Design, synthesis and antiviral evaluation of novel heteroarylcarbothioamide derivatives as dual inhibitors of HIV-1 reverse transcriptase-associated RNase H and RDDP functions

Abstract

In the continuous effort to identify new HIV-1 inhibitors endowed with innovative mechanisms, the dual inhibition of different viral functions would provide a significant advantage against drug-resistant variants. The HIV-1 reverse transcriptase (RT)-associated ribonuclease H (RNase H) is the only viral-encoded enzymatic activity that still lacks an efficient inhibitor. We synthesized a library of 3,5-diamino-N-aryl-1H-pyrazole-4-carbothioamide and 4-amino-5-benzoyl-N-phenyl-2-(substituted-amino)-1H-pyrrole-3-carbothioamide derivatives and tested them against RNase H activity. We identified the pyrazolecarbothioamide derivative A15, able to inhibit viral replication and both RNase H and RNA-dependent DNA polymerase (RDDP) RT-associated activities in the low micromolar range. Docking simulations hypothesized its binding to two RT pockets. Site-directed mutagenesis experiments showed that, with respect to wt RT, V108A substitution strongly reduced A15 IC₅₀ values (12.6-fold for RNase H inhibition and 4.7-fold for RDDP), while substitution A502F caused a 9.0-fold increase in its IC₅₀ value for RNase H, not affecting the RDDP inhibition, reinforcing the hypothesis of a dual-site inhibition. Moreover, A15 retained good inhibition potency against three non-nucleoside RT inhibitor (NNRTI)-resistant enzymes, confirming a mode of action unrelated to NNRTIs and suggesting its potential as a lead compound for development of new HIV-1 RT dual inhibitors active against drug-resistant viruses.

Pyrrolyl heteroarylcarbothioamide derivatives are new dual inhibitors of both ribonuclease H and polymerase HIV-1 RT-associated functions, a new promising resource against drug resistance.

INTRODUCTION

Human immunodeficiency virus type 1 (HIV-1) causes a persistent infection integrating its genome into the host genome and continuously replicating, requiring a life-long treatment to maintain the viral loads under control in order to avoid the progression to AIDS (Cohen et al.2008). High viral mutation rate and suboptimal treatment adherence allow the selection of viral strains resistant to the approved drugs (Gregson et al.2016), hence entailing a continuous effort by the scientific community to identify new inhibitors, possibly targeting viral coded essential proteins with innovative mechanisms.

HIV-1 reverse transcriptase (RT) is a key multifunctional enzyme which combines an RNA-dependent DNA polymerase activity (RDDP), a DNA-dependent DNA polymerase activity (DDDP) and a ribonuclease H (RNase H) function, all essential for the viral genome replication (Schatz et al.1990). Up to now, all approved RT-targeting drugs inhibit only the polymerase activity (Menéndez-Arias, Sebastián-Martín and Álvarez 2016), while no inhibitor has been approved yet targeted at the RNase H, whose essential role played in retroviral replication makes it a very promising target for the development of antiretroviral agents (Corona, Esposito and Tramontano 2014). Small molecules shown to inhibit RNase H activity in preclinical studies can be grouped, according to their mode of action (Corona et al.2013), into (i) metal-chelating active site inhibitors, which can bear different scaffolds including N-hydroxyimides (Klumpp and Mirzadegan 2006), N'-acylhydrazones (Carcelli et al.2017), pyrrolyl and quinolinonyl diketoacids (Costi et al.2014; Corona et al. 2014a, 2016b; Cuzzucoli Crucitti et al.2015), 2-hydroxyisoquinoline-1,3-diones (Billamboz et al. 2008, 2010; Tang et al.2017), 3-hydroxypyrimidine-2,4-diones (Tang et al.2011; Vernekar et al. 2017), hydroxypyridone carboxylic acids (Kankanala et al.2016) and various natural derivatives (Chung et al.2011); (ii) allosteric RNase H inhibitors, which can be distinguished in (a) allosteric selective RNase H inhibitors (Himmel et al.2006); (b) p66/p51 interface inhibitors (Tintori et al.2016) and (c) dual inhibitors of both RT-associated enzyme functions (Wang et al.2017).

