Molecular and Cellular Endocrinology, 14 (1990) 133-141 Elsevier Scientific Publishers Ireland, Ltd.

MOLCEL

133

02395

Occurrence,

quaternary structure and function of G protein subunits in an insect endocrine gland Victoria H. Meller and Lawrence

I. Gilbert

Department of Biology, CB 3280 Coker Ha/l, University of North Carolina, Chapel Hill, NC 27599, U.S.A. (Received

Key words: Adenylate

cyclase;

Prothoracic

gland;

2 April 1990; accepted

28 August

1990)

Drosophila melanogaster; Manduca sexfa

Summary The occurrence, structure and function of the (Yand p subunits of GTP-binding proteins (G proteins) were investigated in the Munduca sexta prothoracic gland, a tissue which possesses a hormonally regulated adenylate cyclase. Subunit-specific antibodies were utilized in immunoblotting studies of tissue from Munduca prothoracic glands, brain, eyes and antennae, and compared to the substrates present in the heads of Drosophila, as well as in a mammalian cell line. All Manduca tissues examined showed putative Gp subunits of 37 and 38 kDa, an unidentified (Y subunit of 41 kDa, in addition to an eye specific (r subunit of 42 kDa. Munduca tissues also produced putative G,, subunits of 48 and 51 kDa which were coupled to prothoracic gland adenylate cyclase as demonstrated by immunoprecipitation. Prothoracic gland G proteins have a definite and limited quaternary structure, consistent with a heterotrimeric model, as demonstrated by crosslinking of prothoracic gland membrane preparations followed by immunoblotting. These studies also yielded data on relative titers of (Y subunits, and suggest that G,, is present in lower amounts than other (Y subunits. The G protein subunits studied in the prothoracic gland appear strikingly similar in molecular weight, function and structure to their mammalian counterparts.

Introduction Guanine nucleotide binding proteins (G proteins) are a family of signal transducing proteins that couple receptors to their effector systems and have a heterotrimeric structure comprised of (Y, p, and y subunits. Upon interaction with an activated receptor, the (Ysubunit exchanges bound GDP for GTP, is thought to dissociate from the j3y complex, and can activate its effector enzyme(s). For

Address for correspondence: Lawrence I. Gilbert, Department of Biology, CB 3280 Coker Hall, University of North Carolina, Chapel Hill, NC 27599, U.S.A.

0303-7207/90/$03.50

0 1990 Elsevier Scientific

Publishers

Ireland,

adenylate cyclase, activation is mediated by the stimulatory G protein, G,, while inhibition is mediated through other G proteins, termed Gil-j, which differ from G, in their 01 subunits. The G proteins appear to be highly conserved, as subunits have been identified in Drosophila melanogaster which are homologous to bovine GO (Yarfitz et al., 1988), Gi, (Provost et al., 1988), GO, (de Sousa et al., 1989; Schmidt et al., 1989; Thambi et al., 1989; Yoon et al., 1989) and G,, (Quan et al., 1989). Insect adenylate cyclase appears to be regulated by G proteins as shown by NaF, cholera toxin and nonhydrolyzable guanine nucleotide analog mediated stimulation of this enzyme (Combest et al., 1985; Guillen et al., 1987; Meller et al., Ltd.

134

1988). These treatments all elicit their effects by persistent activation of the G,, subunit. The prothoracic gland is an insect endocrine gland which produces steroid hormones, ecdysteroids, required for molting and metamorphosis, and is regulated by neuropeptides (prothoracicotropic hormones, or PTTHs) acting through CAMP. Signal transduction in the glands of the tobacco hornworm, Mund~c~ sextu, is dependent on extracellular Ca2+ (Smith et al., 1985) which can regulate a calmodulin (CaM)-dependent adenylate cyclase in this tissue (Meller et al., 1988; see Gilbert et al., 1988). Indications that CaM regulates G protein function in addition to its presumed direct effect on the catalytic unit of adenylate cyclase led us to examine the components of this signal transducing system in the prothoracic gland. The developmental aspects of this system are also intriguing, as some components of adenylate cyclase sensitivity to Ca’+/CaM are lost halfway through the final larval instar, between days 5 and 6 (Meller et al., 1990). The aims of these studies were to identify G protein subunits in the Munduca glands and other Munduca tissues, comparing them to G proteins of a mammalian cell line and another insect, ~rosophiI~ ~ze~~noguster. We specifically sought to identify any developmental differences of these proteins in the prothoracic gland, and establish the relationship of a putative Manduca G,, protein to its proposed effector, adenylate cyclase. Materials and methods

