Design, synthesis and evaluation of 2-(4-(substituted benzoyl)-1,4-diazepan-1-yl)-N-phenylacetamide derivatives as a new class of falcipain-2 inhibitors

1 Introduction

The increasing ratio of malaria in terms of morbidity and mortality in humans turns it into a major public health problem in tropical and subtropical countries. World Health Organization Malaria report 2012 indicates that in 2010, there were 219 million malaria cases leading to approximately 660,000 malaria deaths, mostly among African children. Out of reported 90% malaria casualties in Africa alone the majority were children under five (Vangapandu et al., 2007). Globally, 80% of malaria deaths occur in just 14 African countries. Together, the Democratic Republic of the Congo and Nigeria account for over 40% of the estimated total of malaria death worldwide (World Malaria Report, 2012). Rapid spread of resistance is becoming most apparent in Plasmodium falciparum species toward first-line treatment, chloroquine and sulfadoxine-pyrimethamine (Marfurt et al., 2010). To enhance the efficacy and delayed onset of resistance, the WHO began recommending for use of Artemisinin based combination therapies (ACTs) since 2005 (Rogerson and Menendez, 2006). ACTs have potential clinical efficacy, paradoxically the history of antimalarial chemotherapy predicts that it is a matter of time before parasitic resistance emerges and spreads (Ekland and Fidock, 2008). Nevertheless, safe and cost effective novel chemical classes of antimalarial agents are urgently needed to treat malaria (Guerin et al., 2002).

Among various potential new targets, the cysteine protease falcipain-2 (FP-2) of P. falciparum is the most intensely studied enzyme, and its structural and functional data suggest that it is an attractive target for therapeutic intervention (Pandey et al., 2005; Wang et al., 2006). FP-2 is a principal cysteine protease that plays a major role in parasite food assimilation by its ability to degrade hemoglobin. Many in vitro studies have affirmed that inhibitors of falcipain-2 can inhibit the parasite hemoglobin hydrolysis and prevent the growth of culture parasites (Shenai et al., 2003; Lee et al., 2003; Domínguez et al., 1997; Huang et al., 2002). Some of them were also effective against murine malaria in vivo (Domínguez et al., 2005; Rosenthal et al., 1993; Sajid and McKerrow, 2002). However, the literature indicates that most of the reported FP-2 inhibitors are mainly derived from peptide analogues (Choi et al., 2013; Mallik et al., 2012; Schulz et al., 2007; Schirmeister and Kaeppler, 2003; Shenai et al., 2003; Lee et al., 2003), which have expressed nanomolar IC₅₀ values, due to the formation of covalent interactions between the electrophilic groups, such as aldehydes, nitriles, vinyl sulfones and epoxides with thiolate of the catalytic cysteine amino acid. In addition, most of the existing standard ligands for falcipain-2, and cystine protease inhibitors such as Leupeptin (N-acetyl-l-leucyl-l-leucyl-l-argininal) (Moldoveanu et al., 2004), and E-64 (N-(trans-epoxysuccinyl)-l-leucine 4-guanidinobutylamide) (Moldoveanu et al., 2004; Varughese et al., 1989) possess chiral center(s), (Fig. 1) which increases the synthetic cost of these drugs.

Close

Figure 1 Existing standard ligands for falcipain-2 inhibition.

Export to PPT

This discussion clearly stresses the requirements of non-peptidic inhibitors that would bind non-covalently to the target enzyme in order to reduce toxicity and cost while retaining the potential for high activity and selectivity. The present study demonstrates the design, synthesis and in vitro activity of 2-(4-(substituted benzoyl)-1,4-diazepan-1-yl)-N-phenylacetamides as falcipain-2 inhibitors. Compounds 5(a–t) were synthesized through the route outlined in Scheme 1. Furthermore, docking studies were performed for the active compounds to predict their relative binding affinities and binding modes in the active site of the falcipain-2 enzyme. Molecular docking studies of most active analogues (5g, 5h, 5j, and 5k) revealed that they interacted with Gln 36, Gln 83, Asn 173 and Hip 174 residues of 3BPF protein, via hydrogen bonding. Moreover, in silico pharmacokinetics and toxicity (ADMET) parameters for the selected analogues were estimated using online tool admetSAR and QikProp module of Schrodinger.

