Design and Synthesis of Novel Bis-Imidazolyl Phenyl Butadiyne Derivatives as HCV NS5A Inhibitors

In today’s global plan to completely eradicate hepatitis C virus (HCV), the essential list of medications used for HCV treatment are direct-acting antivirals (DAAs), as interferon-sparing regimens have become the standard-of-care (SOC) treatment. HCV nonstructural protein 5A (NS5A) inhibitors are a very common component of these regimens. Food and Drug Administration (FDA)-approved NS5A inhibitors, although very potent, do not have the same potency against all eight genotypes of HCV. Therefore, this study aims to synthesize NS5A inhibitor analogues with high potency pan-genotypic activity and high metabolic stability. Starting from an NS5A inhibitor scaffold previously identified by our research group, we made several modifications. Two series of compounds were created to test the effect of changing the length and spatial conformation (para-para vs. meta-meta-positioned bis-imidazole-proline-carbamate), replacing amide groups in the linker with imidazole groups, as well as different end-cap compositions and sizes. The frontrunner inhibits genotype 1b (Con1) replicon, with an EC50 value in the picomolar range, and showed high genotypic coverage with nanomolar range EC50 values against four more genotypes. This together with its high metabolic stability (t½ > 120 min) makes it a potential preclinical candidate.


Introduction
Hepatitis C infection remains a public health problem that is widespread throughout the globe. Despite the scientific advances witnessed during the past decade in both the diagnosis and treatment of hepatitis C virus (HCV) [1,2], more efforts are needed for the complete eradication of HCV, given that there is no availability of prevention measures, such as vaccines [3]. HCV is blood-borne and mainly transmitted through contaminated needle sharing and unscreened blood transfusions [4]. It belongs to the Hepacivirus genus of

Design Concept
All reported NS5A inhibitors have a long-established design of highly symmetrical pharmacophore structures, commonly bearing bis-pyrrolidine caps along with other structural features, such as linker length, as well as substituents attached to the caps that affect target activity and spectrum [34].

Design Concept
All reported NS5A inhibitors have a long-established design of highly symmetrical pharmacophore structures, commonly bearing bis-pyrrolidine caps along with other structural features, such as linker length, as well as substituents attached to the caps that affect target activity and spectrum [34].
Pharmaceuticals 2022, 15, 632 4 of 32 even enabled the preparation of injectable aqueous-soluble salts because of the imidazole nitrogen basicity [29]. In compound series (1a-10a), Figure 1, the imidazole-proline-carbamate caps were connected to the core at meta-meta positions, while in compound series (1b-10b) they were connected to the core at para-para positions, in an attempt to reach the ideal length and the most promising spatial orientation. Additionally, we planned to investigate the effect of cap chemical composition and stereochemistry on the potency of synthesized compounds, as it affects the compound's conformation in space, as well as its interactions with the protein. We investigated the effect of the most commonly used amino acid carbamates caps, such as L/D-valine, L/D-leucine, and L/D-phenyl glycine, on the potency of the compounds, while the terminal alkyl carbamate moiety varied between methyl, ethyl, and butyl.

Synthesis
The synthesis of the desired compounds was carried out in seven steps, as summarized in Scheme 2. First, the appropriate iodo-acetophenone was brominated using N-bromosuccinimide (NBS) under acidic conditions in the presence of p-TsOH, where the bromo substituent was installed in the alpha position to the carbon atom then replaced by an N-Boc-L-proline residue using triethylamine (TEA) in acetonitrile. The produced ketoester (B1-2) was used for the formation of the imidazole ring through the reaction with ammonium acetate in refluxing toluene. After that, the deprotection of Boc takes place by trifluoracetic acid (TFA) to give the free amine moiety (D1-2). On the free amine, coupling with amino acid carbamates (1c-10c) took place using HBTU as a coupling agent to get the monomers of the desired dimeric final compounds. Monomeric compounds then underwent Sonogashira cross-coupling reaction with trimethylsilyl acetylene (TMSA) to replace the iodo group, where triphenylphosphine palladium (II) chloride ([PdCl2(PPh3)2]) was used as a catalyst and copper (I) iodide (CuI) as a co-catalyst in the presence of TEA as a base, all together in DMF as solvent under inert conditions at 70 °C. Finally, the desilylation along with dimerization step took using K2CO3 in methanol and distilled water in the presence of a few specks of CuI to afford the final dimeric compounds (Table 1). In compound series (1a-10a), Figure 1, the imidazole-proline-carbamate caps were connected to the core at meta-meta positions, while in compound series (1b-10b) they were connected to the core at para-para positions, in an attempt to reach the ideal length and the most promising spatial orientation. Additionally, we planned to investigate the effect of cap chemical composition and stereochemistry on the potency of synthesized compounds, as it affects the compound's conformation in space, as well as its interactions with the protein.
We investigated the effect of the most commonly used amino acid carbamates caps, such as L/D-valine, L/D-leucine, and L/D-phenyl glycine, on the potency of the compounds, while the terminal alkyl carbamate moiety varied between methyl, ethyl, and butyl.

