Identification of Boronate-Containing Diarylpyrimidine Derivatives as Novel HIV-1 Non-Nucleoside Reverse Transcriptase Inhibitors

In this study, privileged boronic acid ester was introduced into the right wing of etravirine (ETR) to obtain a series of novel boronate-containing derivatives. These newly synthesized derivatives were evaluated for their anti-HIV potency in MT-4 cells using the MTT method, and their inhibitory activity to HIV-1 reverse transcriptase (RT) was assayed by the ELISA method. Most of the synthesized compounds displayed promising antiviral activity against the wild-type and a wide range of HIV-1 mutant strains. In particular, 4a exhibited the most potent activity against the wild-type and a panel of single mutations (L100I, K103N, Y181C, and E138K) with EC50 values ranging from 0.005 to 0.648 μM, which were much superior to those of nevirapine (EC50 = 0.151 μM). Moreover, 4b turned out to be an effective inhibitor against the double-mutant strains F227L + V106A and RES056 with EC50 values of 3.21 and 2.30 μM, respectively. RT inhibition activity and molecular docking were also investigated.


Introduction
Since the first case reported in the United States in 1981 [1], acquired immunodeficiency syndrome (AIDS) has become a major global public health problem which seriously endangers human health. HIV-1, with strong virulence and high mortality, is the main causative agent of AIDS [1]. Reverse transcriptase (RT), which is responsible for reverse transcription of single-stranded RNA into double-stranded DNA, plays an important role in the life cycle of HIV-1. It is unique to the virus and there is no homologous enzyme in human body, making it an ideal target for drug intervention [2]. There are two types of RT inhibitors available on the market: the nucleoside/nucleotide RT inhibitors (NRTIs/NtRTIs) and non-nucleoside RT inhibitors (NNRTIs) [3,4]. Currently, six NNRTIs have been approved by the U.S. Food and Drug Administration (FDA), including the first generation nevirapine (NVP), delavirdine (DLV), and efavirenz (EFV) and the second generation etravirine (ETR), rilpivirine (RPV), and doravirine (DOR) [5]. Moreover, elsulfavirine (ESV) was marketed in Russia for the treatment of HIV in 2017 [6] and ainuovirine (ANV) in China in 2021 [7] ( Figure 1). Although NNRTIs are widely used in clinic due to their potent antiviral activity and high selectivity, their effects are compromised by drug resistance (such as Y181C, E138K, and F227L + V106A) and adverse effects (such as hypersensitivity reactions) [8][9][10][11]. elsulfavirine (ESV) was marketed in Russia for the treatment of HIV in 2017 [6] and ainuovirine (ANV) in China in 2021 [7] (Figure 1). Although NNRTIs are widely used in clinic due to their potent antiviral activity and high selectivity, their effects are compromised by drug resistance (such as Y181C, E138K, and F227L+V106A) and adverse effects (such as hypersensitivity reactions) [8][9][10][11]. Therefore, the development of next-generation NNRTIs with better drug-resistance profiles and lower toxicity is still in high demand. Recently, a considerable amount of literature has illustrated the diverse applications of boronate-containing motifs in the construction of therapeutically useful bioactive molecules [12]. Up to now, four boronate-containing drugs have been approved for clinical use and more are currently in clinical trials [13]. The increasing interest in boronate-containing compounds is due to their unique binding properties to biological targets; for example, boronic acid is a strong Lewis acid because of the open electronic shell of boron, which means that boronic acids can be converted from a triangular planar sp 2 boron form to a tetrahedral sp 3 boron under physiological conditions, thus generating multiple modes when binding to biological targets, such as multiple hydrogen bonding interaction forces and covalent interactions [14]. Therefore, boronic acids and its esters are becoming increasingly prevalent in contemporary drug design. Recent evidence suggests that boronic acid and its esters have been successfully employed in antiviral agents targeting HIV-1 protease and HCV polymerase [15][16][17], which prompted us to shift our antiviral drug research territory to boronate-containing agents [18,19].
As part of our ongoing research, in this work, the privileged phenylboronic acid pinacol ester was introduced in the right wing of the lead ETR to develop stronger interactions with NNRTIs binding pocket (NNIBP), with the hope of improving the activity and drug-resistance profiles ( Figure 2). Meanwhile, according to our previous exploration on the structure-activity relationships (SARs) of NNIBP tolerant region II, various aromatic heterocyclic structures were also introduced into this region, yielding ten novel boronatecontaining derivatives [20][21][22]. Additionally, the privileged 4-cyanovinyl-2,6-dimethylphenyl motif was also introduced to the left wing to develop additional π-π interactions with the highly conserved amino acid W229. The detailed SARs study of these derivatives and molecular docking simulation were also studied. Recently, a considerable amount of literature has illustrated the diverse applications of boronate-containing motifs in the construction of therapeutically useful bioactive molecules [12]. Up to now, four boronate-containing drugs have been approved for clinical use and more are currently in clinical trials [13]. The increasing interest in boronatecontaining compounds is due to their unique binding properties to biological targets; for example, boronic acid is a strong Lewis acid because of the open electronic shell of boron, which means that boronic acids can be converted from a triangular planar sp 2 boron form to a tetrahedral sp 3 boron under physiological conditions, thus generating multiple modes when binding to biological targets, such as multiple hydrogen bonding interaction forces and covalent interactions [14]. Therefore, boronic acids and its esters are becoming increasingly prevalent in contemporary drug design. Recent evidence suggests that boronic acid and its esters have been successfully employed in antiviral agents targeting HIV-1 protease and HCV polymerase [15][16][17], which prompted us to shift our antiviral drug research territory to boronate-containing agents [18,19].
As part of our ongoing research, in this work, the privileged phenylboronic acid pinacol ester was introduced in the right wing of the lead ETR to develop stronger interactions with NNRTIs binding pocket (NNIBP), with the hope of improving the activity and drug-resistance profiles ( Figure 2). Meanwhile, according to our previous exploration on the structure-activity relationships (SARs) of NNIBP tolerant region II, various aromatic heterocyclic structures were also introduced into this region, yielding ten novel boronate-containing derivatives [20][21][22]. Additionally, the privileged 4-cyanovinyl-2,6dimethylphenyl motif was also introduced to the left wing to develop additional π-π interactions with the highly conserved amino acid W229. The detailed SARs study of these derivatives and molecular docking simulation were also studied.

