key: cord-0996619-blvdazxc
authors: Murtuja, Sheikh; Shilkar, Deepak; Sarkar, Biswatrish; Nayan Sinha, Barij; Jayaprakash, Venkatesan
title: A short survey of Dengue protease inhibitor development in the past 6 years (2015-2020) with an emphasis on similarities between DENV and SARS-CoV-2 proteases
date: 2021-09-20
journal: Bioorg Med Chem
DOI: 10.1016/j.bmc.2021.116415
sha: cb9e0cd21303506aef3d39ddf8d416f0e71b0793
doc_id: 996619
cord_uid: blvdazxc

Dengue remains a disease of significant concern, responsible for nearly half of all arthropod-borne disease cases across the globe. Due to the lack of potent and targeted therapeutics, palliative treatment and the adoption of preventive measures remain the only available options. Compounding the problem further, the failure of the only dengue vaccine, Dengvaxia®, also delivered a significant blow to any hopes for the treatment of dengue fever. However, the success of Human Immuno-deficiency Virus (HIV) and Hepatitis C Virus (HCV) protease inhibitors in the past have continued to encourage researchers to investigate other viral protease targets. Dengue virus (DENV) NS2B-NS3 protease is an attractive target partly due to its role in polyprotein processing and also for being the most conserved domain in the viral genome. During the early days of the COVID-19 pandemic, a few cases of Dengue-COVID 19 co-infection were reported. In this review, we compared the substrate-peptide residue preferences and the residues lining the sub-pockets of the proteases of these two viruses and analyzed the significance of this similarity. Also, we attempted to abridge the developments in anti-dengue drug discovery in the last six years (2015–2020), focusing on critical discoveries that influenced the research.

Nearly 390 million people globally are at risk of developing the arthropod-borne viral disease dengue. 1 Dengue virus (DENV) 2 belongs to the genus flavivirus of the Flaviviridae family. 3 The genus flavivirus also includes pathogenic viruses like West Nile virus (WNV), Yellow Fever virus (YFV), and Japanese Encephalitis virus (JEV). 4 There are four distinct but closely related serotypes of dengue viruses: DENV 1, DENV 2, DENV 3, and DENV 4. Further, each serotype can be sub-classified into genotypes and strains. 5 The virus is primarily transmitted by the arthropod vector Aedes aegypti and, to a lesser extent, by Aedes albopictus. 6 It is responsible for varying degrees of clinical manifestations in the infected individuals, including mild flu-like symptoms to severe symptoms such as dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS). Failing to attend DHF and DSS cases immediately can prove to be fatal. 7 Fortunately, less than 1% of dengue patients develop DHF and DSS. 8 Based on the incidence of DHF and DSS in the case of dengue infection, the severity of the four serotypes is found to be in the following order (DENV 2>DENV 1>DENV 3>DENV 4). DENV 4 is rarely fatal. 9 Infection with one serotype does not necessarily offer cross-immunity to other serotypes. Further, any subsequent infection can develop into severe dengue, commonly referred to as "antibody-dependent enhancement." 10, 11 Over the years, dengue has severely affected populations in the tropical and subtropical regions of the world. 12 Some studies investigating the prevalence of dengue have identified 128 countries at risk of dengue 13 epidemics, among which Asia alone accounts for 70% of cases. The year 2019 recorded the highest ever reported dengue cases worldwide, with Afghanistan as the new entrant, witnessing dengue infection for the first time. In the year 2020, cases were on the rise in countries like Bangladesh, Brazil, Cook Islands, Ecuador, India, Indonesia, Maldives, Mauritania, Mayotte (Fr), Nepal, Singapore, Sri Lanka, Sudan, Thailand, Timor-Leste, and Yemen. 14 On the vaccine front, the current status is not highly encouraging.

