Cd
Caldum (lQB0) II,
289307
Q Longman Group UK Lid 1880
Potent and cooperative feedback inhibition of adenylate cyclase activity by calcium in pituitary-derived GM3 cells C.L. BOYAJIAN and D.M.F. COOPER Department of Pharmacology, University of Colorado Health Sciences Center, Denver, CO, USA Calcium (Ca2’) ion concentrations that are achieved intracellularly upon Abstract membrane depolarization or activation of phospholipase C stimulate adenylate cyclase via calmodulin (CaM) in brain tissue. In the present study, this range of Ca2+ concentrations produced unanticipated inhibitory effects on the plasma membrane adenylate cyclase activity of GH3 cells. Ca2+ concentrations ranging from 0.1 to 0.8 @l exerted an Increasing inhibition on enzyme activity, which reached a plateau (3545% inhibition) at around 1 fl. This inhibltory effect was highly cooperative for Ca2+ ions, but was neither enhanced nor dependent upon the addition of CaM (1 fl) to EGTA-washed membranes. The inhibition was greatly enhanced upon stimulation of the enzyme by vasoactive intestinal peptide (VIP) and/or GTP. Prior exposure of cultured cells to pertussis toxin did not affect the lnhlbiilon of plasma membrane adenylate cyclase activity by Ca2’, although in these membranes, hormonal (somatostatin) inhibition was significantly attenuated. Maximally effective concentrations of Ca2+ and somatostatin produced additive inhlbltory effects on adenylate cyclase. The addition of phosphodiesterase inhibitors demonstrated that inhibitory effects of Ca2+ were not mediated by Ca2+-dependent stimulation of a phosphodlesterase actlvlty. These observations provide a mechanism for the feedback inhlbltton by elevated intracellular Ca2’ levels on CAMP-facilitated Ca2+ entry into GHs cells, as well as inhibftory crosstalk between Ca2+-mobilizing slgnals and adenylate cyclase activity. Calcium (Ca’+) ions acting as intracellular messengers, mediate a broad range of physiological processes, including glycogenolysis, cardiac cell pacemaker activity, cell motility, and the activation of Ca2+-dependent proteases [l-4]. In addition, Abbreviations Gi, GS, inhibitory and stimulatory guanine nucleotide regulatory proteins, respectively; CaM, calmodulin, VP, vasoactive intestinal peptide; GppNHp, guanosine S-(fS,y-imido)tiphosphate; ddA. 2’5’dideoxyadenosine; IFMX, isobutylmethylxanthine; FT. pertussis toxin.
Ca2’ ions are integral to the events underlying the secretion of hormones from endocrine cells and the release of transmitters from neurons [5 Given such critical modulatory influences of CaA on cellular responsiveness, it is not surprising that Ca2+ appears to exert considerable influence on major signal transduction systems at the level of the membrane [6-g]. This interaction of Ca signal-generating systems affords the operation of either positive or negative feedback control loops both within and between signalling systems. For instance, in brain, Ca2’ acting via calmodulin (CahQ 299
300
can stimulate adenylate cyclase activi several-fold f2v+ [ 101, which in turn promotes Ca entry via CAMP-mediated phosphorylation of ion channels [2]. Such an interaction provides a feedfotwatd stimulus whereby Ca2’ ions serve to enhance neurotransmitter release 1111. In non-neuronal tissue, the actions of Ca2+ ions on the CAMP-generating system am less clear. There have been reports of CaMdependent Ca2’ regulation of adenylate cyclase activity in peripheral tissues (for review see [12]); however, a range of technical artifacts permits the inference of CaM-dependent stimulation of adenylate cyclase activity by Ca2’ ions, which has not been rigorously confiid in most instances [13]. Studies from intact cells indicate that crosstalk between signal transducing systems is not limited to interactions between the second messengers CAMP and Ca2+. It has been widely documented that treatments that elevate the activity of protein kinase C in intact cells can either enhance or attenuate cyclase adenylate responsiveness 114171. However, few, if any, reports have emerged that demonstrate such effects in isolated plasma In the present study, the effects of membranes. Ca2+ ions and the possible role of CaM in regulating adenylate cyclase activity were investigated in anterior membranes of the rat plasma pituitary-derived GHs cell line. In contrast to previous reports on homogenate or crude particulate preparations from pituitary and GHa adenylate cyclase [IS, 191, but in keeping with one study on purified pituitary membranes by Giannatasio et al. [20], we have encountered a potent, inhibitory effect by Ca2+ on adenylate cyclase activity, which is neither enhanced nor dependent upon the addition of CaM. The regulatory characteristics of this Ca2’ inhibition of GHs adenylate cyclase have been explored with respect to dependence upon GTP, stimulated enzyme states, as well as sensitivity to pertussis toxin, and other agents that interact with the adenylate cyclase system. Materials and Methods Growth of GH3 cells
GHs cells obtained from the American Type Culture
cJsL
CALCIUM
Collection were grown in monolayer culture as described by Tashjian [21], in the presence of Ham’s F-10 medium supplemented with 15% donor horse and 2.5% fetal bovine serum. Preparation of plasma membranes from GH3 cells and rat cerebellum
The medium was removed from cultures of GHs cells and phosphate-buffered saline (12 mM NaHP04, 4 mM KHzPO4, 130 mM NaCl, pH 7.4) containing 0.02% EDTA was added to detach cells. As described in detail by Caldwell et al. [22], the cell suspensions were centrifuged, washed with Phillip’s buffer containing protease inhibitors (20 pg/ml soybean trypsin inhibitor, 4 pg/ml leupeptin, 12 units/ml kallikrein inactivator, 4 p.g/ml antipain), and suspended in Phillip’s buffer diluted by 25% with water. The solutions were gently mixed by inversion and allowed to stand at room temperature for about 10 min. Following centrifugation and lysis of cells in hypo-osmotic buffer containing protease inhibitors, the homogenate was fractionated on a discontinuous gradient of 30 and 40% sucrose solutions in lysis buffer. Material collecting at the 30-4096 interface was removed, washed and resuspended in lysis buffer to a final protein concentration of 0.25-1.0 mg/ml (as determined by the method of Lowry et al. [23], using bovine serum albumin as standard) and stored in liquid nitrogen. Cerebellar plasma membranes were prepared from freshly dissected cerebella of adult male Sprague-Dawley rats as described elsewhere 1241. In addition to the protease inhibitors listed above, dithiothreitol(1 mM) was present in the preparation buffers, and continuous sucrose density gradients (30-38%) were used. Assay of adenylate cyclase activity
The adenylate cyclase activity of purified cerebellar or GHs plasma membranes (generally 66 pg protein per assay) was measured in the presence of the following components (final concentrations): 4 mM phosphocreatine (disodium salt), 20 units/ml creatine phosphokinase, 0.5 mM CAMP, 1 u&ml adenosine deaminase, 1 mM MgCl2, 0.080 mM ATP (disodium salt), 0.01 mM GTP, 3 x lo6
INHIBITION OF ADENYLATE
CYCLASE BY CALClIJM IN GH3 CELLS
cpmassay [a-3aP]-ATP (tetra[triethyl-ammonium] salt), 80 mM Tris-HCl, pH 7.4. The reaction mixture (final volume 100 p.l) was incubated at 30°C for 20 min. Reactions were stopped and the [3aP]-cAME formed was quantified as described by Salomon et al. [25]. Determination offree Ca2’ concentrations Free concentrations of Ca2’ were calculated as previously described in detail [24]. This involved an iterative computing program which solved the equations describing the complexes formed in a mixture comprised of the assay ingredients that affect free divalent cation concentration, i.e., ATE, GTP EGTA H+ Mg2+ Ca” and Na+, using the association cons&us iresented by Martell and Smith [26]. Final assay mixture concentrations of CaC12 (against a background of 200 p.M EGTA) which gave rise to free Ca2* concentrations iu parenthesis, am as follows: 132 (0.080) p.M, 152 (0.10) @I, 1.55 (0.14) @I, 168 (0.22) l&I), 178 (0.33) p.M, 185 (0.49) @I), 191 (0.81) p.M, 197 (1.7) @I, 202 (4.0) @/I, 210 (10) ph4.233 (23) @I, 241 (40) ,uM and 260 (58) /,&I. Treatment of cells with pertussis toxin For pertussis toxin treatment, cell culture media were supplemented with 125 @ml pertussis toxin. Following 24-30 h exposure to the toxin, the cells were harvested and used for preparing membranes for adenylate cyclase assays. Materials Vasoactive intestinal peptide and somatostatin were from Sigma; pertussis toxin from List Biological Laboratories, Inc.; and [a-3+JATP was supplied by ICN Radiochemicals. All other reagents were of the purest grade available from commercial sources.