Animals Manduca sexta were reared individually on artificial diet at 26”C, high hu~dity ( > 60%), and under a nondiapausing light regimen (LD 16 : 8) (Vince and Gilbert, 1977). Insects were staged by a combination of developmental markers, timing and weight (see Goodman et al., 1985). Only day 3 animals weighing between 5 and 7 g and day 4 animals between 8 and 10 g were used. Day 5 animals were chosen by observing the exposed dorsal vessel and pink coloration on the dorsal mid-line, and days 6 and 7 were staged on day 5. All dissections were performed on the days indicated before 1 p.m.

Tissue preparation Prothoracic glands were dissected out under lepidopteran saline (Weevers, 1966), frozen immediately on dry ice, and stored at - 85°C until use. Particulate fractions were obtained by differential centrifugation of prothoracic gland homogenates as described previously (Meller et al., 1988). The human astrocytoma cells (line 132INI) and the purified bovine brain G, was a gift from Dr. M. Martin, University of North Carolina, Chapel Hill. Reagents Nitrocellulose was purchased from Gelman Scientific and MSI, alkaline phosphatase conjugated goat anti-rabbit secondary antibody, NBT (nitroblue tetrazolium) and BCIP (5-bromo-4-chloro-3indoylphosphate) were from Sigma. Immunoprecipitation was done using a preparation of lysed St~phyiococcus aureus cells (Warren and Gilbert, 1988). Antibody and antisera sources Antiserum 5881, raised to a portion of the GTP binding domain of G, subunits, was the gift of Drs. Susanne Mumby and A. Gilman, University of Texas Medical Center, Dallas, TX (see Mumby et al., 1986). It was used at a dilution of 1 : 500. Antiserum 708 R (G,, specific) and antibody 83, an affinity purified preparation of this antiserum (see antibody RM, Simonds et al., 1989), and antiserum J99 (GB specific; see SW/l, Spiegel, 1990) were the gifts of Dr. Alan Spiegel, NIH, Bethesda, MD. Antiserum 708 R was used at a dilution of 1 : 800, J99 at a dilution of 1 : 400, and the purified antibody 83 was used at a concentration of 10 pg,/ml. Western blot techniques Sodium dodecyl sulfate (SDS) gels were blotted onto nitrocellulose, rinsed briefly in Tris buffered saline with 0.2% Tween-20 (TBST), and blocked in this buffer with 0.5% skim milk and 2.5 mM sodium azide (blocking buffer) for 20 min. Blots were then incubated overnight in the appropriate primary antibody diluted in blocking buffer. Blots were rinsed 4 times for 5 min each with TBST and incubated for 1 h in secondary antibody at a 1 : 1000 dilution in blocking buffer. Blots were

135

ice with frequent vortexing for 20 min followed by centrifugation for 20 min at 100,000 X g. The resultant supernatant was used to initiate the adenylate cyclase assay or was subjected to immunoprecipitation.

washed in TBST as before, rinsed in distilled water and developed in an NBT and BCIP color reagent.

Cross&king Crosslinking was preformed by diluting very small samples of washed prothoracic gland particulate fractions (50-100 pg protein) in 0.5 ml 3% formaldehyde. Optimal crosslinking was achieved with a 10 min incubation on ice followed by centrifugation at 30,000 X g for 15 min. The pellet was suspended in SDS sample buffer, boiled for 2 min and run on SDS gels at 40 yg/lane. These gels were blotted onto nitrocellulose and probed with antibodies as described above.

A 70 ~1 aliquot of lubrol extract was incubated for 1 h at 4°C with the indicated amounts of antiserum and preimmune serum. A 20 (11 preparation of lysed S. aureus cells (Warren and Gilbert, 1988) was added to the mixture and incubation continued for another hour. The tubes were then centrifuged at 12,000 X g, the supernatant was removed and the pellet was washed once and resuspended in solubilization buffer. Aliquots (10 ~1) of these preparations were used to initiate the adenylate cyclase assay which was allowed to proceed for 1 h.

Adenylate cyclase assay All reagents for performing the adenylate cyclase assay as well as the assay itself were as described previously (Meller et al., 1988). Briefly, the assay involves the conversion of [a- 32P]ATP to [ 32P]cAMP by the tissue fraction, and the chromatograp~c separation of labeled CAMP from precursor. Recoveries are monitored by reading the OD at 259 nm of an unlabeled CAMP internal standard and are about 70%.