The currently available DENV vaccine, Dengvaxia ® , launched by Sanofi Pasteur in 2015, has only been approved in 20 countries due to concerns about vaccinating seronegative patients. Seropositive and seronegative patients responded differently to the vaccine; the vaccinated seronegative patients were vulnerable to severe dengue contracting the first natural DENV infection. Currently, only individuals in the age group of 9-45 years and with at least one reported previous DENV infection are eligible for the vaccine. 14, 15 Hence, these limitations aggressively demand a parallel attempt at drug development. As our group has previously published the progress made in the development of DENV protease inhibitors. 16 In continuation of our previous work, in this review, along with a discussion on Dengue-COVID 19 coinfection, we have compiled the progress made in the last six years (2015-2020) toward the development of DENV protease inhibitors. We have broadly categorized our discussion into peptide inhibitors, small molecule inhibitors, inhibitors identified through a rational approach, inhibitors identified through modification of previously reported inhibitors, and drug repurposing. A section on patents and information on clinical trials has also been added. Wherever possible, we have discussed the evolution of potent molecules, reported mechanism of action of the inhibitors, and their interactions with the available crystal structures of DENV protease that demonstrated inhibitory activity. Further, the emergence of COVID-19 has stirred fresh concerns in tropical regions due to anticipated threats following a dengue co-infection.

While the data on this topic is relatively premature and inconclusive, we have touched upon Dengue-COVID-19 co-infections and outcomes in patients. This has also inspired us to draw two comparisons, one between the substrate-peptide residue preferences of DENV NS2B-NS3 protease (hereafter referred to as DENV NS2B-NS3 pro) and SARS-CoV 3CL pro and the other between various residues lining the sub-pockets of DENV NS2B-NS3 pro and SARS-CoV 3CL pro .

The DENV genome is approximately 11 kb long, single-stranded, positive-sense RNA, encoding a single polyprotein. The genome is processed into three structural proteins, namely, the capsid (C), envelope (E), and membrane (M) proteins and seven nonstructural proteins, i.e., NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 by the host proteases (Furin and signalase) and the two-component viral NS2B-NS3 protease ( Figure 1A) . [17] [18] [19] The DENV NS2B-NS3 pro (serine protease), which belongs to the trypsin superfamily, contains a catalytic triad in its active site formed by His51, Asp75, and Ser135 residues. 20 The N-terminal domain of NS3 houses NS3 protease, which, together with 40 residues (hydrophilic) of NS2B, 19 exhibits the serine protease activity. 21 The lack of this NS2B-NS3 interaction renders the NS3 protease less active or inactive. 19 Further, this NS2B-NS3 alignment forms the S 2 and S 3 sub-pockets in the protease active site. 22 Since viral replication and propagation depend on the activity of DENV NS2B-NS3 pro, the complex serves as an attractive target for the antiviral drug design against DENV. 23 Figure 1B shows the basic steps involved in the replication cycle of the DENV. 18, 24, 25 infection might have aggravated the severity of dengue. However, this contention needs further exploration as it was difficult to distinguish between the clinical features of Dengue and COVID- 19 . They reported that the patient showed some symptoms like prolonged fever, facial flushing skin, body-ache myalgia, arthralgia, erythema, retro-orbital eye pain, which were consistent with dengue. However, some of these overlap with the clinical features of COVID-19. Besides these, other symptoms, including thrombocytopenia and elevated liver enzymes, were also reported in both diseases. 26 Following this report, several other reports of co-infection have emerged across the globe. Epelbein et al. reported a case of coinfection in a traveler returning from France. 27 Further, Saddique et al. highlighted a case where similarity of symptoms was such that a proper distinction was only possible following a laboratory test. 28 These observations hint at the difficulty in distinguishing the clinical manifestation arising from the co-infection, thus highlighting the need for exploring the co-infections further.