RM&S In lasma membranes prepared from rat cerebella, 8 Ca exerted a biphasic effect on adenylate cyclase activity, which was absolutely dependent upon the
301
presence of CaM (Fig. la). Ca2’ ions, over the concentration range of 0.1 to 1 @I, stimulated the activity of adenylate cyclase by about threefold and began to inhibit the enzyme at concentrations that exceeded -10 @I. In the absence of CaM, Ca2’ elicited a slight stimulation of adenylate cyclase activity, followed by inhibition at concentrations > In contrast, the effect of these 10 p.M. . concentrations of Ca2+ ions on GHs plasma membrane adenylate cyclase was purely inhibitory, although this inhibition was also biphasic (Fig. lb). That is, Ca2’ inhibited adenylate cyclase activity by about 40% over the concentration range of 0.1 to 1 pIvI, such that a plateau between around 1 and 10 p.IvI Ca2’ preceded a further inhibition at It is significant supramicromolar concentrations. that this first inhibitory phase is elicited by concentrations of Ca2’ which correspond precisely to those that are achieved in intact cells upon depolarization or phospholipase C activation [27], and furthemrore, which elicit the CaM-dependent stimulation of cerebelhu adenylate cyclase described Extensive washes of GH3 pIasma above. membranes with EGTA (1mM)containing buffers and addition of a range of CaM concentrations (0.2 to 2 pM) failed either to attenuate or enhance the inhibition. Moreover, the addition of compounds known to antagonize CaM-mediated processes [24] - trifluopemzine (3 pM) and calmidazolium (1 pM) - did not abolish or attenuate the inhibition of enzyme activity induced by 1.7 @I Ca2’ (data not shown). Hill plot analysis of the inhibitory influence of Ca2+ over the 0.1 to 1 pM concentration range from four separate experiments revealed a highly cooperative mechanism (Fig. 2). This cooperative inhibition yielded a Hill coefficient (nB) of 2.21 + 0.47. Since the attenuation of adenylate cyclase activity by neurotransmitters and hormones requires the presence of GTE as an activating ligand of GTE-regulatory proteins, such as Gi, the Ca2+ inhibition of adenylate cyclase was examined as a function of increasing GTP concentrations. As shown in Figure 3. no inhibition was evident in the absence of GTP; as increasing concentrations of GTE were included, the inhibition of adenylate cyclase by 1.7 p.M Ca2’ was progressively revealed.
302
CELL CALCIUM
10-7
10-6 free
10-s
io-‘10-8
10-7
10-6
10-s
10 I4
free [W+] (M)
I=+1 (Ml
Fig. 1 Effects of Ca2’ and CaM on adenylate cyclasc activity in EGTA-washed
membranes prepamd from a) rat cerebellum and b)
GH3 cells. Adenylate cyclase activity was measured: a) in the absence of stimulating hormone; and b) in the presence of 0.1 p.M VIP. Concentrations of free Ca2’ wez~?determined by an iterative computer program, as described in Materials and Methods. Data ate from a representative experiment that was replicated at least four times and arc expressed as the mean activity f standard deviation of triplicate determinations
Fig. 2
Hill plot analysis
of the inhibition
by Ca2” in the
submicromolar range of GH3 adenylate cyclase activity. Data am from four indepcndcnt
experiments
Figure lb, each performed in triplicate
similar to that shown in
Since vasoactive intestinal peptide (VIP, 0.1 j.tM) was present throughout this experiment (Fig. 3), it was not possible to ascertain whether GTP itself or stimulation of the adenylate cyclase activity was responsible for enhancing the extent of Ca2’ Figure 4 demonstrates that in the inhibition. presence of 10 pM GTP, increasing VIP concentrations produced a concentration-dependent stimulation of adenylate cyclase activity. The addition of 117 p.M Ca2’ attenuated this stimulation such that its greatest inhibitory effect was apparent at near maximal stimulatory concentrations of VIP. The inhibitory effect of Ca2’ upon both GTP- and VIP-stimulated adenylate cyclase activity was saturable (Figs 3 and 4). The yfarent GTP- (or stimulation-) dependence of the Ca inhibitory effects on adenylate cyclase is analogous to the receptor-mediated inhibition of GHs adenylate cyclase by hormones (e.g. somatostatin) acting via Gi [28]. Consequently, we examined the effect of pretreating cultured cells with pertussis toxin upon the inhibition of adenylate cyclase activity by somatostatin and Ca2’ ions. Similar levels of VIP-stimulated adenylate cyclase
INHIBITION
-
OF ADENYLATE
500
1
CYCLASE
n
303
IN GH3 CELLS
SW
Control +1.7uMCa2+
a
0
d-4
Control +1.7pM ca2+
!~ n
400
: t
BY CALCIUM
4M1
I
F
d
300
300
,200
’ loo
f-lil
~~
I ..
PJW(4
[GW WI of Ca’+ inhibition
Fig. 3 Dependence activity
on GTP.
Activity
pM VIP at the indicated means
and standard
performed
of GH3 adenylate
was measured
GTP concentration.
deviations
from
cyclase
in the presence Valm
of 0.1 am the
three experiments,
each
+1.7y
4.1pu
Cal+
l1.7pyI
SST
c22*
*alrUsSr
Fig.
5
Differential
effects
of pertussis
hormone-mediated
inhibition
cells were exposed
to 125 ng/ml pertussis
medium
ove-rnighs
described
in Materials
were measured Data
experiments
activities
on Ca2+ and
Cultures of GH3
toxin in their growth
membranes
and Methods.
in the presence
conducted
toxin
of GH3 cyclase.
and plasma
are the mean
were prepared
Adenylate
as
cyclase activities
of 0.1 p.M VIP and 10 pM GTP. f standard
in triplicate.
toxin-treated,
hatched bars, untreated,
*P < 0.02.
by unpaired
Student’s
deviation from three Pertussis SST, somatostatin. opar bars t-test,
when
compared
to
t-test, when compared
to
Control. + 1.7 pM Ca2+ group **P < 0.001,
on VIP.
by unpaired
Control, + 0.1 @vf SST group
Student’s
on Ca2’ inhibition Activity
@vi GTP at the indicated means
and
perfolmed
in triplicate
NO 2ddltlon
Fig. 4 Dependence activity
standard
of G&
was measured
VIP concentmtions.
deviations
from
adenylate
in the presence two
Values experiments,
cyclase of 10 are the each
in triplicate
activity were displayed by plasma membranes derived from cells that had ben either untreated or exposed to pertussis toxin (Fig. 5). This activity inhibited by approximately 23% by WaS somatostatin (0.1 @I), and the inhibition was reduced to about 12% upon pertussis toxin pretreatment. On the other hand, a 38% inhibition elicited by Ca2’ ions was unaltered by treatment with the toxin. In combination9 Ca2’ and somatostatin elicited an additive inhibition of over 5596, of which only the hormone (somatostatin) component was attenuated by pertussis toxin. Since Harden et al. [29] had demonstrated previously (in intact 1321Nl astrocytoma cells) that Ca2’ ions could activate a phosphodiesterase activity to reduce cAh4P levels, the possibility was addressed that the present inhibitory effects of Ca2’ on [3?P]-cAMP formation was media&A by Ca2+-dependentstimulation of a phosphodiesterase activity (that had not been neutralizd by the standard inclusion of 0.5 mM CAMP in the assay). The inhibitory effects of Ca2’ on adenylate cyclase were identical, in either the presence or absence of high concentrations of phosphodiesterase inhibitors (500 p.M IsA@ and 100 pA4 Ro20-1724; data not
304
CELL,cALaJh4
Table 1 Effects of various regulatory agents on the Ca” inhibition of GH3 adenylate cyclase activity Assays were performed as described in Materials and Methods. with 1 mM MgC12 present in each assay. Data are expressed as the mean activity f standard deviation of at least three triplicate samples, from three separate experiments. GppNHp. guanosine S-(fl, y-imido) triphosphate; dd.A. 2’, S-didwxyadenosine Adenykate cyckase activity (pmoUmg.min) Control
- GTP, - VIP No addition +l
pMForskolin
+ 10 ph4GppNHp
+lOpMGTP,-VIP No actdition
+ 10 @l
+ I.7 pM ca2’
% Inhibition
10.7 f 1.1
12 f 6
117 f6.2
93.4 f 4.0
20 f 5
354 f 5.9
204f26
42 f 7
32.9 f 1.9
36f4
12.1 f 0.31
51.2 f 1.5
+2mMMncli!
376fll
338 f21
10f4
GTP, + 0.1 m VIP No addition
435 f 17
243 f 13
44f5
+lOOpMddA
219 f 9.2
137 f 6.3
37f4
Thus, it seems likely that the Ca2+ shown). inhibition is exerted specifically upon the adenylate cyclase activity, either by means of a Ca2+-binding protein, or by dire&y interacting with one of the components of the cyclase system In an attempt to identify the site of action at which Ca2+ inhibition was mediated, a variety of agents that modulate the adenylate cyclase system were examined for their influence on Ca2+ As described earlier (Pig. 3), the inhibition. dependence on GTP for this effect is evident (12% versus 36% inhibition by 1.7 pM Ca2+, in the absence and presence of GTP, respectively; Table I). Furthermore, the addition of VIP (with 10 pM GTP present) enhances the inhibitory effect to 44% (Table 1). The adenylate cyclase stimulatory agents, forskolin, GppNHp and MnCl2 alI stimulate GH3 adenylate cyclase activity to differing extents, but ;y I$C12;$$earst;leY ;ietiichE dideoxyadenosine, a F-site inhibitor of adenylate cyclase [30] inhibited cyclase by 5096, its presence did not preclude a further 37% inhibition by 1.7 pM Ca2’ (Table 1). Thus, it seems unlikely that the inhibitory effect of Ca2’ is mediated at the putative P-site.
