The Role of Cyclic GMP in the Regulation of Cyclic AMP Hydrolysis W. 1. Terasaki
and M. M. Appleman
A rat-heart cyclic nucleotide phosphodiesterase has been chromatographically separated from related enzymes and its kinetic properties have been studied. The enzyme can hydrolyze both cyclicAMP and cyclicGMP and has about the same maximum velocity and apparent K, (groater than lo-” M) for the two nucleotides. Kinetic plots indicate positive coopemtive
behavior for both substmtes. Cyclic GMP at low concentmtions is a potent activator of cyclic AMP hydrolysis and this activation, as well as the coopomtivity, can be abolished by treatment with solvents or sulfydryl reagents under conditions which do not destroy the catalytic function. A kinetic model for this enzyme is proposed and the physjologic role is discussed.
INCE ITS DISCOVERY by Earl Sutherland and Theodore Rall, cyclic AMP has held a place which is unique in the realm of biologic regulation, and a wealth of information has accumulated on the involvement of this nucieotide in cellular function. Cyclic GMP, shown to occur naturally in 1963,’ and subsequently found to be as ubiquitous as its predecessor in mammalian systems 2v3has been much more difficult to assign to any clearly defined functional roles. Recent findings have suggested that cyclic GMP may play a role in controlling the degradation of cyclic AMP by cyclic nucleotide phosphodiesterase. Since the chemical structures of the two naturally occurring cyclic nucleotides are similar, cyclic GMP might be expected to compete against cyclic AMP for catalytic binding sites on phosphodiesterases and thereby inhibit cyclic AMP degradation by these enzymes. Yet Be&o et al.4 first reported the interesting phenomenon that cyclic GMP in micromolar quantities can stimulate the hydrolysis of cyclic AMP in extracts of various rat tissues. Apparently there existed a phosphodiesterase which is activated by cyclic GMP. Since the activating effects were seen only when the substrate concentration was low (below 50 pIM), it was suggested that cyclic GMP might be acting to increase the affinity of a high-K, phosphodiesterase. Soon afterwards, other workers saw similar stimulatory effects on phosphodiesterases from thymic lymphocytes,’ fibroblasts,6 and adipose tissue.’ Since cyclic GMP is effective at concentrations below lo-’ M, it appears feasible that this represents a physiologically important means for the regulation of cyclic AMP levels. No stimulatory effect of cyclic AMP on cyclic GMP hydrolysis has been reported. Lately, it has been suggested that cyclic GMP stimulates a cyclic nucleotide phosphodiesterase which displays positive cooperative kinetic behavior!*g In
S
From the Department of Biological Sciences, University of Southern California. Los Angeles, Calif Received for publication October II, 1974. Supported by the Diabetes Association of Southern California and NIH Biomedical Sciences Support Grant RR-07012-07. Reprint requests should be addressed to M. M. Appleman. Department of Biological Sciences. University of Southern California. Los Angeles, Calif 90007. 0 I975 by Grune & Stratton. Inc. Metobolirm, Vol. 24, No. 3 (March), 1975
311
312
TERASAKI
AND APPLEMAN
the present study, some of the catalytic and regulatory properties losteric form of phosphodiesterase from rat heart will be reported. MATERIALS
of this al-
AND METHODS
The cyclic AMP and cyclic GMP phosphodiesterase assays of Russell et al.,s with minor modifications, have been employed. Incubation mixtures contained appropriately diluted enzyme, 20 mM MOPS-NaOH buffer (pH 7.0). 0.1 M sodium acetate, 1 mM 2-mercaptoethanol, 5 mM magnesium chloride, approximately 150,000 dpm tritiated cyclic nucleotide, and varying amounts of unlabeled cyclic nucleotide. The cyclic GMP-sensitive cyclic nucleotide phosphodiesterase (cNMP-PDE) was prepared from rat heart using a DEAE-cellulose chromatography system similar to that reported by Russell et a1.s for rat liver. Three distinct active fractions were eluted from the columns (Fig. I): Fraction I (cGMP-PDE) has a high affinity for cyclic GMP (K,, 2 x IOm6 M) and is regulatable by a calcium-dependent protein modulating factor. Fraction II (cNMP-PDE), the cyclic GMP-sensitive phosphodiesterase, displays no catalytic preference between cyclic GMP and cyclic AMP at high substrate concentrations. Fraction III (CAMP-PDE) is highly specific for cyclic AMP and gives evidence of negative cooperative regulation. The stability and identity of these enzyme forms were confirmed by the identical elution position and kinetic properties obtained upon rechromatography of the separated fractions (Fig. 2). The work in this paper was carried out solely on doubly chromatographied cNMP-PDE (Fraction II) from rat heart.
RESULTS
Kinetic Properties When represented on a Hill plot, cyclic AMP hydrolysis by cyclic nucleotide phosphodiesterase is cooperative (Hill coefficient, n = 1.4-1.8) with an apparent
80
20 FRACTION
30 NUMBER
40
50
Fig. 1. DEAE-cellulose profile of cyclic nuclaotide phosphodiestemse from rat heart. Sonicated extract was applied to a I .5 x 30-cm column of DEAE-cellulose which was developed with a linear 400 ml gradient from 0.07 to 1.O M sodium acetate (pH 6.5) containing 5 mM 2-mercaptoothanol. Fractions (8.0 ml) were collected at a Row rate of 30 ml/hr. Aliquots (200 Al) won measured directly for enzymatic activity using 1 pM cyclic AMP (A), 1pM cyclic GMP (o), or 100 PM cyclic AMP (A). The major peaks have been labeled fractions I, II, and Ill in order of their elution from the column according to Russell et al.’