Protein determinations The protein content of tissue fractions was determined as described by Lowry et al. (1951).

Results

Lubrol-PX solubilization of particulate proteins

Antibody recognition of G protein subunit homologs

Washed prothoracic gland particulate preparations were suspended in a detergent solubilization buffer containing 0.5% Lubrol-PX, 10 mM MgCl,, 10 mM sodium potassium phosphate buffer (pH 7.2), 1 mM dithiothreitol (DTT), 1 mM EDTA, and 100 PM leupeptin. Concentrations of between 3 and 5 mg particulate protein/ml lubrol buffer were used. When tissue was being prepared for immunoprecipitation, 5 mM NaF was added to the washing buffer. The suspension was kept on

Data are presented in this paper that use three antisera and an antibody that exhibited crossreactivity with Manduca and Drosophila proteins in the anticipated molecular weight range of G protein subunits. These were all raised in rabbits to synthetic peptides representing epitopes of mammalian G proteins, and are described in greater detail in Table 1. The Western blots in Fig. 1 represent the interactions of three of these with washed particulate fractions from a human astro-

TABLE

1

ANTIBODIES

UTILIZED

All antibodies were raised in rabbits to the indicated an affinity purified version of antiserum 708 R. Number J99 708 R 83 (purified) 5881

synthetic

peptides.

Target

Antigen

Ga GW G,, G*

Carboxy terminal 11 Carboxy terminal 10 Carboxy terminal 10 GTP binding domain

All were used as antiserum,

except antibody

83, which was

Reference ammo ammo amino of G,,

acids of Gs acids of G, acids of G, 15 ammo acids

Spiegel (1990) Simonds et al. (1989) Simonds et al. (1989) Mumby et al. (1988)

136

cell line, adult Drosophila heads, and brain and prothoracic glands from last instar larrse as well as eyes and antennae from newly emerged adults. The top panel shows crossreactivity with antiserum J99 to the G0 subunit. A sample of purified bovine brain Gi shows one band at 38 kDa while the astrocytoma preparation yields a single band at 31 kDa. The Drosophila head particulate fraction displays less immunoreactivity, producing a faint band at 38 kDa, and all of the Manduca tissues show two bands at 38 and 37 kDa, with the brain being the best source, followed by the prothoracic glands. The second panel of Fig. 1 summarizes the immunoreactivity of these tissue preparations to antiserum 5881, which recognizes all G, with the exception of G,,. The Gi standard shows a single band at 42 kDa with this antiserum (data not shown). The astrocytoma preparation yields a doublet at 42 kDa and a single band at 41 kDa, while the Drosophila preparation is typified by two very close bands at 42 kDa, with the faster migrating protein staining more heavily. All Manduca tissues show a band at 41 kDa, and a cytoma

Manduca

- 99

- 93 - 87 - 81

- 51 -48

0

1

234567

30

c

at 42 kDa is seen clearly in the particulate preparation. The eye preparation also shows staining due to either nonspecific interactions or crossreacting proteins. One of these, a band at 48 kDa, is visible in lane 6. The results with antibody 83 against G,, are displayed in the bottom panel of Fig. 1. The astrocytoma cell preparation reveals two bands at 50 and 47 kDa and the Drosophila head material yields two broad bands at 54 and 51 kDa, with the former being considerably stronger. Each of these bands appears to be composed of multiple proteins very close in M,. All Manduca tissues show equivalently staining bands at 51 and 48 kDa which appear to resolve into doublets when separated on very low percentage gels (see lane C, Fig. 2). Separation of these antigens on two-dimensional SDS gels confirms that the 48 and 51 kDa bands are each composed of multiple immunoreactive species (Meller, 1990). Manduca

0

10

Fig. 2. Crosslinking of G,, proteins from prothoracic glands. Lanes were loaded with 40 pg of a crosslinked particulate preparation from day 6 glands. Lane 5 is an uncrosslinked sample, and lanes l-4 have been crosslinked with 3% formaldehyde for 0, 5, 10 and 30 min, respectively, as described in Materials and Methods. The proteins were blotted and probed with antibody 83 which recognizes G,,. Molecular weights x 10-s are shown on the right.

second

Fig. 1. G protein subunit immunoreactivity in particulate preparations. Protein samples in lanes are as follows: lane 0, purified Gi standard; lane 1, 15 pg astrocytoma; lane 2, 30 pg Drosophila head; lane 3, 15 pg day 3 Manduca brain; lane 4, 30 pg day 4 Manduca prothoracic gland; lane 5. 30 pg day 7 Manduca prothoracic gland; lane 6, 30 pg Manduca eye; lane 7, 30 pg Manduca antennae. The top panel is probed with antiserum J99 which recognizes G,; the middle panel is probed with antiserum 5881 which recognizes G, subunits; and the lower panel is probed with antibody 83, which recognizes G,,. Molecular weights X 10e3 are shown on the right.

