The Commonly Used cGMP-dependent Protein Kinase Type I (cGKI) Inhibitor Rp-8-Br-PET-cGMPS Can Activate cGKI in Vitro and in Intact Cells*

Small-molecule modulators of cGMP signaling are of interest to basic and clinical research. The cGMP-dependent protein kinase type I (cGKI) is presumably a major mediator of cGMP effects, and the cGMP analogue Rp-8-Br-PET-cGMPS (Rp-PET) (chemical name: β-phenyl-1,N2-etheno-8-bromoguanosine-3′,5′-cyclic monophosphorothioate, Rp-isomer) is currently considered one of the most permeable, selective, and potent cGKI inhibitors available for intact cell studies. Here, we have evaluated the properties of Rp-PET using cGKI-expressing and cGKI-deficient primary vascular smooth muscle cells (VSMCs), purified cGKI isozymes, and an engineered cGMP sensor protein. cGKI activity in intact VSMCs was monitored by cGMP/cGKI-stimulated cell growth and phosphorylation of vasodilator-stimulated phosphoprotein. Unexpectedly, Rp-PET (100 μm) did not efficiently antagonize activation of cGKI by the agonist 8-Br-cGMP (100 μm) in intact VSMCs. Moreover, in the absence of 8-Br-cGMP, Rp-PET (100 μm) stimulated cell growth in a cGKIα-dependent manner. Kinase assays with purified cGKI isozymes confirmed the previously reported inhibition of the cGMP-stimulated enzyme by Rp-PET in vitro. However, in the absence of the agonist cGMP, Rp-PET partially activated the cGKIα isoform. Experiments with a fluorescence resonance energy transfer-based construct harboring the cGMP binding sites of cGKI suggested that binding of Rp-PET induces a conformational change similar to the agonist cGMP. Together, these findings indicate that Rp-PET is a partial cGKIα agonist that under certain conditions stimulates rather than inhibits cGKI activity in vitro and in intact cells. Data obtained with Rp-PET as cGKI inhibitor should be interpreted with caution and not be used as sole evidence to dissect the role of cGKI in signaling processes.

cGMP is a cyclic-nucleotide second messenger with multiple targets and functions. Small-molecule modulators of cGMP generators and effectors are important biochemical tools as well as established and prospective drugs for the treatment of human diseases, such as erectile dysfunction, pulmonary hypertension, and various cardiovascular disorders (1)(2)(3). cGMP is generated by nitric oxide-or natriuretic peptide-stimulated guanylyl cyclases and can bind to and modulate the activity of at least three classes of cGMP effector proteins: cyclic nucleotide-hydrolyzing phosphodiesterases (PDEs), 2 cyclic nucleotide-gated cation channels, and cGMP-dependent protein kinases (cGKs, also known as protein kinase G or PKG) (4).
Based on pharmacological and genetic studies, the cGK type I (cGKI) is considered the major mediator of cGMP signaling in many tissues including the cardiovascular system (5)(6)(7)(8). The mammalian cGKI is a cytosolic Ser/Thr protein kinase comprising an N-terminal regulatory domain with two cGMPbinding sites and a C-terminal catalytic domain. It exists in two isoforms termed cGKI␣ and cGKI␤. The isozymes have identical cGMP-binding sites and catalytic domains but differ in their N-terminal ϳ100 amino acids, which contribute to homodimerization, sensitivity to cGMP activation, and interaction with anchoring and substrate proteins. Recent in vivo studies with transgenic mice demonstrated that both isoforms can induce smooth muscle relaxation and vasodilation (9), but the respective molecular mechanisms behind these effects are controversial (6,10). Similarly, opposing effects of cGMP/cGKI signaling have been reported on the growth and phenotype of vascular smooth muscle cells (VSMCs) (11,12). The inconsistency of the results concerning the function of cGKI might in part be related to unexpected effects of the pharmacological cGKI activators and inhibitors that are commonly used to distinguish between cGKI-dependent and cGKI-independent signaling. For instance, cGMP analogues can bind to multiple * This work was supported by grants from the Deutsche Forschungsgemeinschaft. 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. 1  cGMP receptors and can elevate the cAMP level by inhibiting the cAMP-degrading PDE3 (8,13,14). Moreover, there is evidence that one of the most frequently used cGKI inhibitors, KT5823, does not at all inhibit cGKI activity both in vitro and in intact cells and may in fact inhibit other protein kinases (15)(16)(17). Consequently, other classes of cGKI inhibitors are increasingly used for intact cell studies. These inhibitors are based on cell-permeable modified peptide substrates (18) or on the allosteric activator cGMP (see Fig. 1A) (19). Agonistic cGMP analogues have been converted to antagonists by exchanging one oxygen atom of the cyclic phosphate moiety with sulfur in the equatorial position with respect to the sugar ring (20). The resulting antagonistic Rp-phosphorothioate cGMP analogues are supposed to bind to the cGMP-binding sites of cGKI without inducing the conformational change crucial for allosteric activation of the enzyme (21,22). For instance, the cGMP analogue 8-Br-PET-cGMP (PET) is a cGKI agonist (23), whereas Rp-8-Br-PET-cGMPS (Rp-PET) (see Fig. 1B) competitively inhibits activation of purified cGKI by cGMP (24). Rp-PET is currently considered one of the most permeable, selective, and potent cGKI inhibitors available for intact cell studies (14,19).
