Tur Jovanat or Biotocical Curmistry
Vol. 254, No. 18, Issue of Septeriber 25, pp. 8927-8931, 1979
Printed in U.S.A.

The Fat Cell Adenylate Cyclase System
CHARACTERIZATION AND MANIPULATION OF ITS BIMODAL REGULATION BY GTP*

(Received for publication, November 13, 1978, and in revised form, May 3, 1979)

Dermot M. F. Cooper, Werner Schlegel, Michael C. Lin, and Martin Rodbell

From the Section on Membrane Regulation, National Institute of Arthritis, Metabolism, and Digestive Diseases, National

Institutes of Health, Bethesda, Maryland 20205

GTP evokés both an activatory and an inhibitory
response from adipocyte adenylate cyclase. This paper
describes thé persistence of the bimodal response under
a variety of assay conditions. Additionally, manipula-
tions are described which eliminate one or other of
these actions. Treatment of adipocyte plasma mem-
branes with cholera toxin A; peptide and NAD* abol-
ishes the inhibitory phase of GTP action while preserv-
ing the activating phase. Treatment of the membranes
with p-hydroxymercuriphenylsulfonic acid eliminates
the activatory phase wnile maintaining the inhibitory
action of the nucieotide. Thus it appears that the two
processes mediated by GTP in adipocytes normally co-
exist and operate through different pathways since
either phase can be abolished leaving the other intact.
Adenosine and its purine-modified analogs inhibit fat
cell adenylate cyclase in the GTP inhibitory phase (Lon-
dos, C., Cooper, D. M. F., Schlegel, W., and Rodbell, M.
(1978) Proce. Natl. Acad. Sci. U. S. A. 75, 5362-5366).
When this effect of GTP is abolished by either cholera
toxin or Gpp(NH)p pretreatment, the inhibitory action
of adenosine analogs is also lost. These data sugges. .
central role for GTP in mediating both activation and
inhibition of adenylate cyclase by agents which act
through cell surface receptors.

 

The enhancement of hormonal stimulation and the stimu-
latory effects of GTP on many adenylate cyclase systems have
received widespread attention (see Ref. 1 and references
therein). The adenylate cyclase of rat adipocyte plasma mem-
branes is unusual in that GTP not only enhances activity but
also causes inhibition of the enzyme (2-6). The potential
regulatory significance of this latter, seemingly paradoxical
behavior of GTP has become apparent from recent findings
in this laboratory (7). These studies showed that the putent
inhibitory actions of adenosine and its purine-modified ana-
logs on cyclic AMP' production in intact adipocytes could be
explained by inhibition of adenylate cyclase in the GTP
inhibitory phase.” Yamamura et «.. (6) have suggested that

* The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.

'The abbreviations used are: cyclic AMP, adenosine 3’:5’-mono-
phosphate; ACTH, adrenocorticotropic hormone, Gpp(NH)p, 5’-
guanylylimidodiphosphate; Kos, the concentration of an agent evoking
half of its maximal effect.

? Throughout this paper, the terms “activatory” and “inhibitory”
GTP phases are used as follows: the activatory phase refers to the
progressive increase in activity seen with increasing GTP concentra-
tions from zero GTP to the peak activity. The decline in activity from
this peak with increased GTP levels is referred to as the GTP
inhibitory phase.

the two effects of GTP are mediated by separate proteins
since trypsin treatment of fat cell membranes abolished the
inhibitory effect of GTP.

- It is not clear from the literature whether the biphasic
effects of GTP on acipocyte adenylate cyclase are always
encountered or to what extent assay conditions modify either
phase. It has been suggested that at low temperatures or low
magnesium concentrations GTP selectively inhibits fat cell
cyclase activity (4, 5). Furthermore, Yamamura et al. (6)
recently showed that chelators and sulfhydryl reagenc:, com-
monly included in the assay of adenylate cyclase, modify the
actions of GTP on the fat cell enzyme. A further complication
arises from the contamination of many commercial ATP
preparations with GTP (8) such that the effects of low GTP
concentrations are obscured.

