Recent Advances in Heterocyclic HIV Protease Inhibitors

Abstract

Since the first cases of AIDS, reported in 1980, this disease has become chronic over the years, and researchers have been trying to keep it under control. Despite the development and spread of mutate viruses, HIV protease remains an important pharmacological target. In the development of new HIV protease inhibitors, heterocyclic fragments have proven to be of great importance, owing to their rigid core structure, which may fit better into the enzyme’s hydrophobic pockets, and the presence of a heteroatom, which may increase the number of H-bonding interactions at the active site. According to the concept of targeting the protein backbone, different aromatic or non-aromatic heterocyclic moieties have yielded inhibitors with sufficient activity against mutant viruses. This paper provides an overview of HIV protease inhibitors developed over the last fifteen years, with a focus on the presence of heterocycles in their structure, either in the core or on the side chains, which are crucial for their activity. The rationale behind the design of these new inhibitors, as well as the key synthetic steps involved in their preparation, is also described.

1. Introduction

HIV-1 is the virus responsible for acquired immunodeficiency syndrome (AIDS), which is characterized by an attack and disruption of CD4+T cells, with the consequence of significant vulnerability in humans against different types of infection. This lentivirus integrates irreversibly into the host genome and has a long incubation period, being able to infect non-dividing cells [1,2].

Among the many strategies used to combat the disease, antiretroviral therapy (ART) is considered the most effective current treatment, and it is characterized by a combination of HIV protease inhibitors, reverse transcriptase inhibitors, and/or an integrase inhibitor. Recently, capsid inhibitors, such as Lenacapavir, have shown potential for a long-acting administration regimen [3].

HIV protease plays a crucial role in the viral lifecycle, being essential for the generation of mature, infectious viral particles [4]. This aspartyl protease is involved in the cleavage of peptide bonds in the gag and gag-pol polyproteins, and its inhibitors have been designed to bind to the active site with high affinity. All nine Food and Drug Administration (FDA)-approved inhibitors, except for tipranavir, are competitive peptidomimetic inhibitors, mimicking the natural substrate of the viral protease. Through the incorporation of the hydroxy ethylene core, the cleavage of the inhibitor is prevented. Due to their structural similarities, cross-resistance amongst inhibitors may occur, along with common side effects [5].

Although no new protease inhibitors have been approved or used in a clinical trial since 2007 [6], many small molecules have been described in the last fifteen years, with the aim of targeting the protein backbone. The goal has been to maximize interactions with enzyme backbone atoms [7,8]. Both extensive hydrogen bonding and hydrophobic interactions with enzyme subsites can limit drug resistance, as the geometry of the catalytic site must be conserved to maintain functionality [9,10,11]. In particular, the introduction of heterocyclic fragments has improved the activity of newly designed molecules, resulting in many potent inhibitors, as illustrated by FDA-approved darunavir and related compounds [12]. The introduction of cyclic scaffolds in the design of new chemical entities reduces flexibility and affords more rigid inhibitors. In particular, a cyclic scaffold can be envisaged as having either key interactions with catalytic aspartic acids or with residues belonging to the flap region of the active site [13]. The deep interest in cyclic chemotypes has been motivated by the different types of interactions formed with the enzyme while maintaining a key structural resemblance to a peptide substrate, to improve this profile for multidrug-resistant strains.

This review focuses on the main advances in the synthesis of new heterocyclic HIV protease inhibitors during the last 15 years and in the synthesis of key heterocyclic fragments [14]. The rationale behind the design of such new compounds is described, pointing out the crucial importance of the heterocyclic moiety for their activity. The different approaches are classified according to the type of heterocycle (aromatic or non-aromatic) and the heteroatoms involved (N, O, S, or polyheteroatomic). In the case of different heterocyclic moieties being present, the classification is made according to the group that is most critical for activity. This review aims to highlight ongoing research in the field of heterocyclic synthesis applied to the discovery of new HIV inhibitors.

2. Non-Aromatic Heterocycles

2.1. Non-Aromatic O-Heterocycles

Fused bicyclic bis-tetrahydrofuran (bis-THF) is critical for darunavir’s durable drug resistance profile. Due to its challenging double-ring structure with three chiral centers, many different synthetic approaches have been developed according to the source of the chiral carbons in bis-THF alcohol 2 (Scheme 1) [15]. In this respect, enzyme-catalyzed resolution has been exploited by Khmelnitsky [16], who described the simple and efficient kinetic resolution of the racemic alcohol 2 using immobilized lipase to afford the desired optically pure (R)-bis-tetrahydrofuran (bis-THF) alcohol 3. The reaction solvent, acyl donor, and immobilized biocatalyst proved to be critical factors.

The organocatalytic condensation of 1,2-dihydrofuran 5 (Scheme 2) with glycolaldehyde catalyzed by Schreiner’s thiourea, coupled with enzymatic (Lipase PS) resolution, was described as a one-pot procedure by Itoh for the preparation of enantiopure 3 and applied to the environmentally friendly synthesis of darunavir [17].

A one-pot procedure using 1,2-dihydrofuran 5 and Cbz-protected glycol aldehyde as the starting materials was developed by Opatz through [2+2]-photocycloaddition between both reactants, followed by hydrogenation and lipase-catalyzed kinetic resolution, affording the target compound with a high yield and up to 99% ee (Scheme 3) [18].

In 2020, Ghosh used the kinetic resolution of bis-THF alcohol by lipase (PS-30), in a late stage of the study as an alternative stereoselective approach to cis and trans 2,3-disubstituted tetrahydrofuran derivatives [19]. More recently, Hyster proposed a different chemoenzymatic approach to the use of bis-THF alcohol [20]. β-Ketolactone 8 was easily prepared and subjected to enantio- and diastereoselective dynamic kinetic resolution using suitable ketoreductase (KRED) and glutamate dehydrogenase (GDH) in phosphate buffer (KPi) from metagenomic mining (Scheme 4). Subsequent lactone reduction with diisobutylaluminum hydride and phase-transfer cyclization afforded the desired fragment 3 in an acceptable yield.

Asymmetric catalysis was also exploited in the last period. Sudalai [21] proposed an efficient synthesis of (S)-tetrahydrofuran-3-ol 13, which is a critical fragment in amprenavir and fosamprenavir.

Olefinic tosylate 10 was dihydroxylated under Sharpless asymmetric dihydroxylation conditions, furnishing, in a single step, the desired alcohol in good yield and moderate enantiomeric purity (Scheme 5). Alternatively, epoxidation with mCPBA yielded the racemic epoxide 12, which was subjected to co-catalyzed hydrolytic kinetic resolution, affording the alcohol along with the chiral epoxide in a lower chemical yield but with higher enantiomeric excess.

More and Ramana [22] described a Ru-catalyzed enantio- and diastereoselective dynamic kinetic resolution of benzyloxy/benzoyloxy-α-acyl-γ-butyrolactones via transfer hydrogenation, achieved through the in situ prepared (R,R)-Ru-FsDPEN catalyst. Starting from benzyloxy butyrolactone 14, (3R,3aS,6aR)-hexahydrofuro[2,3–b]furan-3-ol 3 was prepared in good yield, with high de and ee through elaboration of the key intermediate (Scheme 6).

Pd-catalyzed asymmetric hydroalkoxylation of ene-alkoxyallene, followed by ring-closing metathesis (RCM), was successfully employed by Rhee [23] for the preparation of hexahydrofurofuran-3-ol of type E, and was generalized to provide access to pyranofuranol and furopyranol derivatives in good yields and ee (Scheme 7). For the synthesis of 3, readily available allene 17 was coupled with commercial 2-bromo allylic alcohol 18 in the presence of Pd₂(dba)₃ as the palladium source and suitable ligand, affording the chiral adduct 19 in high yield and enantiomeric excess. Subsequent RCM provided the cyclic acetal 20, which underwent a radical, mediated 5-exo cyclization to intermediate 21, promoted by Bu₃SnH in the presence of Et₃B. Ozonolysis of this compound, followed by final reduction with NaBH₄, furnished the target alcohol 3 in high yield with complete diastereoselectivity and no erosion of enantiopurity.

Organocatalytic methods were also successfully used. Ikemoto described an efficient synthesis of bis-THF alcohol 3 and its carbonyloxy-pyrrolidine-2,5-dione derivative taking advantage of diphenylprolinol-catalyzed enantio- and diastereoselective cross aldol reaction of polymeric ethyl glyoxylate 23 with the aldehyde 22 as the key step (Scheme 8) [24]. Optimization of the reaction involved stirring polymeric ethyl glyoxylate in toluene solution with water prior to aldol reaction for an appropriate period. This pre-treatment was highly effective in accelerating the reaction when using 3 mol % of catalyst A. It also ensured excellent reproducibility, even when polymeric ethyl glyoxylate from different manufacturers was utilized. The aldehyde of intermediate 24 was then protected as dimethyl acetal and the ester was reduced, yielding diol 26 in an overall yield of 85%. An acetal exchange reaction catalyzed by H₂SO₄ followed by hydrogenation with Pd/C catalyst afforded 3 in quantitative yield over two steps. Shortly thereafter, the same group described a series of amino perfluoroalkanesulfonamide derivatives of diarylprolinols, which proved to be effective organocatalysts [25].

