Binding thermodynamics of substituted diaminopyrimidine renin inhibitors Ronald W. Sarver a, * , Jeanette Peevers a , Wayne L. Cody a , Fred L. Ciske a , Jim Dyer a , S. Donald Emerson a , Jeanne C. Hagadorn a , Daniel D. Holsworth a , Mehran Jalaie a , Michael Kaufman a , Michelle Mastronardi b , Patrick McConnell a , Noel A. Powell a , John Quin III a , Chad A. Van Huis a , Erli Zhang a , Igor Mochalkin a a Pfizer Global Research and Development, Ann Arbor, MI 48105, USA b Manpower, Ann Arbor, MI 48105, USA Received 19 July 2006 Available online 30 October 2006 Abstract Renin is an aspartyl protease involved in the production of angiotensin II, a potent vasoconstrictor. Renin inhibitors can prevent blood vessel constriction and therefore could be useful for the treatment of hypertension. High-throughput screening efforts identified a small molecule renin inhibitor with a core substituted diaminopyrimidine ring. Parallel medicinal chemistry efforts based on this lead resulted in compound 1. A complex of 1 bound to renin was crystallized, and structural data were obtained by X-ray diffraction. The structure indicated that there were adjacent unoccupied binding pockets. Synthetic efforts were initiated to extend functionality into these pockets so as to improve affinity and adjust pharmacokinetic parameters. Thermodynamics data for inhibitor binding to renin were also collected using isothermal titration calorimetry. These data were used to help guide inhibitor optimization by suggesting molecular alter- ations to improve binding affinity from both thermodynamic and structural perspectives. The addition of a methoxypropyl group extend- ing into the S3 subpocket improved inhibitor affinity and resulted in greater binding enthalpy. Initial additions to the pyrimidine ring template that extended into the large hydrophobic S2 pocket did not improve affinity and dramatically altered the thermodynamic driv- ing force for the binding interaction. Binding of the core template was enthalpically driven, whereas binding of initial inhibitors with S2 extensions was both enthalpically and entropically driven but lost significant binding enthalpy. Additional electrostatic interactions were then incorporated into the S2 extension to improve binding enthalpy while taking advantage of the favorable entropy. Ó 2006 Elsevier Inc. All rights reserved. Keywords: Renin; Inhibitor; Thermodynamics; Affinity; Diaminopyrimidine Analysis of data from the 1999–2002 National Health and Nutrition Examination Survey (NHANES) 1 indicated that an estimated 65 million individuals in the United States have hypertension [1]. This analysis also determined that only 63.4% of the individuals sampled knew they were hypertensive and that 45.3% were being treated. Of these, 29.3% had their blood pressure under control, but 70.7% of the total hypertensive population (46 million people) did not [1]. Comparison of these data with earlier data indi- cated that since 1960 there have been improvements in awareness, treatment, and control of hypertension [2], but there is substantial room to improve both the identification of individuals at risk for hypertension and the treatment options. Current treatment options for hypertension range 0003-2697/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2006.10.017 * Corresponding author. E-mail address: ronald.w.sarver@pfizer.com (R.W. Sarver). 1 Abbreviations used: NHANES, National Health and Nutrition Exam- ination Survey; ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; CHO, Chinese hamster ovary; DMEM, Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; DMSO, dimethyl sulfoxide; MWCO, molecular weight cutoff; TPCK trypsin, N-tosyl-L- phenylalanine chloro-methyl ketone-trypsin; PEG, polyethylene glycol; ITC, isothermal titration calorimetry; S3 sp , S3 subpocket; VDW, van der Waals. www.elsevier.com/locate/yabio Analytical Biochemistry 360 (2007) 30–40 ANALYTICAL BIOCHEMISTRY
Transcript
www.elsevier.com/locate/yabio
Analytical Biochemistry 360 (2007) 30–40
ANALYTICAL
BIOCHEMISTRY
Binding thermodynamics of substituted diaminopyrimidinerenin inhibitors
Ronald W. Sarver a,*, Jeanette Peevers a, Wayne L. Cody a, Fred L. Ciske a, Jim Dyer a,S. Donald Emerson a, Jeanne C. Hagadorn a, Daniel D. Holsworth a, Mehran Jalaie a,Michael Kaufman a, Michelle Mastronardi b, Patrick McConnell a, Noel A. Powell a,
John Quin III a, Chad A. Van Huis a, Erli Zhang a, Igor Mochalkin a
a Pfizer Global Research and Development, Ann Arbor, MI 48105, USAb Manpower, Ann Arbor, MI 48105, USA
Received 19 July 2006Available online 30 October 2006
Abstract
Renin is an aspartyl protease involved in the production of angiotensin II, a potent vasoconstrictor. Renin inhibitors can preventblood vessel constriction and therefore could be useful for the treatment of hypertension. High-throughput screening efforts identifieda small molecule renin inhibitor with a core substituted diaminopyrimidine ring. Parallel medicinal chemistry efforts based on this leadresulted in compound 1. A complex of 1 bound to renin was crystallized, and structural data were obtained by X-ray diffraction. Thestructure indicated that there were adjacent unoccupied binding pockets. Synthetic efforts were initiated to extend functionality into thesepockets so as to improve affinity and adjust pharmacokinetic parameters. Thermodynamics data for inhibitor binding to renin were alsocollected using isothermal titration calorimetry. These data were used to help guide inhibitor optimization by suggesting molecular alter-ations to improve binding affinity from both thermodynamic and structural perspectives. The addition of a methoxypropyl group extend-ing into the S3 subpocket improved inhibitor affinity and resulted in greater binding enthalpy. Initial additions to the pyrimidine ringtemplate that extended into the large hydrophobic S2 pocket did not improve affinity and dramatically altered the thermodynamic driv-ing force for the binding interaction. Binding of the core template was enthalpically driven, whereas binding of initial inhibitors with S2extensions was both enthalpically and entropically driven but lost significant binding enthalpy. Additional electrostatic interactions werethen incorporated into the S2 extension to improve binding enthalpy while taking advantage of the favorable entropy.� 2006 Elsevier Inc. All rights reserved.