CYCLIC GMP IN REGULATION
313
OF CYCLIC AMP HYDROLYSIS
0.8 0.4 0
Fig. 2. Rechromatography of mtphosphoheart cyclic nucleotide diestemses on DEAE-cellulose columns. Isolated fmctions from two DEAE-cellulose columns (as in Fig. 1) were dialyzed to remove the salt and individually rechromatogmphed. Activity was assayed using 1 MM cyclicAMP (o) or cyclic GMP (e). The profiles represent fmction I (upper panel), ffaction II (center panel), and fraction Ill (lower panel).
4
30
n
20
0.8
E E 4
IO
0.4
0
I
1
1
0
FRACTION NUMBER
K, of 6 x 1O-5 M. A plot of cyclic GMP hydrolysis by the enzyme appears as a biphasic curve which bends at a substrate concentration of 10e6 M (Fig. 3). The Hill coefficient in the range of concentrations below this value is nearly identical to the slope of the curve for cyclic AMP hydrolysis for the same enzyme preparation (n = 1.44). At higher concentrations of cyclic GMP, it approaches Michaelis-Mentin kinetics (n = 1.16). The K, for cyclic GMP is 2 x lo-’ M and the V,,, values for the two substrates are similar. Allosteric Activation Low concentrations to cyclic GMP have a pronounced effect on the nature of cyclic AMP hydrolysis only by this form of the phosphodiesterase.‘O Micromolar cyclic GMP stimulates the enzyme only at low concentrations of cyclic AMP substrate, while at nearly saturating substrate cyclic GMP has no effect. In the presence of cyclic GMP, the enzyme loses its cooperativity (the Hill coefficient decreased to 1.O) without any appreciable change in the V,,, or extrapolated K, values for cyclic AMP. Cyclic GMP appears to act as an allosteric activator has a classic K-type cooperative system.” The moderate decrease in the Hill coefficient (from an average of 1.6 to 1.O) which accompanies the loss of cooperativity represents a substantial increase (up to tenfold) in the hydrolysis of 1 PM cyclic AMP. The activator-velocity curves for cyclic GMP are noncooperative, with half-
TERASAKI
314
I.C
AND
APPLEMAN
r
0
n >
I * z > \ >
Fig. 3. Hill plot of cyclic nucleotide hydrolysis by rat-heart phosphodiesterare. Enzyme activity was measured in the range from 100 to 0.1 PM cyclic GMP. The curve for cyclic GMP activity (e) is biphasic, and the dashed line is the best fit for the values below a substrate concentmtion (5) of 10e6 M. For comparison, cyclic AMP activity for the same enzyme preparation is also presented (a). The Hill coefflcients (n) am the sloper of the lines as they are drawn here.
-1.0
u
z -2.c I
I
-7.0
t
I
-6.0
1
-5.0
-4.0
log (S)
maximal effects generally occurring at about 3 x lO-7 M (Fig. 4). Concentrations of 5’ GMP as high as 50 PLM are unable to mimic cyclic GMP, and guanosine (0.5 mM) inhibits activation by cyclic GMP by about 50% without changing the basal (nonstimulated) catalysis of cyclic AMP. While the catalytic activity of phosphodiesterase is strongly dependent on the presence of magnesium ion, the degree of activation by cyclic GMP is not diminished by concentration reductions in this ion from 5.0 to 0.3 mM.
Fig. 4. Activator velocity curve for the effect of cyclic GMP on cyclic AMP hydrolysis. The concentmtion of cyclic AMP was 1 FM. The inset is a double reciprocal representation of the same data (appomnt K, of 3 x 10-7 M cyclic GMP). Velocity (V) is expressed as picomoles cyclic AMP hydrolyzed per minute per micmgmm enzyme protein, and V, is the basal activity measured in the absence of cyclicGMP.
V
I 0
I 0.2
I (
Cyclic
I 0.4
I
I 0.6
GMP ) , pM
I
I 0.8
_
CYCLIC GMP IN REGULATION
OF CYCLIC AMP HYDROLYSIS
315
I
I
I
I
I
-
0
LO/ \\ \ \\
l
0.2tlM
Cyclic GMP
\
0.8 - ‘y \ \
v VO
h
\
0.6 -
0
\
\
IO&
Fig. 5. Effect of cyclic AMP on cyclic GMP hydrolysis. Cyclic GMP hydrolysis was measured at substrata concentrations of 0.2 phi ( l) or 10 WM (0) in the presence of varying concentrations of cyclic AMP. Enzyme activity is expressed as the measuwd velocity in the presence of cyclic AMP (V) with respect to the basal activity in the absence of cyclicAMP (V,).
Cyclic 0
l 4 0
\-
P GMP ‘\
I
I
I
IO
20
30
(
0 .
.
