International Journal of Impotence Research (2006) 18, 251–257 & 2006 Nature Publishing Group All rights reserved 0955-9930/06 $30.00 www.nature.com/ijir

ORIGINAL ARTICLE Tyrosine-612 in PDE5 contributes to higher affinity for over

J Corbin, S Francis and R Zoraghi

Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN, USA

Despite close structural similarity, vardenafil (Levitras) is 32-fold more potent than sildenafil (Viagras) to inhibit cGMP-binding cGMP-specific PDE (PDE5); this is due to differences between their heterocyclic rings. In co-crystals with PDE5, one of the rings of vardenafil or sildenafil interacts with Tyr612, a catalytic site AA, via (1) a hydrogen bond with a water molecule and (2) hydrophobic bonds. For mutant PDE5Y612F, which ablates hydrogen-bonding potential, vardenafil or sildenafil inhibition was strengthened (2.2- or 3.0-fold, respectively), implying that the Tyr612 hydroxyl is a negative determinant for these inhibitors. For mutant PDE5Y612A, which ablates both hydrogen bonding and hydrophobic-bonding potential, vardenafil inhibition was weakened much more than sildenafil inhibition (122- and 26-fold, respectively), suggesting that hydrophobic bonds involving Tyr612 are stronger for vardenafil than for sildenafil. International Journal of Impotence Research (2006) 18, 251–257. doi:10.1038/sj.ijir.3901411; published online 10 November 2005

Keywords: phosphodiesterases (PDEs); GAF; allosteric cGMP-binding sites; noncatalytic cGMP- binding sites; vardenafil; sildenafil; PDE inhibitors

Introduction catalytic activity is also powerfully impacted by unique structural features in each that account for cGMP-binding cGMP-specific phosphodiesterase the difference in IC50 for weak inhibitors, for (PDE5) is an important component of the cGMP- example, IBMX, and potent inhibitors such as hydrolyzing activity in many tissues, including the sildenafil, vardenafil, and . smooth muscle of penile arteries and corpus The molecular basis for the potency difference cavernosum. PDE5 is the target of sildenafil (Via- between vardenafil and sildenafil is enigmatic since grat), vardenafil (Levitrat), and tadalafil (Cialist), X-ray crystallographic structures of PDE5 isolated which are used for treatment of erectile dysfunction catalytic domain in complex with each did not reveal and other maladies associated with vascular dis- any difference in binding mode.18 However, con- ease.1–14 These inhibitors directly compete with tribution of contacts to inhibitor potency cannot be the cGMP substrate for the PDE5 catalytic site, are fully quantified from crystal structures. The atomic highly potent, and share some structural features of structures of sildenafil and vardenafil share struc- cGMP and many nonspecific, low-potency PDE tural similarities that include a heterocyclic pyrazo- inhibitors such as 3-isobutyl-1-methylxanthine lopyrimidinone and imidazotriazinone ring system, (IBMX) (Figure 1).4,9,15–17 All of these compounds respectively (Figure 1), which differ in (1) presence have a heterocyclic core ring structure that is likely of either a carbon or nitrogen at two positions to be essential for competition with cGMP, which (circled in Figure 1) in the five-member pyrazole/ contains a heterocyclic ring. Biochemical imidazole ring and (2) a substituent (methyl or ethyl) potency (IC50) of the inhibitors to block PDE5 of the piperazine moiety (circled in Figure 1). Differences in the five-member rings could also alter electron distribution in the heterocyclic ring and affect the electronegativity of the nitrogen that is Correspondence: Professor J Corbin, Molecular Physiology common to the five-member rings (Figure 1). and Biophysics, Vanderbilt University, 21st & Garland, 612 Nashville, TN 37232-0615, USA. Tyr , a conserved AA in the PDE5 catalytic site, E-mail: [email protected] interacts with the five-member ring in either Received 8 June 2005; revised 20 September 2005; sildenafil or vardenafil (Figure 2) via (1) a hydrogen accepted 20 September 2005; published online 10 bond of the nitrogen (indicated as position 8 in November 2005 Figure 2) with a water that, in turn, forms hydrogen Vardenafil versus sildenafil potency for PDE5 J Corbin et al 252

Figure 1 Molecular structures of sildenafil and vardenafil compared with those for IBMX and cGMP. Circled components denote differences in nitrogen and carbon distribution in the pyrazolopyrimidinone and imidazotrizinone rings of sildenafil and vardenafil, respectively, and in the methyl versus ethyl groups appended to the piperazine moiety in the two inhibitors.