The last class of compounds is of particular interest since the simultaneous inhibition of both RT activities would provide a significant advantage against drug-resistant variants (Distinto et al.2013). The allosteric dual RNase H/RDDP inhibition property has been described for scaffolds characterized by variously substituted aromatic portions bridged by a functionalized spacer containing moieties able to donate or accept hydrogen bonds, either linear or made by heterocyclic rings (Distinto et al.2013; Meleddu et al.2014, 2017). It has been proposed that these compounds are characterized by an innovative mechanism of action since they could bind to two different RT allosteric pockets: the first one located in the DNA polymerase domain (partially overlapping the non-nucleoside RT inhibitor [NNRTI] binding pocket) and the second one below the RNase H active site close to the interface between p66/p51 (Corona et al.2016a).

Among the heterocycles, pyrrole and pyrazole derivatives have been previously investigated as NNRTIs (Patel and Park 2015; Liu et al.2016). Pyrrole derivatives were first identified as NNRTIs within a mass screening of the Parke-Davis Pharmaceutical compound library. Since then the pyrrole moiety has been incorporated in several NNRTI series (Antonucci et al.1995; Ding et al.1995; Mao et al.1998; Almerico et al.2002). In particular, highly substituted pyrroles were proven to be active against HIV replication and considered a promising and versatile scaffold to develop anti-HIV agents (Antonucci et al.1995). Later on pyrrole-based diketo acid derivatives were the first class of compound shown to selectively inhibit the RNase H function (Tramontano et al.2005), interacting with highly conserved amino acidic residues within the RNase H domain (Corona et al.2014a). Benzylpyrazole derivatives were also identified to act as NNRTIs (Mowbray et al.2009), showing an excellent broad-spectrum activity of against RT drug-resistant variants (Corbau et al.2010). Similarly, N-Arylpyrazolecarboxamide derivatives have been described as NNRTI showing in vitro activity against wild-type and NNRTI-resistant HIV-1 isolates species (Wildum et al.2013). More recently, pyrazole derivatives have been reported as dual CCR5/CXCR4 HIV entry and RT inhibitors (Cox et al.2015). Interestingly, despite their lack of HIV-1 RT-associated RDDP inhibitory activity, some N-benzylpyrazole derivatives inhibited the synthesis of early RT products (Kim et al.2015).

Noteworthy, the potential of functionalized pyrrole and pyrazole moieties for the design and the identification of dual function RT allosteric inhibitors have not been explored yet. Hence, in the present work we synthesized a library of 1H-pyrrole-3-carbothioamide and 1H-pyrazole-4-carbothioamide derivatives and tested them on the HIV-1 RNase H activity and viral replication. The pyrazolecarbothioamide derivative A15 was identified as the most promising compound since it was able to inhibit RNase H activity in the low micromolar range and showed to be slightly active against viral replication. Thus, this compound was also tested on the RDDP HIV1-RT activity and shown to be a new dual HIV-1 RT-associated functions inhibitor. Therefore, its possible mechanism of action was investigated by means of biochemical and in silico studies.

MATERIALS AND METHODS

Synthesis and characterization

Compounds were synthesized by slight modification of our previously reported procedure (Cocco, Congiu and Onnis 1995, 2003; Cocco et al. 1999, 2006) (supplementary material S3: Scheme 1 and 2).