Synthesis
The synthesis of the desired compounds was carried out in seven steps, as summarized in Scheme 2. First, the appropriate iodo-acetophenone was brominated using N-bromosuccinimide (NBS) under acidic conditions in the presence of p-TsOH, where the bromo substituent was installed in the alpha position to the carbon atom then replaced by an N-Boc-L-proline residue using triethylamine (TEA) in acetonitrile. The produced ketoester (B 1-2 ) was used for the formation of the imidazole ring through the reaction with ammonium acetate in refluxing toluene. After that, the deprotection of Boc takes place by trifluoracetic acid (TFA) to give the free amine moiety (D 1-2 ). On the free amine, coupling with amino acid carbamates (1c-10c) took place using HBTU as a coupling agent to get the monomers of the desired dimeric final compounds. Monomeric compounds then underwent Sonogashira cross-coupling reaction with trimethylsilyl acetylene (TMSA) to replace the iodo group, where triphenylphosphine palladium (II) chloride ([PdCl 2 (PPh 3 ) 2 ]) was used as a catalyst and copper (I) iodide (CuI) as a co-catalyst in the presence of TEA as a base, all together in DMF as solvent under inert conditions at 70 • C. Finally, the desilylation along with dimerization step took using K 2 CO 3 in methanol and distilled water in the presence of a few specks of CuI to afford the final dimeric compounds ( Table 1). 5

Attachment to Core
As for the preparation of the required intermediate carbamates (1c-10c, summarized in Table 2), amino acids and chloroformates were treated under Schotten-Baumann conditions in the presence of NaOH and dioxane as shown in Scheme 3. Table 2. Cap-specific stereochemistry and substitution.
Pharmaceuticals 2022, 15, 632 6 of 32 As for the preparation of the required intermediate carbamates (1c-10c, summarized in Table 2), amino acids and chloroformates were treated under Schotten-Baumann conditions in the presence of NaOH and dioxane as shown in Scheme 3.

Biological Activity
All synthesized compounds were tested against HCV genotype 1b Con-1 subgenomic replicon in the Huh-5-2 stable cell line. Specifically, their activity against viral replication was estimated by measuring firefly luciferase, which is co-expressed by the bicistronic replicon, and their half-maximal effective concentrations (EC 50 ) were determined. In parallel, the half-maximal cytotoxic concentrations (CC 50 ) of the compounds were estimated, and their selective indexes (SI 50 : CC 50 / EC 50 ) were calculated. All novel synthesized compounds were active against HCV 1b with EC 50 < 1 µM except for compound 4b (Table 3). Table 3. Activity, cytotoxicity, and selectivity of compound series 1a-10a and 1b-10b against genotype 1b (Con1) replicon a .

Molecular Modeling
In addition to the SAR data extracted herein, we tried rationalizing the activity of all our compounds using docking experiments performed over HCV NS5A Gt1b (Con1) protein (PDB entry 3FQM) [37]. As described previously by our group [27] and others [38,39], a possible region of interaction for DAAs is formed between the dimeric protein chain interface at the N-terminus where zinc metal ions also reside ( Figure 2). Here, we focus on analyzing the results of the most active compound (i.e., 10a) vs. daclatasvir. Poses and binding modes for other compounds are available in the supporting information.