In the present study, we intended to validate the selectivity and efficacy of Rp-PET as a cGKI inhibitor in intact cells by comparing its effects on cGKI-expressing and cGKI-deficient VSMCs. Surprisingly, Rp-PET did not efficiently inhibit but rather stimulated cGKI-mediated processes in VSMCs. In vitro experiments with purified cGKI isozymes and an engineered cGKI-based cGMP sensor protein supported these findings, suggesting that Rp-PET is a partial agonist rather than an antagonist of cGKI␣.
To determine cell growth, primary VSMCs were plated on 96-well culture plates (20,000 cells/well) in the absence or presence of 8-Br-cGMP and/or Rp-PET. After 72 h, the cell number was determined by the MTS assay (CellTiter 96 AQ ueous , Promega) or by staining with toluidine blue O. The MTS assay was performed according to the manufacturer's protocol. Briefly, cells were washed once with serum-free medium to remove non-adherent cells. Subsequently, 20 l of the MTS reagent were added to 100 l of serum-free medium in each well. The A 492 was measured after a 30-and 60-min incubation at 37°C in a humidified, 6% CO 2 atmosphere. For toluidine blue O staining, cells were washed once with phosphate-buffered saline and then fixed and stained for 10 min in 100 l of ice-cold toluidine blue O solution (0.5% (w/v) toluidine blue O in phosphate-buffered saline containing 2% (v/v) formaldehyde and 0.2% (v/v) glutaraldehyde). Excess dye was removed by five washes with phosphate-buffered saline. Stained cells were incubated in 100 l of 1% (w/v) SDS to release the dye and the A 620 was determined.
Phosphorylation of the vasodilator-stimulated phosphoprotein (VASP) was detected by Western blotting via the band shift to a higher apparent molecular weight when VASP is phosphorylated at Ser-157 (8). Primary VSMCs were plated on 6-well culture plates (100,000 cells/well) and grown for 7 days to a confluence of 80 -90%. Subsequently, cells were maintained in serum-free medium for a further 48 h. Then cells were preincubated for 30 min with Rp-PET or vehicle followed by a 30-min incubation in the presence or absence of 8-Br-cGMP. Cells were washed once with phosphate-buffered saline and lysed in lysis buffer (20 mM Tris-HCl, pH 8.0, 0.7% (w/v) SDS, 1.7% (v/v) ␤-mercaptoethanol, 0.2 mM phenylmethylsulfonyl fluoride). Cell lysates were incubated for 5 min at 95°C and used for SDS-PAGE and Western blot analysis with polyclonal rabbit antibodies against VASP (Alexis Biochemicals, catalogue number Alx-210-725, 1:2000), Akt (Cell Signaling Technology, catalogue number 9272, 1:1000), and cGKI (1:5000). The rabbit polyclonal cGKI antiserum, termed "cGKI common (DH)," detects both cGKI␣ and cGKI␤. It was raised against recombinant bovine cGKI␣ and affinity-purified using cGKI␣ coupled to BrCN-Sepharose. The cGKI␣ protein was expressed in Sf9 insect cells and purified as described (27).
