ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 278, No. 1, April, pp. 88-98, 1990

Inhibition of Adenylate Cyclase by Polyadenylate’ Mark

Bushfield,

Ilana

Shoshani,

Maria

Cifuentes,

Dorothee

Stiibner,

and Roger A. Johnson”

Department of Physiology and Biophysics, School of Medicine, Health Sciences Center, State University New York at Stony Brook, Stony Brook, New York 11794.8661.

Received May 16,1989, and in revised form November

27,1989

The effects of ribo- and deoxyribonucleic acids on the activity of detergent-dispersed adenylate cyclases from rat and bovine brain were examined. Mn2+ (10 mM)-aCtivated adenylate cyclase was inhibited by micromolar concentrations of poly(A) (I&, N 0.45 MM). This inhibition was directly due to poly(A) and was not mediated by: (a) protein contamination of the poly(A) preparation, (b) metal chelation, (c) formation of an acid-soluble inhibitor of adenylate cyclase, (d) effects on the specific activity of [(r-32P]ATP, (e) competition with MnATP for binding to adenylate cyclase, or (f) diversion of substrate to an alternate polymerase reaction. Inhibition of adenylate cyclase by poly(A) was on the enzyme’s catalytic unit, as purified preparations of the enzyme from bovine brain were inhibited by poly(A). This inhibition by poly(A) was not likely mediated via the enzyme’s “P’‘-site, through which activated forms of the enzyme are selectively inhibited by specific adenosine phosphates. In contrast with inhibition by the “P’‘-site agonist 3’ AMP, inhibition of adenylate cyclase by poly(A) was slow in onset and was not reversible by dilution and showed a different metal-dependence. Inhibition of adenylate cyclase was relatively specific for poly(A) as poly(U) caused less than 50% inhibition and deoxyribonucleic acids had no effect. The potency and specificity of the inhibition of adenylate cyclase by poly(A) imply a biochemically interesting interaction that is possibly also of physiological significance. Q 1990 Academic Press. Inc.

Adenylate cyclases (ATP pyrophosphate-lyase (cyclizing), EC 4.6.1.1) from most tissues are inhibited via 1 This work was supported by Grants DK 38828 and DK 33494 to R.A.J. from the National Institutes of Health and by a fellowship of the Deutsche Forschungsgemeinschaft to D. Stiibner. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ’ Present address: Molecular Pharmacology Group, Institute of Biochemistry, University of Glasgow, Glasgow G12 8QQ, UK. ” To whom correspondence and reprint requests should be sent. 88

of

a unique site by adenosine and analogs of adenosine retaining an intact purine moiety (“P’‘-site) (l-3). This inhibition has been demonstrated in intact cells, membranes, and solubilized preparations of adenylate cyclase (2,4-7). Available evidence suggests that the “P”-site is on the catalytic unit of adenylate cyclase (5-9). Although “P”-site-mediated inhibition of adenylate cyclase does not require the presence of guanine nucleotide-binding regulatory proteins, G, or Gi, inhibition is metal dependent and is evidently only of activated forms of the enzyme, e.g., enzyme activated by stimulatory hormone receptors, activated G,, forskolin, or Mn2+ (5-7). We have used the solubilized, calmodulin-sensitive adenylate cyclase from rat brain as a model system for the study of “P”-site-mediated inhibition (6, 7). With this system, whether adenylate cyclase was activated by Mn2+ or by proteolysis in the presence of GTPyS, the most potent “P’‘-site agonists were 2’d3’AMP, 2’,5’dideoxyadenosine, and 3’AMP, with IC6o’s of 1 to 3 PM. Of these, 2’d3’AMP and 3’AMP are naturally occurring. Sahyoun et al. (10) determined levels of 2’d3’AMP in a number of tissues and suggested that it may be a naturally occurring inhibitor of adenylate cyclase and that it could be derived from the breakdown of DNA. Whether or not there may be a link between DNA breakdown and adenylate cyclase activity is unknown. However, a number of studies have shown that CAMP regulates gene expression in E. coli and in eukaryotes (reviewed by Roesler et al. (11)). This effect may be mediated by enhanced gene transcription and/or mRNA stabilization (12) and results in increased cytosolic levels of poly(A)+ mRNA. Consequently, given the potency of the adenosine phosphates, their possibly being derived from nucleic acids, and our previous observation that various dinucleotides (d(AT), d(AA), d(AC), and d(AG)) were also potent inhibitors of adenylate cyclase (6,7), we undertook a study of the effects of nucleic acids on adenylate cyclase activity. We report here the potent * Abbreviations used: GTPyS, guanosine 5’.O-(3.thiotriphosphate); CAMP, adenosine 3’5’monophosphate; Chaps, 3-[(3-cholamidopropyl)dimethylammonia]propanesulfonic acid. 0003.9861/90

$3.00

Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

POLY(A)

INHIBITION

OF ADENYLATE

inhibition by polyadenylate of adenylate cyclases from several sources. The nature of the inhibition by poly(A) has been characterized and the possibility that it may be via the “P”-site was investigated. EXPERIMENTAL

