and tested its effect on the IGF-Ill manGPR-Gr.2 coupling in phospholipid vesicles reconstituted with the human receptor and Gi.2 (Figure 6). First, RP14AB attenuated the peptide 14-induced stimulation of GTPyS binding to soluble Gr.2 in a dose-dependent manner. The concentration of RP14AB at which 50% inhibition was achieved (ICso) was ~5 uglml using 10 t.rM peptide 14 or 10 pglml at 30 t.rM peptide 14. Second, in vesicles consisting of the human IGFIllmanGPR and Gi-2, 100 nM IGF-II maximally increased GTPyS binding to Gr.2 as much as the effect of 30 uM peptide 14 on soluble G,+ The effect of IGF-II on Gr-2 has been demonstrated to be mediated by the IGF-IllmanGPR in this system (Nishimoto et al., 1989). When RP14AB was introduced into vesicles by means of sonication, the antibody inhibited, in a dose-dependent manner, the IGF-IIinduced Gr-2 activation in the vesicles with an ICm of 10 ug/ml and complete inhibition with 100 t.rg/ml RP14AB. In contrast, preimmune rabbit IgG showed no inhibition of either the Gi.2 activation induced by peptide 14 or that induced by IGF-II in the reconstituted vesicles. Third, the inhibitory effect of RPl4AB was completely abolished by the addition of peptide 14A2 in a dosedependent manner. The ECsO was ~60 PM and reached saturation at 100-300 PM peptide 14A2. Since peptide 14A2 has a primary structure almost identical to peptide 14 with the exception of one residue yet was inactive by itself in either the stimulation or inhibition of Gr-2 (see

$;$otein-Activating

Segment

of IGF-IllmanGP

Receptor

factors whose receptors have fewer than seven transmembrane regions. While cytoplasmic domains of receptors have been inferred to be involved in generating intracellular signals, a significant role of the cytoplasmic domain in receptor function and the ligand-dependent mechanism by which receptors control the function of those domains may be further clarified by investigating the IGF-Ill man6PR. Expsrlmental

E o.oL

b

0

I

*

1 [Antibody;

I

1 I 1 100

Y@ml)

Figure 6. Effect of RP14AB, an Antibody against Peptide 14, on G,.s Activation Induced by Peptide 14 or by IGF-II in the Human IGFIl/manGPR-Gr.s Reconstituted Vesicles Peptide 14, open symbols. IGF-II, solid symbols. Experiments using peptide 14 (broken lines) were performed in a 0.07% CHAPS solution. G,.s (10 nM) was incubated with increasing concentrations of RP14AB in the presence of 10 uM (0) or 30 uM (A) peptide 14 for 5 min. The effect of RP14AB on IGF-II-stimulated Gi.s activation in the human IGFll/man6PR-Gi.n vesicles is indicated with solid lines. Phospholipid vesicles consisting of 3.1 nM human IGF-ll/manGPR and 10 nM Gi.s were incubated with 100 nM IGF-II in the presence of increasing concentrations of RP14AB (0) or preimmune IgG (cl) for 5 min. The effect of preimmune IgG on 30 PM peptide 14-induced stimulation (0) or on 100 nM IGF-II-stimulated Gi.s activation in reconstituted vesicles (R) and effect of RP14AB on solubilized Gi.s in the absence of peptide 14 (0) or on reconstituted Gr.2 in the absence of IGF-II (0) are also indicated as controls. In experiments using vesicles, values represent the mean *SE of three independent experiments.

above), it is highly likely that peptide 14As competed with RPMAB for the peptide 14 segment in the IGF-IllmanGPR, and the competition with peptide MA2 thus attenuated the effect of the antibody on the receptor. The data indicate that the peptide 14 segment is involved in the Gi-zactivating function of the IGF-IllmanGPR in response to IGF-II, although we cannot totally exclude the possibility that the antibody acts by sterically blocking the receptorGi.2 interaction. Conclusion A Gcractivating segment was found in the cytoplasmic domain of the human IGF-lllman6PR. The segment activated Gr.2 in a specific manner quite similar to conventional Gi-Coupled receptors as well as the IGF-ll/manGPR itself. In addition, the function of the segment clearly depended on its primary structure. Although we cannot exclude the possibility that there is still another active segment in the IGF-ll/man6PFi, experiments using the antibody suggest that the identified active segment may play a major role in the Gi.p-activating function of the human IGF-II/man6PR. Because it documents one possible sequence that is active in Gr-2 activation, this study provides a new clue toward understanding the signaling mechanism of not only the IGF-IllmanGPR but also receptors in general. Furthermore, it opens new insight into the investigation of mitogenic pathways triggered by well-characterized growth