5

band

eye

137

Crosslinking of G protein subunits We wished to determine if the quaternary structure of insect G proteins agreed with the heterotrimerit model proposed in vertebrates, and also hoped to gain indirect evidence for the relative amounts of the G, subunits in the prothoracic gland preparations. Chemical crosslinking of tissue preparations was performed and the products analyzed by Western blotting. The presumptive G,, protein was the easiest to follow because of the high quality of antibody 83 and because it reacts with an epitope that is apparently not obscured by crosslinking (the carboxy terminus). However, the two antisera used, J99 and 5881, also provided reproducible data on the molecular weights of crosslinked proteins. Although prothoracic glands from days 1, 5, 6, 7 and 9 were analyzed in this manner, no developmental differences in crosslinked products were detectable with the antibodies used. Fig. 2 is a time course of the crosslinking reaction performed on particulate fractions from day 6 prothoracic glands and probed with the antibody 83. Four products of M, ranging from 81 to 99 kDa are produced with an optimal reaction time of 10 min. Apparently, further crosslinking results in insoluble protein aggregates which do not enter the SDS gel, as demonstrated by silver staining (data not shown). Table 2 summarizes the crosslinking data obtained using all three antibodies and antisera, all the

TABLE

2

PRODUCTS

OF CROSSLINKING

The apparent molecular weights of bands observed after crosslinking and probing with the indicated antibodies are presented along with the estimated molecular weight of protein(s) they are linked to. Numbers represent the mean M, of between three and five observations, all of which have calculated SE d 1 kDa. Antibody

W (kDa)

Calculated M, of linked proteins (kDa)

83

99 93 87 81 12 80 73

48 42 36 30 31 42 35

5881 J99

or or or or

51 45 39 33

or 43 or 36

TABLE

3

RESPONSE

OF LUBROL-PX

EXTRACT

TO 5 mM NaF

The activation ratios were calculated by dividing NaF stimulated activity by unstimulated activity. The value for the Lubrol-PX extract itself was from a separate experiment to assess the extraction method. Activation ratios+SE were calculated for the supernatant and pellet fractions from Fig. 3. Preparation

NaF activation

Lubrol-PX extract Supematant Precipitate

6.3 + 0.3 2.5 + 0.3 2.0 f 0.2

ratio

reported bands being observed in at least three experiments. A single band was observed at 72 kDa using the 5881 antiserum for G,, and two were seen at 73 and 80 kDa when using the J99 antiserum as a probe for Gp. The apparent A4, of the unknown linked proteins in Table 2 is calculated by subtracting the apparent M, of the recognized antigen(s). Immunoprecipitation of adenylate cyclase activity with antibody to G,, The results of Western blotting indicated that the strongest G,, candidates were of M, 48 and 51 kDa, and the following experiments were undertaken as a first step in resolving the function of these putative G,, proteins. Antiserum 708 R and antibody 83 were used as immunoprecipitants in a detergent extract of prothoracic glands, and the fractionated preparations assayed for adenylate cyclase activity. This extract, made by incubating a particulate preparation of day 7 glands with a 0.5% Lubrol-PX buffer and then centrifuging at 100,000 X g, has functional adenylate cyclase-G protein complexes as shown by the response of extracted adenylate cyclase to NaF (first line of Table 3). In an effort to encourage association between G,, and adenylate cyclase, tissue was treated with 5 mM NaF during preparation for the immunoprecipitation experiments, which resulted in the lower NaF activation ratio seen in lines 2 and 3 of Table 3. Western blot analysis of extracts reveal that 50% of the 48 and 51 kDa proteins is solubilized by this method, but that most of the particulate protein remains in the 100,000 X g pellet.

