• cyclic nucleotide-dependent protein kinases;
  • penile erection;
  • sildenafil;
  • vascular smooth muscle


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

To date, relative cellular levels of cGMP and cGMP-binding proteins have not been considered important in the regulation of smooth muscle or any other tissue. In rabbit penile corpus cavernosum, intracellular cGMP was determined to be 18 ± 4 nm, whereas the cGMP-binding sites of types Iα and Iβ cGMP-dependent protein kinase (PKG) and cGMP-binding cGMP-specific phosphodiesterase (PDE5) were 58 ± 14 nm and 188 ± 6 nm, respectively, as estimated by two different methods for each protein. Thus, total cGMP-binding sites (246 nm) greatly exceed total cGMP. Given this excess of cGMP-binding sites and the high affinities of PKG and PDE5 for cGMP, it is likely that a large portion of intracellular cGMP is associated with these proteins, which could provide a dynamic reservoir for cGMP. Phosphorylation of PDE5 by PKG is known to increase the affinity of PDE5 allosteric sites for cGMP, suggesting the potential for regulation of a reservoir of cGMP bound to this protein. Enhanced binding of cGMP by phosphorylated PDE5 could reduce the amount of cGMP available for activation of PKG, contributing to feedback inhibition of smooth muscle relaxation or other processes. This introduces a new concept for cyclic nucleotide signaling.


erythro-9(2-hydroxy-3-nonyl)adenine hydrochloride




cAMP-dependent protein kinase


cGMP-dependent protein kinase





Cyclic GMP plays a key role in the regulation of the contractile state of smooth muscle tissues such as that in the corpus cavernosum of the penis [1–4]. Release of nitric oxide from nerves causes elevation of cGMP in this tissue, which produces relaxation of smooth muscle and accumulation of blood, resulting in penile erection. Inadequate blood flow or nerve function can compromise this process, resulting in male erectile dysfunction. New drugs such as sildenafil (Viagra™) maintain higher levels of cGMP by inhibiting phosphodiesterase-5 (PDE5), an enzyme that degrades cGMP [5]. In addition to its catalytic site that is specific for degradation of cGMP, PDE5 contains allosteric (noncatalytic) binding sites that are highly specific for cGMP [4].

Practically all mammalian signaling pathways contain mechanisms for negative feedback control. This counterbalances the initial signal by dampening the response to this signal and facilitating termination of the signal. Most cAMPsignaling pathways are regulated in this manner. When extracellular signals increase cAMP levels, several negative feedback mechanisms are activated. Some of these involve PDEs. First, cAMP degradation by PDEs is stimulated by mass action, i.e. by increased substrate availability [6]. Second, cellular PDE3 and PDE4 activities are increased acutely by cAMP activation of cAMP-dependent protein kinase (PKA), which phosphorylates these enzymes [7–10]. Third, PDE4 activity is increased by chronic elevation of cAMP, resulting in increased PDE4 gene transcription [11,12].

It is likely that cGMP-signaling pathways as well as cAMP-signaling pathways are regulated by negative feedback. The subject of the present paper is the possibility that overall availability of cellular cGMP could be reduced by being sequestered in the allosteric cGMP-binding sites of PDE5 and it would therefore be unavailable to target proteins such as cGMP-dependent protein kinase (PKG). Intracellular receptors for cGMP in body tissues include PKG, cGMP-binding PDEs, and cGMP-gated channels [13], although the latter may be present at relatively low levels. PKA may act under some conditions as a cGMP receptor, a process known as cross-activation [14]. PKG and PDE5 are known to be present in penile corpus cavernosum. Each contains allosteric cGMP-binding sites. If smooth muscle cells of corpus cavernosum are rich in PDE5, allosteric sites of this enzyme could compete with those of PKG, thereby effectively sequestering a large fraction of cellular cGMP, which would reduce the extent of PKG activation. We have shown that under basal conditions the concentration of PKG in pig coronary arteries exceeds the concentration of cGMP [15], but levels of PDE2 and PDE5 were not considered in that study. Our recent evidence indicates that phosphorylation of PDE5 substantially increases the cGMP-binding affinity of its allosteric sites [6]. After phosphorylation, affinity of PDE5 for cGMP is approximately that of PKG. Increased sequestration of cGMP by PDE5 when PKG becomes activated by this nucleotide in cells would provide negative feedback control of cGMP-stimulated processes. This effect would act in concert with the increase in PDE5 catalytic activity caused by phosphorylation of this enzyme [6] to produce dampening of the signaling pathway. In addition, unknown modulators may further modulate dissociation of cGMP from PDE5, thereby controlling its availability to activate PKG or other cyclic nucleotide receptors.

It has been determined that a major portion of photoreceptor cGMP is tightly associated with the allosteric sites of PDE6, an enzyme present at high concentrations in retinal rod outer segment [16]. Thus, the allosteric cGMP-binding sites of PDE6 comprise a major reservoir for cGMP in that tissue by providing a sequestration site for cGMP which effectively lowers the concentration of free cGMP in the photoreceptor. If PDE5 or any other protein in corpus cavernosum functions effectively as a sequestration site for cGMP, it would be expected that the molar amount of this protein would approach that of cGMP. The present study establishes the quantitative cGMP-binding potential of PKG and PDE5 in corpus cavernosum. Results suggest that PDE5 could bind a significant portion of cGMP in this tissue. While bound, this portion of the cellular cGMP pool would be a relatively inactive fraction, or sequestered form, of total cellular cGMP. Such a reservoir of bound cGMP may be modulated, particularly after cGMP elevation and enzyme phosphorylation, and could represent an important cellular regulatory mechanism not previously recognized. Consideration of relative concentrations of cyclic nucleotides and their intracellular receptors and the impact of these relationships on signaling pathways is a novel concept that has not been addressed experimentally. This report represents an initial step in that direction.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Preparation of extract and DEAE chromatography