Commercially available solvents and reagents were used without further purification unless otherwise stated. ¹H NMR spectra were recorded on a Varian Inova 500 spectrometer. The chemical shifts (δ) are reported in part per million downfield from tetramethylsilane, which was used as internal standard, and the spectra were recorded in DMSO-d₆. Splitting patterns are designated as singlet (s), doublet (d), triplet (t), quartet (q) and multiplet (m). Infrared spectra were recorded on a Bruker Vector 22 spectrometer. The main bands are given in cm⁻¹. Melting points (m.p.) were determined on a Stuart Scientific Melting point SMP1 apparatus and are uncorrected. The purity of tested compounds was determined by combustion elemental analyses conducted by the Microanalytical Laboratory of the Chemistry Department of the University of Ferrara with a Yanagimoto MT-5 CHN recorder elemental analyzer and all values were within 0.4% of the calculated values, which indicates >95% purity of the tested compounds. Analytical thin layer chromatography was performed using 0.25 mm silica gel 60-F plates.

Cyanoacetamidrazones 10, 12, 13, 15–17 (Cocco et al.1996), pyrazoles A1 and A18 (Cocco et al.2006), pyrazoles A2, A9, A14–17(Cocco et al.1999), pyrrolecarbothioamides B7, B18–19 (Cocco, Congiu and Onnis 1995) and pyrrolecarbothioamides B2–3, B5–10, B12–16 (Cocco, Congiu and Onnis 2003) were prepared as previously reported. The preparation and characterization of all the other pyrazole and pyrrole derivatives is described in supplementary material (S4)

Biological activity

Expression and purification of recombinant HIV-1 RTs group M subtype B

p6HRT-prot vector was kindly provided by Dr Stuart Le Grice Laboratory. Heterodimeric RT was expressed essentially as described (Corona et al.2016a). Briefly, Escherichia coli strain M15 containing the p6HRT-prot vector was grown up to an OD600 of 0.7 and induced with isopropyl β-D-1-thiogalactopyranoside (IPTG) 1.7 mM for 4 h. Cell pellets were resuspended in Lysis buffer (50 mM Sodium Phosphate pH 7.8, 0.5 mg/mL lysozyme), incubated on ice for 20 min, added 0.3 M final NaCl, sonicated and centrifuged at 30 000 × g for 1 h. The supernatant was loaded into a Ni²⁺-sepharose column pre-equilibrated with Loading Buffer (50 mM sodium phosphate pH 7.8, 0.3 M NaCl, 10% glycerol, 10 mM imidazole) and washed thoroughly with Wash Buffer (50 mM sodium phosphate pH 6.0, 0.3 M NaCl, 10% glycerol, 80 mM imidazole). RT was gradient-eluted with Elute Buffer (Wash buffer with 0.5 M imidazole). Fractions were collected, and protein purity was checked by SDS-PAGE and found to be higher than 90%. RT-containing fractions were pooled and diluted 1:1 with Dilute Buffer (50 mM sodium phosphate pH 7.0, 10% glycerol) and then loaded into a Hi-trap Heparine HP GE (Healthcare Lifescience) pre-equilibrated with 10 columns volume of Loading Buffer 2 (50 mM sodium phosphate pH 7.0, 10% glycerol, 150 mM NaCl). Column was then washed with Loading Buffer 2 and RT was gradient-eluted with Elute Buffer 2 (50 mM sodium phosphate pH 7.0, 10% glycerol, 150 mM NaCl). Fractions were collected, and protein was dialyzed and stored in buffer containing 50 mM Tris HCl pH 7.0, 25 mM NaCl, 1 mM EDTA, 50% glycerol. Catalytic activities and protein concentration were determined. Enzyme-containing fractions were pooled, and aliquots were stored at −80°C.

Site-directed mutagenesis

Amino acid substitutions were introduced into the p66 HIV-1 RT subunit coded in a p6HRT-prot plasmid using the QuikChange protocol (Agilent Technologies Inc., Santa Clara, CA, USA).