that an estimated 65 million individuals in the UnitedStates have hypertension [1]. This analysis also determinedthat only 63.4% of the individuals sampled knew they werehypertensive and that 45.3% were being treated. Of these,29.3% had their blood pressure under control, but 70.7%of the total hypertensive population (�46 million people)did not [1]. Comparison of these data with earlier data indi-cated that since 1960 there have been improvements inawareness, treatment, and control of hypertension [2], butthere is substantial room to improve both the identificationof individuals at risk for hypertension and the treatmentoptions. Current treatment options for hypertension range
from lifestyle modifications, including lowering salt intakeand reducing weight, to pharmaceutical intervention. Phar-maceutical therapies currently available include diuretics,a-blockers, a/ b-blockers, b-blockers, calcium channelblockers, angiotensin-converting enzyme (ACE) inhibitors,and angiotensin II receptor blockers (ARBs) [3].
ACE inhibitors reduce blood pressure by inhibiting pro-duction of the potent vasoconstrictor angiotensin II, andARB inhibitors block binding of angiotensin II to theangiotensin receptor. Angiotensin II is produced in thefinal step of a two-step enzymatic cleavage of the glycopro-tein angiotensinogen. In the first step, renin, an aspartylprotease, cleaves the decapeptide angiotensin I from angio-tensinogen. Then ACE catalyzes a two-residue cleavage ofangiotensin I to form the octapeptide angiotensin II. Bind-ing of angiotensin II to the angiotensin receptor initiates acascade of events that include vasoconstriction and reten-tion of sodium and water that can lead to hypertension.Inhibition of angiotensin II production thereby providesa rational opportunity for therapeutic treatment of hyper-tension. Several marketed medications inhibit productionof angiotensin II through ACE inhibition, and their clinicalbenefits have been reviewed [4]. Except for dry coughreported in some treated patients [5], ACE inhibitors havebeen widely accepted based on good efficacy and safetyprofiles. Although ACE inhibitors and many of the existingtherapies are effective alone or in combination, as previous-ly mentioned, there remains a portion of hypertensivepatients who do not attain the desired reduction in bloodpressure or do not continue treatment due to undesiredtreatment effects. In addition, some hypertensive patientsdo not respond to a combination of treatments [6] andopportunities remain for improved treatment options.
Because ACE catalyzes the final step in the metabolicpathway of angiotensin II, inhibitors involved furtherupstream in the metabolic pathway, such as renin inhibi-tors, have been sought to help reduce unwanted side effects.Many mechanistic studies of renin have been performed,and the high-resolution structure of renin bound to sub-strate is known. Despite the wealth of information, devel-oping functional nonpeptidic inhibitors of renin hasproven to be challenging. Initial renin inhibitors were pep-tidic and suffered from rapid metabolic clearance and shortplasma half-life. More recently, some nonpeptidic inhibi-tors have been identified, and the renin inhibitor aliskirenhas completed phase III clinical studies and was submittedto the Food and Drug Administration for review [7].
A high-throughput screen for inhibitors of renin wasperformed at Pfizer and resulted in the identification of aunique nonpeptidic small molecule inhibitor with a diami-nopyrimidine core. A round of parallel chemistry resultedin compound 1, which inhibited 50% of renin activity(IC50) in vitro at a compound concentration of 6 lM. Thiswas an improvement over the IC50 of 27 lM for the initiallead, but increased potency still was required to be a viabledrug candidate. Additional chemical modifications weredone to the pyrimidine core to further optimize the lead
and improve affinity for the substrate-binding pocket[8–10]. Many of the synthesized molecules were complexedwith renin and crystallized. Detailed structural informationon the binding of the lead to renin was obtained fromX-ray crystallographic experiments. In addition, thermody-namics for the binding interactions was measured to deter-mine the enthalpic and entropic contributions to bindingaffinity. Structural information and thermodynamic infor-mation were then used in tandem to provide insights intothe structure–activity relationship for ligand interactionswith the binding pockets. This combination of techniqueswas useful in building a unique nonpeptidic small moleculeinhibitor of renin.
Improved sensitivity, reliability, and ease of use of com-mercial microcalorimeters [11] have increased the use ofthermodynamic information in inhibitor design. For exam-ple, thermodynamic evaluation of ligand-binding interac-tions has been used successfully in the design ofimproved inhibitors of HIV protease [12,13]. These studiesshowed that first-generation protease inhibitors werelargely entropically driven and lost significant inhibitoryactivity with mutations in the HIV protease. Subsequentgeneration inhibitors with enthalpically driven bindingretained significant HIV protease inhibitory activity onenzyme mutations. Enthalpically driven inhibitors weremore flexible and able to accommodate protease mutationsthat left the sterically constrained first-generation inhibi-tors inactive.
Materials and methods
Human preprorenin clones
Human preprorenin was cloned from Human Fetal Kid-ney Quick-Clone cDNA (cat. no. 7170-1, Clontech). Prep-rorenin was cloned into pcDNA3.1(+) (cat. no. V790-20,Invitrogen) at the BamHI and XbaI restriction sites. Theclone was confirmed by sequencing and was compared withthe published human preprorenin sequence (Accession No.E01074).
