Development of Novel Peptide-Based Michael Acceptors Targeting

Article pubs.acs.org/jmc

Development of Novel Peptide-Based Michael Acceptors Targeting Rhodesain and Falcipain‑2 for the Treatment of Neglected Tropical Diseases (NTDs) Santo Previti,† Roberta Ettari,*,† Sandro Cosconati,‡ Giorgio Amendola,‡ Khawla Chouchene,§ Annika Wagner,∥,⊥ Ute A. Hellmich,∥,⊥ Kathrin Ulrich,# R. Luise Krauth-Siegel,# Peter R. Wich,∇ Ira Schmid,∇ Tanja Schirmeister,∇ Jiri Gut,○ Philip J. Rosenthal,○ Silvana Grasso,† and Maria Zappalà† †

Department of Chemical, Biological, Pharmaceutical and Environmental Sciences, University of Messina, Viale Annunziata, 98168 Messina, Italy ‡ DiSTABiF, University of Campania Luigi Vanvitelli, Via Vivaldi 43, 81100 Caserta, Italy § Laboratoire de Chimie des Substances Naturelles UR/11-ES-74, Faculté des Sciences de Sfax, Université de Sfax, Route de l’aeroport, 3000 Sfax, Tunisia ∥ Institute of Pharmacy and Biochemistry, University of Mainz, Johann-Joachim-Becherweg 30, DE 55128 Mainz, Germany ⊥ Centre for Biomolecular Magnetic Resonance (BMRZ), Goethe-University Frankfurt, Max-von-Laue-Strasse 9, DE 60438 Frankfurt am Main, Germany # Biochemistry Center, Heidelberg University, Im Neuenheimer Feld 328, DE 69120 Heidelberg, Germany ∇ Institute of Pharmacy and Biochemistry, University of Mainz, Staudingerweg 5, DE 55128 Mainz, Germany ○ Department of Medicine, San Francisco General Hospital, University of California, 1001 Potrero Avenue, San Francisco, California 94110, United States S Supporting Information *

ABSTRACT: This paper describes the development of a class of peptide-based inhibitors as novel antitrypanosomal and antimalarial agents. The inhibitors are based on a characteristic peptide sequence for the inhibition of the cysteine proteases rhodesain of Trypanosoma brucei rhodesiense and falcipain-2 of Plasmodium falciparum. We exploited the reactivity of novel unsaturated electrophilic functions such as vinyl-sulfones, -ketones, -esters, and -nitriles. The Michael acceptors inhibited both rhodesain and falcipain-2, at nanomolar and micromolar levels, respectively. In particular, the vinyl ketone 3b has emerged as a potent rhodesain inhibitor (k2nd = 67 × 106 M−1 min−1), endowed with a picomolar binding affinity (Ki = 38 pM), coupled with a single-digit micromolar activity against Trypanosoma brucei brucei (EC50 = 2.97 μM), thus being considered as a novel lead compound for the discovery of novel effective antitrypanosomal agents.

■

INTRODUCTION

in southern and eastern Africa, which causes a rapid-onset acute form of HAT with a higher mortality rate.3 Malaria is the most widespread and severe tropical infectious disease; in humans, it is caused by several species of the Plasmodium genus, with Plasmodium falciparum being the most dangerous and most prevalent species.4 Although a number of antitrypanosomal and antimalarial agents are available, these suffer from problems with dosing schedules, toxicity, and increasing drug resistance.5−7 Thus, there is an urgent need to identify new effective drugs, ideally directed against novel targets. To address this need, we focused

Neglected tropical diseases (NTDs) are a group of disabling infections particularly endemic in developing regions of Africa, Asia, and the Americas. Over one billion people suffer from one or more NTDs. Two of the most important diseases are human African trypanosomiasis (HAT) and malaria.1 HAT (also known as sleeping sickness) is an endemic parasitic disease which occurs in 36 countries in sub-Saharan Africa, with around 10000 new cases reported each year.2 It is caused by two subspecies of Trypanosoma: Trypanosoma brucei gambiense, particularly widespread in central and western Africa and responsible for a chronic form of the disease, and Trypanosoma brucei rhodesiense, the most common subspecies © 2017 American Chemical Society