Corpus cavernosum was dissected from the penis of New Zealand White male rabbits. Crude extracts of the corpus cavernosum were prepared by suspending the tissue (≈ 2.5 g) in 4 vol. of buffer and homogenizing in 10 mm potassium phosphate buffer (pH 6.8) containing 5 mm EDTA, 5 mm 2-mercaptoethanol (KPEM), 100 µm 3-isobutyl-1-methylxanthine (IBMX), and 100 µm microcystin. Homogenization was carried out for 2 cycles of 30 s each on ice with 1 min pause on ice between cycles using Ultra-Turrax (Texmar Co., Cincinnati, OH, USA) connected to a variable rheostat set at 80% maximum output. The homogenate was centrifuged at 10 000 g for 30 min at 4 °C. The supernatant was diluted in 4 vol. of cold, MilliQ water (Millipore, Bedford, MA, USA) and chromatographed on a DEAE-Sephacel column (8 cm × 0.9 cm) equilibrated in KPEM buffer but without IBMX and microcystin. The pressure head was 45 cm and the flow rate was ≈ 25 mL·h−1. The column was washed with 20 mL of the same buffer containing 20 mm NaCl, and a 60-mL linear gradient of 20–280 mm NaCl in buffer was used to elute the bound proteins. Aliquots from fractions of 0.9 mL were analyzed.

cGMP-dependent protein kinase assay

Activity of PKG was measured as described [17], a method that utilizes phosphotransfer of [32Pi] from [γ-32P]ATP into a synthetic heptapeptide (RKRSRAE; Peninsula Laboratories Inc., Belmont, USA)[18]. The typical assay consisted of 50 µL of an assay mixture comprising 20 mm Tris, pH 7.4, 200 µm ATP (≈ 30 000 c.p.m. per µL assay mixture), 136 µg·mL−1 of the phosphoreceptor heptapeptide, 20 mm magnesium acetate, 100 µm IBMX, 1 µm synthetic peptide inhibitor (Peninsula Laboratories Inc.) of PKA [19] and 10 µL of cGMP (50 µm) or water as indicated. To this mixture, 10 µL of the DEAE fraction was added, and the reaction was incubated at 30 °C for 10 min. The reaction was terminated by spotting 40 µL of the reaction mixture on P 81 phosphocellulose paper (Whatman) as described above.

PDE assays

PDE activity was measured following a modification of assay conditions for PDE2 [20]. Briefly, reactions were initiated by addition of DEAE fractions to a reaction mixture containing 40 mm Mops (pH 7), 0.8 mm EGTA, 15 mm magnesium acetate, 2 mg·mL−1 bovine serum albumin (Sigma), either [3H]cGMP or [3H]cAMP (Amersham) (80 000–150 000 cpm per assay tube) and varying concentrations of unlabeled cGMP or cAMP, respectively. Reactions mixtures were incubated at 30 °C for 15 min or an appropriate time. A mixture with the following ingredients was added to terminate the reaction: 50 mm EDTA, 30 mm theophylline (Sigma), 10 mm cGMP and 10 mm cAMP in 100 mm Tris (pH 7.5). Snake venom 5′-nucleotidase (Sigma) (200 µg) was added to the assay mixture and incubated for 10 min at 30 °C. Assay samples were diluted in 1 mL of dilution solution (0.15 mm EDTA containing 100 µm each of adenosine and guanosine) and applied to QAE-Sephadex columns (8 × 33 mm) pre-equilibrated in 20 mm ammonium formate buffer (pH 7.4). Eluates were collected and the amount of [3H]guanosine or [3H]adenosine measured by counting. In some experiments, the substrate was 0.4 µm cGMP or cAMP. In experiments involving inhibitors or activators, appropriate amounts were added prior to initiation of the reaction.

CGMP-binding assay

The procedure was modified slightly from that described previously [6]. DEAE fraction, 25 µL, was added to 50 µL of a mixture containing 20 mm sodium phosphate (pH 6.8), 0.2 µm[3H]cGMP, 1 mM EDTA, 25 mm 2-mercaptoethanol, and 0.5 mg·mL−1 Type VIII-S histone (Sigma). After 60 min at 4 °C, 1.8 mL cold potassium phosphate (pH 6.8) was added and the sample was filtered onto premoistened Millipore HAWP filters (pore size 0.45 µm), which was then rinsed two times with the cold phosphate buffer, dried, and counted.

Calculations of intracellular concentrations

Catalytic or cGMP-binding activities of enzymes in DEAE fractions were determined to compute the total activities. The DEAE chromatography was used to resolve the multiple peaks of cGMP-binding and kinase activities, as well as the various cAMP-PDE and cGMP-PDE activities, and to minimize effects of factors present in the crude extract. Catalytic activities were converted to molar amounts by assuming a specific PKG activity of 2.5 µmol·min−1·mg−1 with subunit Mr 78 000 [21], and, for PDE5, a specific enzyme activity of 5 µmol·min−1·mg−1 with subunit Mr 100 000 [22]. The catalytic or allosteric sites of PDE5 are highly conserved among mammalian species. Using a similar DEAE chromatographic separation of cellular proteins described in this report, we have determined that the ratio of catalytic activity to cGMP-binding activity is approximately the same for PDE5 from rabbit, dog, and human corpus cavernosum, as well as for homogeneous bovine PDE5. The [3H]cGMP-binding activities were also converted to molar amounts. This calculation was not dependent on bovine protein-specific cGMP-binding activity. [3H]cGMP-binding assays were performed at subsaturating cGMP (0.2 µm). Under the conditions of this particular assay and using saturating levels of [3H]cGMP, purified PKG binds only one cGMP per monomer [22]. At 0.2 µm[3H]cGMP, the purified enzyme binds 0.27 mol cGMP per monomer, i.e. 3.7-fold less cGMP. This value for purified PKG was used to correct cGMP binding in the PKG fractions from the DEAE fractions. It was assumed that PKG binds 1 mol cGMP per subunit at saturation. For PDE5, pmol [3H]cGMP bound per ml was multiplied by ten to correct for cGMP binding of purified PDE5 performed under the same subsaturating concentration as compared with cGMP binding at saturation. Based on results obtained with purified PDE5, it was assumed that PDE5 binds 1 mol cGMP per subunit at saturation.

