Nirmatrelvir: From Discovery to Modern and Alternative Synthetic Approaches

Abstract

The global urgency in response to the COVID-19 pandemic has catalyzed extensive research into discovering efficacious antiviral compounds against SARS-CoV-2. Among these, Nirmatrelvir (PF-07321332) has emerged as a promising candidate, exhibiting potent antiviral activity by targeting the main protease of SARS-CoV-2, and has been marketed in combination with ritonavir as the first oral treatment for COVID-19 with the name of Paxlovid^TM. This review outlines the synthetic approaches to Nirmatrelvir, ranging from Pfizer’s original method to newer, more sustainable strategies, such as flow chemistry strategies and multicomponent reactions. Each approach’s novelty and contributions to yield and purification processes are highlighted. Additionally, the synthesis of key fragments comprising Nirmatrelvir and innovative optimization strategies are discussed.

1. Introduction

The devastating COVID-19 pandemic has prompted scientists to swiftly focus on finding novel molecules or scaffolds exhibiting antiviral properties, aiding in the mitigation of SARS-CoV-2-induced illness. Academic and industrial researchers have delved into molecules previously advanced in the drug-discovery pipeline, rigorously testing their efficacy against COVID-19. In addition to well-known marketed antiviral drugs, such as remdesivir, favipiravir, and umifenovir, inter alia, scientists’ attention focused on PF-00835231, which is endowed with strong antiviral efficacy against SARS-CoV-1 [1]. To establish a structure–activity relationship (SAR) for PF-00835231, researchers synthesized analogs by modifying the constituent amino acids with six notable fragments. Among these, the most promising analog, PF-07321332 (Nirmatrelvir), received FDA approval in December 2021 as part of an antiviral combination pill for home use, aimed at preventing severe illness in patients infected with COVID-19 [2]. Nirmatrelvir exerts its antiviral activity via the covalent inhibition of the main protease of SARS-CoV-2 (SARS-CoV-2 M^pro) [3]. However, its efficacy is limited when administered alone in vivo, due to rapid metabolism by the CYP3A4 enzyme. To overcome this limitation, Nirmatrelvir found suitability for therapeutic application in combination with ritonavir, initially formulated as an HIV protease inhibitor. Ritonavir effectively enhances the bioavailability of Nirmatrelvir by inhibiting CYP3A4. This synergistic combination, developed by Pfizer, is marketed as Paxlovid™ [4]. As stated by real-world data, Paxlovid™ showed proven efficacy in preventing COVID-19 progression among high-risk individuals, significantly reducing the need for hospitalization and supplementary oxygen administration [5,6]. However, ongoing monitoring and research are essential to better understand its long-term efficacy and potential resistance issues in diverse patient populations.

This review will provide the reader with an overview of the most significant synthetic approaches to Nirmatrelvir, starting from the original one developed by Pfizer, and then incorporating new, more sustainable synthetic approaches, such as flow chemistry or multicomponent reactions, taken from the recent literature. The novelty of the methods will be thoroughly explored, emphasizing the distinctive contributions that each step imparts to the final yields or purification processes. Moreover, the syntheses of the main fragments that constitutes Nirmatrelvir and the innovative strategies that facilitated their optimization will also be briefly discussed individually.

2. Development of Nirmatrelvir

Following the global emergency caused by the spread of SARS-CoV-2, leading pharmaceutical companies have embarked on efforts to research efficacious vaccines to contain the virus’s spread among humans and valuable therapies to mitigate the severity of COVID-19 infections. While a wide range of vaccines has been produced during the pandemic, only one orally administered drug has been approved by the FDA and made available on the market for the treatment of severe forms of COVID-19. Paxlovid consists of the combination of two medications: Nirmatrelvir, which inhibits the viral replication process, and ritonavir, which helps boost the levels of Nirmatrelvir in the body. Paxlovid is administered orally and has been authorized for emergency use in several countries as a treatment for individuals with mild to moderate COVID-19 symptoms, who are at high risk of progressing to severe illness [7]. This medication emerged when vaccination campaigns were underway in many countries, while some others had already achieved high vaccination rates. This timing was not coincidental as, during this period, despite vaccines showing beneficial effects on reducing hospitalizations and mortality rates, there was significant concern about potential virus mutations that could diminish vaccine efficacy. Consequently, there was a pressing urgency for orally available, cost-effective antiviral treatments to administer, aiming to minimize COVID-19-associated deaths.

Before Paxlovid marketing, the FDA had repurposed several other therapeutics for treating COVID-19 infection. One such example is the RdRp inhibitor remdesivir, which is intravenously administered to COVID-19-infected hospitalized adults and children. Several challenges were associated with this medication, including its low stability in the blood and limited half-life (necessitating continuous administration). Molnupiravir, another RdRp inhibitor, has obtained full regulatory approval in the UK. The advantage in this case relies in its oral bioavailability. Within this context, there was a growing demand for orally bioavailable therapeutics against COVID-19, which finally led to the development and approval of Nirmatrelvir.

2.1. Target Validation: SARS-CoV-2 M^pro

Nirmatrelvir is a potent inhibitor of SARS-CoV-2 M^pro, thereby blocking viral replication and strongly reducing the severity of COVID-19 infection [8,9]. M^pro plays a crucial role in the viral replication cycle by processing the polyproteins translated from the viral RNA [10]. The protease cleaves these polyproteins at specific sites to generate active viral proteins essential for RNA replication and transcription. A selective inhibition of M^pro effectively alters the synthesis of such vital components, thereby disrupting the virus’s capability to replicate and spread. Due to its indispensable role and the absence of a human counterpart with similar cleavage specificity, M^pro becomes an ideal target for antiviral drug development [11]. This enzyme’s conservation across coronaviruses further enhances its appeal, offering potential broad-spectrum antiviral activity. Consequently, inhibitors of M^pro, such as Nirmatrelvir, hold significant promise in treating COVID-19 and possibly other coronavirus-related diseases. Figure 1 depicts the main structural features of Nirmatrelvir. The terminal -CN functional group is the key component of the entire molecule as it essentially represents its “electrophilic warhead”, which undergoes a covalent reaction with the reactive Cys145 thiol group of the target. Meanwhile, the remaining portion of the molecule, emulating the natural recognition sequence of M^pro, specifically the tripeptide Val-Leu-Gln, ensures Nirmatrelvir’s snug accommodation within the enzyme’s active site, where it establishes supplementary interactions. The recognition motif consists of a P1 unit formed by a Gln analog featuring a cyclic (γ-lactam) side chain and this modification makes noteworthy advantages because the native Gln’s amide group can intramolecularly react with certain types of warheads, leading to inhibitor inefficacy. The cyclic moiety offers greater rigidity compared to the native Gln side chain, which may enhance binding to the target enzyme. Additionally, converting Gln into a cyclic derivative improved synthetic accessibility. The P2 group is made of a dimethylcyclopropyl proline (DMCP) moiety, serving as a leucine analog that predominantly interacts with the enzyme through lipophilic interactions. Finally, the P3 residue is a tertiary leucine, mimicking valine protected at the N-terminal with a trifluoroacetyl group. A covalent reversible interaction occurs between Nirmatrelvir and M^pro thanks to the presence of the reactive electrophilic warhead. The P1’ nitrile group establishes a thioimidate bond with the Cys145 thiol functional group of M^pro via a Pinner-like mechanism. Even if -CN is a less reactive warhead in comparison to aldehydes or Michael’s acceptors, it offers enhanced selectivity and metabolic stability. Moreover, the thioimidate adduct is stabilized within the binding pocket thanks to supplementary interactions, strengthening the affinity. Finally, given that the thiol group of Cys145 is crucial for catalyzing the hydrolysis of peptide bonds, the enzyme’s functionality is interrupted.