In the present study using a simplified incubation medium
and purified ATP we have shown that the biphasic effects of
GTP are present under a wide variety of conditions and should
be considered to represer.t the normal response of the fat cell
enzyme to the nucleotide. In addition, we have established
procedures whereby either the activating or inhibitory phase
can be examined separately. When the inhibitory process is
absent, the enzyme is no longer sensitive to inhibition by
purine-modified adenosine analogs. The regulatory impor-
tance and the molecular basis for this behavior is discussed.

MATERIALS AND METHODS

The sources of materials used in the assay of adenylate cyclase
have been reported (9). L-isoproterenol-p-bitartrate and calf intestinal
adenosine deaminase (230 units/ml) were purchased from Sigma.
ACTH'* (Synacthen) was a gift of Ciba-Geigy. N*-Phenylisopropyl-
adenosine was from Dr. J. N. Fain, Brown University. . “ethy!-3-
isobutyl xanthine was bought from Aldrich Chemical Co. The ATP
used in these studies was either purified according to the method of
Kimura et al. (8) or was ‘ne Sigma product (A-2383) prepared by
phosphorylation of adenosine.

Preparation of Fat Cell Membranes—Plasma membranes were
isolated by a simplification of the method of Avruch and Wallach
(10), suspended in | mm EDTA containing 10 mm Tris-HCl, pH 7.5,
and stored in liquid N; as described by Harwood et al. (3).

Adenylate Cyclase Assay—Adenylate cyc.# ‘tivity was assayed
by the method of Salomon ef al. (9) ina mediu.s ataining 0.1 mt
ATP, 1 Ci of (a-"PJATP, 4 mm MgCl, 0.1 mm cyclic AMP, 2 mm
creatine phosphate, creatine phosphokinase at 25 units/ml, 30 mm
Tris-HCl, pH 7.5, and 0.1% crystalline bovine serum albumin. Reac-
tions were initiated by the addition of approximately 1 pg of mem-
brane protein to give a total volume of 0.1 ml. Incubations were
carried out for 30 min at 24°C, apart from the exceptions 1.oted én the
text. Experiments were performed in duplicate or triplicate = Juree
or more batches of fat cell plasma membranes. Replicates agreed to
within 5%; intraexperimental variation was within 80 to 120% of the
values shown.

Treatment with Cholera Toxin—The A, peptide of cholera toxin
was prepared as described previously (11). Fat cell membranes (200

pa) were incubated with 40 ms Tris-HCI, pH 7.5, 1 ma dithiothreitol,

1 mg/ml of bovine serum albumin, 2 mm NAD’, and 20 yg of A,

8927
8928

peptide of cholera toxin in a final volume of 100 yl. After 5 min at
30°C, 2 ml of 10 mM Tris-HCI (pH 7.5) was added. The sample was
then centrifuged at 30,000 x g (15 min, 4°C) and the pellet was
resuspended in 20 mM Tris-HCI containing 1 mg/ml! of bovine serum
albumin.

Treatment with Mercurial—Fat cell membranes (200 yg) were
incubated with 40 mm Tris-HCI, pH 7.5, and 0.3 mm p-hydroxymer-
curiphenylsulfonic acid (Sigma) in a final volume of 100 ul. After 10
min at 0°C, 2 ml of Tris-HCI, pH 7.5 containing 0.2 mm dithiothreitol
was added; the membranes were sedimented (30,000 x g, 15 min, 4°C)
in 20 mm Tris-HCl contair‘ng 1 mg/ml of bovine serum albumin and
resuspended in the same medium prior to adenylate cyclase assay.

RESULTS

Effects of GTP, ITP, and 2'-Deoxy-GTP—A relatively
symmetrical biphasic (activating and inhibitory) relationship
is apparent between the concentration of the three nucleotides
and hormone-stimulated activity (Fig. 1). GTP is the most
potent compound in terms of both its maximum activation
and its effective concentration range. Stimulation of activity
can be obtained by 2 nm GTP while maximal activity is
evoked by 30 nm GTP. The relative steenness of both sides of
-hese curves should be noted, the concentration range in going
from 1C to 90% (or 90 to 10%) of the maximal effect is only
approximately 10-fold in both c~ ses. Such behavior is strongly
suggestive of positive cooperative interactions (12).

