Functional Analyses of Type V Adenylyl Cyclase

160

ADENYLYLCYCLASES

[ 131

[ 131 Functional Analyses of Type V Adenylyl Cyclase B y T A R U N B . P A T E L , C L A U S W I T T P O T H , A N N J. B A R B I E R ,

YINGES YIGZAW, and KLAUS SCHOLICH Type V and VI forms of adenylyl cyclase (AC) are closely related not only in their amino acid sequences but also in their regulatory properties (reviewed in Sunahara et al. 1 and lyengar2). In brief, both these isoforms of AC are stimulated by the a subunit of stimulatory G protein (G~c~) and are inhibited by the o~ subunit of inhibitory G protein (GLOW)as well as directly by calcium. In this article, we describe the methods to generate the full-length and engineered, soluble forms of canine type V adenylyl cyclase to investigate the regions on this enzyme that interact with each other and modulate the actions of G~c~. In addition, we describe the use of chimeric forms of adenylyl cyclase consisting of domains derived from bovine type I (ACI) and type V (ACV) enzymes to investigate the domains on these enzymes that interact with Gfi7 and Gic~, respectively; ACI, but not ACV, is inhibited by Gfiy (reviewed in Sunahara et a/. 1 and lyengar2). Moreover, we describe the methods utilized to uncover two additional functions of ACV, namely its GTPase-activating protein (GAP) activity against G~c~ and its ability to enhance receptor-mediated G T P - G D P exchange on a subunits of G~ and Gi. Given the aforementioned framework, this article is divided into the following major sections: expression of full-length and two soluble forms of ACV: use of fulllength and soluble forms of ACV to elucidate intramolecular interactions; use of chimeric forms of ACV and ACI to identify Gic*- and Gfiv-interacting sites: GAP activity of ACV: and guanine nucleotide exchange-enhancing activity of ACV. I. E x p r e s s i o n o f F u l l - L e n g t h a n d T w o S o l u b l e F o r m s of Type V Adenylyl Cyclase A. /ntroduction

A common feature shared by the various isoforms of AC is that the predicted structure for all isofomls is similar. Hence the AC molecule traverses the cell membrane 12 times and has 2 major cytosolic domains (referred to as C1 and C2, bracketed in Fig. 1A).t 3 The C I a and C2a subdomains within the two major cytosolic domains share some degree of similarity with each other and also show homology with the catalytic domain of guanylyl cyclases. 3 Notably, the C2b region is present only in ACI, ACII, ACIll, and ACVIII~; the slightly shorter I R. K. Sunahara, C. W. Dessaucr, and A. G. G i h n a n . Amlu. Rcr. Pharmacol. T¢zvicol. 36, 461 t 19961. 2 R. lyengar, FASEB ,/. 7, 768 (1993). W. J. Tang and A. G. G i h n a n , (;ell 71), 869 (1992). ('opyright " 2(102b~ AcademicIhe~ M[{I

HODS IN t!NZYMOLOGY,VO[. ~4"~

All l'ighl~ ol ,cpmdu~thm in [[11}~R~un iCSCT'~ed O(17n 6879/[)2S~.OI)

[ 13]

ANALYSF, S OF TYPE V ADENYLYI. CYCLASE

161

A Extracenular

/

M1

\ /

M2

\

E Intracellular +H3NJUa~

U ~._~

L_C1

C2 -Domain of ACV:

B

C2 I

Clb

R571

A322

CO0

__l

C1- Domain of ACV :

Cla

Membrane

I

F683

C21 I

$933995 10581091 1151 $1184

W e s t e r n Blot

Coomassie kD

97.6

<.,,, C1-C2

ACV

66

Cla-C2 ACV 45

~i:i~!!i!~I

31.5 . . . .

FK~. I. Schematic representation of ACV, its major subdomains, and immunoblot of soluble forms of C I - C 2 and C l a - C 2 R~rnls of ACV. (A) Schematic of the full-length ACV. The locations of the major cytosolic regions C l and C2 are shown in Ihc conlext of the whole molecule. The M1 and M2 regions, which span the membrane six times each, are also denoted. The lower panel depicts the

162

ADENYLYL CYCLASES

[ 13]

ACV does not have a C2b subdomain. Studies of naturally occurring, membranebound mammalian adenylyl cyclases have been difficult for the following reasons. First, because the regulatory properties of the different isoforms are varied I ) and second, because most mammalian cells express a mixture of isoforms, it is difficult to study a given isoform in isolation. Although this problem can be circumvented by expression of the desired AC isoform in insect cell lines such as Spodoptera frugiperda-9 (Sf9), the amount of protein that is expressed is low. Second, the membrane-bound adenylyl cyclases are extremely sensitive to detergents and, therefore, the purification of active enzymes has been difficult. Furthermore, when the active, full-length enzyme has been purified, it has been obtained only in small amounts and then the protein does not store well. Given these limitations, attempts have been made to express the protein subunits and determine whether the enzyme activity can be reconstituted. Interestingly, none of the two halves of the ACI, ACII, or ACV molecule when expressed alone (i.e., M ICI or M2C2 in Fig. I A) exhibit AC activity.45 However, coexpression of the two halves of the ACI, ACII, or ACV molecule 4'5 or the expression of C I a and C2a domains from ACI and ACII, respectively, joined by a linker~''7 demonstrate activity that can be stimulated by forskolin and GTP-bound Gsot.4 7 Given these findings, our objective was to engineer soluble forms of nonchimeric AC in which both the C 1 and C2 regions were derived from type V AC. We also compared the soluble forms of mammalian AC with the full-length ACV. In this section, we describe the methods we used to obtain the full-length canine ACV in Sf9 cells and the soluble counterparts of this enzyme in bacteria.

B. Full-Length Type V Adenylyl Cyclase Plasmid Construction and Generation of Baculovirus. cDNA encoding canine ACV was provided by Y. lshikawa, who originally cloned the enzyme. ~ To express the full-length enzyme using the baculovirus expression system, ACV cDNA in pCDNA39 is treated with Pvull and EcoRl to excise the 3' region of the eDNA. This Pvull-EcoR1 fragment is ligated into pBlueBacHis2A shuttle vector to generate a construct that we have designated 7.pBlueBacHis2A. The BamHI-XbaI fragment from ACV pCDNA3 is excised and inserted into the BamHI-XhoI sites of pBlueBacHis2B; the XbaI site on ACV and the Xhol site on pCDNA3 are blunted to permit ligation. This latter construct is utilized to excise an EcoRI-EcoRl fragment corresponding to the 5' end of the ACV cDNA. This fragment is then cloned into aminoacid residues that delinethe C Ia, C l b, C21, and C211subdomainswithinthc C I and C2 regions. (B) Coomassie-stainedgel and Western analysis of the two purified soluble lk)rmsof ACV, C1 C2 and CIa-C2. The two soluble forms of ACV were expressed in E. coli TP2000 as described in text. Supernatants from cell lysates were applied to Ni-NTA and Mono Q chromatography. Each purilied protein (0.5 ixg) was separated on 10% (w/v) acrylamide gels and Western analysis was perlZ~rmed with Anti-Xpressantibody,which recognizesthe epitope tag in the N terminusof each protein.

[ ] 3]

ANALYSES OF TYPE V AI)ENYI,YL CYCLASE

163

the EcoRl site of 7.pBlueBacHis2A. This ligation results in the insertion of the full-length ACV into the plasmid pBlueBacHis2A (ACV.pBlueBacHis2A). Although this construct is suitable for generation of recombinant virus we have opted to follow the procedure described below because of the convenience of using the Baculogold system (PharMingen, Carlsba& CA). The EcoRV HindllI fragment from this latter construct containing the ACV eDNA in-frame with a hexahistidyl (His(,) lag at the 5' end is then cloned inlo the baculovirus shulfie vector p2Bac. To obtain recombinant baculovirus, the shuttle vector is used in conjunction with Baculogold (PharMingen) to infect Sf9 cells; Sf9 cells are cultured at 27 in serum-free growth medium (GIBCO-BRL Gaithersburg, MD). After three rounds of amplilication, the recombinant virus is selected and used to infec! Sf9 cell cullures in 75-cm 2 flasks. The expression of ACV is monitored at various limes after infection by measuring the basal and forskolin-stimulaled AC activities in uninfecled and infected cells. Typically, cells expressing ACV have forskolin-slimulated AC activities that are at least four times higher than control cells. Once the condilions for a specific pool of virus are identified, larger amounts of cells are infected and membranes are harvested by the method of Kassis and Fishman. it) C. C I - C 2 a m / C l a

C2 fbml.~ (J' 7~vpe V Adeltvlyl Crclase

1. PlasmiJ Conslruction.

Both these forms o f soluble A C V are m o d e l e d on

lhe c h i m e r i c soluble a d e n y l y l cyclase c o m p r i s i n g the C I a region o f A C I linked to the C2a region o f type I1 A C described by Tang and G i h n a n . ~' Figure 1A

schematically represents the location of the C l, C la, alld C2 domains ol: ACV in the context of lhe full-length prolein. The expression vector pTrcHisB (lnvitrogen, San Diego, CA) is used for expression of both IBrms of hexahistidyltagged soluble adenylyl cyclases in E , w h e H c h i a coli. First. lhe (72 region (atnino acid residues 933-1 184 of canine type V adenylyl cyclase) is amplified, using ACV in pCDNA3 (lnvitrogen) as a template. The primers used for amplilication introduce a B qlll site at the 5' end of the polymerase chain reaction (PCR) product and a stop codon followed by a Hindlll reslriction site at lhe 3' end {primer sequences: 5'-ATATATAGATCTTCCACCGCCCGCCTCGAC and 5'-ATATATAAGCTTCTAACTGAGCGGGGG). The PCR product is chined into the B q,/II and Hindlll sites in pTrcHisB. The resulting plasmid is cul by' Bamttl and 4 W. J. Tang, M. Stanzcl, and A. G. Gihnan, Ifio<'h('mi~[cv 34, 14563 (1995). S. Kat>ushika, J.-l. Kay,abe. C. J. Ilomcy. and Y. Ishikavva,J. fliol. ('h(,nl. 268, 2275 (]993). (' W. J. fang mid A. G. Gilman, 5"ci(w~," 268, ]769 ( 1995). 7 C. W. I)cssauer and A. G. Gilm;.m, ./. Biol. C/mm. 271, ]6067 (1096). Y. l~lfikawa, K. Shuichi, L. ('hen. 'X .1. H;.llnon. ,I.-I. Kawabe, amt C. Homey. ,/. l~io/. Ch~'m. 267,

13553 (1992). '~Z. Chcn. It. S. Nicld. H. Sun, A. l:}albicr,and T. B. Pal¢l,,I. Biol. (Tlem. 2711,27525 (1995). ill S. Ka,,q~,and R A. Fishman,,I. Biol. Chem. 257, 5312(1984).