2.2. Lead Discovery and Optimization

The discovery and development of Nirmatrelvir commenced with two hit compounds previously synthesized by Pfizer to counter the SARS-CoV infection: PF-00835231 (referred to as compound 2 in Figure 2), an α-hydroxy ketone-based peptidomimetic, and its corresponding phosphate prodrug, PF-07304814 (referred to as compound 3 in Figure 2).

Compounds 2 and 3 were selected for in vitro and in vivo evaluations against SARS-CoV-2 M^pro and SARS-CoV-2-infected cells and they showed promising results, in particular, PF-00835231 (2) exhibited high inhibitory potency (K_i = 4 ± 0.3 nM) and good antiviral activity (EC₅₀ = 231 nM). One noteworthy aspect of these two hit compounds was the presence of a P1 γ-lactam moiety, which mimics the native substrate’s Gln and serves as a recognition motif for the S1 pocket. The challenge associated with using free glutamine was the potential intramolecular reaction of the side-chain amino group with the electrophilic warhead, resulting in a cyclized inactive product. Indeed, this issue was circumvented by employing lactam functionality, which not only prevents such a side reaction, but also imparts rigidity leading to a reduced loss of conformational entropy upon binding to the target [12]. Despite their activity against SARS-CoV-2 M^pro, the “first generation” of M^pro inhibitors showed poor oral bioavailability, which was in contrast with the aim of Pfizer’s chemists to develop an antiviral oral drug to treat COVID-19 at the first stage of viral infection. Hence, Owen et al. designed a library of analogs of PF-00835231 (2) with the aim of improving the pharmacokinetic properties (Figure 3) [13]. In particular, the focus was on decreasing the number of H-bond donor (HBD) groups to increase gut absorption without affecting drug target recognition or the potency of the drug. The first modification involved the electrophilic warhead, replacing the α-hydroxy ketone moiety with two new functional groups, a benzothiazole-7-yl ketone and a nitrile group, both missing HBD groups. Nitrile derivate 4 showed enhanced oral absorption in rats but a lower antiviral and inhibitory activity with respect to PF-00835231 (2). Further modifications focused on the P2 leucine moiety, which was replaced with a cyclic modified proline moiety able to fit into the S2 pocket. This fragment was also present in serine protease inhibitor Boceprevir: the X-ray crystal structure of SARS-CoV-2 M^pro in a complex with Boceprevir provided a rational structural basis for the incorporation of such a moiety [14]. This modification in combination with the introduction of the benzothiazole-7-yl-ketone warhead produced compound 5, whose permeability was increased but activity decreased, with respect to that of compound 2. Therefore, to enhance potency, subsequent investigations revealed the necessity for a substituent in the P3 position that could more effectively occupy the S3 subsite. The indole moiety was then replaced by both a methanesulfonamide and a trifluoroacetamide functionality in 6 and 7. These compounds showed comparable potency, but 7 proved to be a better antiviral candidate due to its higher antiviral activity and oral absorption in rats and monkeys. The combination of a nitrile warhead with the trifluoroacetamide moiety led to the identification of Nirmatrelvir (1) as a novel clinical candidate. The nitrile warhead was selected above the benzothiazole-7-yl-ketone moiety due to some particular features, including the perceived ability to more easily scale-up, better solubility, and a reduced likelihood of epimerization at the stereocenter at the P1 position.

Nitrile is an established warhead for targeting serine and cysteine proteases [15]. In fact, the nitrile warhead has previously been explored as a reactive group in many FDA-approved drugs [16]. The electron-poor carbon of the nitrile group is able to undergo a nucleophilic attack from the thiol group of Cys, thus affording a reversible covalent thioimidate adduct. The introduction of the nitrile group instead of the HBD group of the α-hydroxymethyl ketone covalent warhead in 2 led to improved oral bioavability.

Table 1. Summary of the advantages and disadvantages of each Nirmatrelvir precursor that led the drug development process.

Compound	Advantages	Disadvantages
	High inhibitory potency Good antiviral activity	Low oral bioavailability
	Lower number of H-bond donors Nitrile warhead enhances oral absorption in rats	Antiviral and inhibitory activities lower than 2
	Benzothiazole-7-yl ketone warhead increases permeability Replacement of leucine with modified proline allows a better fit into the S2 subsite	Antiviral and inhibitory activities lower than 2 and 4
	Replacement of indole with methanesulfonamide allows a better fit into the S3 subsite Antiviral and inhibitory activities higher than 4 and 5	Antiviral and inhibitory activities lower than 2 Oral absorption in rats and monkeys lower than 7
	Replacement of indole with trifluoroacetamide allows a better fit into the S3 subsite Antiviral and inhibitory activities higher than 4 and 5 Oral absorption in rats and monkeys higher than 6	Antiviral and inhibitory activities lower than 2
	Combination of nitrile warhead, modified proline at P2 position, and trifluoroacetamide at P3 position provides antiviral and inhibitory activities higher than 2 Oral bioavailability higher than all precursors

shows all the advantages and disadvantages of each Nirmatrelvir precursor that led the drug development process.