Total mol enzyme subunit (based on the catalytic activities) eluting from the DEAE chromatography was converted to mol subunit per g tissue by dividing that total activity by the weight of tissue used to obtain the supernatant applied to the column. Intracellular concentration of subunit was calculated assuming the tissue contained 0.5 mL intracellular water per ml tissue and that recovery from the column was quantitative. This same intracellular water concentration was assumed earlier for calculating protein and cyclic nucleotide levels in pig coronary artery smooth muscle [15].


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Identification of cGMP-PDEs and cAMP-PDEs after DEAE chromatography of rabbit corpus cavernosum extracts

In crude corpus cavernosum extracts, less than 10% of total PDE (substrate being 0.4 µm or 40 µm[3H]cGMP or [3H]cAMP), or PKG activity, was present in resuspended particulate fraction after centrifugation. After DEAE chromatography of the supernatant fraction, PDE activities in fractions were measured using [3H]cGMP or [3H]cAMP as substrate. At low cGMP substrate concentration (0.4 µm), a single major peak of PDE activity (peak 1) eluted at ≈ 120 mm NaCl and a relatively minor peak eluted at ≈ 250 mm NaCl (peak 2) (Fig. 1). This agrees with the recent findings of Qiu et al. [23]. No other cGMP or cAMP hydrolytic activity was detected using this substrate concentration even when the gradient was extended to 800 mm NaCl (data not shown). In five different DEAE preparations, the level of peak 2 varied from one-twentieth to one-third the level of peak 1. At higher cGMP (40 µm), there was a tenfold increase in activity of peak 1 compared with a 100-fold increase of peak 2 (not shown). This fold difference most likely reflects the relative activities at subsaturating and saturating cGMP substrate concentrations for the two enzymes.


Figure 1. PDE profile using low (0.4 µm) cGMP as substrate after DEAE-Sephacel chromatography of crude extract of rabbit corpus cavernosum. Extract was prepared and chromatographed on DEAE-Sephacel as described in Experimental procedures. Fractions, 0.9 mL, were collected and aliquots of each fraction were assayed for cGMP-PDE using 0.4 µm cGMP as substrate as described in Experimental procedures. Results are representative of experiments performed using five different rabbits.

Download figure to PowerPoint

To ascertain substrate specificity for hydrolysis, a similar analysis using cAMP as substrate was performed on the DEAE fractions (not shown). At 0.4 µm cAMP, a minor peak of cAMP-PDE activity coeluted with peak 1, and the activity in this peak was 240-fold higher when high cAMP (40 µm) was used as substrate. Thus, using low substrate (0.4 µm), cGMP hydrolytic activity was 10–16-fold higher than that for cAMP. Though cAMP hydrolytic activity was low compared with that for cGMP, cAMP hydrolytic activity in peak 1 represented the majority of cAMP hydrolytic activity in the entire profile when using a low, nearly physiological concentration of cAMP. Further analysis revealed that the cGMP and cAMP hydrolytic activities of peak 1 were attributed to a mixture of PDEs in these fractions (see below). Using low substrate concentration, a second peak of cAMP-PDE activity was also detected which coeluted with peak 2. In this case, PDE activity was almost twice as high using cGMP than when using cAMP, whereas at high substrate, PDE activity was fourfold higher with cGMP than with cAMP. Thus, most of the PDE activity in peak 1 was due to a cGMP-specific PDE. This peak could be slightly contaminated by a cAMP-PDE or dual-specificity PDE. Activity in peak 2 represented a dual-specificity PDE.

In order to identify the PDEs in the activity profiles, fractions from peak activities were further analyzed using specific inhibitors. Sildenafil, erythro-9(2-hydroxy-3- nonyl)adenine hydrochloride (EHNA), and cilostamide were chosen because of their selectivities and potencies for PDE5, PDE2 and PDE3, respectively. As shown in Table 1, using low cGMP as substrate, peak 1 (peak fraction) PDE activity exhibited a sildenafil IC50 of ≈ 4 nm. However, with cAMP as substrate, the IC50 increased to 11 nm. Neither EHNA nor cilostamide inhibited the cGMP hydrolytic activity of this fraction. These results indicated that this fraction contains predominantly PDE5.

Table 1.  Effects of PDE inhibitors on cAMP-PDE and cGMP-PDE activities in fractions of DEAE chromatography from Fig. 1. cAMP and cGMP concentrations were 0.4 µm. Representative fractions are peak 1, leading edge (e.g. fraction 33), peak fraction (e.g. fraction 37), and trailing edge (e.g. fraction 41). Peak 2 is, e.g. fraction 55. ND, not determined.
 Inhibitors IC50
 Sildenafil (nm)EHNA (µm)Cilostamide (µm)
Peak 1 (leading edge)10011NDNDNDND
Peak 1 (peak fraction)114ND100ND8
Peak 1 (trailing edge)> 100> 100NDNDNDND
Peak 2 (peak fraction)> 1000> 100010ND8ND

Using low cGMP, activity in peak 2 (peak fraction) was insensitive to sildenafil, although it possessed both cGMP-PDE and cAMP-PDE activities to varying degrees using either low or high substrate concentration, with cAMP being the slightly preferred substrate. These results suggested that activity in this fraction is primarily PDE2. This was confirmed by the finding of a tenfold increase in activity when 1 µm cGMP was added to the low cAMP-PDE assay (Fig. 2.), by inhibition with the PDE2-specific reagent, EHNA, and by lack of inhibition with cilostamide (Table 1). As PDE2 is known to possess a much higher specific enzyme activity than does PDE5 [20], the activity in peak 2 (PDE2) represents a very low molar amount (< 2%) of enzyme as compared with the activity in peak 1 (PDE5). Cilostamide-sensitive PDE3 activity was not detected in these experiments. In some preparations the NaCl gradient was extended from 280 mm to 800 mm at the end of the initial chromatography, but no detectable peak of PDE (substrate being low or high cGMP or cAMP) was detected in these fractions.