Figure 2 Schematic depiction of interactions of Tyr612 with the pyrazole or imidazole moiety of sildenafil or vardenafil. Dotted lines depict the chain of hydrogen bonds between the nitrogen common to both five-member rings (pyrazole/imidazole moieties) of the heterocyclic ring with a water that is also hydrogen-bonded to the phenolic hydroxyl of Tyr612 and another water that forms a water bridge to the catalytic site Zn2 þ . The black arc indicates hydrophobic interaction between the aromatic side chain of Tyr612 and the pyrazole/imidazole ring of the respective inhibitors. Positions in the heterocyclic rings of each inhibitor are numbered 1–9 in keeping with the numbers of these positions in the guanine ring of cGMP. Experimentally determined pKa values for the nitrogen in the five- member ring that is common to both sildenafil and vardenafil are noted.

bonds with both the hydroxyl of Tyr612 and a second accounts for a portion of the difference in potencies. water that is coordinated to the catalytic site Zn2 þ , Contributions of these interactions to the potencies and (2) hydrophobic interactions with the aromatic of sildenafil and vardenafil have been investi- side chain of Tyr612.18 We previously demonstrated gated using site-directed mutagenesis of human 612 612 that substitution of alanine for Tyr (PDE5Y612A) PDE5A1 to replace Tyr with either phenylalanine causes a dramatic decrease in affinity of PDE5 for (PDE5Y612F) or alanine (PDE5Y612A). If the electro- 19 the substrate, cGMP. Affinity of PDE5Y612A for negativity of the nitrogen indicated as position 8 either , UK-122764, or sildenafil is also in sildenafil and vardenafil structures (Figure 2) decreased; compared to PDE5WT which has IC50 differs, then potency of hydrogen-bonding inter- values of B100–300 nM,20nM, and 4 nM, respec- actions involving the water bridge that connect this 612 tively, for the inhibitors, IC50 values of PDEY612A for atom to Tyr may also differ. Such a finding could these inhibitors are decreased 7-, 21-, and 25-fold, also suggest different electron distribution in the respectively.20 Therefore, we hypothesize that dif- five-member rings of the inhibitors; this might also ference in interactions of sildenafil and vardenafil contribute to different potency of interactions with with Tyr612 through either hydrophobic contacts, Tyr612. Therefore, electrochemical characteristics hydrogen bonding, or both types of contacts (pKa) of this nitrogen in the respective pyrazole