Chemistry
The synthetic protocols for the newly designed compounds 3a-j and 4a-c are outlined in Scheme 1. The synthetic routes and experimental procedures for 1a-j and 2a-c are described in more detail in the Supplementary Information. All derivatives were prepared by well-established methods as described in our previous articles [20,21,[23][24][25]. The previously prepared compounds 1a-j and 2a-c were selected as starting materials and reacted with 4-aminophenylboronic acid pinacol ester in the presence of BINAP and Pd2(dba)3 to yield the target compounds 3a−j and 4a−c via the Buchwald−Hartwig reaction. All novel target compounds were fully characterized by proton nuclear magnetic resonance spectroscopy ( 1 H NMR) and carbon-13 nuclear magnetic resonance spectroscopy ( 13 C NMR).

Chemistry
The synthetic protocols for the newly designed compounds 3a-j and 4a-c are outlined in Scheme 1. The synthetic routes and experimental procedures for 1a-j and 2a-c are described in more detail in the Supplementary Informations. All derivatives were prepared by well-established methods as described in our previous articles [20,21,[23][24][25]. The previously prepared compounds 1a-j and 2a-c were selected as starting materials and reacted with 4-aminophenylboronic acid pinacol ester in the presence of BINAP and Pd 2 (dba) 3 to yield the target compounds 3a-j and 4a-c via the Buchwald−Hartwig reaction. All novel target compounds were fully characterized by proton nuclear magnetic resonance spectroscopy ( 1 H NMR) and carbon-13 nuclear magnetic resonance spectroscopy ( 13 C NMR).