Structure-Activity Relationship
Evaluating the structure-activity relationship (SAR) of the synthesized compounds showed that the structure and stereochemistry of the capping groups have a major influence on the activity (EC 50 ).
Regarding the amino acid carbamate derivatives, both aliphatic and aromatic amino acids were used. The aliphatic amino acids were valine and leucine. All leucine derivatives were more potent than their corresponding valine derivatives, irrespective of the stereochemistry of the amino acids. This can be seen in both compound series, the m, m connected caps (1a-10a) and the p, p connected caps (1b-10b), indicating that the extra methylene spacer in leucine increased the lipophilicity/size of the terminal side chain and enhancing the activity. This is exemplified by compounds 5a-8a to compounds 1a-4a, respectively, and compounds 5b-8b to compounds 1b-4b, respectively.
Comparing the effect of using aromatic amino acids, e.g., phenylglycine, to that of aliphatic amino acids, phenylglycine derivatives showed higher activity than both valine and leucine derivatives in series (1a-10a). Therefore, the order of activity for amino acids in capping groups for this series is phenylglycine > leucine > valine. However, in series (1b-10b), the aliphatic caps showed activity better than the aromatic ones, where leucine derivatives showed the highest activity. This is surprising, as phenylglycine derivatives were reported to have EC 50 values lower than the valine derivatives when adopted with other cores [35,36].
As for the stereochemistry of the terminal amino acids used, in series (1a-10a), Denantiomers showed higher biological activity than L-enantiomers. This can be seen when comparing the results of D-enantiomers 6a (EC 50 = 96.6 nM), 8a (EC 50 = 85.4 nM), and 10a (EC 50 = 0.1 nM), to their corresponding L-enantiomers 5a (EC 50 = 113.7 nM), 7a (EC 50 = 97.8 nM), and 9a (EC 50 = 62.0 nM), respectively (Table 3). Compound 10a in particular, the D-phenylglycine derivative, is by far the most potent of all tested compounds with an EC 50 value of 0.1 nM, as well as being the safest with the highest SI 50 > 9990. The only exception to the stereochemistry pattern is the methyl-valine analogues, where the L-valine methyl derivative 1a is two times better in activity than the D-valine methyl derivative 2a. In series (1b-10b), the L-enantiomers were found to be more active than the corresponding D-enantiomers, except for the D-phenylglycine methyl 10b, which showed better activity than the L analogue.
Concerning the terminal carbamate O-substituents of series (1a-10a), the L-valine with ethyl substituent 3a, (EC 50 = 158.8 nM) had almost equal activity to the methyl substituent 1a (EC 50 = 166.4 nM), while the butyl substituent 4a showed much lower activity than both (EC 50 = 901.8 nM). In the case of Land D-leucine capping, the ethyl substituent again showed almost equivalent activity to that of methyl. Therefore, the order of activity for O-substituents is methyl ≈ ethyl > butyl.
For series (1b-10b), the terminal leucine caps showed an increase in activity with a decrease in the bulkiness of their terminal O-substituents. The L-leucine methyl 5b showed activity higher than L-leucine ethyl 7b by more than 3-fold, while D-leucine methyl 6b showed almost double the potency of D-leucine ethyl 8b. Regarding valine caps, the activity increased more clearly as the bulkiness of the terminal O-substituent decreased. The L-valine methyl 1b showed higher activity than the L-valine ethyl 3b and L-valine butyl 4b. Therefore, the order of activity for O-substituents is methyl > ethyl > butyl.
Overall, compound 10a showed the highest activity with an EC 50 value of 0.1 nM among m, m'-compound series (1a-10a), while compound 5b showed the highest activity with an EC 50 value of 16.1 nM among the p, p-compound series (1b-10b).
When comparing series (1a-10a) with m, m connected caps to series (1b-10b) with p, p -connected caps, the m, m -connection exhibited higher activity when combined with the D-enantiomer terminal caps, while the p, p connected compounds favored the Lenantiomer terminal caps. The m, m compounds were more potent with the aromatic caps, while the p, p showed higher potency with the bulky aliphatic terminal groups. As for the carbamate O-substituted terminal group, the activity for both m, m -and p, p -compounds was enhanced when combined with shorter substituents, such as the methyl group.
As per the previous observations, we cannot conclude that there are common patterns that can tell which series shows superior activity over the other; however, the most potent among all compounds was the m, m derivative with the D-enantiomer aromatic cap and the methyl terminal group (compound 10a).