Pmcedures

Synthetic Psptldes and G Proteins The peptides used in this study were synthesized by the solid phase method and highly (95%-99%) purified by high performance liquid chromatography using a Nucleosil 5Cta column with eluants of 0.1% trifluoroacetic acid and l%-60% gradient concentrations of CHsCN. Trifluoroacetic acid in the fraction was replaced with acetic acid using DOWEX-1 column chromatography. The lyophilized synthetic peptide was dissolved in distilled water. As controls, neither water nor acetic acid up to 0.6 mM had any effect on G proteins in buffer A (50 mM HEPES-NaOH [pH 7.41, 100 uM EDTA, and 120 uM MgCl2). Gligomerit G proteins were purified from bovine brain to homogeneity (Bokoch et al., 1964; Katada et al., 1967; ltoh et al., 1966b). Free o subunit of Gi-s purified from bovine lung and G1.s purified from bovine spleen were gifts of Dr. T Asano of the Institute for Developmental Research (Morishita et al., 1969); the 5~ subunit was purified from bovine brain (Katada et al., 1967). G proteins of low molecular weights purified from bovine brain to homogeneity were generously donated by Dr. Yoshimi Takai of Kobe University (Kikuchi et al., 1986; Yamashita et al., 1966). To obtain reproducible results, it was important that the 5~ subunit be present in an amount greater than or equal to Gi.sa. Unless specified, we mixed Gr.so and the 6~ subunit in a molar ratio of 1:1.5. The concentration of Gi.2 described was that of Gt.so. Gi.k was used in a similar fashion. Essentially similar results were obtained when the trimer form of Gt.2 purified from bovine lung or spleen was employed. GTPvS Binding Assay GTPyS binding to G proteins was measured according to the method described (Nishimoto et al., 1969). Incubation of IO nM G proteins was accomplished with or without peptides in buffer A and 60-100 nM [35S]GTPvS at 3pC. GTPTS binding to peptide 14 was negligible. The total amount of G proteins was measured as the maximal GTPyS binding at room temperature. Mg2+ concentrations were set by using Mg-EDTA buffer as described (Birnbaumer et al., 1963).

GDP Relesss Assay and GTP Hydrolysis Assay GDP release from Gi-s was measured as described (Okamoto et al., 1990). GTPase activity of Gr.2 was measured in buffer A supplemented with 1 uM [T-~~P]GTP. The reaction was terminated by adding 2.5% Norit SX Plus (WAKO Pure Chemical, Japan) resolved in 100 mM NaHCOs (pH 6.0). After centrifugation at 20,000 x g for 5 min, the supernatant was counted. Anti-Psptide 14 Antlbody Fresh blood was drawn from a rabbit immunized with 10 mg of peptide 14 and centrifuged. After the anti-serum was precipitated with ammonium sulfate and dialyzed against phosphate-buffered saline, RP14AB (anti-peptide 14 IgG) was purified from the anti-serum by using protein A-Sepharose (Bio-Rad). Before experiments, RP14AB was incubated with peptide 14 or sonicated and incubated with reconstituted IGFIl/manGPR for 90 min at room temperature. Materials The clonal human IGF-ll/manGPR purified from CDNA-overexpressed L cells was a gift of Dr. W. S. Sly of St. Louis University (Nolan et al., 1990). Reconstitution of both the receptor and Gi.2 or Gr.2 alone into phospholipid vesicles was carried out as described (Nishimoto et al., 1969). Rat IGF-II was a generous gift of Dr. S. f? Nissley (National Ins& tute of Health, Bethesda, MD). Human recombinant IGF-II (Amgen)