Close

Scheme 1 Synthetic route for target compounds 5(a–t). Reagents and conditions: (i): Chloroacetylchloride, TEA, DCM, 0 °C-RT, 20 min; (ii): N-Boc-homopiperazine, K₂CO₃, CH₃CN, 100 °C, 5 h.; (iii): TFA/DCM, 0 °C-RT, 6 h; (iv): Various carboxylic acids, EDC·HCl, HOBt/DIPEA-DCM, 0 °C-RT, 6 h.

Export to PPT

2 Materials and methods

3 Results and discussion

Existing falcipain-2 inhibitors, (Desai et al., 2004; Li et al., 2009) were used as a template for making a pharmacophore model for falcipain-2 inhibitors, and some of the compounds (I–VIII), which possess moderate to potent activities are represented in Fig. 2. Among the represented compounds, particular imine (I) and phenyl hydrazones (IV and VIII), probably inhibit the enzyme by covalent interactions.

Close

Figure 2 Structures of some reported falcipain-2 inhibitors.

Export to PPT

The most common features present in the aforementioned falcipain-2 inhibitors are, an aromatic residue (monocyclic/bicyclic) which is attached to the hydrophobic moiety; commonly an aromatic residue through a hydrogen bond donor and acceptor atom(s) as linker. The distance between the aromatic residue and the hydrophobic group ranged from 9 to 14 Å. The hydrogen bond donor (HBD) and hydrogen bond acceptor (HBA) atom(s) are present as either in heterocyclic/alicyclic or open chain form. The numbers of hydrogen bond donor and acceptor atoms range from 0 to 2 and 2 to 6, respectively. The reported molecules are basic in nature due to 2° or 3° amino moiety. By considering these common features as pharmacophore for falcipain-2 inhibitors, a pharmacophore model was built as shown in Fig. 3. Among the broad range of heterocyclic templates particularly, the piperazine core is found as the most encouraging leading heterocyclic ring, present in a number of drugs and clinical candidates that address a broad spectrum of serious targets (Han et al., 2012). Therefore, initially we synthesized some derivatives based on piperazine nucleus, and tested in vitro against falcipain-2 enzyme; unfortunately, the insignificant activity of the compounds (unpublished observation), prompted us to focus our efforts to increase the ring size especially, on utilizing 1,4-diazepam as a core nucleus (Ettari et al., 2009; Micale et al., 2006), with the aim of improving the drug like profile of this novel class of compounds, as shown by basic structure 5(a–t) in Scheme 1.

Close

Figure 3 Pharmacophoric features for falcipain-2 inhibitors.

Export to PPT

The least energy conformation (three minimum energy conformations for each compound) for each designed compound was generated by ACDLABS-12.0 product version 12.01/3D viewer (CHARMM parameterizations), and the pharmacophoric distances were measured from the centroid of an aromatic residue to a hydrophobic residue. The observed distances between the pharmacophoric elements of all the designed compounds are in agreement with our proposed pharmacophore model. To achieve the better pharmacokinetic profile, Lipinski’s Rule of Five (Lipinski et al., 1997) was adopted for the designed molecules. Lipophilicity is an important parameter to be considered while designing ligand to manifest drug-like behavior. Thus, Log P values of all the designed molecules were calculated utilizing JME Molecular Editor (Courtesy of Peter Ertl, Novartis).

The target compounds were prepared as outlined in Scheme 1. First, compound 4 was synthesized in multi-gram scale from the starting material aniline in a sequence of reactions. Chloroacetyl chloride was subjected to nucleophilic substitution reaction with aniline, which afforded compound 2. This intermediate was reacted with N-Boc protected homopiperazine, to obtain compound 3. Subsequently, deprotection of the Boc group with trifluroacetic acid furnished the intermediate 2-(1,4-diazepan-1-yl)-N-phenylacetamide 4. This key intermediate was coupled with appropriate carboxylic acids in the presence of coupling agents 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC·HCl) and 1-hydroxybenzotriazole (HOBt) under nitrogen atmosphere to afford target compounds 5(a–t). Synthesized compounds were isolated as pure and characterized by ¹H NMR, mass spectroscopy, HPLC and elemental analysis. The analytical and spectral data of the compounds confirmed the structures and purities of the final compounds.

All synthesized compounds were evaluated for their in vitro falcipain-2 inhibitor activity. Several compounds showed significant inhibitory activity (>60%), against falcipain-2 at 10 μM presented in Table 1, and their chemical structures and physical constants are shown in Table 2.