164

ADENYLYL CYCLASES

[131

BglII and a PCR product encoding either the C l a region (amino acid residues 964-1713) or the whole C I region (amino acid residues 964-2049) is inserted. The oligonucleotides used to generate the C 1a and C 1 regions introduce a unique BarnHI restriction site at the 5' end and a unique BgllI site and a 14-amino acid linker between CI (or C l a ) and the C2 region at the 3' end. The 5' oligonucleotide sequence is 5 ' - A T A T A T G G A T C C G C C T G A G G T C T C C C A G and the 3' oligonucleotide sequences are 5'-ATATATAGATCTCATACCTCCAGCTGCAGC T G G A G G C A T T C C A C C T G C A G C T G C C C G C T T C T G G G T G C A G C G C A G for C 1, and 5'-ATATATAGATCTCATACCTCCAGCTGCAGCTGGAGGCATTCCAC C T G C A G C T G C G A A T C G G T C A T C C A C C T G C T T for C la. All clones are sequenced to confirm in-frame ligations and correctness of the sequences. 2. Expression of Soluble Adenylyl Cyclases. Escherichia coli strain TP2000, which does not express any endogenous adenylyl cyclase activity, l i, 12 is used for expression of both C 1-C2 and C I a - C 2 forms of ACV. In our experience, the best expression of proteins is obtained when bacteria are freshly transformed. Bacterial cells transformed with a plasmid encoding C I - C 2 ACV or C I a - C 2 ACV are grown in Luria broth containing ampicillin ( 100 #g/ml) at 37' until they reach an OD(,00 nm of 0.4. Isopropyl-B-D-thiogalactopyranoside (|PTG) is then added to a final concentration of 0.1 m M and the cells are incubated ~br 15 hr at 23 = before they are pelleted by centrifugation (6000g for 5 min at 4 ) and stored at - 8 0 ' . Frozen cells are thawed in 1/100 culture volume of medium containing 50 m M Tris-HC1 (pH 8.0), 100 m M NaCI, 1 m M 2-mercaptoethanol, 1 mM benzamidine, and 20 /zg/ml each of aprotinin, leupeptin, and soybean trypsin inhibitor*. Lysozyme is added to a final concentration of 0.1 mg/ml and the cells are incubated on ice for 30 min. MgCl2 and DNase are then added to a final concentration of 5.0 m M and 0.2 mg/ml, respectively. After 5 min of incubation on ice, the lysed cells are centrifuged at 27,000g for 30 min at 4 ' and the protein concentration in the supernatant is determined by using the Bradford reagent (Bio-Rad Hercules, CA) and bovine serum albumin (BSA) as standard. Typically the pellet o f a 2 liter culture of bacteria will receive 20 ml of the lysis buffer and yield a final supernatant volume of approximately 23 ml. To confirm expression of the appropriate proteins, the proteins in the supernatant are separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) [7.5% (w/v) polyacrylamide gels] and transferred onto nitrocellulose, and immunoblots are performed with anti-Xpress antibody (Invitrogen) and the enhanced chemiluminescence (ECL) system (Pierce, Rockford, ILL the proteins contain an anti-Xpress epitope immediately alter the hexahistidyl tag. The supernatant expressing C 1 - C 2 ACV or C l a - C 2 ACV is * The buffer described is for purification of C Ia-C2 and C1-C2 forms of ACV. When supernatants of bacterial cell lysates are used only for activity measurements, the 100 mM NaC1 and I mM 2-mercaptoethanolare substituted with 1 mM EDTA. It A. Roy and A. Danchin, Mol. Gen. Genet. 188, 465 (1982). 12A. Beauve, B. Boeston. M. Crasnier, A. Danchin, and F. O'Gara, J. Bacteriol. 172, 2614 (1990).

[13]

ANALYSES OF TYPE V ADENYLYI, CYCLASE

165

either aliquoted and stored at - 8 0 or used immediately for purification of the recombinant protein. ~ The supematants from cells expressing either C 1-C2 ACV or C 1a-C2 ACV can be used to measure adenylyl cyclase activity, which is stimulated by, both G~ce as well as forskolin and inhibited by Gio~.1:~ D. Purification q [ C 1 - C 2 and C Ia-C2 Forms (?/'Type V Adenylyl Cvclase The bacterial supernatant containing C I - C 2 and C l a - C 2 ACV is incubated

with metal affinity chromatography resin (TALON; Clontech, Palo Alto, CA), which is washed with 50 mM Tris-HCl (pH 8.0), 100 mM NaCI, and 1 mM 2-mercaptoethanol. Typically, 1 ml of TALON resin is used for each 2.5 ml of the supernatant (125 mg of total protein) and the mixture is incubated for 1 hr at 4 with gentle rolling. The resin is then loaded on a 1 x 5 cm Econo column (Bio-Rad) and washed with 20 colunm volumes of buffer containing 50 mM Tris (pH 8.0), 500 mM NaCI, and I mM 2-mercaptoethanol: the flow rate is maintained at 1.2 ml/min. Thereafter, the column is eluted with 100 m M imidazole in 50 mM Tlis-HCl (pH 8.0), 100 mM NaCI, and I mM 2-mercaptoethanol. The flow rate is 1.2 ml/min and soluble cyclases elute between 2 and 6 ml from a column that has 4 ml of the resin. Note thai stora,,e,~ of the soluble adenylyl cyclases with imidazole for even short times inactivates the enzymes: this is also true for the C1 and C2 domains because the mixture of the two does not reconstitttte enzyme activity if stored in imidazole. Therefore, it is imperative to remove the imidazole rapidly. For this purpose, the imidazole eluate from the metal affinity column is immediately applied to a Mono Q 5/5 fast protein liquid chromatography (FPLC) column (Pharmacia. Piscalaway, NJ) and washed extensively' with 20-30 column volumes of buffer containing 50 mM Tris-HCI (pH 8.0), 100 mM NaCI, I mM EDTA, and 1 mM dithiothreilol (DTT), the Mono Q column is equilibrated at a flow rate of 0.5 ml/min with the same buffer. The recombinant proteins are eluted From the column with a 150-450 mM NaCI gradient over 60 rain in the buffer used for column equilibration and wash. The fractions (1 ml) are analyzed on a 10% (w/v) Sl)S-polyacrylamide gel, using Coomassie ,'stalnlnc, " ' " ~, 'rod peak fractions are pooled and concentrated. For concentration, Centricon 50 (Amicon, Danvers, iVlA) spin columns are used according to the manufacturer inslructions. At the same time, the buffer is exchanged with 50 mM Tris (pH 8.0), 1 mM DTT, 1 mM EDTA, and befl)re storage 5c~ (v/v) glycerol is added. The soluble enzymes are stored at 80 . optimum storage concentration is >1 mg of protein/ml. Starting ~ith an 8-liter culture of bacteria, this procedure would typically result in 800 It g of (?la--(?2 and C1-C2 proteins. A Coomassie-stained ,gel showing the purified

i If tile sttp¢lnalants are slored at -80 . they musl be used 'a,ifllh/30 days to ensure Ihat the soluble fornls of A(?Vare sti)l acfi\e. i~ K. Scholich, A. J. Barbier. J. B. Mullenix.and T. B. Patel, Proc. Nat[. Acad. 3'ci. U.S.A.94, 2916 (1997).

166

ADENYLYLCYCLASES

[13]

proteins is presented in Fig. lB. An identical gel is subjected to Western analysis with anti-Xpress epitope antibody (Invitrogen) and is shown in Fig. lB. E. Adenylyl Cyclase Activity Measurements

AC activity is measured on the basis of the principles of separating cAMP by column chromatography 14 and by methods previously described from our laboratory. L~,16 AC assays are performed in the presence of 5 mM Mg=C12. However, the concentration of Mg 2+ is kept at 1 mM in assays involving use of $49 c y c - membranes. II. E l u c i d a t i o n o f I n t r a m o l e c u l a r

Interactions in Adenylyl Cyclase

A. Introduction

In this section, we describe the three methods that identify the intramolecular interaction in ACV that modulates activation by G~oe. The methods described here use the full-length and the two engineered, recombinant, soluble forms of ACV described above and elsewhere, 13 as well as two other approaches, namely the two-hybrid yeast assay and competition by peptides corresponding to sequences in ACV. These findings, which are fully described elsewhere, 17 demonstrate that the C lb region of ACV interacts with a 10-amino acid region within the C2 domain of the enzyme and that this interaction modulates the stimulation of enzyme activity by G ~ . Moreover, these findings provide a plausible model for the biphasic G~o! concentration response that is observed for ACV and ACVI. 17,ts According to this model j7 a high- and low-affinity site for G~o~ may be present on ACV and ACVI and, at least in the case of ACV, the interaction between the C lb region and its C2 domain may alter the affinity of the low-affinity site. Although the crystal structure data of Tesmer et al. ~ demonstrate that a chimeric soluble AC comprising the C I a region of ACV and the C2 domain of ACI1 binds one molecule of G~c~, it should be noted that the C lb region of ACV is not included in this structure, and also that the C 1a of ACV was not studied in the context of the C2 region of ACV. Therefore, how the intramolecular interaction of C l b region of ACV with its C2 domain alters the intermolecular interactions with G~c~ is not entirely clear. Nevertheless, because of the intramolecular interaction within the ACV molecule, the C 1 - C 2 ACV more closely resembles the full-length enzyme in its regulation. t4 y. Salomon, C. Londos, and M. Rodbell, Anal. Biochem. 54, 541 (1974). 15 B. G. Nair, H. M. Rashed, and T. B. Patel, Biochem..1. 264, 563 (1989). 1~, H. Sun, J. M. Seyer, and T. B. Patel, Proc. Natl. Acad. Sci. U.S.A. 92, 2229 (1995). t7 K. Scholich, C. Winpolh. A. J. Barbier, J. B. Mullenix, and T. B. Patel, Proc. Natl. Acad. Sci. U.S.A. 94, 9602 (1997). Is A. Harry, Y. Chen, R. Magnusson. R. lyengar, and G. Weng, J. BioL Chem. 272, 19017 (1997). t9 j. j. G. Tesmer, R. K. Sunahara, A. G. Gilman, and S. R. Sprang, Science 278, 1907 (1997).