-2-0
40
1
_
50
Cyclic AMP ),pM
Activating Eflect of Cyclic AMP It has been reported previously4 that cyclic AMP has no significant activating effect on cyclic GMP degradation by rat liver phosphodiesterase. This is true at 10 PM substrate, where cyclic AMP was strictly an inhibitor of cyclic GMP hydrolysis (Fig. 5). On the other hand, when the cyclic GMP substrate concentration is very low (0.2 PM), cyclic AMP has a small stimulatory effect. In addition, cyclic AMP is a more potent inhibitor of the enzyme when the substrate is high (IO PM) than when it is much lower (0.2 PM). Inhibition The hydrolysis of each cyclic nucleotide is inhibited by high concentrations of the other. Cyclic GMP inhibits the hydrolysis of cyclic AMP in a competitive fashion with a Ki of 26 PM. Cyclic AMP inhibits cyclic GMP hydrolysis also competitively with a K, of 24 PM. These inhibition constants are roughly similar to the respective Michaelis constants. Enzyme Desensitization Treatment of the phosphodiesterase with organic solvents (10% ethanol or acetone) destroys cooperativity with cyclic AMP as the substrate without appreciably decreasing the V,,, (Fig. 6). In addition, the enzyme loses its ability to become activated by cyclic GMP. Incubation of the enzyme with a low amount of the sulphydryl reagent, Nethylmaleimide (NEM, 0.01 moles/g protein), increases basal cyclic AMP hydrolysis, although the maximum attainable activity (cyclic GMP-stimulated) remains unchanged (Fig. 7A). At higher concentrations of NEM, the basal activity, as well as the cyclic GMP-stimulated activity, decreases. The cNMP phosphodiesterase is also fully desensitized by p-chloromercuribenzoate, another sulfydryl inhibitor. The removal of 2-mercaptoethanol by dialysis ac-
TERASAKI
316
AND APPLEMAN
0.6 -
0.4 -
.’
/ L
0.2 -
/
Control
L_JF:T, ,, 0
0.04
0.00
( Cyclic AMP?,
0.120.16
pM_’
Fig. 6. Effects of organic solvents on cocyclic opemtivity of AMP hydrolysis. Phosphodiestemse activity was measured in the pmsence of 10% acetone (II), 10% ethanol (A), or no addition (e) as a control. Velocity (V) is in picomoles cyclic AMP hydrolyzed per minuk per microgmm protein. The data am represented on a Lineweaver-Burk plot.
tivates and desensitizes the enzyme in a manner analogous to NEM (Fig. 7B). It is apparent that when the enzyme is maximally stimulated by cyclic GMP, it cannot be further activated by modification of sulfydryl groups. Relative Hydrolytic
Rates
Table 1 compares the relative catalytic rates of cyclic nucleotide phosphodiesterase in the presence of high and low substrate and allosteric effector. The activities exhibited by the enzyme are strikingly dependent on substrate type and concentration. At low substrate concentration (1 PM) the enzyme hydrolyzes cyclic GMP about five times faster than cyclic AMP. If the substrate concentration is increased, the rates for the two nucleotides become nearly equal, and, in fact, V,,, values indicate that cyclic AMP is the preferred substrate at very high concentrations. Under saturating conditions, degradation of cyclic GMP is SO%-90% of that for cyclic AMP. When cyclic AMP hydrolysis (at 1 PM substrate) is fully activated by cyclic GMP, the magnitude of the stimulation was about sixfold, and the ratio of hydrolysis of cyclic GMP to that of cyclic AMP activity becomes nearly identical to the ratio of their respective maximum velocities (Table 1). Like cyclic GMP, NEM also causes cyclic AMP and cyclic GMP hydrolytic rates to become nearly equal.
CYCLIC
GMP
IN
REGULATION
OF
CYCLIC
AMP
HYDROLYSIS
317
r
+cGMP
+cGMP
0
5 NEM
( moles/g
protein
E-MERCAPTOETHANOL
)
( mM 1
Fig. 7. Effect on N-ethylmaleimide or removal of 2-mercaptoethanol on cyclicGMP activation of mt-heart fraction II. (A) Cyclic nucleotide phosphodiestemse was treated with 0, 0.01, or 0.04 moles NEM per gmm enzyme protein, and then cyclic AMP hydrolysis was measured at 1 pM substmte in the presence of 0 (open ban) or 5 phi (striped bars) cyclic GMP. (B) Doubly chromatogmphed phosphodiestemse was dialyzed overnight against 0.1 M Tris-HCI (pH 7.4) convaining no added 2-mercaptoethanol. After dialysis, cyclic AMP hydrolysis was measured at 1 PM substmte in the presence (closed bars) or absence (open bars) or 1 pM cyclic GMP. The same prepamtion maintained overnight in 5 mM 2-mercaptoethanol sewed as a control.
Table 1. Relative Activities of Rat-heart Cyclic Nucleotide Phosphodiesterase Towards Cyclic GMP and Cyclic AMP Substrate Substrates Compared GMP/cyclic
AMP
at “ma,
0.8
f
0.1
(2)
Cyclic
GMP/cyclic
AMP
lOOBAt
1.2 f
0.2
(6)
Cyclic
GMP/cyclic
AMP
1 PM
4.7
f
1.5
(6)
Cyclic
GMP/cyclic
AMP
(cyclic
TpM
0.9
f
0.1
(4)
Cyclic
AMP
GMP
stimulated)/cyclic
5.9
f
2.0
(8)
shown
preparations
(cyclic are
the
means
in parenthesis.
in addition
double
Ratio (Number)
Cyclic
Ratios
tide
Concentration
reciprocal
to
the
kinetic
cyclic
GMP
f
“Cyclic
stimulated) AMP
standard GMP
AMP-labeled
plots in the range
1 IrM
deviation
stimulated” substrate. of from
with indicates
V,,,
6 to 200
the the
values PM
number presence are
substrate.