Traditional chiral-substrate approaches have also been employed over the past fifteen years. A spirocyclic dioxolane derivative from D-glyceraldehyde 27 was used by Kulkarni as starting chiral substrate for the preparation of all four isomers of bis-tetrahydrofuran alcohol in good overall yield (38%) of the active isomer 3 (Scheme 9) [26].

(S)-Glyceraldehyde was used as a chiral source by Ghosh in the convergent synthesis of various substituted bis-THF derivatives, to find new HIV-protease inhibitors with enhanced binding capacity [27]. The synthesis of the bis-THF ligand began with known compound 30 (Scheme 10) prepared in multigram quantities by Wittig olefination of (S)-glyceraldehyde acetonide with (ethoxycarbonylmethylene) triphenylphosphorane. DIBAL-H reduction yielded the corresponding alcohol 31 in nearly quantitative yield. Subsequently, O-alkylation and subsequent stereoselective [3,4]-sigmatropic rearrangement afforded the alcohol with three chiral centers 33, which was transformed alternatively to bis-THF alcohol 3 or to different functionalized derivatives. The synthesis of alkyl substituted bis-THF ligands 36 required a Mitsunobu reaction on alcohol 33 with PPh₃/DIAD and p-nitrobenzoic acid, to give the corresponding ester 34. This was first hydrolyzed, then alkylation with BnBr or MeI was performed, alternatively. Final reduction, ozonolysis, and cyclization sequence furnished the ligands 36 in good overall yield. Among the inhibitors bearing new bis-THF fragments, compound 37 resulted in being the most potent.

More recently, Ghosh described a new approach to bis-THF alcohol 3, starting from commercially available sugar derivatives [28]. As an example, commercial 1,2-O-isopropylidene-α-D-xylofuranose 38 was converted to α,β-unsaturated ester 39, by sequential selective protection of primary alcohol, Swern oxidation, and Wittig olefination (Scheme 11). Ester 39 was then submitted to highly stereoselective substrate-controlled catalytic hydrogenation, which represented one of the key steps.

Saturated ester 40 was converted to γ-lactone derivative 41 by exposure to BF₃·OEt₂ followed by Et₃SiH. Hydrolysis of the benzoate furnished the corresponding bicyclic alcohol, which was converted to methyl acetal 42 by a three-step sequence involving Dess–Martin oxidation of the primary alcohol to the corresponding aldehyde, Baeyer–Villiger oxidation promoted by m-CPBA, and exposure of the resulting formate to 6% HCl in MeOH. Acetal 42 was obtained in 35% yield over four steps. Final reduction of the lactone and acidification yielded the target bis-THF alcohol.

Regarding sugars as chiral pool materials, Sridhar developed a stereoselective synthesis of carbohydrate-derived perhydrofuro[2,3-b]furan derivatives, starting from sugar-derived allyl vinyl ethers [29]. In particular, bis-THF alcohol 3 was prepared using 3,4-di-O-acetyl-D-arabinal 43 as the chiral starting material (Scheme 12). It was first subjected to Ferrier rearrangement with 2-(phenylselenyl)ethanol, which provided 2,3-unsaturated glycoside 44. After oxidation to selenone 45, a base-mediated thermal fragmentation furnished the allyl vinyl ether 46, which underwent a Claisen [4]-sigmatropic rearrangement, affording the expected 3-C branched derivative 47. Reduction of the aldehyde, followed by ozonolysis of the olefin and acid-mediated acetalization, provided perhydrofuro[2,3-b]furan 50 in good yield. The synthesis of bis-THF alcohol 3 required deprotection of the ester, Mitsunobu inversion with p-nitrobenzoic acid in PPh₃/DIAD system, and final hydrolysis.

The potassium isocitrate, obtained from high-yielding fermentation fed by sunflower oil, was the starting material for the preparation of bis-THF alcohol 3 in the work of Yue (Scheme 13) [30]. After several steps, it was transformed into a tertiary amide, which was reduced along with the ester functionalities to a transient aminal-triol. This latter was converted in situ to the desired bis-THF alcohol. Each step was optimized, resulting in the key alcohol (in the form of activated carbonate) with an overall yield of 43%. Considering the low cost of the starting material and the efficiency of the overall synthesis, the method appeared to be effective in reducing the cost of darunavir.

With the aim of creating a new class of inhibitors with improved pharmacological and drug-resistance profiles, Ghosh described the synthesis of a 6–5–5 ring-fused crown-like tetrahydropyranofurans and their incorporation into new potent inhibitors [31,32]. The synthesis started with an asymmetric Diels–Alder reaction between 57 and cyclopentadiene, using a chiral oxazaborolidinium cation A, as a key step (Scheme 14). The cycloadduct was oxidized with H₂O₂ to afford alcohol 59 in 80% yield over two steps, with 98% ee. After protection as the TBS ether, the intermediate was converted to bicyclic acetal 60 in three-steps, involving reduction with LiAlH₄, one-pot oxidative cleavage of the olefin, and reduction of the resulting aldehyde with DIBAL-H. The lactol 60 (a 2:1 mixture) was treated with trifluoroacetic acid (TFA) to produce the bridged tricyclic derivative 61, and the TBS group was removed with tetrabutylammonium fluoride (TBAF), yielding the desired exo-alcohol. This compound was alternatively epimerized by Dess–Martin oxidation followed by reduction of the resulting ketone with NaBH₄. The inhibitor 64 displayed superior antiviral activity and drug-resistance profiles compared to darunavir.

Shortly afterwards, this fragment was incorporated into new arylboronic derivatives, and they were tested against highly drug-resistant HIV-1 variants [33].

Alternatively, the Diels–Alder reaction of cyclopentadiene with chiral 3-(acyloxy)acryloyloxazolidinone derivatives was later used for the synthesis of 6-5-5 fused crown-like THF ligands (Scheme 15). In particular, 3-(4-methoxybenzoyl)acryloyl oxazolidinone derivative 65 and cyclopentadiene afforded the endo-diastereoselective derivative 66 in 98% yield. Reaction of the cycloadduct in MeOH provided the corresponding methyl ester 67 in excellent yield, which was a key intermediate in the synthesis of the bicyclic alcohol 63 [34].

An optically active (1R, 3aS, 5R, 6S, 7aR)-octahydro-1,6-epoxy-isobenzo-furan-5-ol derivative was recently described as a high affinity ligand for HIV-1 protease inhibitors [35]. The stereoselective synthesis was based on enantioselective enzymatic desymmetrization of meso-1,2(dihydroxy methyl)cyclohex-4-ene 69 using porcine pancreatic lipase (PPL) to produce optically active alcohol 70 (Scheme 16) [28]. It was then converted to a functionalized cyclohexene derivative 71, which underwent Sharpless asymmetric dihydroxylation to yield a 1:1 mixture of diastereomeric diols 72 and 73. The diols were separated and only diastereomer 73 was utilized for the synthesis of ligand alcohol 74. In general, inhibitors bearing such ligands showed very potent activity. Moreover, compound 75, with a difluorophenylmethyl P1 ligand and an aminobenzothiazole as the P2′ ligand, maintained high potency against a panel of highly multidrug-resistant HIV-1 variants.

More recently, a modified procedure was described [36]. Optically active alcohol 70 was converted to methyl acetal 76 (in 75% yield over 3 steps) as the major isomer in a sequence involving, (1) Swern oxidation of alcohol 70 to the aldehyde, (2) treatment of the resulting aldehyde with CH(OMe)₃ in the presence of tetrabutylammonium tribromide (TBABr₃), followed by treatment with 1 N NaOH, and (3) reaction of the resulting methyl acetal with a catalytic amount of CSA (Scheme 17).

Stereoselective syn-1,2-diol functionality was introduced on methyl acetal 76 by substrate-controlled dihydroxylation, with K₂OsO₄/K₃Fe(CN)₆ system, affording syn diol 77, as 95/5 diastereomeric mixture. Swern oxidation provided 1,2-diketone derivative 78, which was reduced with L-Selectride to give the corresponding inverted syn-1,2-diol, stereoselectively. Acid-catalyzed cyclization then furnished tricyclic ligand alcohol 74 with high enantiomeric purity.

The synthesis of enantiomerically pure (3aS,4S,7aR)-hexahydro-2H-furo[2,3-b]pyran-4-ol 83 was originally described by Ghosh [37] starting from known enantiomerically pure lactone 80 (Scheme 18). Subsequent steps involved reduction, selective protection/deprotection, ozonolysis, and reduction of the hemiacetal moiety. Final removal of the silyl group with TBAF furnished the desired ligand.