Transfection and antibiotic selection of CHO cell lines
Transfection of Chinese hamster ovary (CHO) cells, celltype K1 (cat. no. CCL-61, American Type Culture Collec-tion), was performed by using Effectene transfectionreagent (cat. no. 301425, Qiagen). CHO cells, at 70% con-fluency and in a 150-mm tissue culture dish (cat. no. 3025,Falcon), were used for each transfection. A standard trans-fection protocol employed 4 lg renin plasmid DNA and120 ll Effectene reagent in a final volume of 20 ml ofmedia. In addition, 0.4 lg of pcDNA6 V5/HisB plasmidwas transfected with the renin plasmid so as to impart blas-ticidin resistance. The transfection mixture remained on thecells overnight (�20 h).
To establish a stable line, the transfected CHO cellswere grown in Dulbecco’s modified Eagle’s medium
(DMEM)-F12 selection media containing 500 lg/ml genet-icin and 6 lg/ml blasticidin. Following approximately 10days of selection pressure, the surviving cells formed dis-tinct colonies. For clonal selection, flasks of cells werecounted and then diluted to allow for plating of 1 cell perwell using a 96-well plate format. Microscopic examinationwas used to determine wells containing single colonies.Expansion of each colony continued for approximately 7days in a 96-well format. Each colony was expanded intosubsequently larger wells. For long-term storage, cells werecryopreserved in 95% fetal bovine serum (FBS, cat. no.16000-036, Invitrogen) and 5% dimethyl sulfoxide (DMSO)following a standard procedure.
Expression and purification of recombinant human renin
CHO cells expressing preprorenin were grown bothbatchwise and in perfusion mode in a basket bioreactorusing serum-free medium. Batch runs were seeded at1 · 105 per milliliter and grown at 37 �C for 6–7 days,reaching a density of approximately 1.5 · 106 per milliliter.
Perfusion runs (3.5-L bioreactor) were seeded withapproximately 1 · 109 cells, which were allowed to attachto polyester fiber disks contained in a basket within the bio-reactor. The cells were grown at 37 �C and reached a max-imum density of 1.5 · 1011 as estimated from glucoseconsumption. Fresh medium was perfused beginning at72 h at a rate of 2 L per day and gradually increased to amaximum of 20 L per day, with preprorenin beingexpressed for at least 38 days. The resulting secreted prore-nin was glycosylated as indicated by mass spectrometry,but the glycosylation state was not compared directly withnative human prorenin.
Purification was modified from Ref. [14]. All steps wereperformed at 4 �C unless otherwise noted. CHO media con-taining recombinant human prorenin was concentrated150-fold using 10-kDa molecular weight cutoff (MWCO)ultrafiltration, and 40% saturated ammonium sulfate wasadded. The slurry was stirred overnight, and insolublematerial was removed by centrifugation at 16,000g. Thesupernatant was buffer exchanged using 10-kDa MWCOultrafiltration and loaded onto Blue Sepharose 6 Fast Flow(cat. no. 17-0948-02, GE Healthcare) in 20 mM Tris–HCl(pH 8.0) and 0.2 M NaCl. Prorenin was eluted with a lineargradient to 1.4 M NaCl and was buffer exchanged versus20 mM Tris–HCl (pH 7.4) and 0.1 M NaCl. Prorenin wasmixed with immobilized N-tosyl-L-phenylalanine chloro-methyl ketone-trypsin (TPCK trypsin, cat. no. 20230,Pierce Biotechnology,) for 3 h at 20 �C. Mature renin waseluted and loaded onto Blue Sepharose 6 Fast Flow in20 mM Tris–HCl (pH 7.4) and 0.1 M NaCl. The unboundfraction containing renin was buffer exchanged and loadedonto POROS S/20 (cat. no. 1-3022, Applied Biosystems) in10 mM sodium acetate (pH 5.0) and 5.0 mM NaCl. Reninwas eluted with a linear gradient to 1.0 M NaCl and wasdialyzed versus 20 mM Tris–HCl (pH 7.0) and 0.1 M NaCl.Purified renin was concentrated to 12 mg ml�1 and stored
at �70 �C. The final yield typically was 0.2–0.4 mg crystal-lizable renin per liter of prorenin-containing media.
Crystallization of recombinant human renin
Purified renin was crystallized by the vapor diffusion inhanging drop method. Equal volumes (2 ll) of maturerenin (8–12 mg ml�1) and reservoir solution (10–20% poly-ethylene glycol [PEG] 3350, 50 mM sodium citrate [pH4.5], and 0.6 M NaCl) were mixed and equilibrated at20 �C [15]. The crystals belong to the cubic P213 spacegroup with a unit cell dimension of approximately 141 A.Suitable crystals (�0.25 mm on each face) were transferredinto a fresh 4-ll hanging drop containing 2.5 mM com-pound dissolved in well solution with an additional 2–3%PEG 3350. The drop was incubated at 20 �C, typicallyfor 48 h. The crystals were cryoprotected by transferringto a 4-ll drop of 20% ethylene glycol in well solution andthen immersed in liquid nitrogen.
Intensity data collection
X-ray diffraction data of the renin complexes inhibitedwith compounds 1, 3, and 11 (Table 1) were collected at�180 �C at the Advanced Photon Source facility on beam-line 17-ID operated by the Industrial MacromolecularCrystallography Association. The crystals of compounds1 and 11 scattered X-rays to 1.9 and 2.25 A resolution,respectively. The diffraction pattern from the crystal ofcompound 3 was considerably weaker, 2.6 A resolution.Autoindexing and processing of the measured intensitydata were carried out with the HKL2000 software package[16]. The intensity data collection statistics are summarizedin Table 2.