Figure 2. Effect of cGMP on cAMP hydrolysis of selected DEAE fractions. Aliquots of DEAE fractions were assayed for PDE activity using low (0.4 µm) cAMP as substrate in the absence and presence of 1 µm cGMP. Results are reported as standard error of the mean, n = 6.

Download figure to PowerPoint

The finding that at low substrate concentrations cGMP-PDE and cAMP-PDE activities did not exactly coelute suggested that peak 1 contained two different PDEs, with the cAMP-PDE eluting slightly ahead of the cGMP-PDE (not shown). This interpretation was also supported by somewhat higher (11 nm) sildenafil IC50 for the low cGMP-PDE activity in the leading edge of peak 1 than for the IC50 (4 nm) for this activity in the peak fraction (Table 1). We explored the possibility that the leading edge of peak 1 (e.g. fraction 33) represented a Ca2+/calmodulin-dependent PDE1. Addition of Ca2+/calmodulin to this fraction produced a six-fold increase in low cAMP-PDE activity (Fig. 3). Given these results and the elution position of this fraction, it is concluded that a small amount of PDE1 is present at the leading edge of peak 1. The low cAMP-PDE activity of peak 1 is probably primarily due to the presence of PDE1 in the leading half of the peak and by PDE5 in the trailing half of the peak. The known high specific enzyme activity of PDE1 [24] dictates that the molar amount of this enzyme in corpus cavernosum is quite low, but the enzyme could account for a significant proportion of the total cGMP hydrolytic activity in this tissue.


Figure 3. Effect of Ca2+/calmodulin on selected DEAE fractions. Aliquots of DEAE fractions were assayed for PDE activity using low (0.4 µm) cAMP as substrate in the absence and presence of 4 mm Ca2+ and 1 µm calmodulin. Results are reported as standard error of the mean, n = 3.

Download figure to PowerPoint

Identification of PKGs after DEAE chromatography of rabbit corpus cavernosum extracts

Using synthetic heptapeptide as phospho-acceptor, PKG activity in the presence of cGMP was measured in the DEAE fractions (Fig. 4). PKG activity exhibited two peaks (peaks 1 and 2), which showed cGMP dependency of fourfold and eightfold, respectively (not shown). These peaks exhibited the classical cGMP-dependency of PKG Iα and PKG Iβ isoforms and eluted in the same positions of these isoforms from other tissues as determined using the same technique by this laboratory [25]. To confirm the identity of these two isoforms, fractions from each peak were analyzed alongside native PKG Iα and Iβ using cGMP and two isoform-selective cGMP analogs (Table 2). The Ka values for cGMP of both peaks 1 and 2 were very similar to those of purified bovine lung PKG Iα and bovine aorta PKG Iβ, respectively, with the Ka for PKG Iα being significantly lower than that for PKG Iβ. Using 8-butyryl cGMP and 8-p-chlorophenylthio-cGMP, the ratios Ka(cGMP)/Ka(analog) for peak 1 and peak 2 were also very similar to these ratios for purified PKG Iα and PKG Iβ. The Iα/Iβ ratio in the DEAE profile ranged from 0.50 to 0.92 in four experiments. There appears to be species variation in the relative proportion of PKG Iα and PKG Iβ in corpus cavernosum. In dog tissue, total PKG was similar to that in rabbit tissue, but the Iα/Iβ ratio was < 0.1 (Brent Thompson, unpublished results).


Figure 4. PKG kinase profile after DEAE-Sephacel chromatography of crude extract of rabbit corpus cavernosum. Fractions are the same as those used in Fig. 1. Protein kinase activity was determined in the presence of cGMP as described in Experimental procedures. Results are representative of ten experiments.

Download figure to PowerPoint

Table 2. Effects of cGMP analogs on peaks 1 and 2 PKG activity from DEAE chromatography. Identification as PKG Iα and PKG Iβ. Peak fractions were used for assays. Purified PKG Iα and PKG Iβ were used as controls.
  cGMP Ka (µM) K a(cGMP)/Ka(analog)
Purified PKG Iα0.061.01.9
Purified PKG Iβ0.380.31.0
DEAE peak
DEAE peak 20.660.20.9

Identification of cGMP-binding proteins after DEAE chromatography of extracts of rabbit corpus cavernosum

In order to confirm the identities of PDE5 and PKG isoforms, as well as provide independent measurements of enzyme concentrations, cGMP-binding activity was determined in the fractions. The cGMP-binding profile is shown in Fig. 5 (open symbols). As both PKG and PDE5 contain allosteric cGMP-binding sites, and PKA also could bind cGMP, the profile may represent several overlapping activities. In order to block [3H]cGMP binding to PKG and PKA, a combination of PKG-specific unlabeled 8-chlorophenylthio-cGMP (CPTcGMP) and PKA-specific unlabeled cAMP were added to the binding assay. It can be seen that this addition totally blocked the second peak of cGMP-binding activity and partially blocked the trailing edge of the first peak of activity. It was assumed that most of the cGMP-binding activity blocked by this combination of cyclic nucleotide analogs was that of PKG, and that the remaining cGMP-binding activity was that of PDE5. These interpretations were borne out by plotting cGMP-binding data in Fig. 5 versus PKG kinase activity or low cGMP-PDE activity. Figure 6 shows cyclic nucleotide analog inhibited activity (subtraction of the two curves in Fig. 5) plotted along with PKG activity. The two peaks of corrected cGMP-binding activity correlated with the two peaks of PKG activity. This would be expected for the presence of both the Iα and Iβ isoforms of PKG. The Iα isoform appeared to have a slightly higher binding/kinase ratio. Molar concentrations of PKG in peak fractions of Iα and Iβisoforms were calculated separately using either cGMP-binding or PKG activities. By binding activity, the molar concentration of PKG Iα was 4.4 nm, and it was 1.4 nm by kinase activity; PKG Iβ was 2.6 nm by binding activity and 1.4 nm by kinase activity. When cGMP-binding activity remaining after addition of the cAMP/CPTcGMP combination (closed symbols in Fig. 5) was plotted along with low cGMP-PDE activity, the profiles in Fig. 7 were obtained. It can be seen that the corrected cGMP-binding activity coeluted precisely with PDE catalytic activity, confirming that this peak is PDE5. Moreover, the value of 19 nm for PDE5 that was calculated from cGMP-binding activity in the peak fraction of Fig. 7 was very near the value of 20 nm calculated from PDE activity in this fraction. The fact that values for molar levels of both PKG and PDE5 calculated by cGMP-binding activities approximated values calculated using enzyme catalytic activities provided high confidence that relative values shown for these enzymes in Table 3 are reliable.


