Novel Substituted Heteroaromatic Piperazine and Piperidine Derivatives as Inhibitors of Human Enterovirus 71 and Coxsackievirus A16

A series of substituted heteroaromatic piperazine and piperidine derivatives were found through virtual screening based on the structure of human enterovirus 71 capsid protein VP1. The preliminary biological evaluation revealed that compounds 8e and 9e have potent activity against EV71 and Coxsackievirus A16 with low cytotoxicity.


Introduction
Enterovirus 71 (EV71) and Coxsackievirus A16 (CVA16) belong to the genus Enterovirus of family Picornaviridae [1]. Both viruses are common causes of hand, foot, and mouth disease (HFMD), which mostly affects young children [2]. The clinical manifestations associated with HFMD caused by both viruses include fever, sore throat, diarrhea, and papulovesicular rash on the hands, feet, and oropharyngeal mucosa. Moreover, EV71 infections can damage the central nervous system, leading to viral meningitis, encephalitis, or severe myocarditis with or without fatal pulmonary edema [3][4][5]. Many epidemic outbreaks have been reported in Southeast and East Asia because of the high of morbidity and mortality rates associated with such infections. Between January 2009 and May 2011, the EV71 epidemic in China has led to 1,000 deaths [6]. Although few inhibitors and vaccines have been reported, such as compound 1, shown in Supplementary Figure S1 [7][8][9][10], no antiviral agent has been approved by the US Food and Drug Administration for treating HFMD caused by enteroviral infections. Clinical treatments are directed only toward relieving the most prominent symptoms of each clinical syndrome. Therefore, we need to develop novel EV71 and CVA16 inhibitors.
In recent years, there are some organic molecules were identified for HFMD therapy [11], and EV71-related anti-virus activity had been identified by screening traditional Chinese medicines [12,13]. EV71 and CVA16, which are non-enveloped viruses, comprise a positive single-stranded RNA genome packed within an outer capsid composed of four proteins (VP1-VP4) [1]. Mature EV71, like other enteroviruses, has a hydrophobic pocket on VP1 that penetrates from the surface deep into the interior of the VP1 β-barrel, which underlies a canyon-like surface depression [14] and harbors a natural lipid (possibly sphingosine) called the "pocket factor" [15]. Releasing the pocket factor contributes to activation because it is required for initiating the uncoating and liberation of the RNA genome; thus, potential antivirals simulate natural lipids and inhibit uncoating [16], which provides a theoretical basis for virtual screening.
Our laboratory built an in-house chemical database of nearly 10,000 compounds. Based on the crystal structure of the EV71 capsid protein [Protein Data Bank (PDB) ID: 3VBH], we screened our database using DOCK 6.0 and found novel N-containing heterocyclic piperazines and piperidines as EV71 inhibitors and CVA16 inhibitors. Our findings provide a basis for anti-EV71 and anti-CVA16 drug development.

Computer Screening
Identifying a suitable site on the VP1 protein for binding small organic ligands is critical for successfully implementing the computer screening strategy for developing effective anti-EV71 inhibitors. The binding site is a hydrophobic pocket that penetrates from the surface deep into the interior of the VP1 β-barrel and harbors a natural lipid (possibly sphingosine) called the "pocket factor" by Wang [16]. The closure of the VP1 pocket after the pocket factor was expelled initiated EV71 uncoating with pocket collapse; which switches mature EV71 into expanded particles and transforms into a conformation with a high-receptor affinity. Therefore; the release of the pocket factor is required for initiating viral uncoating; and the discovery of potential anti-EV71 inhibitors focuses on compounds that are structurally similar to natural lipids to occupy the hydrophobic pocket; the binding site. Some amino acids are important in EV71 VP1 and the natural lipid complex crystal structure (PDB ID: 3VBH). The hydroxyl group of natural lipids was 2.88 Å away from the N atom of Ile113. It indicates the formation of H-bond. Morever the natural lipid and residues Ile111; Asp112; Thr114; Val192; Met230; and Phe233 interact through molecular electrostatic potential; and Van der Waals force.
The identification of a VP1 capsid-binding pocket facilitated the computational screening of 10,000 organic compounds using our in-house chemical database for potential ligands for this pocket. This screening selected 60 compounds for cell culture cytopathic effect (CPE) assays. Most of the 60 compounds were known as HRV inhibitors before. It is not surprising since the homology of EV71 VP1 coat protein (Sequence ID: Q91PB1) with human rhinovirus (HRV) 2 coat protein (PDB ID: 1FPN) and HRV16 coat protein (PDB ID: 1AYM) from UniProtKB were up to 40% and 38%, respectively. Among the 60 cmpounds identified, 8e is novel and 9e is known to inhibit HRV and their docking models in the hydrophobic pocket are shown in Figure 1 as compared to compound 1 developed by Shih et al. [10] shown in Supplementary Figures S2-S3. Ile111, Val192, Met230, and Phe233 formed hydrophobic interactions with different moieties of compounds 1, 8e, and 9e. The carbonyl group of imidazolidinone in compound 1 was 3.49 Å away from the N atom of Ile113, indicating formation of H-bond. It was also observed that the formation of π-π effects between the compound 1 and residues Phe155 and Trp203 existed. Besides, Phe137, Phe155 and 8e interacted through π-π effect, whereas Tyr201 and Asn228 performed σ-π effect with 9e. It seems that those compounds have more potential for EV71 VP1 inhibition. Fifteen of these compounds displayed significant inhibitory effect (median inhibitory concentration [IC 50 ]: 1 μM to 100 μM), with the structure of the 15 organic inhibitors shown in Table 1. This finding also indicates further structural optimization directions for our research.

