Forskolin Derivatives with Increased Selectivity for Cardiac Adenylyl Cyclase

J Mol Cell Cardiol 30, 97–108 (1998)

Forskolin Derivatives with Increased Selectivity for Cardiac Adenylyl Cyclase Yoshiyuki Toya, Carsten Schwencke and Yoshihiro Ishikawa Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115, USA (Received 24 June 1997, accepted in revised form 15 September 1997) Y. T, C. S  Y. I. Forskolin Derivatives with Increased Selectivity for Cardiac Adenylyl Cyclase. Journal of Molecular and Cellular Cardiology (1998) 30, 97–108. The current study was undertaken to examine whether we can target adenylyl cyclase to regulate b-adrenergic signaling with increased cardiac selectivity. Forskolin, a natural diterpene compound, interacts directly with adenylyl cyclase. We studied the adenylyl cyclase isoform-selectivity of forskolin derivatives using insect cell membranes overexpressing type II, III, and V adenylyl cyclase isoforms. 6-[3-(dimethylamino) propionyl] forskolin (NKH477) stimulated type V more potently (1.87±0.02-fold) than type II (1.04±0.02-fold) and type III (0.89±0.03-fold) relative to forskolin (50 l, P<0.05). Similarly, 6-[3-(dimethylamino)propionyl]-14,15-dihydro-forskolin (DMAPD) stimulated type V (1.39±0.02-fold) more potently than types II (0.66±0.02-fold) and type III (0.31±0.02-fold) relative to forskolin (P<0.05). This selectivity was maintained under different assay conditions – i.e. with different forskolin (0.1–100 l) and Mg (1–10 m) concentrations, with or without Gsa. NKH477 increased cAMP accumulation in HEK293 cells stably overexpressing type V more than forskolin (1.57±0.13-fold) (P<0.05). Examination of multiple tissue homogenates revealed that DMAPD and NKH477 stimulated cardiac adenylyl cyclase more potently than the other tissue adenylyl cyclases (lung, brain, and kidney) relative to forskolin. Our results suggest that a particular side-chain modification of forskolin enhanced the selectivity for the cardiac isoform stimulation. Adenylyl cyclase isoforms may be targeted to increase tissue selectivity in future drug therapy for b-adrenergic regulation.  1998 Academic Press Limited K W: Adenylyl cyclase isoform; Heart; Forskolin derivatives; NKH477; 6-Dimethylaminopropionyl14,15-dihydro-forskolin.

Introduction Activation of the sympathetic nervous system initiates the most potent stimulus to enhancing cardiac output. Blockade or activation of the b-adrenergic receptor (b-AR) system is thus a reasonable pharmacological approach to regulate cardiac function. Over the years, numerous b-agonists and b-antagonists have been developed and widely used in the treatment of cardiovascular diseases. Recent investigations have demonstrated, however, that bARs are subject to dynamic alterations under many pathophysiological conditions (e.g. down-regulation in heart failure). Rapid changes in the properties of b-ARs, including desensitization and

sensitization after b-agonist and b-antagonist therapy, have limited the usefulness of such therapy, in addition to the adverse effects originating from the poor tissue selectivity of b-agonists and bantagonists (Braunwald, 1997). Adenylyl cyclase makes up a large enzyme family consisting of multiple isoforms (types I through IX) that differ in tissue distribution (Taussig and Gilman, 1995; Ishikawa and Homcy, 1997). The type V adenylyl cyclase is the major isoform in the heart (Ishikawa et al., 1992, 1994; Tobise et al., 1994; Espinasse et al., 1995; Yu et al., 1995), types I and VIII are expressed exclusively in the brain (Cali et al., 1994), type II is a major isoform in the lungs (Feinstein et al., 1991), and type III is expressed

Please address all correspondence to: Yoshihiro Ishikawa, Cardiovascular & Pulmonary Research Institute, Allegheny University, Pittsburgh, PA 15212, USA.

0022–2828/98/010097+12 $25.00/0

mc970575

 1998 Academic Press Limited

98

Y. Toya et al.

abundantly in olfactory tissues (Xia et al., 1992). In the diversity of tissue distribution, the adenylyl cyclase isoforms thus make a clear contrast to the other component within the b-AR pathway. Forskolin, like digitalis, is a natural plant extract from Coleus forskolii hitherto used only in traditional medicine. Forskolin acts directly on adenylyl cyclase to increase intracellular levels of cAMP (Seamon et al., 1981; Daly, 1984). Forskolin increases ventricular contractility and evokes vasodilatation in vivo (Lindner et al., 1978). In human heart failure, forskolin augments ejection fraction, stroke volume, and cardiac output and reduces ventricular filling pressures and vascular resistance (Bristow et al., 1984). Despite these beneficial inotropic and vasodilatative effects, poor water solubility and a broad spectrum of non-specific pharmacological activities have hampered its clinical application in modern medicine. 6-[3-(Dimethylamino)propionyl]forskolin (NKH477) is a water-soluble forskolin derivative that was recently introduced as an inotropic agent in heart failure (Hosono et al., 1992; Shafiq et al., 1992; Hirasawa et al., 1993; Mori et al., 1994; Sanbe and Takeo, 1995). NKH477 has significantly reduced bio-distribution in the brain and less neuronal effects than forskolin (Iwase et al., 1996), presumably because of poor passage through the blood–brain barrier. We examined whether we can target adenylyl cyclase to regulate b-adrenergic signaling with inreased cardiac selectivity. We examined forskolin and its analogs, including 6-aminopropionyl derivatives on various adenylyl cyclase isoforms. We demonstrate that 6-aminopropionyl derivatives of forskolin have increased selectivity for type V adenylyl cyclase over types II and III adenylyl cyclase.