PROCEDURES

Preparation of aclenylate cyclase. Washed membrane particles were prepared from rat and bovine brains, and detergent-dispersed adenylate cyclase from rat brain was prepared essentially as described previously (13). Adenylate cyclase was purified from bovine brain essentially as described by Smigel (14) as we described elsewhere (7,15). Rat epididymal fat pads were obtained as described by Rodbell (16) and plasma membranes were prepared as described by Laudat et al. (17). Protein was determined essentially according to Lowry et al. (18). Determination of adenylate cyclase actiuity. Adenylate cyclase activities were determined essentially as previously described (19, 20). Reaction mixtures typically contained 50 mM triethanolamine.HCl, pH 7.5, 1 mM 3.isobutyl-l-methylxanthine, 1 mM dithiothreitol, bovine serum albumin (1 mg/ml), 2 mM purified creatine phosphate, creatine kinase (100 fig/ml), 100 FM ATP, 10 mM MgCl, or 10 mM MnCl,, and [n-“‘P]ATP (2 to 5 X lo” cpm), in a volume of 100 ~1. The formation of [“‘P]cAMP was determined after incubation for 15 min at 30°C. For kinetics experiments MnATP was varied from 5 to 400 FM MnATP with an excess of 10 mM MnCl,, as described by Garbers and Johnson (21). The MnC12 concentrations used in kinetic experiments were determined by atomic absorption spectrometry and ATP concentrations were verified spectrophotometrically. Reactions were terminated by precipitation with zinc acetate, containing 0.5 mM unlabeled CAMP, and sodium carbonate (20). [“‘PIcAMP was isolated by sequential chromatography on Dowex 50 then Al,O, columns, essentially as described by Salomon et al. (22). The unlabeled CAMP served to monitor recovery of the labeled CAMP by measurement of absorbance at 259 nm with an aliquot of the sample. Radioactivity was determined by Cherenkov radiation in a liquid scintillation spectrometer. Purification of nucleic acids. Total RNA was extracted from rat livers as described by Chirgwin et al. (23). Poly(A)’ mRNA was prepared by affinity chromatography with poly(U)-Sephadex, as described in the manufacturer’s suggested working procedures (Bethesda Research Laboratories, Inc.). DNA (from salmon testes), poly(A) (synthetic), poly(dA) (synthetic), and other nucleic acids from commercial sources were purified prior to use. This involved extraction with phenol/chloroform/isoamyl alcohol (25/24/l) followed by precipitation with ethanol, as described by Maniatis et al. (24). After precipitation, the nucleic acids were taken up in sterile Millipore water, at a concentration of 3 to 10 mg/ml. Concentrations of nucleic acids were determined spectrophotometrically. The size and purity of the poly(A) and poly(A)+ mRNA were checked by agarose gel electrophoresis under denaturing conditions (24) and by reverse-phase HPLC. The poly(A) migrated as a single broad band upon agarose gel electrophoresis, corresponding to a size of approximately 250 bases (by use of an RNA ladder from Bethesda Research Laboratories Inc.). The variation in poly(A) chain length implies a maximum possible error of 50% in the concentrations of poly(A) used in this report. Poly(A) was eluted from a Cl8 reverse-phase HPLC column (Beckman) as a single uv(A260nm)-absorbingpeak with a gradient of 5 to 50% acetonitrile in triethanolammonium acetate (100 mM; pH 7). Labeling the 5’ end of poly(A) with 7’4 polynucleotide kinase. Poly(A) (45 pmol), that had been dephosphorylated by treatment with alkaline phosphatase (see below), was incubated for 60 min at 37°C with 20 units of polynucleotide kinase and [y-‘*P]ATP (150 PCi, 50 pmol), in the presence of 20 mM MgCl,, 10 mM dithiothreitol, 0.2 mM EDTA, and 100 mM Tris.Cl, pH 7.6, as described by Maniatis et al. (24). The reaction was terminated by heating at 95°C for 1 min and the labeled poly(A) was isolated by precipitation with ethanol. Calf inLIephosphorylation of poly(A) with alkaline phosphatase. testinal alkaline phosphatase (molecular biological grade) was used to

CYCLASE

89

dephosphorylate both the 3’ and 5’ends of poly(A), as described in the manufacturer’s suggested working procedure (Boehringer-Mannheim). Under these conditions, greater than 97% of the radioactivity was liberated from 5’[3”P] end-labeled poly(A). 5’AMP, S’AMP, and 5’ADP, but not ATP, were also good substrates for alkaline phosphatase and were converted to adenosine under these conditions (as assessed by uv-absorption following anion-exchange HPLC). Materials. [a-“‘P]ATP and [y-““P]ATP were prepared enzymatitally as described by Walseth and Johnson (19), or were purchased from New England Nuclear or from ICN Pharmaceuticals. DEAEagarose, Dowex 50, neutral alumina, and all electrophoresis reagents were from Bio-Rad. ATP, GTP, GTPyS, creatine phosphate, and creatine kinase were obtained from Boehringer-Mannheim. Creatine phosphate was purified as previously described (25). Nucleic acids were obtained either from Sigma or, where stated, from BoehringerMannheim. Ohgo(A),, was from Pharmacia. Proteinase K, calf intestinal alkaline phosphatase, and T4 polynucleotide kinase were from Boehringer-Mannheim. Other reagents were from commercial sources and were of the highest quality available.