International Journal of Impotence Research Vardenafil versus sildenafil potency for PDE5 J Corbin et al 253 and imidazole rings in the inhibitors have been then used directly as virus stock for expression experimentally determined. His-tagged PDE5 pro- without additional purification of recombinant teins were expressed in Sf9 cells, purified, and viruses. Sf9 cells grown at 271C in complete Grace’s characterized for allosteric cGMP-binding proper- insect medium with 10% fetal bovine serum and ties, kinetics of catalytic activity (Km and Vmax), and 10 mg/ml gentamicin (Sigma) in T-175 flasks (Corn- inhibitor potencies (IC50). Results of these studies ing) were infected with B100 ml viral stock/flask provide insight into the mechanism of discrimina- and harvested 92 h postinfection. tion between vardenafil and sildenafil binding to the PDE5 catalytic site. Purification of hPDE5A1 proteins All purification steps were carried out at 41C. Cell Materials and methods pellet from each T-175 flask (B2 Â 107 cells) was suspended in 3 ml lysis buffer (20 mM Tris-HCl, Materials pH 8, 100 mM NaCl) containing protease inhibitors [3H]cGMP was purchased from Amersham Bio- as recommended (Completet; Roche Molecular sciences Inc. (Piscataway, NJ, USA). IBMX, Crotalus Biochemicals) and homogenized in 10- to 20-ml atrox snake venom, cGMP, and histone type II-AS aliquots by 2 Â 6-s bursts in an Ultra Turrex micro- were from Sigma Chemical Co. (St Louis, MO, USA). homogenizer (Tekmar) with 20-s recovery between Sildenafil was purified as described earlier.21 bursts. Homogenate was centrifuged (20 min, 10 000 r.p.m. in a Beckman JA-20 rotor). Supernatant was applied to a Ni-NTA agarose (Qiagen) column Generation of wild-type and mutant hPDE5A1 (1 Â 2 cm) equilibrated with lysis buffer. Column Human cDNA coding for full-length PDE5A1 (cour- was washed with 100 ml lysis buffer followed by tesy of Dr K Omori, Tanabe-Seiyaku Pharmaceutical sequential washes with lysis buffer containing a Co. Ltd., Saitama, Japan) was used as template to stepwise gradient of imidazole (0.8–20 mM). Lysis generate full-length PDE5 by introduction of start buffer containing 100 mM imidazole was soaked into and stop codons at appropriate loci. The resulting the resin, and column was plugged for 2 h before PCR fragment was cloned into pCR 2.1-TOPOs collection of ten 1-ml elutions. Elutions containing vector (Invitrogen), verified by sequencing, and PDE5 were pooled, dialyzed versus 2000 volumes ligated into the EcoRI and NotI unique sites of of 10 mM potassium phosphate, pH 6.8, 25 mM baculovirus expression vector pAcHLT-A (Pharmin- b-mercaptoethanol (KPM) to remove imidazole, gen), which contains a His sequence preceding the flash-frozen in KPM containing 0.15 M NaCl and 6 1 coding region. This step resulted in plasmid pAcA- 10% sucrose, and stored at À70 C. activity in frozen samples was stable over 10 months. PDE5 (M1–N875), which generated N-terminally His- tagged recombinant hPDE5A1. The QuikChange site-directed mutagenesis kit (Stratagene) was used to make point mutations (Y612 to F and Y612 to A) SDS-polyacrylamide gel electrophoresis in the pAcA-PDE5 expression vector according to (SDS-PAGE) the manufacturer’s protocol using the following Purity and integrity of proteins were assessed using pairs of mutagenic oligonucleotides (altered bases SDS-PAGE. Protein samples containing 10% SDS, underlined): (1) for Y612F, 50-GAAGAATGTTGCCT 2 M b-mercaptoethanol, and 0.1% bromphenol blue TCCATAATTGGAGACATG-30 and 50-CATGTCTCCA were boiled for 4 min and subjected to 12% SDS- ATTATGGAAGGCAACATTCTTC-30; (2) for Y612A, PAGE before visualizing by 50-GAAGAATGTTGCCGCTCATAATTGGAGAC-30 and staining. 50-GTCTCCAATTATGAGCGGCAACATTCTTC-30. The presence of the desired mutation was verified by sequencing the entire DNA segment. cGMP binding To measure cGMP binding, Millipore filter binding assays were conducted in a total volume of 50 ml Expression of hPDE5A1 proteins containing 10 mM potassium phosphate buffer, Sf9 cells (BD Pharmingen) were cotransfected with pH 6.8, 1 mM EDTA, 0.5 mg/ml histone type II-AS, BaculoGold linear baculovirus DNA (BD Pharmin- 30 mMDL-dithiothreitol, 0.2 mM sildenafil, and gen) and one of the hPDE5A1 constructs (PDE5WT either 3 mM (stoichoimetry determination) or 0.05– 3 (M1–N875), PDE5Y612F (M1–N875)Y612F, and PDE5Y612A 4 mM (binding-affinity determination) [ H]cGMP. (M1–N875)Y612A) in the pAcHLT-A baculovirus ex- Addition of enzyme initiated the reaction. Follow- pression vector by the phosphate method ing 60 min incubation (41C), 1 ml cold KP buffer according to the protocol from BD Pharmingen. At (10 mM potassium phosphate, pH 6.8) was added; 5 days postinfection, the cotransfection supernatant samples were filtered immediately onto premois- was collected, amplified three times in Sf9 cells, and tened Millipore HAWP filters (pore size 0.45 mm).