of
separately of
obtained
1 pM by
tested unlabeled
extrapolation
enzyme nucleofrom
TERASAKI
318
AND APPLEMAN
DISCUSSION
The data reported here, in addition to that reported earlier, indicate that the cyclic nucleotide phosphodiesterase (fraction II) from rat heart is much more complex an enzyme than would be inferred from its previous title of “low affinity-cyclic AMP-phosphodiesterase.” The enzyme does have a relatively high K, for cyclic AMP but hydrolyzes cyclic GMP with about the same substrate affinity, and each nucleotide is a competitive inhibitor of the other’s hydrolysis. There is good evidence for positive cooperative kinetic behavior for both substrates, but this is most demonstrable at low concentrations for cyclic GMP. Cyclic GMP can activate cyclic AMP hydrolysis; the reverse appears also to be true but is demonstrable only over a very narrow concentration range for both nucleotides. The activation of cyclic AMP hydrolysis by cyclic GMP is due to a loss of cooperativity rather than a change in the K, or maximal velocity. Similar effects can be achieved by treatment with dilute solvents, detergents, sulfydryl reagents, or by the removal of protective sulfydryl compounds. A model, derived from this data, might be considered as involving just two types of cyclic nucleotide binding sites. One would be a catalytic site, with about equal (and low) affinity for cyclic AMP and cyclic GMP in its nonactivated state and with about equal (but high) affinity for the two nucleotides in its activated state. The second type of site would be regulatory; it also could bind both cyclic nucleotides. If the regulatory site is unoccupied, the catalytic site is in its low-affinity state; if the regulatory site is occupied (or if the regulator component is functionally dissociated by chemical treatment), the catalytic site goes to its high-affinity state. The esoteric behavior of the enzyme is then a result of the much greater affinity of the regulatory site for cyclic GMP than for cyclic AMP. Confirmation of this model and of the probable existence of subunits to carry out the allosteric interactions must await further enzyme purification. The physiologic role of this complex phosphodiesterase regulatory system is difficult to assess. The cooperative nature of cyclic AMP hydrolysis could be effective only at very high concentrations of this nucleotide (greater than 10m5 M) such as might occur during a regulatory crisis. This interpretation could be supported by the apparent cytoplasmic location of the enzyme and its very high maximal velocity. As a cyclic GMP phosphodiesterase, the enzyme also leaves much to be desired; the substrate affinity is two to three orders of magnitude below reported levels of this nucleotide, so that even with the cooperative kinetic behavior it could not be very effective. Probably the most appropriate classification of the enzyme is as a cyclic GMP-activated cyclic AMP phosphodiesterase. The lOO-fold greater affinity for cyclic GMP as a regulator when compared to cyclic AMP as a substrate would serve to bring these two functions into line with the reported cyclic nucleotide levels. Concentrations of cyclic GMP as low as 4 x lo-* M will give significant stimulation of cyclic AMP hydrolysis at a 10e6 M substrate concentration. It is then quite possible for cyclic GMP to serve as a regulator of cyclic AMP levels in certain circumstances. This would correlate well with the “Yin-Yang” or dualism theory of Goldberg,3 in which the two nucleotides are viewed as having antagonistic roles in biologic regulation. The question might even be asked whether this
CYCLIC GMP IN REGULATION
OF CYCLIC AMP HYDROLYSIS
319
cyclic GMP regulation of cyclic AMP hydrolysis represents just another example of “Ying-Y ang” phenomena or if it might actually be the mechanism for a number of reported antagonistic actions. REFERENCES 1. Ashman DF, Lipton R, Melicow MM, Price TD: Isolation of adenosine 3’,5’-monophosphate and guanosine 3’,5’-monophosphate from rat urine. Biochem Biophys Res Commun 11:330-334, 1963 2. Robison GA, Butcher RG, Sutherland EW: Cyclic AMP. New York, Academic Press, 1971 3. Goldberg ND, O’Dea RF, Haddox MK: Cyclic GMP, in Advances in Cyclic Nucleotide Research, Vol. 3. New York, Raven Press, 1973, pp 115-223 4. Beavo JA, Hardman JG, Sutherland EW: Stimulation of adenosine 3’,5’-monophosphate hydrolysis by guanosine 3’,5’-monophosphate. J Biol Chem 246:3841-3846.1971 5. Franks DJ, MacManus JP: Cyclic stimulation and inhibition of cyclic AMP phodiesterase from thymic lymphocytes. them Biophys Res Commun 42:844849,
GMP phosBio1971
6. Manganiello V, Vaughan PM: Prostaglandin Et effects on adenosine 3’,5’-cyclic
monophosphate concentration and phosphodiesterase activity in fibroblasts. Proc Nat1 Acad Sci USA 69:269-273, 1972 7. Klotz U, Stock K: Influence of cyclic guanosine-3’,5’-monophosphate on the enzymatic hydrolysis of adenosine-3’,5’-monophosphate. Naunyn Schmiedebergs Arch Pharmakol 27454-62, 1972 8. Russell TR, Terasaki WL, Appleman MM: Separate phosphodiesterases for the hydrolysis of cyclic AMP and cyclic GMP in rat liver. J Biol Chem 248: 1334-l 340, 1973 9. Bevers MM, Smits RAE, Van Riju J, Van Wyk R: Heat treatment of rat liver adenosine 3’,5’-monophosphate phosphodiesterase. Biochim Biophys Acta 341:120-128.1974 10. Appleman MM, Terasaki WL: The regulation of cyclic nucleotide phosphodiesterase, in Advances in Cyclic Nucleotide Research, vol. 5. New York, Raven Press, (in press) 11. Monod J, Wyman J, Changeux JP: Qn the nature of allosteric transitions: A plausible model. J Mol Biol 12:88-l 18, 1965
and M. M. Appleman
A rat-heart cyclic nucleotide phosphodiesterase has been chromatographically separated from related enzymes and its kinetic properties have been studied. The enzyme can hydrolyze both cyclicAMP and cyclicGMP and has about the same maximum velocity and apparent K, (groater than lo-” M) for the two nucleotides. Kinetic plots indicate positive coopemtive
behavior for both substmtes. Cyclic GMP at low concentmtions is a potent activator of cyclic AMP hydrolysis and this activation, as well as the coopomtivity, can be abolished by treatment with solvents or sulfydryl reagents under conditions which do not destroy the catalytic function. A kinetic model for this enzyme is proposed and the physjologic role is discussed.