RESULTS

Effects of poly(A) on detergent-dispersed adenylate cyclase from rat brain. Figure 1 compares the effects of 3’AMP and crude and purified poly(A) on adenylate cyclase activity in the presence of 10 mM MnCIZ. In confirmation of previous reports (6, 7), 3’AMP inhibited Mnzt-activated adenylate cyclase with an I&,, of approximately 3 PM, an effect mediated via the “P”-site. Crude poly(A) from Boehringer-Mannheim reduced activity up to 20% over the concentration range tested (4 to 830 pg/ml), equivalent to 0.04 to 8.3 PM). In other experiments other preparations of this crude poly(A) caused a biphasic effect, with up to 40% inhibition at - 1 PM, but a return to basal values at 3 to 10 PM poly(A). After purification, though, all preparations of poly(A) exhibited a concentration-dependent inhibition of adenylate cyclase with an IC& of ~0.4 PM, equivalent to 0.04 mg/ml (Fig. 1). (Poly(A) from Sigma inhibited adenylate cyclase with a similar IC& (0.45 + 0.04 PM, n = 22), whether or not it had been purified.) The differences in the responses to poly(A) imply the presence of factor(s) in the crude preparations of poly(A) that counteract the effects of the oligonucleotide on adenylate cyclase. The nature and number of such factor(s) has not been pursued further. In all subsequent experiments all nucleic acids were purified before use. Specificity of the inhibition of adenylate cyclase by poZy(A). As shown in Fig. 2, poly(dA) (2.8 to 280 pg/ml) had no effect on adenylate cyclase activity, while tRNA (from yeast, 75% amino acid incorporation) caused up to 25% inhibition at high concentrations (1.5 mg/ml). Poly(U) was only partially inhibitory, reducing activity no more than 48% at 100 pg/ml. Total cellular RNA and RNA which did not bind to poly(U)-Sephadex (i.e., lacking tracts of poly(A)) caused a maximal 25% inhibition of adenylate cyclase at 100 to 800 pug/ml (Fig. 3). Poly(A)+ mRNA caused 50% reduction of adenylate cyclase activity at approximately 250 pg/ml (Fig. 3). In contrast, all other deoxyribonucleic acids tested had no

90

BUSHFIELD

L,

’ 0.01

I 0.1

I 0.03

ET AL.

1 03 1.0 CONCENTRATION (yM)

1 10

3.0

30

FIG. 1. Effects of 3’AMP and crude and purified poly(A) on solubilized adenylate cyclase from rat brain. Adenylate cyclase activity (10.75 fig protein/tube) was determined in the presence of the indicated concentrations of 3’AMP (a), crude poly(A) (from Boehringer-Mannheim) (o), or purified poly(A) (0). Adenylate cyclase activity in the absence of these additions was 30.2 pmol CAMP (15 min. mg protein) ‘. Values are averages of duplicate determinations from a representative from three similar experiments.

significant effect on adenylate cyclase activity. These included: DNA (0.1 to 10 mg/ml), poly(dA).(dT) (25 to 250 pg/ml), and poly(dAT) . (dTA) (12.5 to 125 pg/ml). Hence, this inhibitory effect on adenylate cyclase was confined to ribonucleic acids with poly(A) being the most potent and poly(U) being only partially effective.

o-v-

The importance of the size of poly(A) was determined by examining the effects of short oligomers on adenylate cyclase activity. Oligo(A)rZmIs (i.e., a mixture of oligomers 12 to 18 nucleotides long), at concentrations from 0.1 to 31 PM, had no significant effect on adenylate cyclase activity determined in the presence of 10 mM Mr?+

.-._,

‘/.

POLY(da)

L.

A

-I

x. ‘\.

\

tRNA

/ ‘\

“A

0 0.1

I

I

I

I

I

I

0.3

I

3

10

30

100

300

CONCENTRATION t/q /lOOpI ) FIG. 2. Effects of several nucleic acids on solubilized adenylate cyclase from rat brain. Poly(A) (o), poly(U) (v), poly(dA) (V), or tRNA (A) were present at the indicated concentrations. The molecular sizes (daltons) of the different nucleic acids were: poly(A) - 100,000; poly(dA) in the absence of these additions - 62,000; poly(U) - 250,000; tRNA - 33,000. Control activity of adenylate cyclase (10.8 Kg protein/tube) was 28.7 pmol CAMP (15 min. mg protein))‘. Values are averages of duplicate determinations from one of three similar experiments.

POLY(A)

cl 0.1

INHIBITION

OF ADENYLATE

91

CYCLASE

I

I

I

1

I

0.3

I

3

10

30

CONCENTRATION

I

( w / 100 IJI )

FIG. 3. Effects of poly(A) and RNA on solubilized adenylate cyclase from rat brain. Poly(A) (O), total cellular RNA from rat liver (A), poly(A)+ mRNA (m), and RNA which lacked tracts of poly(A) (A) were present at the indicated concentrations. Poly(A)+ mRNA and RNA lacking tracts of poly(A) (nonbound RNA) were prepared chromatographically with poly(U)-Sephadex, as described under Experimental Procedures. Adenylate cyclase activity in the absence of these additions was 16.0 pmol CAMP (15 min. mg protein) I. Values are averages of duplicate determinations from a representative of two similar experiments.

