Identification of 2-(4-N,N-Dimethylaminophenyl)-5-methyl-1-phenethyl-1H-benzimidazole targeting HIV-1 CA capsid protein and inhibiting HIV-1 replication in cellulo

The capsid (CA) subunit of the HIV-1 Gag polyprotein is involved in several steps of the viral cycle, from the assembly of new viral particles to the protection of the viral genome until it enters into the nucleus of newly infected cells. As such, it represents an interesting therapeutic target to tackle HIV infection. In this study, we screened hundreds of compounds with a low cost of synthesis for their ability to interfere with Gag assembly in vitro. Representatives of the most promising families of compounds were then tested for their ability to inhibit HIV-1 replication in cellulo. From these molecules, a hit compound from the benzimidazole family with high metabolic stability and low toxicity, 2-(4-N,N-dimethylaminophenyl)-5-methyl-1-phenethyl-1H-benzimidazole (696), appeared to block HIV-1 replication with an IC50 of 3 µM. Quantitative PCR experiments demonstrated that 696 does not block HIV-1 infection before the end of reverse transcription, and molecular docking confirmed that 696 is likely to bind at the interface between two monomers of CA and interfere with capsid oligomerization. Altogether, 696 represents a promising lead molecule for the development of a new series of HIV-1 inhibitors. Supplementary Information The online version contains supplementary material available at 10.1186/s40360-022-00581-7.

Page 2 of 13 Alvarez et al. BMC Pharmacology and Toxicology (2022) 23:43 for effective infection [4]. Moreover, CA is interacting with several cellular factors to play different roles during the early phase of the viral replication cycle, from uncoating to trafficking and entering the nucleus of the infected cells [5]. Among others, HIV-1 CA interacts with Cyclophilin A, Nucleoporins 153 and 358, and polyadenylation specificity factor 6 (CPSF6) [6][7][8][9][10]. It can also interact with restriction factor Trim5α to allow the evasion of HIV from immune innate mechanisms [11]. The fact that CA is important for both early and late phases of the infection makes it a target of choice to develop anti-HIV compounds [12]. Moreover, these pleiotropic roles impose a strong selection to maintain all these functions, resulting in low sequence diversity within a single retroviral species and a conserved typical α-helical structure between retroviruses [13][14][15]. It is composed of two domains: the N-terminal domain (NTD) is made up of 7 α-helices while the C-terminal domain (CTD) has 4 α-helices. FIV (feline immunodeficiency virus) and HIV-1 CA also possess an N-terminal β-hairpin which seems to be involved in the entry of nucleotides for the reverse transcription of the viral RNA genome into double-stranded DNA [13,16]. NTD and CTD of CA are separated by a flexible linker which allows the two domains to undergo relative spatial rearrangements during CA oligomerization. HIV-1 CA interactions with adjacent monomers to form the pentamers and hexamers involve both NTD-NTD, CTD-CTD, and NTD-CTD interactions, and CA-CA interactions are also needed to assemble the oligomers and form the fullerene-shaped mature viral core [4,[17][18][19]. Therefore, the emergence of viruses with escape mutations to anti-CA treatments could be hindered by these sequence and structural constraints, reinforcing the interest of CA as a therapeutic target for HIV infection. Peptides, peptidomimetics, or small molecules targeting CA have been developed which could inhibit HIV in vitro and are mostly binding three regions of CA [20]: hydrophobic pockets in the NTD, a groove in the CTD, or a pocket formed between the NTD of one monomer and the CTD of an adjacent monomer in a CA hexamer [21][22][23][24][25]. These various targeted regions are reflecting multiple mechanisms of inhibition, from blocking of CA-CA interactions during virus assembly to interference with post-entry events through the stabilization of the viral core or competition with cellular partners [20,22]. Although a recent paper has demonstrated the clinical potential of CA as a therapeutic target [26], no CA inhibitor is available yet on the market to tackle HIV infection. The low metabolic stability of some compounds [22] and high costs of production are among the main obstacles to their development for clinical use. Thus, the quest for new compounds against HIV-1 CA is still ongoing.
More than 400 compounds from our chemical library described elsewhere have been tested for their ability to interfere with CA assembly. Then from the data previously acquired from those 400 molecules tested in CA from FIV, we selected 10 compounds for their structural relationship with known CA HIV-1 inhibitors [27][28][29][30][31]. From those molecules, 5 compounds demonstrated interference with the assembly of HIV-1 CA in vitro One hit molecule, compound 696, which was recently shown by NMR to bind FIV capsid protein [27], demonstrated an inhibitory effect on HIV-1 replication in cellulo, with an IC 50 of 3 µM. Docking experiments with this hit compound demonstrated its binding to the NTD of HIV-1 CA, in the zone of interaction of CA with its cellular partner Nup153 which is also targeted by the structurally-related inhibitor PF74 [21]. A comparison of the binding site of the PF74 and the possible interaction region of 696 suggests a different inhibition mechanism between PF74 and our compound 696. Also, PF74 accelerated the oligomerization in vitro, and 696 inhibited these. Moreover, 696 shows high metabolic stability and low toxicity, which makes it an interesting lead molecule to optimize for the development of a new class of lowcost HIV CA inhibitors.