A more recent alternative route has been proposed, starting from inexpensive materials and taking advantage of highly enantioselective enzymatic desymmetrization of a meso-diacetate, already developed by the same group [38]. The enzymatic desymmetrization of 84 was optimized by using aqueous 1 N NaHCO₃ for neutralization of acetic acid formed during the reaction (Scheme 19). Swern oxidation followed by protection of the aldehyde furnished the corresponding acetal ent-71. Oxidative cleavage of the olefin by ozonolysis in the presence of pyridine as the organocatalyst, followed by reduction of the resulting dialdehyde with sodium borohydride afforded diol 85. Treatment of the diol with a catalytic amount of CSA furnished bicyclic acetal 86 in 56% yield over 3 steps. After hydrolysis of acetate, the corresponding alcohol was treated with o-nitrophenylselenonitrile and n-tributylphosphine to give the corresponding selenide derivative. Oxidation of the resulting selenide with m-CPBA resulted in elimination of the corresponding selenoxide to afford olefin 87 in 71% yield over 3 steps. Ozonolysis of exo-olefin followed by reduction of the corresponding ketone yielded alcohol 83 in good yield.

Various C3-functionalized cyclopentanyltetrahydrofurans (Cp-THF) were prepared and combined with hydroxyethylsulfonamide isosteres to obtain new inhibitors with high antiviral activity, including against a panel of multidrug-resistant HIV-1 variants [39]. As an example, 3-(R)-acetoxy and 3-(R)-methoxy ligands 92 and 93 were prepared from the known alcohol 89, which was first alkylated to provide methyl ester 90 (Scheme 20). DIBAL-H reduction followed by radical cyclization provided 3-(R)-hydroxy derivative 91 in 10:1 diastereomeric ratio. The major isomer was transformed into the desired ligand 92 by acetylation, followed by removal of the silyl group, in 73% yield over two steps. Alternatively, the methoxy derivative 93 was prepared by methylation followed by removal of silyl group, in 71% yield.

Novel oxatricyclic ligands were designed and synthesized by Ghosh to enhance interactions with the protease backbone [9]. In particular, the inhibitor 101, bearing the tris-THF syn-anti-syn configuration, showed very high activity against a variety of multidrug resistant HIV-1 variants and blocked HIV-protease dimerization about 10-fold better than darunavir (Scheme 21) [40]. Starting from the known acetate 94, the bicyclic enol-ether 95 was easily obtained. Subsequently, after epoxidation with dimethyldioxirane (DMDO) and regio- and stereoselective oxiranyl ring opening with sodium methoxide, the desired endo-alcohol 96 was obtained by Dess–Martin oxidation followed by L-selectride reduction of the resulting ketone. Acylation of 96, followed by treatment with propargyl alcohol, afforded acetals 97 and 98 in a 1:4 ratio. After removal of the acetate, the major diastereomeric alcohol deriving from 97 was converted into the corresponding tricyclic olefin 99. Ozonolysis followed by reduction furnished syn-anti-syn type oxatricyclic tris-THF 100 as a single isomer.

Gem-difluoro-bis-THF ligands were prepared and converted into suitable inhibitors, which exhibited very high activity and better lipophilicity profiles than darunavir, along with significantly improved blood–brain barrier permeability in an in vitro model [41]. The synthesis began with DIBAL-H reduction of optically active methyl ester 102 (Scheme 22), producing dibenzyl-L-glyceraldehyde. This intermediate underwent the Horner–Emmons reaction with sodium hydride and triethyl phosphonoacetate, affording the corresponding α,β-unsaturated ester in 88% yield over two steps. Reduction with DIBAL-H provided the corresponding allylic alcohol, which was treated with chlorodifluoroacetic acid in chloroform at reflux, yielding difluoroacetate derivative 103 in 90% yield over two steps. Subsequently, it underwent a Reformatskii–Claisen reaction by treatment with trimethylsilyl chloride and activated zinc dust. The subsequent acid-catalyzed esterification furnished a 2:1 mixture of diastereomers in 80% yield over two steps. The mixture could be separated after its conversion into the Weinreb amides, by treatment with HN(Me)OMe·HCl and n-BuLi. These amide diastereomers were separated by silica gel chromatography, providing the syn diastereomer 104 as the major product and the anti diastereomer 105 as the minor product, in a 2:1 ratio and with an overall yield of 80%.

Reduction of the Weinreb amide 105 was carried out with lithium aluminum hydride, and the resulting crude aldehyde was reduced with sodium borohydride in a one-pot operation to provide difluoro alcohol 106 in near quantitative yield. Ozonolysis, followed by reductive cleavage with PPh₃, provided cyclic acetal 107 upon cyclization. Final catalytic hydrogenation and treatment with CSA of the triol afforded difluoro-bis-THF 108.

To maximize the ligand-binding site interactions in the protease active site, Ghosh described cyclohexyl-derived ligands within a 6-5-5 fused ring system [42].

Starting from 1,3-diketone 109, reaction with dihydrofuran 110 in the presence of Mn(OAc)₃·2H₂O afforded the corresponding tricyclic derivative 111 (Scheme 23). Enone 111 was first hydrogenated, and the resulting ketone was reduced with NaBH₄, to give racemic endo alcohol 112. This racemic alcohol was subjected to enzymatic resolution using lipase PS-30, which provided the optically active acetate derivative 114 (45%yield) and alcohol 113 (45% yield). Acetate 114 was converted back to the alcohol 113 in 89% yield by transesterificaton using NaOMe in MeOH. A series of ligands was prepared and incorporated into new inhibitors. Notably, compound 115 exhibited very impressive enzyme and antiviral potency (Ki = 10 pM, antiviral IC₅₀ = 1.9 nM).

Novel isosorbide scaffolds were proposed by Liu as P2 ligands of HIV-1 protease inhibitors bearing the hydroxyethylamine core 116 (Scheme 24) [43]. They showed very high inhibition activity with IC50 in the nanomolar or picomolar range. From a preparative perspective, they leveraged the commercial availability of starting isosorbide mononitrate 117, which was easily elaborated and activated as a mixed carbonate. Specifically, it was transformed into mixed carbonate 118 with p-NO₂PhOCOCl in good yield or converted into TBS-protected alcohol 119 via TBS ether formation, followed by cleavage of nitrate under reductive conditions. The alcohol was then methylated; the TBS group was cleaved, and final carbonate formation afforded the key intermediate 120.

Chiral 4,4-dimethyltetrahydrofuran-3-ol was prepared by Srinivasa Reddy and used for the synthesis of new amprenavir analogs [44]. Starting from pantolactone 122 (Scheme 25), which is available in both enantiomeric forms, lactol 123 was obtained by protection and reduction. Subsequent reduction with BF₃Et₂O and triethylsilane, followed by final deprotection with TBAF, afforded the desired THF-alcohol.

Different O-heterocycles were also proposed by Yajima as HIV-1 protease inhibitors [45]. Fornicin A, a meroterpenoid with a γ-butylolactone moiety in its side chains showed weak anti-HIV-1 protease activity without cytotoxicity against E-PR293 cells. The synthesis involved the preparation of optically active alcohol 130 from the starting material 2,5-dihydroxybenzaldehyde 126 (Scheme 26), using Sharpless asymmetric dihydroxylation on the suitable benzothiazolsulfide 129 as the key step. Oxidation with mCPBA and a final Smiles rearrangement of β-hydroxysulfone yielded the desired alcohol 130, which was coupled with the suitable acid segment prepared from the known aldehyde 131. Ring-closing metathesis using Grubbs 1st catalyst and a final deprotection step furnished Fornicin A.

Flexible macrocycles between the P1′-side chain and a suitable P2′-ligand were proposed by Ghosh to broaden the activity of the inhibitor by achieving a better fit in the S2 hydrophobic pocket, which increases in size upon certain mutations [46]. New inhibitors bearing 16- to 19-membered macrocyclic rings, connecting a nelfinavir-like P2 ligand and a tyrosine side chain containing a hydroxyethylamine sulfonamide isostere, were synthesized and found to be more potent than their corresponding acyclic counterparts. In particular, compound 134 showed the best enzyme inhibitory and antiviral activity (K_i = 0.2 nM, IC₅₀ = 0.21 mM) (Scheme 27). The synthesis of the desired tyrosine-derived hydroxyethylamine sulfonamide isostere 138 began with butadiene monoxide 135, which was transformed into the corresponding allylic alcohol. This was subjected to Sharpless asymmetric epoxidation, and the resulting epoxide 136 was regioselectively opened by TMSN₃/Ti(O-iPr)₄ system to afford the corresponding azido diol, which was subsequently converted to epoxide 137. After introducing the sulfonamide moiety, a Boc-protected amine and a free phenol were formed in one pot process via catalytic hydrogenation in the presence of Boc₂O. Finally, the introduction of an allyl chain furnished the desired intermediate 138.

The P2 fragment was synthesized as 3-hydroxy-2-alkenylbenzoic acid from maleic anhydride 140 via formation of phosphorane 141, followed by Wittig reaction with known aldehyde 142. The resulting dienoic acid was then alkylated and subjected to 1,6-electrocyclization by exposure to TFAA/Et₃N, followed by reduction with NaBH₄ (Scheme 28). Final saponification yielded the desired benzoic acid 144. After amine deprotection and coupling with the benzoic acid, the resulting acyclic diene was subjected to ring-closing metathesis (RCM) with Grubbs’ 1st generation catalyst, furnishing the desired macrocycles.

2.2. Non-Aromatic N-Heterocycles

Among non-aromatic N-heterocycles, the structure of 1,3-diazacycloalkan-2-ones is of interest due to its presence as the core of HIV-protease inhibitor DMP 450 (Figure 1).