Figure 5. cGMP-binding profile after DEAE-Sephacel chromatography of crude extract of rabbit corpus cavernosum, showing the effect of a mixture of cAMP and CPTcGMP in the cGMP-binding assay. Fractions are the same as those used in Fig. 1. cGMP-binding assays were performed as described in Experimental procedures. Assays were performed in the absence and presence of a combination of 0.5 µm cAMP and 0.25 µm 8-CPTcGMP.

Download figure to PowerPoint


Figure 6. PKG-specific cGMP-binding profile after DEAE-Sephacel chromatography of crude extract of rabbit corpus cavernosum and comparison with the PKG activity profile. Fractions are the same as those used in Fig. 1. Values in Fig. 5 for cGMP-binding activity in the presence of cAMP/CPTcGMP were subtracted from those in the absence of this mixture to obtain the cGMP-binding profile shown. PKG activity is from Fig. 4.

Download figure to PowerPoint


Figure 7. PDE5-specific cGMP-binding activity profile after DEAE-Sephacel chromatography of crude extract of corpus cavernosum and comparison with low cGMP-PDE profile. cGMP-binding activity is that shown in Fig. 5 in the presence of cAMP/CPTcGMP. PDE activity is from Fig. 1.

Download figure to PowerPoint

Table 3.  Calculated intracellular concentrations of cGMP and binding sites of PKG and PDE5 in rabbit corpus cavernosum. Intracellular concentrations of PKG and PDE5 were calculated based on specific enzyme activities of pure proteins and enzyme activities in the DEAE fractions. Intracellular water was assumed to be 0.5 g·g tissue−1. PKG, PKG Iα and PKG Iβ. Values for PKG and PDE5 were calculated as cGMP-binding sites assuming four cGMP-binding sites per holoenzyme molecule for each protein.
 Calculated intracellular concentration (nm)
PKG58 ± 14
PDE5188 ± 6
cGMP18 ± 4

cGMP in rabbit corpus cavernosum extracts

Boiled samples of the original extract were mixed with tracer [3H]cGMP and chromatographed on Sephadex G-25 before assay of cGMP using our published method [26]. The [3H]cGMP served to identify the cGMP peak after chromatography as well as for correcting assay values for recovery. This procedure removes proteins and contaminating nucleosides and nucleotides that might interfere with cyclic nucleotide assays, and it sufficiently separates cGMP from cAMP. This was particularly important as presence of high cAMP could potentially give a falsely high value for cGMP. We utilized a kinase activation assay for cGMP that was developed in this laboratory, which is highly sensitive and selective for each nucleotide. From assay values, the intracellular cGMP concentration was determined to be 18 ± 4 nm(Table 3). This compares with a value of 50 nm (assuming 50% cell water) reported by Bush et al. [27].

Stoichiometric ratio among cGMP and cGMP-binding proteins

Using data from this study, intracellular concentration of cGMP was compared with those of PKG and PDE5, which were calculated as the theoretical total cGMP-binding sites of each protein, as shown in Table 3. The capacity for binding of cGMP by either PKG or PDE5 exceeded the amount of cGMP measured in unstimulated corpus cavernosum. This might be expected for PKG, which is the main target of cGMP for smooth muscle relaxation and penile erection in this tissue. There was an even larger excess of cGMP-binding sites over cGMP for PDE5 (≈ tenfold) than for PKG (≈ threefold).

cAMP and PKA in corpus cavernosum

We have also measured cAMP in the crude extract, as well as PKA and cAMP-binding activities in DEAE fractions of extract of rabbit corpus cavernosum. Using the same procedures described above for cGMP, PKG, and PDE5, intracellular cAMP was calculated to be 100 ± 14 nm (n = 6) as compared with a cGMP level of 18 ± 4 nm (n = 4) (see above). cAMP-binding sites of PKA were determined to be 240 nm, which was considerably higher than those of PKG (58 nm); however, free regulatory subunit of PKA was found to be in nearly twofold excess of the holoenzyme and PKA was approximately equally divided into PKAI and PKAII isozymes. It could not be ruled out that, due to a phenomenon termed cross-activation [14], the large amount of regulatory subunit of PKA binds a small amount of the total cGMP in corpus cavernosum. It should also be emphasized that the regulatory subunit could sequester a high proportion of cAMP in this tissue. Whether or not the free regulatory subunit could act to sequester cAMP for regulation of this signaling pathway should be explored.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The present finding of significant levels of cGMP, PKG, and PDE5 in rabbit corpus cavernosum are consistent with the important role of cGMP signaling in smooth muscle tissues. Levels of cAMP and PKA also suggest important roles for these factors, although cAMP-PDE activity is low compared with cGMP-PDE activity. Whether or not cAMP could ‘cross-activate’ PKG in this tissue should be explored, but this nucleotide binds to PDE5 allosteric sites very weakly [28,29]

Results herein demonstrate that cGMP-binding sites of PKG are 3.2 times higher than cGMP molecules in unstimulated rabbit corpus cavernosum, which is close to the value of five times higher that we measured earlier in pig coronary artery smooth muscle [15], but considerably lower than the value of 37 times for these factors in platelets [30]. Given the high affinity of cGMP for PKG, and the high level of PKG in smooth muscle cells, the intracellular relationship between cGMP and PKG would be expected to be nearly stoichiometric. Thus, PKG could be one-fifth to one-third saturated with cGMP in the unstimulated state. This does not provide a large window for activation of PKG for smooth muscle relaxation or penile erection as threefold to fivefold elevation in cGMP should produce maximum effect. This is in good agreement with our finding that the concentration of sodium nitroprusside which produces maximum relaxation of pig coronary artery smooth muscle raises cGMP by only threefold [14]. A narrow window for cGMP regulation of penile erection or other processes should necessitate tight cellular control of cGMP levels. Over-stimulation of cGMP levels would be deleterious to many physiological processes. A narrow window for cAMP/PKA regulation of heart rate and force has already been demonstrated [31].