Cell 716

yielded similar results. Soybean phosphatidylcholine (azolectin) taining 20% of phosphatidylcholine and crude phospholipids from Sigma (P-5638).

conwas

We thank Dr. Tomiko Asano for oligomeric G proteins; Dr. William S. Sly for the clonal human IGF-ll/manGPR; Dr. Yoshimi Takai for critical advice and small G proteins; Drs. Shintaro lwashita and Yuichiro Hayashi for discussion; Drs. Tatsuya Haga, Akira Kikuchi, Yoshihiro Ohkuni, Tarho liri, Tomokazu Ohtsuka, and Yoshiomi and Yumi Tamai for support. Very special thanks to Drs. Chris Martens, Phyllis Gardner, Paul Nghiem, Kazuyoshi Yonezawa, and Shuji Takahashi for critical reading of this manuscript; and Drs. Ron Rosenfeld, Richard A. Roth, Yoshito Kaziro, Kumao Toyoshima, and Ken-ichi Arai for criticism. We gratefully acknowledge support by grants from the Ministry of Education, Science, and Culture of Japan, the lchiro Kanehara Foundation, the Naito Foundation, the Uehara Memorial Foundation, Toyo Jozo Co., and Teijin Institute for Biomedical Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “edvertisement” in accordance with 18 USC. Section 1734 solely to indicate this fact. May 3, 1990; revised

June

14. 1990.

References Birnbaumer, L., Stengel. D., Desmier, skolin regulation of liver membrane them. 136, 107-112.

M., and Hanoune, J. (1983). Foradenylate cyclase. Eur. J. Bio-

Bokoch, G. M., Katada, T., Northup, J. K., Ui, M., and Gilman, A. G. (1984). Purification and propertiesof the inhibitory guanine nucleotidebinding regulatory component of adenylate cyclase. J. Biol. Chem. 259, 3560-3567. Bonner, T I., Buckley, N. J., Young, A. C., and Brann. M. R. (1987). Identification of a family of muscarinic acetylcholine receptor genes. Science 237, 527-531. Carty, D. J., Padrell, E., Codina, J., Birnbaumer, L., Hildebrandt, J. D., and lyengar, R. (1990). Distinct guanine nucleotide binding and release properties of the three Gi proteins. J. Biol. Chem. 265, 6268-6273. Cerione, R. A., Regan, J. W., Nakata, H., Codina, J., Benovic, J. L., Gierschik, P, Somers, R. L., Spiegel, A. M., Birnbaumer, L., Lefkowitz, R. J., and Caron, M. G. (1988). Functional reconstitution of the az-adrenergic receptor with guanine nucleotide regulatory proteins in phospholipid vesicles. J. Biol. Chem. 261, 3901-3909. Chedid, M., Shirakawa, F., Naylor, P, and Mizel. S. 8. (1989). Signal transduction pathway for IL-l: involvement of a pertussis toxin-sensitive GTP-binding protein in the activation of adenylate cyclase. J. Immunol. 742, 4301-4306. Church, J. G., and Buick, R. N. (1988). G-protein-mediated epidermal growth factor signal transduction in a human breast cancer cell line: evidence for two intracellular pathways distinguishable by pertussis toxin. J. Biol. Chem. 263, 4242-4246. Gilman, signals.

from G-protein a-subunit genes and cDNAs with parsequences from purified proteins. FEES Lett. 230,

Katada, T., Oinuma, M., Kusakabe, K., and Ui, M. (1987). A new GTPbinding protein in brain tissue serving as the specific substrate of isletactivating protein, pertussis toxin. FEBS Lett. 213, 353-358.

Acknowledgments

Received

acid sequences tial amino acid 85-89.