Frain described the synthesis of novel tetrahydropyrimidinones, starting from tetraols, that are easily prepared from suitable alditols [47]. Tetraol 146, from D-(+)-arabitol, was first protected in primary hydroxyl groups as TBS ethers (Scheme 29). The introduction of the amines was performed by mesylate activation of alcohols, followed by displacement with azide and reduction by hydrogenation. Extended hydrogenation periods led to the isolation of the mono TBS-protected diaminodialcohol 148. Subsequent reaction with carbonyl diimidazole in the presence of base provided the crude tetrahydropyrimidinone, which was treated with acid to remove the TBS group. Final perbenzylation furnished the functionalized tetrahydropyrimidinone 150, which exhibited modest HIV-protease inhibition activity.

Unsymmetrical 1,3-diazacycloalkan-2-ones were prepared by Jamir [48] from thiocarbamate salts using sodium percarbonate as an oxidant. These compounds were evaluated as HIV-1 protease inhibitors via in silico approach. Secondary amines 152, such as morpholine, piperidine, 4-hydroxy piperidine, diethylamine, and di-isopropylamine, were reacted with phenyl isothiocyanate and other substituted phenyl isothiocyanates 151 to afford the corresponding unsymmetrical 1,3-diazacycloalkan-2-one in high yields (Scheme 30). Computational assessment of IC₅₀ values using known references satisfactorily confirmed the inhibitory activity of the selected compounds against HIV-1 protease.

In the work of Jadhav [49], bis-allylidene-4-piperidones analogues were prepared by Claisen–Schmidt condensation between 2,6-substituted piperidin-4-ones and cinnamaldehyde in a basic medium (Scheme 31). In particular, 3,5-bis (3-phenyl-allylidene)-2,6-diphenyl-piperidine-4-one 155 was obtained from the parent compound 154. The N-tosyl derivative 156 was prepared by reacting 155 with tosyl chloride. Compound 156 showed moderate HIV-1 protease inhibitor activity compared to the standard, pepstatin-A.

Indolin-2-one moieties were proposed by Eissenstat to interact with the S20 subsite of the HIV protease binding pocket [50]. Several of these inhibitors were synthesized and exhibited sub-nanomolar Ki values and antiviral IC50s in the low nM range against wild-type (WT) HIV and a panel of multi-drug resistant (MDR) strains. The darunavir sulfonamide was replaced by an indolin-2-one moiety, reacting bis-THF amino analog 157 with indolin-2-one sulfonyl chloride 158, affording compound 159. Subsequent heating in the presence of DMF dimethylacetal or related acetals led to the formation of indolin-2-one 160 (Scheme 32). The amine portion of the enamines could be readily varied through an exchange reaction of dimethamino analog 161 with excess amine.

1,4-benzodiazepine derivatives have been recognized as mimicking the β-hairpin flap of HIV-1 protease, thereby interfering with the flap–flap protein–protein interaction which controls access to the active site. Recently, the Konvalinka group screened a variety of chemical structures for inhibition of HIV replication [51]. This led to the identification of several new compounds that inhibited HIV PR in the low micromolar range. In particular, compound 166 resulted in being the most potent and selective for HIV PR. The monomer was prepared by reacting o-phenylenediamine 162 with dimedone in the presence of catalytic amount of TFA, affording the corresponding enamine 164 which was then reacted with benzaldehyde to obtain monomer 165. Treatment with oxalyl chloride at low temperature, furnished the desired dimer 166 in modest yield as a diastereoisomeric mixture (Scheme 33), whose racemate (11R, 11′R + 11S, 11′S) showed inhibition activity in the nanomolar range (IC₅₀ = 30 nM).

New bioactive compounds containing 1,4-benzodiazepine scaffolds were proposed by Fletcher, with the most potent compound, 167, inhibiting the protease with a modest Ki of 11 µM (Scheme 34) [52]. It was prepared starting with regioselective nitration of benzoic acid 168, followed by benzylic oxidation, which afforded ketone 169. After esterification, catalytic hydrogenation yielded aniline 170. Amide 171 was prepared coupling aniline 170 with Fmoc-protected pre-activated D-Leu, and then subjected to base-mediated cyclization, furnishing the key benzodiazepine nucleus. After deprotection and N-alkylation, the final compound 167 was obtained. Different compounds were obtained by deprotection and/or alkylation. Methyl imine could be replaced by a phenyl imine, resulting in the formation of different compounds.

The tetrahydropyridine ring system is a widely distributed structural framework involved in numerous pharmaceuticals and natural products. Mohammadi [53] recently described an efficient, one-pot, catalyst-free, four-components procedure for the synthesis of novel 10β-hydroxy-4-nitro-5-phenyl-2,3,5,5a-tetrahydro-1H-imidazo[1,2-a]indeno[2,1-e]pyridin-6(10bH)-one derivatives 177 from the corresponding diamine 173, nitro ketene dithioacetal 174, aryl aldehydes 175, and 1,3-indandione 176 (Scheme 35). The overall transformation consisted of a Knoevenagel condensation, Michael addition, tautomerism, and cyclisation sequence. All newly synthesized compounds were screened for molecular docking studies. Some of them showed minimum binding energy and good affinity toward the active pocket of HIV protease enzyme compared to Saquinavir.

A flexible piperidine moiety was introduced into a novel class of HIV-1 protease inhibitors as the P2 ligand by Wang Y., Cen, and Wang J [54]. In particular, inhibitor 181, which features (R)-piperidine-3-carboxamide as the P2 ligand and 4-methoxybenzenesulfonamide as the P2′ ligand, showed more than a 6-fold enhancement of activity compared to darunavir (IC₅₀ 0.13 ± 0.01 nM). Furthermore, there was no significant change in potency against darunavir-resistant mutations and HIV-1_NL4-3 variant. For the synthesis, easily prepared (R)-piperidine-3-carboxylic acid 178 was reacted with (Boc)₂O to obtain the corresponding Boc-amino derivative 179 in excellent yield. Coupling of the acid with a suitable amine, under the catalysis of EDCI, HOBt, and DMAP, furnished the amide 181 (Scheme 36). Removal of the Boc group provided the target compound.

Macrocyclic peptides have emerged as a new class of drug discovery modalities because they are considered more likely to acquire strong interactions with target proteins, owing to their restricted molecular motion and rigidity compared to linear peptides. Recently, Mikamiyama [55] discovered a novel HIV-1 protease inhibitor, 182 (Figure 2), with potent antiviral activity (EC₅₀ 37 nM) and oral bioavailability, using a structure-based drug design approach via X-ray crystal structure analysis. This approach started from hit macrocyclic peptides identified by mRNA display against HIV-1 protease. In particular, the improvement of the proteolytic stability of macrocyclic peptides by introducing a methyl group at the α-position of certain amino acids, such as proline, is crucial for exhibiting strong antiviral activity.

The five membered hydantoin scaffold was already recognized as a key fragment in HIV protease inhibitors. Recently, Zhang reported a new general iridium-catalyzed asymmetric hydrogenation of hydantoin and thiazolidinedione-derived exocyclic alkenes using BiphPHOX as a ligand [56]. The transformation, which showed good functional group tolerance, high yields, and enantioselectivities in the hydrogenated products, was applied to the synthesis of 189, a key intermediate in the preparation of an HIV protease inhibitor 190 (Scheme 37) [57]. A gram-scale reaction was carried out starting from alkene 186, yielding the hydrogenated product 187 with 92% ee and 97% yield. This product was then reduced with LiAlH₄, providing chiral 1,3-diazacycloalkan-2-one 188. The reduced product 188 was transformed into the key intermediate 189 + dechlorination in the presence of Pd/C.

A densely functionalized iminohydantoin, bearing a quaternary stereocenter, was recently described and developed as an HIV protease inhibitor [58,59]. Ischay and Hoang described process development efforts that enabled the first scale-up of compound 196 (Scheme 38). Several challenges were addressed, including reaction optimization, purification, stabilization of the intermediates, and deprotection procedures. For the reaction of iminohydantoin 195 formation, the addition order was evaluated. It was found that the presence of the amine coupling partner, in the form of BSA (benzensulphonic acid) salt, was required during the loading of the EDC·HCl reagent. Reagent stoichiometry was then investigated, showing that the loading of EDC·HCl could be reduced to 1.5 equivalents and 6·BSA to 1.1 equivalents without negatively affecting reaction performance. The use of 2–3.5 equiv. i-Pr₂NEt proved optimal for scale-up to 687 g. A MIBK (methyl isobutylketone) solvate was identified as a crystalline form that provided an additional isolable intermediate, serving as a purity control point prior to Cbz-group removal and API isolation. The Cbz protecting group was removed from iminohydantoin 195 using HCl in acetic acid, that mitigated hydrolysis of the iminohydantoin. After work-up, crystallization afforded 196 1-succinate (Scheme 38).