3. Synthetic Approaches to Nirmatrelvir

3.1. Pfizer’s Synthesis

The original Nirmatrelvir synthesis developed by Pfizer is reported in the patent US11,351,149. The document, filled on 5 August 2021 and published as the PCT application WO2021/250648, presents several examples of the synthesis of Nirmatrelvir [17]. In particular, the procedure described in Scheme 1 allows us to adapt Nirmatrelvir’s synthesis also on a larger multigram scale. The synthetic route starts with the bicyclic pyrrolidine derivative 8, which is reacted with Boc-protected L-tert-leucine 9 in a Hexafluorophosphate Azabenzotriazole Tetramethyl Uronium (HATU)/N,N-Diisopropylethylamine (DIPEA)-mediated amide coupling in DMF to afford 10 in a 50% yield after column chromatography purification. The methyl ester is hydrolyzed in basic conditions to afford 11. After Boc-deprotection, the reaction of 12·HCl with ethyl trifluoroacetate produces intermediate 13 in a quantitative yield. The latter reacts with hydrochloric salt 14·HCl, in the presence of 2-Hydroxypyridine-N-oxide (HOPO) and N-(3-dimetilaminopropil)-N′-etilcarbodiimide crystalline (EDCI·HCl) in DMF, providing 15 in a 84% yield. Primary amide 15 is finally dehydrated with the Burgess reagent to furnish the target Nirmatrelvir (1) in an 81% yield. In a recent paper, the possibility of using isobutyl acetate as a solvent to obtain key intermediate 15 as a solvate with higher purity in comparison to its production with other existing synthetic methods is shown [18].

3.2. Structural Components of Nirmatrelvir

Based on the structural analysis, the chemical structure of Nirmatrelvir is considered a peptidomimetic composed of two main fragments, termed the “western fragment” and “eastern fragment”, linked via an amide bond (Figure 4). The western fragment comprises a bicyclic proline residue connected to an L-tert-leucine capped with a trifluoroacetyl group, while the eastern fragment consists of a γ-lactam analog of Gln, bearing the terminal electrophilic warhead. Algera et al. discussed the development of the first ever commercial process of Nirmatrelvir, describing the difficulties and challenges due to the extremely accelerated timeline that was needed to face the emergency [19,20,21]. This paragraph will highlight the strategies for the construction of both fragments, analyzing how the main synthetic challenges have been approached to obtain to the most effective chemical conditions.

3.2.1. Synthesis of the Bicyclic [3.1.0] Proline Building Block

One of the key fragments of Nirmatrelvir is the bicyclic [3.1.0] proline derivative 8 (Scheme 1), which is essential for the synthesis of the key western fragment intermediate 13 (Scheme 1). Several strategies have been attempted to provide this moiety with high yields, and the identification of an efficacious synthetic strategy for the bicyclic proline moiety has been essential for the development of a scalable pathway that ultimately led to the commercial process of the industrial synthesis of the drug. Multiple strategies have been reported during the years for the synthesis of this moiety, but, after evaluating all the possible approaches to afford the bicyclic [3.1.0] proline building block, researchers at Pfizer focused on the cyclopropanation of olefine 17 (Scheme 2) [19].

This strategy starts from commercially available compound 16, which is inexpensive and already contains the correct stereocenter required for 8. Compound 16 is converted into its derivative (17) following a strategy that was already reported on a multigram scale. The key step in the synthesis of the bicyclic [3.1.0] proline building block is the cyclopropanation of 17 (Scheme 3). Pfizer’s researchers envisioned that this reaction could proceed with good enantioselectivity due to the steric hindrance provided by the gem-dimethyl group. Methods for the gem-dialkyl cyclopropanation of inactivated or weakly activated olefins reported in the literature were not very effective, until 2018, when Werth and Uyeda reported the cobalt-catalyzed dimethylcyclopropanation of olefins, where they employed a cobalt precatalyst in a complex with pyridine diimine (PDI) ligand L1 (L1CoBr₂), 2,2-dichloropropane as a carbene equivalent and zinc as a reductant [22]. Pfizer thus decided to focus on the development and optimization of a scalable process for the synthesis of 8 using Co-catalyzed dimethylcyclopropanation as the key reaction. Primarily, it is important to note that L1 was used as aligand in the initial campaign to adhere to the stringent project timeline. However, alternative ligands could be explored for further process optimization. Co-catalyzed cyclopropanation required the use of a zinc halide, as zinc salts seem to accelerate the reduction of the cobalt catalyst L1CoBr₂ to the active species L1CoBr. Following screening, zinc bromide was initially employed, but it faced reproducibility issues and handling difficulties owing to the hygroscopic nature of the salt. To encounter the need of an operationally friendly activation procedure and facile workup, the zinc halide was generated in situ, in order to eliminate the problems related to the direct handling of the salts. A comparison of zinc bromide, zinc iodide, and iodine revealed that the rate of cyclopropanation was higher in the presence of the iodide counteranion, but iodine resulted in a higher conversion than zinc iodide. Based on this observation, Pfizer’s researchers moved forward with iodine as an activator for commercial production.

While scaling up the process from a 100 g scale to kilograms, a reduction in conversion rates from 95% to 60–70% was observed. Suspecting catalyst instability as the cause of the conversion drop, they concentrated their efforts on gaining a deeper understanding of the process. Their investigations revealed that catalyst degradation started immediately after the introduction of the 2,2-dihalopropane starting material into the reactor. They also observed that, during the kilogram-scale reactions, zinc was not uniformly suspended in the reactor. The instability of the catalyst was therefore attributed to the rate of regeneration of the reduced form of the catalyst, which was lower than the rate of oxidative addition of 2,2-dihalopropane. This problem was due to a low stirring rate. In fact, after increasing the agitation rate in the reactor, they observed a huge increase in the conversion. Even if 2,2-dichloropropane was initially used by Pfizer as the starting material, its difficult obtainment in big quantities led to the decision to move forward with 2,2-dibromopropane as the starting material for the kilogram-scale process. One of the challenges that they had to face with the use of this reactant was the more rapid degradation of the catalyst, with respect to that observed with 2,2-dichloropropane. The problem was solved by reducing the rate of addition of 2,2-dibromopropane in the reactor, in order to allow the Co catalyst to convert to its reduced form and to avoid the accumulation of 2,2-dibromopropane in the reactor. After optimizing the cyclopropanation step of the synthesis, they focused on the deprotection of 18 and formation of 8. The main problem was the purification of the product, which was addressed by crystallizing 8 and performing the final deprotection and isolation directly from the crude mixture after an acidic workup (Scheme 4).

Compound 18 was then treated with HCl in THF leading to the removal of the Boc group and finally to the crystallization of 8·HCl at a high purity, with a 50−60% isolated yield over two steps (i.e., cyclopropanation and deprotection). The addition of Methyl tert-butyl ether (MTBE) allowed us to improve the solubility of 8·HCl, increasing the yield to >70%. The overall process was then executed on a 310 kg scale to produce 8·HCl with a 73% yield over two steps.