Anti-HIV-1 Activity Evaluation
All of the newly synthesized DAPY derivatives 3a-j and 4a-c were evaluated agai the wild-type HIV-1 (IIIB) and the double-mutant strain RES056 (K103N/Y181C) in MT-4 cell line using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium brom (MTT) method. The approved NVP, EFV, and ETR were selected as control drugs. T values of EC50 (anti-HIV-1 potency), CC50 (cytotoxicity), and SI (selectivity ind CC50/EC50 ratio) of the target compounds are summarized in Tables 1-3.   As depicted in Table 1, the newly synthesized compounds exhibited moderate to high potency against the HIV-1 IIIB strain with EC 50 values ranging from 0.009 to 6.17 µM. Among them, most of the target compounds displayed low sub-micromolar EC 50 values ranging from 0.009 to 0.215 µM, being more potent or equipotent compared to that of NVP (EC 50 = 0.151 µM). Unfortunately, all derivatives showed weaker efficacy or lost activity against the double-mutant strain RES056.
To design additional NNRTIs with improved activity against the mutant strain RES056, we replaced the cyano group in the left wing of 3a-c with cyanovinyl motif, hoping that it would extend deeper into the hydrophobic channel and develop stronger interactions with the highly conserved residue W229. As shown in Table 1, all the derivatives (4a-c) exhibited increased potencies with EC 50 values ranging from 0.009 µM to 0.014 µM. Among them, 4a was turned out to be the most effective inhibitor against HIV-1 IIIB with an EC 50 value of 0.009 µM, being much superior to that of NVP (EC 50 = 0.151 µM) and comparable to those of ETR (EC 50 = 0.004 µM) and EFV (EC 50 = 0.003 µM). Moreover, 4a displayed much lower cytotoxicity (CC 50 > 238 µM), which was found to be up to 51-fold less than that of ETR (CC 50 > 4.6 µM). For RES056 mutant strain, 4a still showed no activity (EC 50 > 238 µM), but 4b and 4c both exhibited better bioactivity (EC 50 = 2.30 and 13.9 µM, respectively). Surprisingly, 4b was 75-fold more potent than 3b (EC50 = 2.30 vs. 174 µM). This proved that the substitution of cyano group for cyano vinyl group can indeed improve the biological activity of the compounds against RES056 strains.
Furthermore, 3a, 3c, and 4a-c were selected to evaluate the ability to inhibit a variety of NNRTIs-resistant strains, including L100I, K103N, Y181C, Y188L, E138K, and F227L + V106A. As is depicted in Table 2, all the tested compounds maintained modest to high potencies against the whole mutant strains with EC 50 values ranging from 0.005 to 9.10 µM, with the exception of 4a in the F227L + V106A mutant strain. In the case of the K103N mutant strain, the most prevalent resistance to NVP and EFV, all these compounds displayed potent inhibitory activities with EC 50 values ranging from 0.005 to 0.120 µM, being superior or comparable to that of NVP (EC 50 = 3.93 µM). Particularly, compounds 4a (EC 50 = 0.005 µM) and 4b (EC 50 = 0.008 µM) provided the highest potencies towards K103N, which were far superior to those of NVP and EFV (EC 50 = 0.632 and 0.071 µM, respectively) and were comparable to those of ETR (EC 50 = 0.003 µM). Moreover, 4a also maintained high potency against L100I mutant strain with EC 50 value of 0.186 µM, being 3.3-fold more potent than that of NVP (EC 50 = 0.623 µM). More interestingly, all these selected compounds displayed highly potent inhibitory activities against the E138K mutant strain (EC 50 = 0.021-0.193 µM), among which 4a (EC 50 = 0.021 µM) and 4b (EC 50 = 0.030 µM) were the most potent derivatives and exhibited higher activities compared to those of NVP (EC 50 = 0.168 µM). Moreover, 4a also exhibited the most potent inhibitory activity against the Y181C mutant strain (EC 50 = 1.61 µM), being superior to that of NVP (EC 50 = 5.06 µM). It is worth noting that compounds 3a and 4b (EC 50 = 4.95 and 3.21 µM, respectively) showed the prominent inhibitory activities against the F227L + V106A strain, which were superior to NVP (EC 50 = 8.05 µM). In addition, the selectivity index (SI) and resistance fold (RF) of the tested compounds were summarized in Table 3.

HIV-1 RT Inhibition Assay
To validate the binding target of these derivatives, 3a, 3c, and 4a-c were selected to test their inhibitory activity against the recombinant wild-type HIV-1 RT enzyme. As displayed in Table 4, most compounds showed high binding-affinity with the wild-type HIV-1 RT (IC 50 = 0.047-0.354 µM). The most potent compound 4a exhibited the highest RT inhibitory activity (IC 50 = 0.047 µM), being much more potent than that of NVP (IC 50 = 0.638 µM) and comparable to that of ETR (IC 50 = 0.011 µM). The preliminary results demonstrated that these newly synthesized derivatives could bind to HIV-1 RT and behave as typical NNRTIs.