Received: March 16, 2017 Published: August 1, 2017 6911

DOI: 10.1021/acs.jmedchem.7b00405 J. Med. Chem. 2017, 60, 6911−6923

Journal of Medicinal Chemistry

Article

rhodesain and an EC50 value of 14.8 μM against cultured T. brucei.

our attention on parasite cathepsin L-like cysteine proteases, notably rhodesain and falcipain-2 (FP-2), which have been recognized as novel promising targets for the treatment of HAT and malaria because of their key roles for parasite survival.8−11 Rhodesain is a Clan CA, family C1 (papain-family) cathepsin L-like cysteine protease that is essential for Trypanosoma brucei survival. Its importance stems from several functions, such as its role in crossing the blood brain barrier, thus inducing the neurological stage of HAT;12 other functions include also turnover of variant surface glycoproteins (VSGs) that coat trypanosomes,13 degradation of host immunoglobulins to reduce the host immune response,14 and degradation of parasite and imported host proteins within lysosomes.15 FP-2, a Clan CA, family C1 cysteine protease of P. falciparum, hydrolyzes hemoglobin to provide amino acids that are essential to the parasite for protein synthesis. FP-2 may also be responsible for the cleavage of the cytoskeletal proteins ankyrin and band-4.1 to facilitate rupture of the red-cell membrane.16,17 Both proteases are characterized by a left (L) and a right (R) domain, with the catalytic triad (Cys/His/Asn) located in a cleft between the two domains. In recent years, our group has been actively involved in the development of novel rhodesain and FP-2 inhibitors for the treatment of HAT18−25 and malaria.26−33 Peptidyl vinyl sulfones were previously identified as potent inhibitors of rhodesain8,9 and FP-2.10 In the present work, we designed a new series of peptide-based inhibitors 1−4a−d (Chart 1) as potential rhodesain and FP-2 inhibitors, starting from the vinyl sulfones K11777 (5) and K11002 (6) (Chart 2) as lead compounds.34−37

Chart 3. Structure of Model Compound 7

Michael acceptors were of interest as inhibitor warheads due to their ability to covalently trap the active site thiol function of cysteine proteases. Herein we report the synthesis and biological investigation of compounds 1−4a−d against rhodesain and FP-2 and against cultured T. b. brucei and P. falciparum. In silico studies were also performed to elucidate at the molecular level the binding mode of our compounds and rationalize the activity of the newly discovered antiparasitic agents.

■

RESULTS AND DISCUSSION Chemistry. The synthesis of Michael acceptors 1−4a−d (Scheme 1) was accomplished starting from the acids 8−11 and the amine 12, synthesized according to literature procedure.38−40 The carboxylic acids 8−11 were coupled to the amine 12 in the presence of HOBt and EDCI, as coupling reagents, to get the dipeptides 13−16 spanning the P1−P3 region and bearing the required terminal olefin for further functionalization via cross-metathesis methodology. These intermediates 13−16 were reacted with the cross-metathesis (CM) partners 17a−d containing the appropriate electronwithdrawing (EWG) group. This synthetic approach allowed us to recover the CM products in excellent yields, particularly for the highly reactive vinyl esters and ketones. The reaction was carried out by employing the Hoveyda−Grubbs second generation catalyst, a homogeneous ruthenium carbene complex easy to handle in air and tolerant toward a large variety of functional groups.41 Furthermore, the use of microwave irradiation to promote the CM reaction resulted in reduced reaction times (from 24 h to 2−4 h) and byproduct formation. Biological Activity. All compounds 1−4a−d were tested against recombinant rhodesain by using Cbz-Phe-Arg-AMC as a fluorogenic substrate.42 First, a preliminary screening at a fixed inhibitor concentration of 50 μM was performed. An equivalent volume of DMSO was used as negative control, and E-64 (18),43 the irreversible standard inhibitor of clan CA family C1 cysteine proteases (papain family), was used as positive control. All compounds except 2d inhibited the enzyme activity by 85−100%. Continuous assays were then performed at seven different concentrations ranging from minimally inhibitory to fully inhibitory concentrations, to determine the first-order rate constants of inhibition kinac (min−1), the dissociation constants Ki (nM), and the second-order rate constants of inhibition k2nd (M−1 min−1), as k2nd = kinac/Ki (Table 1). All compounds were shown to inhibit rhodesain in an irreversible manner, as demonstrated by analysis of the progress curves at seven different concentrations (see, e.g., 1b, Figure 1) and in agreement with literature data.42