3.2.2. Synthesis of the Western Fragment

As already mentioned in the previous paragraph, a key intermediate in the synthesis of Nirmatrelvir is the western fragment 13, whose synthesis has been described by Algera et al. [19]. The discovery of a synthetic route for 13 led to the production of Nirmatrelvir in 4 steps in the first ever kilogram-scale manufacture of the drug shown in Scheme 5, which was further developed into the commercial route.

The western fragment 13 was eventually obtained from 19 and 20 through a convergent approach consisting of an amidation step.

Compound 19 (Scheme 6) is obtained through hydrolysis from the bicyclic [3.1.0] proline derivative 8·HCl, bearing a methyl ester moiety. The high solubility of both 19 and its salts in aqueous conditions made the isolation of 19 challenging, so hydrolysis was performed differently by dissolving 8·HCl in a THF and water mixture, followed by treatment with LiOH causing the precipitation of the lithium carboxylate 19·Li (Scheme 6).

19·Li was used for the screening of the amidation reaction with 20, obtained from the reaction of commercially available L-tert-leucine with ethyl trifluoroacetate and sodium methoxide in methanol (Scheme 7). Following an aqueous work up, compound 20 was further purified by recrystallization from heptane (Scheme 7).

An initial screening of amidation conditions showed that the treatment of 20 with 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) and N-hydroxy succinimide produced a succinimidyl ester, which could react with 19·Li to produce 13 with good yields and high stereochemical purity. So, after high-throughput screening to optimize the process and a laboratory scale-up, it was found that the use of 4-Toluenesulfonyl chloride (TsCl) in combination with 4-Dimethylaminopyridine (DMAP) in THF as the solvent afforded 13 in a high yield with minimal impurities (Scheme 8). These conditions also presented advantages like the availability and cost of TsCl and the ease of using THF in industrial plants. DCM was not further evaluated due to its low environmental compatibility.

The reaction was performed by dissolving 20 in THF. TsCl and DMAP were then added to activate 20, before the introduction of 19·Li to avoid the tosylation of 19. The reaction was then quenched with HCl and brine to facilitate the aqueous workup. The isolation of the reaction product directly from THF was difficult, but it could be crystallized in iPrOAc with the addition of heptane as an antisolvent. Further process development adjustments were made on both the amidation reaction and the hydrolysis of methyl ester 8. In particular, while the lithium salt of 19 was simple to obtain, it was not suitable for a scale-up. In fact, 19·Li was found to be rich in LiCl impurities and hygroscopic. For these reasons, different alternatives were screened. In the end, the best solution appeared to be the sodium salt that was chosen for ongoing development, thanks to its more favorable physical characteristics compared to the lithium salt and because it could be easily obtained replacing LiOH with NaOH in the hydrolysis of 8. The conditions tested for the preliminary attempts of the amidation process with 19·Na were the same identified for the reaction with 19·Li. However, despite the effectiveness of the process, expensive and undesirable reagents, like DMAP, and the solvent exchange from THF to isopropyl alcohol (IPA) were required. In addition, the preferable use of liquid reagents led to the replacement of TsCl with Methanesulfonyl chloride (MsCl), which led to a complete conversion without the use of DMAP, but using only Triethylamine (TEA). Scheme 9 shows the optimized conditions used for the conversion of 19·Na to 13.

3.2.3. Synthesis of the Eastern Fragment

Another key building block of Nirmatrelvir is represented by the primary amide 14·HCl, also referred to as the eastern fragment, obtained via the aminolysis of 24·TsOH [20]. The first-generation synthesis of 14·HCl (Scheme 10) provides 24·TsOH starting from the diastereoselective alkylation of the lithium dianion of Boc-dimethyl glutarate 21. The alkylation using bromoacetonitrile resulted in the predominant formation of a single diastereomer of nitrile 22. The nitrile group was then subjected to reduction, leading to the formation of lactam 23. Subsequent treatment with p-toluenesulfonic acid facilitated deprotection, yielding 24·TsOH as a crystalline intermediate, thus obviating the necessity for column chromatography purification. Subsequently, 24·TsOH was subjected to Boc-protection to yield 23, which was isolated and reacted with ammonia, resulting in the formation of the aminolysis product 25. Treatment of this intermediate with hydrochloric acid in IPA allowed the deprotection of the amine, yielding 14·HCl.

Even if this synthetic strategy was suitable for the large-scale production of 14·HCl, researchers at Pfizer recognized the opportunity to streamline the process. In particular, the main problems with the first-generation route were: (i) the need for protection and deprotection steps to reduce the formation of side products; (ii) the formation of 23 twice in the synthesis, with the need of isolating the intermediate the second time it appeared, resulting in additional processing time and waste; and (iii) the fact that 24·TsOH was not suitable for direct aminolysis due to the sluggishness of the reaction with ammonia and the production of different impurities (Scheme 11). Moreover, the obtained 14·HCl still contained residual traces of isopropyl alcohol, contributing to some of the remaining mass balance. The use and storage of 14·HCl also presented safety risks due to the breaking of its agglomerates, which led to the release of flammable solvents. For all these reasons, researchers started to look for an alternative synthetic strategy to afford 14·HCl.

Given that the supply chain for 24·TsOH had already been established at Pfizer, the optimization of the route began with the exploration of conditions enabling the direct aminolysis of 24·TsOH. The objective was to achieve high yields, minimize impurity formation, and reduce reaction times. It was envisioned that the use of a Lewis acid during the reaction could accelerate aminolysis reaction times, increasing the reactivity of the ester toward substitution and preventing the formation of pyroglutamate 27 and dimer 28 (Scheme 11). An evaluation of various Lewis acids was ultimately performed, leading to the implementation of MgSO₄ as the preferred reagent for this reaction, as it is an insoluble salt easily removable via the filtration of the reaction mixture. The reaction performed in the presence of MgSO₄ resulted in high yields and minimal formation of the hydrolysis by-product 26.