centrations, forskolin (50 or 100 l) was dissolved easily in the final reaction mixture without precipitation and adenylyl cyclase activity was not affected by DMSO. Our preliminary data showed that the catalytic activity of adenylyl cyclase was not affected in the presence of an increasing concentration of DMSO (0–2%). For example, type V catalytic activity was 4.120 pmol/mg.min at 0%, 4.132 pmol/mg.min at 1%, 4.196 pmol/mg.min at 2%.

Overexpression of adenylyl cyclase isoforms in insect cells A 3.5-kb Bam HI fragment from type II adenylyl cyclase cDNA (Feinstein et al., 1991), a 4.0-kb Bam HI–Hind III fragment from type III (Bakalyar and Reed, 1990) (both kindly provided by Dr R. Reed), and a 4.0-kb Bam HI–Ssp I fragment from type V (Ishikawa et al., 1992) were subcloned into pBluBac vectors (Invitrogen, San Diego, CA, USA) as previously described (Kawabe et al., 1994a; Ebina et al., 1997). Recombinant shuttle vectors were transfected into High Five cells, an insect cell line derived from Trichoplusia ni egg cell homogenates (Invitrogen). The plaques were then purified as described previously (Kawabe et al., 1994b). High Five cells (1–2×109) were infected with recombinant baculovirus and incubated in InsectXPRESS medium (BioWhittaker Inc., Walkersville, MD, USA) containing 6% fetal bovine serum at 27°C for 3 days.

Insect cell membrane preparations

Materials and Methods Forskolin and its derivatives Forskolin was purchased from Calbiochem-Novabiochem Corp. (San Diego, CA, USA). Forskolin derivatives: 6-[(dimethylamino)acetyl]forskolin; 6[3-(dimethylamino)propionyl]forskolin (NKH477); 6-[3-(dimethylamino)propionyl]-7-deacetyl-forskolin; 6-[3-(dimethylamino)propionyl]-14,15-dihydroforskolin; 6-[3-(methylamino)propionyl]forskolin; and 6-(3-aminopropionyl)forskolin; were kindly provided by Nippon Kayaku Co., Ltd (Tokyo, Japan) (Tatee et al., 1996). Forskolin and its derivatives were dissolved in dimethylsulfoxide (DMSO) to a final concentration of 0.5 or 1.0%. At these con-

High Five cells were washed twice with ice-cold phosphate-buffered saline and homogenized in a buffer containing 50 m Tris/HCl (pH 8.0), 1 m EGTA, 1 m EDTA, 1 m dithiothreitol, 200 m sucrose, and a protease inhibitor mixture containing 20 lg/ml 1-chloro-3-tosylamido-7-amino-2-heptanone, 10 lg/ml leupeptin, 0.1 m phenylmethylsulfonyl fluoride, 50 U/ml egg white trypsin inhibitor, and 2 lg/ml aprotinin (Kawabe et al., 1994a). Cells were disrupted with a sonicator and centrifuged at 500×g for 10 min at 4°C. The supernatants were further centrifuged at 100 000×g for 40 min at 4°C. The resultant pellets were resuspended in the same buffer without EGTA. The crude membrane preparations were stored at −80°C until use.

Forskolin and Cardiac Adenylyl Cyclase

Tissue preparation C57/BL6 mice were purchased from Taconic (Germantown, NY, USA). Tissues (heart, brain, lung, and kidney) were minced in a buffer containing 50 m Tris/HCl (pH 8.0), 1 m EGTA, 1m EDTA, 1 m dithiothreitol, 200 m sucrose, and a protease inhibitor mixture (10 lg/ml leupeptin, 1 m phenylmethylsulfonyl fluoride, 50 units/ml egg white trypsin inhibitor, 20 lg/ml 1-chloro-3-tosylamido-7-amino--2-heptanone, and 2 lg/ml aprotinin). The tissues were then homogenized with a Polytron for 3×10 s, followed by centrifugation at 500×g for 10 min at 4°C. The supernatants were retained and further centrifuged at 100 000×g for 40 min at 4°C. The crude membrane preparations were made by resuspending the pellet in the same buffer without EGTA and stored at −80°C until use. Animals used in this study were maintained in accordance with the guidelines of the Committee on Animals of the Harvard Medical School and the ‘‘Guide for the Care and Use of Laboratory Animals’’ (DHHS Publication No. [National Institute of Health] 83-23, revised 1985).