Chart 1. Structure of the Michael Acceptors 1−4a−d

Chart 2. Structure of Model Compounds 5 and 6

Considering that rhodesain and FP-2 can tolerate a range of bulky hydrophobic residues within the S2 pocket and exhibit a strong preference for P2 Phe or Leu, respectively, we introduced these residues at the P2 site, with a hPhe inserted at the P1 site, analogous to the structures of both 5 and 6. The N-terminal amino group was protected with a carbobenzyloxy (Cbz) or p-nitro-benzyloxycarbonyl (pNZ) group, the latter in agreement with the structure of inhibitor 7 (Chart 3), synthesized by our group,25 with a Ki value of 6.81 μM toward 6912



Article

Scheme 1a

Reagents and conditions: (a) HOBt, EDCI, dry DMF/CH2Cl2 (1:1), 0 °C, N2, then 10 min DIPEA, 12, rt, 12 h; (b) Hoveyda−Grubbs second catalyst, dry CH2Cl2, 100−150 °C, MW, 2−4 h.

a

Table 1. Activity of the Michael Acceptors 1−4a−d towards Rhodesain

rhodesain

a

compd

R

R′

EWG

kinac (min )

Ki (nM)

k2nd (× 103 M−1 min−1)

1a 2a 3a 4a 1b 2b 3b 4b 1c 2c 3c 4c 1d 2d 3d 4d 18

CH(CH3)2

H NO2 H NO2 H NO2 H NO2 H NO2 H NO2 H NO2 H NO2

COOMe

0.0038 ± 0.0002 0.0023 ± 0.00004 0.0018 ± 0.0001 0.0012 ± 0.0001 0.0016 ± 0.0004 0.0027 ± 0.0009 0.00255 ± 0.00075 0.0017 ± 0.0006 0.0028 ± 0.0003 0.0043 ± 0.0005 0.0033 ± 0.0004 0.0024 ± 0.0005 0.001 ± 0.0001 nda 0.0005 ± 0.0002 0.0004 ± 0.0005 0.009 ± 0.0004

7.5 ± 0.4 17 ± 6 2.0 ± 0.2 3.6 ± 0.02 0.5 ± 0.2 0.9 ± 0.1 0.038 ± 0.011 0.074 ± 0.032 9.1 ± 2.6 14 ± 3 5.0 ± 0.6 7.2 ± 0.03 53 ± 1 nda 46 ± 3 1743 ± 3 35 ± 5

509 ± 2 152 ± 20 919 ± 51 347 ± 44 3795 ± 834 2969 ± 1152 67000 ± 432 24354 ± 1873 342 ± 129 304 ± 20 663 ± 15 334 ± 71 19 ± 2 nda 12 ± 6 0.26 ± 0.03 261 ± 27

Ph CH(CH3)2 Ph CH(CH3)2 Ph CH(CH3)2 Ph

COMe

SO2Ph

CN

−1

Compound 2d did not pass the initial screening.

than the corresponding inhibitors with a pNZ group at the Nterminal moiety (e.g., 3b vs 4b, 3c vs 4c, 3a vs 4a). Overall, the most potent compounds toward rhodesain were the vinyl ketones 3b and 4b which showed k2nd values of 67000 × 103 M−1 min−1 (3b) and 24354 × 103 M−1 min−1 (4b), coupled with the highest binding affinity expressed by Ki values of 38 and 74 pM, respectively. Selectivity assays were also performed by testing the active inhibitors against a papain-family human cysteine protease, i.e.,

With the exception of nitriles 1d, 3d, and 4d, which inhibited rhodesain with poor k2nd values ( sulfones > nitriles. Regarding the P2 site, generally, inhibitors with a Phe residue were more active than the corresponding Leu derivatives (e.g., 4b vs 2b, 3b vs 1b, 3a vs 1a). Concerning the N-protecting group, the Cbz-derivatives were more potent 6913