Effect of Temperature—The effect of increasing GTP con-
centration on basal, ACTH, and isoproterenol-stimulated ade-
nylate cyclase activity is compared at 24°C and 36°C in Fig.
2.° Previous studies (3, 5) had shown that the inhibition caused
by GTP at 25°C could be either reduced or altered to activa-
tion by raising the temperature to 37°C. Indeed Pairault (5)
had suggested that the only action of GTP was inhibition at
low temperatures which was changed to activation at high
temperatures. The basis for these observations can be appre-
ciated by reference to Fig. 2, where it is apparent that the
entire biphasic curve is present at both temperatures but is
shifted 3- to 4-fold to the right on elevating the temperature.
The total dependence of ACTH and the relative independence
of isoproterenol on GTP for stimulation of activity, which had
been previously pointed out by Yamamura et al. (6), is main-
tained at both temperatures.

Effects of [Mg’*] and [Mn**] on GTP Inhibition—Rodbell
(4) previously showed that at high magnesium concentrations
the inhibition caused by 0.1 mm GTP (3) was changed to
activation. Fig. 3 demonstrates that raising [Mg’*] from 4 to
20 mm caused no selective effect on either the inhibitory or
the activatory GTP phase. Londos and Preston (13) have
shown Mn” to be 50 to 100 times more potent than Mg”* in
activating hepatic adenylate cyclase. When 2 mm Mn?* was
included with 20 mm Mg”* in the incubation medium, the
inhibitory phase of GTP action on the fat cell cyclase system
was eliminated and activation by GTP was reduced to the
extent that ACTH stimulation was at ~~at 10% above basal
activity (Fig. 3), These results suggest that full activation of
adenylate cyclase at a putretive metal ion site (13) renders the
enzyme relatively insensitive to regulation by GTP, as has
been noted previously for hormonal activation of the adipo-
cyte adenylate cyclase system (14).

Effects of p-Hydroxymercuriphenylsulfonate—Generally,
mercurials inhibit adenylate cyclase in a manner that can be
reversed by thiol reagents (15-18). However, pretreatment of
fat cell membranes with p-hydroxymercuriphenylsulfonate

“The GTP curve obtained at 36°C in the present study is consid-
erably more sensitive than that previously reported (6). We attribute
this difference to the use of purified ATP and to the exclusion of

ascorbate from the assay. We find that ascorbate can diminish the
GTP inhibitory phase under --me circumstances.

Bimodal Reguiation of Adenylate Cyclase

 

 

500 “GTP
o 2-deoxy GTP
“|TP
AS
E 400
8
a
2 1
= 300
S
<
oO
0 et
E 200
1001...
0 9 8 7 6 5

LOG [nucleotide], M
Fic. 1. Effects of nucleotide triphosphates on isoprotere:.ol-
stimulated fat cell adenylate cyclase activity. Adenylate cyclase

activity was assayed in the presence of | um isoproterenol under
standard conditions described under “Materials and Methods.”

followed by neutralization with dithiothreitol resulted in a 3-
fold stimulation of basal adi aylate cyclase activity (Fig. 4).
Accompanying the stimulation of basal activity by the mer-
curial was complete loss of the stimulatory effects of GTP on
the enzyme system. Instead, GTP caused frank inhibition of
enzyme activily even at concentrations as low as 2 nM (Fig. 4);
note that the apparent Ko; for GTP was decreased by mcr-
curial treatment relative to that seen in nontreated mem-
branes (Fig. 1). Vhe unique effects of p-hydroxymercuriphen-
ylsulfonic acid pretreatment will be dealt with in detail in a
future report.

Cholera Toxin Treatment—Following pretreatment of fat
cell membranes with cholera toxin (the A-1 subunit) and
NAD*, the typical biphasic effects of GTP were converted
largely to a monophasic activating relationship both in the
absence and presence of hormones (Fig. 5). NAD* was re-
quired for this effect of the A, subunit. It can be seen thet the
stimulatory effect of GTP was enhanced by toxin treatment
as has been reported for other cyclase systems (11, 19-21). It
has been reported that the toxin inhibits the breakdown of
GTP at a specific GTPase associated with the nucleotide
activation process and that this effect of the toxin explains its
activating effects (21). The finding that the toxin abolishes
the GTP inhibitory effect on the fat cell cyclase system raises
the possibility that the toxin also may influence- clase activ-
ity by eliminating a competing GTP-depende:.. -hibitory
process. .