Purification of Gsa Gsa was overexpressed in High Five insect cells and was purified as previously described (Jones et al., 1993), with some modification. After cell lysis in a buffer containing 20 m Tris/HCl (pH 7.5), 1 m EDTA, 1 m EGTA, 5 m b-mercaptoethanol, 100 l GDP, 100 l AlCl3, 10 m NaF, and a protease inhibitor mixture, the lysates were centrifuged at 100 000×g for 40 min at 4°C. The supernatant fraction was applied to a DEAE-MenSep 1500 column (Millipore, Bedford, MA, USA) preequilibrated with a loading buffer containing 20 m Tris/HCl (pH 7.5), 1 m EDTA, 5 m b-mercaptoethanol, 2 m MgCl2, and 0.1 m PMSF. After washing, the protein was eluted with a gradient from 0 to 200 m NaCl in the loading buffer. Fractions with high GTPcS binding activity were pooled and then applied to a pre-equilibrated hydroxyapatite column, BioGel HTP (Bio-Rad Laboratories, Hercules, CA, USA). The protein was eluted with a gradient from 10 to 300 m potassium phosphate. Fractions with high GTPcS binding activity were pooled and concentrated with a Centricon-30 (Amicon Inc., Beverly, MA, USA) and was resuspended in a buffer containing 20 m HEPES (pH 8.0), 0.5 m EDTA, 1 m b-mercaptoethanol, 0.5 m PMSF, and 20% glycerol. This operation for concentration-dilution was repeated three times.

99

Final samples were stored at −80°C until use. For activation, the purified Gsa was incubated at 30°C for 30 min in the presence of 100 l GTPcS and 5 m MgCl2.

Adenylyl cyclase assay Adenylyl cyclase activity was measured as previously described, with some modification (Kawabe et al., 1994a). In brief, the reaction mixtures contained 20 m HEPES (pH 8.0), 0.5 m EDTA, 0.2 m ATP containing a-32P-ATP (1×106 ct/min), 0.2 m cAMP containing 3H-cAMP (1×104 ct/ min), 1 m creatine phosphate, 8 units/ml creatine phosphokinase, 0.5 m 1-methyl-3-isobutylxanthine (IBMX), and 2–5 lg of membrane protein in a final volume of 100 ll. The assay was performed at 30°C for 20 min and terminated by the addition of 100 ll of ice-cold 2% SDS, followed by heating at 80°C for 10 min. cAMP formed during the incubation was separated with Dowex and neutral alumina columns (Salomon, 1979) and corrected for recovery with 3H-cAMP. Protein concentration was measured with the Bio-Rad protein assay system (Bio-Rad Laboratories, Hercules, CA, USA).

Stable transfection of type V adenylyl cyclase into human embryonic kidney (HEK) 293 cells HEK 293 cells were grown at 37°C in Dulbecco’s modified Eagle’s medium supplemented with 5% fetal bovine serum in a humidified 95% air/5% CO2 incubator. Cells were transfected with either the type V adenylyl cyclase cDNA in pcDNAneo or pcDNAneo alone by the calcium phosphate method (Kawabe et al., 1996). Neomycin-resistant cells were selected in culture medium containing G418 (500 lg/ml).

3

H-adenin labeling and cAMP accumulation assay

cAMP accumulation in intact cells was measured according to the method of Wong et al. (1991). The cells were placed in a serum-free culture medium containing 0.1% bovine serum albumin and 3Hadenine (0.5 lCi/well) for 18 h. The cells were then washed once with a HEPES balanced salt solution (HBSS) (137 m NaCl, 5 m KCl, 1.2 m CaCl2, 1 m MgCl2, 20 m HEPES (pH 7.5), 10 m Dglucose and 0.2% phenol red) and pretreated with

100

Y. Toya et al.

HBSS containing 0.5 m IBMX for 10 min. Reactions were started by the addition of 50 l forskolin derivatives and terminated by the addition of 12% (w/v) trichloroacetic acid, 0.25 m ATP and 0.25 m cAMP. The 3H-ATP and 3H-cAMP were separated by the method of Salomon (1979). The cAMP production is expressed as 3H-cAMP/ (3H-cAMP+3H-ATP)×103.

Data analysis and statistics All data were expressed as mean±... Infection of insect cells and their membrane preparations were performed independently for each assay. We performed statistical analysis using non-paired Student’s t-test. A probability value <0.05 was considered indicative of a significant difference.