The next challenging task involved devising a method for isolating 14·HCl after Boc-deprotection. Analysis of the reaction mixture indicated the presence of the correct product, together with magnesium ions and tosylate as impurities that required removal. The decision to isolate HCl salt instead of tosylate salt was due to the difficulties faced during the attempt to obtain 14·TsOH as a crystalline solid. With this purpose, DMF was found to be an efficient solvent to purge undesired by-products, including magnesium chloride and p-toluenesulfonic acid, while facilitating the high recovery of 14·HCl. Moreover, to shift the equilibrium from 14·TsOH toward 14·HCl, an excess of HCl was used in the reaction. Furthermore, HCl in IPA could be used without the risk of a high level of residual Mg (OTs)₂, which is poorly soluble in this solvent. So, after aminolysis, filtration through Celite was employed to remove insoluble materials and methanol was exchanged with DMF to solubilize the side products. HCl in IPA was added to prompt the crystallization of 14·HCl, and finally dichloromethane was introduced as an antisolvent to enhance product recovery (Scheme 12, path A). Although this process offered notable advantages, it also presented several challenges that needed to be addressed. Initially, 14·HCl isolated through this method contained organic impurities and residual DMF, which posed difficulties in effective removal, due to the stability concerns of 14·HCl at elevated temperatures required for solvent evaporation. Furthermore, the use of DCM as an antisolvent was undesirable for environmental reasons. An investigation for alternative solvent systems was pursued, eventually leading to the use of methanol and water as the process medium—methanol was found to be a potential candidate because it was the same solvent used in the previous aminolysis reaction. Since 14·HCl remained moderately soluble in methanol, an antisolvent was required to increase product recovery. Isopropyl alcohol has been evaluated as the best antisolvent to improve the recovery of solids. Crystallization using MeOH and water as the process solvents and IPA as the antisolvent resulted in the crystallization of 14·HCl at high purity and with minimal residual solvent (Scheme 12, path B).

Further optimizations after the scale-up of the process led to reducing hold times after HCl addition in IPA and a lower temperature (10 °C) for the isolation of 14·HCl. These modifications were implemented because the stability of 14·HCl appeared to be dependent on the hold time and temperature.

3.3. Application of Flow Chemistry in Nirmatrelvir Synthesis

The industrial approach used by Pfizer for Nirmatrelvir’s synthesis involved expensive and challenging chemicals or required extended reaction times, which heightened the risk of side-product formation due to epimerization. Flow chemistry offers a promising approach for the synthesis of bioactive compounds, like Nirmatrelvir, due to its numerous advantages over traditional batch synthesis methods [23]. Veeramani’s group figured that the formation of the amide bond between the western fragment and lactam 29, which already embodies the -CN moiety, would be more efficient; therefore, they directed their efforts toward synthesizing fragment 29 (Scheme 13).

The main difficulty of this pathway was the obtainment of 29 at sufficient purity to respect the quality requirements defined for advanced intermediates taken forward to the final API.

The optimized synthesis of 29 is depicted in Scheme 14 and starts with the conversion of commercially available ester 23 to carboxamide 25 using a concentrated solution of ammonia in methanol. This solution was prepared by passing ammonia gas through a pre-cooled solution of 23 in methanol. Dehydration of 25 was initially attempted with trifluoroacetic anhydride (TFAA), but the formation of heavy suspension in the reaction mixture led us to prefer the use of propanephosphonic acid anhydride (T3P), which allows the obtainment of the nitrile-bearing fragment 29. The Boc deprotection step was challenging because of the formation of the impurity 14·HCl, obtained from the hydrolysis of the cyano group to an amide. Several reagents and reaction conditions were attempted with the aim of minimizing the hydrolysis of 29·HCl to 14·HCl and obtaining 29 in a good and persistent yield. The best conditions resulted in the use of aqueous HCl in THF. Purification of 29 included the simple precipitation of hydrochloride salt (29·HCl) from the reaction mixture with > 95% purity, even if 1–2% of the amino-amide 14·HCl persisted as an impurity.

Having obtained 29 with sufficient purity, T3P was used to carry out the amidation of 13 with 29 to produce 1. The main advantage of using T3P for amide bond formation is that the small quantity of impurity 14 was not detrimental to the final yield. In fact, the amidation of 13 with 14 produced impurity 15, which then underwent dehydration by the same T3P present in the reaction mixture to produce the desired final product. Nirmatrelvir (1) was obtained with a ~97% yield overall (Scheme 15)

The main advantage of this procedure is that it avoids the use of the Burgess reagent in favor of T3P, which presents low toxicity, long shelf-life stability, and easy handling. Veeramani’s group further improved the synthetic process, performing the final dehydration of impurity 15 to 1 using continuous flow chemistry. The main advantage of flow reactors, in comparison to batch reactors, is their more favorable surface-to-volume ratio, which enhances heat and mass transfer to obtain desired products with higher yields and better selectivity. Flow chemistry also allows us to easily perform sensitive reactions, to have better control of exothermic reactions, and to easily scale-up the process for large-scale production.

Specifically, the enhancement of the dehydration stage involved preventing the liquid stream from evaporating. This allowed us to use higher temperatures beyond the boiling points of the solvents, consequently reducing reaction durations. Figure 5 shows the schematic representation of the flow experimental setup that was employed.

A complete conversion of carboxamide 15 to Nirmatrelvir 1 was obtained when the reaction was performed at 100 °C for a residence time of 30 min, using 2.5 equiv of N,N-Diisopropylethylamine (DIPEA) and 2.0 equivalents of T3P. The yield and purity obtained are comparable to those obtained by the conventional batch process.

It is possible to conclude that the continuous-flow process can create advantages for the synthesis of Nirmatrelvir, as it allows us to reduce the reaction times from the 12–16 h necessary for the batch process to just 30 min. Moreover, the flow process allows us to avoid the use of high temperatures and extreme pressures. Currently, more studies are being performed to obtain Nirmatrelvir from the reactions of 13 and 14 through a T3P-mediated tandem amidation dehydration in flow.

3.4. Multicomponent Approach

An alternative way to improve the synthetic process to obtain Nirmatrelvir is the use of multicomponent reactions. This class of reactions is being increasingly appreciated in medicinal chemistry, because it enables us to synthesize biologically active molecules in a more sustainable, cost-effective, and rapid way.

Preschel et al. proposed the synthesis of Nirmatrelvir based on the multicomponent assembly of N-trifluoroacetyl-tert-leucine 33, chiral bicyclic imine 31, and isocyanide 32 through a diastereoselective Ugi-type three-component reaction (U-3CR) (Scheme 16) [24].

Chiral imine 31 was prepared by the monoamine oxidase N (MAO-N) biocatalyzed oxidation of amine 34. This reaction proceeded with an enantiomeric excess higher than 99% (Scheme 17). Biocatalysis has become a relevant tool in the pharmaceutical industry, because it can provide a more sustainable and efficient method to produce APIs and pharmaceutical intermediates. Enzymes are in fact renewable and biodegradable materials, they can react in aqueous conditions and they have high chemo- and enantioselectivity, therefore avoiding protection or deprotection steps.

The isocyanide moiety was obtained from the Boc-protected amino ester 23 (Scheme 18), which is commercially available. Compound 23 was converted into primary alcohol 35 using LiBH_4. Compound 35 was then protected as a benzoate ester to produce 36. Boc-cleavage, followed by the formylation of the resulting amine, produced formamide 37, which was then dehydrated with triphosgene (0.66 equiv) in the presence of TEA (10 equiv) in DCM at −78 °C to produce isocyanide 32 as an isolable, stable, crystalline solid (Scheme 18). The same results can alternatively be obtained by the treatment of 37 with TFAA (1.9 equiv), TEA (8 equiv), and THF at 0 °C for 30 min. The obtained crystals were analyzed through X-ray diffraction; thus, the structure and absolute configuration of compound 32 were confirmed.