Adenosine Action—Adenosine and analogs such as N°.
phenylisopropyladenosine inhibit fat cell adenylate cyclase
through a receptor that reacts competitively with inethylxan-
thines; with isolated membrane preparations,. inhibitior. by
N*-phenylisopropyladenosine is observed only in the prese-. -
of inhibitory concentrations of GTP (Ref. 7 and Table 1).
Mercurial treatment, shown above to convert GTP action to
a purely inhibitory mode, does not affect the ability of the
adenosine analog to inhibit cyclase activity (Table I). How-

‘No effect of N*-phenylisopropyladenosine was observed at low
(activatory range; 6 x 10°° m) GTP concentrations following any of
the treatments described in Table I.
Bimodal Regulation of Adenylate Cyclase 8929

 

 

36°

Fic. 2. Effect of temperature on

\ the GTP dependency of basal and

{- SN hormone-stimulated activity. Stan-
dard assays were performed for 30 min

3 : ao ZL at 24°C (left panel) and for 7% min at
100 200 A 36°C (right panel) in the absence (@) or
poe ® — presence of either 5 uM isoproterenol

(FSO) (O) or 1 ym ACTH (4).

300
* BASAL 24
250} °ISO 500
€ * ACTH
3 400
@ 200
E
x 150 300
a
50 100
Lg L ee ee ee
0 39 8 7 4 5 0 39 8
LOG [GTP], M
350

pmol cAMP/mg prot/min

 

uh. nylate cyclase activity was determ.sed
1, © in the absence (@) or presence of either

A B Cc YX
300
250 ia .
Fic. 3. Effects of metal ion con-
1 i, f centration on the GTP titration. Ade-

10 um isoproterenol (A) or 2 pu ACTH
(©). Standard conditions were employed
using A, 4mm Mg**; B, 20 mm Mg**; C,
20 mm Mg’* plus 4 mm Mn’.

 

 

“9 8 7 6
LOG [GTP], M

 

ae

ADENYLATE CYCLASE ACTIVITY
pmot cAMP/mg PROTEIN/MIN
8
T

 

 

fl

i
a

1 L 1
1o* 10* 10? 10+
CONCENTRATION OF GTP, M

Fic. 4. Effects of mercurial pretreatment on the GTP titra-
tion of basal adenylate cyclase activity. Cyclase activity was
determined under standard conditiona with membranes which had
been pretreated with p-hydroxymercuriphenyl sulfonate (O) or not
(@) under conditions described under “Materials and Methods.”

 

O1
wF
&

002

cyclic AMP (pmol/mg/min)

 

 

Ly wad ee ed ed
9 8 7 6 + 4

log [GTP], M

Fic. 5. Effecta of cholera toxin pretreatment on the bimodal
actions of GTP. Aliquots (1 yg) of membranea which had been
pretreated with either cholera toxin and NAD* (A, A) or NAD* alone
(@, ©) were assayed under standard conditions in the absence (@, A)
or presence (O, A) of 0.4 pm isoproterenol. Pretreatment with NAD*
alone had no effect on control activity (results not shown).
8930

Tase I

Modification of N”-phenylisopropyladenusine inhibition of fot cell
adenylate cyclase activity by various procedures

The adenylate cyclase activity of 0.5- to 1.5-yg aliquots of variously
treated membranes was determined under standard asaay conditions
in the presence of 2.5 units/ml of adenosine deaminase to reduce
endogenous adenosine levels. Where indicated, GTP and N*-phenyl-
isopropyladenosine (PIA) were present at final concentrations of 9
um and 0.4 pM, reapectively.