A. G. (1987). G proteins: transducers Annu. Rev. Biochem. 56, 615-649.

of receptor-generated

Goren, H. J., Northup, J. K., and Hollenberg, M. D. (1985). Action of insulin modulated by pertussis toxin in rat adipocytes. Can. J. Physiol. Pharmacol. 63, 1017-1022. Higashijima, T., Uzu, S., Nakajima, T., and Ross, E. M. (1988). Mastoparan, a peptide toxin from wasp venom, mimics receptors by activating GTP-binding regulatory proteins (G proteins). J. Biol. Chem. 263, 64918494. Itoh, H., Toyama, R., Kozasa, T., Tsukamoto, T., Matsuoka, M.. and Kaziro, Y. (1988a). Presence of three distinct molecular species of G, protein u subunit: structure of rat cDNAs and human genomic DNAs. J. Biol. Chem. 263, 6656-6664. Itoh, H., Katada, T., Ui, M., Kawasaki, H., Suzuki, K., and Kaziro, Y. (1988b). Identification of three pertussis toxin substrates (41, 40 and 39kd proteins) in mammalian brain: comparison of predicted amino

Kikuchi. A., Yamashita, T., Kawata, M., Yamamoto, K.. Ikeda, K., Tanimoto, T., and Takai, Y. (1988). Purification and characterization of a novel GTP-binding protein with a molecular weight of 24,000 from bovine brain membranes. J. Biol. Chem. 263, 2897-2904. Kobilka, B. K., Caron, M. G., quencing, and azadrenergic

Matsui, H., Kobilka, T. S., Yang-Feng. T. L., Francke, U., Lefkowitz, R. J., and Regan, J. W. (1987). Cloning, seexpression of the gene coding for the human platelet receptor. Science 238, 650-658.

Kobilka, B. K., Kobilka, T. S., Daniel, K., Regan, J. W.. Caron, M. G., and Lefkowitz, R. J. (1988). Chimeric az-, pz-adrenergic receptors: delineation of domains involved in effector coupling and ligand binding specificity. Science 240, 1310-1316. Kojima, I., Nishimoto, I., liri, T., Ogata, E., and Rosenfeld, Ft. G. (1988). Evidence that type II insulin-like growth factor receptor is coupled to calcium gating system. Biochem. Biophys. Res. Commun. 754, 9-19. Kubo, T., Fukuda, K., Mikami. A., Maeda, A., Takahashi, A., Mishina, M., Haga, T., Haga, K., Ichiyama, A., Kangawa, K., Kojima, M., Matsuo, H., Hirose, T., and Numa, S. (1986). Cloning, sequencing and expression of complementary DNA encoding the muscarinic acetylcholine receptor. Nature 323, 411-416. Kurose, H.. Katada, T., Haga, T., Haga, K., Ichiyama, A., and Ui, M. (1988). Functional interaction of purified muscarinic receptors with purified inhibitoryguanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J. Biol. Chem. 267, 64236428. Kyte, J., and Doolittle, hydropathic character

R. F. (1982). A simple method for displaying of a protein. J. Mol. Biol. 757, 105-132.

the

Lobel. P, Dahms, N. M., and Kornfeld. S. (1987). Cloning and sequence analysis of the cation-independent mannose aphosphate receptor. Proc. Natl. Acad. Sci. USA 84, 2X3-2570. Luttrel, L. M., Hewlett, E. L., Romero, G., and Rogol, A. D. (1988). tussis toxin treatment attenuates some effects of insulin in BC3H-1 rine myocytes. J. Biol. Chem. 263, 6134-6141.