Results Effects of forskolin derivatives on adenylyl cyclase isoforms The overexpression of the recombinant adenylyl cyclase isoforms increased catalytic activity by approximately 12-fold (for type II), 11-fold (for type III), and 57-fold (for type V) over that of control cell membranes as determined in the presence of forskolin (50 l). Thus, under these conditions, each isoform was examined as the dominant positive adenylyl cyclase isoform in these cells. We examined a representative isoform from each subgroup as the dominant positive isoform in insect cells, i.e. type II from the type II/IV/VII subgroup, type III from the type III subgroup, and type V from the type V/VI subgroup because it was difficult to examine all the adenylyl cyclase isoforms (Ishikawa and Homcy, 1997). It is known that biochemical properties are relatively well conserved within each subgroup (Taussig and Gilman, 1995). The concentration of forskolin and its derivatives used in our assays was higher than that used in in vivo studies (Iwase et al., 1996), because biochemical analysis of adenylyl cyclase catalytic activity requires higher concentrations of activators including forskolin and b-agonists than physiological analysis. Table 1 summarizes the effect of forskolin and its derivatives on the catalytic activity of adenylyl cyclase isoforms. The potency of each compound was compared as the percentage of forskolin-stimulated activity. An R6-N, N-dimethylaminoacetyl

derivative of forskolin {compound A, 6-[(dimethylamino)acetyl]forskolin} had a slightly weaker effect relative to forskolin on all types of adenylyl cyclase (type II, 76%; type III, 90%; type V, 92%). However, the conversion at the same position into 3-(N, N-dimethylamino)propionyl group increased the stimulation of type V (187%), while it maintained a similar effect on types II (89%) and III (104%) (compound B, 6-[3-(dimethylamino)propionyl]forskolin or NKH477). These findings indicate that a simple modification at R6 position from N, N-dimethylaminoacetyl group to 3-(N, N-dimethylamino)propionyl group increased the selectivity for the type V isoform. To test whether 3-(N, N-dimethylamino)propionyl group at R6 position alone is sufficient for the increased selectivity, we examined additional modifications of the same position or other positions. 6-[3-(Methylamino)propionyl]forskolin (compound C) and 6-(3-aminopropionyl)forskolin (compound D), which reduce one or two N-methyl groups compared with NKH477, maintained similar effects as NKH477. Modification at the positions R7 and R13 decreased the potency, but maintained the similar selectivity as NKH477 {compound E, 6[3-(dimethylamino)propionyl]-7-deacetylforskolin; compound F, 6-[3-(dimethylamino)propionyl]-14, 15-dihydroforskolin or DMAPD}. In particular, DMAPD possessed greater effect on type V (139%) than forskolin, while having much less effect on the other two isoforms (type II, 66%; type III, 31%). Thus, so long as 3-(N, N-dimethylamino)propionyl group at R6 position is present, the compounds stimulated type V greater than types II and III.

Dose–response of 6-aminopropionyl derivatives To confirm greater selectivity of 6-aminopropionyl derivatives for type V over types II and III than forskolin, we conducted dose–response studies of forskolin, NKH477, and DMAPD on these isoforms. As shown in Figure 1, 100 l of forskolin stimulated type II, type III, and type V adenylyl cyclase activity by 6-, 27-, and 18-fold, respectively. NKH477 had an effect similar to that of forskolin on types II and III, and a much stronger effect on type V; it stimulated type II, type III and type V adenylyl cyclase activity by 7-, 29-, and 34-fold. DMAPD had a weaker effect than forskolin on types II and III, but a stronger effect on type V; it stimulated type II, type III and type V adenylyl cyclase activity by 4-, 9-, and 27-fold. Thus, the two 6-aminopropionyl derivatives clearly showed increased selectivity for type V adenylyl cyclase.

101

Forskolin and Cardiac Adenylyl Cyclase

Table 1 The chemical structure of forskolin derivatives and their effect on adenylyl cyclase isoforms (types II, III, and V). High Five cell membranes overexpressing types II, III, and V adenylyl cyclase were prepared as described in ‘‘Materials and Methods’’. Assays were performed in the presence of 50 l various forskolin derivatives; A, 6[(dimethylamino)acetyl]forskolin; B, 6-[3-(dimethylamino)propionyl]forskolin (NKH477); C, 6-[3-(methylamino) propionyl]forskolin; D, 6-(3-aminopropionyl)forskolin; E, 6-[3-(dimethylamino)propionyl]-7-deacetylforskolin; F, 6-[3(dimethylamino)propionyl]-14,15-dihydroforskolin (DMAPD). All assays were performed in duplicate. Data are shown as the percentage of the catalytic activity with forskolin. The absolute value of catalytic activity in the presence of 50 l forskolin was 970±49 pmol/mg.min for Type II, 832±58 pmol/mg.min for type III, and 4422±275 pmol/ mg.min for type V. Each value represents mean±... from four independent experiments Compound

Forskolin A B (NKH477) C D E F (DMAPD)

Adenylyl cyclase activity (% of Fsk-stimulated activity)

Position R6

R7

R13

Type II

Type III

Type V

Hydroxy (Dimethylamino) acetoxy 3-(Dimethylamino) propionyloxy 3-(Methylamino) propionyloxy 3-Amino propionyloxy 3-(Dimethylamino) propionyloxy 3-(Dimethylamino) propionyloxy

Acetoxy Acetoxy Acetoxy

Vinyl Vinyl Vinyl

100% 76±2 104±2

100% 90±2 89±3

100% 92±3 187±2

Acetoxy

Vinyl

124±2

108±2

245±3

Acetoxy

Vinyl

126±1

117±4

247±4

Hydroxy

Vinyl

43±2

26±2

82±2

Acetoxy

Ethyl

66±2

31±2

139±2

CH3

R13

O OH CH3

CH3

O

OH

R7

H3C

CH3

Compounds C and D, which are metabolites of NKH477, stimulated the adenylyl cyclase isoforms in a similar manner as NKH477; they stimulated type V more potently than type II and III at both high (50 l, see Table 1) and low concentrations (1 l; C, 203% on type V; 105% on type II, and 104% on type III; D, 198% on type V; 113% on type II, and 109% on type III, see also Fig. 1). Importantly, the isoform selectivity of these for-