(data not shown). The data suggest that long chains of poly(A) are required to inhibit adenylate cyclase under these conditions. of some potential mechanisms by which Elimination adenylate cyclase could be inhibited by poly(A). Several potential mechanisms were considered by which poly(A) could bring about an apparent reduction in adenylate cyclase activity. Formation of adenosine, 3’AMP, or other nucleotides from poly(A) during the adenylate cyclase incubation could result in inhibition of the enzyme via the “P’‘-site. To test for the formation of such acid soluble inhibitors, the adenylate cyclase was incubated with 1 PM poly(A) or 30 PM 3’AMP for 15 min at 30°C under normal assay conditions except without labeled ATP. This incubation was stopped by the addition of perchloric acid, which was then neutralized with potassium carbonate. The mixture was centrifuged and the acid soluble supernatant was then tested for its effect on adenylate cyclase or was injected onto an anion-exchange HPLC column (Synchropak AX300). As shown in Table I, 1 FM poly(A) caused a 75% inhibition of adenylate cyclase, whereas the acid extract of the poly(A) subsequent to the incubation had no significant effect on enzyme activity. Under these conditions there was also no detectable formation of adenosine, 3’AMP, or any other nucleotide as determined by HPLC techniques. By comparison, 30 FM 3’AMP inhibited adenylate cyclase by 90% and this inhibitory effect was only slightly di-

minished by the preincubation with the solubilized brain preparation and the acid extraction procedure (Table I). The unlikely possibility that inhibition of adenylate cyclase might be mediated by residual protein contaminants in the purified poly(A) preparation was also considered. As shown in Fig. 4, treatment of the purified

TABLE Test

I

for the Formation of Acid-Soluble Inhibitors Control

Acid extract V hb,tor/ VC,,,,“,

S’AMP pol~(N

0.10 0.25

0.33 0.95

Note. Adenylate cyclase activity was determined in the absence or presence of 30 pM 3’AMP or 1 pM poly(A) (control column). In parallel incubations, [o-s’P]ATP was omitted from the incubation mixture and the reaction was terminated by the addition of 10 al of 3 M perchloric acid. Samples were neutralized by the addition of 15 ~1 of 1.5 M potassium carbonate. Following centrifugation (full speed for 5 min at room temperature in an IEC Microfuge), 50 ~1 of the supernatant was added to a second incubation in order to determine the effects on adenylate cyclase activity. Note that due to these additions, assays with the acid extracts contained 40% of the original concentration of 3’AMP (i.e., 12 FM). The control adenylate cyclase activity was 14.1 pmol CAMP (15 min. mg protein) -‘. Values are averages of duplicate determinations from one of two similar experiments.

92

BUSHFIELD I

I

I

ET AL. I

I

I

. CONTROL POtY(A) 0 PROTEINASE K-TREATED POLYIA) A ALKALINE PHOSPHATASETREATED FOLY(A) V NOS;;fiEGENERATING

1.0 -

.

g iTi 9 .

0

:’

I

0.1

0.3

I

1.0

I

3.0

I

IO

I

30

CONCENTRATION1pg 1100 $1 FIG. 4. Poly(A) inhibition of adenylate cyclase after treatment with proteinase K or dephosphorylation with alkaline phosphatase and with omission of the ATP-regenerating system. Adenylate cyclase (10.8 pg protein/tube) activity was determined in the presence of purified poly(A), in the absence (‘J) or presence (0) of the ATP-regenerating system, creatine phosphate and creatine phosphokinase, or in the presence of purified poly(A) that had been treated with proteinase K (0) or with alkaline phosphatase (A) (cf. Experimental Procedures). Adenylate cyclase activity in the absence and presence of the ATP-regenerating system was 18.5 and 17.4 pmol CAMP (15 min. mg protein))‘, respectively. Values are averages of duplicate determinations from a representative of three similar experiments.

poly(A) with proteinase K (26), to digest contaminating proteins, had no significant effect on the inhibition of adenylate cyclase by poiy(A). The possibility that adenylate cyclase, or a contaminating enzyme, could be acting as a polymerase to divert radiolabel from [ CX-~~P]ATP into poly(A) instead of CAMP was examined by measuring the incorporation of radioactivity into acid-precipitable material. Under standard assay conditions, approximately 1% of added radioactivity was found in the precipitate after the addition of 0.3 M perchloric acid, either in the absence or in the presence of 0.1 to 1 FM poly(A). Hence, inhibition of adenylate cyclase by poly(A) is not an artifact caused by the diversion of ATP from the adenylate cyclase reaction into a polymerase reaction pathway. Since commercially available poly(A) was prepared from ADP by polynucleotide phosphorylase (27), it was possible that a pyrophosphate group was present at the 5’ end. If this were the case, it is conceiveable that poly(A) could be further phosphorylated to form a 5’-triphosphate by the ATP-regenerating system (creatine phosphate and creatine phosphokinase) present in the adenylate cyclase incubation. Such a compound could, conceivably, compete with the substrate (MnATP) for adenylate cyclase. However, omission of the ATP-regenerating system from the adenylate cyclase incubation did not affect the ability of poly(A) to inhibit activity (Fig. 4). The slight rightward shift in the concentrationresponse curve in the absence of the ATP-regenerating system (Fig. 4) was not consistently seen in three sepa-