Permu-

MacDonald, R. G., Pfeffer, S. R., Coussens, L., Tepper, M. A., Brocklebank, C. M., Mole, J. E., Anderson, J. K., Chen. E.. Czech, M. P, and Ullrich, A. (1988). A single receptor binds both insulin-like growth factor II and mannose 6-phosphate. Science 239, 1134-1136. Matsunaga, H., Nishimoto, I., Kojima, I., Yamashita, N.. Kurokawa, K., and Ogata, E. (1988). Activation of a calcium-permeable cation channel by insulin-like growth factor II in BALB/c 3T3 cells. Am. J. Physiol. 255, C442-C446. Morgan, D. O., Edman, J. C., Standring, D. N., Fried, V. A., Smith, M. C., Roth, R. A., and Rutter, W. J. (1987). Insulin-like growth factor II receptor as a multifunctional binding protein. Nature 329, 301-307. Morishita, R., Asano, T., Kato, K., Itoh, H.. and Kaziro, Y. (1989). Purification and identification of two pertussis-toxin-sensitive GTP-binding proteins of bovine spleen. Biochem. Biophys. Res. Commun. 767, 1280-1285. Nishimoto, I., Hata, Y., Ogata, E., and Kojima, I. (1987a). Insulin-like growth factor II stimulates calcium influx in competent BALE/c 3T3 cells primed with epidermal growth factor. J. Biol. Chem. 262, 1212012126. Nishimoto, I., Ogata, E., and Kojima, I. (1987b). Pertussis toxin inhibits the action of insulin-like growth factor I. Biochem. Biophys. Res. Commun. 748, 403-411. Nishimoto, I., Murayama, Y., Katada, T., Ui, M., and Ogata, E. (1989). Possible direct linkage of insulin-like growth factor-11 receptor with guanine nucleotide-binding proteins. J. Biol. Chem. 264, 14029-14038. Nissley, S. P, Haskell, J. F., Sasaki, N., De Vroede, M. A., and Rechler, M. M. (1985). Insulin-like growth factor receptors. J. Cell. Sci. (Suppl.) 3, 39-51. Nolan, C. M., Kyle, J. W., Watanabe, H., and Sly, W. S. (1990). Binding of insulin-like growth factor II (IGF-II) by human cation-independent

G$tein-Activating

mannose deficient

Segment

of IGF-IllmanGP

Receptor

Sphosphate receptor/lGF-II receptor expressed mouse L cells. (1990). Cell Reg. 7, 197-213.

in receptor-

Oinuma, M., Katada, T., and Ui, M. (1967). A new GTP-binding protein in differentiated human leukemic (HL-60) cells serving as the specific substrate of islet-activating proteins, pertussis toxin. J. Biol. Chem. 262, 0347-0353. Okamoto, T., Nishimoto, I., Murayama, Y., Ohkuni, Y., and Ogata, E. (1990). Insulin-like growth factor-lllmannose Sphosphate receptor is incapable of activating GTP-binding proteins in response to mannose Sphosphate, but capable in response to insulin-like growth factor-II. B&hem. Biophys. Res. Commun. 168, 1201-1210. Richman, R. The effect of blast growth: Cell. Physiol.

A., Weiss, J. P, Roberts, S. B., and Florini. J. R. (1960). serum and multiplication stimulating activity on L6 myothe lack of correlation with cyclic nucleotide changes. J. 103, 63-69.

Roth, R. A. (1966). Structure of the receptor for insulin-like growth factor II: the puzzle amplified. Science 239, 1269-1271. Rothenberg, P. L., and Kahn, C. R. (1968). Insulin inhibits pertussis toxin-catalyzed ADP-ribosylation of G-proteins: evidence for a novel interaction between insulin receptors and G-proteins. J. Biol. Chem. 263, 15546-15552. Rubenstein, R. C., Wong, S. K.-F., and Ross, E. M. (1967). The hydrophobic tryptic core of f3-adrenergic receptor retains G, regulatory activity in response to agonists and thiols. J. Biol. Chem. 262, 16655 16662. Sibley, D. R., Benovic, J. L., Caron, M. G., and Lefkowitz, R. J. (1967). Regulation of transmembrane signaling by receptor phosphorylation. Cell 48, 913-922. Weiss, E. R., Kelleher, D. J.. and Johnson, G. L. (1988). of interaction between rhodopsin and transducin using tipeptide antibodies. J. Biol. Chem. 283, 6150-6154.

Mapping sites rhodopsin an-

Yamashita, T., Yamamoto, K., Kikuchi, A., Kawata. M.. Kondo, J.. Hishida, T., Teranishi, Y., Shiku, H., and Takai, Y. (1986). Purification and characterization of c-Ki-ras p21 from bovine brain crude membranes. J. Biol. Chem. 283, 17161-17166. Zimniak, P, and Racker, E. (1976). Electrogenicity catalyzed by the Caz+ ATPase from sarcoplasmic Chem 253, 4631-4637.

of Ca*+ transport reticulum. J. Biol.