[ ] 3]

ANALYSES ()F TYPE V ADENYLYI~CYCLASE A

t6

B

Full-length ACV

C1-C2 ACV

C

Cla-C2 ACV

25-

14

30

2,

~12

25

~ ~10

1:

20

~

15

1,

,

4

2~

167

100 200 [ Os(z] (nM)

300

0

100 200 [ Gs(z] {nM)

300

0

100 200 [ Gs(~] (nM)

500

FI(;. 2. Aciivution of tile fLlil-leiiglh and soluble Iorill'~ ()l A(~V I'J\. hlcreasin 7 coneenlration~ of (i~c~ . (,,\) Stinlulalion oflile full-length I \ C \ " in Sf9 ceil nicmlbrancs by ",,aryin7 concenlr:Jlioil~, oi(i,os . (RI Slimulalion el lile (71 C2 soluble form of A ( ' V in the prcseilme of varioti~; (]~cl,~' conccnhalion~,. ((') ~lhi/tll;.ilion of the C la ( ' ] <,oltlbie form of A C V by \ar~,mg concellh'alion~; of O,u ,~. Menlblaile,'~ of SIO eolb, (2()/Lg of proleil]) or ~;LiFlerilat;.il]ts of lystltes (20 /~g of I~rotein) froii] bacioiia cxpresshlg either die C I C2 or ( ' l u C'2 I~.)rnl of ,,\CV were tlssciyecl []w AC t/ctiviiy in lhe pl'e~>ence o f v~uiotis COlleell|l'{lliOll!4 {)f (J~(,7 I {is dcscl-ibed in lext. t\clivllies LIFe i)l'escntod ;.is the i11c;.111~, :: S[~1 o[ tilt'co oxperilneill!,. IThose dala have boeli pul~lishedol~o\~heie: ~ee Rcf. 17.]

B. U.vc ql' FMl-Lenc, th aml Sohtbh" Froms q/' 7~vpe V Adenylyl Cy~'la.se When the full-length and two soluble forms (CI C2 mid C I a - C 2 ) of ACV are lested for stimulation of enzyme activity in response to different concentrations of the GTPase-deficient, constitutively active mulant (Q213L) form of G~o, (G~*), an interesiin- difference is observed. As illustrated by data in Fig. 2A, the activity of the full-length ACV expressed in S f9 cells is stimulated by G~ce~ in a concentrationdependent manner. However, the G~c~~ concentration-response curve is biphasic with an inflexion at approximately 50 nM G@~~ (Fig. 2A). A similar biphasic stimulation of G~c~~-inediated AC activity is also observed when the C 1 - C 2 form

of soluble ACV is used in the G~c~~ COl]centratioll-resl-lonse curves (Pig. 2B). Prolrl the shapes of the G~os' concerllration response curves observed with full-length clnd C 1-C2 forms of A C V it would appear ihut two s;Jttnal[oil curves for G ~ ~

;.ire superiinposed and that there ~lle probably two O~cv~-interacling sites on these enzymes. One site thai hus a higher affinity for Gsc~~:apparently begins to become saturated al ~ 5 0 - 6 0 nM concentrations of G
C2 f o r m o f A C V

is stimulated.

168

ADENYLYLCYCLASES

[13]

However, concentrations of Gpt* between 60 and 80 nM do not activate the enzyme (Fig. 2C). Further increases in the G~oe* concentration (>100 nM) elevate activity of the C Ia-C2 form of ACV (Fig. 2C). Assuming saturation of the high-affinity site for G ~ * at the inflexion observed in Fig. 2A and B (~50 nM G~ce) and the peak observed with C Ia-C2 ACV at 50 nM G~oe*,the calculated apparent 50% effective concentration (ECso) values of Gpe ~ for the high-affinity site on all three forms of ACV are similar (39 ± 14 nM for lull length: 41.7 4- 6 nM for C1-C2; 44.3 4- 5,6 nM for Cla-C2). On the other hand, the apparent ECs0 values of Gs~* for the low-affinity site on the C la C2 form of ACV is markedly higher than for the full-length or C 1-C2 ACV (cf. 128.3 4- 14.4 nM for C I a-C2 vs. 72.3 4- 5 nM for C 1-C2 and 63 4- 9 nM for full-length enzyme). These data suggest that the C l b region, which is missing in the C I a - C 2 form of ACV, is important for modulating the affinity of the second, low-affinity, site on ACV for G~*. Thus, in the absence of the C l b region, as seen with the C I a - C 2 ACV, the affinity of the second site for G~oe* is decreased and the two saturation curves are separated by a trough (Fig. 2C). The decrease in G~oe*-mediated stimulation of activity of the C 1a-C2 form at concentrations of the G protein between 50 and 80 nM (Fig. 2C) may be due to GpE* making contact at two loci on ACV that form the high-affinity site for the G protein and is elaborated on in a previous publication. 17 Importantly, however, the data in Fig. 2 also demonstrate that the C 1-C2, and not the C l a - C 2 , form of soluble adenylyl cyclase more closely resembles the full-length enzyme. Therefore, the longer C1-C2 ACV is the preferred form to substitute for the full-length enzyme in experimental systems.

C. Yeast Two-Hybrid Assay Because data in Fig. 2 indicate that the C lb region of ACV modulates the ability of Gpt* to stimulate activity, using the yeast two-hybrid assay, we have investigated whether the C lb region of ACV is involved in interactions with G~ot* or with another region of ACV that modulates the actions of Gp~*. Essentially, by the yeast two-hybrid assay, we do not observe any interactions between the C lb region of ACV and G~oe*. Therefore, only the intramolecular interactions of C I b are described below. For the purpose of yeast two-hybrid assay, the Matchmaker kit (Clontech) is used. Employing cDNA encoding the full-length ACV as template and BamHltagged (5') and SalI-tagged (3') primers corresponding to nucleotides 1861-1875 and 2178-2199, the C lb domain of ACV (amino acids 572-683) is generated by PCR. The BamHI and SalI sites in the 5' and 3' primers, respectively, facilitate the directional cloning of cDNA encoding the C lb region into the plasmids pGBT9 and pGAD424. Likewise, cDNA encoding the C2 domain (amino acids 933-1184, nucleotides 2797-3555) of ACV is amplified by PCR, using ACV cDNA as

[13]

ANALYSES OF TYPE V AI)ENYLYL CYCLASE

169

template, and cloned into the BamHI and Sail sites of the plasmids pGBT9 and pGAD424. The subdomains of C2, C2I (amino acids 995-1058; nucleotides 2985 3174) and C211 (amino acids 1091 1151; nucleotides 3271 3453), are amplified by PCR to introduce an EcoRI site at the 5' terminus and a Sa/I site at the 3' terminus. Both subdomains are cloned inlo the EcoRl and Sail sites of plasraids pGAD424 and pGBT9. All constructs are sequenced to confirm the correct sequences and reading frames. The plasmids pGAD424 and pGBT9 contain the GAI,4 activation domain and binding domains, respectively, and expression of proteins is under the control of the yeast alcohol dehydrogenase promoter. The HF7c yeast strain provided in the Matchmaker kit (Clontech) is translormed with the pGAD424 and pGBT9 constructs. Growth conditions, media, and transfl)rmation protocols are perfl)rmed according to the manufacturer instructions. T r a n s f o r m e d yeast cells are g r o w n on plates c o n t a i n i n g e i t h e r m e d i u m dew~id ol: I A e u e i n e and L q r y p t o p h a n ( L e u /Trp

) or m e d i u m in w h i c h L - h i s t i d i n e

as w e l l as I.-leucine and i A r y p t o p h a n have been o m i t t e d ( L e u /Trp / H i s

). T h e

plates are incubated at 30 for 3 days. Several o f the c o l o n i e s f l o m t r a n s f o r m a n t s are then i n d i v i d u a l l y streaked out onto new plates c o n t a i n i n g the c o r r e s p o n d i n g

medium. To verify the inferences l'rom data in Fig. 2, the expression of lacZ gene (fl-galactosidase) is monitored. After overnight growth of at least six individual colonies of the various transformants in liqnid Trp /Leu /His medium (3 ml). the density of cells is determined by measuring the OD(,00 ,,,,,. The cultures are each diluted to an OD(,ii0,~m of 0.3 and allowed to grow lk)r an additional 3 hr in con> plete liquid medium. Thereafter, fi-galactosidase activity in equal numbers of cells (determined by O1)(,o{),,,,,) is assayed by the chemiluminescence assay ~ of Jain and Magrath > as described in our previous publications.iv21 To permit comparisons, HF7c cells are translk)rmed with the various plasmid constructs at the same time and simultaneously assayed for growth and fi-galactosidase activity. Activity of fi-galactosidase is corrected lbr cell number. Control HF7c yeast cells transformed with C2 in either pGBT9 or pGAD424 and the corresponding other plasmid without any eDNA insert do not grow in His nledium and. therelbre, fi-galactosidase activity in these cells cannot be monitored (not shown). However. control cells transformed with pGAD424-CIb and pGBT9 alone show enough growth on His medium to allow fi-galactosidase activity measurements that are not above background (Fig. 3). On the other hand, cells trans[ormed with pGBTg-C2 and pGAD424-C lb show a significantly higher level of fi-galactosidase activity (Fig. 3). :i The widely used replica [liter assa~ u~dng X-Oal as substrate [o monitor # galactosidase acti',ity, is nol ,;en'dtivc enough to observe Ihese illlCracliOllS. 2(I V. K. ,lain and 1. T. Magrath. An,/. bhochem. 199, I 19 ( 1991). 21 H. Sun, Z. Chert, H. Popplelon, K. Scholich, J. Mullenix, G. J. Weipz, D. L. Fulgham, P. J. Berries. and T. B. Patel. J. t~io[. Chem. 272, 5413 (1997).

170

ADENYLYLCYCLASES

40-J

I

P < 0.006

[13]

I

-~ e mo

~

30-

p < 0.000 I - - 1

0

~

~o

~.-~

pGBT9 DGAD424

-

C2

C21

C211

Clb

Clb

Clb

Clb

FIG. 3. Interactions between the C 1b and C2 or C2I domain of ACV. Expression of fl-galactosidase activity in HF7c cells cotransformed with plasmids pGBT9 and pGAD424 either containing the indicated cDNAs corresponding to regions within ACV or devoid of any eDNA insert (denoted by "-"). After growth of the various translk)rmants on medium devoid of t,-histidine for 3 days, fl-galactosidase activity in six colonies each of the transfornmnts was measured after growing cells for 24 hr in medium devoid of L-histidine followed by' incubation for 3 hr in complete medium. Equal numbers of cells as monitored by OD6o0 nm were used to assay fl-galactosidase activity. Student unpaired t test analyses were employed to assess the significance of differences shown. [These data have been published elsewhere; see Ref. 17.]