HIV-1 DNA polymerase-independent RNase H activity determination

The wt and mutated HIV RT-associated RNase H activity was measured as described (Corona et al.2014b) in 100-μL reaction volume containing 50 mM Tris HCl pH 7.8, 6 mM MgCl₂, 1 mM dithiothreitol (DTT), 80 mM KCl, hybrid RNA/DNA (5^΄-GTTTTCTTTTCCCCCCTGAC-3^΄-Fluorescein, 5^΄-CAAAAGAAAAGGGGGGACUG-3^΄-Dabcyl) and different amount of enzymes according to a linear range of dose–response curve: 20 ng of wt RT; 60 ng K103N; 75 ng V108A; 5 ng Y181C; 30 ng Y188L; 100 ng A502F. The reaction mixture was incubated for 1 h at 37°C, the reaction was stopped by addition of EDTA and products were measured with a Victor 3 (Perkin) at 490/528 nm.

HIV-1 RNA-dependent DNA polymerase activity determination

The HIV-1 RT-associated RDDP activity was measured as described (Meleddu et al.2014), briefly in 25 μL volume containing 60 mM Tris‐HCl pH 8.1, 8 mM MgCl₂, 60 mM KCl, 13 mM DTT, poly(A)-oligo(dT), 100 μM dTTP and 6 ng wt RT (or 30 ng K103N RT; 19 ng V108A RT; 1,5 ng Y181CRT; 15 ng Y188L RT; 15 ng A502F RT). After enzyme addition the reaction mixture was incubated for 30 min at 37°C and enzymatic reaction was stopped by addition of EDTA. Reaction products were detected by picogreen addition and measured with a Victor 3 (Perkin) at 502/523 nm.

HIV replication

Drug-mediated inhibition of virus-induced cytotoxicity was assayed in MT-2 cells as described (Esposito et al.2011) with minor modifications. Briefly, triplicate wells of 96-well plates containing 1 × 10⁴ MT-2 cells were infected with HIV-1 IIIB strain at a multiplicity of infection of 0.1. Serial dilutions of drug were added immediately after infection. Cell viability was quantified 5 days after infection with the MTT-dye reduction method.

Detection of protein inhibitor interactions by differential scanning fluorimetry

Thermal stability assays were performed according to Nettleship et al. (2008). To a LightCycler 480 96-well plate (Roche), 1 μL of 500 μM inhibitor in DMSO was added, followed by 49 μL of 300 nM HIV-1 RT in reaction buffer containing 20 mM HEPES, pH 7.5, 10 mM MgCl₂, 100 mM NaCl and a 1:1000 dilution of Sypro Orange dye (Invitrogen). The mixture was heated from 30°C to 90°C in increments of 0.2°C. Fluorescence intensity was measured using excitation and emission wavelengths of 483 and 568 nm, respectively. Changes in protein thermal stability (ΔTm) upon inhibitor binding were analyzed by using LightCycler 480 software. All assays were performed in triplicate.

Molecular modeling

Ligand preparation

The ligand was built within the Maestro platform, and the geometry was optimized with Macromodel (Mohamadi et al.1990) using the Merck Molecular Force Fields (Halgren 1996), the GB/SA solvation model (Still et al.1990) and the Polak-Ribier Coniugate Gradient (PRCG) method converging on gradient with a threshold of 0.05 kJ/(molÅ).

Protein preparation

Starting crystal coordinates of the complex RT-RNase H inhibitor were downloaded from the Protein Data Bank (http://www.rcsb.org/) pdb accession code 3lp2 (Su et al.2010). Afterward, the protein was then prepared using the Schrödinger protein preparation wizard (Schrödinger 2013): hydrogen atoms were added to the system. Partial atomic charges were assigned according to the Optimized Potential for Liquid Simulations (OPLS-2005) (Jorgensen et al.1996) force field. Asn103 was mutated in Lys. A minimization was performed to optimize hydrogen atoms and to remove any high-energy contacts or distorted bonds, angles and dihedrals. Starting from wt protein, the mutated enzymes V108A-RT and A502F-RT were generated. Both mutated RTs were minimized considering OPLS force field in GB/SA implicit water (Still et al.1990), setting 10 000 steps interactions analysis with the PRCG method and a convergence criterion of 0.1 kcal/mol.