Article

Table 3. Antitrypanosomal Activity of the Most Active Michael Acceptors T. b. brucei EC50 (μM) compd 1a 3a 4a 1b 2b 3b 4b 1c 3c 4c 19 (chlorhexidine)

Figure 1. Progress curves of substrate hydrolysis in the presence of the inhibitor 1b. Inhibitor concentrations (from top to bottom): 0, 0.005, 0.01, 0.025, 0.05, 0.1, 0.5, 1 μM.

a

cathepsin L, on the basis of the high structural homology between cathepsin L and rhodesain. An equivalent volume of DMSO was used as negative control, and 18, also in this case, was used as positive control. The results of the evaluation indicated that all tested compounds inhibited cathepsin L to a lower extent with respect to rhodesain (Table 2). The best selectivity was generally

compd

kinac (min )

Ki (nM)

1a 2a 3a 4a 1b 2b 3b 4b 1c 2c 3c 4c 1d 2d 3d 4d 18

0.0032 ± 0.0002 0.0014 ± 0.00005 0.0040 ± 0.0004 0.0016 ± 0.00005 0.0013 ± 0.0001 0.0026 ± 0.00005 0.002 ± 0.0004 0.0016 ± 0.0002 0.0009 ± 0.00005 0.005 ± 0.00005 0.0014 ± 0.0009 0.003 ± 0.0002 0.0012 ± 0.00005 nd 0.0022 ± 0.00005 nd 0.0032 ± 0.0001

40 ± 7 41 ± 3 27 ± 7 25 ± 0.2 110 ± 40 3.5 ± 0.1 0.4 ± 0.1 2.2 ± 0.2 0.61 ± 0.05 250 ± 20 290 ± 87 45 ± 5 330 ± 67 nd 240 ± 11 nd 30 ± 3

k2nd (× 10 M

−1

5.02 ± 2.67 8.28 ± 6.41 a 3.20 ± 1.77 2.48 ± 1.33 2.97 ± 1.19 4.98 ± 1.45 2.62 ± 0.76 18.5 ± 9.81 4.07 ± 0.85 0.53 ± 0.16

1.72 7.24 60.32 2.61 1.98 2.16 1.36 0.46 9.45 0.49 0.13

The compound was unstable after 24 h.

Table 4. Activity of Compounds 1−4a−d against FP-2 and P. falciparum

human cathepsin L 3

48 h

± ± ± ± ± ± ± ± ± ± ±

reactivity was similar to that against rhodesain: ketones > sulfones > esters > nitriles. The active compounds displayed IC50 values in the range of 0.11−50 μM, with the strongest inhibitors the vinyl ketones 1b and 2b (IC50 values of 0.25 and 0.11 μM, respectively). Generally, all compounds active against FP-2 showed activity against cultured P. falciparum in the low micromolar range, indicating correlation between protease inhibitory and antiparasitic properties (Table 4). Further, inhibition of FP-2

Table 2. Activity of the Michael Acceptors 1−4a−d towards Human Cathepsin-L −1

6.08 12.9 97.2 3.85 3.23 3.18 6.05 2.50 23.7 5.48 0.67

24 h

−1

min )

81 ± 10 35 ± 2 153 ± 23 64 ± 1 13 ± 4 739 ± 2 5754 ± 523 737 ± 14 1571 ± 213 22 ± 1 4.4 ± 1.9 65 ± 2 3.8 ± 0.6 nd 9.3 ± 0.2 nd 110 ± 6

shown by vinyl ketones 1b, 3b, and 4b and by vinyl sulfones 2c and 3c; in these cases, in fact, the second-order rate constants of inhibition for cathepsin L are 1−2 orders of magnitude lower than those observed for rhodesain inhibition. The most potent rhodesain inhibitors were tested for their antiparasitic activity against cultured T. b. brucei and, with the exception of compounds 4a and 3c, all derivatives displayed single-digit micromolar activity (Table 3). The most active compounds against cultured trypanosomes were the vinyl ketones 2b and 3b, with IC50 values of 2.48 and 2.97 μM, respectively. Compounds 1−4a−d were also tested against recombinant FP-2, the obtained IC50 values (Table 3) revealed a weaker interaction compared to that with rhodesain. The trend of

compd

FP-2 IC50 μM

P. falciparum EC50 μM

1a 2a 3a 4a 1b 2b 3b 4b 1c 2c 3c 4c 1d 2d 3d 4d 18 20 (Artemisinin) 21 (Chloroquine)