R6

skolin derivatives was not subject to the amount of the enzyme protein in assay tubes [Fig. 1(d)]; at different concentrations of the insect cell membrane protein (0.5–2.0 lg/tube), the selectivity of the isoforms was unchanged. It should be pointed out, however, that we do not know the absolute amount of each adenylyl cyclase isoform protein in insect cells. This is because of the lack of common antibodies with the requisite sensitivity to detect isoforms with a similar sensitivity.

102

Y. Toya et al.

(b)

(a)

1.5

1.5

1.0

*

0.5 *

cAMP production (nmol/mg/min)

cAMP production (nmol/mg/min)

*

1.0

0.5

†

* †

0

0 0

10–7 10–6 10–5 Concentration (mol/l)

10–4

0

10–4

(d)

(c)

AC V

†

10

† 8

† *

6

†

4

AC activity (% of FSK-stimulated activity)

200

†

cAMP production (nmol/mg/min)

† * * 10–7 10–6 10–5 Concentration (mol/l)

150

AC II

AC III

100

50

* 2

0 0

0

–7

–6

–5

10 10 10 Concentration (mol/l)

–4

10

0.5

2

10

0.5 2 10 0.5 Protein ( g/tube)

2

10

Forskolin and Cardiac Adenylyl Cyclase

103

Effect of MgCl2

Effect of forskolin derivatives in intact cells

Forskolin-stimulated catalytic activity of adenylyl cyclase changes with Mg concentration (Pieroni et al., 1995). To exclude the possibility that the selectivity for type V occurs only at a high Mg concentration (5 m) used in the above assays, we examined the effect of NKH477 at variable concentrations (1–10 m). Forskolin stimulation of types II and V by an increasing concentration of MgCl2 was saturable by 2–5 m [Figs 2(a) and (c)]. In contrast, type III was stimulated in a concentration-dependent manner to the highest concentration examined (10 m) [Fig. 2(b)]. NKH477 stimulated type II slightly more than, type III the same as, and type V twice as much as forskolin. The potency of NKH477 to stimulate type V, however, was decreased at higher Mg concentrations [Fig. 2(c)]. Thus, NKH477 stimulated type V greater than forskolin at any Mg concentrations.

To exclude the possibility that the greater selectivity of 6-aminopropionyl derivatives for type V occurs only when assayed using membrane preparations, we examined the effect of forskolin and NKH477 on cAMP accumulation in intact HEK293 cells overexpressing type V adenylyl cyclase. Forskolinstimulated cAMP accumulation in this cell line was five times as great as that in control cells, as previously described (Kawabe et al., 1996). As shown in Figure 4, NKH477 increased the cAMP accumulation more than forskolin (>1.5-fold), i.e. a pattern of stimulation similar to that observed with membrane preparations.

Effects of Gsa-stimulation Forskolin effect is potentiated when adenylyl cyclase is stimulated with Gsa upon hormonal activation (Seamon et al., 1981). Thus, we examined whether the selectivity of 6-aminopropionyl derivatives for type V is maintained in the presence of Gsa stimulation. As shown in Figure 3, the purified GTPcSGsa increased basal activity of types II, III, and V by approximately 10-, 4-, and 7-fold, respectively. The synergistic activation of type V by forskolin and Gsa was modest, in contrast to the greater synergism with types II, which agrees with previously reported results (Sutkowski et al., 1994). The effect of forskolin on type III dramatically increased with Gsa stimulation. An identical pattern of synergistic stimulation was observed with NKH477. Further, with or without Gsa activation, NKH477 was twice as stimulatory to type V as was forskolin. Thus, the selectivity for type V by NKH477 was maintained in the Gsa-stimulated condition.

Effect of forskolin derivatives in tissues Finally, we examined whether the greater selectivity of 6-aminopropionyl derivatives for type V is maintained in tissues (Fig. 5). We used multiple mouse tissues (heart, kidney, lung, and brain) that are known to express a distinct mixture of adenylyl cyclase isoforms (Ishikawa and Homcy, 1997). NKH477 stimulated all tissue adenylyl cyclases more potently than forskolin and stimulated cardiac adenylyl cyclase most potently among the tissue adenylyl cyclases. In contrast, DMAPD stimulated tissue adenylyl cyclase more potently than forskolin only in the heart. Thus, both NKH477 and DMAPD stimulated cardiac adenylyl cyclase more potently than forskolin. The selectivity of DMAPD for cardiac adenylyl cyclase seemed greater than that of NKH477.