INCE ITS DISCOVERY by Earl Sutherland and Theodore Rall, cyclic AMP has held a place which is unique in the realm of biologic regulation, and a wealth of information has accumulated on the involvement of this nucieotide in cellular function. Cyclic GMP, shown to occur naturally in 1963,’ and subsequently found to be as ubiquitous as its predecessor in mammalian systems 2v3has been much more difficult to assign to any clearly defined functional roles. Recent findings have suggested that cyclic GMP may play a role in controlling the degradation of cyclic AMP by cyclic nucleotide phosphodiesterase. Since the chemical structures of the two naturally occurring cyclic nucleotides are similar, cyclic GMP might be expected to compete against cyclic AMP for catalytic binding sites on phosphodiesterases and thereby inhibit cyclic AMP degradation by these enzymes. Yet Be&o et al.4 first reported the interesting phenomenon that cyclic GMP in micromolar quantities can stimulate the hydrolysis of cyclic AMP in extracts of various rat tissues. Apparently there existed a phosphodiesterase which is activated by cyclic GMP. Since the activating effects were seen only when the substrate concentration was low (below 50 pIM), it was suggested that cyclic GMP might be acting to increase the affinity of a high-K, phosphodiesterase. Soon afterwards, other workers saw similar stimulatory effects on phosphodiesterases from thymic lymphocytes,’ fibroblasts,6 and adipose tissue.’ Since cyclic GMP is effective at concentrations below lo-’ M, it appears feasible that this represents a physiologically important means for the regulation of cyclic AMP levels. No stimulatory effect of cyclic AMP on cyclic GMP hydrolysis has been reported. Lately, it has been suggested that cyclic GMP stimulates a cyclic nucleotide phosphodiesterase which displays positive cooperative kinetic behavior!*g In
S
From the Department of Biological Sciences, University of Southern California. Los Angeles, Calif Received for publication October II, 1974. Supported by the Diabetes Association of Southern California and NIH Biomedical Sciences Support Grant RR-07012-07. Reprint requests should be addressed to M. M. Appleman. Department of Biological Sciences. University of Southern California. Los Angeles, Calif 90007. 0 I975 by Grune & Stratton. Inc. Metobolirm, Vol. 24, No. 3 (March), 1975
311
312
TERASAKI
AND APPLEMAN
the present study, some of the catalytic and regulatory properties losteric form of phosphodiesterase from rat heart will be reported. MATERIALS
of this al-
AND METHODS
The cyclic AMP and cyclic GMP phosphodiesterase assays of Russell et al.,s with minor modifications, have been employed. Incubation mixtures contained appropriately diluted enzyme, 20 mM MOPS-NaOH buffer (pH 7.0). 0.1 M sodium acetate, 1 mM 2-mercaptoethanol, 5 mM magnesium chloride, approximately 150,000 dpm tritiated cyclic nucleotide, and varying amounts of unlabeled cyclic nucleotide. The cyclic GMP-sensitive cyclic nucleotide phosphodiesterase (cNMP-PDE) was prepared from rat heart using a DEAE-cellulose chromatography system similar to that reported by Russell et a1.s for rat liver. Three distinct active fractions were eluted from the columns (Fig. I): Fraction I (cGMP-PDE) has a high affinity for cyclic GMP (K,, 2 x IOm6 M) and is regulatable by a calcium-dependent protein modulating factor. Fraction II (cNMP-PDE), the cyclic GMP-sensitive phosphodiesterase, displays no catalytic preference between cyclic GMP and cyclic AMP at high substrate concentrations. Fraction III (CAMP-PDE) is highly specific for cyclic AMP and gives evidence of negative cooperative regulation. The stability and identity of these enzyme forms were confirmed by the identical elution position and kinetic properties obtained upon rechromatography of the separated fractions (Fig. 2). The work in this paper was carried out solely on doubly chromatographied cNMP-PDE (Fraction II) from rat heart.
RESULTS
Kinetic Properties When represented on a Hill plot, cyclic AMP hydrolysis by cyclic nucleotide phosphodiesterase is cooperative (Hill coefficient, n = 1.4-1.8) with an apparent
80
20 FRACTION
30 NUMBER
40
50
Fig. 1. DEAE-cellulose profile of cyclic nuclaotide phosphodiestemse from rat heart. Sonicated extract was applied to a I .5 x 30-cm column of DEAE-cellulose which was developed with a linear 400 ml gradient from 0.07 to 1.O M sodium acetate (pH 6.5) containing 5 mM 2-mercaptoothanol. Fractions (8.0 ml) were collected at a Row rate of 30 ml/hr. Aliquots (200 Al) won measured directly for enzymatic activity using 1 pM cyclic AMP (A), 1pM cyclic GMP (o), or 100 PM cyclic AMP (A). The major peaks have been labeled fractions I, II, and Ill in order of their elution from the column according to Russell et al.’