Chemistry
Compounds 9a-h (Table 1) were previously reported as HRV inhibitors [17,18]. Because compounds 8a-f have never been reported, we describe their synthesis in this paper (Scheme 1). To prepare piperazine-1-yl-alkyl alcohol 4a-b, ethyl piperazine-1-carboxylate 2 was coupled with compounds 3a-b in the presence of K 2 CO 3 in acetonitrile and exposed the ethoxycarbonyl group in the presence of aq NaOH in ethanol. The 2-bromo group of 2-bromothiazole (5) was substituted with compounds 4a-b in the presence of Na 2 CO 3 in dimethylformamide (DMF) to yield compounds 6a-b. Finally, compounds 8a-f were synthesized through the Mitsunobu reaction by coupling compounds 6a-b and 4-substituted phenols 7a-e.

Biology
The anti-EV71 activity, anti-CVA16 activity, and cytotoxicity of the compounds on the EV71 test strain (Shenzhen 98, shzh01-8) and CVA16 (laboratory viral culture) were evaluated by using cell culture cytopathic effect assays. The results are listed in Table 1.

Results for Anti-EV71 Activity
In the series of imidazolidinone compounds reported by Shia [10], the length of the alkyl linker does not contribute significantly to antiviral activity. However, we found such length to be important, as evidenced by our compounds 8b and 8c. The p-monosubstituted phenyl ring crucially affects antiviral activity [10]. Similarly, the steric requirement of 4-electron donating substitution may have affected the antiviral activity against EV71 via 8d-f. Moreover, the larger the alkoxy group, the higher the activity, such as with compounds 8a-b and 9g-h. Compared with that of its corresponding pyridine isomer 9f, the antiviral activity of the benzene ring of 9e was increased 25-fold. This effect might be due to the strict steric requirement on the phenyl ring or to the increased in basicity of the pyridine nitrogen. Although chloropyridine derivatives 9a-h have been known as potent human rhinovirus capsid-binding inhibitors [17,18], it hasn't been reported that the series of compounds manifests anti-EV71 activity. We first discovered compounds 9a-h showed a wide range of anti-EV71 activity. 9a-f were the derivatives of pirodavir that was known as the anti-HRV capsid-binding inbititor [19] (Supplementary Figure S1, IC 50 = 1.2 ± 0.2 μM, TC 50 = 31 ± 2.2 μM, SI = 25 [20]). However, compound 9e exhibited high antiviral potency and an excellent selectivity index (IC 50 = 1.0 μM, TC 50 > 512.97 μM, SI = 512.8).

Results for anti-CVA16 Activity
The evaluation of the series of compounds 8b-c revealed that the length of the alkyl linker was not sensitive to anti-CVA16 activity. 9b exhibited moderate anti-EV71 activity, but it was inactive at the highest concentration (29.74 μM, TC 0 = 29.74 μM). The compounds with the alkoxy, ester, or alkyl group at position 4 of the phenoxyl ring (compounds 8a-f and 9g-h) revealed that the suitable steric volume, neither too large nor too small, may be significant for anti-CVA16. Compound 9e was similar to its corresponding pyridine isomer 9f in terms of anti-coxsackie activity. These effects might be due to distinct differences in the hydrophobicity of the compounds. It was the first time that their anti-CVA16 activity was reported.