Discussion Although all the forskolin derivatives used in our study stimulated adenylyl cyclase type II, III, and V as well as multiple tissue adenylyl cyclases, some derivatives showed a clearly distinct pattern of stimulation from that of forskolin. In particular,

Figure 1 Effect of forskolin, DMAPD, and NKH477 on adenylyl cyclase isoforms. Forskolin derivatives dose-response effect [(a)–(c)]. High Five cell membranes overexpressing type II (a), type III (b), and type V (c) adenylyl cyclases were prepared as described in Materials and Methods. Adenylyl cyclase activity was measured in the presence of 5 m of MgCl2 with various concentrations of forskolin (Χ), DMAPD (Α) or NKH477 (Β). Means±... from four independnt experiments are shown. ∗, P<0.05 and †, P<0.01 differences from the value with forskolin. Protein dose-response effect (d). Different amounts of H5 cell membrane proteins overexpressing each isoform were used in assays. Adenylyl cyclase activity was measured in the presence of 50 l of DMAPD (––––), NKH477 (– – – –) or forskolin. Data are shown as the percentage of the catalytic activity with 50 l forskolin. Assays were performed in duplicate. Similar results were obtained in two independent assays.

104

Y. Toya et al. (a) 1.5

cAMP production (nmol/mg/min)

*

1.0

0.5

0

2

4 6 Mg (mmol/l)

(b)

8

10

(c) 1.5

10

† †

8

†

cAMP production (nmol/mg/min)

cAMP production (nmol/mg/min)

† 1.0

0.5

6

4

2

0

2

4 6 Mg (mmol/l)

8

10

0

2

4 6 Mg (mmol/l)

8

10

Figure 2 Effect of MgCl2 on forskolin- or NKH477-stimulated adenylyl cyclase activity. High Five cell membranes overexpressing type II (a), type III (b), and type V (c) adenylyl cyclases were prepared as described in Materials and Methods. Adenylyl cyclase activity was measured in the presence of an increasing concentration of MgCl2 with 50 l forskolin (Χ) or 50 l NKH477 (Β). Means±... from five or six independent experiments are shown. ∗, P<0.05 and †, P<0.01 differences between the value with forskolin and that with NKH477.

105

Forskolin and Cardiac Adenylyl Cyclase (a)

cAMP production (nmol/mg/min)

8

N. S.

N. S.

6

4

2

0

Basal

Fsk NKH477 Gs

(b)

Gs Gs + + Fsk NKH477

(c) 5

N. S.

14

N. S.

†

†

12

cAMP production (nmol/mg/min)

cAMP production (nmol/mg/min)

4

3

2

10

8

6

4

1 2

0

Basal

Fsk NKH477 Gs

Gs Gs + + Fsk NKH477

0

Basal

Fsk NKH477 Gs

Gs Gs + + Fsk NKH477

Figure 3 Effect of Gsa on forskolin- or NKH477-stimulated adenylyl cyclase activity. High Five cell membranes overexpressing type II (a), type III (b), and type V (c) adenylyl cyclases were prepared as described in Materials and Methods. Adenylyl cyclase activity was measured in the presence of 5 m MgCl2 with 50 l forskolin (hatched bars), or 50 l NKH477 (solid bars). Purified recombinant Gsa was activated in the presence of Mg-GTPcS and diluted to a final concentration of 10 nM at the time of assay. Means±... from four independent experiments are shown. ∗, P<0.05 and †, P<0.01 differences between the value with forskolin and that with NKH477.

106

Y. Toya et al.

500

* *

cAMP accumulation ( 1000)

400

* 300

200

100

0

Basal

Fsk

NKH477

Figure 4 Effect of forskolin and NKH477 in intact cells. HEK293 cells overexpressing type V adenylyl cyclase were treated with 50 l forskolin or NKH477 as described in Materials and Methods. The cAMP production is expressed as 3H-cAMP/(3H-cAMP+3H-ATP)×103. All assays were done in duplicate. Means±... from four independent experiments are shown. ∗, P<0.01 differences.

derivatives with a dimethylaminopropionyl modification at R6-position significantly increased the selectivity for the type V isoform and cardiac adenylyl cyclase. The increased selectivity was preserved under different assay conditions, with different samples, and with different preparations. DMAPD and NKH477 stimulated cardiac adenylyl cyclase greater than the other tissue adenylyl cyclases. This agrees with that DMAPD and NKH477 stimulated the type V isoform more potently than the other isoforms in insect cells because the dominant isoform in adult ventricular myocytes is type V at least at the level of mRNA (Ishikawa et al., 1992, 1994; Tobise et al., 1994; Yu et al., 1995). Although we do not know the protein expression of each adenylyl cyclase isoform in tissues because of the lack of antibodies with the requisite sensitivity, the biochemical property of cardiac adenylyl cyclase is most consistent with the known properties of the type V isoform (Ishikawa and Homcy, 1997), and thus it is reasonable to suggest that the type V isoform importantly determines the