9.53 ± 0.095 9.63 ± 7.15 >50 >50 0.25 ± 0.03 0.11 ± 0.04 3.01 ± 0.22 0.78 ± 0.17 0.78 ± 0.069 1.32 ± 0.22 7.33 ± 1.02 >50 >50 >50 >50 >50 0.25 ± 0.06

4.40 ± 1.62 3.97 ± 0.72 >10 >10 3.16 ± 1.14 2.74 ± 1.02 7.03 ± 1.65 7.18 ± 0.81 1.53 ± 0.091 2.31 ± 0.01 >10 >10 >10 >10 >10 >10 0.023 ± 0.002 0.057 ± 0.009

and cultured parasites was consistently better with a leucine rather than phenylalanine moiety at P2, consistent with the known inhibitory specificity of the protease.44,45 Effect of Compound 3b on the Thiol Level of T. brucei. In trypanosomes, trypanothione [T(SH)2] is the predominant cellular thiol. Under culture conditions, bloodstream (BS) T. brucei contain about 350 μM T(SH)2 and 100 μM glutathione (GSH).46 To assess the effect of the Michael acceptors on the low molecular weight thiols, parasites were treated for 24 h with compound 3b. The cells showed a concentration-dependent 6914



Article

growth retardation (Figure 2a) corresponding to the EC50 value of 3 μM (Table 3). Cellular GSH and T(SH)2 concentrations

Table 5. Cytotoxicity of 1b−3b and 1c towards HeLa Cell Lines and Selectivity Index (SI)a compd 1b 2b 3b 1c a

CC50 [μM]

SI

± ± ± ±

3 3.3 2.6 4.6

9.58 8.20 7.90 12.17

1.16 0.08 0.26 0.90

SI: EC50 (T. b. brucei)/CC50 (HeLa cells).

compounds covalently binds to the protease, a covalent docking protocol devised by Olson and colleagues was employed.50 In particular, AutoDock4.2 software51,52 was used by exploiting the possibility to include a flexible side chain during the docking simulation. This method requires the intended ligand to be joined in an arbitrary conformation with the residue which participates in a covalent bond, in this case Cys25, and then the calculations to be run by treating the modified residue as flexible. To probe the predictive power of the protocol, the redocking of the cocrystal 6 ligand was performed. The software correctly managed to recapitulate the conformation of the crystal structure with a root-mean-squared deviation (rmsd) of 2.37 Å compared to the crystal structure; notably, the sole significant discrepancy, between the docked binding pose and the experimental one, was in the position of the hPhe in P1 (data not shown), which, by pointing outward, does not establish any relevant contact with the nearby residues in the crystal or in the docked pose. Therefore, we were encouraged to proceed with the in silico tests of 1−4a−d. Analysis of the results achieved for these docking calculations revealed that the binding modes of our compounds closely resemble that of 6, interacting with the same set of key residues. As representative of the obtained binding poses, the docked solution achieved for the most active ligand 3b is represented in Figure 3a,b.

Figure 2. Thiol levels in BS T. brucei treated with compound 3b for 24 h. Cells were seeded to 1 × 105 cells/mL. (a) Cell density after 24 h cultivation. (b) Cellular content of free thiols (GSH, T(SH)2) and free thiols plus disulfides (GSSG, TS2) quantified by HPLC analyses of mBBr-derivatized samples. All values are the mean ± SD of three independent experiments.