Structure determination and refinement
The crystal structure of renin complexed with com-pound 1 was determined by molecular replacement usingrenin coordinates of the previously determined renin struc-ture (PDB code 2I4Q) as an initial model. The rotation andtranslation searches were carried out with the MOLREPprogram [17], Collaborative Computational Project, num-ber 4. The rotation search provided two outstanding solu-tions with Rfree/r values of 8.54 and 5.68, corresponding toeach of two molecules in the asymmetric unit related by anoncrystallographic twofold rotation axis. The positionof one molecule was determined with a translation functionthat provided an R value of 0.540 (score of 0.60), and thesecond molecule improved the R value to 0.482 (score of0.704). The coordinates from the molecular replacementsolution were further optimized by rigid body refinementfollowed by slow cooling simulated annealing and individ-ual B-factor refinement using CNX 2002 (Accelrys) [18].This decreased the Rwork and Rfree values to 0.275/0.295.Electron (2Fo � Fc) and difference (Fo � Fc) density mapswere used for interactive fitting of protein structures into
electron density using COOT [19] and x-Build [20]. Proteinmodel building was alternated with coordinate minimiza-tion and individual B-factor refinement using Refmac 5.0[21]. The parts of the protein structure that underwent con-formational changes with respect to the initial model were
fixed based on an analysis of the electron and differencedensity maps after deletion and then reconstruction. Thecalculated (2Fo � Fc) electron and (Fo � Fc) difference den-sity maps showed well-defined electron density correspond-ing to the active site inhibitor, which was omitted from the
Table 2X-ray intensity data collection of renin crystals
Compound 1 Compound 3 Compound 11
Space group P213 P213 P213
Cell constantsa = b = c (A) 140.86 141.16 141.53a = b = c (�) 90.0 90.0 90.0Complexes/asymmetric unit 2 2 2Resolution (A) 1.90 2.60 2.24Outmost range (A) 1.97–1.90 2.69–2.60 2.32–2.24Observations 216,560 120,225 333,042Unique reflections 70,222 28,798 45,325Outmost range 6373 2858 4278Redundancy 3.1 4.2 7.3Rsym 0.063 0.125 0.073Outmost range 0.606 0.539 0.627Completeness (%) 96 99 99Outmost range (A) 88 99 94I/r(I) 16.98 11.86 31.48Outmost range (A) 1.01 2.08 3.28
Fig. 1. (A) Costructure of 1 bound to renin showing the Connelly surfaceof the binding site color coded according to electrostatic potential. Redindicates negatively charged surface area, and blue indicates positivelycharged surface area. A large hydrophobic pocket S2 and the mouth of thesmaller hydrophobic S3 subpocket are identified. (B) View of the OMIT(Fo � Fc) electron density map of compound 1 (rotated several degreesfrom panel A) in the renin complex. Contoured at the 2.5r level,compound 1 is shown in atom colors: carbon, green; nitrogen, blue;oxygen, red; and fluorine, orange. Hydrogen atoms are not shown. Ringnumbering of the diaminopyrimidine ring is provided. Also shown areresidues involved in the S2 pocket and S3sp, residues involved in H-bondformation with compound 1, H-bond lengths for the interactions (in A),and the relation of the renin-binding pockets to the compound. (Forinterpretation of the references to color in this figure legend, the reader isreferred to the Web version of this article.)
calculations (Fig. 1B). Placement of compound 1 to elec-tron density was carried out with X-LIGAND (Accelrys)[22]. Solvent water molecules were added periodicallybased on examination of difference density maps usingCOOT and X-SOLVATE. Final protein coordinates werevalidated using PROCHECK [23].
Crystal structures of renin complexed with compounds 3
and 11 were determined by the molecular replacementmethod. Renin coordinates from the refined structure ofcompound 1 were used as the starting model. Protocolsand crystallographic software packages used in structurerefinement and electron density fitting of complexes 3 and11 were identical to those used in structure determinationof renin complexed with compound 1. The OMIT (Fo � Fc)electron density map for compounds 3 and 11 are shown inFigs. 3B and 4B The final refinement statistics are present-ed in Table 3. The coordinates of the structures have beendeposited in the Research Collaboratory for StructuralBioinformatics Protein Data Bank (compound 1, accessPDB code: 2IKO; compound 3, access PDB code: 2IKU;compound 11, access PDB code: 2IL2).
Isothermal titration calorimetry
Isothermal titration calorimetry (ITC) experiments wereperformed using a Microcal VP isothermal titrating micro-calorimeter. Data collection, analysis, and plotting wereperformed using a Windows-based software package (Ori-gin, version 7.0) supplied by Microcal. The titrating micro-calorimeter consisted of a sample and reference cell held inan adiabatic enclosure. The calorimeter was calibrated bycomparing the measured areas of applied heat pulses withknown values. Known and experimentally measured valuesagreed to within 2%. To minimize heat of dilution effectsresulting from differences in buffer composition betweenligand and protein, ligands were dissolved in dialysate buff-er from the final step in the renin purification. Ligand and
Fig. 2. ITC results for 8-ll injections of 204 lM renin into 10 lM compound 3 in 20 mM Tris (pH 7.0) and 100 mM NaCl at 28 �C. Top panel shows thechange in enthalpy per injection of renin into 3. Bottom panel shows the integrated enthalpies and the results from nonlinear regression fitting of theintegrated enthalpies using the equation described in Materials and methods. Thermodynamic parameters from nonlinear regression analysis: N = 0.86,Ka = 2.10 · 107 M�1, DH = �9.44 kcal/M.