Molecular Modeling
In addition to the SAR data extracted herein, we tried rationalizing the activity of all our compounds using docking experiments performed over HCV NS5A Gt1b (Con1) protein (PDB entry 3FQM) [37]. As described previously by our group [27] and others [38,39], a possible region of interaction for DAAs is formed between the dimeric protein chain interface at the N-terminus where zinc metal ions also reside ( Figure 2). Here, we focus on analyzing the results of the most active compound (i.e., 10a) vs. daclatasvir. Poses and binding modes for other compounds are available in the supporting information. Compound 10a exhibits a therapeutic potential similar to the known drug daclatasvir, both having a picomolar EC50 value. Its chemical structure incorporates a core connected to m, m' cap substitution that showed a rich hydrogen bond (HB) network and several hydrophobic and/or π-π interactions ( Figure 3) with docking scores of −4.04 Chemgauss4 in the OpenEye software and −8.80 kcal/mol in the PyRx software, respectively. The binding mode witnessed in both software programs is balancing between the two Tyr93 A/B residues in either subunit. The imidazole ring forms H-bonds (HBs), and the phenyl ring of the caps is engaged in hydrophobic interactions. On the other hand, daclatasvir introduces a p, p' cap substitution with similar binding modes and interacting residues ( Figure 4). The respective docking score of −5.89 Chemgauss4 in the OpenEye software and −8.70 kcal/mol in the PyRx software. Although the p, p' substitution is not favored in compound 10b, this is expected if we consider the difference in core length between 10a and daclatasvir. The latter shows a distance of 13.8 Å from sp 2 N-imidazole to sp 2 N'-imidazole ring and 14.3 Å for 10a, accordingly. Compound 10a exhibits a therapeutic potential similar to the known drug daclatasvir, both having a picomolar EC 50 value. Its chemical structure incorporates a core connected to m, m' cap substitution that showed a rich hydrogen bond (HB) network and several hydrophobic and/or π-π interactions ( Figure 3) with docking scores of −4.04 Chemgauss4 in the OpenEye software and −8.80 kcal/mol in the PyRx software, respectively. The binding mode witnessed in both software programs is balancing between the two Tyr93 A/B residues in either subunit. The imidazole ring forms H-bonds (HBs), and the phenyl ring of the caps is engaged in hydrophobic interactions. On the other hand, daclatasvir introduces a p, p' cap substitution with similar binding modes and interacting residues ( Figure 4). The respective docking score of −5.89 Chemgauss4 in the OpenEye software and −8.70 kcal/mol in the PyRx software. Although the p, p' substitution is not favored in compound 10b, this is expected if we consider the difference in core length between 10a and daclatasvir. The latter shows a distance of 13.8 Å from sp 2 N-imidazole to sp 2 N'-imidazole ring and 14.3 Å for 10a, accordingly.
Importantly, as mentioned in the present study and our previous results [27] highly active NS5A inhibitors tend to occupy the centered interface ( Figure 5) between the two Tyr93 A/B residues, while solutions traversing down the rifts in either side "due to the protein's antiparallel symmetry" match with less active compounds (see Supplementary Materials ( Figures S1-S21)). Moreover, based on the docking score comparison (see Supplementary Materials Table S1) between two different software programs used in our study, the models represent an almost 90% probability that most active compounds will likely be included in top spots. Thus, models showcase a desired compatibility between in vitro and in silico results, providing further information on possible NS5A inhibitors mode of activity concerning the previously specified region [27,39] of NS5A genotype 1b. Concluding to somewhat higher confidence for future rational drug design efforts.