138

pi antiserum )II prslmmune s8rum

0

0.1

0.3

1

1

0.9

0.7

0

Fig. 3. Immunoprecipitation of adenylate cyclase activity with antibodies to G,,. A Lubrol-PX solubilized preparation of day 7 glands was immunoprecipitated with the indicated amounts of antiserum 708 R and preimmune serum. The precipitate (0) and supernatant (0) were then assayed for NaF stimulated adenylate cyclase activity, and the results plotted as percent of recovered activity. The insert shows the same procedure done with 1.72 ng of antibody 83 (no serum added); S, enzyme activity in supernatant; P, enzyme activity in precipitate.

The data revealed an antiserum-dependent concentration of adenylate cyclase activity in the precipitate (Fig. 3). The insert in Fig. 3 shows the results of performing the immunoprecipitation with 1.72 ng of affinity purified antibody in the absence of serum. Both the antiserum and the affinity purified antibody strongly inhibit adenylate cyclase activity, with purified antibody and the highest concentrations of serum blocking the total recoverable activity by 70% and 40%, respectively. Despite this, there is a striking increase in enzyme activity associated with the pellet fraction when 1 ~1 of antiserum is used (over 20-fold in the experiment shown), indicating the possibility of a functional coupling between the 48 and/or 51 kDa proteins and adenylate cyclase. Discussion Antibodies raised to mammalian G protein peptides crossreact with proteins in insect tissues The crossreactivity observed here is expected in light of the very close homologies that have been reported between mammalian G proteins and their insect homologs. Indeed, antibodies raised to conserved regions of mammalian G proteins have been shown to crossreact with a variety of insect

tissues from animals of diverse orders (Raming and Breer, 1990; Wolfgang et al., 1990; Quan and Forte, 1990). Antiserum J99 to Gp reacts very strongly with a pair of Manduca proteins, but a two amino acid difference between the Drosophila and bovine protein in the carboxy terminal 10 amino acids (Yarfitz et al., 1988) might explain the apparently reduced immunoreactivity of Drosophila tissue with J99, i.e. some blots fail to show the band seen in lane 2, Fig. 1. The fact that only a single band can be detected in Drosophila is consistent with reports that only a single Gp gene exists in Drosophila (Yarfitz et al., 1988). Since Manduca tissues yield two bands of M, 37 and 38 kDa, it appears that not all insects produce a single form of Gp. The conservation of (Y subunits, including Drosophila G,, (de Sousa et al., 1989; Schmidt et al., 1989; Thambi et al., 1989; Yoon et al., 1989) and G,, (Provost et al., 1988) in the region of the guanine nucleotide binding domain to which the antiserum 5881 is raised guarantees a broad range of targets for this antiserum. This antiserum has a low affinity for G,,, probably due to the substitution of arginine for lysine at a site corresponding to the carboxy terminal position of the antigenic peptide (Mumby et al., 1986), a change shared by Drosophila G,, (Quan et al., 1989). As G,, is usually present at much lower levels than G,, and G,,, 5881 is not expected to recognize G,, in tissue preparations such as used here. A purified standard consisting of bovine brain G, yields one 42 kDa band when probed with this antiserum, but multiple proteins were recognized by 5881 in the astrocytoma preparation, consisting of a very close doublet at 42 kDa and a single band at 41 kDa. The doublet at 42 kDa observed here is probably a resolution of the Gi, previously reported in this cell line (Buss et al., 1987) and may represent the two Gi, subunits common in brain, Gi, and Gi, (40-41 kDa, see Goldsmith et al., 1988). Although the 40 kDa protein has the approximate M, and abundance of G,, (39 kDa, see Spiegel et al., 1988) its identity is conjectural at this point. A G,, subunit has been previously identified by immunological and biochemical methods in insect neural tissues (Homburger et al., 1987; Hopkins et al., 1988; Wolfgang et al., 1990). Mes-