were measured by HPLC analysis of the monobromobimane (mBBr)-labeled thiols. The untreated control contained about 340 μM T(SH)2 which dropped to 160 μM upon treating the cells with 3 μM of compound 3b. The GSH level was only marginally affected. The higher reactivity of T(SH)2 may be explained by its thiol pK value of 7.4 which is much lower than that of GSH.47 In all samples, both thiols were almost exclusively present in reduced form, indicating that compound 3b did not cause oxidation of the free thiols. The level of protein-bound GSH, which is ∼5 μM in BS T. brucei,46 was not affected (data not shown). The drop of free thiols, mainly T(SH)2, does not, however, explain the EC50 value of 3 μM observed for compound 3b (Table 3). Depletion of trypanothione synthetase in BS T. brucei results after 24 h in a decrease of T(SH)2 to about 50 μM without affecting cell growth.46 This is supported by previous work showing that a severe depletion of T(SH) 2 is accompanied by only minimal growth retardation.48 Clearly, the parasites can survive with very low levels of T(SH)2, although under these conditions, they are more sensitive toward oxidative stress.49 Finally, cytotoxicity of the most active compounds at cellular level 1b−3b and 1c was evaluated in HeLa cell lines (Table 5): these data demonstrated that the tested compounds have a selectivity index ranging from 2.6 to 4.6, in agreement with the profile of irreversible inhibitors.42 Molecular Modeling. To rationalize the structure−activity relationship data (SARs) of the tested compounds, we performed a series of in silico experiments, consisting of docking calculations and molecular dynamics (MD) simulations. To this end, the crystal structure of rhodesain bound to the inhibitor 6 (PDB 2P86)35 was selected. As this set of

Figure 3. (a) Superimposition of cocrystal ligand 6 (green) and 3b docked conformation (pink) in rhodesain (blue). (b) Rhodesain/3b complex. The enzyme is depicted as blue ribbons and sticks and the ligand as pink sticks. Important residues are labeled. H-bonds are shown in green dashed lines.

Notably, the binding of 3b, as well as of the other compounds, seems to be governed by the occurrence of several H-bond interactions with the residues lining the enzyme binding cavity. In particular, the backbone atoms of the Phe and hPhe residues of compound 3b are involved in H-bonds with the backbone atoms of rhodesain Gly66 and Asp161, respectively. In addition, most of the warhead moieties are favorably positioned to accept two H-bonds from the Trp184 and Gln19 side chains. Indeed, the rigid nitrile warhead (1−4d) would not adopt a pose which is conducive to the formation of these two H-bonds, likely explaining the considerably lower 6915



Article

Table 6. H-Bonds Statistics from the MD Simulation enzyme residue atom

ligand atoms

Focca

AvgDistb (Å)

AvgAngc (deg)

Asp161−backbone CO Gly66−backbone CO Gln19−side chain NH2 Trp184−side chain NH Gly66−backbone NH

hPhe NH Phe NH warhead CO warhead CO Phe CO

0.8554 0.7246 0.6410 0.5489 0.4391

2.82 2.84 2.84 2.86 2.88

163.64 156.88 160.35 158.56 154.55

a

Frequency of occurrence of the H-bond in the frames of the simulation. bMean distance (in Å) of the H-bond along the simulation. cMean angle (in deg) H-bond recorded along the simulation.