To delineate more precisely the region(s) in the C2 domain that interact with the C lb region, using the two-hybrid assay, we have tested the ability of the C lb region to interact with two subdomains (C2I and C2II; see Fig. 1) of C2, which are highly conserved among all known isoforms of mammalian AC. Transformants carrying the plasmids pGBT9-C21 and pGAD424-CIb show robust cell growth on His medium (not shown) and significant fi-galactosidase activity over background, indicating interaction between the C2I region and the C l b domain (Fig. 3). In contrast, no significant fi-galactosidase activity above background can be monitored for the C21I region in any of the combination of plasmids (see, e.g., pGBT9-C2II and pGAD424-CIb; (Fig. 3). It is noteworthy that by both criteria, growth on His medium and fi-galactosidase activity, the interaction between the C2 or C2I region of ACV and C l b is observed only when the C l b domain is expressed as a fusion protein with the GAL4 activation domain. This observation is not unique for the proteins being tested in this system. Similar phenomena have previously been reported by us 21 and others (reviewed

[ 13]

ANALYSES OF TYPE V ADENYLYLCYCLASE

171

in Fields and Sternglanz 22) and underscore the necessity to test the interacting proteins as chimeras of both the binding and activation domain of GAL4. Nonetheless, the tindings with the two-hybrid assay suggest that the C l b domain of ACV interacts with the C2 region and, more specifically, a 64-amino acid (C21) suhdomain of C2.

I). U~e ql'PeptMe,~ Because the yeast two-hybrid data suggest that the C lb region of ACV interacts with its C2I subdomain and because the lack of the C I b region in C I a-C2 ACV altered the G,oe* concentration-response curve (Fig. 2), we postulate that disruption of the C Ib-C2I interaction would modulate stimulation of enzyme activity by G~o!. To test this postulate, peptides corresponding to the C2I region are synthesized and used in activity assays with the full-length and two soluble forms of ACV. The rationale in these experiments is that if a given peptide interferes with the C Ib-C2I interaction then the G,oF concentration response curve with either the full-length ACV or CI C2 ACV should resemble that observed with the C I a-C2 ACV in Fig. 2. Moreover, the peptide should not alter the prolile of the G,o~ concentration-response curve observed with C I a-C2 ACV. The following four peptides corresponding to the C2I region of ACV are used: peptide 1 (PI) corresponding to residues 1013 1033 (sequence NNEGVECLRVLNE1IADFDEI), peptide 2 (P2) corresponding to amino acids 1042-1058 (sequence LEKIKTIGSTYMAASGL), peptide 5 (P5) corresponding to the N terminus of P2 (sequence LEKIKTIGST), and peptide 4 (P4) corresponding to the C terminus of P2 (sequence YMAASGLNDS). None of the peptides, at concentrations up to 100 lzM, alter basal or forskolinstimulated activity of either the full-length or two soluble lorms of ACV (not shown). However, when the activity of C1-C2 ACV is stimulated O~a*, peptides 2 and 5, but not 1 and 4, decrease AC activity such that maximal inhibition is observed at 3 and 10 tzM concentrations of P2 and P5, respectively fsee Scholich e; ol. 13 for details). Thereafter, experiments have been performed in which the AC activity of the full-length and soluble lk~rms of ACV is monitored in the presence o1: maximally effective concentrations of P2 or P5 and varying concentrations of G~o,*. Controls are performed under similar condilions Ju the presence of pcptide P4 or PI, which do not alter AC activity. In the prcscnce of either P2 (not shown) or P5 (Fig. 4A and B) the profile of the G,c~' concentration response curve with either C I - C 2 ACV or the full-length ACV is altered to resemble the prolile observed with C I a - C 2 ACV in the absence o1" any peptides (see Fig. 2C). Notably, peptide P4 does not alter the G~a* concentration response prolile of C I - C 2 ACV (not shown) or the full-length -~-~S. Fields and R. Slernglanz, 7)z, ml~ Gem,1. I0, 286 (19941.

172

ADENYLYL CYCLASES

A

B C1-C2 ACV + P5

[131 C

Full-length ACV + P5

Full-length ACV + P4

~" 120

12C

120 100

~

loo

100

,~ E

80

80

80

E

60

60

60

~'$

40

40

40

'~ ~'~ 20

20

2O

R~

50

100 150

[ Gs~] (nM)

200

250

0

50

100 150 200 [ Gso~] (nM)

250

0~

50

100

150 200

[ Gsa] (nM)

250

FIG. 4. Peptide P5 converts the profile of Gsce concentration response curves of C I - C 2 and lulllength ACV to mimic the effects of G~o!* on Cla-C2 ACV. (A) Effect of peptide 5 (10 I~M) on the ability of different G~c~* concentrations to stimulate the C I-C2 soluble form of ACV in supernatants of bacterial cell lysates (20 txg of protein). (B) Stimulation of the lull-length ACV in Sf9 cell membranes (20 #g of protein) by the indicated varying concentrations of G~*, in the presence of 10 txM P5. (C) Same as (B) except 10 t~M peptide 4 was used. To facilitate comparisons between different forms of ACV, AC activities are presented as a percentage of maximal activity measured. Each value is the mean ± SEM of three determinations. [These data have been published elsewhere; see Ref. 17.]

ACV (Fig. 4C). Moreover, none of the peptides alter the G~o~* concentrationresponse curves observed with C la-C2 ACV (see Scholich et al. 13 for details). These data, along with other similar experiments performed in the presence of forskolin and varying concentrations of Gsoe* (see Scholich el al. t3) as well as the finding that the Clb region of ACV interacts with its C21 domain (Fig. 3), permit the following conclusions. First, peptides P2 and P5 alter the profile of the Gsot* concentration-response curves of the full-length and C I - C 2 ACV by disrupting an interaction of their C l b regions with the C2I domain, and this renders the two enzymes similar to their C l a - C 2 counterpart. Second, because the N terminus of P2 (peptide P5) but not the C terminus of P2 (peptide P4) alters the profile of the G~oe* concentration-response curves, it may be concluded that the C lb region of ACV interacts with a 10-amino acid region (Ll°42-Tl°51) in the C2 domain. Third, the intramolecular interactions of the C lb region with these 10 amino acids in the C2 domain modulate stimulatory actions of G,c~* on ACV. The cumulative data obtained from the three approaches, namely, utilizing different forms of the soluble ACV, the two-hybrid yeast assay, and peptides, also permit us to hypothesize about a two-site model for G~c~* interactions with ACV. Such a model would certainly explain the biphasic concentration-response curve that is observed for ACV (see Ref. 17 and Fig. 2) and ACVI) 8 Perhaps the C l b interaction with the C21 domain is unique for the two isoforms of the enzyme (ACV and ACVI), which are closely related to each other, and it is entirely possible that in AC isoforms where the G~c~concentration-response curve is not biphasic l~ only

[13]

A N A I , Y S E S OF T Y P E V A D E N Y L Y I . C Y C L A S E

173

one G,c~ interaction site is present on the molecule as demonstrated by the crystal structure of the C la domain of ACV and the C2 region of ACII. I~) III. U s e o f C h i m e r i c F o r m s o f T y p e V a n d T y p e I A d e n y l y l C y c l a s e to I d e n t i f y Gic~- a n d G f l ? / - I n t e r a c t i n g S i t e s A. Introduction

One of the unique features of ACV is that this enzyme is inhibited by Gift (reviewed in Sunahara et al. R and Iyengar2). Likewise, the type I adenylyl cyclase (ACI) is unique in that its activity is inhibited by Gfi)/subunits (reviewed in Sunahara et at. I and lyengar2). Interestingly, ACI activity is also stimulated by Ca -~+ plus cahnodulin (CAM) (reviewed in Sunahara et al. t and lyengar 2 and the region on ACI that is necessary to observe stimulation of enzyme activity is the C lb regJon.-~? 25 Moreover, it has previously been demonstrated that ACI activity can be inhibited by Gift when the enzyme is stimulated by Ca2+/CaM, but not when its activity is enhanced by G~c~.2(' Given these unique features of ACV and ACI, we describe here the method of using individual and chimeric domains of these enzymcs to identil}, Gfi?/- and Gic~l-interacting regions. Because the C I and C2 domains of AC i soforms are sufficient to observe enzymatic activity (~13,t7 and because these domains do not have to be linked to each other to reconstitute enzyme activity 772s our methods involved the use of wild-type and chin]eric C I regions of ACI and ACV mixed with their C2 domains. Figure 5A illustrates the regions of ACI and ACV that were used, the chimeric C I domains, and the nomenclature that is utilized here to refer to the different soluble AC forms that were derived by mixing the subunlts. •

t3. Generation o/'cDNA C(m.s-tructs Encoding Wihl-Type and Chinwric Adeny/yl Cvchtse Domain,',

All constructs are created in the plasmid pTrcHisB (lnvitrogen). The fulllength cDNA encoding bovine ACI was provided by A. G. Gihnan (Department Because the AC isol'ormx u~ed in lhis ~cclion are derived by mixin o AC subunit~,, in our nomenchlturc, a period is placed between the ~,ubunits. The only exception to this rule is the IC la, VC2 chimera, which is linked by the arlilicial linker described in Section I. 23 Z. L. Wu. S. T. Wong, and D. R. Storm, ,/. BioL Chem. 168, 23766 (1993). 24 T. Vorhcrr. L. Knop~l, K Hofmann, S..Mollncr. T. Pfeuf'fcr. and E. Carafoli, Biochemi.~ltT 32, 608 I I1993L 2~ L. Levm and R. R. Reed. J. Biol. ('hem. 270, ?.573 (1995). =~' C. Wittpoth, K. Scholich. Y. Yigzaw, T. Stringfield, and T. B. Patel, Prec. Natl. Acad. Sci. U.S.A. 96, 9551 ( 1999L 27 R. K. Sunahara. C. W. Dessauer. R. E. Whisnant, C. Kleuss, and A. G. Gihnan, ,I. 13iol. Che,,n. 272, 22265 (1997). 2,~ C. W. I)essauer and A. G. Gihnan, .L Bio[. Chem. 271, 16967 (1996).