Pharmacochemical Evaluation and Drug-Likeness
Moreover, with the use of the FAF4 online server [40], pharmacochemical profiling (see Supplementary Materials Table S2a,b) was performed for all new compounds, including the reference compound (daclatasvir). In an attempt to extract valuable information from the data, we see that the number of tolerated rotatable bonds (RB) in DAAs shows a range between 13-18 since greater than 18 witnessed in the case of compounds 4a and 4b (i.e., RB = 20) with either m, m'or p, p'-substitution was included in the least active compounds. For the sum of the compounds, we did not find any correlation regarding values on tPSA or HBD/HBA count since all are identical with each other. Focusing on daclatasvir and 10a, which both exhibited a picomolar antiviral activity on NS5A Gt1b, we see an acceptable range in values on both ionic or not partition coefficients (i.e., logD = 4.16 vs. 6.23/logP = 5.09 vs. 6.01) and flatness (i.e., Fsp 3 = 0.45 vs. 0.27) for daclatasvir vs. 10a, respectively.  Importantly, as mentioned in the present study and our previous results [27] highly active NS5A inhibitors tend to occupy the centered interface ( Figure 5) between the two Tyr93 A/B residues, while solutions traversing down the rifts in either side "due to the protein's antiparallel symmetry" match with less active compounds (see Supplementary Materials (Figures S1-S21). Moreover, based on the docking score comparison (see Supplementary Materials Table S1) between two different software programs used in our study, the models represent an almost 90% probability that most active compounds will likely be included in top spots. Thus, models showcase a desired compatibility between in vitro and in silico results, providing further information on possible NS5A inhibitors mode of activity concerning the previously specified region [27,39] of NS5A genotype 1b. Concluding to somewhat higher confidence for future rational drug design efforts.  Table S2a,b) was performed for all new compounds, including the reference compound (daclatasvir). In an attempt to extract valuable information from the data, we see that the number of tolerated rotatable bonds (RB) in DAAs shows a

Validation of Compound Activity with Additional Assays
The potency of compound 10a measured by the luciferase assay in HCV GT 1b was confirmed at the level of viral RNA and NS5A protein expression. Huh5-2 replicon cells were treated with 10a in serial dilutions, or its solvent DMSO (control), and viral RNA was quantified by reverse transcription-quantitative polymerase chain reaction (RT-qPCR). NS5A protein levels were evaluated by indirect immunofluorescence combined with confocal microscopy analysis. HCV RNA replication was reduced in the presence of the compound with EC 50 = 0.25 nM ( Figure 6A), which is close to the one determined in the luciferase assay (EC 50 = 0.10 nM). A consistent reduction of NS5A expression was detected, as shown in Figure 6B, in a dose-dependent manner. Nuclei were stained with propidium iodide (PI) as a cell viability control.

Evaluation of Metabolic Stability
Phase I and phase II metabolic stability were measured for our most potent compound 10a to further evaluate its applicability to go into in vivo studies. This was done using human liver S9 fractions, following the previously reported protocol [41,42]. A number of samples were taken at defined time points, and using LC-MS/MS, the remaining percentage of the parent compound was determined. The calculated half-life time was more than 120 min, indicating high metabolic stability.

Molecular Modeling
A library of the newly synthesized compounds 1a-10a, 1b-10b and daclatasvir serving as control substances were compiled in both a smile formatted file (*.smi) and a Sybyl MOL2 file (*.mol2) using the free program Open Babel v3.1.1 [43]. Following this step, the SMILES library was transformed to SDF format, and their conformers were generated using

Ligand and Protein Preparation
The AB dimer crystal structure of HCV NS5A protein genotype 1b (PDB entry 3fqm [37]) was downloaded from the protein databank [51]. The PDB file was cleaned of water molecules and prepared with the OEDocking Suite program MAKE RECEPTOR v4.0.0.0 (OpenEye Scientific Software, Inc.) [52,53] to give the respective input oed extension file to be used in simulation experiments. The search space was centered at the zinc regions of the dimer in each case equally for subunits A and B. This generated an initial box of~70,000 Å 3 , which after a balanced site-shape creation resulted in an outer docking space of approximately~27,000 Å 3 for the protein. Neither residue modifications nor any constraints were implemented on the protein in the docking. The compiled SDF library contained all of the compounds 1a-10a, 1b-10b and daclatasvir. Conformer generation took place with the use of Omega v4.1.0.0 software (OpenEye Scientific Software, Inc.) [44,45] by setting a threshold of 600 structures with the flipping option turned off since all molecules possess specific stereochemistry, and the docking was performed with the OEDocking suite program FRED v4.0.0.0 (OpenEye Scientific Software, Inc.) [46,53]. The model calculations performed are produced by an Exhaustive Search Algorithm. Refinement of results was additionally performed to sort poses with standard options by the OEDocking suite program Scorepose v4.0.0.0 (OpenEye Scientific Software, Inc.) [46,47].