CYCLIC GMP IN REGULATION
313
OF CYCLIC AMP HYDROLYSIS
0.8 0.4 0
Fig. 2. Rechromatography of mtphosphoheart cyclic nucleotide diestemses on DEAE-cellulose columns. Isolated fmctions from two DEAE-cellulose columns (as in Fig. 1) were dialyzed to remove the salt and individually rechromatogmphed. Activity was assayed using 1 MM cyclicAMP (o) or cyclic GMP (e). The profiles represent fmction I (upper panel), ffaction II (center panel), and fraction Ill (lower panel).
4
30
n
20
0.8
E E 4
IO
0.4
0
I
1
1
0
FRACTION NUMBER
K, of 6 x 1O-5 M. A plot of cyclic GMP hydrolysis by the enzyme appears as a biphasic curve which bends at a substrate concentration of 10e6 M (Fig. 3). The Hill coefficient in the range of concentrations below this value is nearly identical to the slope of the curve for cyclic AMP hydrolysis for the same enzyme preparation (n = 1.44). At higher concentrations of cyclic GMP, it approaches Michaelis-Mentin kinetics (n = 1.16). The K, for cyclic GMP is 2 x lo-’ M and the V,,, values for the two substrates are similar. Allosteric Activation Low concentrations to cyclic GMP have a pronounced effect on the nature of cyclic AMP hydrolysis only by this form of the phosphodiesterase.‘O Micromolar cyclic GMP stimulates the enzyme only at low concentrations of cyclic AMP substrate, while at nearly saturating substrate cyclic GMP has no effect. In the presence of cyclic GMP, the enzyme loses its cooperativity (the Hill coefficient decreased to 1.O) without any appreciable change in the V,,, or extrapolated K, values for cyclic AMP. Cyclic GMP appears to act as an allosteric activator has a classic K-type cooperative system.” The moderate decrease in the Hill coefficient (from an average of 1.6 to 1.O) which accompanies the loss of cooperativity represents a substantial increase (up to tenfold) in the hydrolysis of 1 PM cyclic AMP. The activator-velocity curves for cyclic GMP are noncooperative, with half-
TERASAKI
314
I.C
AND
APPLEMAN
r
0
n >
I * z > \ >
Fig. 3. Hill plot of cyclic nucleotide hydrolysis by rat-heart phosphodiesterare. Enzyme activity was measured in the range from 100 to 0.1 PM cyclic GMP. The curve for cyclic GMP activity (e) is biphasic, and the dashed line is the best fit for the values below a substrate concentmtion (5) of 10e6 M. For comparison, cyclic AMP activity for the same enzyme preparation is also presented (a). The Hill coefflcients (n) am the sloper of the lines as they are drawn here.
-1.0
u
z -2.c I
I
-7.0
t
I
-6.0
1
-5.0
-4.0
log (S)
maximal effects generally occurring at about 3 x lO-7 M (Fig. 4). Concentrations of 5’ GMP as high as 50 PLM are unable to mimic cyclic GMP, and guanosine (0.5 mM) inhibits activation by cyclic GMP by about 50% without changing the basal (nonstimulated) catalysis of cyclic AMP. While the catalytic activity of phosphodiesterase is strongly dependent on the presence of magnesium ion, the degree of activation by cyclic GMP is not diminished by concentration reductions in this ion from 5.0 to 0.3 mM.
Fig. 4. Activator velocity curve for the effect of cyclic GMP on cyclic AMP hydrolysis. The concentmtion of cyclic AMP was 1 FM. The inset is a double reciprocal representation of the same data (appomnt K, of 3 x 10-7 M cyclic GMP). Velocity (V) is expressed as picomoles cyclic AMP hydrolyzed per minute per micmgmm enzyme protein, and V, is the basal activity measured in the absence of cyclicGMP.
V
I 0
I 0.2
I (
Cyclic
I 0.4
I
I 0.6
GMP ) , pM
I
I 0.8
_
CYCLIC GMP IN REGULATION
OF CYCLIC AMP HYDROLYSIS
315
I
I
I
I
I
-
0
LO/ \\ \ \\
l
0.2tlM
Cyclic GMP
\
0.8 - ‘y \ \
v VO
h
\
0.6 -
0
\
\
IO&
Fig. 5. Effect of cyclic AMP on cyclic GMP hydrolysis. Cyclic GMP hydrolysis was measured at substrata concentrations of 0.2 phi ( l) or 10 WM (0) in the presence of varying concentrations of cyclic AMP. Enzyme activity is expressed as the measuwd velocity in the presence of cyclic AMP (V) with respect to the basal activity in the absence of cyclicAMP (V,).
Cyclic 0
l 4 0
\-
P GMP ‘\
I
I
I
IO
20
30
(
0 .
.