protein solutions were degassed prior to analysis. The ref-erence cell was filled with dialysate buffer. Because aqueoussolubilities of ligands was poor, buffered renin at 300 lMwas placed in a 250-ll syringe and 20 lM ligand was placedin the 1.375-ml sample cell. The concentration of renin wasdetermined using an A280 of 1.06 for a 1-mg/ml solution.Typically, 30 injections (4 ll/injection) of renin were madeby a computer-controlled injector into the sample cell filledwith ligand solution. The syringe stir rate was 300 rpm.Heat adsorbed or released with each injection was mea-sured by the calorimeter. Titration isotherms for the bind-ing interactions were composed of the differential heat flowfor each injection. These were integrated to provide theenthalpy change with each injection. Heats of dilutionobtained by injecting renin into final purification bufferwere insignificant. Isotherms fit well to a single-site modelusing an iterative nonlinear least-squares algorithm [11].All parameters were floated during the iterations. Bindingisotherms fit by this method provided the equilibrium asso-ciation or binding constant (Ka), the change in enthalpy(DH), and the stoichiometry of binding (N). Binding stoi-chiometry was 1:1 within experimental error. The changein free energy (DG) and change in entropy (DS) were thendetermined using the following equation:
DG ¼ �RT InKa ¼ DH � TDS; ð1Þ
where R is the universal gas constant, T is the temperaturein degrees Kelvin, and other parameters are as definedpreviously.
Results
The structure and thermodynamic binding parametersof a small molecule renin inhibitor compound 1 are provid-ed in Table 1. IC50 values reported in the table weredetermined in duplicate using a previously described fluo-rescence tGFP assay [24]. This 3,5-difluoro analog containsthe core diaminopyrimidine ring system of the lead identi-fied from high-throughput screening. Gibbs free energy ofbinding, DG = �7.5 kcal/M, for 1 was the weakest of thecompounds listed, as calculated from the equilibrium disso-ciation constant (Kd) of 3.6 lM determined by ITC. Weakaffinity of the lead was consistent with weak renin inhibi-tion of the compound (IC50 = 6.6 lM). The active sitecostructure of 1 bound to renin is shown in Fig. 1 as deter-mined from X-ray crystallography to a resolution of 1.9 ASeveral hydrogen bonds were formed between the diamino-pyrimidine and renin, including H-bonds from the 2-aminogroup to Thr77 and Ser76, an H-bond between Asp215 andthe 4-amino group, an H-bond between Asp32 and the 4-amino group, and an H-bond between Asp32 and thepyrimidine ring nitrogen at position 5. This strong networkof H-bonds contributed to the significant binding enthalpyof �9.5 kcal/M for 1. The large hydrophobic S2 (residuenumbers Ser76, Thr77, Phe242, His290-Met292, andTyr220) and smaller S3 subpocket (S3sp) (residue numbersVal30, Tyr14, Thr12, Tyr155, Ser219, and Gly217) wereunoccupied by 1 (Fig. 1A). Alterations to the core diamino-pyrimidine template were then explored to take advantage
Fig. 3. (A) Costructure of 3 bound to renin showing the Connelly surfaceof the binding site color coded according to electrostatic potential.Extension of the methoxypropyl ether substituent into the S3sp is shown.(B) View of the OMIT (Fo � Fc) electron density map of compound 3 inthe renin complex. Contoured at the 2.5r level, compound 3 is shown inatom colors: carbon, green; nitrogen, blue; oxygen, red; and fluorine,orange. Also shown are residues involved in the S2 pocket and S3sp,residues involved in H-bond formation with compound 3, H-bond lengthsfor the interactions (in A), and the relation of the renin-binding pockets tothe compound. (For interpretation of the references to color in this figurelegend, the reader is referred to the Web version of this article.)
Table 3Refinement summary of deviations from ideality and final refinementparameters and the final R/Rfree values
of these nearby binding pockets to improve affinity andpharmacokinetic parameters of the lead.
Molecular modeling studies indicate that an appropri-ately positioned methoxypropyl group could be extendedinto the S3sp by attaching it to a tetrahydroquinoline ringtethered to the diaminopyrimidine ring of the lead.Compound 2, shown in Table 1, was synthesized basedon modeling results [9]. Extension of the methoxypropylof 2 into S3sp improved ligand affinity significantly, as indi-cated by the thermodynamic parameters listed in Table 1.The equilibrium dissociation constant (Kd) or binding con-stant was 530 nM, and the change in binding enthalpy was�14.5 kcal/M. Improved free binding energy resulted froma DDH of �5 kcal/M compared with 1. Improved affinityalso translated into improved inhibition of enzyme activity,renin IC50 = 691 nM, approximately a 10-fold increasecompared with 1.
Compound 3, resulting from the addition of a phenylring to 2, showed further improvements in binding affinityand enzyme inhibition. Fig. 2 shows a representative bind-ing isotherm obtained by ITC for the interaction of 3 withrenin. Nonlinear least-squares analysis of the these data, as
described in Materials and methods, resulted in the follow-ing thermodynamic parameters: stoichiometry of binding,N = 0.86; Ka = 2.1 · 107 M�1 or Kd = 48 nM; and DH =�9.44 kcal/M. An average Kd of 79 nM for multiple bind-ing experiments was measured by ITC and agreed closelywith the renin IC50 of 58 nM. Binding enthalpy wasimproved only slightly compared with 1, but there was1.8 kcal/M less entropic penalty for the binding interaction.Fig. 3 shows the costructure of 3 bound to the renin activesite as determined by X-ray crystallography at a resolutionof 2.6 A Extension of the methoxypropyl ether into theS3sp was evident. H-bonding to the diaminopyrimidine ringwas similar to 1, and the S2 pocket was unoccupied. Bind-ing thermodynamics for several other molecules with exten-sions into the S3sp were examined, and the results are listedin Table 1. All compounds with S3sp extensions boundtighter and inhibited renin greater than 1, but none ofthe compounds tested bound significantly tighter by ITCthan 3. Substitution of an N-ethyl acetamide for the meth-oxypropyl, compound 6, resulted in a small increase inrenin inhibition and binding affinity. For comparison, thethermodynamics for 5 can be examined because it differedfrom 6 by just a methyl substituent and the S3sp substitu-ent. For all of the compounds analyzed by ITC and listed
Fig. 4. (A) Costructure of 11 bound to renin showing the Connelly surfaceof the binding site color coded according to electrostatic potential.Extension of the methoxypropyl ether substituent into the S3sp is shown.(B,C) View of the OMIT (Fo � Fc) electron density maps of compound 11
bound to renin in the two different binding orientations identified incrystallographic experiments. Contoured at the 2.5r level, compound 11 isshown in atom colors: carbon, green; nitrogen, blue; oxygen, red; andsulfur, yellow. Also shown are residues involved in the S2 pocket and S3sp,residues involved in H-bond formation with compound 11, H-bondlengths for the interactions (in A), and the relation of the renin-bindingpockets to the compound. (For interpretation of the references to color inthis figure legend, the reader is referred to the Web version of this article.)