The three fragments were combined together in a U-3CR, as reported in Scheme 19. The enantioselectivity of the reaction is driven by imine 31, because the gem dimethyl moiety shields the concave face of the bicyclic system. In a first attempt, isocyanide 32, imine 31, and carboxylic acid 33, obtained from the trifluoroacetylation of L-tert-Leucine, were reacted in DCM to produce 38 as a 5:1 mixture of diastereomers. The low enantioselectivity was initially connected to incomplete stereo-induction at the newly formed stereocenter. After further examination, it was found that the procedure used to obtain 33 led to some erosion of its stereochemistry. A new attempt was therefore made using commercially available acid 33 (>98 e.e.), which reacted with imine 31 and isocyanide 32 to produce compound 38 in a 68% isolated yield and > 25:1 dr. Methanolysis of the benzoate ester produced primary alcohol 39, which was then deprotected and oxidized with a one-pot procedure involving PhI (OAc)₂/TEMPO with amonium acetate as the nitrogen source, to produce Nirmatrelvir in an 83% yield (Scheme 19).

It was observed that, by streamlining the synthetic pathway, compound 32 could be obtained without intermediate purifications and with minimal mass loss. The U-3CR reaction was performed in methanol, allowing a one-pot combination of the Ugi reaction and of the saponification step. The product was directly used for the final step to produce Nirmatrelvir in a 70% yield, which is higher than the one obtained over the three last steps with intermediate purifications (48%) (Scheme 20).

3.5. Sustainable Synthesis

The immediate high demand of Nirmatrelvir during the COVID-19 pandemic led to environmental concerns connected to the industrial process, which makes large use of peptide coupling reagents and organic solvents. Since the beginning, it has been clear that there was a need to optimize the existing synthetic pathways to Nirmatrelvir to make them greener and cheaper, two characteristics that are also fundamental for the distribution of Paxlovid in the Global South. For these reasons, many efforts were put into the development of a new industrial process that could have a lower environmental impact while being economically attractive. An interesting approach to the synthesis of Nirmatrelvir has been suggested by Kincaid et al., with a seven-step, three-pot route, reaching the targeted product in a 70% overall yield [25]. The key points of this sustainable approach are (i) the amide bond formation steps, obtained with green technologies that allow us to avoid the use of common coupling reagents (e.g., HATU, DCC, COMU, etc.), and (ii) the dehydration step of the primary amide into the corresponding nitrile without using the Burgess reagent and chlorinated solvents, which are employed in the parental Pfizer procedure. The entire process is reported in Scheme 21.

The sequence features a one-pot thioesterification/amide bond formation using di-2-pyridyldithiocarbonate (DPDTC). This reagent activates carboxylic acids and, beyond avoiding the use of classical coupling reagents, it also prevents epimerization while allowing the easy removal of the 2-mercaptopyridine by-product with an aqueous alkaline solution. Furthermore, the produced 2-mercaptopyridine can be readily turned back into DPDTC, showcasing a crucial characteristic for a reagent employed in an environmentally friendly process. Intermediate thioesters 40 and 41 are stable and potentially isolable, simplifying potential large-scale use and manufacture. However, in this route, isolation and purification were not required. After the formation of thioester 40, in the presence of DMAP, the reaction mixture went through an in-flask treatment with 8 and N-Methylmorpholine (NMM) in EtOAc affording the desired amide intermediate 10. Compound 10 was not isolated, but directly dissolved in aqueous THF in the presence of LiOH. After the reaction, the mixture was neutralized with aqueous HCl and extracted with EtOAc obtaining 11 with a 78% yield over three steps. The resulting intermediate 11 did not require purification, since the small amounts of impurities present, primarily 2,2’-dipyridyldisulfide, had no synthetic consequences. Carboxylic acid 11 was subjected to a second thioesterification under the same reaction conditions previously discussed. Finally, the eastern fragment was introduced by reacting 41 with 29·HCl in EtOAc in the presence of NMM, affording 42 with a 94% yield over two steps. The subsequent N-Boc deprotection of 42 was performed using a concentrated solution of HCl/dioxane in MeCN. It seems that HCl is preferable to trifluoroacetic acid (TFA) as the latter leads to lower yields and is involved in the formation of multiple side products, including various materials resulting from epimerization. Solvent and an excess of HCl were removed in vacuo and the amine hydrocloride salt obtained was treated with trifluoroacetic anhydride (TFAA) and NMM to install the trifluoroacetamide moiety. TFAA and NMM were then removed via aqueous washes, affording Nirmatrelvir in a 95% yield over two steps with 95% purity confirmed by HPLC. A brief focus on the synthesis of reagent 29 is summarized in Scheme 22. Compound 29 was obtained from 23, a readily accessible methyl ester, which underwent conversion into the corresponding primary amide 25 utilizing a documented procedure [13]. At this stage, the primary amide underwent a dehydration step employing a more environmentally friendly approach compared to Pfizer’s use of the Burgess reagent with a chlorinated solvent. In fact, the nitrile-bearing intermediate has been obtained by applying technology based on “amide exchange” [26,27]. This technology uses commercially available fluoroacetonitrile as the sacrificial acceptor of water under palladium-catalyzed conditions. The treatment of 25 in these conditions produced 30, with a 93% yield. Crystallization of 30 was needed to eliminate impurities and by-products coming from the use of fluoroacetonitrile. Subsequent N-Boc deprotection was achieved using HCl in the presence of a sacrificial nitrile, such as MeCN, to reduce undesired competitive hydrolysis and, at the same time, performing the azeotropic removal of residual water in 30 using recoverable toluene under a high vacuum. These two precautions minimized the formation of undesired by-products 29a and 29b affording 29 as HCl salt with a 95% yield and only 3% hydrolysis product.