rate experiments. In addition, complete dephosphorylation of the poly(A) by alkaline phosphatase had no effect on the inhibitory activity of poly(A) (Fig. 4). Metal-dependence and kinetics of inhibition of adenylate cyclase by poly(A); comparison with “F-site inhibition. Since both free divalent cation (Mg” or Mn’+) and metal-ATP (MgATP or MnATP) are required for the bireactant mechanism of catalysis by adenylate cyclase (28), it is possible that poly(A) could reduce adenylate cyclase activity through the formation of a complex with the free cation. To address this, we examined the effects of varying the concentration of free Mg2+ or Mn2+ on the inhibition of adenylate cyclase by poly(A) or by 3’AMP. Inhibition via the “P’‘-site is known to be metal-dependent (29). As shown in Fig. 5A, increasing the concentration of free Mg2+ from 0.5 mM to 10 mM increased adenylate cyclase activity and increased sensitivity to inhibition by 10 pM 3’AMP (Fig. 5B). Maximal inhibition was seen once the Mg2+ concentration had reached -6 mM. This corresponds well with the Mg2+dependence of “P”-site-mediated inhibition of platelet adenylate cyclase by adenosine (29). By comparison, 0.7 pM poly(A) caused little or no inhibition of adenylate cyclase in the presence of Mg2+ (Fig. 5B) and inhibition by higher concentrations of poly(A), up to 3.3 pM, caused no more than 15% reduction in adenylate cyclase activity measured with Mg2+ (data not shown). Consistent with previously published data (28), the solubilized brain adenylate cyclase was strongly activated by 0.03 to 10 mM free Mn2+ (Fig. 5A). Activation

POLY(A)

[FREE METAL] ,

6

1

I

INHIBITION

OF ADENYLATE

(mM) I

I

\poly(A)

t

I

. 1

1

‘\ l\.\ % \

/wb (Al. 3’AMP

/

o-0-0 I

I

0

2 [FREE M:NGANE:E]

,

I

0 ,

IO (mM;

FIG. 5. Metal ion dependence for inhibition of adenylate cyclase by 3’AMP and poly(A). (A) The activity of solubilized adenylate cyclase (10.8 lg protein/tube) was determined in the presence of the indicated concentrations of free Mg”+ or Mn ‘+, in excess of a fixed concentration of the respective metal-ATP (100 PM). (B) EfIects of 10 pM 3’AMP or 0.7 pM purified poly(A) on adenyloate cyclase activity were determined in the presence of the indicated concentrations of free Mg’+ in excess of a fixed concentration of 100 pM MgATP. Adenylate cyclase activities in the absence of 3’AMP or poly(A) are shown in (A). (C) Effects of 10 pM 3’AMP or 0.7 ~.LM purified poly(A) on adenylate cyclase activity were determined in the presence of the indicated concentrations of free Mn*+ in excess of a fixed concentration of 100 fiM MnATP. Control adenylate cyclase activities are shown in (A). Values are averages of duplicate determinations from one of two similar experiments.

was maximal with eO.3 mM free Mn2+ and the response reached a plateau at a slightly lower level with higher Mn*+ concentrations (Fig. 5A). Under these conditions both poly(A) and 3’AMP showed marked dependence on the concentration of free Mnzf for inhibition of adenylate cyclase (Fig. 5C). Inhibition by 3’AMP was maximal at ~2 mM free Mn’+, whereas maximal inhibition by poly(A) required higher concentrations of free Mn’+ (Fig. 5C). This difference in the metal-dependence of inhibition by poly(A) and 3’AMP was also observed at roughly equipotent concentrations of 3’AMP (1 PM) and poly(A) (0.3 FM) (data not shown). In separate experiments (not shown) the concentration of free Mnzf was

CYCLASE

93

increased to 20 mM without significantly affecting the maximal inhibition of adenylate cyclase by poly(A). These data indicate that the inhibition of adenylate cyclase by poly(A) is not due to metal chelation. The effects of poly(A) on the kinetic behavior of adenylate cyclase with respect to MnATP were investigated with a fixed concentration of 10 mM free Mn’+ (see footnote 5), with both crude and purified preparations of the enzyme (Fig. 6). Double-reciprocal plots for the crude enzyme are shown in Fig. 6A, with no additions or with 0.1, 0.33, or 1.0 PM poly(A), or with 10 PM 3’AMP. The lines shown in Fig. 6A were generated by linear regression analysis of the data obtained with increasing MnATP concentrations up to 80 PM, above which there was an apparent substrate-dependent inhibition. The apparent substrate inhibition was not observed in the presence of 3’AMP, nor was it observed at lower concentrations of free Mn2+ (28). Data obtained without and with 10 PM 3’AMP resulted in lines which intersected to the left of the ordinate on the abscissa. These data are consistent with the previously published kinetic behavior of “P”-site-mediated inhibition by adenosine, which results in noncompetitive inhibition with respect to metal-ATP (29). Consistent with inhibitory behavior being noncompetitive, sensitivity to inhibition by 3’AMP was not influenced by substrate concentration (not shown). By comparison, varying concentrations of poly(A) resulted in a series of lines which intersected to the left of the ordinate above the abscissa (Fig. 6A). However, the sensitivity to inhibition by poly(A) decreased with increasing substrate concentration; the ICSO for poly(A) increased from 0.28 to 0.62 PM as MnATP increased from 5 to 400 PM (data not shown). These characteristics of poly(A) inhibition kinetics were also observed with the purified enzyme (Fig. 6B). Increasing concentrations of poly(A) (20 to 160 nM) resulted in a pattern of lines intersecting to the left of the ordinate and above the abscissa and sensitivity to inhibition also diminished with increasing substrate concentration. At 80 PM MnATP the ICSO for poly(A) was 20 nM, whereas at 400 PM MnATP it was greater than 160 nM (not shown). Not surprisingly, plots of l/velocity versus [poly(A)] at increasing substrate concentrations resulted in a family of hyperbolic curves (not shown). The experiments represented in Fig. 6B were conducted with enzyme that had been preincubated with poly(A) for 30 min at 30°C. Under these conditions, inhibition by poly(A) is optimal and stable, and the overall kinetic behavior is not due to nor influenced by the slow onset of inhibition (see below). Hence, inhibition of adenylate cyclase by poly(A) was mixed (changes in slope and in’ Although the concentration of free Mn” was not also varied, precluding the determination of true K,,, and V,,,,, values, with this high concentration of free Mn*+, at least two orders of magnitude greater than the KmMn, the apparent kinetic constants will be close approximations of actual kinetic constants (see (28, 301).