The initial efforts in the search for DENV protease inhibitors were based on the residue sequences at various cleavage sites. Further, this approach was supported by the finding that NS2B-NS3 proteases of the Flavivirus family have a preference for substrates having dibasic residues (Lys or Arg) at the P1 and P2 site, and further residue preference at P1'site was a small amino acid (Gly, Ala, and Ser). 29 Table 1 presents the outcome of their work. SARS-CoV-2, which caused the COVID-19 pandemic, belongs to the family Coronaviridae and genus Betacoronoviruses. Coronaviruses are enveloped, single-stranded, positive-sense RNA viruses responsible for various diseases, including respiratory diseases in both animals and humans. [32] [33] [34] [35] The genome of SARS-CoV-2 is~29.9 kb in length, and it has 14 open reading frames (ORF's) encoding for 27 proteins. 36 This viral genome encodes two proteases, besides Papain-like protease (PL pro ), the Chymotrypsin-like cysteine protease, the 3C-like protease (3CL pro ) is considered to be the main protease (also called M pro ) responsible for polyprotein processing. In addition, among the coronaviruses, the sequence for 3CL pro is highly conserved. 37, 38 The fact that none of the human host cell proteases have such substrate specificity [39] [40] [41] further strengthens the rationale behind the design and development of 3CL protease inhibitors. The catalytic dyad of SARS-CoV 3CL pro is composed of Cys145 and His41, which is different from the serine protease catalytic triad (Ser-His-Asp). 42 Here, we compared the substrate-peptide residue preferences of both DENV NS2B-NS3 pro and the SARS-CoV 3CL pro of coronaviruses. Since these two proteases have already been established as potential targets, 23, 43 it would be interesting to see the outcome of this comparison. In 2010, Chuk et al. presented the substrate-peptide residue preferences of SARS-CoV 3CL pro ( Table 1) . From the table, it is evident that Val>Phe>Thr residues were favored at P5 position in that order, while small hydrophobic residues like Cys and Val were more favored at P4 position, and positively charged residues with a tendency to form β-sheet structure were more favored (Arg>Val) at the P3 position. Further, at the P2 position, hydrophobic residue without β branch was most favored, and the order of residue preference was Leu>Met>Phe. Moreover, β-branch residues like Ile and Val were least preferred. At the P1 position, Gln was most favored, with Gln>His>Met as the order of preference for this position. For the prime side residue preferences, P1' favored small residues; the order of preference was Ser>Ala>Cys>Gly. The P2' and P3' positions did not exhibit any specific residue preferences; however, it was observed that small residues like Ser, Gly, and Ala were more prone to cleavage by the protease than the larger residues. It was also observed that solvent-exposed sites like P5, P3, and P3' favored positively charged residues suggesting that the electrostatic interactions being responsible for the catalytic activity of the 3CL pro . The P1 site was most selective as cleavage was only observed in the presence of Gln, Met, or His. Substituting all the favored residue on positions P5 to P1 resulted in a 2.8 fold higher activity than the wild-type substrate sequence. 44 Table 1 shows the substratepeptide residue preferences of DENV 1-4 NS2B-NS3 pro and the SARS-CoV 3CL pro . 31, 44 In summary, substrate residue specificity of DENV NS2B-NS3 pro is mainly seen for the P2-P1-P1' regions. While P2-P1 had dibasic residue preference (Lys, Arg, except for Gln as P2 residue at the NS2B/NS3 cleavage point in DENV 1-4 pro), the P1'region had a preference for smaller sized residues (Ser, Gly). In the case of SARS-CoV 3CL pro, specificity was seen for P2, P1, and P1' region, 45 where it was observed that Gln was indispensable for the P1 site. Leu was favored most at P2 site, and similar to DENV pro , Ser was favored at P1.' 

As mentioned above, the substrate specificities for both the proteases have been well defined in this section, and the corresponding sub-pocket occupancies of the residues involved have been discussed. Table 2 summarizes the residues lining the corresponding S4, S3, S2, S1, and S1' pockets of both DENV pro 31 and SARS-CoV-2 3CL Pro . [46] [47] [48] [49] The overlap of residues within the pockets is due to the common wall between the pockets. It is evident from the comparison that there is no exact overlap of residues lining the sub-pockets of both the proteases, except for Ser and Leu in the S1 pocket, Gln and