International Journal of Impotence Research Vardenafil versus sildenafil potency for PDE5 J Corbin et al 254 Filters were then washed twice with 2 ml cold KP pounds from acidic to basic buffer and from basic to buffer, dried, and counted. Bound protein counts acidic buffer and following UV changes. were corrected by subtraction of nonspecific binding ( þ 1mM unlabeled cGMP). Blanks lacking protein were run for each [3H]cGMP concentration. Results

Catalytic activity of hPDE5A1 proteins Experimental determination of pKa for the nitrogen PDE activity was determined using a modification that is common to both five-member rings of the procedure described previously;22 reaction in sildenafil and vardenafil mixture contained 50 mM Tris/HCl, pH 7.5, 10 mM Measurements of pKa for the nitrogen common to both the pyrazole of sildenafil and the imidazole of MgCl2, 0.3 mg/ml bovine serum albumin, 0.5 mM cGMP, [3H]cGMP (B100 000 c.p.m./tube), and one vardenafil were performed as described in Materials of the PDE5 proteins in a total volume of 100 ml. and methods. The mean of four measurements was Incubation time was 10 min at 301C. Less than 10% calculated. For vardenafil, two pKa values were 3 of [ H]cGMP was hydrolyzed during the reaction. found; the pKa of the nitrogen located at position 8 Apparent Km and Vmax values were determined by (Figure 2) was 3.6. For sildenafil, only the pKa for nonlinear analysis of data using the one-site binding the nitrogen at position 1 was detected. A pKa for the model (hyperbola) in Prism Graphpad Software. nitrogen located at position 8 in the pyrazole ring of Biochemical potency (IC50) for sildenafil or var- sildenafil was not observed. As the measurements denafil inhibition of PDE5 catalytic activity was could not be performed below pH 2, the pKa of this determined using a range of inhibitor concentrations nitrogen in sildenafil was inferred to be o2. (1–1 000 000 pM). IC50 values were determined using the nonlinear regression analysis of the sigmoidal Expression and purification of hPDE5 proteins dose response Prism Graphpad Software (variable Recombinant His-tagged wild-type human PDE5A slope). All values represent three measurements, (PDE5 ) and two mutants (PDE5 and each in triplicate. Gibbs free energy change (DG)of WT Y612F PDE5Y612A) were expressed, purified to near homo- binding upon formation of enzyme–inhibitor com- geneity (498%), and detected on SDS-PAGE (12% plex was calculated from the IC50 of each inhibitor (w/v) gel) as described in Materials and methods for the respective proteins using Equation (1): (Figure 3). Each migrated as a 99-kDa band that DG ¼ÀRT ln KA where KA ¼ IC50 ð1Þ correlated well with the predicted molecular weight based on AA composition of human PDE5. Identities Contribution of a substituted AA side chain to the of the recombinant proteins were verified by free energy of binding in the enzyme–inhibitor Western blot analysis using a polyclonal PDE5- complex was calculated from differences in the specific antibody (not shown). Protein yields were DG values for PDE5WT and PDE5mutant using Equa- B2.7 mg/l infected cell media. tion (2): DD ¼ D À D ð2Þ G GWT Gmut Structural integrity of PDE5 mutants DDG is the change in free energy of binding in Overall structural integrity of the PDE5 proteins was assessed using the Millipore filtration assay to enzyme–inhibitor complexes attributable to the 3 substituted group,23,24 R is the ideal gas constant measure the stoichiometry and affinity of [ H]cGMP (1.98  10À3 kcal/degree/mol), and T is the tempera- binding to the allosteric cGMP-binding sites in the ture at which the assay was done (3031K). regulatory domain of each PDE5WT, PDE5Y612F, and PDE5Y612A bound cGMP with comparable affinities Attempts were made to confirm IC50 values by 3 (KD ¼ 190–210 nM) and stoichiometry (0.50–0.55 mol measuring direct binding affinity (KD)of[H]silde- 3 nafil and [3H]vardenafil to PDE5. However, because bound [ H]cGMP/PDE5 monomer) (data not shown). These values compared well to those previously of poor binding affinity of PDE5Y612A, exceptionally high concentrations of the tritiated PDE5 inhibitors reported for native and non-His-tagged recombinant bovine PDE5,25,26 indicating that overall structures were required for determination of KD. At concen- trations of PDE5 inhibitor of 420 nM, the blank was of the mutant proteins were preserved, and that excessive and the low signal to noise ratio precluded differences in kinetic parameters were not due to determination of binding affinity. nonspecific conformational effects.