94

BUSHFIELD

ET AL.

I

I

I

POLY(A)

I

I

I

0

0.05

0.10

1

0.15

I I

0.20

l/MnATP (PM)-’ e 0.3

PURIFIED BOVINE BRAIN ADENYLATE CYCLASE

POLY[A] ( nM

)

z 0

0.2

0 i > ’ v

0.1

A@

1 / [MnATP],

(1 / pM)

FIG. 6. Double-reciprocal plot of kinetics of inhibition of adenylate cyclase by poly(A) with respect to MnATP. Adenylate cyclase activities were determined with concentrations of MnATP that varied from 5 to 400 pM MnATP at a fixed concentration of 10 mM excess MnCI,, in the absence or presence of the indicated concentrations of purified poly(A). (A) Crude detergent-dispersed adenylate cyclase from bovine brain. Inhibition by poly(A) is compared with inhibition by 10 pM 3’AMP (*). Values are averages of duplicate determinations in one of two similar experiments. (B) Purified bovine brain adenylate cyclase. Enzyme was preincubated for 30 min at 30°C with poly(A) and a reaction mixture that was complete except for [w~‘P]ATP. After the preincubation ((u-32P]ATP was added and the formation of [32P]cAMP was determined in a second 15.min incubation at 30°C. Values are averages of duplicate determinations from one of three similar experiments.

tercepts) with respect to MnATP and was distinguishable from “P”-site-mediated inhibition in that inhibition by poly(A) was influenced by substrate binding. Time- and protein-dependence and reversibility of the inhibition by poly(A). With the reaction conditions used, adenylate cyclase activity was linear with respect to protein concentration for both crude and purified preparations of the enzyme. This was true in the absence or presence of 10 PM 3’AMP (data not shown), but was

not true for both crude and purified preparations of the enzyme in the presence of poly(A) (Fig. 7). Poly(A) (1 PM) more effectively inhibited the crude enzyme at low protein concentrations (>80% at 37 pg/ml) than at high protein concentrations (-12% at 740 pg/ml) (Fig. 7). In other experiments with the crude enzyme we noted that raising the protein concentration from 2.7 to 215 pg/ml resulted in an increase in the IC& for inhibition by poly(A) from 100 nM poly(A), but exhibited an IC& N 40 nM under these conditions (Fig. 10). As was pointed out previously, though, sensitivity to inhibition diminished with increasing substrate concentration and ICSOvalues for the purified enzyme > 160 nM were noted at 400 PM MnATP and approached 10 nM at 5 PM MnATP (cf. Fig. 6). Whether adenylate cyclase is inhibited by poly(A) may be a question of access by poly(A). It may be that detergents and enzyme purification allow poly(A) access to sites that would not be accessible in the environment of intact membranes. To address this in another way, experiments were conducted with alamethicin. Alamethicin, a channel-forming antibiotic ionophore that has been used by others to unmask latent adenylate cyclase activity in membrane vesicles (32), increased adenylate cyclase activity - 75% in membranes from rat TABLE

II

Comparison of Poly(A) Inhibition Solubilized,

of Particulate, and Purified Forms of Adenylate No additions

DetergentCyclase

0.1% Lubrol-PX V poly,A,/ Vcontrol

Rat brain washed particles Bovine brain washed particles Rat fat cell plasma membranes Solubilized rat brain particles Purified bovine brain catalytic unit

0.80 f 0.01

0.47 f 0.02

0.81 + 0.01

0.34 + 0.06

0.71 f 0.05

0.45 f 0.01 0.27 f 0.01 0.36 f 0.03

Note. Adenylate cyclase activity was measured in the absence or presence of 1 FM poly(A) and/or 0.1% Lubrol-PX. The activities in the absence of poly(A) or detergent were (as pmol CAMP (15 min’mg protein))‘) rat brain membranes (10.2 pg protein/tube), 4.91 + 0.83; bovine brain membranes (9.7 fig/tube), 2.64 k 0.22; rat fat cell membranes (11.9 pg/tube), 11.3 _+ 0.6. The enzyme activities in the presence of 0.1% Lubrol-PX were rat brain membranes, 8.97 f 0.26; bovine brain membranes, 5.73 f 0.01; rat fat cell membranes, 12.2 f 0.3; solubilked rat brain enzyme (10.8 fig), 26.5 k 1.0; purified catalytic subunit, 6.77 f 0.51 pmol CAMP per tube. All values shown are the averages of duplicate determinations f range from two separate experiments.