His in the S2 pocket, and Arg and Asp in the S3 pocket. However, although the residues are different, residues with similar properties overlap the sub-pockets of both the proteases. The S1 sub-pocket residues (Tyr, Leu, Ile, Phe, and Met) lining both the proteases have hydrophobic side chains. Similarly, in the S2 pocket, residues (Arg, His) from both the proteases have positively charged basic side chains. In addition, residues with polar uncharged side chains (Gln, Asn, and Thr) line both the proteases. In the S3 subpocket, residues (Leu, Val, Met, and Tyr) lining both the proteases have hydrophobic side chains. It can also be seen that residues with positively charged basic side chains (Arg, Lys, and His) constitute the subpockets. Finally, the S4 sub-pocket of both the proteases are lined by residues with hydrophobic side chains (Val, Met, and Leu). Besides this, residues like Cys, Pro, and Gly lining different sub-pockets of both the proteases have tendencies to form various interactions, including hydrogen bond and hydrophobic interactions. While much needs to be explored before arriving at a conclusion, we believe that this similarity could be exploited in the future for developing dual inhibitors against both proteases. Cys 145

Thr26, Thr25 Cys145

Thr190 Leu27 Ser144

Note: Residues with similar color share properties by virtue of their side chains

This section summarizes the research on inhibitors reported in the last six years (2015-20) and summarizes the predicted interactions of a few key residues. Most interactions were common among reported peptide inhibitors. Inhibitors other than peptides were also critically examined for their predicted interactions with the proteases. Besides, any common attributes from different chemical structures were also examined.

Since we have already compared the substrate-peptide residue preferences and residues lining the proteases for DENV and SARS-CoV-2 in the aforementioned sections, the insights provided below could guide the development of potent dual inhibitors. However, experimental validations remain an ultimate tool for any in-silico favored results.

Anti-DENV drug development began with the design of substrate-based peptide inhibitors against DENV NS2B-NS3 pro. However, the relatively flat topography of the DENV protease active site posed numerous challenges in drug design. 22 In addition, there was a significant possibility of developing compounds with compromised pharmacokinetic properties as the non-prime side favored positively charged residues which would ultimately impart a negative charge to the active site. Despite these apprehensions, attempts were made to design peptide inhibitors and peptidomimetics. Eventually, inhibitors with nanomolar affinities were obtained. The peptide inhibitors reported in the last six years have been discussed in the subsections, namely substrate-peptide analogs of non-prime side, peptide conjugates, and cyclic peptides. In this section, we discuss three inhibitors developed by modifying a common tripeptide with varying cap (Cap-Arg-Lys-Phg-NH 2 ) reported by Behnam et al. 50 The first modification was done by Weigel et al..

They observed that previous modifications of peptide hybrids did not optimize the S 2 pocket, considering the target affinity. [50] [51] [52] Hence, arginine mimetic moieties were screened to find a suitable fit. This work revolved around already established tripeptide Bz-Arg-Lys-Phg-NH 2 . 50 They synthesized and performed in-vitro characterization of potent peptidic inhibitors of DENV pro by incorporating phenylalanine and phenylglycine derivatives as arginine-mimicking groups. The second modification-based study in this series was conducted by Behnam et al. (2015) , who developed substrate-based peptide inhibitors and extended their toward the development of potential dual inhibitors of DENV and WNV pro by synthesizing and performing an extensive biological evaluation of inhibitors containing benzyl ethers of 4-hydroxyphenylglycine as non-natural peptidic building blocks. Figure 3 outlines the critical inhibitors from their work and displays the best compound obtained at each modification step. Their work was centered around retro-peptide sequences (Bz-Arg-Lys-Nle-NH 2 ), having good inhibitory activity against DENV2 protease. 51 The replacement of Nle with Phg (Bz-Arg-Lys-L-Phg-NH 2, IC 50 =3.32 µM) improved the inhibition profile. 50 Figure 4 shows the predicted occupancy of different pockets by inhibitor 9. The occupancy was similar to that of the previous inhibitor and an identified, predicted key interaction was that of Alkyl side chain residue Val155 which exhibited a π-σ interaction with the phenyl ring of ether. 54 The common observation about the occupancies of the residues of the potent molecules in different pockets is shown in Table 3 . It can be seen that they have almost a common pattern of occupancy. Arginine mimetic modification of Weigel et al. however, influences the position of Phg wherein it gets displaced to the S 1 ' pocket. The best inhibition (IC 50 =0.018 µM) of protease was witnessed when Phg occupied the S 1 pocket. This tripeptide was so far, the best reported protease inhibitor identified. However, as with previous cases, pharmacokinetic challenges remained a major concern, for which the probable solutions have been discussed later on in this review. 