Wang and coll. recently reported a new strategy for the synthesis of an iminohydantoin derived from chiral quaternary α-aryl amino acids [60]. Such fragments are present in certain HIV protease inhibitors [58]. They took advantage of newly developed chiral Karady–Beckwith dehydroalanines, which were prepared for use in a photoredox-mediated highly stereoselective Giese-type reaction with carboxylic acids and tertiary amines. A final stereoselective Clayden rearrangement afforded chiral quaternary α-aryl amino acid derivatives. A new chiral dehydroalanine bearing an N-methyl, N-phenyl-urea moiety 202 was synthesized starting from S-benzyl-L-cysteine 197, which was first transformed into amino amide 198 (Scheme 39). Cyclization with pivalaldehyde was then performed, and the free NH group was converted into the corresponding urea moiety, affording intermediate 201. An exocyclic double bond was obtained by oxidation of sulfide and subsequent elimination of the sulfoxide. Compound 202 underwent a Giese-type reaction with pivalic acid under the irradiation with a blue light-emitting diode (LED), using 4CzIPN as the photocatalyst (PC) and Cs₂CO₃ as the base in DMF, affording the desired adduct 203 with a high diastereoisomeric ratio (>20:1). This intermediate smoothly underwent a Clayden rearrangement, giving rise to the quaternary amino acid derivative 204 in high yield and diastereoselectivity. Final deprotection afforded free amino acid 205, suitable for incorporation into the HIV protease inhibitor 206.

2.3. Non-Aromatic Heterocycles with Multiple Heteroatoms

Starting from KNI-764-derived isostere prepared by Mimoto and coworkers [61] Ghosh [62] proposed a new series of protease inhibitors in which cyclic ethers were incorporated as P2-ligands. As an example, GRL-0355 (Scheme 40) displayed impressive antiviral properties, with improved potency and efficacy against multidrug-resistant HIV-1 variants. The key intermediate 207, prepared through two alternative pathways, was protected as the corresponding acetonide, after which the acetate group was hydrolyzed to afford the corresponding alcohol. This intermediate was subjected to oxidation using ruthenium chloride hydrate and sodium periodate, resulting in the formation of the target carboxylic acid 208. Compound 208 was then transformed into amide 214 by activation to the corresponding mixed anhydride, followed by reaction with heterocyclic amine 213 [63]. The latter was derived from (R)-BocDMTA 212, which was synthesized with 99.4% ee via enantioselective hydrolysis of methyl (±)-5,5-dimethyl-1,3-thiazolidine-4-carboxylate 210 by a Klebsiella oxytoca hydrolase.

The work of Guarna and his group on small peptidomimetics bearing 6,8-dioxa-3-azabicyclo[3.2.1]-octane scaffold resulted in the identification of HIV protease inhibitors with IC₅₀ values in the sub-micromolar range [64]. They proposed the bicyclic acetal portion as a potential transition state analog in interactions with the enzyme active site. In particular, the best results were obtained with the glycine-derived scaffold BTG(O)-A (bicyclices derived from tartaric acid and glycine), which bears an α-amino alcohol fragment at position 7, and with phenylalanine-derived scaffold BTF(O)-B (bicyclices derived from tartaric acid and phenylalanine), which bears small aliphatic chains at the same position.

The synthesis began with benzylamine 215, which was treated with bromoacetaldehyde dimethylacetal. Amine 216 was then acylated with (2R,3R)-di-O-acetyltartaric anhydride, yielding amide 217. This was subsequently transformed into cyclic acetal 218 by treatment with MeOH/HCl. The 6,8-dioxa-3-azabicyclo[3.2.1]-octane 219 was obtained by treating 218 under reflux in toluene in the presence of H₂SO₄ over silica gel. Amide–ester exchange reactions allowed the preparation of the desired amides BTG(O)-A (Scheme 41).

A novel HIV protease inhibitor was designed by Bungard using a morpholine core as the aspartate binding group [65]. Analysis of the crystal structure of the initial lead bound to HIV protease enabled optimization of enzyme potency and antiviral activity. This resulted in a series of potent orally bioavailable inhibitors, among which MK-8718 (Scheme 42) was identified as a compound with a favorable overall profile. The preparation of a suitable morpholine intermediate started with the reaction of amino alcohol 220, derived from D-serine, with (R)-epichlorohydrin to give morpholine 221. Swern oxidation afforded aldehyde 222. Subsequently, Seyferth–Gilbert homologation yielded alkyne 223. Sonagashira coupling, followed by alkyne reduction, gave aniline 224. Cbz protection, followed by TBS deprotection, afforded the corresponding alcohol. Carbamate formation, followed by Cbz-deprotection, produced the desired aniline 225. Final elaboration by coupling with a suitable diaryl-α-aminopropanoic acid furnished MK-8718.

Morpholine was also proposed by Zhou, Cen, and Wang [66] as a P2 ligand of HIV-1 protease inhibitors with a flexible heterocyclic structure, which can fit into the minimally distorted active site of darunavir-resistant HIV-1 variants. Starting from easily available N-substituted morpholines, a new series of inhibitors was synthesized and tested. As an example, 2-Morpholinoethan-1-ol 226 was activated as carbonate, then a suitable aminoalcohol fragment was alkoxycarbonylated, providing carbamate inhibitor 229, which exhibited almost 4-fold superior activity against wild-type HIV protease, compared to darunavir, along with appreciable antiviral activity against darunavir-resistant HIV-1 variants (Scheme 43).

2.4. Heteroaromatics

Based on Ritonavir, Kaye designed novel, structurally simplified, truncated analogs bearing heteroaryl groups linked to the peptidomimetic chain [67]. Starting from pyridine-2-carbaldehydes 230, the indolizine-2-carboxylate esters 232 (Scheme 44) were synthesized by Baylis–Hillman reaction followed by cyclization promoted by acetic anhydride at reflux, through acetylation of the alcohol, which facilitated the reaction. Saponification of the esters yielded the corresponding carboxylic acids 233. Different amides were prepared by coupling with suitable amines in presence of carbonyl diimidazole (CDI). According to the authors, inhibition activity was not measured; however, enzyme-binding, enzyme inhibition, and in silico docking studies of those compounds were planned.

The incorporation of substituents with hydrogen bond donor and acceptor groups at the P1 position of a symmetry-based HIV protease inhibitor series was described by DeGoey and resulted in significant improvement in potency against the resistant mutants [68]. Overall, compound 240 (Scheme 45) demonstrated the best balance of potency against drug resistant strains of HIV and exhibited oral bioavailability in pharmacokinetic studies. The synthesis began with benzylated tyrosine 234 and reaction with sodio acetonitrile yielded nitrile 235, which was then treated with a benzyl Grignard reagent to give the enaminone 236. A high degree of stereocontrol was observed during the stepwise reduction of 236 with NaBH₄, followed by Boc protection of the resulting amine. Removal of the benzyl protecting groups through hydrogenolysis gave the amine 238. The corresponding triflate was generated using N-phenyl-bis(trifluoromethanesulfonimide) affording 239. Palladium-mediated coupling with 5-fluoro-3-pyridylboronic acid furnished the bis aryl intermediate, which was finally coupled with a suitable t-butyl-glycine derivative, yielded the desired inhibitor 240.

During the study of novel inhibitors against South African wild-type (C-SA) HIV protease, Makatini [69] described the first pentacycloundecane (PCU) diol peptoid-derived inhibitors, with IC₅₀ values ranging from 6.5 to 0.075 µM. Starting from inhibitor 241, which showed very good activity (IC₅₀ 0.075 µM), it was derivatized by substituting the carbobenzyloxy group with the (2-pyrimidinylthio)acetic acid group in order to produce compound 242 for improved solubility (Scheme 46). Unfortunately, this molecule exhibited significantly less binding affinity to the enzyme (IC₅₀ 1.0 µM).

The discovery of HIV-1 protease inhibitors, which exhibit high potency against both HIV-1 wild-type and multi-PI-resistant HIV-mutants is always a key focus. The work of Kesteleyn [70] was oriented towards discovering new PIs suitable for a long-acting injectable drug applications. In this regard, they described new compounds bearing a heterocyclic 6-methoxy-3-pyridinyl (MP) or a 6-(dimethylamino)-3-pyridinyl (DMAP) (R₃) group at the para-position of the P1′ benzyl fragment, which showed antiviral activity in the low nanomolar range. The introduction of a heteroaryl moiety was performed via a Suzuki coupling reaction on polyfunctionalized bromoaryl intermediate 249. Bromo lactone 248 was recognized as a key intermediate. For the synthesis of 248, a new enantioselective methodology was developed. As shown in Scheme 47, the first step involved converting N-Boc-protected (S)-4-bromophenylalanine 243 into the Weinreb amide 244. Treatment with 3-butenylmagnesium bromide then furnished ketone 245, which was oxidized at the terminal double bond to afford the γ-ketocarboxylic acid 246. The desired bromo-lactone 247 was obtained by initially esterifying the carboxylic acid and then performing a reductive cyclization with N-selectride. Final benzylation, using lithium hexamethyldisilazane as a base at low temperature, produced the alkylated lactone 248, stereoselectively. Hydrolysis of 248 provided the corresponding carboxylic acid derivative. To prevent re-lactonization, silylation of the alcohol group with TBDMSCl was performed. The introduction of the P2 amine was achieved via the HATU-mediated coupling. Subsequently, the heterocyclic aromatic group R₃ (MP or DMAP) was introduced through palladium-assisted Suzuki coupling of boronic acids, producing compounds 250. Finally, N-Boc deprotection under acid conditions was followed by HATU-mediated coupling of N-(methoxycarbonyl)-L-tert-leucine, resulting in the desired HIV-1 PIs 252 with high overall yields and enantioselective purity (ee > 95%).