139

senger RNA from the ~r~~o~~~la homolog of G,, is preferentially expressed in neural tissue of this insect (de Sousa et al., 1989; Schmidt et al., 1989; Thambi et al., 1989; Yoon et al., 1989), a finding which has been verified by immunolocalization (Wolfgang et al., 1990). Probing of Drosophila head preparations with antiserum J881 yields a doublet at 42 kDa, with the faster migrating band being much more intense. Manduca tissues yield two different bands with this antiserum, a 41 kDa protein found in all tissues examined, but most prominently in the brain and prothoracic glands, and a 42 kDa protein which is the most intense band of the eye fraction. By analogy with previously identified G protein homologs in insect neural tissue, it appears reasonable that the prominent lower band of the doublet seen in Drosophila, and the 41 kDa band seen in Munduca tissues is insect G,,. The identity of this 41 kDa protein in the prothoracic glands is important, as it may ultimately yield clues as to how this tissue is regulated. There have been no identified inhibitory regulators of gland adenylate cyclase, although the opening of ligand sated Ca2+ channels has been proposed as the primary action of PTTH on this tissue (see Gilbert et al., 1988 for review), and involvement of phosphatidylinositol hydrolysis has not been ruled out. Both G,, and G,, have been implicated in the regulation of ion channels and phospholipase C activity (Freissmuth et al., 1989). The presence of signal transducing G proteins in the eyes of insects and other invertebrates has been established bioche~cally (Saibil and MichelVillaz, 1984; Blumenfeld et al., 1985; Fein, 1986). For example, octopus photoreceptors yield a 41 kDa protein that is labeled by pertussis toxin in a light-inhibited manner similar to bovine transducin (Tsuda et al., 1986), and a 41 kDa protein was identified in housefly (Musca) eye membranes by light activated binding of a GTP analog (Devary et al., 1987). This suggests that the immunoreactivity associated with the 42 kDa protein from Drosophila heads and Manducu eyes observed in the present study may be due to a signal transducing G protein invoIved in vision. The carboxy terminal region of the G,, protein which served as the antigen for antiserum 708 R is identical in Drosophila and bovine brain (Quart et

al., 1989). This antiserum labels multiple bands in all the tissues examined here, the astrocytoma preparation showing bands of 50 and 47 kDa, Drosophila head fractions producing very strong bands at 54 kDa and less intense ones at 51 kDa, and Munduca tissue preparations producing bands of 48 and 51 kDa. By alternate splicing of mRNA, the mammalian G,, gene gives rise to four proteins migrating as two bands on one-dimensional SDS (Robishaw et al., 1986; Graziano et al., 1989; Mattera et al., 1989). Alternately spliced mRNAs have also been demonstrated for the Drosophila G,, gene, which produce products with apparent mobilities of 48 and 51 kDa when translated in vitro (Quan and Forte, 1990). The consistent observation of two bands in Munducu tissues and the resolution of these bands into numerous proteins by two-dimensional electrophoresis suggests that either multiple genes or alternate splicing of mRNA is occurring in this insect also. Crosslinking patterns of G protein homoiogs The experiments in which crosslinking was fol-

lowed by immunoblott~g provided two potentially important types of information about G proteins in the prothoracic gland. The first was an estimation of the quaternary structure of G proteins in the Munduca prothoracic gland. There is a definite crosslinking pattern produced by each of the subunits examined which indicates that they have a limited number of close associations with other proteins. A general absence of products larger than 100 kDa leads to the conclusion that prothoracic gland G proteins do not form large, oligomeric complexes, as has been proposed in some systems (see Rodbell, 1980). The anomalous migration of crosslinked proteins makes analysis based on the molecular weights of the linked proteins difficult. However, the deduced sizes of crosslinked proteins are consistent with the (rj?y heterotrimeric structure. For example, the putative G,, subunits shown in Fig. 2 and Table 2 are linked to a protein the size of Gp (37 or 38 kDa) or the Gpu complex (about 48 kDa). A recent report indicates that purified G,, also binds very tightly to tubuhn (Wang et al., 1990), raising the possibility that some of the bands noted here could result from the crosslinking to tubulin rather than to By subunits, both of which are expected to

140

produce mobility shifts of about 50 kDa. If so, this would indicate that a large portion of the G,, in the ~an~~cu preparations binds endogenous tubulin. In addition to information on the quaternary structure of the insect G proteins, crosslinking and immunoblotting experiments provided information on the relative abundances of G, protein subunits, even in the absence of information on the antibody sensitivity to insect proteins. In mammalian tissues G,, and Gi, are in much greater abundance than G,, (Gilman, 1987). G,, is especially abundant in brain, being in &fold excess over Gi, (Mil~gan et al., 1987). If prothoracic gland G proteins are in heterotrimeric form and share a colon pool of G, protein as has been proposed for vertebrate systems (see Gilman, 1987; Spiegel et al., 1988; Freissmuth et al., 1989), then crosslinking followed by probing with the G, antiserum should show Gfi linking primarily to the most abundant G, subunits. Our results reveal that Gp crosslinks to proteins of M, ranging from 35 to 43 kDa, much smaller than is expected for G,,, and suggests a stoichiometric excess of smaller subunits, perhaps Gi, and G,,, over the larger Gsa subunits. Thus, it appears that both the quatemary structure and the relative amounts of the different G, subunits are conserved between invertebrates and vertebrates in addition to the striking homologies seen within the sequences of the subunits themselves. Antibody to the 48 and 51 kDa proteins immunoprecipitates adenylute cyclase activity The immunoprecipitation experiments were performed to identify the effector of the 48 and 51 kDa G,, protein homologs identified in ~~nducu prothoracic glands. The fact that antibodies to these proteins interacted with the functional G,,adenylate cyclase complex to precipitate it, as evidenced by increased adenylate cyclase activity in the precipitate, demonstrates that these G,, subunits do regulate adenylate cyclase. The lower NaF activation ratios in the supematant and precipitate are presumably due to the fact that 5 mM NaF is a component of the buffer used in tissue preparation, and therefore the G proteins involved are partially preactivated. A similar paradigm was used to demonstrate the coupling of G,, to adenyl-