174

[ 13]

ADENYLYLCYCLASES

A ACI C1 Cla

225

ACV C2

Clb

Abbreviations Used VC1 .VC2 VCla.VC2 VClalClb.VC2 ICla.VC2 IC1 .VC2 VClalClb.IC2 VC1 .IC2 IC1 .IC2

C1 E~3 Cla CEDE]

r--] ET~ ~ I J ~ ~ /

0

C2

o 0 O 0 O ' ~ '

"~

200

~

175

-~

150

:~

125

-~::,,

75

~

50

g

25

,<

t

0.5

1.0

[VC 1] constant stant

1.5 , 2.0 2.5 [VC 1]/[VC2] ratio

3.0

3.5

FIG. 5. Reconstitution of enzyme activity fi-om mixtures of soluble subdomains of ACV and AC[. (A) Schematic representation of the various forms of soluble adenylyl cyclases derived by mixing the cytosolic C 1 or C 1a or chimeric C 1 regions with C2 domains of either ACV or ACI. The various domains of ACl (solid) and ACV (open) are represented in the boxes. The amino acid residues encompassing the domains in bovine ACI are as follows: C 1,236-607; C 1a, 236~-71 ; C 1b, 472 607:C2,809 I 133. In canine ACV the domains shown comprise the tbllowing amino acid residues: C l, 322 683; C la, 322 571; C2, 933 1184. In the list of abbreviations the roman numeral preceding the C1 and C2 domains and their submgions denotes the AC isoform from which that particular region is derived. (B) Titration profile of the reconstituted enzyme activity of VCI and VC2. Increasing amounts of lysates containing one of the subdomains were mixed with I 0 l~g (total protein) of the complementary subdomain and stimulated with G~o~* ( 100 nM) and forskoliu ( 100/IM ). As shown, the optimal activity was achieved at a concentration ratio of 1 : I [~r either subdomain.

of Pharmacology, University of Texas Southwestern medical Center, Dallas, Tx). Complementary DNAs encoding the C la [ICla; amino acid (aa) 236-471], C I (ICI; aa 236-607), and C2 (IC2; aa 809-1133) regions of ACI are obtained by PCR, using ACI cDNA as template and the following primers:

ICla, primer A: 5'- ATATATGGATCCGGCTGAGCGCGCCCAG-3' primer B: 5'-ATATATAGCGCTATGAGTTTTCAGAAAACTGTTCCTCTC-3' [CI, primer C: 5'-ATATATGGATCCGGCTGAGCGCGCCCAG-3' primer D: 5'- ATATATAAGCTTCTAGTCCTGAAGCTGGTGGTACTTTCGCTCTCG-3' IC2, primer E: 5'-ATATATAGATCTGTCAAGCTGCGGCTG-T primer F: 5'-ATATATAAGCTTCTAAGCCTCCTTCCCAGAGGC-3'

[ 13]

ANALYSES OF TYPE V ADENYI,YLCYCLASE

175

The PCR products are cloned into the BamHl HindlIl sites (IC 1 and IC I a regions) and Bglll Hindlll sites (1C2 domain) of the plasmid pTrcHisB. To generate the chimeric C I domain ( V C I a I C I b ) consisting of the C la region (aa 322 571 ) from ACV and the C lb region (aa 472 607) of ACI, unique Eco47Ill [silent T ~ - C mutation at nucleotide (nt) 1674] and B~,III restriction sites are introduced at the 3' end of VC I a, using PCR methodology; the 5' primer contains a BamHI site. This PCR fragment is then cloned into BamHI-BgllI-treated plasmid pTrcHisB, which contains cDNA encoding the C1 and C2 regions of ACV joincd by an artificial linker (Scholich el al. t3, see Section 1). The resulting construct containing ACV C l a C2 with the engineered, unique Eco47111 site is digested with kk'o47Ill and the PCR-generated IC 1b eDNA (encoding aa 472-607 of ACI), engineered to contain EcoRV sites at the 5' and 3' ends, is then inserted into this new restriction site. This generates VC I alC I b. VC2. Using PCR methodology, this latter construct is then used ,:is a template to generate the individual, chimeric C 1 region in which the C I a portion is derived from ACV and the C I b domain is from ACE the 5' and 3' primers are designed to include unique BamHl and ttindlll sites, respectively: a stop codon is placed after the Hindlll site. The cDNA cncoding the chimeric C1 regions is cloned into the BamHl-HindllI sites in plasmid pTrcHisB. C. /=~7)res,~ionqf Recombinant Proleins The AC subunits ICI, IC2, VCI, VC2, and V C l a l C I b are expressed in the TP2000 strain of E. coil Expression of proteins is induced with IPTG (100 izM) and is performed at 2 3 for 21 hr as described above under Expression of Soluble Adenylyl Cyclases. Gic~l protein is expressed in E. coli JM 109(DE3) that has been colransformed with pBB 1312') to express N-myristoyltransferase and ensure synthesis of myristoylated G protein. 3° The expression of both N-myristoyltransferase and Gic~l is induced by addition of 100 izM IPTG and incubation at 30 for 16 hr. The bacterial pellets are harvested as described above and GioeI protein is purilied by, the method of Mumby and Linder) I Bovine brain fly subunits of heterotrimeric G proteins are purified to homogeneity as described by Mulnby el a/. 32 with the modilications of Neer el a/. 3~ Reconstitution (~/"Enzyme Activity.fiom TyFe V and Type I Aden.vlyl Qvc/a.ve Sld~domains. One of the unique features of using unlinkcd AC domains to reconstitute enzyme activity is that the two regions (CI and C2) must be titrated against 2'~R. J. Duronio. 1). A. Rudnick. S. P. Adan>. D. A. 'l~*wlef,and J. I. Gordon. J. Biol. Chem. 266, 10498 (1991). 3o M. E. Lindcr, 1. H. Pang, R. ,J. DLironio. J. I. (]ordon. P. C. Sternwcis, and A. O. Gilnmn. ,I. Biol.

C/win. 266, 4654 ( 1991 I. ~1S. M. Mumbyand M. E. l,inder, MeHiod.sEnomol. 237, 254 (1994). ~: S. Mumby,l.-lt. Pang, A. G. Gihnan,and [: C. Stcrnweis.J. I¢i~1.('hem. 263, 2020 (I 9,~8). ~ E. J. Neer, J. M. l~ok.and L. G. Wolf.,/. Biol. Chem. 259, 14222 (1984).

176

ADENYLYL CYCLASES

[ 131

each other. In the example shown in Fig. 5B, supernatant of lysates from bacteria expressing one of the ACV domains is kept constant ( 10/xg of total protein) and the amount of supernatant from bacterial lysates containing the complementary domain of ACV is varied. The amount of C I or C2 proteins in the lysates is estimated by densitometry of Western blots perlbrmed with anti-Xpress (Invitrogen) antibody, using anti-Xpress epitope-tagged pure G~o~ as standard. Moreover, to facilitate the optimization of conditions, enzyme activity is measured in the presence of both G ~ (100 nM) and forskolin (100/~M). As shown in Fig. 5B, the ratio of the two domains of ACV increases the activity of AC until a peak is observed when similar amounts of each subunit are present (ratio of 1.0 in Fig. 5B). Thereafter, increasing concentrations of C2 decrease enzyme activity. Therefore, it would appear that for optimal activity the C 1 and C2 regions must be in a I : 1 ratio and at higher concentrations of either domain, homodimeric forms interfere with the formation of a heterodimeric catalytically active enzyme. After optimizing the amounts of the various subunits II that need to be mixed to reconstitute activity in the presence of both G~ot and forskolin, the activity # of various combinations of subunits is determined in the absence and presence of G ~ or forskolin ( 100 # M ) . With the exception of IC I.IC2, all other forms of soluble AC as well as full-length ACI expressed in Sf9 cells are stimulated by G~c~ (Fig. 6A). On the other hand, forskolin is able to stimulate the activity only of mixtures containing either C 1 or C I a and C2 domains of ACV (Fig. 6A). Therefore, forskolin is not used to stimulate enzyme activity in our experiments to elucidate the actions of Gi~ and Gfl?/subunits. Because the C l b region of ACI binds Ca2+/CaM 23-25 it would follow that C1 domains of AC containing the C l b region from ACV should be stimulated by Ca:+/CaM. Except for I C I . I C 2 all other forms of soluble ACs and the fulllength ACI that contains the C l b region of AC! are stimulated by Ca2+/CaM (Fig. 6B). Note that the substitution of the C lb region of ACV with the C lb region of ACI permitts stimulation of enzyme activity by Ca2+/CaM. The inability of either G ~ or Ca2+/CaM to stimulate the ICI.IC2 form of soluble AC cannot be attributed to the lack of interaction between the two domains of the enzyme because the basal activity of this mixture is measurable and higher than that of the other combinations of C I and C2 domains. More likely, the complex formed by ICI and IC2 does not allow access to the G~o!- and Ca2+/CaM binding sites. It should be noted that the optimal concentration of G~c~ required to stimulate enzyme activity in Fig. 6A is different and determined empirically II For optimization, the amount of AC subdomain proteins in the snpematant of bacterial lysates is measured by densitometry of Western blots performed with the anti-Xpress (lnvitrogen) epitope antibody. For standards, the anti-Xpress epitope-tagged pure Gsol is used. # In this section, for expression of specific activity of enzyme, the amount of protein corresponding to the two AC domains is measured by Western analyses using the anti-Xpress epitope antibody (lnvitrogen) as described in footnote II

A 350 300 "~ ,_

• Basal Gsc~ [] Forskolin

250

ii

_~ ~ ~oo cz_

>,~

150 lOO

/ 1 / f/ / I / / / / /

m

0 >

0

>

>

¢

50H

O <

O

O

>

-

O >

O -

>

B 300

>.

250

< o

200 !