In Vitro Kinase Assay-A radioactive assay was used to determine the kinase activity of purified recombinant bovine cGKI␣ and cGKI␤. Both isozymes were expressed in Sf9 insect cells and purified by affinity chromatography (27,28). The phosphorylation reaction was carried out at 30°C in a total volume of 100 l. The reaction mix contained 50 mM Mes, 0.4 mM EGTA, 1 mM magnesium acetate, 10 mM NaCl, 10 mM dithiothreitol, 0.1% (w/v) bovine serum albumin, 0.1 mM ATP (ϳ100 cpm/ pmol [␥-32 P]ATP), 40 M substrate peptide GRTGRRNSI-amide, and various concentrations of cGMP and/or Rp-PET. Reactions were started by adding 10 ng of purified enzyme. After 5 min, 80 l of the reaction mix were spotted onto Whatman P81 phosphocellulose paper (2.5 ϫ 3.0 cm). Then the filter papers were washed three times for 10 min in 85 mM phosphoric acid, dried, and put into scintillation vials to measure 32 P incorporation. Activity was calculated as mol of phosphate transferred per minute and mg of kinase. K a values for reaching half-maximal activity were determined from the inflection points of the activation curves. K i values for inhibition of the enzyme to halfmaximal activity were determined by Dixon plots (29).
Electrospray Ionization Mass Spectrometry (ESI-MS)-Aqueous Rp-PET and PET stock solutions (10 mM, sodium salt) were stored at Ϫ20°C. Before MS analysis, they were allowed to thermally equilibrate at room temperature for 30 min. For ESI-MS, stock solutions were diluted with water/acetonitrile (50:50, v/v) to a final concentration of 0.5 mM. For spiking, PET was added to the Rp-PET aliquot in a 30:70 (v/v) ratio. Solutions were infused (4 l/min) into the ESI source using a Parmer Infusion 74900 series syringe pump (Cole-Parmer Instrument Co.). Mass spectra were acquired in the negative ion mode using a HCT Plus ion trap mass spectrometer equipped with a standard ESI source (Bruker-Daltonics). Spectra (50 -1000 m/z) were acquired in the standard enhanced mode (scan rate 8100 m/z per second). Dry gas (5 liters/min) temperature was set to 300°C, the nebulizer was set to 10.0 p.s.i., and the electrospray voltage was set to 4000 V. Maximal accumulation time was set to 200 ms. Loading of the trap was controlled by the instrument (ICC 70000).
Fluorescence Resonance Energy Transfer (FRET) Measurements-A fusion protein (cGi-500, Ref. 30) consisting of the tandem cGMPbinding domains of bovine cGKI sandwiched between the cyan and yellow fluorescent proteins CFP and YFP, respectively, was transiently expressed in HEK-293 cells using the FuGENE 6 transfection reagent according to the instructions of the manufacturer (Roche Applied Science). 1 ϫ 10 7 cells were lysed in homogenization buffer (25 mM triethanolamine/HCl, pH 7.4, containing 2 mM dithiothreitol and a 100-fold dilution of protease inhibitor mixture, Sigma-Aldrich) by sonication (1 pulse, 5 s), and a cytosolic fraction was obtained (100,000 ϫ g, 40 min, 4°C). Fluorescence measurements were performed on a Cary eclipse spectrofluorometer equipped with a microplate accessory (Varian Inc., excitation at 436 nm) in white half-area microplates (Greiner, catalogue number 675075) using 5 l of the cytosol containing the indicator in a total volume of 100 l of buffer A (25 mM triethanolamine/HCl, pH 7.4, 2 mM dithiothreitol, 10 mM MgCl 2 ). Concentration-response curves for the compounds were assessed by recording of CFP and YFP emissions at 475 and 525 nm, respectively, for 5 min, subtracting the background emission of a water-filled well and calculating the emission ratio of 475 to 525 nm.

RESULTS AND DISCUSSION
The effects of Rp-PET (Fig. 1B) on intact cells were studied in murine primary aortic VSMCs, which express both cGKI␣ and cGKI␤ (31). VSMCs obtained from control or cGKI-deficient mice (25) were compared. This cell culture system has been proven useful to identify cGKI-dependent functions. Previous studies have shown that the stimulation of cell growth and VASP phosphorylation by the membrane-permeable cGMP analogue 8-Br-cGMP is indeed mediated via activation of cGKI  's t test). Similar results were obtained by the MTS assay (data not shown). Phosphorylation of VASP was measured after a 30-min preincubation with or without Rp-PET followed by 30 min with or without 8-Br-cGMP. Phospho-VASP (p-VASP) was monitored by immunodetection of the band shift to a higher apparent molecular weight when VASP is phosphorylated at Ser-157 (8). Staining of cGKI confirmed the presence and absence of the kinase in control and cGKI-deficient cells, respectively. Akt was used as a loading control. rel. cell number, relative cell number. (31)(32)(33). Therefore, it was tested whether Rp-PET could inhibit cGMP/cGKI-stimulated VSMC growth and phosphorylation of VASP. As expected, the cGKI agonist 8-Br-cGMP (100 M) induced cell growth and VASP phosphorylation in cGKI-expressing cells (Fig. 2, left panels) but was ineffective in cGKIknock-out cells (Fig. 2, right panels). Surprisingly, Rp-PET (100 M) had no significant effect on 8-Br-cGMP (100 M)-induced growth and phospho-VASP. When applied under basal conditions, i.e. in the absence of 8-Br-cGMP, Rp-PET even stimulated cell growth and VASP phosphorylation in control VSMCs (Fig. 2, left panels). Growth stimulation by Rp-PET alone was reproducible and highly significant, but clearly weaker than with 8-Br-cGMP. Importantly, Rp-PET had no effects on cGKI-deficient cells (Fig. 2, right panels), demonstrating that its apparent agonistic activity in control VSMCs was indeed mediated by cGKI.