ate cyclase in bovine brain preparations, with very comparable results (Simonds et al., 1989). The carboxy terminus of Dro~ophiIa (Quan et al., 1989) and bovine G,, are virtually identical, a property which is responsible for the avid labeling of insect Gscvby antibody 83. As this is the region of the G protein proposed to interact with the receptor (Bimbaumer et al., 1987) we propose that the Manduca prothoracic gland GsU homolog described here is also receptor regulated. To date, no activators of this gland other than the P’ITHs have been described, and these do not appear to be coupled directly to adenylate cyclase, but as noted previously, have been proposed to open Ca2+ channels, thus stimulating adenylate cyclase indirectly (see Gilbert et al., 1988; unpublished observations). This leads to the postulate that an as yet unidentified regulator of prothoracic gland function is coupled to adenylate cyclase through the identified G,, subunits of M, 48 and 51 kDa, and that PTTH either regulates ion channels or phospholipase C activity, possibly by way of the 41 kDa putative G,,. Antigenically recognizable G,, homologs from Manduca prothoracic glands are of almost identical molecular weight to those of astr~ytoma cells and are coupled to the insect’s adenylate cyclase. We have presented evidence that their patterns of expression, i.e. relative amounts of the different G, subunits, are similar to those proposed for vertebrates (Gilman, 1987). These observations suggest that insect signal transduction systems are strikingly similar to those of vertebrates, both in composition and function. The unique advantages of systems exploiting insect development and genetics will hopefully be brought to bear on the questions of signal transduction.

We thank Drs. S. Mumby, A. Gilman and A. Spiegel for their generous gifts of antibodies and helpful suggestions, Pam Kitchen and Dr. W. Combest for their assistance and advice, Susan Whitfield for graphics and Dr. M. Martin for the gifts of astrocytoma cells and purified G protein subunits. This research was supported by grants DCB-8802108 from the National Science Foundation and DK-30118 from the National Institutes of Health.

141

References Bimbaumer, L., Codina, J., Mattera, R., Yatani, A., Scherer, N.. Toro, M.-J. and Brown, A. (1987) Kidney Int. 32, s-14-s-37. Blumenfeld, A., Erusahmsky, J., Heichal, O., Selinger, Z. and Minke. B. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7116 7120. Buss, J.E., Mumby, S.M., Casey, P.J., Gilman, A.G. and Sefton, B.M. (1987) Natl. Acad. Sci. U.S.A. 84, 7493-7497. Comb&, W., Sheridan, D. and Gilbert, L.I. (1985) Insect Biochem. 15,579-588. de Sousa, S., Hoveland, L., Yarfitz, S. and Hurley, J. (1989) J. Biol. Chem. 264, 18544-18551. Devary, O., Heichal, O., Blumenfeld, A., Cassel, D., Suss, E., Barash, S., Rubinstein, C., Minke, B. and Sehnger, Z. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 6939-6954. Fein, A. (1986) Science 232, 1543-1545. Freissmuth, M., Casey, P. and Gilman, A. (1989) FASEB J. 3, 2125-2131. Gilbert, L.I., Combest, W., Smith, W., Meiler, V. and Rountree, D. (1988) BioEssays 8, 153-157. Gilman, A. (1987) Annu. Rev. Biochem. 56, 615-649. Goldsmith, P., Rossiter, K., Carter, A., Simonds, W., Unson, C., Vinitsky, R. and Spiegel, A. (1988) J. Biol. Chem. 263, 647666479. Goodman W., Carlson, R. and Nelson, K. (1985) Ann. Ent. Sot. Am. 78,70-80 Graziano, M., Freissmuth, M. and Gilman, A. (1989) J. Biol. Chem. 264, 409-418. Guillen, A., Haro, A. and Municio, A. (1987) Arch. B&hem. Biophys. 254. 234-240. Hornburger, V., Brabet, P., Audigier, Y., Pantaloni, C., Bockaert, J. and Rouot, 8. (1987) Mol. Pharmacol. 31, 313-319. Hopkins, R., Stamnes, M., Simon, M. and Hurley, J. (1988) Biophys. Biochim. Acta 970, 355-362. Lowry, D., Rosebrough, M., Farr, A. and Randall, R. (1951) J. Biol. Chem. 193, 265-275. Mattera, R., Graziano, M., Yatani, A., Zhou, Z., Graf, R., Codina, J., Bimbaumer, L., Gihnan, A. and Brown, A. (1989) Science 243, 804-807. Meller, V. (1990) Regulation of Adenylate Cyclase by Calcium and Calmodulin in an Insect Endocrine Gland, Thesis, University of North Carolina, Chapel Hill. Meller, V., Combest, W., Smith, W. and Gilbert, L.I. (1988) Mol. Cell. Endocrinol. 59, 67-76. Meller, V., Sakurai, S. and Gilbert, L.I. (1990) Cell Regul. (in press). Milligan, G., Streaty, R., Gierschik, P., Spiegel, A. and Klee. W. (1987) J. Biol. Chem. 262, 8626-8630.