General
Melting points (mp; °C, uncorrected) were determined in open glass capillaries using a YRT-3 electrothermal melting point apparatus. The 1 H-NMR and 13 C-NMR were recorded on a Bruker ARX 400 MHz spectrometer (Karlsruhe, Germany). Chemical shifts are expressed in δ (ppm) with reference to tetramethylsilane. The mass spectra were processed through a Waters Xevo G2 QTof electrospray ionization (ESI) spectrometer (Denver, CO, USA). All solvents and reagents were purchased commercially and used without further purification.

Computer Screening
The high-resolution X-ray structure of EV71 VP1 with the capsid-binding pocket was used as a receptor for ligand docking. DOCK 6.0 is an automatic computerized method used to screen small molecule databases for ligands that bind to a given receptor [21]. DOCK 6.0 defined the EV71 VP1 binding site with a set of overlapping spheres, the centers of which became the potential locations for ligand atoms. The binding of an organic ligand to the VP1 hydrophobic pocket was evaluated based on shape complementarity and simplified interaction energy (force field energy). Our in-house database was chosen as the small-molecule database because it included approximately 10,000 available small organic compounds. The molecular structures were generated using the CONCORD heuristic algorithm (developed by R. Pearlman at the University of Texas). The parameters used for docking were set to the defaults. For example, the maximum orientation was 1500, and the bump_max was 3. Following four rounds of inflexible and flexible virtual screens, 50 molecules with the best shape complementarity scores and 50 with the best force field scores were selected from the DOCK 6.0 screening. The resulting 60 compounds were then chosen for the cell culture CPE assay. Of these 60 compounds, 37 were from the shape list, 23 were from the force field list, and 9 were on both lists. 4a,b, and 6a,b

Biological Evaluation
3.4.1. Neutralization Test [22] This assay measured the ability of the test compounds to inhibit the CPE induced by EV71 and CVA16 on Vero cells. The 96-well tissue culture plates were seeded with 20,000 Vero cells/mL. The plates were incubated for 24 h at 37 °C. The virus (100 TCID 50 ) mixed with different concentrations of the test compounds was added to the cells, which were incubated at 37 °C for 2 h. After adsorption, the infected cells were overlaid with 50 μL of Dulbecco's modified Eagle's medium with 5% fetal bovine serum, and 2% DMSO. The concentration of the test compound required to reduce the virus-induced CPE to 4+ relative to the virus control was expressed as IC 50 according to the Reed-Muench method. All assays were performed in triplicate.

Cytotoxicity Assay [23]
The Vero cells treated with the test compounds were incubated at 37 °C for 96 h. After incubation, the cells were harvested and counted. The concentration of a test compound required to reduce cell viability to 50% of the tested control culture was expressed as TC 50 .

Conclusions
A computer-based database screening strategy was employed to identify a novel series of aromatic heterocyclic substituted piperazine and piperidine derivatives as ligands for the EV71 VP1 hydrophobic pocket that exhibits antiviral activity against EV71 and CVA16. Further mechanistic studies on this new class of anti-EV71 inhibitors are in progress. Our study on the structure-activity relationships of anti-EV71 and anti-CVA16 indicated that the space volume of the 4-electron donating group substituent at the phenoxyl ring largely influenced the in vitro anti-EV71 activity of the new class of potent inhibitors, and the appropriate steric requirement on the p-monosubstituted phenyl ring may exhibit activity against CVA16. Interestingly, the anti-EV71 activity of this series of compounds was reduced when the phenoxyl ring was substituted for the corresponding pridine isomer. Moreover, the length of the alkyl linker influenced the anti-EV71 activity. However, both factors did not affect the anti-CVA16 activity. The piperidine derivative 9e exhibited the most potent antiviral activity against EV71 (IC 50 = 1 μM), with no apparent cytotoxicity against Vero cells (TC 50 > 512.97 μM), compared with compound 1(IC 50 = 1 μM, TC 50 = 77.91 μM). Compound 8e exhibited excellent activity against CVA16 (IC 50 = 1.24 μM) and moderate activity against EV71 (IC 50 = 4.28 μM), with weak cytotoxicity on Vero cell lines (TC 50 = 43.08 μM). Therefore, these compounds are very promising candidates for further optimization towards a drug for HFMD treatment.