This report is the first to address the stoichiometric relationships among cGMP, PKG, and PDE5 in any tissue. These three factors play fundamental roles in the regulation of smooth muscle relaxation in corpus cavernosum and other tissues. The impetus for the present study was our recent finding that cGMP binding to PDE5 allosteric sites is regulated by phosphorylation [6]. This binding activity may have more than a single physiological role, but one suggested here is that it could sequester cGMP away from its targets, such as PKG, for stimulation of smooth muscle relaxation in corpus cavernosum and other tissues. This sequestration would increase after phosphorylation. In order for sequestration of cGMP by PDE5 to be a significant mode of cell regulation, the stoichiometric amount of this enzyme should be relatively close to that of cGMP. Results herein demonstrate that molar ratio of PDE5 allosteric cGMP-binding sites to cGMP is about ten, which is clearly sufficient for PDE5 to bind a significant portion of cGMP, even after elevation of this nucleotide by nitric oxide or other agents. A similar level of PDE5 was obtained when quantified by either catalytic or cGMP-binding activity, verifying this stoichiometry.

As the specific enzyme activity of PDE5 is 5 µmol·min−1·mg−1[22], the kcat would be 8 mol·mol−1·s−1. According to calculations above, catalytic sites (half the number of binding sites) would exceed cGMP by a molar ratio of 5 : 1. Thus, assuming equal intracellular distributions of cGMP and PDE5, cellular cGMP would turn over every 1/40 s [(1/8 s·mol cGMP−1)/(1/5 mol cGMP·mol PDE5 catalytic site−1)] if cGMP were saturating for PDE5 catalysis. However, cGMP is only 18 nm in corpus cavernosum smooth muscle cells. As the Km of PDE5 for cGMP is 5.6 µm[22], the rate of cGMP hydrolysis would be about 0.018 µm/(5.6 µm × 2) = 0.0016 of the maximum rate. It could therefore be estimated that the time required for turnover of total cellular cGMP by PDE5 catalytic activity would be 1/40 s/0.0016 = 16 s. This is probably a minimum value as it assumes that binding of cGMP to PDE5 and PKG would not influence turnover rate. The value for basal cellular cGMP reported here (18 nm) is somewhat less than that reported by Bush et al. (50 nm) for rabbit corpus cavernosum [1].

It should be emphasized again that most cGMP molecules under intracellular conditions would be bound to PDE5 or PKG given the high affinities of these two proteins for cGMP. Using the KD value of 30 nm that we determined recently for phospho-PDE5 [6], and the intracellular values of 188 nm PDE5 and 18 nm for cGMP (Table 3), we used the simple equation for ligand (L) binding to receptor (R) Eqn (1) to calculate KD Eqn (2):

  • image(1)
  • image(2)

Inserting values for each entity, a concentration of 3.1 nm free cGMP (L) is calculated. As total intracellular cGMP is 18 nm, the PDE5-bound concentration is 83% of the total. This assumes that PKG does not bind cGMP at all. Because of the relatively low molar amount of PDE2 in corpus cavernosum it is doubtful that there is sufficient cGMP-binding capacity of this enzyme to play a significant role in sequestration of cGMP. This is probably also the case for cyclic nucleotide-gated channels [32,33], which are believed to be present at low molar levels in tissues. No peak of cGMP-binding activity other than those associated with PDE5 and PKG was obtained using the assay adopted for the present investigation. The model in Fig. 8 depicts our proposal for cellular negative feedback regulation of cGMP based on stoichiometric relationships of cGMP, PKG, and PDE5 determined in this study, as well as effects of phosphorylation of PDE5. The increased binding affinity of PDE5 for cGMP following phosphorylation of this enzyme would favor sequestration of cGMP in these sites, resulting in decreased activation of PKG. Increases in cellular cGMP level stimulate phosphorylation of PDE5 both by activation of PKG and by a substrate-directed effect, i.e. by cGMP binding to the allosteric sites of PDE5. Therefore, cGMP elevation would cause increased sequestration, resulting in dampening of the cGMP signal and facilitating termination of this signal. This process is apparently part of a concert of negative feedback processes for cGMP that have evolved for tight regulation of penile erection and other physiological events. These have all been described in this laboratory and include: (a) increased PDE5 catalytic activity due to mass action of elevated cGMP; (b) increased cGMP binding to PDE5 allosteric sites due to mass action of elevated cGMP; (c) increased PDE5 catalytic activity due to phosphorylation and activation of PDE5 by activated PKG [6], and (d) increased cGMP binding to PDE5 allosteric sites due to this phosphorylation. A fifth possible process is direct stimulation of the PDE5 catalytic site by allosteric cGMP binding to the enzyme, which would be predicted by the principle of reciprocity as discussed earlier [34]. The presence of such an array of mechanisms for negative feedback of the cGMP pathway suggests that cells cannot readily tolerate excessive activation of PKG or other target proteins.


Figure 8. Model for physiological negative feedback regulation of cGMP by increased catalytic and cGMP-binding activities of PDE5. (A) unstimulated cells (low cGMP). (B) stimulated cells with elevated cGMP. Stoichiometric relationships are estimated from measurements in rabbit corpus cavernosum in this report. Elevated cGMP may be exaggerated in order to illustrate cGMP-binding proteins in different states of bound cGMP.