The peptides discussed above targeted the non-prime side and had their own merits and demerits. Lin et al. (2017) observed that the currently available designs focused on the non-prime side of the DENV protease active site and almost always led to low lipophilic and nonspecific scaffolds. This prompted them to exploit the advantages of aprotinin (58 amino acids protein) for designing new cyclic peptides as aprotinin showed a high affinity for DENV2 pro (K i =26.9 nM). Their strategy of designing new cyclopeptides was based on the structure of aprotinin targeting the active site pockets S 3 -S 4 ' of the DENV pro. After analyzing aprotinin's binding loop, they found that the prime side significantly modulated DENV pro binding affinity. 56 Their designed cyclic peptides showed interaction with both sides of the active site. The design was based on the two loops of aprotinin, linked together with a glycine linker. The first loop, called the binding loop, was significant in binding to DENV pro, while the second loop was incorporated into the design to maintain binding loop structure and rigidity. Also, their previous work identified the key factors affecting binding affinity, which were hydrogen bonding contributed by P1 and P2' residues, the P3' and P4' residue hydrophobic packing, and maintaining the aprotinin binding loop conformation intact in the designed cyclic peptide. 56 Analogous to the structure of aprotinin, there was a need to optimize the binding loop from residue 12 to 18 (Pro13 to Ile18/Ile19) of aprotinin, the linker between the binding loop, and the second loop of aprotinin spanning from residue 35 to 39 (Tyr35/Gly36

to Arg39) in the newly designed cyclopeptides. Also, the linker glycine residues were either retained or were omitted entirely between Ile18/Ile19 and Tyr35/Gly36. Aprotinin's disulfide bond was retained by keeping a disulfide bond between Cys14 of the first loop and Cys38 of the second loop. The length of the binding loop was optimized and compounds with and without linker (one or two Gly residues) were observed, the length of the second loop was optimized and eventually these modifications yielded cyclic peptides of varying lengths. Also, the preferences of different residues at a different position were understood and correlated with the obtained K i values. The best inhibitor showed a K i value of 2.9 µM against wild-type DENV3 pro. Also, the preference of residues for the P1 and P2' positions was established. The key finding from this effort was that the preference between two basic amino acids Lys and Arg, was context-dependent at P1, and Arg was favored at P2' position as it appeared to have more interactions with other residues through hydrogen bonding and van der Waal forces. The choice of optimal C-terminal residue was dependent on the length of the cyclic peptide since different lengths favored different residues at the C-terminal. However, in the native aprotinin, Arg39 was replaced with alanine.

The flexibility was limited to alanine in most of the derivatives, and results were based on binding interactions. Similarly, the significance of P3 residue (N-terminal) was analyzed, and it was found that proline was more favored than a benzoyl cap at this position. However, the presence of proline was found to be context-dependent in the entire experiment as the removal of proline in some cyclic peptides did not change the activity significantly. Table 4 shows the binding affinity of cyclic peptides resulting from the incorporated changes in the optimization process. Although the most effective cyclic peptide obtained had a K i value of 2.9 µM against wild-type DENV3 pro, this value was quite high when compared to the aprotinin inhibition constant of 29 nM against DENV2 pro. Hence, few strategies including, making the molecule more rigid and less bulky, improving the permeability and substituting cleavable peptide bonds between P1 and P1' with non-cleavable bonds to ensure cyclic structure intact were recommended and further it was predicted that any combination of these set of changes could yield a more rigid molecule with good binding affinity. 57 Although nearly all the parameters were incorporated for optimization in the above study, the results were not significantly encouraging. It was suggested that reducing the molecule's flexibility could provide a more potent inhibitor as high flexibility was previously attributed to the low protease inhibition. In addition to improving the protease inhibitory profile, it is also essential to optimize the compound for antiviral activity. However, studies in the past made no such attempts. Takagi Markedly, Compound 15 was not an as effective antiviral agent as intended. The high hydrophilicity due to the side chains of the residues would have inhibited the membrane permeability of the cyclopeptide.