Discussion In the present study, we examined the effects of Ca2’ in the concentration range that is reached intracellularly, following membrane depolarization or phospholipase C activation, upon the adenylate cyclase activity of GH3 plasma membranes. Ca2’ ions exerted a potent and cooperative, inhibitory effect on GH3 adenylate cyclase activity. These results are similar to those obtained in rat anterior pituitary membranes by Giannatasio et al. 1201, who described a biytasic inhibition of adenylate cyclase activity by Ca over concentration ranges identical to those used in the present report Also consistent with the present findings the Ca2’ inhibition of anterior pituitary adenylai cyclase did not require CaM and was not attenuated by anticahnodulin drums. These investigators examined the effect of Ca ’ on the stimulation by Mg2+ of adenylate cyclase; a kinetic analysis of their data afforded the interpretation that Ca2+ interacts competitively with an allosteric Mg2+ site on adenylate cyclase. This kinetic approach does not address the structural or mechanistic features of the site of Ca2+ inhibition on cyclase. This point is particularly relevant since Mg2+ plays many roles in the regulation of
INHIBITION OF ADENYLATE CYCLASE BY CALCIUM IN GH3 CELLS
adenylate cyclase - at the level of receptors, G-proteins and the catalytic unit of cyclase itself [3 11. The present paper sought to address the regulatory significance of Ca2+ inhibitory effects on cyclase, in terms of GTP-dependency, pertussis toxin-sensitivity, or sensitivity to agents known to influence the activity of the enzyme. In exploring the regulatory characteristics of the inhibition of GHa adenylate cyclase by Ca2+ ions, we discovered that inhibition is not observed in the absence of GTP, but is revealed as the concentration of the nucleotide is increased to about 1 r_LM.In addition, in the presence of 10 @vi GTE, significant Ca2+ inhibition is not apparent until VIP elicits a substantial stimulation of basal cyclase activity. From these results, it is not possible to determine whether GTP is specifically required, or whether there is both a GTP dependency and a requirement for VIP-stimulated states of adenylate cyclase. In GHa cells, hormones that act through G-proteins to inhibit adenylate cyclase, also require GTE and stimulated states of the enzyme [28]. In this context, the inhibition of adenylate cyclase by low analogous to Ca2+ concentrations appears hormone-mediated inhibition of cyclase. However, there are apparent differences in their mechanisms. For example, the magnitude of the inhibition by Ca2+ (35-45%) exceeds that elicited by inhibitory hormones. and Ca2+ inhibition is not sensitive to pretreatment of cells with pertussis toxin, which Moreover, renders Gi proteins dysfunctional. maximally effective concentrations of Ca2+ and somatostatin elicit au additive inhibition (> 55%) on adenylate cyclase activity, which indicates that these agents do not share common mechanisms of action. These data do not rule out the possibility that a G-protein that is not a pertussis toxin substrate may mediate Ca2’ effects on adenylate cyclase. The adenylate cyclase activity of GHs is inhibited by Ca2’ ion concentrations that are achieved intracellularly upon either membrane depolarization or activation of receptors coupled to the phosphatidylinositol signalling s stem [27]. In It brain, identical concentrations of Ca serve instead to elevate CAMP production about threefold, in an entirely CaM-dependent manner [7, 101. The effect of these low concentrations of Ca2+ on cyclase activity, whether stimulatory or inhibitory, provides
305
Ca2+-mobilizing and link between a CAME-mediated signal transduction systems 111, 20 . Especially noteworthy is the finding that while Ca1+ stimulation of neuronal adenylate cyclase depends absolutely upon CaM, equivalent Ca2+ concentrations in GHs cells act independently of CaM to reduce cyclase activity beyond that achieved The highly even by inhibitory hormones. cooperative nature of this Ca2+ inhibitory effect suggests that a Ca2+-binding protein which has adapted to this purpose may be involved. Nevertheless, since several CaM antagonists failed to affect the ability of Ca2+ to inhibit cyclase activity, the Ca2+ ion binding site may not be analogous to those found on proteins displaying E-helix-loop-F-helix structures 1341. In keeping with the present results, which suggest that [Ca2+]i in the physiological range can inhibit plasma membrane adenylate cyclase, studies in intact cells suggest that hormones or neurotransmltters that elevate [Ca2+]1 can decrease the production of CAMP. In some instances, at least part of this effect was due to the stimulation of a phosphodiesterase activity. Harden et al. [29] demonstrated that in 1321Nl astrocytoma cells, the phosphodiesterase inhibitor isobutylmethylxanthine reversed muscarinic cholinergic receptor-mediated inhibition of agonist-stimulated CAMP accumulation. However, as reported by others in intact NCB-20 cells 1351, as weIl as in the present study, the inclusion of high concentrations of phosphodiesterase inhibitors did not perturb Ca2’ inhibition of adenylate cyclase. These data suggest that Ca2+ can act directly on a component of the adenylate cyclase system Although we interpret the present data to indicate direct effects of Ca2+ on adenylate cyclase activity, the possibility could be entertamed that given the GTP and VIP enhancement of the inhibitory effects - phospholipase C was being stimulated in these preparations. This could result in the liberation of diacylglycerol and activation of endogenous protein kinase C, which might then mediate the inhibition in a Ca2+-dependent manner. This possibility seems highly unlikely for a number of reasons: i) this plasma membrane preparation from a sucrose gradient is not likely to contain significant amounts of protein kinase C; and ii) the
306
effect is seen directly in a plasma membrane preparation, whereas the majority of reports of protein kinase C modulating adenylate cyclase activity rely on a pretreatment of cells with protein kinase C stimulants, prior to the preparation of a membrane fraction [see e.g. 11-141. Indeed in another cell line (NCB-20) in which we have encountered similar actions of Ca’+, neither inhibitors of protein kinase C, such as sphingosine, nor treatment of the membranes with protein kinase C and diacylglycerol, mod&d the inhibitory effects of Ca2’ (Boyajian and Cooper, in preparation). In contrast to the present fiidings and those of Giannatasio et al. [20], Greenlee and Okada [ 181 reported a modest (-20%) activation of adenylatc cyclase by Ca2+ in a particulate preparation from rat anterior pituitary, which was dependent upon the presence of supramicromolar concentrations of CaM. Relatively crude preparations utilized in the latter studies, unlike the purified plasma membrane preparations used in the present case, may have permitted stimulatory effects to be mediated by soluble factors loosely associated with the plasma There is some controversy as to membrane. whether Ca2’/CaM stimulation of adenylate cyclase occurs outside of the central nervous system [131. Several methodological factors may underlie misleading inferences of Ca2+ or Ca2+/CaM stimulatory effects on adenylate cyclase activity. ‘disinhibition’ by include possible These supramicromolar Cah4 of adenylate cyclase activity due to chelation of Ca2’ ions activation of Ca2+-dependent protease activities’ that stimulate adenylate cyclase [32, 333, or activation of Ca2+- or CM-dependent protein kinase in untractionated membrane preparations. The present inhibition of adenylate cyclase activity by Ca2’ ions is saturable and of sufficient magnitude to suggest that Ca2’ signalling systems may play a significant role in the modulation of the activity of the cAA@generating system, and its dependent physiological processes. For instance, both protein kinase A, via phosphorylation mechanisms [36], and Gs of the adenylate cyclase systems [37, 381, can stimulate the activity of voltage-gated Ca2+ channels in GHs cells. Resulting elevations in intracellular Ca2’ could provide feedback inhibitory control of such CAMP-
CELL CALCIUM
or G,facilitating actions. In support of this view, Dufy and colleagues 1391have noted previously that the elevation of intracellular Ca2+ via voltage-gated channels appears to result in a feedback inhibition of the activity of the channel. It is interesting that very recently a similar observation has been made for another GTP regulatory protein-mediated signal transduction system, i.e. the retinal guanylate c yclase enzyme 1401. In this instance, submicromolar concentrations of Ca2’ inhibited enzyme activity independently of CaM in a highly cooperative manner (nH = 3.9), by about 40%. In addition, Ca2’ has recently been reported to inhibit adenylate cyclase in MA-10 Leydig cells [41]. Thus, it may transpire that the potent inhibition of adenylate cyclase activity by Ca2’ ions observed presently reflects a widespread intracellular Ca2+-sensing device for the feedback inhibition of elevated intracellular Ca2+ resulting in turn in the modulation of various celhtlar signal transduction systems. Acknowledgements The authors thank Ms Lusnne Esle-r for maintaining GH3 cells in culture, and Ms Kathy Fassler for expert secretsrid assistance. These studies were supported by NH-IgrantGM 32483.