Molecular Docking Replication Experiments
The corresponding compiled Sybyl MOL2 library was subjected to MM2 energy minimization [54] with PerkinElmer Chem3D (Waltham, MA, USA) before in silico reproduction using the virtual screening tool PyRx v0.8 (The Scripps Research Institute) [49] and specifically AutoDock Vina [55,56], which incorporates an alternate algorithm to the calculations (i.e., Genetic Algorithm). Similar to the previously documented conditions mentioned herein, docking was performed in the same region of the protein Hepatitis C virus NS5A genotype 1b. Centered at the zinc regions (i.e., axis X: 23.6268, Y: −9.8409, Z: 9.0372) of the dimer's subunits A and B. This generated an initial box of >60,000 Å 3 (i.e., dimensions X: 41.3353, Y: 33.7403, Z: 45.1190). Protein was treated as rigid and the sole parameter adjustment was on "exhaustiveness", which value was set to 20 instead of the default 8.

Pharmacochemical Profiling
Compounds were compiled on a smile formatted file (*.smi) following the generation of their 3D structure file (*.sdf) all with the use of the program Open Babel v3.1.1 [43]. The later SDF file served as the required input file for the pharmacochemical properties calculations performed over the FAF-Drugs4 server [40]. Selected descriptors were XLOGP3 [57] for the predicted lipophilicity and the server's built-in filter for drug-likeness [58][59][60][61], which takes into account common drug-design properties following Lipinski's rule of five and others providing predicted calculations for polar surface area (tPSA), flatness (Fsp 3 ) hydrogen bond acceptor/donor count, etc. (see Supplementary Materials Table S2a,b). Moreover, compound 3 was positively checked for the detection of undesirable moieties [62], as well as PAINS [63].

Chemistry
Solvents and reagents were obtained from commercial suppliers and were used without further purification. 1 H-NMR spectra were recorded at 400 MHz and 13 C-NMR spectra were run at 101 MHz in deuterated chloroform (CDCl 3 ). The chemical shifts are referenced in the residual protonated solvent signals. The purities of the tested compounds were determined by HPLC, coupled with mass spectrometry. Mass spectrometric analysis (UPLC-ESI-MS) was performed using a Waters ACQUITY Xevo TQD system, which consisted of an ACQUITY UHPLC H-Class system and Xevo™ TQD triple-quadrupole tandem mass spectrometer with an electrospray ionization (ESI) interface (Waters Corp., Milford, MA, USA). An Acquity BEH C18 100 mm × 2.1 mm column (particle size, 1.7 µm) was used to separate analytes (Waters, Dublin, Ireland). The solvent system consisted of water containing 0.1% formic acid (A) and 0.1% formic acid in acetonitrile (B). HPLC method: flow rate 200 µL/min. The percentage of B started at an initial of 5% and maintained for 1 min, then increased up to 100% during 10 min, kept at 100% for 2 min, and flushed back to 5% in 3 min, then kept at 5% for 1 min. The MS scan was carried out under the following conditions: capillary voltage 3.5 kV, cone voltage 20 V, radio frequency (RF) lens voltage 2.5 V, source temperature 150 • C, and desolvation gas temperature 500 • C. Nitrogen was used as the desolvation and cone gas at flow rates of 1000 and 20 L/h, respectively. System operation and data acquisition were controlled using MassLynx 4.1 software (Waters). HRMS was measured as previously described [64].

General Synthetic Methods and Experimental Details for All Compounds General Procedure for Carbamate Synthesis
In a 250 mL round bottom flask, 1 M NaOH (75 mL) was added and left to cool to 0 • C in an ice bath. After that, the respective amino acid (24 mmol) was added, and the solution was left to stir until it was homogeneous. Then, the respective chloroformate (33 mmol) and 1,4-dioxane (30 mL) were added drop by drop. The reaction mixture was then allowed to stir overnight at room temperature. The solution was extracted with Et 2 O (3 × 50 mL). The aqueous layer was cooled to 0 • C in an ice bath and concentrated HCl was added dropwise until pH = 2. The aqueous solution was extracted again with Et 2 O (3 × 100 mL). The organic layers were combined, dried over anhydrous Na 2 SO 4 , filtered, and concentrated in vacuo to give a viscous oily product. The compound was used for the next step without further purification.
(Methoxycarbonyl)-D-phenylglycine (10C). The compound was synthesized according to the procedure for carbamate synthesis using D-phenylglycine amino acid and methyl chloroformate to give a clear oily product: yield 2.6 g (54%); C 10 H 11 NO 4 .