 

 

-G > eThD wre
Pretreatment Of £00 tgorroterenol “ota HA “SPH
uM amol cyclic AMP/mg/30 min
Normal None 1.75 1.55 0.39 (25)"
0.4 2.67 460 2.52 (48)
Cholera toxin pre- None 1.70 451 3.97 (88)
treated” 0.4 3.73 7.33 6.36 (86)
GppNHp preacti- 0.4 24.57 2580 23.49 (92)
vated’
Mn®* (2 mm), Mg’* 0.4 10.49 14.34 14.09 (98)
(20 mm) in as-
say”
Mercurial _treat- None 2.01 0.39 0.18 (48)
ment’

 

 

“Values in brackets express the percentage of activity in the
presence of PIA compared to that in its absence. The inhibition
caused by PIA could in all cases be reversed by 250 um 1-methyl-3-
isobutyl xanthine (results not shown).

* Pretreatment of membranes with cholera toxin A, peptide plus
NALD and the mercurial MPS as described under “Materials and
Methods.”

* Membranes (56 yg) were incubated for 60 min at 24°C in the
presence of 10 um Gpp(NEDP ar:! all adenylate cyclase assay com-
ponents except [a-"PJATP. The activity of aliquots (1.4 pg) of this
material was then determined over 10 min by its addition to a mixture
containing (a-"PJATP, GTP, isoproterenol, and PIA as indicated.

“ Activities were determined in the presence of 20 mm Mg’* con-
taining 2mm Mn’* under the conditions indicated in the table.

ever, any treatment which caused loss of GTP inhibition,
including cholera toxin treatment, incubation with 2 mm Mg’*
and 20 mm Mn” or preincubation with Gpp(NH)p,° resulted
in loss of the inhibitory effects of both GTP and N"-phenyli-
sopropyladenosine. These findings indicate an intimate link-
age between the GTP inhibitory process and. the process
through which adenosine inhibits adenylate cyclase.

DISCUSSION

The activating and inhibiting phases of GTP regulation of
fat cell adenylate cyclase, which are particularly pronounced
in the presence of hormones and adenosine, have been clearly
established. A striking feature of the GTP titration curves is
that in most instances each phase occurs over a narre:
concentration range (c/. Fig. 1), which might indicate positive
cooperative interactions (12). However, such indications dis-
appear when one process is abolished o1 masked. Thus, when
the activating phase is abolished by mercurial treatment, the
inhibitory effects of GTP occur over « croader concentration
range and the sensitivity of GTP action is enhanced (cf. Fig.
4). Conversely, when the inhibitory phase is obscured by
cholera toxin treatment, the stimulatory effect of GTP is
enhanced. Such findings lead us to conclude that, in unper-
tuibed conditions, both processes are active and mutually
repressing, possibly through competition for some limited
catalytic activity. Thus, although assay conditions (tempera-
ture, divalent metal ions, and type of hormone) may modify

* Gpp(NH)>p inhibits basal adenylate cyclase in short incubations;
however with longer incubations (>10 min), Gpp(NH)p activates at
all concentrations (4, 22). Foilowing a 60-min preincubation of fat cell
membranes with 10 pm Gpp(NH)p, adenylate cyclase becomes per-
sistently activated and insensitive to GTP inhibition (unpublished
results).

Bimodal Regulation of Adenylate Cyclase

the sensitivity, or the amplitude of the GTP effects, or both,
both phases normally co-exist.

Although the precise mechanism remains obscure, the im-
portance of the bimodal behavior of GTP action is becoming
increas.ngly apparent in terins of the regulatory flexibility it
provides. The activatory phase amplifies hormonal stimiula-
tion (Figs. 2 and 3). In the inhibitory phase, adenosine and
purine-modified adenosine analogs, acting through so-called
R-type receptors inhibit both cyclic AMP production and
lipolysis in the fat cell (7). Here we have shown that the
inhibitory actions of a potent R-site adenosine analog ure
abolished when the GTP inhibitory phase is abolished or
obscured by pretreatment with Gpp(NH)p, cholera toxin, or
by incubating in the presence of high concentrations of diva-
Jent cations. Such total dependence of the analog’s action on
the integrity of the GTP inhibitory phase underlines the
regulatory importance of the latter and elevates it to the same
degree of importance as the role of GTP in the stimulatory
actions of hormones such as catecholamines, ACTH, and
glucagon on the fat cell system. The facility of GTP to regulate
an enzyme system such that in one mode it is insensitive to
regulation by adenosine and in another mode extremely sen-
sitive to such regulation hints that the bimodal effects of the
nucleotide are mediated by separate proteins. Consistent with
this notion is the finding of Yamamura e/ al. (6) that trypsin
treatment of fat cell membranes abolished the GTP inhibitory
process, leaving the GTP stimulatory process intact.