References 1. Rasmussen H. (1980) calcium and CAMP in stimuhu response coupling. Ann. NY Acad. Sci., 356,346-353. 2. Meldolesi J. Possan T. (1987) Pathways of Ca2’ influx at the plasma membrane: voltage-, receptor-, and second messenger-operated channels. Exp. Cell Res., 17.271-283. 3. Benidgc UT. Galone A. (1988) Cytosolic calcium oscillators. FASEB J., 2, 3074-3082. 4. Kater SB. Mattson MP. Cohau C. Connor J. (1988) Calcium regulation of the neuronal growth core. Trends Neurosci., 11,315-321. 5. Smitb SJ. Augustine GJ. (1988) Calcium ions, active zones and synaptic transmitter release. Trends Neurosci.. 11, 458-464. 6. Tomlinson S. MacNeil S. Brown BL. (1985) Calcium, cyclic AMP and hormone action. Clin. Endocrinol.. 23. 595-610. 7. Cooper DMF. Ahlijanisn MK. Perez-Reyes E. (1988) Calmodulin plays a dominant role in determining neumtmnsmitter regulation of neuronal ade-nylatecyclaae. J. Celh11.B&hem.. 36,417-427. 8. Nahorski SR. (1988) Inositol polyphosphates and neuronal calcium homeostasis. Trends Neumsci.. 11.444-448. 9. Eberhard DA. Halz RW. (1988) Intracellular Ca2+ activates phospholipase C. Trends Neurosci., 11,5 17-520. 10. Brostrom Co. Huang Y. Bceckemidge BMcL. WolfTDL.
INHIBITION OF ADJZNYLATE CYCLASE BY CALCIUM IN GH3 CELLS
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26. 27.
(1975) Idemitication of a calcium-binding protein as a calcium-dependent regulator of brain adenylate cyclase. Proc. Natl. Acad. Sci. USA, 72,646s. Cooper DMF. Caldwell KK. Pemz-Reyes E. Ahlijanian MK. Schlegel W. (1988) Interactions between neutottansmitters that regulate CAMP and intracellular Ca” levels in the CNS. In: eds. Kito S. Segawa T. Kuriyama K. Tohyama M. Olsen RW. Neuromceptors and Signal Transdix%ion. Advances in Experimental Medicine and Biology, vol. 236, pp. 217-227. MacNeil S. Lakey T. Tomlinson S. (1985) Cahnodulin regulation of adenylate cyclase activity. Cell Calcium, 6, 213-226. Cooper DMF. Caldwell KK. (1988) Does cahnodulin regulate non-neummtl adenylate cyclam? B&hem. J., 254, 935-936. Summers ST. Cronin MJ. (1988) Phorbol esters induce two distinct changes in GH3 pituitary cell adenylate cyclase activity. Arch. Biochem. Biophya., 262, 12-18. Norstedt C. Jondal M. Fredholm B. (1987) Activation of protein kinase C inhibits prcwtaglandin - and potentiates adenosine receptor-stimulated accumulation of cyclic AMP in a human T-cell leukemia lioe. FEBS L&t., 220,57&O. Wiener E. Scatpa A. (1989) Activation of protein kinase C modulates the adenylate cyclase effector system of B-lymphocytes. J. Biol. Chem., 264, 43244328. Bell JD. Brunton LL. (1987) Multiple effects of phorbol esters on hormone-sensitive adenylate cyclase activity in S49 lymphoma cells. Am. J. Physiol., 15, E783E789. Greenlee DV. Okada S. (1987) Calmodulin activates adenylate cyclase from rat anterior pituitary. Mol. Pharmacol., 32.743-748. Brostrom MA. Btotman LA. Brostrom Co. (1982) Calcium-dependent adenylate cyclase of pituitary tumor cells. B&him. Biophys. Acta, 721.227-235. Giannatasio G. Bianchi R. Spada A. Vallar L. (1987) Effect of calcium on adenylate cyclsae of rat anterior pituitary gland. Endocrinology, 120,2611-2619. Tashjian AH. Jr. (1979) Clonal stmins of hormone-producing pituitary cells. Methods Enzymol.., 58, 527-535. Caldwell KK. Newell MK. Cambier JC. et al. (1988) Evaluation of methods for the isolation of phtsma membranes displaying guanosine 5’-hiphosphate-dependence for the regulation of adenylate cyclase activity: Potential application to the study other guanosine S’-triphosphate-dependent transduction systems. Anal. Biochem., 175,177-190. Lowry OH. Rosebtough NJ. Farr AL. Randall RJ. (1951) Protein measumment with the Folin phenol reagent. J. Biol. Chem., 193,265-275. Ahlijanian MK. Cooper DMF. (1987) Antagonism of cahnodulin-stimulated adenylate cyclase of trifluoperazine, cahnidazolium and W-7 in rat cembellar membranes. J. Phannacol. Exp. Ther., 241, 407-414. Salomon Y. Londos C. Rodbell M. (1974) A highly sensitive adenylate cyclase assay. Anal. Biochem., 58, 541-548. Matte.11AE. Smith RM. (1975) Critical Stability Constants. New York, Plenum Press, vol. 2. pp. 269-284. Tsien RY. Pozzan T. Rink TJ. (1984) Measuring and manipulating cytosolic Ca’+ with trapped indicators. Trends
307
Bionhem. Sci.. 9.263-266. 28. Cooper DMF. Caldwell KK. Boyajian CL. Petcoff DW. Schlegel W. (1989) Adenosine Al receptors inhibit both adenylate cychise activity and TRH-activated Ca2+channels by a pertussis toxin-sensitive mechanisms in GHz cells. Celhtl. Signal., 1.85-97. 29. Harden TK. Evans T. Hepler JR. et al. (1985) Regulation of cyclic AMP metabolism by muscarinic cholinergic receptors. In: eds, Cooper DMF. Seamon KB. Advances in cyclic micleotide and protein phosphorylation research. New York, Raven Press, voL 19, pp, 207-220. 30. Londos C. Wolff J. (1977) Two distinct a&no&z-sensitive sites on adenylate cycbtse. Proc. Natl. Aced. Sci. USA, 77,2551-2554. 31. Iyengar R Bimbaumer L. (1981) Hysteretic activation of adenylyl cyclase. I. Effects of Mg ion on the rate of activation of guanine nucleotides. J. Biol. Chem., 256, 11036-11041. 32. Adnot S. Poirier-Dupuis M. Franks DJ. Hamet P. (1982) Stimulation of rat platelet adenylate cyclase by an endogenous calcium-dependent pmtease-like activity. J. Cyclic Nucleotide Res.. 8, 103-l 18. 33. Wallace MA. Fain JN. (1985) Guanosine 5’-G-thiotriphosphate stimulates phospholipase C activity in plasma membranes of tat hepatocytes. J. Biol. Chem., 260, 9527-9530. 34. Cox JA. (1988) Interactive properties of calmodulin. B&hem. J., 249,621-629. 35. Hollingsworth EB. Daly JW. (1987) Inhibition of receptor-mediated stimulation of cyclic AMP accumulation in neuroblastoma-hybrid NCB-20 cells by a phorbol ester. Biocbim. Biophys. Acta, 930,272-278. 36. Armstrong D. Eckert R (1987) Voltage-activated calcium channels that must be phosphotylated to respond to membrane depolarization. Proc. Natl. Acad. Sci. USA, 84, 2518-2522. 37. Hedlund B. Dufy B. Bat& JL. (1988) Vasoactive intestinal polypeptide alters G&/B6 pituitary cell excitability. Pfluger’s Arch., 411. 173-179. 38. Bjoro T. Ostberg BC. Sand 0. et al. (1987) Vasoactive intestinal peptide and peptide with N-terminal hi&line and C-terminal isoleucine increase prolactin secretion in cultured rat pituitary cells (GI-LtCt)via a CAMP-dependent mechanism which involves transient elevation of intracellular Ca2+. Mol. Cellul. Bndocrinol., 49, 119-128. 39. Dufy B. Jaken S. Barker JL. (1987) Intraoellular Ca’+-dependent protein kinase C activation mimics delayed effects of thyrotropin-releasing hormone on clonal pituitary cell excitability. Endocrinology, 121,793~802. 40. Koch K-W. Stryer L. (1988) Highly cooperative feedback control of retinal rod guanylate cyclase by calcium ions. Nature, 334,64-66. 41. Pereira MB. Segaloff DL. Ascoli M. (1988) Ca2+ is an inhibitor of adenylate cyclase in MA-10 Leydig tumor cells. Endocrinology, 122,2232-2239. Please send reprint requests to : Dr D.M.F. Cooper, Department of Pharmacology, University of Colorado Health SciCenter, 4200 E. Ninth Avenue. Denver, CG 80262, USA Received : 9 October 1989 Revised : 7 November 1989 Accepted r 22 November 1989
Caldum (lQB0) II,
289307
Q Longman Group UK Lid 1880
Potent and cooperative feedback inhibition of adenylate cyclase activity by calcium in pituitary-derived GM3 cells C.L. BOYAJIAN and D.M.F. COOPER Department of Pharmacology, University of Colorado Health Sciences Center, Denver, CO, USA Calcium (Ca2’) ion concentrations that are achieved intracellularly upon Abstract membrane depolarization or activation of phospholipase C stimulate adenylate cyclase via calmodulin (CaM) in brain tissue. In the present study, this range of Ca2+ concentrations produced unanticipated inhibitory effects on the plasma membrane adenylate cyclase activity of GH3 cells. Ca2+ concentrations ranging from 0.1 to 0.8 @l exerted an Increasing inhibition on enzyme activity, which reached a plateau (3545% inhibition) at around 1 fl. This inhibltory effect was highly cooperative for Ca2+ ions, but was neither enhanced nor dependent upon the addition of CaM (1 fl) to EGTA-washed membranes. The inhibition was greatly enhanced upon stimulation of the enzyme by vasoactive intestinal peptide (VIP) and/or GTP. Prior exposure of cultured cells to pertussis toxin did not affect the lnhlbiilon of plasma membrane adenylate cyclase activity by Ca2’, although in these membranes, hormonal (somatostatin) inhibition was significantly attenuated. Maximally effective concentrations of Ca2+ and somatostatin produced additive inhlbltory effects on adenylate cyclase. The addition of phosphodiesterase inhibitors demonstrated that inhibitory effects of Ca2+ were not mediated by Ca2+-dependent stimulation of a phosphodlesterase actlvlty. These observations provide a mechanism for the feedback inhlbltton by elevated intracellular Ca2’ levels on CAMP-facilitated Ca2+ entry into GHs cells, as well as inhibftory crosstalk between Ca2+-mobilizing slgnals and adenylate cyclase activity. Calcium (Ca’+) ions acting as intracellular messengers, mediate a broad range of physiological processes, including glycogenolysis, cardiac cell pacemaker activity, cell motility, and the activation of Ca2+-dependent proteases [l-4]. In addition, Abbreviations Gi, GS, inhibitory and stimulatory guanine nucleotide regulatory proteins, respectively; CaM, calmodulin, VP, vasoactive intestinal peptide; GppNHp, guanosine S-(fS,y-imido)tiphosphate; ddA. 2’5’dideoxyadenosine; IFMX, isobutylmethylxanthine; FT. pertussis toxin.