O >

>

• Basal ~I Ca2+IOaM T

go

_o~

o~

150

~

~00

<

50'

<

>

--

>

--

--

>

-I

I:t¢~. 6. S/imuhttion o1 adun_~lyl c~clusc acti\it~ rcuon~,litutcd b~ mixin~ the cyios¢~lic C I and C2 rc~ion<~ of ACI zmd/or ACV. (.A) A d e n y l y l c,,.clasc a c t h i t y \~.as m e a s u r e d u n d e r basal c o n d i t i o n s or hi i11¢ p r e s e n c e e l ( ] ~ f or fl~rskolhL Eilllei membralleS of SfO cells cxpren, usccl. Mc'Llns ± SI~]M are h]lo\t ii (/i

] ex[*~criillClllS). (B) Melnl~railc',, o [ S l g c'<'ll~ OXplCssin7

ihe I'ull-len.~lh .,M]~I and ~aiiou', I n i x l u i e s o f C I or c M m e r i c ('1 and ( ' 2 donlains o1 & ( ' l or A ( ' \ " wore a->>>Li)od I])r adenylbl c)chp, e a c i h i l ) , in lilo plchenc¢ and absence o f ('a~'| (500 n M ). ('a 2

was added

Ii) lhcsc ti<4hLl~
2f~.l

S E M (H

]~ ¢xporhl/cnls) arc ,,llo~ n [ these clala h a \ e been i~ubli~hed ol-,c\~]~ore:

178

ADENYLYLCYCLASES

[ 13]

as described for the soluble forms of ACV in Fig. 2. Thus, VC 1.IC2 requires 120 nM G~o~; all other combinations require a 80 nM concentration of the G protein. Using G~ot-stimulated activity of the various mixtures of AC domains and the full-length ACI, the effects of purified bovine brain G13y subunits are monitored. As demonstrated previously by others/~'34'35 /47, subunits, in a concentrationdependent manner, inhibit the activity of full-length ACI (Fig. 7A). Likewise, consistent with previous findings that ACV is not inhibited by 13y subunits (reviewed in Sunahara et aL I and Iyengar2), the soluble form of ACV (VCI.VC2) is not inhibited by fly subunits (Fig. 7A). To determine which portions of ACt is (are) necessary to observe the inhibition of activity mediated by t37' subunits, experiments are performed with the C 1 or C 1a and C2 regions of ACI mixed with the complementary domains of ACV. When AC activity is reconstituted with the CI domain or its N-terminal C l a region from ACI and the C2 region of ACV (IC1.VC2 and ICla.VC2, respectively), the G~o!*-stimulated activity is inhibited by G13y subunits in a manner similar to that observed with the full-length ACI (Fig. 7A). On the other hand, when the C2 region of ACI is reconstituted with the C 1 domain o f A C V (VC 1.IC2), G~*-stimulated activity is not altered by G137' subunits (Fig. 7A). These data demonstrate that the C 1a region of ACI is sufficient to observe inhibition of enzyme activity by G13y subunits of G proteins. To determine whether the C lb region of ACI also contributes to inhibition of activity by G137' subunits of G proteins, the chimeric C I region comprising C la of ACV and C lb of ACI (VC I alC 1b) is reconstituted with the C2 region of ACV (VC 1alC I b.VC2). The G~c~*-stimulated AC activity of this enzyme is not altered by G137" subunits, indicating that the C l b region alone is not sufficient to observe 13?'-mediated inhibition. Interestingly, however, when the C2 region of type I AC is used to reconstitute AC activity with VC1 alC l b, G/J?' subunits inhibit the G~ce*-stimulated enzyme as effectively as that observed with the l~fll-length ACI and ICIa.VC2 (Fig. 7A). These findings, coupled with the observations that neither the C I b nor C2 region of ACI by itself is sufficient to observe G13?,,-mediated inhibition of enzyme activity, suggest that the C l b and C2 domains of ACI interact with each other to form a Gfl),-interacting site. This contention is also supported by the data in Fig. 7B. Hence, when the activity of V C I a l C l b . I C 2 is elevated by Ca 2 ~/CaM, 137' subunits of heterotrimeric G proteins inhibit activity. However, Ca2+/CaM-stimulated activity of the V C I a I C l b . V C 2 enzyme is not altered by G137' subunits (Fig. 7B). As expected, Ca2+/CaM stimulates the activity of IC1.VC2 AC and this activity is also inhibited by G137' subunils in a concentration-dependent manner (Fig. 7B). Taken together, the data in Fig. 7 demonstrate that the C l a region of ACI is sufficient to observe G13?'-mediated inhibition of enzyme activity and that the C I b and C2 domains of ACI cooperate to form a G13y-interacting site that also permits 34 R. Taussig, L. M. Quarmby,and A. G. Gilman, .1. Biol. Chem. 268, 9 (1993). 35 W.-J. Tang and A. G. Gilman, J. Biol. Chem. 266, 8595 ( 1991).

[ 13]

ANALYSES OF TYPE V ADENYLYI. CYCLASE B 140

140

120

120

179

10(J ~. ea m~

80

2o ~T-,

60

ca

.<

~

C

2

+

IC1" VC2

• ~--

VClalClb- VC2 VCl alC1b" IC2 IC!a" VC2 VC1"iC2 Sf9 ACI

•

40

at

20 0

50

100 150 200 [{~2-subunit] (nM)

250

300

-7-'v°gTJ0.

+ ~cJ 60 ~_~ o>,

~g 40

o.< ~-

2o

,

,~,1 -F-V--- ~ .... 50

100 150 200 [l}y-subunit] {ha)

250

300

FI(;. 7. I n h i b i l i o n o f acti',ity o f d i f l c r c n l A C [orms by' /~y subunits o f h c t e r o t r i n l c r i c G proteins, iA} AC activity in either illelllbrLlllCS of Sf t) coils e x p r e s s i n g the I'ull-lenglll ACI or in mixtures ~.)l" s u b d o n m i n s lroln ACI and ACV w e r e slimulated with O,a'*. T h e effect o1: various conccntralions o I /~F subunits to m o d u l a t e / \ C acfi~ity, was monitored. Data arc presented as percent inhibition oiG~c~*stimulated activib and represent m e a n s ± SEN] 01 -- 3 experiments). (B) S a m e as ( a ) . except that the A C a c l i v i l y ,aas stimulated by addilion o f C a M (500 n M ). Ca e+ was presenl at ;.t fimil conccnlration of 30 IM'I. Percenl inhibition of ('ti e !/(':.lM-stimulated acli'~ity is s h o w n as nleans k SEM ( # z - - 3 cxpcrimems). IThese data haxe been published e l s e w h e r e : see Ref. 26. I

these latter G protein subunits to inhibit enzyme activity. The requirement for both C 1b and C2 regions of ACI to observe Gfiy effects is reminiscent of our previous lindings that the C lb region of ACV interacts with a 10-amino acid region on its (72 domain and that this intramolecular interaction modulates the ability olG,c~ to s t i m u l a t e enzyme activity.IV Using the subdomains and chimeric domains described in Fig. 7, w e have also been able to show that the Gioe interaction sites on ACV and ACI are located in their C la and C I regions, respectively. Furthermore, these regions from the two AC isofornls are suflicient to observe inhibition of enzyme activity. These data are described in detail in Wittpoth ef al. 2(' and. Ibr the ACV, are similar to those of Dessaucr e t a[. 3(' Moreover, we have also observed thal the presence of the C lb region of either ACI or ACV in conjunction with the C I a region of ACV enhances the sensitivity of AC to inhibition by Oioe (see Wittpoth e t a/. > Ik)rdetails). Overall. therefore, the examples of data described above show the utility of the experimental methodology invoh, ing AC domains and their chimeras. IV. G T P a s e - A c t i v a t i n g P r o t e i n A c t i v i t y o f T y p e V A d e n y l y l C y c l a s e A. Inlrodzmlion A s reviewed in several articles in v{)lume 344 of this series, the GTPase activity of several c~ subunits of heterotrimeric G proteins is regulated by regulators of G protein signaling (RGS proteins). To dale, however, ilo RGS protein that

180

ADENYLYLCYCLASES

[ 13]

modulates Gsc~ function has been identified. Therefore, we propose the hypothesis that AC, the effector of G ~ , may serve as a GAP for the G protein. Essentially our data (see Scholich eta/. 37) have demonstrated that the C 1 - C 2 ACV, its C2 domain, and the C2I and C2II subdomains act as GAPs for G~ce. In this section, we describe the methods that we have used to identify the GAP function of ACV and its subdomains.

B. Expression and Purification of C1, C2, C2L and C2H Domains of Type V Adenylyl Cyclase The cDNAs encoding the C 1 and C2 domains of ACV are generated by PCR, using restriction enzyme site-tagged primers, and cloned into plasmid pTrcHisB. These proteins are expressed and purified exactly as described above for C 1-C2 and C I a - C 2 ACV (Section I). During the concentration step when buffer is exchanged, Centricon 10 (Amicon) is used for the C2 protein and Centricon 30 (Amicon) is used for the CI protein. C2I and C2II regions of ACV are also expressed as His6-tagged proteins. The cDNAs encoding the C2I and C2II regions (nt 2983-3174 and 3271-3453, respectively, of canine ACV) are generated by PCR, using the appropriate restriction enzyme site-tagged oligonucleotide primers, and ligated into the plasmid pTrcHisB. The C21 and C2II proteins are expressed as described for soluble AC forms in Section I, except that the C2I protein is expressed at 14Z The C2I protein is purified exactly as described above for the larger domains of ACV. For the purification of the C2II subdomain the TALON column is washed with 20 column volumes of 100 mM imidazole, pH 8.0, and then eluted with 5 ml of 250 m M imidazole, pH 8.0. Significant loss of C2I and C21I proteins occurs during the concentration of samples with the Centricon spin columns. To minimize this loss, a 1-mg/ml solution of BSA is placed in the Centricon 10 (Amicon) columns for 5-10 rain at room temperature. Unbound BSA is removed by washing the columns five times with 50 m M Tris (pH 8.0), 1 mM DTT, and 1 mM EDTA before use. By this method, the C21I subdomain is purified without any contamination.

C. GTPase-Activating Protein Assay,~)r G~e~ Monomeric Gsc~ (37 nM) or Gic~l (54 nM) is incubated for 10 rain at room temperature with [y-32p]GTP (I00 nM) in medium containing 25 mM H E P E S NaOH (pH 8.0), 0.1 mM EDTA, and 1 mM DTT. The final incubation volume is 1.5 ml. The temperature is then lowered in a refrigerated water bath to 14 ° and the incubation proceeds for another 5 rain. At this point (time 0), GTP (0.1 mM ~6C.W. Dessauer, J. J. G. Tesmer, S. R. Sprang, and A. G. Gilman, J. Biol. Chem.273, 25831 (1998). 37 K. Scholich, J. B. Mullenix, C. Wittpoth, H. M. Poppleton, S. C. Pierre, M. A. Lindorfer, J. C. Garrison, and T. B. Pate1, Science 283, 1328 (1999).

[ 13]

ANALYSES OF TYPE V ADENYLYLCYCLASE

18 1

tinal concentration)in the presence or absence of the indicated protein (1,1 /~M) is added. These concentrations of C 1 - C 2 ACV and its subdomains are similar to the concentration of RGS4 required to stimulate GTPase activity of GiOt1.3s Aliquots (100 i~1) are withdrawn at various times and reactions are terminated by mixing with 750 #1 of an ice-cold 5% suspension of activated charcoal in 50 mM NaHePO4, pH 7.4. The charcoal is pelleted in a tabletop centrifuge at 14,000g for 10 rain at 4 . Aliquols (650 #1) of the supernatant are pipetted in fresh Eppendorf tubes and the spin is repeated at least once to remove residual charcoal. Alter the final spin, 625/~1 of the supernatant is removed and the free 3_,p is determined in a scintillation counter. Because this assay essentially measures the GTP hydrolysis from a single cycle of GTPase activity, below we refer to this assay as single cycle GTPase activity. Using the aforementioned assay, the effect of C 1 - C 2 ACV and the proteins corresponding to its C I and C2 domains on the single-cycle GTPase activity of G~c~ is monitored. As demonstrated in Fig. 8A, the addition of C 1-C2 ACV to G~c~ lhat has been loaded with [V-~ePIGTP increases the rate of single-cycle GTP hydrolysis such that the tt/2 for complete hydrolysis of [V-32P] GTP bound to G~o~is decreased fl'om 55.1 ± 3.08 to 1 I. 1 :t: 2.52 sec. Similarly, protein corresponding to the C2 domain of ACV also increases the GTPase activity of G~o~and decreases the tile of single-cycle GTP hydrolysis by the G protein (Fig. 8A). In contrast, lhe C I domain of ACV, at 50-fold molar excess over G~o~, does not alter the rate of single-cycle GTP hydrolysis by G~c~ (Fig. 8A). These data demonstrate that the C 1 - C 2 ACV and its C2 domain act as GAPs for G~o~. To identify lhe minimal region(s) on the C2 domain lhat can interact with G~o! and retain the G~o, GAP activity, the following approach is taken.