in Table 1, there was good agreement of the renin-bindingaffinities determined by ITC and renin inhibition, exceptfor 7 where renin IC50 = 7.1 nM and Kd = 186 nM.
Initial ideas for extending structure into the S2 pocketwere provided by NMR binding experiments [25]. Severalsmall compounds that bound renin in the presence of 2
were identified. One of the compounds identified was amethoxyaminophenylbenzamide. Additional NMR experi-ments indicated that the small molecule bound in the S2pocket in the presence of 2 and was positioned closeenough to be synthetically tethered. Compound 9 was syn-thesized based on these data [25]. Compared with the sim-ilar compound 5 (minus the S2 substituent) with a reninIC50 of 132 nM, 9 was less potent with an IC50 of336 nM. This was contrary to what was initially expectedgiven that the substituent should occupy a significant por-tion of the hydrophobic S2 pocket. But thermodynamicsdata for the binding interaction of 9 with renin helped toexplain the loss of activity. Binding affinity of 9 was272 nM compared with the tighter binding affinity of178 nM for compound 5. Interestingly, there was also asubstantial loss in binding enthalpy for 9, only �2.1 kcal/M, compared with �9.3 kcal/M for 5. Based only on thereduced binding enthalpy, binding affinity would have beenexpected to drop more substantially, but the loss was large-ly offset by favorable binding entropy of 6.9 kcal/M for 9.
Analysis of the thermodynamics data for 9 suggestedthat the polar substituents that were buried in the S2pocket were not positioned to form H-bonds; therefore,significant losses in binding enthalpy occurred due to des-olvation of the polar groups that were not recovered onbinding. Thermodynamics data suggested that improvedaffinity could be obtained by incorporation of substituentsin the S2 pocket that correctly position polar functional-ities to interact with the negatively and positively chargedareas in the S2 pocket. Compounds 10 and 11 resultedfrom these initial synthetic efforts, and the thermodynam-ics data are listed in Table 1. Compound 11 with a naph-thylsulfonamide substituent in the S2 pocket resulted inimproved affinity, Kd = 79 nM, and renin inhibition,IC50 = 27 nM. Fig. 4 shows the costructure of 11 boundto the renin active site obtained by X-ray crystallographyat a resolution of 2.24 A Positioned in the back of thebinding pocket, the diaminopyrimidine is positioned sim-ilar to previous analogs making H-bonds to Asp215 andAsp32, with the methoxypropyl substituent extending intoS3sp. The naphthylsulfonamide extends into the largehydrophobic S2 pocket, positioning the oxygen atoms ofthe sulfonamide close to the positive charge at the bottomof the S2 pocket. Two different orientations of thenaphthylsulfonamide ring were detected in the crystalstructure. In one binding orientation (Fig. 4B), the sulfon-amide oxygen atoms form H-bonds with the amide back-bone of Ser219 and the hydroxyl side chain of Thr77. Apeptide flap (renin residues Lys238–Tyr244) near thebinding site was also in a closed conformation. In theother ligand orientation (Fig. 4C), the sulfonamide oxy-gen atoms form H-bonds with the amide backbone ofSer219 and Tyr220. His 290 flipped its orientation toaccommodate the naphthyl ring of 11, and the binding
site flap was in an open conformation with a 4.2-A shiftin the position of Phe242. The electron density was clearfor all of the inhibitor atoms in both orientations,although the electron density of the outermost ring ofthe naphthyl group was not well refined.
Discussion
High-throughput screening of renin uncovered a substi-tuted diaminopyrimidine that inhibited renin activity. Ini-tial synthetic modifications produced 1 with a renin IC50
of 6.6 lM and a Kd of 3.6 lM, as determined by ITC. Sev-eral hydrogen bonds were formed between the diaminopyr-imidine and renin, including H-bonds from the 2-aminogroup to Thr77 and Ser76, an H-bond between Asp215and the 4-amino group, an H-bond between Asp32 andthe 4-amino group, and an H-bond between Asp32 andthe pyrimidine ring nitrogen at position 5. In addition,the binding pocket is negatively charged in the area sur-rounding the pyrimidine ring system and complementsthe basic nitrogen ring system of the inhibitor. There is alarge hydrophobic pocket, S2, and a smaller hydrophobicpocket, S3sp, that are adjacent to the catalytic site, asshown in Fig. 1.