General Procedure for Alpha Carbon Bromination
N-bromosuccinimide (6.25 mmol, 1.11 g) was added to the appropriate iodoacetophenone (5 mmol, 1.22 g) in acetonitrile (30 mL), the mixture was stirred at room temperature for 10-15 min, then p-toluene sulfonic acid (p-TsOH) (10 mmol, 1.90 g) was added to the mixture and refluxed for 2 h. The reaction mixture was concentrated under reduced pressure, washed with a saturated solution of Na 2 CO 3 and extracted with EtOAc. The organic layer was separated and dried over MgSO 4 , filtered then concentrated under reduced pressure.

2-Bromo-1-(3-iodophenyl) ethan-1-one (A 1 ).
Synthesized according to the method outlined above using 3 -iodoacetophenone; orange oil; yield: 1.5 g (92%); 1  General Procedure for Pyrrolidine Dicarboxylate Formation. Boc-L-proline (4.6 mmol, 1.5 g) was added to Compound A 1 or A 2 in Acetonitrile (15 mL) Following that, TEA (25 mmol, 3.50 mL) was added, and the reaction mixture was left to stir at room temperature for 3 h. Then, it was concentrated under vacuum, washed with distilled water, and extracted with EtOAc (50 mL × 3). The combined organic layers were separated and dried over MgSO 4 and filtered then concentrated under vacuum. The compound was confirmed by MS analysis and used in the next step without further purification.

General Procedure for Boc Deprotection
Trifluoroacetic acid (2 mL) was added to compound C 1 or C 2 in dichloromethane (15 mL) at room temperature, and the reaction was left for 3 h. The reaction mixture was concentrated under reduced pressure, neutralized with 5 M NaOH, and extracted with EtOAc. The organic layer was separated, dried over MgSO 4 , and filtered then concentrated under reduced pressure.
To the resulting compounds from the Sonogashira reaction step, distilled water (2 mL), methanol (10 mL), and potassium carbonate (4 mmol, 0.55 g) were added. The reaction mixture was stirred at 55 • C overnight. Then, it was concentrated under vacuum, washed with distilled water, and extracted with EtOAc (50 mL × 3). The combined organic layers were dried over MgSO 4 , filtered then concentrated under vacuum. The final product was later purified using silica gel in column chromatography.

In Vitro Transcriptio
H77S.3/GLuc2A was linearized with XbaI and used for in vitro transcription, as described previously [70].

Transfection with In Vitro Transcribed RNA
Electroporation with the full-length viral RNA H77S.3/GLuc2A into Huh7-Lunet cells was performed as described elsewhere [71]. In brief, 4 × 10 6 cells were detached by trypsin and resuspended in Cytomix [72] containing 2 mM ATP and 5 mM glutathione, mixed with 10 µg of viral RNA, and electroporated with a Gene Pulser system (Bio-Rad, Hercules, CA, USA).

Anti-HCV Assay
The anti-HCV assay in replicon cells was performed by seeding 1 × 10 4 cells per well in a 96-well plate. After 24 h of incubation at 37 • C (5% CO 2 ) and medium removal, serial dilutions of the test compounds in complete DMEM were added. Cells were lysed after 72 h, and F-Luc activity was measured. The anti-HCV assay for the full-length viral construct H77S.3/GLuc2A was performed after electroporation of Huh7-Lunet cells with viral RNA. Cells were seeded in 96-well plates, the medium was removed 24 h after electroporation, and serial dilutions of the compounds in complete DMEM were added. At 72 h posttreatment, cell supernatants were collected, and G-Luc activity was measured. In addition, the cells were lysed for total protein quantification.
Relative luminescence units (RLU) were expressed as a percentage of the respective units derived from DMSO-treated cells (control). The half-maximal effective concentration (EC 50 ), defined as the concentration of compound that reduces luciferase signal by 50%, was calculated by dose-response curve analysis using Prism 5.0 software (GraphPad Software Inc., San Diego, CA, USA).