Ca2’ ions are integral to the events underlying the secretion of hormones from endocrine cells and the release of transmitters from neurons [5 Given such critical modulatory influences of CaA on cellular responsiveness, it is not surprising that Ca2+ appears to exert considerable influence on major signal transduction systems at the level of the membrane [6-g]. This interaction of Ca signal-generating systems affords the operation of either positive or negative feedback control loops both within and between signalling systems. For instance, in brain, Ca2’ acting via calmodulin (CahQ 299
300
can stimulate adenylate cyclase activi several-fold f2v+ [ 101, which in turn promotes Ca entry via CAMP-mediated phosphorylation of ion channels [2]. Such an interaction provides a feedfotwatd stimulus whereby Ca2’ ions serve to enhance neurotransmitter release 1111. In non-neuronal tissue, the actions of Ca2+ ions on the CAMP-generating system am less clear. There have been reports of CaMdependent Ca2’ regulation of adenylate cyclase activity in peripheral tissues (for review see [12]); however, a range of technical artifacts permits the inference of CaM-dependent stimulation of adenylate cyclase activity by Ca2’ ions, which has not been rigorously confiid in most instances [13]. Studies from intact cells indicate that crosstalk between signal transducing systems is not limited to interactions between the second messengers CAMP and Ca2+. It has been widely documented that treatments that elevate the activity of protein kinase C in intact cells can either enhance or attenuate cyclase adenylate responsiveness 114171. However, few, if any, reports have emerged that demonstrate such effects in isolated plasma In the present study, the effects of membranes. Ca2+ ions and the possible role of CaM in regulating adenylate cyclase activity were investigated in anterior membranes of the rat plasma pituitary-derived GHs cell line. In contrast to previous reports on homogenate or crude particulate preparations from pituitary and GHa adenylate cyclase [IS, 191, but in keeping with one study on purified pituitary membranes by Giannatasio et al. [20], we have encountered a potent, inhibitory effect by Ca2+ on adenylate cyclase activity, which is neither enhanced nor dependent upon the addition of CaM. The regulatory characteristics of this Ca2’ inhibition of GHs adenylate cyclase have been explored with respect to dependence upon GTP, stimulated enzyme states, as well as sensitivity to pertussis toxin, and other agents that interact with the adenylate cyclase system. Materials and Methods Growth of GH3 cells
GHs cells obtained from the American Type Culture
cJsL
CALCIUM
Collection were grown in monolayer culture as described by Tashjian [21], in the presence of Ham’s F-10 medium supplemented with 15% donor horse and 2.5% fetal bovine serum. Preparation of plasma membranes from GH3 cells and rat cerebellum
The medium was removed from cultures of GHs cells and phosphate-buffered saline (12 mM NaHP04, 4 mM KHzPO4, 130 mM NaCl, pH 7.4) containing 0.02% EDTA was added to detach cells. As described in detail by Caldwell et al. [22], the cell suspensions were centrifuged, washed with Phillip’s buffer containing protease inhibitors (20 pg/ml soybean trypsin inhibitor, 4 pg/ml leupeptin, 12 units/ml kallikrein inactivator, 4 p.g/ml antipain), and suspended in Phillip’s buffer diluted by 25% with water. The solutions were gently mixed by inversion and allowed to stand at room temperature for about 10 min. Following centrifugation and lysis of cells in hypo-osmotic buffer containing protease inhibitors, the homogenate was fractionated on a discontinuous gradient of 30 and 40% sucrose solutions in lysis buffer. Material collecting at the 30-4096 interface was removed, washed and resuspended in lysis buffer to a final protein concentration of 0.25-1.0 mg/ml (as determined by the method of Lowry et al. [23], using bovine serum albumin as standard) and stored in liquid nitrogen. Cerebellar plasma membranes were prepared from freshly dissected cerebella of adult male Sprague-Dawley rats as described elsewhere 1241. In addition to the protease inhibitors listed above, dithiothreitol(1 mM) was present in the preparation buffers, and continuous sucrose density gradients (30-38%) were used. Assay of adenylate cyclase activity
The adenylate cyclase activity of purified cerebellar or GHs plasma membranes (generally 66 pg protein per assay) was measured in the presence of the following components (final concentrations): 4 mM phosphocreatine (disodium salt), 20 units/ml creatine phosphokinase, 0.5 mM CAMP, 1 u&ml adenosine deaminase, 1 mM MgCl2, 0.080 mM ATP (disodium salt), 0.01 mM GTP, 3 x lo6
INHIBITION OF ADENYLATE
CYCLASE BY CALClIJM IN GH3 CELLS
cpmassay [a-3aP]-ATP (tetra[triethyl-ammonium] salt), 80 mM Tris-HCl, pH 7.4. The reaction mixture (final volume 100 p.l) was incubated at 30°C for 20 min. Reactions were stopped and the [3aP]-cAME formed was quantified as described by Salomon et al. [25]. Determination offree Ca2’ concentrations Free concentrations of Ca2’ were calculated as previously described in detail [24]. This involved an iterative computing program which solved the equations describing the complexes formed in a mixture comprised of the assay ingredients that affect free divalent cation concentration, i.e., ATE, GTP EGTA H+ Mg2+ Ca” and Na+, using the association cons&us iresented by Martell and Smith [26]. Final assay mixture concentrations of CaC12 (against a background of 200 p.M EGTA) which gave rise to free Ca2* concentrations iu parenthesis, am as follows: 132 (0.080) p.M, 152 (0.10) @I, 1.55 (0.14) @I, 168 (0.22) l&I), 178 (0.33) p.M, 185 (0.49) @I), 191 (0.81) p.M, 197 (1.7) @I, 202 (4.0) @/I, 210 (10) ph4.233 (23) @I, 241 (40) ,uM and 260 (58) /,&I. Treatment of cells with pertussis toxin For pertussis toxin treatment, cell culture media were supplemented with 125 @ml pertussis toxin. Following 24-30 h exposure to the toxin, the cells were harvested and used for preparing membranes for adenylate cyclase assays. Materials Vasoactive intestinal peptide and somatostatin were from Sigma; pertussis toxin from List Biological Laboratories, Inc.; and [a-3+JATP was supplied by ICN Radiochemicals. All other reagents were of the purest grade available from commercial sources.