Significant efforts were made to improve its permeability 58,59 while maintaining protease inhibition and antiviral efficiency. In the process, 11 cyclic peptides obtained were further subjected to antiviral assay.

Finally, the optimization efforts yielded a cyclopeptide (P4' D-Arg, P3' L-Arg, P2' L-Lys, P1' L-Lys, P1

D-2NaI, P2 L-hPhe, P3 L-Phe, and P4 D-NaI), which showed promising antiviral activity with an EC 50 value of 2.2 µM without significantly compromising the IC 50 value (1.1 µM). 60 The cyclic peptide was designed keeping the substrate-peptide residue preferences into consideration. The aromatic residues, introduced into the design improved the activity. Notably, the permeability issues were addressed without compromising the protease activity. Such attempts viz. introduction of amphipathic sequence motifs and the proper combination of arginine and hydrophobic aromatic residues could be applied to future cyclic molecules to improve the permeability issues and eventually obtain the desired potent antiviral agent. 

Small molecules offer significant advantages over large peptides in terms of complexity, feasibility, and scalability. Small molecule inhibitors can introduce large conformational changes in DENV NS2B-NS3

pro. 61 This section describes various attempts to develop non-peptide small molecule inhibitors. These have been sub-classified into inhibitors from structure-guided small molecule optimization, obtained from natural sources, High-throughput screening (HTS), HTVS, and a rational design approach. Further, in these subsections, the chemical class is also mentioned. In this section, we discuss various efforts made toward optimizing the potential inhibitors by recognizing the scaffolds involved in the protease inhibitory activity as well as gaining knowledge about protease residues predicted to be involved in the interaction with the inhibitor. compounds. 62 The structure-guided small molecule optimization effort improved compound potency and selectivity for DENV protease. studies using the X-ray crystal structure of DENV2 pro (2FOM) identified two flavonoids, apigenin, and luteolin, with the best docking score (-7.7 kcal/mol). However, luteolin (23, Figure 9 ) was considered a more preferred lead owing to its predicted interaction with the residues of the catalytic triad. Luteolin was also predicted to be involved with two identified key residues (Asp75 and Gly153) via hydrophobic interactions. 68 Figure 9 . DENV NS2B-NS3 pro inhibitors from natural sources. for molecular docking, they reported the binding affinities of three triterpenoids viz. Nimbin, desacetylnimbin, and desacetylsalannin to be 5.56, -5.24, and -3.43 kcal/mol, respectively. The high binding affinity of Nimbin (30, Figure 10 ) was attributed to four predicted hydrogen bonds of Nimbin with the enzyme, which included three bonds with the residues of the catalytic triad (His51, Asp75, Ser135) and one with Asn152 residue. In addition, a hydrophobic interaction with six other residues, including the key residues Val154, Pro132, and Gly153, also contributed toward binding affinity. 72 Bharadwaj et al. Figure 10 . DENV NS2B-NS3 pro inhibitors from natural sources.

In a previous HTS study, Balasubramanian Figure 11 . DENV NS2B-NS3 pro inhibitors from natural sources.

To identify anti-dengue agents from natural products, Lee However, flavonoids are known to yield false-positive activities in biochemical assays. We observed that none of the original studies have discussed assay interference properties of compounds obtained from natural sources. Upon analyzing all the compounds among those compounds obtained from natural resources, we found that seven out of total of 15 compounds were found to have PAINS groups. Of these, four compounds had catechol moiety while three others had ene-one-ene configuration. Considering these findings, we believe that although promising, the data obtained after studying natural compounds, especially those from the flavonoid class cannot be relied upon completely due to false findings. 