Molecular modeling indicated that a methoxypropylsubstituent could be tethered to the lead template to posi-tion the substituent in the S3sp. The addition of this substi-tuent, resulting in 2, improved renin affinity to 530 nM andIC50 to 690 nM. Compound 2 was also selective versus twoother aspartyl proteases: cathepsin and pepsin. Renin haslow active site homology to other human aspartyl proteas-es. This low homology and lack of cathepsin and pepsininhibition suggested that side effects from inhibition ofother aspartyl proteases would be less likely but that stillneeds to be addressed with in vivo studies. The S3sp exten-sion of 2 altered the binding thermodynamics. Extension ofthe methoxypropyl group into the hydrophobic S3sp wasexpected to expel ordered solvent molecules from the pock-et, resulting in favorable binding entropy. But it was alsoexpected that entropic contributions from the solventexclusion would be partially offset by restricting the fourrotatable bonds of the methoxypropyl ether group.Regarding the enthalpic contribution to binding free ener-gy for the S3sp extension, the formation of favorablehydrophobic van der Waals (VDW) interactions with theligand was expected to increase binding enthalpy. Actualthermodynamics for 2 indicated that restricting the rotat-able bonds outweighed the entropic gain from solventexpulsion, but the enthalpic contribution from hydropho-bic interactions outweighed that entropic penalty, resultingin a net overall gain from the extension into S3sp. Com-pound 2 had a 5-kcal/M increase in binding enthalpy rela-tive to 1, but this gain was offset by a 3.9-kcal/M loss inbinding entropy for a net overall gain of 1 kcal/M.
Binding thermodynamics was also examined by ITC fora compound with a different substituent in S3sp. The N-eth-yl formamide substituent (compound 6) extending into S3sp
provided a small gain in affinity. Compared with 5, the IC50
also improved twofold. But additional gains in affinitycould be obtained for other groups extending into the S3pocket if functionality could be added to the end of thehydrophobic extension that could interact with Tyr155 atthe bottom of the S3 pocket. Compound 4 with a tricyclicring off the diaminopyrimidine bound renin with thermo-dynamics similar to 5.
Compound 3 had one of the best renin affinities of theinhibitors tested. Relative to compound 2, it had approxi-mately a 10-fold gain in renin inhibition and a 7-fold gainin affinity. The change in binding enthalpy for the interac-tion of 3 with renin was �10 kcal/M, which was 4.5 kcal/Mless than that for 2, but TDS was �0.2 kcal/M comparedwith �5.9 kcal/M. Improved binding entropy likely is relat-ed to displacement of additional ordered solvent moleculesfrom the binding pocket by the phenyl substituent makingVDW contact with the protein.
Compound 7 with a 3,5-difluorophenyl substituent hadthe best IC50 of the compounds tested by ITC. Affinitydetermined by ITC was weaker than expected based onIC50, but 7 was tightly bound with a Kd of 186 nM. Com-pound 7 was bound tighter and inhibited renin better com-pared with 8, although the only difference was the methylsubstituent. Structural data indicated that the 3,5-difluor-ophenyl ring of 7 was positioned in closer contact withthe hydrophobic surface of the binding pocket than thatof 8, where the 3,5-difluorophenyl ring was more solventexposed. The 3,5-difluorophenyl group of 7 was positionedto make VDW contact with Pro111. Therefore, there wasincreased binding enthalpy from VDW interactions. Inaddition, the entropic cost of restraining the phenyl groupin a better binding orientation was paid at synthesis, sothere was less entropic penalty on binding.
Perhaps the most interesting substituent effect was thesignificant loss of binding enthalpy and increase in bindingentropy for compounds 9 and 10 that had extensions intothe S2 pocket. Increased binding entropy likely is due todisplacement of ordered solvent molecules from the largehydrophobic S2 pocket that occurs on burial of nonpolarsurface area for the inhibitor. Loss of binding enthalpycould be explained by burial of polar functional groupsinto the S2 pocket that did not form productive interac-tions or H-bonds. Unlike compounds 9 and 10, compound11 with an S2 extension did not lose as much bindingenthalpy but also did not add as much favorable bindingentropy. The sulfonamide oxygen atoms on the naphthyl-sulfonamide group form a polar interaction with the S2pocket; therefore, binding of 11 does not pay the sameenthalpic penalty as does binding of 9 and 10. However,the added polar interaction removes some of the flexibilityof the inhibitor and protein backbone, resulting in lessfavorable binding entropy. In addition, nonoptimal bind-ing interactions of the naphthylsulfonamide group of com-pound 11 in the S2 pocket apparently contributed to thealternative binding modes of this group, as detected inthe crystal structure. In one inhibitor-binding mode, a
peptide flap (residues Lys238–Tyr244) near the renin activesite is in a closed conformation; in the other inhibitor-bind-ing orientation, the flap is in an open conformation. Themixed population of open and closed conformations maynegate a portion of entropic gain due to expulsion ofordered water molecules from the closed loop conforma-tion. Entropy/Enthalpy compensation, such as that seenin compounds 9, 10, and 11, can be difficult to balance inoptimizing functionality for binding interactions in the S2pocket. Results from 11 suggest that further optimizationof substituents to form better positioned polar interactionsin the S2 pocket could enhance binding and thereforeenzyme inhibition.
The current work has shown that a combination oflibrary screening, molecular modeling, and analysis of struc-tural, thermodynamic, and inhibition data provided a valu-able tool for structure-based drug design. Even whenstructural data are not available, thermodynamic data canprovide insight into drug design; however, interpretationof thermodynamic data is more insightful when structuraldata are available. Some insight into drug design that ther-modynamic data provide is intuitive to skilled medicinalchemists, but there are situations that are not apparent with-out thermodynamic measurements. This work used a combi-nation of structural, thermodynamic, and inhibition data todesign unique potent small molecule inhibitors of renin.
Acknowledgment
The authors thank John Bryant for collecting IC50 data.