D. Yeast Two-Hybrid Assay to Determim: Minimal Regions qf C2 Domain ~/" T37~e V Adeizylyl Qvclase That ]nlero~'t with G,ot The sn'uctura[ data from cocrys|allization of G~c~ with the C la domain of ACV and the C2 region of AC|I demonstrate that the G protein interacts with amino acids in AC that reside in two highly conserved regions within the C2 domain of all iso|orms of the enzymel9: in ACV. these highly conserved regions in the C2 domain comprise aa 995-1058 and 1091-1151. These two regions are referred to here as C21 and C2IL respeclively. To determine whether these regions, when taken out of the context of the ACV molecule or the complete C2 region, can interact with G~of, the yeast two-hybrid assay is employed. However. before using cDNAs encoding small subdomains of the C2 region of ACV. we have investigated ,Maelher the proteins corresponding to larger domains in ACV interact with G~c~. Essentially, the same system (Matchmaker: Clontech) as described above in q'~D. M. Bcrnlall. T. M. \¥'ilkie, and A. G. Gilman, Cell 86, 445 (1996).

182

ADENYLYL CYCLASES

B

A 120•

6O

o

lOOI 04

[ 131

50

80-

40

oo I

O

4o~

30

~ 20

2o ~

"N

30

60

90

120

150

180

10 0

Gs~ +C1C2

+01

+C2

+C21

+C211

Time (s)

FIG. 8. CI-C2 ACV and its C2 domain act as GAPs for G~c~.(A) G~o, (37 nM) was incubated flw I 0 min at room temperature with [?/-32p]GTP ( 100 nM) and then at 14 as described in text. At time 0, 0.1 mM GTP and the indicated protein I I. 1 tzM) were added. Aliquots were withdrawn at the indicated times and fi'ee PO4 was measured as described in text. The open circles represent G~c~alone, closed squares indicate G~oein the presence of C I-C2 ACV, closed circles represent G~c¢ in the presence of C2, and open squares indicate G~c~in the presence of C1 ACV. An'ows depict the half-time (q/2) for complete hydrolysis of GTP bound to G~c~ under control conditions and in the presence of C1-C2 ACV or its C2 domain. Values are presented as a percentage of total 32pi released. Each experiment was repeated at least three times. Total Pi released under the different conditions was as follows: G~ee alone, 5(1 fmol: G~c~+ CI-C2 ACV, 48 fmol; G~ce + C2, 48 fmol; G~c( + CI, 42 fmol. (B) tl/2 of complete hydrolysis of GTP bound to G~o, in the absence and presence of the indicated protein leach at 1.1 ItM). Experimental conditions were the same as in (A). Each value is the mean + SEM of at least four determinations. *p < 0.001 as compared with Gso, alone (Student unpaired t test). [These data have been published elsewhere: see Ref. 37. I S e c t i o n II is utilized. Our data d e m o n s t r a t e that as m o n i t o r e d by both H1S3 and lacZ r e p o r t e r genes, the C2 d o m a i n o f A C V interacts better with Gsc~* as c o m p a r e d with w i l d - t y p e G~o! (not s h o w n ) . Notably, in the s a m e t w o - h y b r i d s y s t e m , using the e p i d e r m a l g r o w t h factor r e c e p t o r and either w i l d - t y p e or constitutively active G ~ , w e have p r e v i o u s l y s h o w n that the w i l d - t y p e G ~ interacts better with the rec e p t o r as c o m p a r e d with G~o~*.21 Thus, in the t w o - h y b r i d assay, the interactions o f Gsc~ and Gs~* are c o n s i s t e n t with the notion that the active form o f G ~ interacts better with r e c e p t o r and that Gsot* interacts better with the e f f e c t o r (AC). M o s t importantly, the f i n d i n g s d e s c r i b e d a b o v e d e m o n s t r a t e that the yeast t w o - h y b r i d m e t h o d is useful to m o n i t o r interactions b e t w e e n C2 d o m a i n o f A C V and Gs~*.

E. Interactions o f C2 Subdomains with G,~ F u r t h e r e x p e r i m e n t s are p e r f o r m e d to define the s u b d o m a i n ( s ) within this region that are involved in the interactions with Gs~*. Therefl)re, by e m p l o y i n g

[ l 3]

ANALYSES OF "I'YPI~2V ADENYLYL CYCLASE

183

the plasmids pGBT9 and pGAD424 containing the C21 (aa 995-1058) and C211 {aa 1091 1151 ) subdomains of ACV (Fig. 1) and G~c~~ or G~c~, experiments simihn to those described above are performed. The constructs used for C2] and C21] are described above under Section I]. The G~ce conslructs have been described previously,el The C21 and C2ll subdomains of ACV are highly homologous in all forms of mammalian AC cloned and characterized. Although amino acids in this region of ACII have been shown to interact with G~a, 19 ill yeast cells transformed with the C21I region and G~oe~ or O~a no interaction is evident (not shown}. However, in identical experiments with HF7c cells transforlned with the C2I subdomain of ACV and G~c~orG~a ~, growth of cells is observed in Trp /Leu /His medium; controls transfl}rmed with C2I in pGBT9 and pGAD424 without cDNA insert or ','ice versa do not grow on the latter n]edium, although transformation has occurred as demonstrated by growth of all cells in the Trp /Leu medium (not shown). To m{mit{}r expression of the IllS.? gene more quantilatively, six colonies cach of HF7c cells transforlned with the vari{}us constructs in pGBT9 and pGAD424 are grown in liquid Trp /Leu-/His medium. After 24 hr of growth in this medium, the OD(,{}0,,,,~ is determined tor each culture (Fig. 9A). Similarly, fi-galactosidase activity is measured in an equal i]umber {}f cells (Fig. 9B). As shown in Fig. 9, by bolh criteria, that is~ HIS3 and LACZ expression, the C21 region interacts better with G,c~: than with G~oe. As described in Section 111, t{} facilitate the comparisons, all translormations and analyses are pcrlornled simullaneousIy {)l] three separate occasions. Moreover, by both criteria, growth on Trp /Leu /His medium and /~-galactosidase activity, the interaction between the C21 region of ACV and G~ce is best observed when the former protein is expressed as a chimera with the G A L 4 binding domain and G~c~ as chimera will] the GAL4-activating d{}main (Fig. 9). As mentioned previously, this is not unique to our studies, and a similar phenomen{}n has previously been observed with other proteins {reviewed in Fields and Slcrnglan×-- ). The data obtained with the two-hybrid assay demonstrate thai C2 as well as thc 04-amhlo acid-long region of C2 (C2I) can interact with active O~c~and provide the lationale for expression of the smaller protein fragments for use in the G A P assay.

b: Gl'l'ase AcliYati<~ Proteill ActiY#r o/C21 a . d C21I S.l~domain.~ r?/ 7~rpc V Aden.Yl3"l Cvc/ase To determine whether the C21 subdomain thal interacts with G~o~~ in the twohybrid assay also exhibits G A P activity against the G protein, the C21 region purified as described above is used. Although the C2[I region of ACV does not interact with (-}~{z~ in the two-hybrid assay, becatlse alllillO acids ]11 this region interact with G~a,. ~9 the purHied C2I[ region is also employed. Using the C21 and C211 proteins, the G~o~ G A P assay described above is performed. In these experiments, the half-lime for complelc GTP hyrolysis from the single-cycle GTPase

184

ADENYLYL CYCLASES A

[ 13]

B

1.2l 1.0

p<0000

"~gE g~,== gg8 ~ga

~0

o

p < 0.108

pGBT9 pGAD424

C2 I Gsa*

C2 I Gsa

Gs(~* Gsa C2 I C2 I

pGBT9 pGAD424

C2 I C2 I Gsa* Gs(x

Gs(x* Gsa C2 I C2 I

FIG. 9. Interactions between the O,c* or Gsoe* and C21 domain of ACV. (A) After growth of the various transformants in medium devoidof L-Hisfor 24 hr the cell growth was determined by measuring the OD at 600 nm for six cohmieseach of the transformanls. (B) fi-Galactosidaseactivityin six colonies each of the transformants was measured aftergrowingcells for 24 hr in mediumdevoidol'L-Hisfollowed by incubation lot 3 hr in complete medium. Equal numbers of cells as monitored by OD¢,II0nm werc used to assay fi-galactosidase activity. All transformations and assays were performed simultaneously. Student unpaired t test analyses was employed to assess significanceof differences shown. assay is measured in several experiments. As shown in Fig. 8B, both the C21 and C2II regions of ACV, like the larger C2 domain, exert G~ot G A P activity. Thus, the small domains in ACV, which include the amino acids that interact with G~oe, are by themselves sufficient to act as G~ot GAPs. Notably, although the C21I subdomain of ACV does not interact with G~oe* in the two-hybrid assay, it is capable of interacting with G~ce in the GA P assay. This is a classic example of why the negative data froin the two-hybrid system should be cautiously interpreted. Overall, the approaches described above have permitted us to identify the G~c~ G A P activity of ACV and its subdomains. V. G u a n i n e N u c l e o t i d e E x c h a n g e of Type V Adenylyl Cyclase

Factor-Enhancing

Activity

Because the C2 region and its subdomains in ACV can serve as Gsa GAPs and thereby expedite the termination of signal, we have examined whether signal onset can also be modulated by ACV and its subdomains. For adenylyl cyclase activation via G~oe, the G protein in its heterotrimeric form must be activated by a receptor (see Birnbaumer and Birnbaumer > for review). We and others 1(~4°-41 have previously ~ h Birnbaumer and M. Birnbaumer, J. l?ecepmr 5"iqu ~ I Tra ~d 'zion l?es. 15, 213 (I 995). 4o T. Okamoto, Y. Murayanla, Y. Ha+~ashi. M. ]nagaki, E. Ogata, and I. Nishimoto, Cell 67, 723 ( 1991 ). 4] T. Okamoto and 1. Nishimoto, .L BioL C/win. 262, 8342 (1992).