Molecular Docking (MD) Simulation
In order to verify the binding interactions of these newly synthesized compounds and to obtain further insights into their binding modes in the NNIBP of RT, 4a was selected as the representative compound in further molecular docking studies. The co-crystal structures of HIV-1 WT RT/K-5a2 (PDB code: 6C0J) and HIV-1 K103N/Y181C mutant RT/25a (PDB code: 6C0R) were used as the input structure for docking calculation (K-5a2 and 25a were two potent diarypyrimidine (DAPY) derivatives with similar center core to 4a found in our previous work [21,25]). PyMOL was used to visualize the results. The docking protocol is described in the experimental section.
The binding mode of 4a with HIV-1 WT RT is shown in Figure 3. The key amino acid residues surrounding the ligand appear as light gray sticks. The overlapping conformations of 4a with the lead compound ETR is shown in Figure 3A. The results indicated that 4a binds with NNIBP in a horseshoe-like conformation, which is similar to that seen with NNRTIs of the DAPY family, and fully occupies the cavity in NNIBP and remains adaptable to adapt to the cavity due to its high flexibility. As shown in Figure 3B, detailed binding interactions reveal the following features. First, the left 4-cyanovinyl-2,6-dimethylphenyl group of 4a fully occupies the hydrophobic cavity surrounded by hydrophobic aromatic amino acid residues Y181, Y188, F227 and W229, exhibiting a π−π interaction with these residues. Second, the N-atom of the center core and NH linker connecting the central pyrimidine ring and the right wing are involved in double hydrogen bonding with the backbone of K101 through water bridges or directly, which are conserved hydrogen bonds among the secondgeneration NNRTIs/RT complexes. Moreover, the phenyl-linked 4,4,5,5-tetramethyl-1,3,2dioxaborolane is directed to the tolerant region I and can develop extensive interactions with surrounding lipophilic amino acids. In regard to RT carrying the K103N/Y181C doublemutation, Y181C mutation abolishes the favorable π−π stacking interactions between the Y181 side chain and the 4-cyanovinyl-2,6-dimethylphenyl group of 4a, which greatly reduces the binding interface between 4a and C181. Additionally, the dramatic changes in NNIBP result in a decrease of the buried interface between 4a and residue 103, which leads to the lack of the extensive hydrogen-bonding network between 4a and the backbone of K101 ( Figure 3C). The overlapping conformations of 3a and 4a are shown in Figure 3D. The cyanovinyl group of 4a is able to reach deeper into the hydrophobic pocket than the cyano group of 3a and forms a π−π interaction with surrounding hydrophobic amino acid residues, such as F227 and W229. This may have contributed to the superior biological activity of 4a compared to that of 3a.
dramatic changes in NNIBP result in a decrease of the buried interface between 4a and residue 103, which leads to the lack of the extensive hydrogen-bonding network between 4a and the backbone of K101 ( Figure 3C). The overlapping conformations of 3a and 4a are shown in Figure 3D. The cyanovinyl group of 4a is able to reach deeper into the hydrophobic pocket than the cyano group of 3a and forms a π−π interaction with surrounding hydrophobic amino acid residues, such as F227 and W229. This may have contributed to the superior biological activity of 4a compared to that of 3a.

In Silico Prediction of Physicochemical Properties
The drug-like properties of representative compounds 4a, 4b and control drugs ETR and RPV were characterized using a free online software (https://admetmesh.scbdd.com/, accessed on 30 September 2022). As shown in Table 5, the results indicated that the various parameters of 4a and 4b, including hydrogen bond acceptors (nHA), hydrogen bond donors (nHD), topological polar surface area (tPSA), and rotatable bonds (nRot) were all in the optimal range. However, due to the presence of multiple aromatic rings and cyclic

In Silico Prediction of Physicochemical Properties
The drug-like properties of representative compounds 4a, 4b and control drugs ETR and RPV were characterized using a free online software ADMETlab 2.0 (https: //admetmesh.scbdd.com/, accessed on 30 September 2022). As shown in Table 5, the results indicated that the various parameters of 4a and 4b, including hydrogen bond acceptors (nHA), hydrogen bond donors (nHD), topological polar surface area (tPSA), and rotatable bonds (nRot) were all in the optimal range. However, due to the presence of multiple aromatic rings and cyclic borate ester structure in 4a and 4b, their LogP values are higher than those of ETR and RPV.

Synthesis
All melting points were determined on a micro melting point apparatus (RY-1G, Tianjin Tian Guang Optical Instruments). 1 H NMR and 13 C NMR spectra were recorded in DMSOd 6 on a Bruker AV-400 spectrometer with tetramethylsilane (TMS) as the internal standard; signals are abbreviated as s (singlet), d (doublet), t (triplet), and m (multiplet). Chemical shifts are reported in δ values (ppm) from TMS and coupling constants are given in hertz (Hz). The mass spectra were measured in AG1313A Standard LC Autosampler (Agilent). All reactions were routinely monitored by thin layer chromatography (TLC) on Silica Gel GF254 for TLC (Merck) and spots were visualized with iodine vapor or by irradiation with UV light (λ = 254 and 356 nm). Flash column chromatography was performed on columns packed with Silica Gel (Qingdao Haiyang Chemical Company). Solvents were purified and dried by standard methods. The concentration of the reaction solutions involved the use of rotary evaporator at reduced pressure.