Docking and post-docking

Compound A15 was docked into wt and mutated RTs applying QM-polarized ligand (QMPL) docking protocol applying default settings (Schrödinger). The enzyme was divided into boxes of the same size (46 × 46 × 46 Å) covering overall the whole p66 subunit. The docking grids were defined by centering on W229 and Q500. In order to better take into account the induced fit phenomena, the most energy favored generated complexes were fully optimized with the OPLS force field in GB/SA implicit water (Still et al.1990), setting 10 000 steps interactions analysis with the PRCG method and a convergence criterion of 0.1 kcal/(molÅ). The resulting complexes were considered for the binding modes graphical analysis with Pymol Version 1.5.0.4. Molecular Graphics System (Schrödinger, LLC) and Maestro (Schrödinger 2015).

RESULTS AND DISCUSSION

We considered a small in-house library of 38 compounds, 18 3,5-diamino-N-aryl-1H-pyrazole-4-carbothioamide derivatives and 20 4-amino-5-benzoyl-N-phenyl-2-(substituted-amino)-1H-pyrrole-3-carbothioamide derivatives. All compounds own the described pharmacophoric requirements for RNase H/RDDP inhibition (Distinto et al.2013; Meleddu et al.2015), with two or three substituted aromatic rings linked by a functionalized spacer containing moieties able to donate or accept hydrogen bonds (Fig. 1).

Figure 1.

Schematic representation of synthesized compounds. (A) 3,5-diamino-N-aryl-1H-pyrazole-4-carbothioamide derivatives. (B) 4-amino-5-benzoyl-N-phenyl-2-(substituted-amino)-1H-pyrrole-3-carbothioamide derivatives.

Biological activity

To identify dual inhibitors, our approach was to first test compounds against the RT-associated RNase H function. Results showed that 15 out of 38 tested compounds inhibited the RNase H activity with IC₅₀ values ≤20 μM (Tables 1 and 2).

Table 1.

HIV-1 RT-associated RNase H inhibition by 3,5-diaminopyrazole derivatives.

Table 2.

HIV-1 RT-associated RNase H inhibition by 4-amino-1H-pyrrole-3-carbothioamide derivatives.

In particular, among the pyrazole-4-carbothioamide derivatives, the diphenyl derivative A1 showed an IC₅₀ value of 98 μM (Table 1), while the introduction of a phenylacetyl substituent in R led to a consistent increase in potency (A2, IC₅₀ value of 19 μM; Table 1). The introduction of a 4-chlorine into the phenylacetyl moiety led to compound A5, showing a further improved activity (IC₅₀ value of 4 μM). The anti-RNase H activity was maintained by replacing the carbothioamide moiety with the carboxamide group (compound A6, IC₅₀ 5 μM), while the introduction of a second halogen into the 4-Cl phenylacetyl (A7) led to a complete loss of activity. Interestingly, the introduction of substituents on the benzoyl moiety in 4-(chlorophenyl) acetyl carbothioamide derivatives was detrimental for the activity (compounds A8-A14), with the exception of A13 (IC₅₀ 21 μM), whereas the cotemporaneous presence of a 4-chlorobenzoyl moiety and a phenyl carbothioamide led to an increase in potency (compound A15 IC₅₀ of 7 μM). Moreover, the introduction a methylene spacer between the carbothioamide nitrogen and the phenyl ring (compound A16) or the presence of a small acyl group on pyrazole N-1 (compounds A17–18) led to a consistent loss in activity.