pKa measurements Kinetic analysis of catalytic activity of PDE5 Measurement of the pKa of the respective nitrogens proteins at the position indicated number 8 in sildenafil and Specific enzyme activity of each purified PDE5 vardenafil (Figure 2) was performed on an SGA- construct was measured as described under Materi- profiler (SIRIUS UK: analyzing software: Profiler als and methods. Km and Vmax for cGMP were SGA 3.0) in two replicates by titrating the com- determined by nonlinear regression analysis of data

International Journal of Impotence Research Vardenafil versus sildenafil potency for PDE5 J Corbin et al 255 (Prism Graphpad Software) (Table 1). PDE5WT had and suggested that the main interaction between 612 Km of 2.970.8 mM and Vmax of 1.070.2 mmol/min/mg cGMP and Tyr involves hydrophobic contacts. for cGMP. To examine components of the interaction of Tyr612 with substrate and inhibitors, this residue was replaced by either phenylalanine (PDE5Y612F) IC50 of sildenafil and vardenafil for PDE5 proteins or alanine (PDE5Y612A). PDE5Y612F showed a slight IC50 values (Figure 4, Table 1) were determined as 7 change in affinity for cGMP (Km ¼ 6.1 0.8 mM) described in Materials and methods using 0.5 mM 3 (Table 1), and no change in Vmax (1.070.2 mmol/ [ H]cGMP, which is a significantly lower concentra- min/mg). PDE5Y612A exhibited profound loss in tion than the Km for each protein. Under these B affinity for cGMP as substrate, with 15-fold conditions, IC50 approaches Ki. Vardenafil inhibited increase in Km (4272.1 mM) (Table 1), whereas human PDE5A1 with 32-fold greater potency than 7 specific enzyme activity (Vmax ¼ 0.50 0.03 mmol/ did sildenafil, with IC50 values for these two min/mg) was slightly lower than that of PDE5WT. inhibitors of 0.1270.02 nM and 3.970.4 nM, respec- These results agreed with our previous finding that tively (Table 1). The contribution of the hydrogen- substitution of alanine for Tyr612 mainly affects bonding potential between the hydroxyl of Tyr612 substrate-binding affinity, not maximum catalysis, and the water bridge with the nitrogen that is

Figure 3 SDS-PAGE of purified human PDE5A1 wild-type and mutant proteins. In all, 10 mg of wild-type or mutant human Figure 4 Potency of sildenafil or vardenafil inhibition of wild- PDE5A1 proteins that had been expressed in Sf9 cells and type or mutant PDE5 proteins. Experiments to determine IC50 purified to apparent homogeneity through the Ni-NTA chromato- values were performed as described in Materials and methods. graphy step was applied to each lane. MW is the molecular mass Data shown are representative of three experiments and values are of standards. Proteins were visualized by staining the gel with mean7s.e.m. of triplicate determinations. (a) Sildenafil concen- Coomassie Brilliant Blue dye. tration curve; (b) vardenafil concentration curve.

Table 1 Effect of PDE5 Tyr612 mutations on affinity for cGMP as substrate, as well as on potency and selectivity of inhibition by sildenafil and vardenafil

cGMP Km cGMP Km Sildenafil Sildenafil Vardenafil Vardenafil Selectivity for (mM) (fold effect) IC50 (nM) (fold effect) IC50 (nM) (fold effect) Vardenafil (fold)

WT 2.9070.80 1 3.8570.38 1 0.1270.02 1 32 Y612F 6.1070.80 2 1.3070.25 0.34 0.05570.01 0.45 24 Y612A 41.6072.10 15 98.374.7 26 14.771.60 122 6.7

All values are mean7s.e.m. of three experiments of triplicate determinations. To facilitate comparison, the Km for cGMP and IC50 values for sildenafil and vardenafil were taken as 1.0 and the corresponding values for each construct were calculated as a fold change for each parameter. Selectivity for vardenafil was determined by dividing IC50 for sildenafil by that for vardenafil for each of the proteins.