-2-0
40
1
_
50
Cyclic AMP ),pM
Activating Eflect of Cyclic AMP It has been reported previously4 that cyclic AMP has no significant activating effect on cyclic GMP degradation by rat liver phosphodiesterase. This is true at 10 PM substrate, where cyclic AMP was strictly an inhibitor of cyclic GMP hydrolysis (Fig. 5). On the other hand, when the cyclic GMP substrate concentration is very low (0.2 PM), cyclic AMP has a small stimulatory effect. In addition, cyclic AMP is a more potent inhibitor of the enzyme when the substrate is high (IO PM) than when it is much lower (0.2 PM). Inhibition The hydrolysis of each cyclic nucleotide is inhibited by high concentrations of the other. Cyclic GMP inhibits the hydrolysis of cyclic AMP in a competitive fashion with a Ki of 26 PM. Cyclic AMP inhibits cyclic GMP hydrolysis also competitively with a K, of 24 PM. These inhibition constants are roughly similar to the respective Michaelis constants. Enzyme Desensitization Treatment of the phosphodiesterase with organic solvents (10% ethanol or acetone) destroys cooperativity with cyclic AMP as the substrate without appreciably decreasing the V,,, (Fig. 6). In addition, the enzyme loses its ability to become activated by cyclic GMP. Incubation of the enzyme with a low amount of the sulphydryl reagent, Nethylmaleimide (NEM, 0.01 moles/g protein), increases basal cyclic AMP hydrolysis, although the maximum attainable activity (cyclic GMP-stimulated) remains unchanged (Fig. 7A). At higher concentrations of NEM, the basal activity, as well as the cyclic GMP-stimulated activity, decreases. The cNMP phosphodiesterase is also fully desensitized by p-chloromercuribenzoate, another sulfydryl inhibitor. The removal of 2-mercaptoethanol by dialysis ac-
TERASAKI
316
AND APPLEMAN
0.6 -
0.4 -
.’
/ L
0.2 -
/
Control
L_JF:T, ,, 0
0.04
0.00
( Cyclic AMP?,
0.120.16
pM_’
Fig. 6. Effects of organic solvents on cocyclic opemtivity of AMP hydrolysis. Phosphodiestemse activity was measured in the pmsence of 10% acetone (II), 10% ethanol (A), or no addition (e) as a control. Velocity (V) is in picomoles cyclic AMP hydrolyzed per minuk per microgmm protein. The data am represented on a Lineweaver-Burk plot.
tivates and desensitizes the enzyme in a manner analogous to NEM (Fig. 7B). It is apparent that when the enzyme is maximally stimulated by cyclic GMP, it cannot be further activated by modification of sulfydryl groups. Relative Hydrolytic
Rates
Table 1 compares the relative catalytic rates of cyclic nucleotide phosphodiesterase in the presence of high and low substrate and allosteric effector. The activities exhibited by the enzyme are strikingly dependent on substrate type and concentration. At low substrate concentration (1 PM) the enzyme hydrolyzes cyclic GMP about five times faster than cyclic AMP. If the substrate concentration is increased, the rates for the two nucleotides become nearly equal, and, in fact, V,,, values indicate that cyclic AMP is the preferred substrate at very high concentrations. Under saturating conditions, degradation of cyclic GMP is SO%-90% of that for cyclic AMP. When cyclic AMP hydrolysis (at 1 PM substrate) is fully activated by cyclic GMP, the magnitude of the stimulation was about sixfold, and the ratio of hydrolysis of cyclic GMP to that of cyclic AMP activity becomes nearly identical to the ratio of their respective maximum velocities (Table 1). Like cyclic GMP, NEM also causes cyclic AMP and cyclic GMP hydrolytic rates to become nearly equal.
CYCLIC
GMP
IN
REGULATION
OF
CYCLIC
AMP
HYDROLYSIS
317
r
+cGMP
+cGMP
0
5 NEM
( moles/g
protein
E-MERCAPTOETHANOL
)
( mM 1
Fig. 7. Effect on N-ethylmaleimide or removal of 2-mercaptoethanol on cyclicGMP activation of mt-heart fraction II. (A) Cyclic nucleotide phosphodiestemse was treated with 0, 0.01, or 0.04 moles NEM per gmm enzyme protein, and then cyclic AMP hydrolysis was measured at 1 pM substmte in the presence of 0 (open ban) or 5 phi (striped bars) cyclic GMP. (B) Doubly chromatogmphed phosphodiestemse was dialyzed overnight against 0.1 M Tris-HCI (pH 7.4) convaining no added 2-mercaptoethanol. After dialysis, cyclic AMP hydrolysis was measured at 1 PM substmte in the presence (closed bars) or absence (open bars) or 1 pM cyclic GMP. The same prepamtion maintained overnight in 5 mM 2-mercaptoethanol sewed as a control.
Table 1. Relative Activities of Rat-heart Cyclic Nucleotide Phosphodiesterase Towards Cyclic GMP and Cyclic AMP Substrate Substrates Compared GMP/cyclic
AMP
at “ma,
0.8
f
0.1
(2)
Cyclic
GMP/cyclic
AMP
lOOBAt
1.2 f
0.2
(6)
Cyclic
GMP/cyclic
AMP
1 PM
4.7
f
1.5
(6)
Cyclic
GMP/cyclic
AMP
(cyclic
TpM
0.9
f
0.1
(4)
Cyclic
AMP
GMP
stimulated)/cyclic
5.9
f
2.0
(8)
shown
preparations
(cyclic are
the
means
in parenthesis.
in addition
double
Ratio (Number)
Cyclic
Ratios
tide
Concentration
reciprocal
to
the
kinetic
cyclic
GMP
f
“Cyclic
stimulated) AMP
standard GMP
AMP-labeled
plots in the range
1 IrM
deviation
stimulated” substrate. of from
with indicates
V,,,
6 to 200
the the
values PM
number presence are
substrate.