Download figure to PowerPoint

Negative feedback control of cyclic nucleotide-stimulated pathways is not unique to cGMP pathways. We described a negative feedback pathway for cAMP [9] that involves stimulation of PDE3 by phosphorylation catalyzed by activated PKA [8]. PDE4 is also activated by a similar mechanism in some tissues [35], and this enzyme is induced by chronic elevation of cAMP [12].

The fact that sildenafil, when taken as Viagra™ tablets, is rather selective for penile erection and does not have strong effects on other tissues, may have several possible explanations [4,36]. One that is particularly viable is that sexual arousal causes relatively specific nerve stimulation directed to the penis. Elevation of cGMP due to guanylyl cyclase activation by selective nitric oxide release in this organ would be potentiated by sildenafil. Another possibility presented here is that PDE5 is by far the predominant PDE that hydrolyzes nearly physiological concentrations of cGMP in crude extract of rabbit corpus cavernosum. Also, PDE5 is relatively abundant in this tissue, being at a level (94 nm PDE5 catalytic sites in rabbit) that is nearly stoichiometric with the free blood level of sildenafil, which is believed to be ≈ 50 nm in humans after a 100-mg tablet [37,38]. Under this condition, the high affinity of sildenafil for PDE5 would dictate that a high proportion of total cellular PDE5 would contain bound sildenafil. This would cause substantial inhibition of this enzyme and result in significant cGMP elevation and enhanced penile erection.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We are grateful to A. Beasley and K. Grimes for excellent technical assistance. We also thank N. Mount and S. Ballard of Pfizer Central Research (Sandwich, UK) for several informative discussions. This work was supported by Pfizer Central Research, Sandwich, UK; National Institutes of Health (DK40029); American Heart Association; American Heart Association South-east Affiliate; and E. Bronson Ingram Cancer Center at Vanderbilt University.