POLY(A)

INHIBITION

OF ADENYLATE

(POLY (A)]

CYCLASE

97

M

of detergent-dispersed and partially purified forms of adenylate cyclase from bovine brain to inhibition by poly(A). FIG. 10. Sensitivity Activities were determined with 100 pM MnATP and 10 mM Mn”. Values are averages of duplicate determinations from one of two similar experiments. I&, values from both experiments are given for poly(A) in Table II.

brain, a slightly greater extent than seen with 0.1% Lubrol-PX (-50%). Alamethicin did not, however, affect the sensitivity of the enzyme to inhibition by polyadenylate (not shown). These effects of alamethicin are not surprising, though, in that while the pores formed by the antibiotic may be sufficient to allow passage of the relatively smaller substrate molecules, consistent with the behavior in heart (32), the pores may not have been sufficiently large to allow passage of molecules the size of the polyadenylate (- 100 kD mean mass) to inhibit adenylate cyclase. DISCUSSION

We report here that poly(A) inhibited the detergentsolubilized adenylate cyclase from rat brain with an IC& of eO.45 PM. The inhibition was remarkably selective for long chain poly(A), with poly(dA) and short chain poly(A), for example, being completely without effect. To help establish the validity of the inhibitory effect of poly(A) on adenylate cyclase, several possible and obvious alternative mechanisms were tested. We established that inhibition by poly(A) was not due to (a) protein contamination of the poly(A) preparations, (b) the formation of acid-soluble inhibitors, (c) effects on the specific activity of substrate ATP, (d) incorporation of ATP into poly(A) by a polymerase reaction, (e) competition with ATP for binding to the catalytic unit, or (f) an action at the inhibitory “P”-site, through which specific adenosine and deoxyadenosine phosphates inhibit adenylate cyclase. Adenylate cyclase that had been purified to near homogeneity by affinity chromatography on forskolin-agarose was inhibited by polyadenylate with an IC,, N 40

nM, roughly an order of magnitude lower than that found with the unpurified solubilized preparation (I(& N 0.45 PM). This difference may be due to one or more of several alternatives: (a) a high degree of nonspecific binding of poly(A) in the crude preparation, (b) the catalytic unit in the crude preparation was not fully accessible for poly(A) binding, or (c) components or factors (e.g., subunits of G-proteins or other ribonucleotides) may be removed during purification that otherwise would impair the binding of poly(A). More importantly, the inhibition of the purified catalytic unit by poly(A) suggests that it is this component of the adenylate cyclase system to which poly(A) binds to inhibit the enzyme. Since the inhibition by poly(A) was substantially more potent for the purified enzyme than for the crude enzyme, it is not likely that the small amount (~5%) of G, contaminating this preparation (consistent with Smigel(14)) mediates inhibition by poly(A). Whether the catalytic unit contains the putative consensus sequence for poly(A) binding that has been found in poly(A)-binding proteins (33) awaits comparisons with the sequences of catalytic components of the several adenylate cyclases. The slow onset and poorly reversible nature of the inhibition of adenylate cyclase by poly(A) would argue against a simple interaction between poly(A) and the catalytic unit. These characteristics would be consistent with a high-affinity binding of poly(A) to adenylate cyclase and/or a poly(A)-dependent enzymatic process occurring prior to inhibition of adenylate cyclase. However, since inhibition was reversed by high concentrations of salts, the inactivation of adenylate cyclase was not due to denaturation. It is not known whether inhibition of the purified enzyme is reversible.

98

BUSHFIELD

Poly(A) exerts a very potent inhibition of adenylate cyclases from several sources and the potency of inhibition of the detergent-solubilized and especially the purified forms of the enzyme from brain make polyadenylate the most potent nonhormonal inhibitor of adenylate cyclase. While the potency of this inhibition would be consistent with a potential physiological role, the lack of effect of poly(A) on adenylate cyclase in membrane preparations would argue against an obvious in vivo role. It is conceiveable, though, that since poly(A) inhibited particulate adenylate cyclases in the presence of membrane perturbants, other cellular factors could also modify or modulate the sensitivity of the enzyme to inhibition by poly(A) in viuo. Alternatively, it may be that poly(A) would only interact with forms of adenylate cyclase that are not associated with membranes, as might occur during the synthesis, migration, and/or degradation of the enzyme. A number of functions have been ascribed to the poly(A) sequences found at the 3’ end of most eukaryotic mRNAs (34-40) and the average steady state size of poly(A) is known to change and be controlled by external factors in other systems (41-43). If one assumes a tissue content of total RNA of 20 mg per gram wet weight (23), 5% of which is mRNA with an average size of 2000 nucleotides, the intracellular mRNA concentration would be approximately 2.4 PM and the normal cellular concentration of poly(A)+ mRNA would be in a range comparable to those found here to inhibit adenylate cyclases. ACKNOWLEDGMENTS We express thanks to Dr. Yaacov Hod for his help in obtaining and purifying several of the nucleic acids used in these studies and to Drs. Alison Marker and Raafat El-Maghrabi for their help in the chromatographic and electrophoretic purification of poly(A). We acknowledge the help of Dr. Siu-Mei Helena Yeung who provided the solubilized rat brain adenylate cyclase. The critical reading of the manuscript and helpful comments by Dr. Joel G. Hardman, Vanderbilt University, were greatly appreciated.

REFERENCES 1. Haslam, R. J., and Lynham, J. A. (1972) Life Sci. Part ZZ Biochem. Gen. Mol. Biol. 11, 1143-1154. 2. Fain, J. N., Pointer, 247,6866-6872.