HTVS has become an essential part of drug discovery efforts, saving time and providing an opportunity for screening molecules rationally. HTVS, over the years, has helped researchers identify potential HITs and, eventually, potential lead compounds. In the last decade, many antivirals have emerged as a result of this in-silico approach. Boesenbergia rotunda were reported to be potent inhibitors of DENV2 serine pro. 64 In addition, analogs of 4-hydroxypanduratin were also reported as DENV pro inhibitors. 95 Osman et al. (2017) attempted to exploit the benefits of α, β-unsaturated ketones. They synthesized ten compounds by incorporating a piperidone ring to the α, β-unsaturated ketone system and evaluated their inhibitory activity against DENV2 pro. They performed in-silico studies and correlated the in-silico and in-vitro results. Docking studies using DENV2 pro (PDB 2FOM) identified two inhibitors (52 and 53, Figure 19 ) with good docking scores. These molecules also showed good in-vitro affinities in agreement with the in-silico binding scores. Nitro derivatives of 3, 5-bis (arylidene)-4-piperidones (52, 53) had IC 50 values of 15.22 and 16.23 µmol/L, respectively, as compared to the standard panduratin A, which had an IC 50 value of 57.28 µmol/L. In molecular docking studies, both compounds appeared to occupy the active site.

Compound 52 was predicted to form five hydrogen bonds with the His51, Pro132, Ser135, Gly153, and Arg54 residues. A π-π stacking interaction was also observed between the phenyl ring and His51. 

Considering Also, the replacement of the ester group with amide and the introduction of carbamates between two aromatic rings was associated with inhibition of serine protease activity. 103 These changes were incorporated in the lead structure, and 19 analogs of the lead were synthesized and evaluated for DENV2

and WNV pro inhibition. These results have been summarized in (Table 5-7) . The 4-guanidinobenzoate modifications ( Table 5) showed that the activity decreased when the aromatic ring was replaced with the aliphatic chain. The 4-guanidinophenol modifications ( Table 6 ) also did not yield encouraging results.

However, they were more effective when compared to the 4-guanidinobenzoate modifications. The introduction of the carbamate group between the aromatic groups also resulted in less active compounds ( These two modifications yielded successful results. Although the results were not highly encouraging, the insights from these studies certainly improve our understanding of the dos and don'ts in the drug design process. 

>95% Ki=79 ± 21 µM  --- --- --- 15 Selamectin Sigma- Aldrich (USA) >95% Ki=63 ± 18 µM ` --- --- ---

Drug refocusing as a concept has gained significant popularity with the emergence of the COVID-19

pandemic. Similar approaches have also been used for DENV drug discovery. Bhakat Figure 23 ..DENV NS2B-NS3 pro inhibitors identified using the drug repurposing approach.