This convergent route to Nirmatrelvir has been accomplished in a 70% overall yield, a considerable improvement on the 48% reported by Pfizer for the first-generation process. After achieving these interesting results, the process to obtain Nirmatrelvir was further streamlined. The initial seven-step strategy could in fact be reduced by two steps by eliminating the necessity of removing the protecting group from the starting N-Boc-L-tert-leucine (Scheme 23) [28]. To achieve this, the trifluoroacetamide derivative of L-tert-leucine (20) was used as the starting material. This approach allowed us to perform with the same reagent both the deprotection step and the subsequent insertion of the trifluoroacetamide fragment. The main issue encountered when using the trifluoroacetamide derivative of L-tert-leucine 20 as starting material was the challenging generation of the corresponding thioester. This thioester was essential for the reaction with the sodium salt of bicyclic proline 10·Na to yield intermediate 13. Following the low yields obtained using the corresponding HCl salt of 20, attributed to the compound’s instability, several amide coupling reagents were screened. This resulted in the adoption of propanephosphonic acid anhydride (T3P) as the coupling reagent. T3P was chosen due to its efficiency, practicality in usage and workup, absence of stereochemical issues regarding diastereomer formation, environmental considerations, and economic accessibility. Therefore, carboxylic acid 20 was activated by T3P in the presence of DIPEA at −10 °C over a 1 h period. Bicyclic proline sodium salt 10·Na was then added portionwise at 0 °C, followed by dilution with anhydrous EtOAc. The resulting mixture was then warmed slowly to rt with continuous stirring for 20 h. Aqueous acid was added to remove side products generated from T3P. After removing the aqueous medium and concentrating the ethyl acetate, 13 was obtained in high yields (from 93% to 96% using a concentration of 0.5 M in EtOAc). After obtaining 13, this intermediate had to be reacted with 29·HCl to afford Nirmatrelvir (1). T3P was also evaluated for this amide bond formation: 29·HCl carries a primary amine with high reactivity, so the mixing of all reagents in a single reaction flask in the presence of DIPEA (T < 0 °C) with vigorous stirring led to complete starting-material conversion within three hours, as evidenced by TLC analysis. After this time, the reaction was diluted with EtOAc and subjected to an aqueous wash using 1 M HCl. Solvent removal in vacuo produced crude 1, which was then exposed to EtOAc/MTBE (1:10). Subsequent solvent removal provided solvated Nirmatrelvir MTBE.

In conclusion, this optimization campaign of the previous seven-step, three-pot route that produced Nirmatrelvir with a 70% overall yield resulted in a streamline of the process and the development of a three-step, one-pot sequence that allowed us to create Nirmatrelvir as MTBE solvated with an overall 64% yield.

4. Industrial Scale-Up of Nirmatrelvir Synthesis

Pfizer’s proposed synthesis of Nirmatrelvir offers several opportunities for optimization toward industrial-scale production. Key areas for improvement include addressing cost-intensive components, such as starting materials 8, 9, and 13; the coupling reagent HATU; the Burgess reagent used for amide dehydration; and enhancing the yield of trifluoroacetylation to obtain compound 13. Additionally, the initial step required purification via column chromatography.

Appasaheb et al. made a screening of several coupling reagents and reaction conditions to improve the yield of the coupling reaction between the bicyclic proline and N-Boc-L-tert-leucine, without using the expensive HATU as the coupling reagent [29]. The most favorable outcomes were observed when employing a combination of N,N′-Dicyclohexylcarbodiimide (DCC) and Hexafluorophosphate Benzotriazole Tetramethyl Uronium (HBTU); however, the use of HBTU was ruled out due to safety concerns. Employing 2.1 equivalents of DCC, 0.5 equivalents of 1-Hydroxy-7-azabenzotriazole (HOAt), and 3 equivalents of NMM in DCM compound 10 was obtained with a 88% yield. Proline 8, identified as a primary cost contributor, is deliberately utilized as the limiting reagent in this process, with the reaction proceeding for 24 h at room temperature. The use of DMF as a co-solvent allowed us to increase the yield only from 88% to 90%, so the improvement was not considered significant and therefore not necessary.

To avoid column chromatography, intermediate 10 on the crude mixture was used directly for the successive saponification step. After an aqueous workup, intermediate 11 was precipitated from a solution of acetonitrile by the addition of water. The best results on this telescoped procedure allowed us to obtain 11 with a 75% yield over two steps and with 73% purity confirmed by HPLC (Scheme 24).

After the deprotection of the amine, the yield of the trifluoroacetylation step was improved up to 98% using TFAA in MeOH, instead of trifluoro ethyl acetate, with an excess of DIPEA (Scheme 25).

The dehydration of the amide functional group using the Burgess reagent presented several challenges concerning cost, availability and reagent stability during storage, as well as the removal of by-products. Additionally, the removal of dichloromethane (DCM) in the amidation reaction was recommended for environmental consideration. After the screening of different conditions, Appasaheb et al. found TFAA and T3P to be similarly effective. Increasing the quantity of T3P, some impurities formed in the reaction mixture, but they were removed through crystallization from MTBE/EtOAc. TFAA was found to provide better results and allowed us to obtain Nirmatrelvir (1) in a 90% overall yield. A screening of the reaction conditions was made and the best results were obtained using NMM as a base. Crystallization led to an MTBE solvated product with a 85% yield. MTBE-free crystals were obtained through crystallization in heptane/iPrOAc with the same conditions reported in the Pfizer procedure (Scheme 26).

Due to the ability of TFAA to both dehydrate the amide and produce the trifluoroacetylated product, a one-pot process was attempted. The first attempt to obtain Nirmatrelvir’s MTBE solvated product in a telescoped manner led to the obtainment of Nirmatrelvir with a 63% yield (Scheme 27).

The possibility to perform the final amidation dehydration step through a telescoped process was studied in more detail by Algera et al., starting from the new screening of different reaction conditions for both the amidation and dehydration steps [30].

The coupling reaction involving 13 and 14 was optimized using EDCI with HOPO in methyl ethyl ketone (MEK) and T3P with N-methylimidazole (NMI) in MeCN, which yielded the most favorable outcomes and were thus subjected to further investigation. These reaction conditions exhibited compatibility with various solvents and bases. T3P facilitated a high conversion rate and, as previously mentioned, it was also effective in the dehydration process. However, notable quantities of impurities 43 and 44 were generated under these conditions (Figure 6). Additionally, the use of MeCN as the solvent posed challenges for the subsequent aqueous workup step.

This led to the selection of EDCI and HOPO in MEK as the best conditions for the amidation step.

For the amide dehydration step, Burgess-type reagents (Figure 7) were excluded for atom economy reasons, as three equivalents of reagent were needed.

After the screening of different dehydration conditions, T3P and NMI in MeCN and TFAA and NMM in iPrOAc were found as the most promising. T3P caused the formation of impurity 44 reported in Figure 6, and was found to be more inclined toward the epimerization of the nitrile stereocenter; so, TFAA was selected for use in the dehydration step.