In the last period, studies on the synthesis of key heterocyclic fragments, particularly those oriented towards the pharmaceutical industry, were described. Kappe and his group developed multistep continuous flow reaction method for the synthesis of the biaryl-hydrazine unit of atazanavir [71]. The synthesis involved Pd-catalyzed Suzuki–Miyaura cross-coupling, followed by hydrazone formation and a hydrogenation step. At the end, an additional liquid–liquid extraction step was performed. The method yielded the desired product with an overall yield of 74%, which exceeded the 53% overall batch yield previously described in the literature.

A useful method was developed by Lindhardt and Skrydstrup for the synthesis of active esters by palladium-catalyzed alkoxycarbonylation of (hetero)aromatic bromides (Scheme 48) [72]. The protocol was general for a range of oxygen nucleophiles including N-hydroxysuccinimide (NHS) 252 and showed high functional group tolerance. The method enabled the synthesis of an important precursor to the HIV protease inhibitor saquinavir, through the formation of an NHS ester followed by acyl substitution.

Quinoline and isoquinoline moieties have been extensively studied as scaffolds for HIV inhibition and have been recently reviewed [73]. In the work of Sarveswari [74], two diverging series of water-soluble, non-peptidic quinoline analogs were designed and synthesized through the sequential attachment of piperazine and various amino acids to the quinoline scaffold using peptide coupling procedures. All synthesized compounds were subjected to in silico screening against HIV protease-1 and the cytotoxicity of all compounds was examined on HCT116 cells. In particular, compound 259 (Scheme 49), which bears serine and piperazine in sequence at C2 of quinoline, showed significant HIV protease inhibition properties and demonstrated cytotoxicity IC₅₀ value at 22.7 ± 0.59 nM. The synthesis began with the construction of quinoline core via HCl-promoted coupling between ethyl 4-chloroacetoacetate 254 and 2-amino-5-chlorobenzophenone 253 in methanol, yielding 94% of the desired compound 255. In the second step, Boc-piperazine was introduced by nucleophilic substitution, leading to compound 256. After removing the Boc protecting group, the piperazine amine was coupled with L-serine using HOBt, EDC. Final deprotection of the amino acid in methanolic HCl afforded the target compound 259.

On the way to extending hydrogen bonding interactions between inhibitor and HIV-1 protease, pyrimidine bases were recognized as suitable as P2 ligands capable of enhancing the activity of the inhibitors. In Wang’s work [75], inhibitor 264 (Scheme 50), bearing N-2-(2,4-Dioxo-3,4-dihydropyrimidin-1(2H)-yl) acetamide and a 4-methoxylphenylsulfonamide, showed high enzyme inhibitory activity, with an IC₅₀ of 2.53 nM in vitro and an inhibition ratio with 68% against wild-type HIV-1 in vivo, with low cytotoxicity. Its antiviral activity was also effective against DRV-resistant HIV-1 variants. The syntheses began with the preparation of substituted 2, 4-dioxopyrimidin-1(2H)-yl acetic acid 262 from uracil, via N-alkylation with ethyl bromoacetate (38%), followed by saponification with sodium hydroxide (61%). Coupling with the suitable amine 263 under EDCl/HOBt/DMAP conditions yielded the final product in high yield (98%).

Atazanavir is one of the most prescribed HIV-1 protease inhibitors approved by the FDA. It was the first protease inhibitor approved for once-a-day dosing to treat AIDS, owing to its good oral bioavailability and favorable pharmacokinetic profile. Reddy’s work [76] resulted in a new multistep synthesis for biaryl-hydrazine unit {tert-butyl 2-[4-(2-pyridinyl)benzyl]hydrazinecarboxylate} of atazanavir 268 on a large scale (Scheme 51). The synthesis began with a palladium catalyzed Suzuki–Miyaura coupling of readily available 2-chloropyridine 264 and (4-cyanophenyl)boronic acid 265, yielding the biaryl derivative 266. Next, the cyano group was reduced to the aldehyde 267 using DIBAL-H. Treatment with tert-butyl carbazate, followed by in situ reduction with NaBH₄-furnished hydrazone 268. The entire process was scaled-up and completed in three steps with an overall yield of 71%.

Several naturally occurring triterpenes and their semisynthetic analogs have demonstrated potent anti-HIV activities. Recently, Zheng [77] synthesized thirteen nitrogen-containing derivatives of 3,11-dioxo-olean-12- en-30-oic acid by introducing various amino acids and nitrogen-containing heterocyclic groups at the 30-carboxyl group, starting from 18β-glycyrrhetinic acid (GA) (Scheme 52). Among them, compound 270 displayed relatively moderate inhibitory activity, with IC₅₀ values below 0.24 mM. Molecular docking studies revealed favorable hydrophobic–hydrophobic and hydrogen bonding interactions in the active site of HIV-1PR. These findings underscore the potential of such derivatives as promising candidates for the development of HIV-1 PR inhibitors. The synthesis commenced with the oxidation of the 3-hydroxyl group of GA to a carbonyl group (97%), yielding compound 269. The target compounds 270 were synthesized using an EDCI/HOBt/DMAP/TEA system in the presence of a suitable nucleophile.

Certain thiazolyl and benzothiazolyl guanidines have been reported as exhibiting a wide range of pharmacological and antimicrobial activities. In the field of carbohydrates, glycosyl isothiocyanates are versatile synthetic intermediates for the synthesis of biologically active carbohydrate derivatives. Cao et al. synthesized and evaluated the bioactivity of some new N-glucosyl-N′-(4-arylthiazol-2-yl) aminoguanidines (Scheme 53) [78] The starting materials, 2-amino-4-arylthiazoles of type 271, were refluxed with substituted glycosyl isothiocyanates in dry benzene. Subsequent desulfurization of the resulting thioureas with HgCl₂ in DMF, in the presence of hydrazine hydrate and TEA, furnished the target compounds. In particular, compounds 274 and 275 showed moderate activity against HIV-1 protease (IC₅₀ between 22 and 107 mg/mL).

While ritonavir is approved as an HIV PI, it is hardly used as such, being more frequently used as a pharmacokinetic enhancer. Unfortunately, its use is often associated with various side effects. Therefore, novel derivatives have been described. Jonckers [79] proposed thiazol-5-ylmethyl (2S,3R)-4-(2-(ethyl(methyl)amino)-N-isobutylbenzo[d]oxazole-6-carboxamido)-3-hydroxy-1-phenylbutan-2-ylcarbamate 281 as a lead candidate for this class (Scheme 54). This compound, together with structurally similar analogues, demonstrated excellent ‘boosting’ properties when tested in dogs. These findings made it attractive in the search for novel pharmacokinetic enhancers. The synthesis of the lead compound 279 began with commercially available epoxide 276, which was reacted with an excess of isobutyl amine, yielding monoprotected bis-aminoalcohol 277. This was then coupled with acid derivative 278 using HATU or BOP as activating agent, yielding intermediate 279. After mild Boc-deprotection, the intermediate amine was coupled with suitable thiazolyl carbonate 280, affording target compound 281 in high yield. The in vitro antiviral activity against wild-type HIV-1 (EC₅₀ = 71 mM) was evaluated in an acutely infected lymphoblastic cell line (MT4-LTR-EGFP) and CYP 3A4 inhibition (IC₅₀ = 0.031 mM) was determined in vitro using a human liver microsome (HLM)-based assay.

Multiple heteroaromatic fragments were introduced into novel inhibitors. In particular, Houpis [80] described the convergent synthesis of clinical candidate 282 (Scheme 55), a protease inhibitor specifically designed to allow for long acting-controlled release formulations. Central disconnection generated two synthons 283 and 284 bearing heteroaromatic moieties.

The preparation of epoxide 284 started from protected tyrosine 285 (Scheme 56). The phenol functional group was activated as the triflate, using triflic anhydride in the presence of pyridine. The resulting product was then coupled with a suitable pyridyl boronic acid in the presence of commercially available ferrocenyl-based catalyst, Cl₂Pd (dppf)-DCM, yielding the bis-aryl intermediate 286. The key intermediate, epoxide 284, could be obtained from either metal-catalyzed or enzymatic process in high yield by exposing 287 to aqueous potassium hydroxide in tert-amyl alcohol.

The optically pure amino alcohol (R,R-293) was prepared starting from ketone 289 (Scheme 57), which was brominated selectively at the α-aliphatic position using CuBr₂ in refluxing EtOAc to give rac-290. An NaBH₄ reduction followed by a Ritter reaction afforded the racemic cis-amino alcohol 292. The optically active derivative (R,R-293) was obtained in good yield via a resolution with S-mandelic acid. The optically active amide 283 was prepared using well established methods. Subsequent coupling with epoxide 284 was performed using ZnCl₂/TMEDA system, followed by the addition of LiHMDS. Deprotection of the BOC group and coupling with the methyl carbamate tert-butylglycine furnished the desired inhibitor 282.