  1. Top of page
  2. Abstract
  3. Experimental procedures
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Bush, P.A., Aronson, W.J., Buga, G.M., Rajfer, J., Ignarro, L.J. (1992) Nitric oxide is a potent relaxant of human and rabbit corpus cavernosum. J. Urol. 147, 16501655.
  • 2
    Burnett, A.L., Lowenstein, C.J., Bradt, D.S., Chang, T.S.K., Snyder, S.H. (1992) Nitric oxide in the penis: physiology and pathology. Science 257, 401403.
  • 3
    Wagner, G. & de Tejada, S. (1998) Update on male erectile dysfunction. Br. Med. J. 316, 678682.
  • 4
    Corbin, J.D. & Francis, S.H. (1999) Cyclic GMP phosphodiesterase 5: target for sildenafil. J. Biol. Chem. 274, 1372913732.
  • 5
    Boolell, M., Allen, M.J., Ballard, S.A., Gepi-Attee, S., Muirhead, G.J., Naylor, A.M., Osterloh, I.H., Gingell, C. (1996) Sildenafil: an orally active type 5 cyclic GMP-specific phosphodiesterase inhibitor for the treatment of penile erectile dysfunction. Int. J. Impotence Res. 8, 4752.
  • 6
    Corbin, J.D., Turko, I.V., Beasley, A., Francis, S.H. (2000) Phosphorylation of phosphodiesterase-5 by cyclic nucleotide-dependent protein kinase alters its catalytic and allosteric cGMP-binding activities. Eur. J. Biochem. 267, 27602767.DOI: 10.1046/j.1432-1327.2000.01297.x
  • 7
    Loten, E.G. & Sneyd, J.G. (1973) Evidence for separate sites of action for the antilipolytic effects of insulin and prostaglandin E1. Endocrinology 93, 13151322.
  • 8
    Degerman, E., Belfrage, P., Manganiello, V.C. (1997) Structure, localization, and regulation of cGMP-inhibited phosphodiesterase (PDE3). J. Biol. Chem. 272, 68236826.
  • 9
    Corbin, J.D., Beebe, S.J., Blackmore, P.F. (1985) cAMP-dependent protein kinase activation lowers hepatocyte cAMP. J. Biol. Chem. 260, 87318735.
  • 10
    Sette, C., Iona, S., Conti, M. (1994) The short-term activation of a Rolipram-sensitive, cAMP-specific phosphodiesterase by thyroid-stimulating hormone in thyroid FRTL-5 cells is mediated by a cAMP-dependent phosphorylation. J. Biol. Chem. 269, 92459252.
  • 11
    Conti, M., Iona, S., Cuomo, M., Swinnen, J.V., Odeh, J., Svoboda, M.E. (1995) Characterization of a hormone-inducible, high affinity adenosine 3′-5′-cyclic monophosphate phosphodiesterase from the rat Sertoli cell. Biochemistry 34, 79797987.
  • 12
    Swinnen, J.V., Joseph, D.R., Conti, M. (1989) The mRNA encoding a high-affinity cAMP phosphodiesterase is regulated by hormones and cAMP. Proc. Natl Acad. Sci. USA 86, 81978201.
  • 13
    Shabb, J.B. & Corbin, J.D. (1992) Cyclic nucleotide-binding domains in proteins having diverse functions. J. Biol. Chem. 267, 57235726.
  • 14
    Jiang, H., Colbran, J.L., Francis, S.H., Corbin, J.D. (1992) Direct evidence for cross-activation of cGMP-dependent protein kinase by cAMP in pig coronary arteries. J. Biol. Chem. 267, 10151019.
  • 15
    Francis, S.H., Noblett, B.D., Todd, B.W., Wells, J.N., Corbin, J.D. (1988) Relaxation of vascular and tracheal smooth muscle by cyclic nucleotide analogs that preferentially activate purified cGMP-dependent protein kinase. Mol. Pharmacol. 34, 506517.
  • 16
    Gillespie, P.G. (1990) Phosphodiesterases in visual transduction by rods and cones. In Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds), pp. 163184. John Wiley and Sons, Chichester, UK.
  • 17
    Francis, S.H., Wolfe L., Corbin, J.D. (1991) Purification of type I alpha and type I beta isozymes and proteolyzed type I beta monomeric enzyme of cGMP-dependent protein kinase from bovine aorta. Methods Enzymol. 200, 332341.
  • 18
    Glass, D.B. & Krebs, E.G. (1982) Phosphorylation by guanosine 3′:5′-monophosphate-dependent protein kinase of synthetic peptide analogs of a site phosphorylated in histone H2B. J. Biol. Chem. 257, 11961200.
  • 19
    Cheng, H.C., Kemp, B.E., Pearson, R.B., Smith, A.J., Van Misconi, L., Patten, S.M., Walsh, D.A. (1986) A potent synthetic peptide inhibitor of the cAMP-dependent protein kinase. J. Biol. Chem. 261, 989992.
  • 20
    Martins, T.J., Mumby, M.C., Beavo, J.A. (1982) Purification and characterization of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from bovine tissues. J. Biol. Chem. 257, 19731979.
  • 21
    Wolfe, L., Corbin, J.D., Francis, S.H. (1989) Characterization of a novel isozyme of cGMP-dependent protein kinase from bovine aorta. J. Biol. Chem. 264, 77347741.
  • 22
    Thomas, M.K., Francis, S.H., Corbin, J.D. (1990) Characterization of a purified bovine lung cGMP-binding cGMP phosphodiesterase. J. Biol. Chem. 265, 1496414970.
  • 23
    Qiu, Y., Kraft, P., Lombardi, E., Clancy, J. (2000) Rabbit corpus cavernosum smooth muscle shows a different phosphodiesterase profile than human corpus cavernosum. J. Urology 164, 882886.
  • 24
    Sonnenburg, W.K., Seger, D., Kwak, K.S., Huang, J., Charbonneau, H., Beavo, J.A. (1995) Identification of inhibitory and calmodulin-binding domains of the PDE1A1 and PDE1A2 calmodulin-stimulated cyclic nucleotide phosphodiesterases. J. Biol. Chem. 270, 3098931000.
  • 25
    Wolfe, L., Francis, S.H., Corbin, J.D. (1989) Properties of a cGMP-dependent monomeric protein kinase from bovine aorta. J. Biol. Chem. 264, 41574162.
  • 26
    Corbin, J.D., Gettys, T.W., Blackmore, P.F., Beebe, S.J., Francis, S.H., Glass, D.B., Redmon, J.B., Sheorain, V.S., Landiss, L.R. (1988) Purification and assay of cAMP, cGMP, and cyclic nucleotide analogs in cells treated with cyclic nucleotide analogs. Methods Enzymol. 159, 7482.
  • 27
    Bush, P.A., Aronson, W.J., Rajfer, J., Ignarro, L.J. (1993) The l-arginine-nitric oxide-cyclic GMP pathway mediates inhibitory nonadrenergic-noncholinergic neurotransmission in the corpus cavernosum of human and rabbit. Circulation 87, V30V32.
  • 28
    Francis, S.H., Lincoln, T.M., Corbin, J.D. (1980) Characterization of a novel cGMP binding protein from rat lung. J. Biol. Chem. 255, 620626.
  • 29
    Turko, I.V., Francis, S.H., Corbin, J.D. (1999) Studies of the molecular mechanism of discrimination between cGMP and cAMP in the allosteric cyclic nucleotide-binding sites of the cGMP-binding cGMP-specific phosphodiesterase. J. Biol. Chem. 274, 2903829041.
  • 30
    Eigenthaler, M., Nolte, C., Halbrugge, M., Walter, U. (1992) Concentration and regulation of cyclic nucleotides, cyclic- nucleotide-dependent protein kinases and one of their major substrates in human platelets. Estimating the rate of cAMP-regulated and cGMP-regulated protein phosphorylation in intact cells. Eur. J. Biochem. 205, 471481.
  • 31
    Keely, S.L. & Corbin, J.D. (1977) Involvement of cAMP-dependent protein kinase in the regulation of heart contractile force. Am. J. Physiol. 233, H269H275.
  • 32
    Biel, M., Altenhofen, W., Hullin, R., Ludwig, J., Freichel, M., Flockerzi, V., Dascal, N., Kaupp, U.B., Hofmann, F. (1993) Primary structure and functional expression of a cyclic nucleotide-gated channel from rabbit aorta. FEBS Lett. 329, 134138.
  • 33
    Yao, X.Q., Segal, A.S., Welling, P., Zhang, X.Q., McNicholas, C.M., Engel, D., Boulpaep, E.L., Desir, G.V. (1995) Primary structure and functional expression of a cGMP-gated potassium channel. Proc. Natl Acad. Sci. USA 92, 1171111715.
  • 34
    Weber, G. (1975) Energetics of ligand binding to protein. Adv. Protein Chem. 29, 183.
  • 35
    Sette, C. & Conti, M. (1996) Phosphorylation and activation of a cAMP-specific phosphodiesterase by the cAMP-dependent protein kinase. Involvement of serine 54 in the enzyme activation. J. Biol. Chem. 271, 1652616534.
  • 36
    Ballard, S.A., Gingell, C.J., Tang, K., Turner, L.A., Price, M.E., Naylor, A.M. (1998) Effects of sildenafil on the relaxation of human corpus cavernosum tissue in vitro and on the activities of cyclic nucleotide phosphodiesterase isozymes. J. Urol. 159, 21642171.
  • 37
    Jeremy, J.Y., Ballard, S.A., Naylor, A.M., Miller, M.A., Angelini, G.D. (1997) Effects of sildenafil, a type-5 cGMP phosphodiesterase inhibitor, and papaverine on cyclic GMP and cyclic AMP levels in the rabbit corpus cavernosum in vitro. Br. J. Urol. 79, 958963.
  • 38
    Zusman, R.M., Morales, A., Glasser, D.B., Osterloh, I.H. (1999) Overall cardiovascular profile of sildenafil citrate. Am. J. Cardiol. 83, 35C44C.
  1. Enzymes: cyclic nucleotide phosphodiesterase (EC; cAMP-dependent protein kinase (EC; cGMP-dependent protein kinase (EC