International Journal of Impotence Research Vardenafil versus sildenafil potency for PDE5 J Corbin et al 256 common to the five-member rings of sildenafil and Recently solved X-ray crystal structures of isolated vardenafil was tested using PDE5Y612F; this sub- PDE5 catalytic domain revealed that vardenafil and stitution retained the aromatic side chain of tyro- sildenafil have similar binding modes. Among sine, but lacked the hydrogen-bonding potential. several contacts, this binding involves interactions PDE5Y612F mutant had increased sensitivity to inhi- of the five-member ring of each of the inhibitors 612 bition by both inhibitors compared to PDE5WT through (1) hydrophobic interactions with Tyr (Figure 4, Table 1). The IC50 of sildenafil or varde- and (2) a hydrogen-bond interaction between the nafil for PDE5Y612F (1.370.3 or 0.0670.01 nM, shared nitrogen (at position 8 in Figure 2) of the respectively) was B3.0- or B2.2-fold lower than pyrazole/imidazole moieties of sildenafil or varde- that for PDE5WT (Table 1). Furthermore, the differ- nafil with a water molecule that also forms a ence in biochemical potency of vardenafil compared hydrogen-bond bridge with both the hydroxyl of to sildenafil was narrowed slightly from 32-fold to Tyr612 and another water molecule that coordinates B24-fold. The fact that this Tyr612 to Phe mutation, to the catalytic site Zn2 þ (Figure 2).18 which increased hydrophobicity, strengthened affi- The present biochemical study has determined nity for both inhibitors suggested that higher affinity that Tyr612 in PDE5, which is critical for binding the of vardenafil in PDE5WT is not due exclusively to cGMP substrate, contributes a component of the greater vardenafil hydrophobic interactions. binding affinity (potency) for both sildenafil and Potency of each inhibitor was significantly less for vardenafil as well as to higher affinity of PDE5 for PDE5Y612A compared to PDE5WT, but the effect was vardenafil over sildenafil. This study also quantifies greater for vardenafil than for sildenafil. Sildenafil this contribution. Based on mutation of Tyr612 inhibited PDE5Y612A with an IC50 ¼ 98.374.7 nM to Ala, the contribution to vardenafil affinity is (26-fold lower potency (or affinity)). The effect of B5-fold greater than that for sildenafil affinity. This this same mutation on vardenafil potency was B5 appears to primarily involve hydrophobic inter- times greater than that on sildenafil; vardenafil actions between the aromatic side chain of Tyr612 inhibited PDE5Y612A mutant with an IC50 ¼ 157 and the five-member ring structures of the inhibi- 612 1.6 nM, that is, 122-fold lower potency compared to tors. Hydrophobic interaction of Tyr with varde- PDE5WT. Therefore, this mutation decreased varde- nafil may be more optimal than that with sildenafil. nafil potency from 32-fold greater than that of Lower binding affinity of sildenafil compared to sildenfail to 4.7-fold greater than that of sildenafil vardenafil for PDE5 catalytic site could have several (Table 1). possible molecular mechanisms. Sildenafil binding could impose a slight strain on the binding pocket; 612 this could occur in part through less than ideal Contribution of Tyr to sildenafil and vardenafil interactions between the Tyr612 side chain and the potency pyrazole group. More optimal interaction when Contributions of the side chain components of 612 vardenafil is bound could contribute to greater Tyr were quantified by calculating the change potency of vardenafil. in free energy of binding (DDG) of sildenafil or In sum, the results herein establish that Tyr612 in vardenafil in the mutants following the standard PDE5 is important for binding both sildenafil and procedure used for assessing potential bonding 24 vardenafil although it is significantly more favorable interactions of ligands with proteins. for vardenafil. The hydrophobicity of this tyrosine The DDG value for PDE5Y612F inhibition by sildena- is critical for potency of both inhibitors. From the fil and vardenafil for each inhibitor (À0.7 and crystal structure, the Tyr612 hydroxyl is indicated to À0.5 kcal/mol, respectively) indicated a modest form a bond with a water bridge between the enhancement of binding affinity. This suggested nitrogen at position 8 and the catalytic site Zn2 þ . that the tyrosine hydroxyl could be a slightly nega- However, the present studies indicate that this tive determinant for both inhibitors. In contrast, hydroxyl group is not essential for binding of DDG values for PDE5Y612A inhibition by sildenafil or either sildenafil or vardenafil. The different electro- vardenafil (2.7 and 2.9 kcal/mol, respectively) were negativity found at position 8 of the two inhibitors in the range expected for an important role for 612 could also play a role in affinity and selectivity, but interaction of the aromatic ring of Tyr with both the data do not suggest a mechanism for either inhibitors in PDE5WT. property.