Among the pyrrole derivatives, 7 out of 20 compounds exhibited IC₅₀ values in the low micromolar range (Table 2). Results showed that the 4-methylbenzoyl moiety in the 2-amino group led to compounds (B1–B6) showing IC₅₀ values ranging from 6 to 8 μM, with the exception of 4-chlorophenylamino and of 3,4-dichlorophenylamino derivatives. The replacement of the 4-methylbenzoyl group with a 4-methoxybenzoyl was detrimental for the activity, since among compounds B7–B13 only the 4-phenylamino derivative B8 retained some RNase H inhibition properties (IC₅₀ of 19 μM). Differently, in the 4-chlorobenzoyl-pyrrole series (B14–B17), the presence of 4-chlorophenylamino group did not affect RNase H potency of inhibition, since B17 showed the same potency of inhibition observed for 4-methylphenylamino and 4-methoxyphenyl analogs B15 and B16 (IC₅₀ values of 8–9 μM). Finally, the replacement of the aroyl moiety with a small alkylamino group (B18) or a pyridylamino moiety (B20) was detrimental for the anti-RNase H activity in the pyrrole derivatives bearing an unsubstituted benzoyl moiety, while its replacement whit a phenyl acetyl moiety (B19) preserved the activity.

Next, all compounds showing IC₅₀ values below 20 μM were screened for antiviral activity in cell-based assays. Results showed that the pyrazole-4-carbothioamide derivatives were not cytotoxic and that one of them, compound A15, inhibited HIV-1 replication in cell-based assays with an EC₅₀ value of 25 μM, showing a selectivity index of 1.8 (Table 3). Differently, all tested pyrrole derivatives were cytotoxic.

Table 3.

Inhibition of HIV-1 replication inhibition by selected 3,5-diaminopyrazole and 4-amino-1H-pyrrole-3-carbothioamide derivatives.

	HIV-1 aEC50 (μM)	b CC50 (μM)
Compounds		TZMbl	CEM cells
A2	>50	>50 (100%)c	> 50 (98%)c
A5	> 50	>50 (100%)c	> 50 (94%)c
A6	>50	>50 (100%)c	> 50 (92%)c
A15	25 ± 2	44 ± 3	> 50 (79%)c
B1	>30	30 ± 5	> 50 (85%)c
B2	>6.7	6.7 ± 1.2	11 ± 2
B3	>8.7	8.7 ± 1.4	7.5 ± 0.6
B5	>30	9.0 ± 1.8	>50 (57%)c
B8	>30	8.5 ± 1.3	50 ± 2
B14	>30	30 ± 4	>50 (80%)c
B15	>30	30 ± 2	54 ± 3
B16	>25	25 ± 3	36 ± 5
B17	>18	18 ± 2	24 ± 3
B19	>7.2	7.2 ± 0.6	13 ± 2
EFV	0.14 ± 0.02	>50 (100%)c	> 50 (100%)c

^aCompound concentration (± standard deviation) required to decrease viral replication in TZMbl cells by 50%.

^bCompound concentration (± standard deviation) required to reduce cell viability by 50%.

^cPercentage of cells viability in presence of 50 μM inhibitor.

Searching for dual function RT inhibitors, we then tested compound A15 on the RT-associated RDDP activity showing that A15 inhibited an IC₅₀ value of 17 μM (Fig. S1, Supporting Information). Hence, compound A15 was shown to be a dual RNase H/RDDP RT inhibitor with a 2-fold higher potency for RNase H inhibition and, to acquire information useful to optimize the scaffold, we investigated its mechanism of action and binding mode.

A15 derivative metal chelating and dimer stability interfering properties

To gain more insights into the A15 mode of action, we first asked whether it could bind to the RT active site chelating the magnesium. For this, we investigated its potential to chelate the magnesium by measuring its UV spectra in the absence and in the presence of magnesium ions. Results showed that A15 maximum of absorbance did not shift in the presence of 6 mM MgCl₂, excluding the involvement of chelation in the mechanism of action (Fig. S2, Supporting Information). Second, since previous studies showed that molecules targeting the RT heterodimer dimerization could inhibit both RT-associated enzymatic activities (Tintori et al.2016), we examined by differential scanning fluorimetry whether A15 could alter the HIV-1 RT thermal stability (Cummings, Farnum and Nelen 2006). Results showed that A15 did not affect RT thermal stability, differently from the positive controls used, i.e. the active site metal chelating beta-thujaplicinol (BTP), which has been shown to stabilize the RT against thermal denaturation (Su et al.2010), and the allosteric RNase H inhibitor 2-(3,4-dihydroxyphenyl)-5,6-dimethylthieno[2,3-d]pyrimidin-4(3H)-one (VU) that has been shown to destabilize the HIV-1 RT (Masaoka et al.2013) (Fig. 2). Of note, A15 showed an RT thermal stability pattern similar to the one previously reported for the NNRTI efavirenz (EFV) and RMNC6 (Corona et al.2016a), hence suggesting an allosteric mode of action (Fig. 2).