of
separately of
obtained
1 pM by
tested unlabeled
extrapolation
enzyme nucleofrom
TERASAKI
318
AND APPLEMAN
DISCUSSION
The data reported here, in addition to that reported earlier, indicate that the cyclic nucleotide phosphodiesterase (fraction II) from rat heart is much more complex an enzyme than would be inferred from its previous title of “low affinity-cyclic AMP-phosphodiesterase.” The enzyme does have a relatively high K, for cyclic AMP but hydrolyzes cyclic GMP with about the same substrate affinity, and each nucleotide is a competitive inhibitor of the other’s hydrolysis. There is good evidence for positive cooperative kinetic behavior for both substrates, but this is most demonstrable at low concentrations for cyclic GMP. Cyclic GMP can activate cyclic AMP hydrolysis; the reverse appears also to be true but is demonstrable only over a very narrow concentration range for both nucleotides. The activation of cyclic AMP hydrolysis by cyclic GMP is due to a loss of cooperativity rather than a change in the K, or maximal velocity. Similar effects can be achieved by treatment with dilute solvents, detergents, sulfydryl reagents, or by the removal of protective sulfydryl compounds. A model, derived from this data, might be considered as involving just two types of cyclic nucleotide binding sites. One would be a catalytic site, with about equal (and low) affinity for cyclic AMP and cyclic GMP in its nonactivated state and with about equal (but high) affinity for the two nucleotides in its activated state. The second type of site would be regulatory; it also could bind both cyclic nucleotides. If the regulatory site is unoccupied, the catalytic site is in its low-affinity state; if the regulatory site is occupied (or if the regulator component is functionally dissociated by chemical treatment), the catalytic site goes to its high-affinity state. The esoteric behavior of the enzyme is then a result of the much greater affinity of the regulatory site for cyclic GMP than for cyclic AMP. Confirmation of this model and of the probable existence of subunits to carry out the allosteric interactions must await further enzyme purification. The physiologic role of this complex phosphodiesterase regulatory system is difficult to assess. The cooperative nature of cyclic AMP hydrolysis could be effective only at very high concentrations of this nucleotide (greater than 10m5 M) such as might occur during a regulatory crisis. This interpretation could be supported by the apparent cytoplasmic location of the enzyme and its very high maximal velocity. As a cyclic GMP phosphodiesterase, the enzyme also leaves much to be desired; the substrate affinity is two to three orders of magnitude below reported levels of this nucleotide, so that even with the cooperative kinetic behavior it could not be very effective. Probably the most appropriate classification of the enzyme is as a cyclic GMP-activated cyclic AMP phosphodiesterase. The lOO-fold greater affinity for cyclic GMP as a regulator when compared to cyclic AMP as a substrate would serve to bring these two functions into line with the reported cyclic nucleotide levels. Concentrations of cyclic GMP as low as 4 x lo-* M will give significant stimulation of cyclic AMP hydrolysis at a 10e6 M substrate concentration. It is then quite possible for cyclic GMP to serve as a regulator of cyclic AMP levels in certain circumstances. This would correlate well with the “Yin-Yang” or dualism theory of Goldberg,3 in which the two nucleotides are viewed as having antagonistic roles in biologic regulation. The question might even be asked whether this
CYCLIC GMP IN REGULATION
OF CYCLIC AMP HYDROLYSIS
319
cyclic GMP regulation of cyclic AMP hydrolysis represents just another example of “Ying-Y ang” phenomena or if it might actually be the mechanism for a number of reported antagonistic actions. REFERENCES 1. Ashman DF, Lipton R, Melicow MM, Price TD: Isolation of adenosine 3’,5’-monophosphate and guanosine 3’,5’-monophosphate from rat urine. Biochem Biophys Res Commun 11:330-334, 1963 2. Robison GA, Butcher RG, Sutherland EW: Cyclic AMP. New York, Academic Press, 1971 3. Goldberg ND, O’Dea RF, Haddox MK: Cyclic GMP, in Advances in Cyclic Nucleotide Research, Vol. 3. New York, Raven Press, 1973, pp 115-223 4. Beavo JA, Hardman JG, Sutherland EW: Stimulation of adenosine 3’,5’-monophosphate hydrolysis by guanosine 3’,5’-monophosphate. J Biol Chem 246:3841-3846.1971 5. Franks DJ, MacManus JP: Cyclic stimulation and inhibition of cyclic AMP phodiesterase from thymic lymphocytes. them Biophys Res Commun 42:844849,
GMP phosBio1971
6. Manganiello V, Vaughan PM: Prostaglandin Et effects on adenosine 3’,5’-cyclic
monophosphate concentration and phosphodiesterase activity in fibroblasts. Proc Nat1 Acad Sci USA 69:269-273, 1972 7. Klotz U, Stock K: Influence of cyclic guanosine-3’,5’-monophosphate on the enzymatic hydrolysis of adenosine-3’,5’-monophosphate. Naunyn Schmiedebergs Arch Pharmakol 27454-62, 1972 8. Russell TR, Terasaki WL, Appleman MM: Separate phosphodiesterases for the hydrolysis of cyclic AMP and cyclic GMP in rat liver. J Biol Chem 248: 1334-l 340, 1973 9. Bevers MM, Smits RAE, Van Riju J, Van Wyk R: Heat treatment of rat liver adenosine 3’,5’-monophosphate phosphodiesterase. Biochim Biophys Acta 341:120-128.1974 10. Appleman MM, Terasaki WL: The regulation of cyclic nucleotide phosphodiesterase, in Advances in Cyclic Nucleotide Research, vol. 5. New York, Raven Press, (in press) 11. Monod J, Wyman J, Changeux JP: Qn the nature of allosteric transitions: A plausible model. J Mol Biol 12:88-l 18, 1965