Discussion Abbreviations

0 0 Previous reports showed that the IC50 for sildenafil PDE, 3 ,5 -cyclic phosphodiesterase; inhibition of PDE5 is B4nM versus a value of PDE5A, cGMP-binding cGMP-specific PDE; KPEM, 17,27 B0.1 nM for vardenafil, and that differences in 10 mM potassium phosphate; pH 6.8, 1 mM EDTA, the heterocyclic ring systems of these inhibitors and 25 mM b-mercaptoethanol; IBMX, 3-isobutyl-1- 28,29 define the higher potency of vardenafil (Figure 1). methylxanthine; DTT, DL-dithiothreitol.

International Journal of Impotence Research Vardenafil versus sildenafil potency for PDE5 J Corbin et al 257 References 16 Jeremy JY, Ballard SA, Naylor AM, Miller MA, Angelini GD. Effects of sildenafil, a type-5 cGMP phosphodiesterase 1 Hamet P, Coquil JF, Bousseau-Lafortune S, Franks DJ, inhibitor, and on cyclic GMP and cyclic AMP Tremblay J. Cyclic GMP binding and phosphodiesterase: levels in the rabbit corpus cavernosum in vitro. Br J Urol 1997; implication for platelet function. Adv Cyclic Nucleotide 79: 958–963. Protein Phosphorylation Res 1984; 16: 119–136. 17 Saenz de Tejada I, Angulo J, Cuevas P, Fernandez A, 2 Corbin JD, Francis SH. Cyclic GMP phosphodiesterase-5: Moncada I, Allona A et al. The phosphodiesterase inhibitory target of sildenafil. J Biol Chem 1999; 274: 13729–13732. selectivity and the in vitro and in vivo potency of the 3 Francis SH, Turko IV, Corbin JD. Cyclic nucleotide phospho- new PDE5 inhibitor vardenafil. Int J Impot Res 2001; 13: diesterases: relating structure and function. Prog Nucleic Acid 282–290. Res Mol Biol 2001; 65: 1–52. 18 Sung BJ, Yeon Hwang K, Ho Jeon Y, Lee JI, Heo YS, Hwan Kim 4 Corbin JD, Francis SH, Webb DJ. Phosphodiesterase type 5 as a J, Moon J et al. Structure of the catalytic domain of human pharmacologic target in erectile dysfunction. Urology 2002; phosphodiesterase 5 with bound drug molecules. Nature 60: 4–11. 2003; 425: 98–102. 5 Rotella DP. Phosphodiesterase 5 inhibitors: current status and 19 Turko IV, Francis SH, Corbin JD. Potential roles of conserved potential applications. Nat Rev Drug Discov 2002; 1: 674–682. amino acids in the catalytic domain of the cGMP-binding 6 Montorsi F, Salonia A, Deho F, Cestari A, Guazzoni G, Rigatti P cGMP-specific phosphodiesterase. J Biol Chem 1998; 273: et al. Pharmacological management of erectile dysfunction. 6460–6466. BJU Int 2003; 91: 446–454. 20 Turko IV, Ballard SA, Francis SH, Corbin JD. Inhibition of 7 Hellstrom WJ, Gittelman M, Karlin G, Segerson T, Thibonnier cyclic GMP-binding cyclic GMP-specific phosphodiesterase M, Taylor T et al. Sustained efficacy and tolerability of (Type 5) by sildenafil and related compounds. Mol Pharmacol vardenafil, a highly potent selective phosphodiesterase type 1999; 56: 124–130. 5 inhibitor, in men with erectile dysfunction: results of 21 Francis SH, Sekhar KR, Rouse AB, Grimes KA, Corbin JD. a randomized, double-blind, 26-week placebo-controlled Single step isolation of sildenafil from commercially available pivotal trial. Urology 2003; 61: 8–14. Viagra tablets. Int J Impot Res 2003; 15: 369–372. 