References
[1] T. Thom, N. Haase, W. Rosamond, V. Howard, J. Rumsfeld, T.Manolio, Z. Zheng, K. Flegal, C. O’Donnell, S. Kittner, D. Lloyd-Jones, D. Goff Jr., Y. Hong, Heart disease and stroke statistics—2006update: a report from the American Heart Association StatisticsCommittee and Stroke Statistics Subcommittee, Circulation 113(2006) e696.
[2] T.J. Wang, R.S. Vasan, Epidemiology of uncontrolled hypertensionin the United States, Circulation 112 (2005) 1651–1662.
[6] N.M. Kaplan, Resistant hypertension, J. Hypertens. 23 (2005) 1441–1444.
[7] A.H. Gradman, R.E. Schmieder, R.L. Lins, Y. Chiang, M.P.Bedigian, Aliskerin, a novel, orally-effective renin inhibitor, providesantihypertensive efficacy and placebo-like tolerability similar to AT1-receptor blocker in hypertensive patients, Circulation 111 (2005)1012–1018.
[8] N. Powell, F. Ciske, C. Cai, W. Cody, D. Downing, D. Holsworth, K.Mennen, C. Van Huis, M. Jalaie, M. Ryan, J. Bryant, W. Collard, S.
Ferreira, P. McConnell, E. Zhang, J. Edmunds, Rational design of 6-(2,4-diaminopyrimidinyl)-1,4-benzoxazinones as small molecule renininhibitors, Presented at the meeting of the American Chemical SocietyMEDI-89, Atlanta, GA, 2006.
[9] D. Holsworth, T. Belliotti, J. Bryant, C. Cai, W. Cody, D. Downing,M. Jalaie, T. Li, N. Powell, P. McConnell, M. Ryan, E. Zhang, J.Edmunds, Discovery of 6-ethyl-2,4-diaminopyrimidine-based smallmolecule renin inhibitors, Presented at the meeting of the AmericanChemical Society MEDI-87, Atlanta, GA, 2006.
[10] D. Holsworth, M. Stier, J. Edmunds, W. He, S. Place, S. Maiti, Anexpeditious synthesis of 6-alkyl-5-(40-amino-phenyl)-pyrimidine-2,4diamines, Synthetic Commun. 33 (2004) 3467–3475.
[11] T. Wiseman, S. Williston, J.F. Brandts, L.N. Lin, Rapid measure-ment of binding constants and heats of binding using a new titrationcalorimeter, Anal. Biochem. 179 (1989) 131–137.
[12] A. Velazquez-Campoy, I. Luque, E. Freire, The application ofthermodynamic methods in drug design, Thermochim. Acta 380(2001) 217–227.
[13] P.C. Weber, F.R. Salemme, Applications of calorimetric methods todrug discovery and the study of protein interactions, Curr. Opin.Struct. Biol. 13 (2003) 115–121.
[14] K.F. Geoghegan, A.J. Lanzetti, M.J. Ammirati, D.E. Danley, B.A.O’Conner, P.M. Hobart, Structure and Function of the AsparticProteases, Plenum, New York, 1991.
[15] L.W. Lim, R.A. Stegeman, N.K. Leimgruber, J.K. Gierse, S.S.Abdel-Meguid, Preliminary crystallographic study of glycosylatedrecombinant human renin, J. Mol. Biol. 210 (1989) 239–240.
[16] Z. Otwinowski, W. Minor, Processing of X-ray Diffracting DataCollected in Oscillation Mode, Academic Press, San Diego, 1997.
[17] A. Vagin, A. Teplyakov, MOLREP: an automated program formolecular replacement, J. Appl. Crystallogr. 30 (1997) 1022–1025.
[18] A.T. Brunger, P.D. Adams, G.M. Clore, W.L. DeLano, P. Gros,R.W. Grosse-Kunstleve, J-S. Jiang, J. Kuszewski, M. Nilges, N.S.Pannu, R.J. Read, L.M. Rice, T. Simonson, G.L. Warren, Crystal-lography and NMR system: a new software suite for macromolecularstructure determination, Acta Crystallogr. D Biol. Crystallogr. 54(1998) 905–921.
[19] P. Emsley, K. Cowtan, COOT: model-building tools for moleculargraphics, Acta Crystallogr. D Biol. Crystallogr. 60 (2004) 2126–2132.
[20] T.J. Oldfield, A number of real-space torsion-angle refinementtechniques for proteins, nucleic acids, ligands, and solvent, ActaCrystallogr. D Biol. Crystallogr. 57 (2001) 82–94.
[21] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Refinement of macro-molecular structures by the maximum-likelihood method, ActaCrystallogr. D Biol. Crystallogr. 53 (1997) 240–255.
[22] T.J. Oldfield, X-LIGAND: an application for the automated additionof flexible ligands into electron density, Acta Crystallogr. D Biol.Crystallogr. 57 (2001) 696–705.
[23] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton,PROCHECK: a program to check the stereochemical quality ofprotein structures, J. Appl. Crystallogr. 26 (1993) 283–291.
[24] D. Holsworth, N. Powell, D. Downing, C. Cai, W. Cody, J.Edmunds, M. Ryan, M. Jalaie, G.L. Bryant, R. Ostroski, E. Zhang,Discovery of novel non-peptidic ketopiperazine-based renin inhibi-tors, Bioorg. Med. Chem. 13 (2005) 2657–2664.
[25] M. Jalaie, W. Cody, S. Emerson, D. Holsworth, P. McConnell, D.Alessi, J. Day, J. Dyer, J.C. Hagadorn, M. Mastronardi, P.Sahasrabudhe, V. Thanabal, E. Zhang, Design of novel pyrimidinebased renin inhibitors by utilizing NMR auxiliary binding assay: S2pocket analogs, Presented at the meeting of the American ChemicalSociety MEDI-86, Atlanta, GA, 2006.