RM&S In lasma membranes prepared from rat cerebella, 8 Ca exerted a biphasic effect on adenylate cyclase activity, which was absolutely dependent upon the
301
presence of CaM (Fig. la). Ca2’ ions, over the concentration range of 0.1 to 1 @I, stimulated the activity of adenylate cyclase by about threefold and began to inhibit the enzyme at concentrations that exceeded -10 @I. In the absence of CaM, Ca2’ elicited a slight stimulation of adenylate cyclase activity, followed by inhibition at concentrations > In contrast, the effect of these 10 p.M. . concentrations of Ca2+ ions on GHs plasma membrane adenylate cyclase was purely inhibitory, although this inhibition was also biphasic (Fig. lb). That is, Ca2’ inhibited adenylate cyclase activity by about 40% over the concentration range of 0.1 to 1 pIvI, such that a plateau between around 1 and 10 p.IvI Ca2’ preceded a further inhibition at It is significant supramicromolar concentrations. that this first inhibitory phase is elicited by concentrations of Ca2’ which correspond precisely to those that are achieved in intact cells upon depolarization or phospholipase C activation [27], and furthemrore, which elicit the CaM-dependent stimulation of cerebelhu adenylate cyclase described Extensive washes of GH3 pIasma above. membranes with EGTA (1mM)containing buffers and addition of a range of CaM concentrations (0.2 to 2 pM) failed either to attenuate or enhance the inhibition. Moreover, the addition of compounds known to antagonize CaM-mediated processes [24] - trifluopemzine (3 pM) and calmidazolium (1 pM) - did not abolish or attenuate the inhibition of enzyme activity induced by 1.7 @I Ca2’ (data not shown). Hill plot analysis of the inhibitory influence of Ca2+ over the 0.1 to 1 pM concentration range from four separate experiments revealed a highly cooperative mechanism (Fig. 2). This cooperative inhibition yielded a Hill coefficient (nB) of 2.21 + 0.47. Since the attenuation of adenylate cyclase activity by neurotransmitters and hormones requires the presence of GTE as an activating ligand of GTE-regulatory proteins, such as Gi, the Ca2+ inhibition of adenylate cyclase was examined as a function of increasing GTP concentrations. As shown in Figure 3. no inhibition was evident in the absence of GTP; as increasing concentrations of GTE were included, the inhibition of adenylate cyclase by 1.7 p.M Ca2’ was progressively revealed.
302
CELL CALCIUM
10-7
10-6 free
10-s
io-‘10-8
10-7
10-6
10-s
10 I4
free [W+] (M)
I=+1 (Ml
Fig. 1 Effects of Ca2’ and CaM on adenylate cyclasc activity in EGTA-washed
membranes prepamd from a) rat cerebellum and b)
GH3 cells. Adenylate cyclase activity was measured: a) in the absence of stimulating hormone; and b) in the presence of 0.1 p.M VIP. Concentrations of free Ca2’ wez~?determined by an iterative computer program, as described in Materials and Methods. Data ate from a representative experiment that was replicated at least four times and arc expressed as the mean activity f standard deviation of triplicate determinations
Fig. 2
Hill plot analysis
of the inhibition
by Ca2” in the
submicromolar range of GH3 adenylate cyclase activity. Data am from four indepcndcnt
experiments
Figure lb, each performed in triplicate
similar to that shown in
Since vasoactive intestinal peptide (VIP, 0.1 j.tM) was present throughout this experiment (Fig. 3), it was not possible to ascertain whether GTP itself or stimulation of the adenylate cyclase activity was responsible for enhancing the extent of Ca2’ Figure 4 demonstrates that in the inhibition. presence of 10 pM GTP, increasing VIP concentrations produced a concentration-dependent stimulation of adenylate cyclase activity. The addition of 117 p.M Ca2’ attenuated this stimulation such that its greatest inhibitory effect was apparent at near maximal stimulatory concentrations of VIP. The inhibitory effect of Ca2’ upon both GTP- and VIP-stimulated adenylate cyclase activity was saturable (Figs 3 and 4). The yfarent GTP- (or stimulation-) dependence of the Ca inhibitory effects on adenylate cyclase is analogous to the receptor-mediated inhibition of GHs adenylate cyclase by hormones (e.g. somatostatin) acting via Gi [28]. Consequently, we examined the effect of pretreating cultured cells with pertussis toxin upon the inhibition of adenylate cyclase activity by somatostatin and Ca2’ ions. Similar levels of VIP-stimulated adenylate cyclase
INHIBITION
-
OF ADENYLATE
500
1
CYCLASE
n
303
IN GH3 CELLS
SW
Control +1.7uMCa2+
a
0
d-4
Control +1.7pM ca2+
!~ n
400
: t
BY CALCIUM
4M1
I
F
d
300
300
,200
’ loo
f-lil
~~
I ..
PJW(4
[GW WI of Ca’+ inhibition
Fig. 3 Dependence activity
on GTP.
Activity
pM VIP at the indicated means
and standard
performed
of GH3 adenylate
was measured
GTP concentration.
deviations
from
cyclase
in the presence Valm
of 0.1 am the
three experiments,
each
+1.7y
4.1pu
Cal+
l1.7pyI
SST
c22*
*alrUsSr
Fig.
5
Differential
effects
of pertussis
hormone-mediated
inhibition
cells were exposed
to 125 ng/ml pertussis
medium
ove-rnighs
described
in Materials
were measured Data
experiments
activities
on Ca2+ and
Cultures of GH3
toxin in their growth
membranes
and Methods.
in the presence
conducted
toxin
of GH3 cyclase.
and plasma
are the mean
were prepared
Adenylate
as
cyclase activities
of 0.1 p.M VIP and 10 pM GTP. f standard
in triplicate.
toxin-treated,
hatched bars, untreated,
*P < 0.02.
by unpaired
Student’s
deviation from three Pertussis SST, somatostatin. opar bars t-test,
when
compared
to
t-test, when compared
to
Control. + 1.7 pM Ca2+ group **P < 0.001,
on VIP.
by unpaired
Control, + 0.1 @vf SST group
Student’s
on Ca2’ inhibition Activity
@vi GTP at the indicated means
and
perfolmed
in triplicate
NO 2ddltlon
Fig. 4 Dependence activity
standard
of G&
was measured
VIP concentmtions.
deviations
from
adenylate
in the presence two
Values experiments,
cyclase of 10 are the each
in triplicate
activity were displayed by plasma membranes derived from cells that had ben either untreated or exposed to pertussis toxin (Fig. 5). This activity inhibited by approximately 23% by WaS somatostatin (0.1 @I), and the inhibition was reduced to about 12% upon pertussis toxin pretreatment. On the other hand, a 38% inhibition elicited by Ca2’ ions was unaltered by treatment with the toxin. In combination9 Ca2’ and somatostatin elicited an additive inhibition of over 5596, of which only the hormone (somatostatin) component was attenuated by pertussis toxin. Since Harden et al. [29] had demonstrated previously (in intact 1321Nl astrocytoma cells) that Ca2’ ions could activate a phosphodiesterase activity to reduce cAh4P levels, the possibility was addressed that the present inhibitory effects of Ca2’ on [3?P]-cAMP formation was media&A by Ca2+-dependentstimulation of a phosphodiesterase activity (that had not been neutralizd by the standard inclusion of 0.5 mM CAMP in the assay). The inhibitory effects of Ca2’ on adenylate cyclase were identical, in either the presence or absence of high concentrations of phosphodiesterase inhibitors (500 p.M IsA@ and 100 pA4 Ro20-1724; data not
304
CELL,cALaJh4
Table 1 Effects of various regulatory agents on the Ca” inhibition of GH3 adenylate cyclase activity Assays were performed as described in Materials and Methods. with 1 mM MgC12 present in each assay. Data are expressed as the mean activity f standard deviation of at least three triplicate samples, from three separate experiments. GppNHp. guanosine S-(fl, y-imido) triphosphate; dd.A. 2’, S-didwxyadenosine Adenykate cyckase activity (pmoUmg.min) Control
- GTP, - VIP No addition +l
pMForskolin
+ 10 ph4GppNHp
+lOpMGTP,-VIP No actdition
+ 10 @l
+ I.7 pM ca2’
% Inhibition
10.7 f 1.1
12 f 6
117 f6.2
93.4 f 4.0
20 f 5
354 f 5.9
204f26
42 f 7
32.9 f 1.9
36f4
12.1 f 0.31
51.2 f 1.5
+2mMMncli!
376fll
338 f21
10f4
GTP, + 0.1 m VIP No addition
435 f 17
243 f 13
44f5
+lOOpMddA
219 f 9.2
137 f 6.3
37f4
Thus, it seems likely that the Ca2+ shown). inhibition is exerted specifically upon the adenylate cyclase activity, either by means of a Ca2+-binding protein, or by dire&y interacting with one of the components of the cyclase system In an attempt to identify the site of action at which Ca2+ inhibition was mediated, a variety of agents that modulate the adenylate cyclase system were examined for their influence on Ca2+ As described earlier (Pig. 3), the inhibition. dependence on GTP for this effect is evident (12% versus 36% inhibition by 1.7 pM Ca2+, in the absence and presence of GTP, respectively; Table I). Furthermore, the addition of VIP (with 10 pM GTP present) enhances the inhibitory effect to 44% (Table 1). The adenylate cyclase stimulatory agents, forskolin, GppNHp and MnCl2 alI stimulate GH3 adenylate cyclase activity to differing extents, but ;y I$C12;$$earst;leY ;ietiichE dideoxyadenosine, a F-site inhibitor of adenylate cyclase [30] inhibited cyclase by 5096, its presence did not preclude a further 37% inhibition by 1.7 pM Ca2’ (Table 1). Thus, it seems unlikely that the inhibitory effect of Ca2’ is mediated at the putative P-site.