Screening of the chemical library for CA assembly inhibition in vitro
More than 400 compounds from our chemical library described elsewhere [28][29][30] have been tested for their ability to interfere with CA assembly. As the assembly inhibition assay requires high protein and molecule concentrations, with the concomitant solubility problems, a dose-response curve could not be determined for this assay as the 1:1 protein: compound ratio is unreachable. However, it is a common first-line screening strategy for anti-assembly molecules to select molecules to be tested in cellulo and/or in vivo [26]. Then from the data previously acquired from those 400 molecules tested in CA from FIV, we selected 10 compounds for their structural relationship with known CA HIV-1 inhibitors [30]. From those molecules, 5 compounds demonstrated interference with the assembly of HIV-1 CA in vitro (Fig. 1), namely compounds 305, 553, 878, 696, and 1310 (Fig. 2).
Compound 314 showed an inconsistent behavior because of its insolubility at concentrations higher than 10 µM. Compounds 314 and 305 are thiazolidene hydrazines designed as antiparasitic compounds targeting Trypanosoma cruzi (Fig. 2). Compounds 1310 and 878 are derivatives from the same rational design as compound 314, but these were not active against T. cruzi and have no recognized biological activity [32]. Notably, compounds 696 and 553 are structurally related to CA inhibitor PF74. However, 553 had high unspecific cytotoxicity in murine macrophages 100% at 50 µM in 24 h. Also, it has a nitro group which is generally associated with toxicity and interference with liver metabolism and was not considered for in cellulo experiments, also it was tested positive in the Ames test in silico (Fig. 2).

Toxicology profile of the active compounds
The cytotoxicity of the compounds was assayed in different types of cells (TZM-bl cells, U87-CD4-CCR5, murine macrophages (J774.1), and THP-1 monocytes) at 25, 50, and 100 µM (Fig. 2). All of the compounds demonstrated EC 50 of more than 50 µM, except for compound 553. Compound 314 has the most complete toxicology profile because was preclinically developed for Chagas disease infection, as an antiparasitic drug. This one also demonstrates good toxicology profiles, with no mutagenicity and no genotoxicity, also low oral toxicity in vivo. Compound 696 also demonstrates low unspecific toxicity. It has lower oral toxicity than PF74. Was predicted as no mutagenic by the TEST software.

Inhibitory activity of selected compounds against HIV-1 replication in cellulo
Based on the compounds which showed an inhibitory effect in our assembly assay ( Fig. 1), we selected molecules from our chemical library related to these lead compounds, favoring those with good solubility and low toxicity. This led us to consider 22 molecules (Table S2) in total that were tested in cellulo for their ability to inhibit HIV-1 replication. Among the 22 compounds, only two compounds showed a clear dose-response inhibition of HIV-1 replication, compounds 314 and 696. Compound 696 showed inhibition of HIV-1 replication in U87-CD4-CCR5 cells of more than 80% at the two highest concentrations tested after 11 days (Fig. 3a). As demonstrated using MTT assay, this effect was not due to the possible toxicity of compound 696 in the range of concentration we used (Fig. 3b). Thus, with an IC 50 of 2.3 ± 0.6 µM on day 11, 696 demonstrates a selective activity of at least 10 times between mammalian cells (EC 50 > 50 µM, Fig. 2) and antiviral activity (EC 50 /IC 50 > 21). Also compound 696 had low toxicity on murine macrophages EC 50 > 50 µM (at 48 hs), EC 50 inU87-CD4-CCR5 (11 days) and TZM-bl cells > 50 mM (48hs), EC50 in human macrophages > 100 mM (48hs), EC 50 in human red blood cells > 100 mM (24hs). Among the other compounds, 314 showed inhibition only at 10 µM, which was not sufficient for the calculation of an IC 50 . A trend towards inhibition of HIV-1 replication at low concentrations for compounds 878 and 1310 was observed. The results obtained with 878 and 1310 could not be confirmed at high concentrations because of solubility problems. In particular, compound 1310 has two ionizable groups, which may interfere with its permeability into the cells. It will be interesting to consider these two compounds for future chemical modifications, such as esterification with methanol of compound 1310 or a different solubilization approach with compound 878. The remaining compounds tested showed no We have recently demonstrated the binding and inhibitory effect of 696 on FIV capsid assembly in vitro [27]. Thus, we investigated the interaction of 696 with HIV-1 CA by microscale thermophoresis (MST). Interestingly, specific binding of compound 696 to monomeric HIV-1 CA was detectable with an estimated K D of 69 ± 6 µM ( Fig. 4). On the opposite, although showing some inhibitory effects at high concentrations, compound 314 displayed no specific binding for HIV-1 CA at 200 µM (data not shown). This suggests that the effect of 314 on HIV-1 replication might not be specific to CA inhibition and that this compound might interfere with another step of viral replication. As it is an efficient inhibitor of some parasite proteases [33], an inhibitory effect of 314 against HIV-1 protease will be worth exploring. Thus, we focused on compound 696 as it showed a clear doseresponse inhibition of HIV-1 replication in cellulo and a specific affinity for HIV-1 CA in the micromolar range.