Figure 2.

Effect of A15 on the thermal stability of p66/p51 HIV-1 RT. The melting temperature of HIV-1 RT was measured by differential scanning fluorimetry in presence of increasing concentrations of (inverted filled triangles) BTP, (filled circles) VU and (open triangles) A15.

A15 derivative docking studies

To further investigate the RT−A15 interactions, we then carried out QMPL blind docking experiments. QMPL docking workflow combines docking with ab initio methods for ligand charges calculation within the protein environment (Schrödinger). Subsequently, the best poses were subjected to post-docking minimization to consider induced-fit protein conformation change (that takes place after ligand binding) and implicit water solvation (Mohamadi et al.1990). These studies suggested that A15 could be able to bind into two RT pockets: one close to the polymerase catalytic site (pocket 1) and one close to RNase H catalytic site (pocket 2).

According to previous reports (Himmel et al.2006), A15 binding into the first pocket could affect both RDDP and RNase H activities, the first by restricting RT thumb domain flexibility and impeding correct nucleic acid positioning, and the latter by deviating the trajectory of the nucleic acid which could not be properly positioned within the RNase H active site. In agreement with this binding model, docking results showed that A15 could occupy the area between the polymerase catalytic site and the NNRTI binding pocket (Fig. 3a and b). Of note, the A15-RT complex seems to be stabilized by hydrogen bonds with D186, K223 and by π–π interaction with W229, which are important and conserved residues that are prime target for the rational design of NNRTIs (Pelemans et al.2000). A15 binding was also proposed to be stabilized by van der Waals interactions with other RT residues such as L100, V108, V111, Y188, F227 and L234. Some of these residues are included in the so called ‘primer grip’ which plays a key role in directing the movement of RT along the template primer. To confirm the proposed binding mode, further molecular docking studies were performed using single point mutated V108A RT (Fig. 3c and d). Results showed that when docked with V108A RT, A15 changed its binding mode, as compared to the wt RT complex, entering more deeply within pocket 1 cavity since steric hindrance of alanine is lower than valine.

Figure 3.

Putative binding mode of A15 in RT pocket 1. (a) A15 depicted in pocket 1 of wt-RT; (c) A15 depicted in pocket 1 in complex of RT-V108A; (b-d) 2D representation of A15 in complex with wt and V108A RT, respectively, with indication of binding pocket interacting residues.

As also previously proposed for other dual RT inhibitor derivatives BHMP07 (Felts et al.2011) and RMNC6 (Felts et al.2011; Corona et al.2016a), blind docking experiments showed that A15 could also bind to a second pocket (pocket 2), located below the RNase H catalytic site (Fig. 4a and b). To confirm this hypothesized binding model, in silico long-energy minimization studies were performed using single point mutated A502F RT. Of note, residue A502 is located in the alpha helix close to the putative binding pocket and has been reported to have a critical role also in the binding of other dual inhibitors (Corona et al.2016a). Substitution of Ala with Phe caused a movement of alpha helix that reduced the space between the two subunits p51 and p66, hindering the entrance of the compound (Fig. 4c and d). Docking experiments performed on A502F RT showed that A15 was not able to bind into the cavity between the two subunits, mostly lying in the enzyme surface exposed to the solvent. The complex A502F RT-A15 was hence overall less stable than the complex wt RT-A15.