Discussion In the present study, we examined the effects of Ca2’ in the concentration range that is reached intracellularly, following membrane depolarization or phospholipase C activation, upon the adenylate cyclase activity of GH3 plasma membranes. Ca2’ ions exerted a potent and cooperative, inhibitory effect on GH3 adenylate cyclase activity. These results are similar to those obtained in rat anterior pituitary membranes by Giannatasio et al. 1201, who described a biytasic inhibition of adenylate cyclase activity by Ca over concentration ranges identical to those used in the present report Also consistent with the present findings the Ca2’ inhibition of anterior pituitary adenylai cyclase did not require CaM and was not attenuated by anticahnodulin drums. These investigators examined the effect of Ca ’ on the stimulation by Mg2+ of adenylate cyclase; a kinetic analysis of their data afforded the interpretation that Ca2+ interacts competitively with an allosteric Mg2+ site on adenylate cyclase. This kinetic approach does not address the structural or mechanistic features of the site of Ca2+ inhibition on cyclase. This point is particularly relevant since Mg2+ plays many roles in the regulation of
INHIBITION OF ADENYLATE CYCLASE BY CALCIUM IN GH3 CELLS
adenylate cyclase - at the level of receptors, G-proteins and the catalytic unit of cyclase itself [3 11. The present paper sought to address the regulatory significance of Ca2+ inhibitory effects on cyclase, in terms of GTP-dependency, pertussis toxin-sensitivity, or sensitivity to agents known to influence the activity of the enzyme. In exploring the regulatory characteristics of the inhibition of GHa adenylate cyclase by Ca2+ ions, we discovered that inhibition is not observed in the absence of GTP, but is revealed as the concentration of the nucleotide is increased to about 1 r_LM.In addition, in the presence of 10 @vi GTE, significant Ca2+ inhibition is not apparent until VIP elicits a substantial stimulation of basal cyclase activity. From these results, it is not possible to determine whether GTP is specifically required, or whether there is both a GTP dependency and a requirement for VIP-stimulated states of adenylate cyclase. In GHa cells, hormones that act through G-proteins to inhibit adenylate cyclase, also require GTE and stimulated states of the enzyme [28]. In this context, the inhibition of adenylate cyclase by low analogous to Ca2+ concentrations appears hormone-mediated inhibition of cyclase. However, there are apparent differences in their mechanisms. For example, the magnitude of the inhibition by Ca2+ (35-45%) exceeds that elicited by inhibitory hormones. and Ca2+ inhibition is not sensitive to pretreatment of cells with pertussis toxin, which Moreover, renders Gi proteins dysfunctional. maximally effective concentrations of Ca2+ and somatostatin elicit au additive inhibition (> 55%) on adenylate cyclase activity, which indicates that these agents do not share common mechanisms of action. These data do not rule out the possibility that a G-protein that is not a pertussis toxin substrate may mediate Ca2’ effects on adenylate cyclase. The adenylate cyclase activity of GHs is inhibited by Ca2’ ion concentrations that are achieved intracellularly upon either membrane depolarization or activation of receptors coupled to the phosphatidylinositol signalling s stem [27]. In It brain, identical concentrations of Ca serve instead to elevate CAMP production about threefold, in an entirely CaM-dependent manner [7, 101. The effect of these low concentrations of Ca2+ on cyclase activity, whether stimulatory or inhibitory, provides
305
Ca2+-mobilizing and link between a CAME-mediated signal transduction systems 111, 20 . Especially noteworthy is the finding that while Ca1+ stimulation of neuronal adenylate cyclase depends absolutely upon CaM, equivalent Ca2+ concentrations in GHs cells act independently of CaM to reduce cyclase activity beyond that achieved The highly even by inhibitory hormones. cooperative nature of this Ca2+ inhibitory effect suggests that a Ca2+-binding protein which has adapted to this purpose may be involved. Nevertheless, since several CaM antagonists failed to affect the ability of Ca2+ to inhibit cyclase activity, the Ca2+ ion binding site may not be analogous to those found on proteins displaying E-helix-loop-F-helix structures 1341. In keeping with the present results, which suggest that [Ca2+]i in the physiological range can inhibit plasma membrane adenylate cyclase, studies in intact cells suggest that hormones or neurotransmltters that elevate [Ca2+]1 can decrease the production of CAMP. In some instances, at least part of this effect was due to the stimulation of a phosphodiesterase activity. Harden et al. [29] demonstrated that in 1321Nl astrocytoma cells, the phosphodiesterase inhibitor isobutylmethylxanthine reversed muscarinic cholinergic receptor-mediated inhibition of agonist-stimulated CAMP accumulation. However, as reported by others in intact NCB-20 cells 1351, as weIl as in the present study, the inclusion of high concentrations of phosphodiesterase inhibitors did not perturb Ca2’ inhibition of adenylate cyclase. These data suggest that Ca2+ can act directly on a component of the adenylate cyclase system Although we interpret the present data to indicate direct effects of Ca2+ on adenylate cyclase activity, the possibility could be entertamed that given the GTP and VIP enhancement of the inhibitory effects - phospholipase C was being stimulated in these preparations. This could result in the liberation of diacylglycerol and activation of endogenous protein kinase C, which might then mediate the inhibition in a Ca2+-dependent manner. This possibility seems highly unlikely for a number of reasons: i) this plasma membrane preparation from a sucrose gradient is not likely to contain significant amounts of protein kinase C; and ii) the
306
effect is seen directly in a plasma membrane preparation, whereas the majority of reports of protein kinase C modulating adenylate cyclase activity rely on a pretreatment of cells with protein kinase C stimulants, prior to the preparation of a membrane fraction [see e.g. 11-141. Indeed in another cell line (NCB-20) in which we have encountered similar actions of Ca’+, neither inhibitors of protein kinase C, such as sphingosine, nor treatment of the membranes with protein kinase C and diacylglycerol, mod&d the inhibitory effects of Ca2’ (Boyajian and Cooper, in preparation). In contrast to the present fiidings and those of Giannatasio et al. [20], Greenlee and Okada [ 181 reported a modest (-20%) activation of adenylatc cyclase by Ca2+ in a particulate preparation from rat anterior pituitary, which was dependent upon the presence of supramicromolar concentrations of CaM. Relatively crude preparations utilized in the latter studies, unlike the purified plasma membrane preparations used in the present case, may have permitted stimulatory effects to be mediated by soluble factors loosely associated with the plasma There is some controversy as to membrane. whether Ca2’/CaM stimulation of adenylate cyclase occurs outside of the central nervous system [131. Several methodological factors may underlie misleading inferences of Ca2+ or Ca2+/CaM stimulatory effects on adenylate cyclase activity. ‘disinhibition’ by include possible These supramicromolar Cah4 of adenylate cyclase activity due to chelation of Ca2’ ions activation of Ca2+-dependent protease activities’ that stimulate adenylate cyclase [32, 333, or activation of Ca2+- or CM-dependent protein kinase in untractionated membrane preparations. The present inhibition of adenylate cyclase activity by Ca2’ ions is saturable and of sufficient magnitude to suggest that Ca2’ signalling systems may play a significant role in the modulation of the activity of the cAA@generating system, and its dependent physiological processes. For instance, both protein kinase A, via phosphorylation mechanisms [36], and Gs of the adenylate cyclase systems [37, 381, can stimulate the activity of voltage-gated Ca2+ channels in GHs cells. Resulting elevations in intracellular Ca2’ could provide feedback inhibitory control of such CAMP-
CELL CALCIUM
or G,facilitating actions. In support of this view, Dufy and colleagues 1391have noted previously that the elevation of intracellular Ca2+ via voltage-gated channels appears to result in a feedback inhibition of the activity of the channel. It is interesting that very recently a similar observation has been made for another GTP regulatory protein-mediated signal transduction system, i.e. the retinal guanylate c yclase enzyme 1401. In this instance, submicromolar concentrations of Ca2’ inhibited enzyme activity independently of CaM in a highly cooperative manner (nH = 3.9), by about 40%. In addition, Ca2’ has recently been reported to inhibit adenylate cyclase in MA-10 Leydig cells [41]. Thus, it may transpire that the potent inhibition of adenylate cyclase activity by Ca2’ ions observed presently reflects a widespread intracellular Ca2+-sensing device for the feedback inhibition of elevated intracellular Ca2+ resulting in turn in the modulation of various celhtlar signal transduction systems. Acknowledgements The authors thank Ms Lusnne Esle-r for maintaining GH3 cells in culture, and Ms Kathy Fassler for expert secretsrid assistance. These studies were supported by NH-IgrantGM 32483.