Ghosh [81] described the design, synthesis, and biological evaluation of a series of novel HIV-1 protease inhibitors bearing isophthalamide derivatives as the P2–P3 ligands. In particular, compound 295 showed an enzyme Ki of 0.17 nM and antiviral IC₅₀ of 14 nM (Scheme 58). The general synthetic strategy involved the coupling of isophthalic acid derivative 296 with the amine of the hydroxyethylamine sulfonamide isosteres 297 to produce HIV-1 protease inhibitors. Various isophthalic acid derivatives can be obtained by coupling readily available isophthalic monoacid 298 with various amines. In particular, amine 299 was prepared from oxazole ester 300, which was first reduced with excess of DIBAL-H to afford the corresponding alcohol 301. Subsequently, azidation of the alcohol with diphenylphosphoryl azide in the presence of DBU provided the corresponding azide. Finally, reduction of the azide with triphenyl phosphine in THF yielded amine 299.

Ghosh and his group described the effect of a single atom change or a scission of a single bond on the activity of a panel of compounds structurally like darunavir. Seven novel PIs were synthesized, and the modifications involved an exchange of sulfur for oxygen, a scission of a single bond in P2′-cyclopropylaminobenzothiazole (or -oxazole), and/or introduction of P1-benzene ring with mono- or bis-fluorine atoms (Scheme 59) [82]. X-ray structural analyses of the PIs complexed with wild-type protease (PR_WT) and highly multi-PI-resistant PR_{DRV P51} revealed that the PIs adjusted themselves to the protease with resistance-associated amino acid substitutions. Inhibitors containing a benzothiazole moiety at the P2′ position showed greater anti-HIV-1 activity than those with a benzoxazole moiety. It was claimed that this greater potency is attributed to the capacity of sulfur atoms to form bidirectional σ-hole potentials with the carbonyl oxygen of G48 [83]. On the other hand, substitution of cyclopropyl with isopropyl at the distal part of the inhibitor’s P2′ moiety resulted in a reduction in the antiviral activity. The membrane penetration data confirmed previous findings that the addition of two fluorine atoms greatly boosts the activity of the inhibitors.

More recently, Ghosh et al. [84] described new inhibitors bearing various pyridyl-pyrimidine, aryl thiazole, or alkylthiazole moieties as P2 ligands in darunavir-like hydroxyethylamine sulfonamide isosteres, with the aim of promoting hydrogen bonding interactions with the backbone atoms in the S2 subsite of HIV-1 protease. The different ligands were introduced using suitable parent benzoic acids, whose synthesis was described. The new inhibitors showed sub-nanomolar levels of protease inhibitory activity and low nanomolar levels of antiviral activity. In particular, pyridyl-pyrimidine benzamide derivative 303a (Scheme 60) displayed an enzyme inhibitory Ki of 28 pM and antiviral activity of 154 nM, whereas compound 303b displayed potent antiviral activity with an IC₅₀ value of 66 nM. Among the thiazole-derived inhibitors, compound 308, with a pyridyl thiazole as the P2 ligand, showed the best result, exhibiting an HIV-1 protease inhibitory Ki of 8.7 nM and antiviral activity of 580 nM.

For the synthesis of 303, commercially available 4-methyl-3-[[4-(3- pyridinyl)-2-pyrimidinyl]amino]benzoic acid 302 was reacted with the previously reported hydroxyethylene isosteres 297 in the presence of HATU and Et₃N in DMF. The synthesis of 308 started from 2-methyl 5-aminomethylbenzoate 304, which was reacted with potassium thiocyanate and acetyl chloride in acetone to yield the thiourea derivative 305. Exposure of 305 to potassium carbonate in methanol, followed by the addition of α-bromo ketone 306, furnished the pyridinylthiazole derivative 307. Saponification of methyl ester with aqueous LiOH in THF afforded the corresponding carboxylic acid. Final coupling with the hydroxyethylsulfonamide isosteric amine 297a produced inhibitor 308 in good yields.

In 2010, we synthesized a series of new thienyl analogues of nelfinavir and saquinavir with different substitution patterns, derived from suitable enantiopure diols [85]. Their inhibitory activity against wild-type recombinant HIV-1 protease was evaluated. In general, thienyl groups spaced from the core by a methylene group yielded products with IC₅₀ values in the nanomolar range, regardless of the heterocycle’s type or substitution pattern. Notably, compounds 309 and 310 (Scheme 61) were the most active, and their activity was substantially maintained or even increased against two common mutants under drug pressure, such as V32I and V82A. The synthesis began with dihydroxybutylesters 311, which were reduced by BH₃SMe₂, producing triols 312 in nearly quantitative yield. Freshly prepared 3,3-dimethoxypentane was used for selective protection of the hydroxyl groups on carbons 1 and 2. The free hydroxyl was then transformed into the mesylate 313 and subsequently substituted with an azido group. Final hydrolysis yielded azidodiols 314 with modest to good overall yield.

The perhydroisoquinolinic fragment was linked via selective activation of the hydroxyl group as a mesitylene sulfonyl derivative 316, followed by reaction with the commercially available substituted perhydroisoquinoline (Scheme 62). Finally, the azidoalcohols were transformed into amino alcohols 318a and 318b by Pd-catalyzed hydrogenation in excellent yield.

For the synthesis of the nelfinavir thienyl analog, amino alcohol 318a was reacted with 3-acetoxy-2-methylbenzoic acid. Final deacetylation of the phenolic group yielded the target compounds 309 in good yield (Scheme 63). The characteristic dipeptide unit of saquinavir derivatives 319 was prepared by coupling asparagine tert-butyl ester with quinolinic acid. Hydrolysis with TFA afforded acid 319, which was subsequently coupled with amino alcohol 318b, resulting in the desired saquinavir derivatives 310.

Investigating the synthesis and biological evaluation of new structurally simplified non-peptidic heteroaromatic molecules as PIs, we described the synthesis and biological evaluation of a new series of potential HIV-1 protease inhibitors of types 330–333, incorporating different benzofused heterocycles (Scheme 64) [86]. The variation in heteroatoms in such molecules affected biological activities and a benzothiophene containing inhibitor 333 exhibited high potency against wild-type HIV-1 protease with an IC₅₀ of 60 nM, thanks to the lower desolvation penalty paid by this hydrophobic moiety. The synthesis began with commercially available (S)-glycidol (98% ee), which was reacted with 3-nitrobenzenesulfonyl chloride (NsCl) and triethylamine (Et₃N) at −10 °C to afford 320 in 80% yield. Subsequently, it underwent regioselective displacement of the nosylate with the appropriate 5-hydroxyheteroarene and K₂CO₃ to yield the corresponding epoxides 321, 324, and 325 in 70%, 78%, and 82% yield, respectively. The indole derivative 321 was alkylated with methyl iodide and benzyl chloride to produce compounds 322 and 323 in 90% and 86% yield. The opening of the oxiranyl ring with i-BuNH₂ in i-PrOH provided aminoalcohols 326–329 in quantitative yield. These compounds were then reacted with 3,4-dimethoxybenzenesulfonyl chloride and Et₃N in dry DCM to afford target compounds 330–333 in high yield.

With the aim of facilitating access to new HIV-1 protease inhibitors bearing heteroaryl moieties as P1-ligands, we speculated on a convenient synthetic route to introduce diversity into the common hydroxyethylamino core present in several approved PIs [87] In a straightforward retrosynthetic approach, variously functionalized aromatic groups could be incorporated via Suzuki coupling between an activated C(sp3) bromide (allylic electrophile) and an array of arylboronic acids, thereby furnishing methyl 4-arylcrotonates. An effective ligand-free Suzuki coupling protocol was described for coupling methyl (E)-4-bromobut-2-enoate with several arylboronic acids. Given the strong interest in methodologies that produce polyarylated frameworks, we subsequently reported a nickel-catalyzed double phenylation of methyl 4-bromocrotonate, which furnished suitable doubly phenylated building blocks [88].

Different aspects regarding the preparation of peptidomimetic and pseudopeptidic structures containing heterocycles were also reviewed in 2012 by us, with particular focus on novel tricyclic structures as potential drugs [89]. Following the concept of targeting the protein backbone, we systematically investigated various substitution patterns on the common stereo-defined isopropanolamine core. In 2014, we described the synthesis of new, structurally simple indolic non-peptidic HIV-protease inhibitors from (S)-glycidol using regioselective methods [90], varying the type and/or the position of the functional group on the indole and the nature of the nitrogen containing group (sulfonamides or perhydroisoquinoline). The systematic study of in vitro inhibition activity of these compounds confirmed the general beneficial effect of the 5-indolyl substituents in the presence of arylsulfonamide moieties, which showed activities in the micromolar range. Oxyindoles and carbamoyl indoles showed general good activity, whereas simple aminoindoles were much less active (Scheme 65). For the synthesis of oxyindoles, epoxide 322 was opened with perhydroisoquinoline, yielding inhibitor 334, or alternatively with i-BuNH₂, giving intermediate 338. This last amino alcohol was reacted with 3,4-diOMe-phenylsulfonyl chloride, affording compound 335 in excellent yield.