Compound 696 does not block HIV-1 infection before the end of the reverse transcription
Various mechanisms of action have been described so far for CA inhibitors (reviewed in [20]): inhibition of the formation of the viral core in the late steps of viral replication by inhibition of CA/CA interactions will lead to the production of non-infectious viral particles, while inhibition of early stages of infection, during uncoating and/or nuclear import of the capsid will lead to abortive infection before integration and will block further viral replication.
As inhibition of infection of U87-CD4-CCR5 cells can reflect inhibition at both early and late steps of infection, we evaluated specifically the effect of compound 696 on the early phase of the viral cycle. To check if the addition of the CA inhibitor inhibited HIV-1 reverse transcription, qPCR was performed on cells preincubated with 696 for 18 h (Fig. 5) before infection.
The nucleoside analog Zidovudine (AZT) was used as a positive control for inhibition of reverse transcription.

Modelling the mechanisms of inhibition by 696
To understand the molecular mechanisms of inhibition, compound 696 was computationally docked against a monomer of HIV-1 CA (Fig. 6). The molecular docking for 696 suggests that this compound binds to the NTD of CA, in a zone that is involved in protein: protein interactions during HIV-1 CA oligomerization [19]. Thus, we positioned compound 696 at the interface between two monomers of an HIV-1 CA hexamer. Compound 696 appears to bind at the interface between monomers (Fig. 7). This suggests that this compound may inhibit HIV-1 replication in cells by interfering with Gag:Gag interactions and/or capsid transportation during the viral cycle. Noteworthy, PF74 and 696 are structurally related and PF74 has been shown to inhibit HIV-1 by targeting the same area in a rather similar orientation (Fig. 7). The main difference between the two compounds is the extrusion of PF74, which presents a larger interface of interaction with the second monomer of CA (130 Ǻ2) than 696 (16 Ǻ2), as calculated by PISA [35].
Interestingly, this zone is the binding site of cellular partners Nup153 and CPSF6, which are involved in the import of the CA or its content from the cytoplasm to the nucleus before HIV-1 integration [8][9][10].
Indeed, PF74 has been described as inhibiting the uncoating step by stabilizing the assembled viral capsid, thereby targeting early steps of viral replication and resulting in the blockade of HIV-1 integration [21,36]. Our molecular docking strategy, which proved successful in identifying the binding site of 696 of FIV CA [27], suggests that 696 is binding at the same site as PF74 on HIV-1 CA However, we showed that this compound does not inhibit the early steps of HIV-1 replication, before reverse transcription (Fig. 5). This could come from a weaker interaction of 696 with HIV-1 CA, suggested by the smaller surface of interaction with the second monomer of CA compared to PF74 (Fig. 7). It could also reflect that 696, as PF74, could interfere with the import of the HIV-1 capsid in the nucleus at the pre-integration step [34]. Structure-based optimization of 696 to increase this interaction and more detailed experiments to address which step of the HIV-1 viral cycle is inhibited, either pre-integration events and/or assembly itself, will be sought to clarify this point.
Noteworthy, PF74 has been described as having low metabolic stability [22] (Table 1), which is not the case here as 696 was stable for 4 h under microsomal and cytosolic fraction derived from rat liver, as checked by Thin Layer Chromatography by the absence of new metabolites ( Figure S2 in supporting material). 696 demonstrated also a selective activity of at least 10 times between mammalian cell and antiviral activity (Figs. 2 and 3). These observations, together with the low toxicity effects predicted in vivo (see supporting information) make this compound a safe hit candidate for a drug development process. Moreover, 696 is a benzimidazole derivative, which is encouraging as such derivatives are among the most frequently used ring systems in drugs listed by the US FDA [37]. Furthermore, 696 has the advantages of easy and low-cost preparation (see supporting information). This is a crucial point, as the affordability of prescription drugs is a limitation in low-income countries with high HIV-1 prevalence, which host most of the HIVinfected individuals worldwide. Nowadays, that is the real limitation in HIV control because most people cannot afford the costs of the actual treatments. This needs to be kept in mind for the upcoming optimization based on compound 696 as a hit molecule, before being used as an anti-HIV-1 therapeutic molecule.