property of bAR signaling in the heart (Ishikawa et al., 1994). Other isoforms including types IV and VI have also been detected by either RT-PCR (type IV) or Northern blotting (type VI) in the heart, although their mRNA levels appear to be lower than type V mRNA levels (Ishikawa and Homcy, 1997). We did not examine these isoforms in this study because the catalytic activity in insect cells overexpressing these isoforms was not high enough to become the dominant adenylyl cyclase isoform in insect cells. Type II is a major lung isoform (Ishikawa and Homcy, 1997). Therefore, less potent stimulation of lung adenylyl cyclase than that of cardiac adenylyl cyclase by DMAPD and NKH477 agrees with our findings obtained from type II adenylyl cyclase overexpressed in insect cell membranes. Agents that can directly and selectively activate the type V adenylyl cyclase isoform may mimic super-selective b-agonist in terms of cardiac specificity and will be useful in the treatment of heart failure. In this regard, DMAPD and NKH477 may serve as the prototype of such super-selective compounds. It needs to be determined, however, whether the long-term activation of the b-AR signaling pathway has a favorable clinical response (Lambertz et al., 1984; Packer et al., 1991). The availability of adenylyl cyclase isoform clones has permitted structure-function studies of forskolin derivatives to examine binding and activation of adenylyl cyclase (Robbins et al., 1996; Sutkowski et al., 1994, 1996). Although such efforts have focused on a neuronal adenylyl cyclase isoform (type I), and the findings indicate that the potency of binding and activating adenylyl cyclase may not always be interrelated (Robbins et al., 1996), 6-isothiocyanation of forskolin successfully inhibited the activation of the type I adenylyl cyclase isoform (Sutkowski et al., 1996). Departing from earlier approaches, we have addressed the isoform selective activation by forskolin derivatives using various adenylyl cyclase isoforms in this study. We did not examine the neuronal adenylyl cyclase isoforms (types I/VIII) in this study because these isoforms are not expressed in peripheral tissues including the heart. Putting together, altering the adenylyl cyclase catalytic activity of particular isoforms by forskolin derivatives may operate as an alternative pharmacological therapy to the conventional b-agonist or b-antagonist therapy. Increased selectivity of 6aminopropionyl derivatives for type V in our study has clearly demonstrated the feasibility of developing such drugs. The current study also demonstrated that screening such reagents using molecular clones of adenylyl cyclase isoforms is an operable method.

107

Forskolin and Cardiac Adenylyl Cyclase (a)

(b) 200 cAMP production (pmol/mg/min)

cAMP production (pmol/mg/min)

4000 * 3000

2000

1000

0

Fsk

NKH477

Fsk

NKH477

DMAPD

(d) 300 cAMP production (pmol/mg/min)

200 cAMP production (pmol/mg/min)

100

0

DMAPD

(c)

*

*

100

0

Fsk

NKH477

DMAPD

* * 200

100

0

Fsk

NKH477

DMAPD

Figure 5 Comparison of DMAPD and NKH477 with forskolin in tissues. Mouse tissue homogenates from the brain (a), lung (b), kidney (c), and heart (d) were prepared as described in Materials and Methods. Tissue adenylyl cyclase activity was determined in the presence of forskolin (open bars), DMAPD (solid bars), and NKH477 (striped bars) at 50 l. All assays were done in duplicate, and the means±... from four independent experiments are shown. ∗, P<0.01 differences from the values with forskolin.

Acknowledgment We thank H. Kinoshita (Nippon Kayaku Co., Ltd) for helpful discussion and J. Olivo for editorial work. This study was supported by grants from the US Public Health Service HL59139, HL54895 and the American Heart Association #13-533-945. CS was supported by the Deutsche Forschungsgemeinschaft.

References B HA, R RR, 1990. Identification of a specialized adenylyl cyclase that may mediate odorant detection. Science 250: 1403–1406. B MR, G R, S A, M W, M W, 1984. Pharmacology and inotropic potential of forskolin in the human heart. J Clin Invest 74: 212–223.

C JJ, Z JC, M N, C DMF, K J, 1994. Type VIII adenylyl cyclase. J Biol Chem 269: 12190–12195. D JW, 1984. Forskolin, adenylate cyclase, and cell physiology: an overview. Adv Cyclic Nucleotide Res 17: 81–89. E T, T Y, O N, K J-I, S C, I Y, 1997. Isoform-dependent activation of adenylyl cyclase by proteolysis. FEBS Lett 401: 223– 226. E I, I V, D N, S F, H J, M JJ, 1995. Type V, but not type VI, adenylyl cyclase mRNA accumulates in the rat heart during ontogenic development. Correlation with increased adenylyl cyclase activity. J Mol Cell Cardiol 27: 1789– 1795. F PG, S KA, B HA, T WJ, K J, G AG, R RR, 1991. Molecular cloning and characterization of a Ca2+/calmodulininsensitive adenylyl cyclase from rat brain. Proc Natl Acad Sci USA 88: 10173–10177. H A, A T, H M, H A, H