8 Lue TF. Erectile dysfunction. N Engl J Med 2000; 342: 22 Francis SH, Lincoln TM, Corbin JD. Characterization of a 1802–1813. novel cGMP binding protein from rat lung. J Biol Chem 1980; 9 Porst H, Padma-Nathan H, Giuliano F, Anglin G, Varanese L, 255: 620–626. Rosen R. Efficacy of tadalafil for the treatment of erectile dys- 23 Wilkinson AJ, Fersht AR, Blow DM, Winter G. Site-directed function at 24 and 36 hours after dosing: a randomized con- mutagenesis as a probe of enzyme structure and catalysis: trolled trial. Urology 2003; 62: 121–125; discussion 125–126. tyrosyl-tRNA synthetase cysteine-35 to glycine-35 mutation. 10 Goldstein I, Lue TF, Padma-Nathan H, Rosen RC, Steers WD, Biochemistry 1983; 22: 3581–3586. Wicker PA. Oral sildenafil in the treatment of erectile 24 Andrews PR, Craik DJ, Martin JL. Functional group contribu- dysfunction. Sildenafil Study Group. N Engl J Med 1998; tions to drug–receptor interactions. J Med Chem 1984; 27: 338: 1397–1404. 1648–1657. 11 Michelakis ED, Tymchak W, Noga M, Webster L, Wu XC, Lien 25 Thomas MK, Francis SH, Corbin JD. Characterization of a D et al. Long-term treatment with oral sildenafil is safe and purified bovine lung cGMP-binding cGMP phosphodiesterase. improves functional capacity and hemodynamics in patients J Biol Chem 1990; 265: 14964–14970. with pulmonary arterial hypertension. Circulation 2003; 108: 26 Turko IV, Haik TL, McAllister-Lucas LM, Burns F, Francis SH, 2066–2069. Corbin JD. Identification of key amino acids in a conserved 12 Ghofrani HA, Wiedemann R, Rose F, Schermuly RT, cGMP-binding site of cGMP-binding phosphodiesterases. A Olschewski H, Weissmann N et al. Sildenafil for treatment putative NKXnD motif for cGMP binding. J Biol Chem 1996; of lung fibrosis and pulmonary hypertension: a randomised 271: 22240–22244. controlled trial. Lancet 2002; 360: 895–900. 27 Blount MA, Beasley A, Zoraghi R, Sekhar KR, Bessay EP, 13 Kloner RA, Brown M, Prisant LM, Collins M. Effect of Francis SH et al. Binding of tritiated sildenafil, tadalafil, or sildenafil in patients with erectile dysfunction taking anti- vardenafil to the phosphodiesterase-5 catalytic site displays hypertensive therapy. Sildenafil Study Group. Am J Hypertens potency, specificity, heterogeneity, and cGMP stimulation. 2001; 14: 70–73. Mol Pharmacol 2004; 66: 144–152. 14 Carson CC. Erectile dysfunction: evaluation and new treat- 28 Corbin JD, Beasley A, Blount MA, Francis SH. Vardenafil: ment options. Psychosom Med 2004; 66: 664–671. structural basis for higher potency over sildenafil in inhibiting 15 Boolell M, Allen MJ, Ballard SA, Gepi-Attee S, Muirhead GJ, cGMP-specific phosphodiesterase-5 (PDE5). Neurochem Int Naylor AM et al. Sildenafil: an orally active type 5 cyclic 2004; 45: 859–863. GMP-specific phosphodiesterase inhibitor for the treatment 29 Francis SH, Corbin JD. Molecular mechanisms and pharma- of penile erectile dysfunction. Int J Impotence Res 1996; 8: cokinetics of phosphodiesterase-5 antagonists. Curr Urol Rep 47–52. 2003; 4: 457–465.

International Journal of Impotence Research