The telescoping of the amidation and dehydration steps was favored by the elimination of MTBE. This also allowed us to increase the reproducibility. The solvents for the two steps were then evaluated, considering the need for an aqueous workup between the amidation and the dehydration steps and the preferability to use a single solvent in the two steps. MEK allowed us to control moisture in the dehydration step, important for the sensibility of TFAA to water, but the dehydration step was slow, creating concerns regarding TFAA accumulation. iPrOAc favored the dehydration step, but moisture control was more difficult. Moreover, 14 is poorly soluble in iPrOAc. Considering all these aspects, iPrOAc was finally chosen as the preferred solvent.

The main impurities formed during the telescoped sequence were acylureas 50 and 51 (Figure 8), caused by EDCI-promoted amidation, but their formation could be prevented by increasing the amount of HOPO used in the reaction (0.75 equiv). Basic conditions could lead to the rearrangement impurity 52 and subsequently 53. After the dehydration step, a small quantity of starting material 15 was still present. The epimerization that caused the formation of 44 could be avoided by controlling the pH and temperature. TFA adducts 54 and 55 formed during the dehydration step were accurately controlled defining plant parameters (Figure 8).

The optimized procedure reported in Scheme 28 was first tested at the laboratory scale with 5 g (60% yield), then 100 g (65% yield), and finally 53 kg in the manufacturing plant (75% yield, 95 kg of 1). The increase in the yield with bigger quantities is due to the better control of moisture.

5. Variants and Modifications to Nirmatrelvir

Even after the conclusion of the COVID-19 pandemic, a strong demand persists for new therapeutics targeting SARS-CoV-2, due to the development of drug resistance (ACS central science, 2023, 9.8: 1658–1669) [31]. Moreover, numerous global drug discovery initiatives have been undertaken to develop treatments for post-infection scenarios and to pre-emptively address potential future pandemics. The literature extensively documents endeavors aimed at enhancing the potency of Nirmatrelvir through structural modifications or synthesizing compounds with improved properties compared to existing patented and commercial drugs. To this purpose, notable findings in this regard have been reported, and we decided to include three examples of chemical modifications of Nirmatrelvir, made in order to improve potency and reduce drug resistance.

One of the proposed modifications is the electrophilic warhead. In 2024, Elshan et al. described the discovery of CMX990 (57, Figure 9), a potent SARS-CoV-2 M^pro inhibitor bearing a novel trifluoromethoxymethyl ketone warhead [32]. The authors explored SAR across all regions of this peptidomimetic in order to reach the final product, now going through Phase 1 clinical trials. This compound showed an excellent balance of physicochemical and ADME (absorption, digestion, metabolism, and excretion) properties, allowing for high potency in target engagement as well as favorable PK properties. Compond 57 was validated in a biochemical assay for SARS-CoV-2 M^pro, where it was demonstrated that the compound acts as a reversible inhibitor (IC₅₀ = 23.4 nM). Compond 57 also exhibited the inhibition of replications of eight SARS-CoV-2 variants, including alpha, delta, and omicron.

In a recent study, Ghosh et al. focused on modifying P1 and P4 of Nirmatrelvir to enhance inhibitory activity [33]. Their goal was to improve binding within the S1 subsite and fill the hydrophobic pocket in the S4 subsite of the M^pro. Various alkyl, aryl, halogenated acetamide, carbamate, and urea derivatives were evaluated as P4 ligands, combined with five-membered and six-membered lactams as P1 ligands. The findings show that P4 amide derivatives are significantly more potent in comparison to carbamate and urea ones. Notably, several inhibitors with a six-membered lactam ring and halogenated P4 amides showed a marked increase in inhibitory and antiviral activities compared to Nirmatrelvir. Specifically, compound 58 (Figure 9), featuring a P1 six-membered lactam and P4 trifluoro acetamide groups, demonstrated a three-fold improvement in antiviral activity in VeroE6 cells and retained efficacy against variants such as delta and omicron, comparable to remdesivir. High-resolution X-ray crystallography of 5e-bound SARS-CoV-2 M^pro revealed that the six-membered lactam ring forms stronger hydrogen bonds and fills the hydrophobic pocket more effectively than the five-membered lactam ring in Nirmatrelvir. These results highlight the potency-enhancing effects of the P1 six-membered lactam, providing valuable insights into the optimization of Nirmatrelvir derivatives and related compounds for M^pro inhibition.

Another interesting modification to Nirmatrelvir’s P4 residue was made in 2023 by Brewitz et al., who decided to replace the nitrile warhead with an alkyne group [34]. Nirmatrelvir-derived alkyne 59 (Figure 9) was found to inhibit M^pro ~5-fold less efficiently than Nirmatrelvir (1) with an IC₅₀ ~0.14 μM, but its inhibition of SARS-CoV-2 progression in cells is ~11-fold less efficient than Nirmatrlevir (EC50 ~25.7 μM) (Figure 9). Compound 59 was also tested through MTT cells’ viability assay and was found to be non-cytotoxic, even if further studies should be performed on the metabolism of alkynes in animals. The alkyne warhead was also functionalized with aryl groups that could fit into the S’ pocket, but all the synthetized derivatives showed a ~50-fold to ~80-fold decreased inhibition of M^pro compared to that of 59. The functionalization of the alkyne with a –CF₃ group instead presented interesting results: compound 60 (Figure 9) showed, in fact, a similar potency to 59 with an IC₅₀ ~0.22 μM, but the inhibition of the progression of SARS-CoV-2 in infected VeroE6 cells was ~5-fold more efficient than that of 59 (EC₅₀ ~5.1 μM) and only ~2-fold less efficient than Nirmatrelvir (1) (EC₅₀ ~2.2 μM). Even if both 59 and 60 are less efficient than Nirmatrelvir, this study shows that appropriately functionalized alkynes have potential for the inhibition of SARS-CoV-2 M^pro and SARS-CoV-2 progression in cells.

6. Conclusions

The global urgency to counteract the COVID-19 pandemic has provided significant advancements in antiviral research, leading to the discovery of Nirmatrelvir as a potent inhibitor of SARS-CoV-2 M^pro. Combined with ritonavir and marketed as Paxlovid™, this compound represents the first oral treatment for COVID-19. This review has described various synthetic approaches for the synthesis of Nirmatrelvir from Pfizer’s original strategy to innovative and more sustainable approaches, such as flow chemistry and multicomponent reactions. Moreover, several examples of Nirmatrelvir modifications, produced to overcome drug resistance, have been discussed. Looking forward, the synthesis and optimization strategies outlined in this review not only enhance the production efficiency of Nirmatrelvir, but also pave the way for future therapeutic interventions. By adopting more sustainable and scalable synthetic methods, the industry can better respond to current and future pandemics. Additionally, the innovative strategies discussed here can serve as a foundation for developing new antiviral agents with improved efficacy and safety profiles. Continued research and developments in this area hold promise for expanding known antiviral therapies.