Retaining the common 2-(4-(substituted benzoyl)-1,4-diazepan-1-yl)-N phenylacetamide framework, compounds 5a, 5b were screened, initially. Fortunately, these two compounds 5a and 5b exhibited inhibitory activity against falcipain-2 enzyme at 10 μM concentrations with inhibition values of 49% and 61%, respectively. Based on its moderate potency and synthetic feasibility, compound 5b served as a solid starting point for the future drug discovery program. Thus, various substitutions were introduced around the scaffold based on 5b specifically in the aromatic moiety. The effect of the fluorine group, an electron withdrawing substituent, was investigated at position 2 on the phenyl moiety (compound 5c). This modification did not result in any improved potency, and compound exhibited 60% enzyme inhibition value close to the compound 5b. Placement of another electron withdrawing substituent such as chloro in the aromatic ring, generated molecules (5d, 5e, 5f) with less potency (47% inhibition by 5d and 5e, and 36% by 5f). When a strong electron withdrawing group, trifluoromethyl was introduced at the 3 position to get the compound 5g (72% inhibition) and position 4 to get the compound 5h (68% inhibition), both compounds showed higher potency than the hit compound 5b. However, investigation of a trifluoromethyl group at the 2 position (compound 5i) of the phenyl ring showed lesser potency (50% inhibition) compared to compound 5b.

Consequently, a methoxy group, an electron releasing substituent was introduced at 2 and 3 positions of the phenyl ring resulting in compounds 5j and 5k, with improved inhibition potency (70% and 64%), greater than the hit compound 5b from this series. Attachment of the methoxy group in the 4 position of the phenyl ring gave rise to compound 5l, with inhibition potency (55%), which was lesser than that of 5b. Compounds 5m, 5n and 5o achieved by replacement of the methoxy group with the ethoxy group lose (32%, 36% and 30%) potency markedly. Replacement of the methoxy group in the phenyl ring with another weak electron releasing group such as methylene (5p, 5q and 5r) at 2, 3 and 4 positions and the ethylene (5s and 5t) group at 3 and 4 positions leads to a decrease in the potency.

This discussion clearly indicates that the strong electron withdrawing group (trifluoromethyl) at 3 and 4 positions and a strong electron releasing group (methoxy) at 2 and 3 positions generated compounds that displayed marked inhibition as evidenced by compounds 5g, 5h, 5j, and 5k compared to the non-substituted derivatives i.e., compound (5a). Higher homologation of electron releasing groups, methoxy to ethoxy (5m, 5n, 5o) and methylene to ethylene (5s, 5t) in the phenyl ring generated compounds with lesser inhibition.

The prepared ligands were docked with proteins using extra precision mode (XP). The best docked pose obtained from Glide was analyzed. Results of docking simulations using Glide are summarized as XP Gscore, lipophilic energy, H-bond energy and electrophoretic energy in Table 3. Comparison of interacting residues of these analogues (5g, 5h, 5j, 5k) shows that the amino acids Ala 175, Trp 43, Ile 85, Leu 84, Cys 42, Trp 206, Val 152 and Asp173 of falcipain-II protein are most commonly involved as shown in Fig. 4(A–D). Among these, the residue Asp 173 was the hydrogen bonding common interacting residue of falcipain-II with ligands (5g, 5h, 5j, 5k). Compound 5g showed eight hydrophobic interactions, Ala 175, Trp 43, Ile 85, Leu 84, Leu172, Cys 42, Trp 206, Val 152; two hydrogen bonds, Gln 36 (2.08 Å), Asn 173 (2.19 Å). Compound 5h showed nine hydrophobic interactions, Ala 175, Trp 43, Ile 85, Leu 84, Leu 172, Cys 42, Trp 206, Val 152, Ile 84; two hydrogen bonds, Asn 173 (2.25 Å), Gln 36 (2.27 Å). Compound 5j showed seven hydrophobic interactions, Ala 175, Trp 43, Ile 85, Leu 84, Cys 42, Trp 206, Val 152; two hydrogen bonds, Asn 173 (2.34 Å), Hip 174 (2.06 Å). Compound 5k showed nine hydrophobic interactions, Ala 175, Trp 43, Ile 85, Leu 84, Leu 172, Cys 42, Trp 206, Val 152, Val 150; three hydrogen bonds, Asn 173 (2.51 Å), Hip 174 (1.93 Å), Gln 83 (2.49 Å).

Close

Figure 4 (A–D) Predicted binding poses of 5g, 5h, 5j, 5k respectively docked to chain A of falcipain-II protein (3BPF.pdb). The red dashed line represents the possible hydrogen bond and green dashed line represents the possible hydrophobic interaction.