Description of the compounds
The herein studied compounds were selected using the following criteria: i) agents inhibiting CA assembly of FIV in vitro, ii) compounds with strong activity in other pathogens with low-cost production, or iii) compounds with low unspecific toxicity. Then from the chemical collection with around 2000 exclusive compounds designed by our team, we selected 10 compounds belonging to diverse structural families: steroids, quinoxalines, curcuminoids, thiazolidene hydrazines, and benzimidazole. Then, based on the compound which showed an inhibitory effect in our assembly assay, we extended our selection with additional molecules related to these lead compounds to finally consider 22 molecules in total for their inhibitory effect on HIV-1 replication in U87 cells. The chemical synthesis of these molecules followed the previously reported procedures [28,30,33,38,39]. Synthesis and in-depth characterization of the two compounds 696 [27] and 314, which are thoroughly analyzed in this study, are available as supporting information.

Liver fraction stability studies
For the determination of liver fraction stability, rat liver microsomal and cytosolic proteins were used (MOL-TOX ™ manufactures products, North Carolina, US). Rat liver microsomal and cytosolic proteins were used according to the manufacturer's recommendations  and the methodology previously described [40]. The protein content of the microsomal and cytosolic fractions was determined by the bicinchoninic acid assay from Sigma (Schnelldorf, Germany), as suggested by the manufacturer. The final concentration of 696 in the aqueous medium was 100 µM and was prepared from a stock solution in DMSO. The solutions were further homogenized and incubated at 37 o C for 1 to 4 h following the OECD recommendation for drug development. After that, thin-layer chromatography of ethyl acetate extracts was performed to evaluate the presence of metabolites.

Calculation of the pharmacokinetic parameters [40, 41]
The predictions were performed as described before [42] with the open-access SwissADME software (http:// www. swiss adme. ch, accessed on 25 February 2020), a tool that allows the prediction of different pharmacokinetic parameters such as water solubility, lipophilicity, gastrointestinal absorption, bioavailability, etc. (see supporting information). The oral median lethal dose (LD 50 ) in rats was estimated using the Toxicity Estimation Software Tool (TEST). For both of them, the SMILES codes, generated with the ChemBioOffice 2010 program, were used.

Red blood cell lysis assay [32]
Human blood (blood discarded from the blood bank) collected in sodium citrate solution (3.8%) was centri-  The experiments were done in quintuplicate. AmpB was used as a positive control. The EC 50 was defined as the drug concentration at which 50% of the cells were viable relative to the control (0.4% DMSO), and was determined by analysis using Origin-Lab8.5 ® sigmoidal regression (% of viable cells compared to the logarithm of the compound concentration) [39]. Tests were performed in triplicate, in two independent experiments.