Rp-8-Br-PET-cGMPS Can Activate cGKI
To further investigate the potential partial agonistic activity of Rp-PET, its effects on the activity of purified cGKI␣ and cGKI␤ were examined. Both enzymes displayed cGMP-dependent kinase activity with characteristic K a values for stimulation with cGMP (K a ϳ0.1 and ϳ1 M for cGKI␣ and cGKI␤, respectively; Fig. 3, no inhibitor, and Fig. 4). Increasing concentrations of Rp-PET caused a right shift of the cGMP activation curves for both the cGKI␣ isoform (Fig. 3A, upper panel) and the cGKI␤ isoform (Fig.  3B, upper panel), as expected for an inhibitor. By using Dixon plots (29), K i values of 0.03 and 0.05 M were determined for cGKI␣ (Fig. 3A, lower panel) and cGKI␤ (Fig. 3B, lower panel), respectively. These K i values were similar to K i values reported previously for the inhibition of the cGKI isozymes by Rp-PET in vitro (24). Interestingly, inhibition of cGMP-activated cGKI␣ by Rp-PET was less complete than inhibition of cGKI␤ (Fig.  3, upper panels, black arrows). Moreover, when added in the absence of cGMP, Rp-PET appeared to increase the activity of cGKI␣ but not cGKI␤ above basal levels (Fig. 3,  upper panels, broken arrows). These findings were consistent with the partial agonistic effect of Rp-PET observed in intact cells and suggested that Rp-PET might preferentially activate the cGKI␣ isozyme. Indeed, Rp-PET alone increased the kinase activity of cGKI␣ in a concentration-dependent manner with a K a value of 1 M (Fig. 4A). In line with a partial agonistic activity, the maximal kinase activity that could be induced with Rp-PET was lower than  with cGMP, reaching ϳ38% of the enzyme activity in the presence of saturating cGMP concentrations. In contrast, Rp-PET did not significantly alter the kinase activity of the cGKI␤ isoform (Fig. 4B).
It seemed possible that the apparent activation of purified cGKI␣ by Rp-PET was caused by spontaneous exchange of the sulfur atom of the cyclic phosphorothioate moiety of Rp-PET with oxygen, resulting in conversion of the Rp-analogue to the corresponding agonistic compound, PET ( Fig. 1B; see Technical Information about Rp-8-Br-PET-cGMPS, update October 15, 2007, Biolog Life Science Institute). To exclude this possibility and to analyze the homogeneity of the Rp-PET compound, the aliquot used for the kinase assays was examined by ESI-MS. Only peaks at m/z ratios corresponding to the molecular mass of Rp-PET could be detected (Fig. 5, left panel), indicating that the compound was pure and that exchange of sulfur with oxygen or other chemical alterations had not occurred. Spiking of Rp-PET with the potential conversion product PET, which is 16 Da lighter, confirmed that these compounds could clearly be resolved by the experimental setup (Fig. 5, right panel). Thus, the mass spectrum confirmed that Rp-PET itself exerted the partial agonistic effect on cGKI␣ that was observed in the kinase assays.
To further verify its partial agonistic activity, the effects of Rp-PET on the isolated cGMP-binding domains of cGKI were studied by FRET measurements. For these measurements, the two cGMP-binding sites of bovine cGKI were sandwiched between CFP and YFP, respectively (30). This construct contains only the cGMP-binding sites, which are identical in cGKI␣ and cGKI␤ but lacks the N-terminal region that differs between the isozymes. Binding of the agonist cGMP to the FRET construct induces a conformational change that can be monitored as altered FRET signal (30), whereas binding of an antagonist does not produce a FRET change. In line with a partial agonistic activity of Rp-PET, increasing concentrations of Rp-PET caused a similar, albeit weaker, FRET change as the known agonists cGMP and PET (Fig. 6).