activity of the compounds featuring nitriles (compare 1−4b vs 1−4d). As for the P2 moiety (Leu or Phe residues of the inhibitors), our docking studies confirmed that the bulky Phe residue provides a better fit into the hydrophobic S2 binding site which comprises Ala208, Leu160, Ala138, Met68, and Leu67 residues.53 This likely explains the increased activity of the Phe-containing derivatives against rhodesain and cultured trypanosomes, compared to those featuring a Leu in the P2 site (see 1b vs 3b). Of note, consistent with data for FP-2 bound to peptidyl inhibitors,45 Leu was preferred at P2 for inhibition of FP-2 and cultured malaria parasites. The carbobenzyloxy group is lodged in the S3 pocket,53 establishing a T-shaped chargetransfer interaction with the rhodesain Phe61 side chain. Interestingly, in compounds 2,4a−d, the p-nitro substituent of the carbobenzyloxy group seems to have two opposite effects on the stability of the ligand/protein interactions. Specifically, while enhancing the above-described charge-transfer contacts, the same partially negative group is unfavorably pointing toward a negatively charged region of the S3 pocket (Asp60), which likely explains why the nitro group does not substantially contribute to ligand potency. As seen in the cocrystal ligand and for 3b and its structural congeners, the hPhe residue in P1 does not have a critical role in stabilizing the ligand/protein interactions because it points toward the outer part of the protein active site. When analyzing the Ki constants (Table 1), it is clear that, at least for our set of compounds, the different potencies cannot be entirely ascribed to the different reactivities of the employed warheads. Instead, we suggest that the contacts established by the warheads (e.g., H-bonds) are instrumental in stabilizing the ligand in a pose which is conducive for the formation of the covalent bond with the reactive cysteine residue (Cys25). In this regard, the rigid nitrile-containing analogues were the least active ones because they cannot adopt the proper orientation within the enzyme cleft. On the other hand, the ketone moiety seems to be well positioned to fulfill steric and electrostatic demands, which likely explains the higher potencies of these compounds 1−4b. Additionally, to overcome the hurdle of protein rigidity in docking and to probe the stability of the predicted ligand/ enzyme interaction, the covalent 3b/rhodesain complex was subjected to an 80 ns long MD simulation. Table 6 reports the data regarding the stability of key H-bonds that 3b forms with the enzyme residues. Out of five H-bonds, four of them were maintained for more than 50% of the frames of the simulation, further supporting the predicted contacts. To rationalize the recorded selectivities of the compounds for rhodesain over FP-2 we also covalently docked 3b into FP-2 using the crystal structure of the enzyme in complex with an inhibitor (PDB 2OUL).54 The obtained binding pose is presented in Figure 4.

Figure 4. Predicted FP-2/3b complex. The enzyme is depicted as yellow ribbons and sticks, and the ligand is depicted as pink sticks. Important residues are labeled. H-bonds are shown as green dashed lines.

Interestingly, while the overall binding pose of 3b with FP-2 is virtually superimposable to that of rhodesain, a number of changes were found in the S2 pocket of FP-2, which is known to play a key role in determining selectivity in this class of proteases.55 Specifically, Ala208, Ala138, and Gly163 of rhodesain are replaced in FP-2 by aspartate, serine, and alanine residues, respectively. These modifications make the cleft narrower and indeed more polar, thereby inducing the bulky and hydrophobic 3b Phe in the P2 position to establish less favorable contacts with the enzyme. Varied interactions with these residues may explain the observed loss of potency against FP-2 compared to that against rhodesain. Interestingly, similar results were also achieved by other authors.44,56,57 To assess the drug-likeness of compound 3b, its physicochemical and pharmacokinetic (PK) features were profiled through the Qikprop software (Schrödinger. LLC New York) by employing the same parameters used in a previous work24 Qikprop predicts a number of PK parameters and compares them with the ranges calculated for the 95% on the marketed drugs. The results, summarized in Table 7, show that 3b would feature a substantial gut/blood barrier permeability (see QPPCaco values) and an acceptable probability of passing the blood/brain barrier according to its QPlogBB and QPPMDCK values. Furthermore, 3b is predicted to have a 100% human oral absorption. According to this analysis, our inhibitor is well suited to be a lead compound for the discovery of novel antitrypanosomal agents.

■

CONCLUSIONS In summary, we designed and synthesized novel peptide-based Michael acceptors, which were shown to be nanomolar/ picomolar irreversible inhibitors of rhodesain with antitrypanosomal activity. Activity against FP-2 was lower, but multiple compounds demonstrated low micromolar activity against 6916


cultured P. falciparum. The vinyl ketone 3b, with a picomolar binding affinity and impressive k2nd value toward rhodesain, in addition to its good antiparasitic activity, represents a new chemotype for the development of novel antitrypanosomal agents. Further structural variations may be necessary to improve the cell permeability of test compounds, potentially leading to an enhancement of antiparasitic activity.

maximum is 4 1

Number of violations of Lipinski’s rule of five. The rules are mol_MW < 500, QPlogPo/w < 5, donorHB ≤ 5, accptHB ≤ 10. Compunds that satisfy these rules are considered drug-like.

−2.0−6.5 500 great −3.0−1.2 500 great >80% is high,

Development of Novel Peptide-Based Michael Acceptors Targeting

Recommend Documents