Figure 4.

Putative binding mode of A15 in RT pocket 2. (a) A15 depicted in pocket 2 of wt RT; (c) conformational change of the pocket 2 upon mutation of A502F RT (in yellow); wt RT is colored in pink and the A502F RT in orange, residues interacting with the compound are in sticks; d) A15 depicted A15 in RT-A502F pocket 2; (b-e) 2D representation of A15 in complex with wt and A502F RT, respectively, with indication of binding pocket interacting residues.

A15 derivative effects on mutated RTs

In order to confirm the hypotheses suggested by the in silico studies and to compare the mode of action of A15 with the one of classical NNRTIs, site-directed mutagenesis was performed independently introducing the amino acid substitutions V108A, K103N, Y181C and Y188L. In addition, mutated A502F RT was also obtained. Derivative A15 was tested on both RNase H and RDDP activities of these mutated RTs (Table 4). Results showed that A15 retained a good potency of inhibition against all the three NNRTI-resistant enzymes, confirming a binding to RT different from NNRTIs. Furthermore, in agreement with the docking model, the V108A substitution strongly affected the A15 potency of RT inhibition (12.6-fold for RNase H IC₅₀ value and 4.7-fold for RDDP IC₅₀ value, respectively). Also in agreement with the docking model, the A502F substitution caused a 9.0-fold increase in the RNase H IC₅₀ value respect to wt RT.

Table 4.

Inhibition of HIV-1 mutated RT-associated RNase H and RDDP activities by A15.

	A15				BTP		Efavirenz
	RNase H		RDDP		RNase H		RDDP
RT	IC50 (μM)a	fold	IC50 (μM)b	fold	IC50 (μM)a	fold	IC50 (nM)b	fold
wt	7.7 ± 2.0	1	17 ± 4	1	0.19 ± 0.03	1	23 ± 4	1
K103N	22 ± 4	2.9	37 ± 2	2.2	0.22 ± 0.08	1.2	176 ± 25	7.6
Y181C	28 ± 1	3.6	32 ± 3	1.9	0.23 ± 0.05	1.2	50 ± 9	2.2
Y188L	14 ± 1	1.8	32 ± 2	1.9	0.08 ± 0.05	0.4	198 ± 60	8.6
V108A	96 ± 18	12.6	82 ± 2	4.8	0.18 ± 0.04	0.9	21 ± 3	0.9
A502F	69 ± 14	9.0	24 ± 3	1.4	0.17 ± 0.03	0.9	25 ± 2	1.1

^aCompound concentration (± standard deviation) required to inhibit HIV-1 RT-associated RNase H activity by 50%.

^bCompound concentration (± standard deviation) required to inhibit HIV-1 RT-associated RDDP activity by 50%.

Overall, we described the identification of the of 1H-pyrrole-3-carbothioamide-derived scaffold as a promising starting point for the synthesis of dual RT inhibitors, identifying the 3,5-diamino-N-benzoyl-1H-pyrazole-4-(4-chloro-phenyl)carbothioamide derivative, compound A15, as an efficient RT and viral replication inhibitor. Mode-of-action studies showed that A15 acts through a dual RT inhibition, achieved by binding to two different sites on the RT, and retained activity against enzymes carrying mutations commonly conferring resistance to NNRTIs, hence showing a potential for A15 as a hit compound whose scaffold improvement would allow to develop new RT dual inhibitors more active against viral replication and possibly against selected NNRTI-resistant viruses and less susceptible to the selection of drug-resistant viruses.

SUPPLEMENTARY DATA

Supplementary data are available at FEMSPD online.

FUNDING

AC, FE and ET thank Italian MIUR for financial support (PRIN 2010, 2010W2KM5L_003); the National Foundation for Cancer Research grant number PO1CA154295.

Conflict statement. None declared.

Supplementary Material

Supplemental material

Supplementary data are available at FEMSPD online.