Luciferase and Bradford Assays
The F-Luc activity was measured using the Luciferase Assay System (Promega, Madison, WI, USA), as recommended by the manufacturer. The G-Luc activity was measured using 12 µM coelenterazine (Promega) in assay buffer (50 mM potassium phosphate, pH 7.4, 500 mM NaCl, 1 mM EDTA). Measurements were carried out using a GloMax 20/20 single tube luminometer (Promega) for 10 s. The F-Luc activity was normalized to the total protein amount determined using the Bradford assay reagent (Pierce, Waltham, MA, USA).

Cytotoxicity Assay
Intracellular ATP levels were measured to determine the cytotoxicity of the compounds in the treated cells. In 96-well plates, 10 4 cells per well were seeded, incubated with serial dilutions of the compounds or DMSO as a control for 72 h at 37 • C (5% CO 2 ) and lysed for ATP measurement. ATP levels were measured using the ViaLight HS BioAssay kit (Lonza, Basel, Switzerland) in a GloMax 20/20 single-tube luminometer (Promega) for 1 s, according to the manufacturer's protocol. Values were normalized to total protein amounts and used to determine the compound concentration causing 50% cell death (CC 50 ). Nonlinear regression analysis was performed using Prism 5.0 software (GraphPad Software Inc., San Diego, CA, USA).

Indirect Immunofluorescence
An indirect immunofluorescence analysis of Con1 NS5A was performed as described previously [71]. Propidium iodide (Sigma-Aldrich, St. Louis, MO, USA) was used to stain the DNA. Images were obtained with a Leica TCS-SP8 Confocal Microscope (Wetzlar, Germany).

Total RNA Extraction and Quantification of Viral Replicons
Extraction of total RNA from Huh5-2 cells was performed using TRIzol reagent (Ambion, Austin, TX, USA), according to the manufacturer's instructions. Replicon RNA was quantified with reverse transcription (RT) and quantitative real-time polymerase chain reaction (qPCR). RT was performed using a Con1 IRES-specific primer (IRES-R:

Statistical Analysis
In all diagrams, bars represent the mean values of at least two independent experiments in triplicate. Error bars represent standard deviations. Only results subjected to statistical analysis using Student's t-test with p ≤ 0.05 was considered statistically significant and presented. Statistical calculations were performed using Excel Microsoft Office ® .

Conclusions
In the context of eradicating HCV infections, globally focused efforts are required with wider research approaches targeting not only drugs' potency but equally important the emerging resistance and the pan-genotypic spectrum of such compounds. Evolving our previously identified potent DAAs bearing the core scaffolds 3,3 -(buta-1,3-diyne-1,4diyl)dianiline and 4,4 -(buta-1,3-diyne-1,4-diyl)dianiline linked to different capping groups to act as NS5A inhibitors, we introduced cap modifications. Specifically, we replaced existing amide groups with imidazole rings, hoping to achieve better potency, but also increase genotypic coverage, as well as to attain good metabolic stability. Different capping groups were also linked to these modified cores to try various possible interactions inside the NS5A pocket. Among those, the m, m' compound 10a with the aromatic phenylglycine cap proved the most potent in the picomolar range (EC 50 = 0.10 nM), while the second most active compound 5b, p, p' with the aliphatic leucine cap was in the low nanomolar range (EC 50 = 16.12 nM) among this series. Additionally, combining in silico and pharmacochemical drug-like calculations, we started setting solid foundations on revealing all concealed aspects of the HCV NS5A activity and offering rationalized interpretation of data. Proving that spatial conformation of the molecule inside the NS5A pocket and molecule length affects the interactions that take place between caps and protein. Since our data indicate the core length being in close correlation with whether an mor psubstitution will be tolerated, higher length cores with p-,p'substitution was not favoured in our hands. Therefore, it is important to study the change in caps together with the molecule conformation to achieve the best combination that gives the highest potency. Additionally, 10a, the most potent among all, was tested against GT 1a, GT 2a, GT 3a, and GT 4a and showed an EC 50 in the nanomolar range against all of them, showing very good coverage of different genotypes. These outcomes and the fact that it has good metabolic stability make compound 10a a promising candidate for further study and testing.