Mumby, S., Kahn, R., Manning, D. and Gilman, A. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 265-269. Provost, N., Somers, D. and Hurley, J. (1988) J. Biol. Chem. 263, 12070-12076. Quan, F. and Forte, M. (1990) Mol. Cell. Biol. 10, 910-917. Quan, F., Wolfgang, W. and Forte, M. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 4321-4325. Ram.& K. and Breer, H. (1990) Comp. B&hem. Physiol. 95B, 861-864. Robishaw, J., Smigel, M. and Gilman, A. (1986) J. Biol. Chem. 216, 9587-9590. Rodbell, M. (1980) Nature 284, 17-22. Saibil, H. and Michel-Villaz, M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 5111-5115. Schmidt, C., Garen-Fazio. S., Chow, Y.-K. and Neer, E. (1989) Cell Regul. 1, 125-134. Simonds, W., Goldsmith, P., Woodard, C., Unson, C. and Spiegel, A. (1989) FEBS Lett. 249, 189-194. Smith, W., Gilbert, L.I. and Bollenbacher, W. (1985) Mol. Cell. Endoctinol. 39, 71-78. Spiegel, A.M. (1990) Antibodies as probes of the structure and function of heterotrimeric GTP-binding proteins. in ADPRibosylation Toxins and G-Proteins, in press (Moss, J. and Vaughn, M., eds.), American Society for Microbiology, Washington, DC. Spiegel, A., Carter, A., Brann, M., Collins, R., Goldsmith, P., Simonds, W., Vinitsky, R., Eide, B., Rossiter, K., Weinstein, L. and Woodard, C. (1988) Recent Prog. Harm. Res. 44, 337-375. Thambi, N., Quan, F., Wolfgang, W., Spiegel, A. and Forte, M. (1989) J. Biol. Chem. 264, 18552-18560. Tsuda, M., Tsuda, T., Terayama, Y., Fukada, Y., Akino, T., Yamanaka, G., Stryer, L., Katada, T., Ui, M. and Ebrey, T. (1986) FEBS Len. 198, 5-10. Vince, R. and Gilbert, L.I. (1977) Insect B&hem. 7, 115-120. Wang, N., Yan, K. and Rasenick, M. (1990) J. Biol. Chem. 26.5, 1239-1242. Warren, J. and Gilbert, L.I. (1988) in Immunological Techniques in Insect Biology (Gilbert, L.I. and Miller, T.A., eds.), pp. 181-214, Springer, New York. Weevers, R. (1966) 3. Exp. Biol. 44, 163-175. Wolfgang, W., Quan, F., Goldsmith, P., l&son, C., Spiegel, A. and Forte, M. (1990) J. Neurosci. 10, 1014-1024. Yarfitz, S., Provost, N. and Hurley, J. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 7134-7138. Yoon, J., Shortridge, R., Bloomquist, B., Schneuwly, S., Perdew, M. and Pak, W. (1989) J. Biol. Chem. 264, 1853618543.