The combined results of the in vitro kinase assays with purified cGKI isoforms (Figs. 3 and 4) and of the FRET measurements (Fig. 6) suggest the following model. Rp-PET binds to the cGMP-binding sites of both cGKI isozymes. Although it acts as an antagonist of cGKI␤, its binding to cGKI␣ induces a conformational change that partially activates the enzyme. To confirm that Rp-PET activates cGKI␣ but not cGKI␤ in intact cells, growth assays were performed with so-called SM-I␣ or SM-I␤ smooth muscle rescue VSMCs, which express only the cGKI␣ or the cGKI␤ isoform, respectively (9). Indeed, Rp-PET (100 M) stimulated the growth of SM-I␣ but not SM-I␤ cells (Fig.  7). 8-Br-cGMP (100 M) was effective in both cell preparations,   demonstrating that, in principle, both cGKI isozymes can promote VSMC growth.
The differential sensitivity of cGKI␣ versus cGKI␤ to the partial agonistic effect of Rp-PET must be related to their different N termini, which apparently differ in their ability to couple ligand binding to activation of the catalytic region. Indeed, the N termini determine the differential sensitivity of the isozymes to cGMP activation, the cGKI␣ isoform being ϳ10 times more sensitive to cGMP than cGKI␤ (34). Target-specific effects have also been reported for the Rp-phosphorothioate cGMP analogue Rp-8-pCPT-cGMPS, which is an antagonist of the olfactory cyclic nucleotide-gated channel but an agonist of the photoreceptor cyclic nucleotide-gated channel (35). Because Rp-8-pCPT-cGMPS is also used as cGK inhibitor (36), we tested its effect on cGKI activity in the absence of cGMP. Similar to Rp-PET, Rp-8-pCPT-cGMPS did not alter the basal kinase activity of cGKI␤ but partially activated the cGKI␣ isozyme to ϳ35% of the maximal cGMP-stimulated activity with a K a value of 1 M (data not shown). In the original reports on the use of Rp-8-pCPT-cGMPS (36) and Rp-PET (24) as cGK inhibitors, the effects of these compounds on cGKI activity in the absence of cGMP were not determined. However, in line with the present results, other investigators have noticed partial agonistic effects of both Rp-PET and Rp-8-pCPT-cGMPS on cGKI (37,38). Partial agonistic activity has also been reported for Rp-phosphorothioate cAMP analogues that are used as inhibitors of the cAMP-dependent protein kinase (39).
As a partial agonist, Rp-PET should, in principle, be able to inhibit cGMP-activated cGKI in intact cells. The failure of our initial attempts to inhibit cGKI-mediated VSMC growth and VASP phosphorylation in the presence of 100 M 8-Br-cGMP by an equimolar concentration of Rp-PET (Fig. 2) might be related to an inappropriate intracellular ratio of 8-Br-cGMP to Rp-PET. Indeed, when the nominal 1:1 ratio was decreased to 1:4, i.e. 50 M 8-Br-cGMP and 200 M Rp-PET, cGMP/cGKImediated cell growth and VASP phosphorylation were significantly inhibited by Rp-PET (Fig. 8). These results are consistent with a previous study showing that Rp-PET can inhibit cGMPactivated cGKI in intact cells under certain conditions (24). Considering the 46-fold higher lipophilicity (14) and the 4-fold higher concentration of Rp-PET versus 8-Br-cGMP used in our experiment, it appears that the inhibitory potential of Rp-PET in intact cells is quite moderate and that it must be present in large excess to inhibit the effect of a strong agonist such as 8-Br-cGMP.
Taken together, the present study indicates that the cGMP analogue Rp-PET, which is frequently used as cGKI inhibitor, is a cGKI␤ antagonist and partial cGKI␣ agonist and, therefore, affects cGKI activity in a complex isoform-dependent manner. In intact cells, Rp-PET may either inhibit or activate cGKImediated pathways depending on the intracellular cGMP level and the prevalence of the cGKI␣ or cGKI␤ isoform. Other studies have shown that Rp-PET can also exert cGKI-independent effects (40,41), perhaps via modulation of PDEs (3,14). The unpredictable behavior of Rp-PET complicates the interpretation of intact cell studies. Rp-PET should be used with caution as cGKI inhibitor and not as sole proof for the involvement of cGKI in signaling pathways.