Expression and purification of HIV-1 CA
The HIV-1 CA protein was expressed and purified as described for FIV CA with adaptations [31,43]. Escherichia coli cells (BL2I (DE3) pLysS, Lucigen, WI, US) transformed with HIV-1 pNL4-3 CA expressing plasmid (cordially provided by Professor Michael Summers, University of Maryland, US) were grown in Terrific Broth Medium (Sigma-Aldrich, Schnelldorf, Germany), supplemented with 50 mg/mL of ampicillin, at 37 °C. When cells reached an OD value between 0.3 and 0.4 at 600 nm, the expression of CA was induced by the addition of 1 mM isopropyl-D-1-thiogalactopyranoside (IPTG, Euromedex, Souffelweyersheim, France) for 20 h at 25 °C, then cells were harvested by centrifugation (7000 × g) and the pellets were stored overnight at -20 °C. The pellet was then treated with lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, pH 8.5, lysozyme 1 mg/mL, DNase I 1 mg/mL, protease inhibitor cocktail 80 mg/mL, MgCl 2 4 mM) for 1 h at 0 °C, followed by 4 steps of sonication (1 min then 30 s of rest) at full power (Omni Sonic Ruptor 400 ultrasonic Homogenizer, PerkinElmer, MA, US). The clarification step was performed twice by centrifugation at 10,000 × g for 20 min and the supernatant was filtered on a 0.45 µm filter. The purification of HIV-1 CA protein was performed by nickel affinity chromatography using batch incubation with Ni 2+ TED resin (0.15 g of resin per g of pellet, Macherey-Nagel, Düren, Germany), then loaded onto a gravity column. The column was washed three times with LEW buffer (50 mM NaH 2 PO4, 300 mM NaCl, pH 8.5), and the elution was performed with LEW buffer containing increasing concentrations of imidazole (5 mM, 50 mM, 500 mM). The concentration of CA was quantified by spectrophotometry at 280 nm using a Nanodrop (Thermo Fisher, MA, US). The purity of the protein at homogeneity was checked using SDS-PAGE analysis before further processing. Buffer exchange, using

Assembly inhibition assay
The CA assembly assay was performed in a final volume of 70 µL in 384-well plates, with a final concentration of 5 mg/mL of purified recombinant HIV-1 CA in Phosphate/NaCl Buffer (NaH 2 PO 4 /Na 2 HPO 4 50 mM, NaCl 100 mM, pH 7.4), as described for FIV CA [27,31]. incubator at 37 °C. THP-1 monocytes were obtained from ATCC. Human red blood cells were purified from blood discarded from the local blood bank.

Virus
HIV-1 NL4-3 Ba-L was produced by transient transfection of HEK293T cells with pNL4-3 Ba-L using the calcium phosphate method. The infectious virus was harvested at 48 and 72 h after transfection and filtered through a 0.22 μm filter. NL4-3 Ba-L titers were determined on TZM-bl cells.

Inhibition of HIV-1 replication in U87-CD4-CCR5 cells
Compounds were first dissolved in DMSO to reach a concentration of 20 mM. They were further diluted to a 200 µM working stock and then further diluted to reach a starting concentration of 10 µM in the culture plate. U87-CD4-CCR5 cells were seeded in quadruplicate at 4000 cells/well in a final volume of 100 µL. The next day, 24 µL of the compound was added in serial 1:2 dilutions to the cells. Two hours later, 25 µL of infectious virus (NL4-3 Ba-L) were added to the cells with a TCID 50 of 250. AZT at 10 µM was included as an inhibition control. The medium was replenished in the presence of compounds once every 3 to 4 days. The supernatant was collected every 3-4 days and analyzed for virus production by an in-house p24 capture ELISA. IC 50 was calculated using "Quest Graph ™ IC 50 Calculator" (AAT Bioquest, Inc., https:// www. aatbio. com/ tools/ ic50-calcu lator) [44].
Cytotoxicity of the compound was tested on U87-CD4-CCR5 cells in parallel by MTT assay (at the same protocol as in Sect. 3.5 but after 11 days of incubation)) [39].

Molecular docking
The molecular docking experiments were mostly performed as previously described [31]. The hexameric crystal structure of the HIV-1 capsid protein CA was used as the target (PDB entry 4XFX [19]). Compounds were modeled using the smiles code from the ChemOffice software. Then, CA and compounds were prepared using AutoDockTools v1.5.6 [49]. The polar hydrogen atoms were added, the non-polar hydrogens were merged, and the Gasteiger partial atomic charges were computed. Finally, all the possible rotatable bonds were assigned for each compound molecule. A "blind docking" was then carried out with the program AutoDock Vina v1.1.2 [50]. Compounds were treated as flexible while the target protein was treated as rigid. The search space was defined to encompass the entire surface of the target protein, and a large exhaustiveness value of 500 was used regarding the wide search space (70 × 60 × 70 Å 3 ). A visual examination of the resulting docking poses was carried out using PyMOL (Schrödinger, LLC).

Conclusions
As it shows good metabolic stability and low unspecific toxicity, as it targets a region of CA which appears promising in a therapeutic goal, as it inhibits HIV-1 replication in vitro even though it has not been optimized, and given its low-cost synthesis, compound 696 represents an interesting lead molecule which warrants further optimization to generate potential anti-HIV therapeutic derivatives. Moreover, as compounds with closely related structures did not demonstrate any inhibition of viral replication, we concluded that this benzimidazole system is very specific in its antiviral effects.
Additional file 1. Details of the synthesis and characterization of the compounds, and structures of all the assayed molecules.