[ 13]

ANALYSES OF TYPE V ADENYLYL CYCI,ASE

185

shown that peptides corresponding to short regions in G~-coupled receptors can aclivate the heterotrimeric G protein in vitro. Therefore, these peptides can be regarded as constitutively active receptor. ~ a l ll] our ;,tssays we have employed the peptide /7111-2 corresponding to amino acids 259-273 in the fie-adrenergic receptor, a° The purified G~oe is mixed with purified bovine brain Gill subunils in a I : 5 molar ratio and incubated on ice for 60 rain. Note that to achieve proper reconstitulion of recolnbinant G~0~ with bovine brain G f i / s u b u n i t s , GI)P (1 I,M) should be inchlded. The purified soluble form of ACV, C 1 - C 2 ACV, is used as the active enzyme. Receptor-n~ediated activation of heterolrimeric G proteins results in an augmentation of G T P - G D P exchange on the c, subunit. > This increase in G T P - G D P exchange can be conveniently measured by monitoring thc steady state GTPase activity of G pl'oteins. I~a~.24~ Therelore, in our studies we have monitored the steady state GTPase activity of O~. To measure receptor mimetic peptide-mediated increase in steady state GTPase activity, the following experimental conditions are used. The reconstituted heterotrimeric G~ (27.7 nM) is incubated with or without the peptide and/or proteins of interest at 25 in medium containing I00 nM [/-:~-~P]GTP, 25 mM HEPES (pH 8.0), tl)0 txM EDTA, 120 IzM MgCI2, and I mM I)TT. Aliquots (50 p l) are withdrawn and hydrolysis o f [ / - 3 e P ] G T P (100 nM). which is linear for more than 30 m/n, is monitored in duplicate at the 20-rain lime point as described above for lhe (Zcv G A P assay. As dcmonstraled previously I~4° the peptidc filll-2, by enhancing the rate of GTP GDP exchange, increases activation of tile heterotrin]cric G~ in a concentration-dependent maimer( Fig. IOA). In the presence of either C 1-C2 ACV ( 1. t /~M ) or the C2 domain e l ACV (I.I tzM: see Scholich el a/.~:), lhe concenmition response curve for fillI-2-n]cdiaied activation o1" G~ is shifted to the loll such that in the presence ol'C I - C 2 A C V or its ('2 domain. 100-fold lower concentrations of lhe peptide ;,lle reqtiired io activate (]~ to the s;,lllle extent :,is that obseived without AC or its C2 domain (Fig. 10A). Notably. in cimlio] experiincnts perforined in lhc abscl]CC of pepiidc tJlll-2, neither C I C2 ACV llOl"the C2 dolllain of the ellZyllle aher,, slcadv stale GTPase actixitv o1 the heteroh'imeric G, (see, e.g., Fig. 10A). Moreover. in the presence of peptidc/7111-2 :.tl i.l concentration (I 00 nM) Ihal by i tsell does not increase steady stale GTPasc aclivily of heterolriineric (is, the effect of (71 C2 A C V on /glll-2-mediated aclivatioll of (]~ is collccntlatiOll dependent iscc Scholich el a/. ~: Ior details), hitereslhlgly, like the C2 don]ain, Ihc C21 and (7211 regions retail] tile capacity Io enhance the /~lll-2-mediated increase in the slcadv slate GTPaso acti\iiy of (7,.~7 However, the protein corl-cspoilding to the (7i region of A C V clots not ;,illcl lhe ability of/-Jill-2 to increase slcady stale (7FPase a c l i v i i \ of (7,,. ~7 I). M. Bcrman, T. M. Wilkic. and A. (]. Gi]nmn, ('c/l 86, 445 119961. ~ [ t ti~w,hiiima. K. M. Fcrgu',on.M. I). 5;migcl.~md ..\. G. (iihmm. J. Biol. (lw.,n. 262, 7~7 (I 9,";7}.

186

ADENYLYL CYCLASES

A

B

08

5-

-> r~ a. 0; -~

4 ¢~

[ 13]

g~

Gs+ C1-C2ACV

3

06

co~ gE 04.

E

,

2

G

0l

S .01

~ .1

.......

1 10 100 1000 I3111-2(#M)

02.

0011 0

20 40 Time(minutes)

60

FiG. I(1. CI-C2 ACV enhances the fi2-adrenergic receptor peptide (fllll-2)-mediated actiwLtion of heterotrimeric Gs. (A) C I-C2 ACV augments the ability of different concentrations of filll-2 to increase steady state GTPase activity of G~. Steady state GTPase activity of reconstituted G~ (27.7 nM) was monitored in the presence of different concentrations of peptide fill[-2 with (lilled squares) or without C I-C2 ACV ( 1.4 I*M ) as described in text. The rates of Pi release per microgram of G~cein the heterotrimer are presented and represent means ± SEM (n -- 31. (B) C I C2 ACV increases the rate and extent of [:~sS]GTPyS binding to reconstituted G> 135S]GTPvS binding to G~ ( I(1 nM ) alone (open square), in the presence of 200 nM C1-C2 ACV (open circle), in the presence of 0. I #M fil[l-2 peptide (closed square), or in the presence of both CI C2-ACV and filll 2 (closed circle) was monitored in buffer containing 100 nM GTPvS, 25 mM Hepes (pH 8.0t, 10[) I*M EDTA, I mM DTT. and 150 I~M MgCI> Data are presented as moles of [~sSIGTPvS bound per molc of Gsce in the heterotrimer and are representative of two experiments. (These data have been published elsewhere; see Ref. 37.] A s a s e c o n d m e t h o d to e v a l u a t e the ability o f C 1 - C 2 A C V to a u g m e n t g u a n i n e n u c l e o t i d e e x c h a n g e activity of fllII-2, the b i n d i n g o f G T P v S to G~ is m o n i t o r e d . T h e s e e x p e r i m e n t s are p e r f o r m e d u n d e r c o n d i t i o n s that do not favor rapid a n d m a x imal G T P v S b i n d i n g to the G protein. Essentially, h e t e r o t r i m e r i c G~ is i n c u b a t e d in b u f f e r c o n t a i n i n g 100 n M [-~sS]GTPvS, 25 m M H E P E S (pH 8.0), 1 0 0 / z M E D T A , 1 m M DTT, a n d 150 /~M MgC12. For the t i m e c o u r s e o f [ 3 5 S I G T P v S b i n d i n g , a l i q u o t s ( 1 0 0 / ~ 1 ) are w i t h d r a w n at the t i m e s i n d i c a t e d and b i n d i n g is t e r m i n a t e d by m i x i n g with 2.0 ml o f ice-cold stop buffer c o n t a i n i n g 20 m M T r i s - H C l (pH 7.4), 10 m M NaCI, and 25 m M MgC12. T h i s m i x t u r e is then filtered t h r o u g h B A 8 2 nitroc e l l u l o s e disks ( 2 5 - m m d i a m e t e r ; S c h l e i c h e r & Schuell Keene, N H ) u n d e r v a c u u m . T h e n i t r o c e l l u l o s e disks are w a s h e d twice with 2.0 ml of stop buffer w h i l e still in the v a c u u m m a n i f o l d . T h e B A 8 2 filters are d i s s o l v e d in 2 lnl o f e t h y l e n e glycol m o n o m e t h y l e t h e r and t h e n c o u n t e d i-\)r [ ~ s S ] G T P v S b o u n d to the G~oe. N o n s p e cific b i n d i n g is d e t e r m i n e d in parallel i n c u b a t i o n s that are identical e x c e p t that e x c e s s u n l a b l e d G T P v S ( 100 l , M ) is present.** U n d e r these c o n d i t i o n s , w h i c h are

~* The background or nonspeciIic binding ol- [3 ~ SIGTPvS may be high if fresh DTT is nnt used in tile assay.

[1 3]

ANALYSES OF TYPE V ADENYI,YI,CYCLASE

187

ideal to monitor receptor-mimetic peptide-induced activation of G T P v S binding to the heterotrimeric G~, receptor-mimetic peptides alter both the rate and extent of G T P v S binding to the G protein. 1~'374° Thus, in the presence of C 1-C2 ACV, the rate and extent of threshold concentrations ( 100 nM; Fig. 10A) of filll-2-mediated G T P y S binding to G~ is augmented (Fig. 10B). C I - C 2 ACV by itself does not bind any G T P v S and does not by itself modulate G T P v S binding to G~ (Fig. 10B). Similar results are obtained with the C2 domain of ACV (see Scholich e t al. 37 f o r details). These data along with the steady slate GTPase assays demonstrate that the C1 C2 ACV and its C2 domain can augment the ability of the fi-adrenergic receptor mimetic peptide to augment guanine nucleotide exchange on G~.tt Using a peptide corresponding to amino acids 382-400 in the Ma muscarinic receptor (MIll-4), we have investigated whether the CI domain of ACV. which interacls with Gi, alters the ability of the muscarinic receptor mimetic peptide to activate Gi. Essentially, using approaches similar to those described above for G~, we have shown that the C 1-C2 ACV and the C 1 domain of ACV augments MIII-4mediated activation of Oi .44 Thus, while the C2 region of ACV increases receptormediated activation of G~ by receptors coupled to this G protein, the C I region of the enzyme enhances the activation of Gi-coupled receptors to actiw~te the inhibitory O protein. These actions of ACV and its subdomains would ensure that the respective G proteins thal modulate its activity are activated in the presence of low amounts of active receptors or in the presence of small amounts of receptor ligands. This mechanism would serve to amplify the signal initiated by activated receptors and at the level of AC ensure the rapid onset of signal. Thus, besides catalyzing the synthesis of cAMP, AC also has two other activities, namely G~o, GAP and enhancer of receptor function. Because of our lindings with Gi, 44 this paradigm is applicable to both the stimulatory and inhibitory GTP-binding proteins of AC.

i l For both CI-C2 ACV and the C2 clomainto increase the guaninenucleotideexchange actix,ity of filll-2, the proteins have to be ucli,,e. Thus. if C I-('2 ACV has losl enzymaticactivityor if the C2 domain has It)st the ability to reconstituteenzymeactivitywhen reconstitutedwith the C I region of ACV. then lhese proteinscannotenhancefilll 2 mediatedGTP-GDPexchange on G~. a4 C. Winpoth. K. Scholich.J. D. Bilycu.and T. B. Patel,.I. Biol. Chem. 275, 25915 (2000).