References 1. Rasmussen H. (1980) calcium and CAMP in stimuhu response coupling. Ann. NY Acad. Sci., 356,346-353. 2. Meldolesi J. Possan T. (1987) Pathways of Ca2’ influx at the plasma membrane: voltage-, receptor-, and second messenger-operated channels. Exp. Cell Res., 17.271-283. 3. Benidgc UT. Galone A. (1988) Cytosolic calcium oscillators. FASEB J., 2, 3074-3082. 4. Kater SB. Mattson MP. Cohau C. Connor J. (1988) Calcium regulation of the neuronal growth core. Trends Neurosci., 11,315-321. 5. Smitb SJ. Augustine GJ. (1988) Calcium ions, active zones and synaptic transmitter release. Trends Neurosci.. 11, 458-464. 6. Tomlinson S. MacNeil S. Brown BL. (1985) Calcium, cyclic AMP and hormone action. Clin. Endocrinol.. 23. 595-610. 7. Cooper DMF. Ahlijanisn MK. Perez-Reyes E. (1988) Calmodulin plays a dominant role in determining neumtmnsmitter regulation of neuronal ade-nylatecyclaae. J. Celh11.B&hem.. 36,417-427. 8. Nahorski SR. (1988) Inositol polyphosphates and neuronal calcium homeostasis. Trends Neumsci.. 11.444-448. 9. Eberhard DA. Halz RW. (1988) Intracellular Ca2+ activates phospholipase C. Trends Neurosci., 11,5 17-520. 10. Brostrom Co. Huang Y. Bceckemidge BMcL. WolfTDL.
INHIBITION OF ADJZNYLATE CYCLASE BY CALCIUM IN GH3 CELLS
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26. 27.
(1975) Idemitication of a calcium-binding protein as a calcium-dependent regulator of brain adenylate cyclase. Proc. Natl. Acad. Sci. USA, 72,646s. Cooper DMF. Caldwell KK. Pemz-Reyes E. Ahlijanian MK. Schlegel W. (1988) Interactions between neutottansmitters that regulate CAMP and intracellular Ca” levels in the CNS. In: eds. Kito S. Segawa T. Kuriyama K. Tohyama M. Olsen RW. Neuromceptors and Signal Transdix%ion. Advances in Experimental Medicine and Biology, vol. 236, pp. 217-227. MacNeil S. Lakey T. Tomlinson S. (1985) Cahnodulin regulation of adenylate cyclase activity. Cell Calcium, 6, 213-226. Cooper DMF. Caldwell KK. (1988) Does cahnodulin regulate non-neummtl adenylate cyclam? B&hem. J., 254, 935-936. Summers ST. Cronin MJ. (1988) Phorbol esters induce two distinct changes in GH3 pituitary cell adenylate cyclase activity. Arch. Biochem. Biophya., 262, 12-18. Norstedt C. Jondal M. Fredholm B. (1987) Activation of protein kinase C inhibits prcwtaglandin - and potentiates adenosine receptor-stimulated accumulation of cyclic AMP in a human T-cell leukemia lioe. FEBS L&t., 220,57&O. Wiener E. Scatpa A. (1989) Activation of protein kinase C modulates the adenylate cyclase effector system of B-lymphocytes. J. Biol. Chem., 264, 43244328. Bell JD. Brunton LL. (1987) Multiple effects of phorbol esters on hormone-sensitive adenylate cyclase activity in S49 lymphoma cells. Am. J. Physiol., 15, E783E789. Greenlee DV. Okada S. (1987) Calmodulin activates adenylate cyclase from rat anterior pituitary. Mol. Pharmacol., 32.743-748. Brostrom MA. Btotman LA. Brostrom Co. (1982) Calcium-dependent adenylate cyclase of pituitary tumor cells. B&him. Biophys. Acta, 721.227-235. Giannatasio G. Bianchi R. Spada A. Vallar L. (1987) Effect of calcium on adenylate cyclsae of rat anterior pituitary gland. Endocrinology, 120,2611-2619. Tashjian AH. Jr. (1979) Clonal stmins of hormone-producing pituitary cells. Methods Enzymol.., 58, 527-535. Caldwell KK. Newell MK. Cambier JC. et al. (1988) Evaluation of methods for the isolation of phtsma membranes displaying guanosine 5’-hiphosphate-dependence for the regulation of adenylate cyclase activity: Potential application to the study other guanosine S’-triphosphate-dependent transduction systems. Anal. Biochem., 175,177-190. Lowry OH. Rosebtough NJ. Farr AL. Randall RJ. (1951) Protein measumment with the Folin phenol reagent. J. Biol. Chem., 193,265-275. Ahlijanian MK. Cooper DMF. (1987) Antagonism of cahnodulin-stimulated adenylate cyclase of trifluoperazine, cahnidazolium and W-7 in rat cembellar membranes. J. Phannacol. Exp. Ther., 241, 407-414. Salomon Y. Londos C. Rodbell M. (1974) A highly sensitive adenylate cyclase assay. Anal. Biochem., 58, 541-548. Matte.11AE. Smith RM. (1975) Critical Stability Constants. New York, Plenum Press, vol. 2. pp. 269-284. Tsien RY. Pozzan T. Rink TJ. (1984) Measuring and manipulating cytosolic Ca’+ with trapped indicators. Trends
307
Bionhem. Sci.. 9.263-266. 28. Cooper DMF. Caldwell KK. Boyajian CL. Petcoff DW. Schlegel W. (1989) Adenosine Al receptors inhibit both adenylate cychise activity and TRH-activated Ca2+channels by a pertussis toxin-sensitive mechanisms in GHz cells. Celhtl. Signal., 1.85-97. 29. Harden TK. Evans T. Hepler JR. et al. (1985) Regulation of cyclic AMP metabolism by muscarinic cholinergic receptors. In: eds, Cooper DMF. Seamon KB. Advances in cyclic micleotide and protein phosphorylation research. New York, Raven Press, voL 19, pp, 207-220. 30. Londos C. Wolff J. (1977) Two distinct a&no&z-sensitive sites on adenylate cycbtse. Proc. Natl. Aced. Sci. USA, 77,2551-2554. 31. Iyengar R Bimbaumer L. (1981) Hysteretic activation of adenylyl cyclase. I. Effects of Mg ion on the rate of activation of guanine nucleotides. J. Biol. Chem., 256, 11036-11041. 32. Adnot S. Poirier-Dupuis M. Franks DJ. Hamet P. (1982) Stimulation of rat platelet adenylate cyclase by an endogenous calcium-dependent pmtease-like activity. J. Cyclic Nucleotide Res.. 8, 103-l 18. 33. Wallace MA. Fain JN. (1985) Guanosine 5’-G-thiotriphosphate stimulates phospholipase C activity in plasma membranes of tat hepatocytes. J. Biol. Chem., 260, 9527-9530. 34. Cox JA. (1988) Interactive properties of calmodulin. B&hem. J., 249,621-629. 35. Hollingsworth EB. Daly JW. (1987) Inhibition of receptor-mediated stimulation of cyclic AMP accumulation in neuroblastoma-hybrid NCB-20 cells by a phorbol ester. Biocbim. Biophys. Acta, 930,272-278. 36. Armstrong D. Eckert R (1987) Voltage-activated calcium channels that must be phosphotylated to respond to membrane depolarization. Proc. Natl. Acad. Sci. USA, 84, 2518-2522. 37. Hedlund B. Dufy B. Bat& JL. (1988) Vasoactive intestinal polypeptide alters G&/B6 pituitary cell excitability. Pfluger’s Arch., 411. 173-179. 38. Bjoro T. Ostberg BC. Sand 0. et al. (1987) Vasoactive intestinal peptide and peptide with N-terminal hi&line and C-terminal isoleucine increase prolactin secretion in cultured rat pituitary cells (GI-LtCt)via a CAMP-dependent mechanism which involves transient elevation of intracellular Ca2+. Mol. Cellul. Bndocrinol., 49, 119-128. 39. Dufy B. Jaken S. Barker JL. (1987) Intraoellular Ca’+-dependent protein kinase C activation mimics delayed effects of thyrotropin-releasing hormone on clonal pituitary cell excitability. Endocrinology, 121,793~802. 40. Koch K-W. Stryer L. (1988) Highly cooperative feedback control of retinal rod guanylate cyclase by calcium ions. Nature, 334,64-66. 41. Pereira MB. Segaloff DL. Ascoli M. (1988) Ca2+ is an inhibitor of adenylate cyclase in MA-10 Leydig tumor cells. Endocrinology, 122,2232-2239. Please send reprint requests to : Dr D.M.F. Cooper, Department of Pharmacology, University of Colorado Health SciCenter, 4200 E. Ninth Avenue. Denver, CG 80262, USA Received : 9 October 1989 Revised : 7 November 1989 Accepted r 22 November 1989