A similar bimodal regulation of adenylate cyclase by GTP
may exist in other cells. Ir human platelets, prostaglandins
stimulate adenylate. -.ase activity whereas :atecholamines,
acting through an a-type receptor, inhibit cyclase activity.
Both actions require GTP in the incubation medium when
membrane fragments containing cyclase activity are examined
for the effects of prostaglandins and catecholamines (23).
Thus, it is possible that those extracellular agents which
inhibit the production of cyclic AMP in cells may do so, in
part, through a GTP-dependent process that inhibits adeny!-
ate cyclase. As a model, the apparent paradoxical bimodal
responses of the fat cell adenylate cyclase system to GTP may
yet prove to be the key for understanding how hormones or
other ext-acellular agents can promote either increased or
decreased levels of cyclic AMP in their target cells.

Acknowledgment—We wish to express our thanks to D.C. Londos
for his careful reading of the manuacript and helpful su; .estions.

REFERENCES

1, Rodbell, M., Lin, M. C., Salomon, Y., Londos, C., Harwood, J. P.,
Martin, B. R., Rendell, M., and Berman, M. (1975) Adv. Cyclic
Nucleotide Res. 5, 3-29

2. Cryer, P. E., Jarrett, L., and Kipnis, D. M. (1969) Biochim.
Biophys. Acta 177, 586-590

3. Harwood, J. P., Low, H., and Rodbell, M. (1973) J. Biol. Chem.
248, 6239-6245

4. Rodbell, M. (1975) J. Biol. Chem. 250, 5826-580 .

5. Pairault, J. (1976) Eur. J. Biochem. 62, 323-334

6. Yamamura, H., Lad, P. M., and Rodbell, M. (1977) J. Biol. Chem.
252, 7964-7966

7, Londos, C., Cooper, D. M. F., Schlegel, W., and Rodbell, M. (1978)
Proc, Natl. Acad, Sci. U.S. A. 75, 5362-5366

8. Kimura, N., and Negata, N. (1977) J. Biol. Chem. 252, 382-3835

9. Selomon, Y., Londos, C., and Rodbell, M. (1974) Anal. Bice. :.
58, 541-548

10. Avruch, J., and Wallach, D. F. M. (1971) Biochim. Biophys. Acta
233, 334-337

11. Lin, M. C., Welton, A. F., and Berman, M. F. (1978) J. Cyclic
Nucleotide Res. 4, 159-168

12. Newsholme, E. A., and Start, C. (1973) Regulation in Metabolism,
pp. 34-87, John Wiley and Sons, London

13. Londos, C., and Preston, M. S. (1977) J. Biol. Chem. 252, 5957-
5961
4

15.
16.

17.

18.

Bimodal Regulation of Adenylate Cyclase

. Pohl, S. L., Birnbaumer, L., and Rodbell, M. (1971) J. Biul. Chem.
246, 1849-1856

Schramm, M., and Naim, E. (1970) J. Biol. Chem. 245, 3225-3231

Storm, D. R., and Chase, R. A. (1975) J. Biol. Chem. 250, 2539-
2545

Spiegel, A. M., Brown, E. M., and Aurbach, G. D. (1976) J. Cyelic
Nucleotide Rex. 2, 393-404

Kather, H., Geiger, M., and Simon, B. (1976) Naunyn-Schmie-
debergs Arch. Pharmacol, 292, 177~182

19,
20,

21.

22.

23.

8931

Bennett, V., and Cuatrecasas, P. (1975) J. Membr. Biol. 22, 1-28

Levinson, S. L., and Blume, A. J. (1977) J. Biol. Chem. 252, 3766-
3774

Cassel, D., and Selinger, %. (1977) Proc. Natl. Acad. Sci. U.S. A.
74, 3307-3311

Ebert, R., and Schwabe, U. (1974) Naunyn-Schmiedebergs Arch.
Pharmacol. 286, 297-313

Jakobs, K. H., Saur, W., and Schultz, G. (1978) FEBS Lett. 85,
§€7-170