R. H., and Ward, W. F. (1972) J. Biol. Chem.

3. Londos, C., and Wolff, 5482-5486.

J. (1977) Proc. Natl. Acad. Sci. USA 74,

4. Wolff, J., Londos, C., and Cook, G. H. (1978) Arch. Biochem. Biophys. 191,161-168. 5. Florio, 202.

V. A., and Ross, E. M. (1983) Mol. Pharmacol.

24, 195-

6. Johnson, R. A. (1986) Fed. Proc. 44(3), 874. 7. Johnson, R. A., Yeung, S.-M. H., Stiibner, D., Bushfield, Shoshani, I. (1989) Mol. Pharmacol. 35,681-688. 8. Premont, J., Cuillon, G., and Bockaert, phys. Res. Commun. 90,513-519.

M., and

J. (1979) Biochem. Bio-

9. Johnson, R. A. (1982) FEBS Z&t. 140,80@84.

ET AL. 10. Sayhoun, N., Schmitges, C. J., Siegal, M. I., and Cuatrecasas, P. (1976) Life Sci. 19,1961-1970. 11. Roesler, W. J., Vandenbark, G. R., and Hanson, R. W. (1988) J. Biol. Chem. 263,9063%9066. 12 Hod, Y., and Hanson, R. W. (1988) J. Biol. Chem. 263, 77477752. 13. Johnson, R. A., and Sutherland, E. W. (1974) in Methods in Enzymology (Hardman, J. M., and O’Malley, B. W., Eds.), Vol. 38, pp. 135-143, Academic Press, New York. 14. Smigel, M. D. (1986) J. Biol. Chem. 261,1976-1982. 15. Stiibner, D., and Johnson, R. A. (1989) FEBSLett. 248,1555161. 16. Rodbell, M. (1964) J. Biol. Chem. 293,375-380. 17. Laudat, M. H., Pairault, J., Bayer, P., Martin, M., and Laudat, P. (1972) Biochim. Biophys. Acta 255,100551008. 18. Lowry, D. H., Rosenbrough, N. J., Farr, A. L., and Randall, R. J. (1951)J Biol. Chem. 193,26&275. 19. Walseth, T. F., and Johnson, R. A. (1979) Biochim. Biophys. Acta 562,11-31. 20. Jakobs, K. H., Saur, W., and Schultz, G. (1976) J. Cyclic Nucleotide Res. 2, 381-392. 21. Garbers, D. L., and Johnson, R. A. (1975) J. Biol. Chem. 250, 8449-8456. 22. Salomon, Y., Londos, C., and Rodbell, M. (1974) Anal. Biochem. 58,541-548. 23. Chirgwin, J. M., Przybla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18,5294-5299. 24. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, p. 202, Cold Spring Harbor. 25. Johnson, R. A. (1980) J. Biol. Chem. 255,8252-8258. 26. Davis, L. G., Dibner, M. D., and Battey, J. F. (1987) Basic Methods in Molecular Biology, p. 44, Elsevier, New York. 27. Grunberg-Manago, M., and Ochoa, S. (1955) J. Amer. Chem. Sot. 77,3165-3166. 28. Johnson, R. A., and Garbers, D. L. (1977) in Receptors & Hormone Action (O’Malley, B. W., and Birnbaumer, L., Eds.), pp. 549-572, Academic Press, New York. 29. Johnson, R. A., Sam, W., and Jakobs, K. H. (1979) J. Biol. Chem. 254,1094~1101. 30. Cleland, W. W. (1970) in The Enzymes (P. D. Boyer, Ed.), Vol. 2, 3rd ed., pp. l-65, Academic Press, New York. 31. Jakobs, K. H., Saur, W., and Johnson, R. A. (1979) Biochim. Biophys. Acta 583,409-421. 32. Besch, H. R., Jr., Jones, L. R., Fleming, J. W., and Watanabe, A. M. (1977) J. Biol. Chem. 252,7905-7908. 33. Sachs, A. B., Davis, R. W., and Kornberg, R. D. (1987) Mol. Cell Biol. 7,3268-3272. 34. Schroder, H. C., Rottmann, M., Wenger, R., Bachmann, M., Dorn, A., and Miiller, W. E. G. (1988) Biochem. J. 252,777-790. 35. Nudel, U., Soreq, H., and Littauer, U. Z. (1976) Eur. J. Biochem. 64,115-121. 36. Palatnik, C. M., Storti, R. V., Capone, A. K., and Jacobson, A. (1980) J. Mol. Biol. 141,99-118. 37. Doel, M. T., and Carey, N. H. (1976) Cell 8,51-58. 38. Parets Soler, A., Gozalbo, D., Zueco, J., and Sentandreu, R. (1987) Biochem. J. 246,575-581. 39. Galili, G., Kawata, E. E., Smith, L. D., and Larkins, B. A. (1988) J. Biol. Chem. 263,5764-5770. 40. Jacobson, A., and Favreau, M. (1983) Nucleic Acids Res. 11,63536368. 41. Brawerman, G., and Diez, J. (1975) Cell 5, 271-280. 42. Slater, I., Gillespie, D., and Slater, D. W. (1973) Proc. Natl. Acad. Sci. USA 70,406-411. 43. Zingg, H. H., Lefebvre, D. L., and Almazan, G. (1988) J. Biol. Chem. 263,11,041-11,043.