108

Y. Toya et al.

K, 1993. Effects of a new forskolin derivative, NKH477, on canine ventricular arrhythmia models. J Cardiovasc Pharmacol 22: 847–851. H M, T T, F A, F R, I O, T T, N K, 1992. Cardiovascular and adenylate cyclase stimulant properties of NKH477, a novel water-soluble forskolin derivative. J Cardiovasc Pharmacol 19: 625–634. I Y, H CJ, 1997. The adenylyl cyclases as integrators of transmembrane signal transduction. Circ Res 80: 297–304. I Y, K S, C L, H NJ, K J, H CJ, 1992. Isolation and characterization of a novel cardiac adenylylcyclase cDNA. J Biol Chem 267: 13553–13557. I Y, S S, K K, S RP, K K, K S, V DE, V SF, H CJ, 1994. Down-regulation of adenylylcyclase types V and VI mRNA levels in pacing-induced heart failure. J Clin Invest 93: 2224–2229. I M, I Y, S RP, S Y-T, S N, G P, V DE, V SF, 1996. Neurally mediated cardiac effects of forskolin in conscious dogs. Am J Physiol 271: H1473–1482. J TL, W C, S AM, 1993. Baculovirus expression and purification of a soluble, mutant Gprotein alpha subunit. Protein Expr Purif 4: 64–71. K J, E T, I S, K D, H CJ, I Y, 1994a. A novel peptide inhibitor of adenylyl cyclase (AC): a peptide from type V AC directly inhibits AC catalytic activity. J Biol Chem 269: 24906– 24911. K J, I G, E T, O S, K T, U Y, H CJ, I Y, 1994b. Differential activation of adenylylcyclase by protein kinase C isoenzymes. J Biol Chem 269: 16554–16558. K J, E T, T Y, O N, S C, D E, I Y, 1996. Regulation of type V adenylyl cyclase by PMA-sensitive and -insensitive protein kinase C isoenzymes in intact cells. FEBS Lett 384: 273– 276. L H, M J, E R, 1984. Long-term hemodynamic effects of prenalterol in patients with severe congestive heart failure. Circulation 69: 298–305. L E, D AN, B BK, 1978. Positive inotropic and blood pressure lowering activity of a diterpene derivative isolated from Coleus forskohlii: forskolin. Arzneim Forsch 28: 284–289. M M, T M, T H, H K, H Y, Y M, 1994. Effect of NKH477, a new water soluble forskolin derivative, on arterial-ventricular coupling and mechanical energy transduction in patients with left ventricular systolic dysfunction: comparison with dobutamine. J Cardiovasc Pharmacol 24: 310–316. P M, C JR, R RJ, I RJ, DB R, Z SM, H GH, B WJ, E U, K ML, M GI, S RN, S J, T PK, DM DL, G PS, 1991.

Effect of oral milrinone on mortality in severe chronic heart failure. N Engl J Med 325: 1468–1475. P JP, H A, C J, J O, M RP, I R, 1995. Distinct characteristics of the basal activities of adenylyl cyclases 2 and 6. J Biol Chem 270: 21368–21373. R JD, B DL, T WJ, S R, S KB, 1996. Forskolin carbamates: binding and activation studies with type I adenylyl cyclase. J Med Chem 39: 2745–2752. S Y, 1979. Adenylate cyclase assay. Adv Cyclic Nucleotide Res 10: 35–55. S A, T S, 1995. Effects of NKH477, a watersoluble forskolin derivative, on cardiac function in rats with chronic heart failure after myocardial infarction. J Pharmacol Exp Ther 274: 120–126. S KB, P W, D JW, 1981. Forskolin: unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc Natl Acad Sci USA 78: 3363–3367. S J, S S, I T, K H, 1992. Mechanisms of vasodilation induced by NKH477, a watersoluble forskolin derivative, in smooth muscle of the porcine coronary artery. Circ Res 71: 70–81. S EM, T W-J, B CW, R JD, S KB, 1994. Regulation of forskolin interaction with type I, II, V and VI adenylyl cyclases. Biochemistry 33: 12852–12859. S EM, R JD, T WJ, S KB, 1996. Irreversible inhibition of forskolin interactions with type I adenylyl cyclase by a 6-isothiocyanate derivative of forskolin. Mol Pharmacol 50: 299–305. T T, N A, N K, I G, T T, S M, F A, H M, Y K, E K, S A, 1996. Forskolin derivatives. I. Synthesis, and cardiovascular and adenylate cyclasestimulating activities of water soluble forskolins. Chem Pharm Bull 44: 2274–2279. T R, G AG, 1995. Mammalian membranebound adenylyl cyclases. J Biol Chem 270: 1–4. T K, I Y, H SR, I M-J, N JB, Y H, F M, S EE, H CJ, 1994. Changes in type VI adenylylcyclase isoform expression correlate with a decreased capacity for cyclic AMP generation in the aging ventricle. Circ Res 74: 596– 603. W YH, F A, P AM, Z I, E T, P J, B HR, 1991. Mutant a subunit of Gia2 inhibits cyclic AMP accumulation. Nature 351: 63–65. X Z, C EJ, W F, S DR, 1992. The type III calcium/calmodulin-sensitive adenylyl cyclase is not specific to olfactory sensory neurons. Neurosci Lett 144: 169–173. Y HJ, U JR, G RD, 1995. Determination and cellular localization of adenylyl cyclase isozymes expressed in embryonic chick heart. FEBS Lett 374: 89–94.