The preparation of the corresponding carbamoyl derivatives 336 and 337 was straightforward. 5-Aminoindole 339 was first reacted with p-nitrophenylchlorocarbonate to afford the activated carbamate 340 (Scheme 66). Glycidol was then introduced via a substitution reaction, yielding the oxiranyl carbamate 341 in good yield. The introduction of i-BuNH₂, followed by sulfonylation with a suitable ArSO₂Cl, furnished the final compounds 336 and 337.

Different spacers were studied, connecting heteroaryl moieties to the hydroxyethylamine core. Thus, in 2017, new heteroaryl HIV protease inhibitors, bearing a carboxy–amide spacer, were synthesized in our lab in a few steps and with high yield, starting from commercially available homochiral epoxides [91]. Onto a given hydroxyethyl-amino-isopropanoyl-sulfonamide core, we introduced different heteroarenes and modified the central core, with the presence of either H or a benzyl group (structure A in Scheme 67). For the synthesis of the simple, unsubstituted isopropanolamine core (R = H), we took advantage of an established route, starting from the commercially available bidentate electrophile (S)-glycidol. The epoxide was opened by i-PrNH₂; the aryl sulfonyl moiety was introduced, and the primary OH group was substituted by NH₂, furnishing the key aminoalcohol 343 in good overall yield. Coupling with suitable 5-heteroaryl acids afforded the corresponding amides 344.

The synthesis of benzyl derivatives 347 was even shorter (Scheme 68) and started from the commercially available homochiral N-Boc-protected amino-epoxide 345. After opening with iBuNH₂ and the subsequent introduction of the arylsulfonyl moiety, the N-Boc group was displaced by treatment with trifluoroacetic acid in dichloromethane. The resulting ammonium trifluoroacetate was treated with NEt₃, affording amine 346, which was then reacted with suitable 5-heteroarylcarboxylic acids previously activated with N,N′-carbonyldiimidazole. Thus, the final products 347 were obtained in four steps and excellent overall yield.

In general, the presence of a carboxyamide moiety showed a positive effect on in vitro inhibition activity against recombinant protease. The IC₅₀ values ranged between 1 and 15 nM. In particular, benzofuryl derivatives 344a and 347a displayed some of the best IC₅₀ values among such structurally simple inhibitors. Docking analysis supported the experimental results regarding activity, demonstrating that these benzofuryl derivatives exhibited a favorable number of interactions within the active site. The inhibitory activity of these molecules was also evaluated in HEK293 cells.

The study of new heteroaryl HIV protease inhibitors was extended to those bearing a carbamoyl spacer [92]. In particular, we focused on new derivatives with the general structure B (Scheme 69), in which the heterocycle is spaced from the core by a carbamoyl function, like the arrangement in darunavir and TMC-126. Their synthesis was straightforward from commercially available homochiral epoxides. Different substitution patterns were introduced onto a given isopropanoyl-sulfonamide core, which could have either H or benzyl group.

Both the carbamoyl moiety and benzyl group displayed a general beneficial effect on the in vitro inhibition activity against recombinant protease. The IC₅₀ values ranged from 11 to 0.6 nM. In particular, benzofuryl and indolyl derivatives 351a and 351c showed some of the best IC₅₀ values (IC₅₀ 0.6 nM for both). Regarding the amide inhibitors, their activity was also confirmed in HEK293 mammalian cells and was maintained against protease mutants. Furthermore, the metabolic stability of all compounds was studied and found to be comparable to that of commercially available inhibitors. More recently, following the concept of repositioning HIV protease inhibitors as cancer therapeutics, compound 349b was also evaluated for its ability to induce cytotoxicity in hepatocellular carcinoma cell lines [93].

Considering the structure of HIV protease as a C2-symmetric homodimer in its active form, we recently reported on the synthesis, enzyme inhibition, and structure–activity relationship of a new class of HIV-1 protease inhibitors containing a pseudo-symmetric hydroxyethylamine core and heteroarylcarboxyamide moieties [94]. To obtain a pseudo-symmetric hydroxyethylamine core, a benzyl group was placed on the sulfonamide nitrogen, resulting in the general structure C (Scheme 70). A straightforward synthetic pathway yielded nine compounds in a few steps with high yields. Potent inhibitory activity, with nanomolar IC50 values measured with a standard fluorimetric test, was achieved. In particular, compounds 353a–c (whose synthesis is described in Scheme 70), which contain the indole ring in P1, exhibited HIV-1 protease inhibitory activity that was more potent than darunavir in the same assay.

Recently, considering all our data on heteroaryl non-peptidic inhibitors, specifically their high HIV protease inhibitory activity and easy of synthetic approach, we reported novel simple heteroaryl carboxamides (Scheme 71), bearing p-NO₂ electron withdrawing group or p-OMe electron releasing group on arylsulfonamide. We compared their in vitro activity in HEK293 cells with that of our previously described compounds [95].

Benzofuryl-, benzothienyl-, and indolyl rings were introduced via efficient synthetic procedures. All compounds showed inhibitory activity comparable to the commercial drug darunavir, with particularly potent examples such as 356a, 356b, and 355b (IC₅₀ < 0.6 nM). These compounds were effective against both wild-type HIV-1 protease and mutants containing V32I or V82A mutations. In silico evaluation of their absorption, distribution, metabolism, and excretion (ADME) properties was also conducted, comparing the results with the predicted properties of darunavir. As a result, 12 out 27 compounds (including 355b) performed better than or equal to darunavir across all ADME prediction models, demonstrating the potential of these compounds for further drug development.

3. Conclusions

In conclusion, the introduction of heterocyclic moieties into HIV protease inhibitor scaffolds remains a powerful tool to enhance their activity and overcome enzyme mutations. Both non-aromatic heterocycles, owing to their rigid structures, and heteroaromatics, due to their planar structures combined with modulation of hydrogen bonding interactions, can expand the pool of inhibitors by following the concept of targeting the protein backbone.

Although much has been described regarding the interaction between heterocyclic fragments and the enzyme, the field remains open to further investigations, particularly concerning heterocycles with multiple heteroatoms (aromatic and non-aromatic) and simplified structures that reduce synthetic costs.

Table 1. Most significant new HIV-protease inhibitors of the last 15 years.

Category	Sub-Category	Chemical Structure (Examples)	Biological Properties	References
Non-aromatic O-heterocycles	Bis-THF derivatives		Ki = 0.013 nM IC50 = 2.8 nM (antiviral)	[17]
			Ki = 2.9 pM ID50 = 2.4 nM	[27]
			Ki = 0.04 nM (HIVP) IC50 = 0.25 nM (Antiviral)	[31,32]
			Ki = 54 pM (HIVP) IC50 = 86 pM (Antiviral)	[28]
			Ki = 10 pM IC50 = 1.9 nM	[42]
			IC50 = 0.05–0.43 nM	[43]
Non-aromatic N-heterocycles	1,3-diazacycloalkan-2-ones		Ki = 0.30 nM IC90 = 140 nM	[47]
			ND	[47]
	Indolin-2-ones		Ki = <0.10 nM (WT) IC50 = from 3 to >400 nM (WT)	[50]
	Benzodiazepines		Ki = 12 nM IC50 = 0.03 µM	[51]
			Ki = 11 µM	[52]
	Tetrahydropyridines		Estimated Ki = 0.74–2.29 µM	[53]
	Piperidines and Pyrrolidines		IC50 0.13 nM	[54]
			EC50 = 37 nM	[55]
	Hydantoins			[58,59]
				[58,60]
Non-aromatic heterocycles with multiple heteroatoms	1,3-Thiazolidine		Ki = 5.2 pM IC50 = 9.2 nM	[62]
	6,8-dioxa-3-azabicyclo[3.2.1]-octane		IC50 in the sub-micromolar range	[64]
	Morpholines		IC50 = 0.8 nM (Enzyme) IC95 = 49 nM (Antiviral)	[65]
			Ki = 0.09 nM IC50 = 0.41 nM	[66]
Heteroaromatics	Pyridines		EC50 = 4 nM	[68]
			EC50 = Low nM range	[70]
	Pyrimidines		IC50 = 1.0 µM	[69]
	Pyrimidines		IC50 = 2.53 nM	[75]
	Quinolines		IC50 = 22.7 nM	[73,74]
	Thiazoles		IC50 = between 22 and 107 mg/mL	[78]
			IC50 = 0.031 mM EC50 = 71 mM (WT)	[79]
			IC50 = 0.0094 nM (WT)	[82]
			Ki = 8.7 nM IC50 = 508 nM	[84]
	Thiophenes		IC50 = 2.9 nM	[85]
	Benzothiophenes		IC50 = 0.6 nM	[85]
			IC50 = 60 nM	[86]
	Indoles		IC50 = 14 µM	[90]
			IC50 = 2–8 µM	[90]
	Benzofurans		IC50 < 0.6 nM	[95]

summarizes the most significant new HIV-protease inhibitors appearing in the literature during the last 15 years.