Export to PPT

In the present study, validation of the docking was done by removing co-crystallized ligand E-64 from the active site and subjecting it again to dock into the binding pocket in the conformation found in the crystal structure. It is important to note that E-64 is a covalent inhibitor of several cysteine proteases and is potent non-selective due to the formation of covalent bond with the active site thiol (Li et al., 2009; Schiefer et al., 2013). Therefore, to validate the docking protocol, we used E-64 with an open epoxide ring as reported in the literature (Mugumbate et al., 2013). Furthermore, a set of studies were performed to compute the RMSD value between the pose of co-crystallized inhibitor present in the enzyme and its best ranked pose (Fig. 5), in the protein. As a result, the RMSD value calculated for co-crystallized ligand, E 64 was <2 Å against the target enzyme. The best scoring pose of the reference drug inside the FP-2 showed that interactions are in resemblance with the reported X-ray pose interactions (Kerr et al., 2009). These results suggested that our docking procedure, and software protocol could be relied on to predict the experimental binding mode of the designed analogues.

Close

Figure 5 Best docked pose of E-64 inside falcipain-2. The pink dashed line represents the possible hydrogen bond and green dashed line represents the possible hydrophobic interaction.

Export to PPT

As discussed earlier, physicochemical properties such as molecular weight (180–500), log P_o/w (−0.4 to 5.6), of all the designed analogues follow Lipinski’s Rule of Five. Furthermore, Qik–prop module of Schrodinger, and online software admetSAR (http://www.admetexp.org/predict/), were used to generate in silico pharmacokinetics and physicochemical parameters for compounds 5g, 5h, 5j and 5k in order to estimate their drug-likeness properties. The recommended range (Ntie-Kang et al., 2013; Cheng et al., 2012) of some physicochemical and pharmacokinetic parameters, which plays an important role to both pre-clinical research and clinical development, such as: log S (−6.0 to 0.5), log BB (−3.0 to 1.0), % Human oral absorption (⩽25% is poor and ⩾80% is high) was predicted (Table 4). In addition, AMES toxicity, acute oral toxicity and carcinogenicity were also predicted and represented as qualitative terms with probability (in bracket) in Table 4. Range of Caco-2 cell permeability gives a view to select a molecule as lead compound at early stage by determining the capability to transport about across the cell membrane of the gut. The physicochemical properties and in silico predicted ADMET properties as mentioned in Tables 2 and 4, for selected analogues are in the acceptable range of druggable profile.

Although, representatives of this new series did not exhibit strong falcipain-2 activities as compared to the existing nonpeptidic FP-2 inhibitors such as chalcones (Li et al., 1995; Liu et al., 2001), which are biosynthetic precursors of flavonoids, thiosemicarbazones (Chiyanzu et al., 2003) and isoquinolines (Sajid and McKerrow, 2002; Batra et al., 2003). However, this work shows that indirect drug design protocol can be successfully expedited for the discovery of small molecule cysteine protease inhibitors. Overall, the discussed efforts toward the design of inhibitors of the malarial cysteine proteases clearly demonstrate the effectiveness of combining computational techniques with organic synthesis in delivering novel potential inhibitors and in providing directions for further improvements.

4 Conclusion

In summation, a series of 2-(4-(substituted benzoyl)-1,4-diazepan-1-yl)-N-phenylacetamide with twenty novel derivatives, using ligand based approach were synthesized and screened for falcipain-2 inhibition. The structures of the compounds were assigned on the basis of ¹H NMR and mass spectral data. The preliminary SARs indicate that, overall, the compounds having 3-F (5b), 3-CF₃ (5g), 4-CF₃ (5h), 2-OCH₃ (5j) and 3-OCH₃ (5k) substituent in the aromatic ring displayed marked inhibition. Compound 5g is the most potent compound from this series, with lesser toxicity predicted by softwares (admetSAR and QikProp module of Schrodinger), and it can be utilized as a potential lead compound in the designing of new candidates to optimize the inhibitory potencies of this class of compounds, with potent antimalarial activity. Docking result provided the fundamental structural information to maintain the inhibitory activity, and was helpful for future inhibitor design. Moreover, to get the conclusive results, plasmodia study has to be done which will be the futuristic extension of the work. Thus, the present approach could be an excellent starting point for the development of novel cysteine protease inhibitors, and in general, for the development of new drugs for malaria.