We searched for dengue fever in the clinical trials database (https://clinicaltrials.gov) and found only 19 records reflecting drug trials. A summary of the patents available for dengue protease inhibitors is provided in Table 9 ; none of these records reflected any protease inhibitors. Despite best efforts by various research groups over the years, none of the DENV protease inhibitors has entered clinical trials. The failure of the only vaccine development has further increased the demand for the search for a potent drug. The structure and function of DENV protease is well understood. Based on the efforts made by researchers in the industry and academia, it is well understood that aiming only to develop inhibitors with sub-micromolar or nanomolar affinities will not serve the purpose. The fundamental challenge in developing an effective drug lies in dealing with the toxicity, stability, and permeability of DENV protease inhibitors. Improving the permeability of the peptide inhibitors while maintaining the protease inhibitory profile requires a balanced selection of functional groups to be incorporated in the design. One such attempt was made by Takagi et al. while designing the cyclic peptides, where they found the introduction of aromatic residues in the design improved the inhibitory profile. Also, in general, the peptide inhibitors showed marked improvement in protease inhibition by aromatization of N & C-terminals. An insight into the protease inhibitor interaction provided by the different research groups in their work lead us to identify certain key residues which could be targeted using a rational design approach. Since these residues of DENV protease appear to be involved in a variety of interactions as shown in Table 10 (hydrogen bonding interactions, hydrophobic interactions, π-π interaction, π-cation interaction, glycine interaction, polar interaction, π-σ interaction) and since the chemical scaffolds with which they interacted belonged to diverse chemical classes, this advantage could be leveraged in drug design approach, both for the designing the small molecule inhibitors as well as peptide inhibitors. Importantly the identified key residues showed interaction with more than half of the reported inhibitors. Table 10 shows the key residues of the proteases which interacted with the inhibitors having a wide range of chemical scaffolds in their architecture. It can be seen that the residues of the catalytic triad and residues An152, Pro132, Lys73 were commonly engaged in hydrogen bonding and hydrophobic interactions. In contrast, certain residues like Tyr161 (π-π interaction), His51 (π-π interaction), Gly153 (Glycine interaction), Lys74 (π-cation interaction), Val155 (π-σ interaction) showed some characteristic interactions confined to these residues. Exploring the occupancies of S1,' S1, S2, S3, and S4 pockets by different scaffolds (Table 10 ) and the residues interacting with these pockets could help us develop novel scaffolds (peptides and small molecules). At the same time, a rational attempt could be made to address the pharmacokinetic issues of the tripeptide inhibitors by proper substitution of basic residues at P1 and P2 position of tripeptide molecule with different heterocyclic/aromatic/cyclic/bicyclic rings, which were predicted to occupy different positions in the S1 and S2 pockets of different reported inhibitors. Table 10 also provides few hydrophobic scaffolds as possible substitutes of the tripeptide cap group, often appearing to occupy the S3 and S4 sub-pockets of the protease. Hence the scaffolds occupying the S3 and S4 sub-pockets could prove to be good substitutes or other possible substitutes of the cap of the tripeptides, and their utility in enhancing the pharmacokinetic properties could be further tested. Each scaffold with its substation offers a definite property to the molecule, which is finally responsible for the inhibitory activity. Table 10 summarizes scaffolds present in various compounds. Since we are far from obtaining a small molecule inhibitor with a nanomolar affinity, any combination of scaffolds with moderate inhibition could be rationally selected for further development. While assembling active portions from these scaffolds could yield active molecules, their synthetic feasibility should also be considered. 

The authors declare no conflicts of interest.

The global distribution and burden of dengue

Experimental investigations regarding the etiology of dengue fever

Towards the design of antiviral inhibitors against flaviviruses: The case for the multifunctional NS3 protein from Dengue virus as a target

The medicinal chemistry of dengue fever

Biochemistry and medicinal chemistry of the dengue virus protease

Dengue virus infection: Epidemiology, pathogenesis, clinical presentation, diagnosis, and prevention

Current concepts: Dengue

Barriers to preclinical investigations of anti-dengue immunity and dengue pathogenesis

Global spread and persistence of dengue

Secondary infection as a risk factor for dengue hemorrhagic fever/dengue shock syndrome: An historical perspective and role of antibodydependent enhancement of infection

Neutralization and Antibody-Dependent Enhancement of Dengue Viruses

Refining the Global Spatial Limits of Dengue Virus Transmission by Evidence-Based Consensus

World Health Organisation. Dengue and severe dengue. WHO Fact Sheet

A recombinant live attenuated tetravalent vaccine for the prevention of dengue

Progress and prospects on DENV protease inhibitors

Flavivirus genome organization, expression, and replication

Progress for dengue virus diseases: Towards the NS2B-NS3pro inhibition for a therapeutic-based approach

Purified NS2B/NS3 serine protease of dengue virus type 2 exhibits cofactor NS2B dependence for cleavage of substrates with dibasic amino acids in vitro

Detection of a trypsin-like serine protease domain in flaviviruses and pestviruses

Both nonstructural proteins NS2B and NS3 are required for the proteolytic processing of dengue virus nonstructural proteins

Structural basis for the activation of flaviviral NS3 proteases from dengue and West Nile virus

Dengue protease activity: The structural integrity and interaction of NS2B with NS3 protease and its potential as a drug target

A structural perspective of the Flavivirus life cycle

Structure-based discovery of dengue virus protease inhibitors

Co-infection of dengue and covid-19: A case report

COVID-19 and dengue co-infection in a returning traveller

Emergence of co-infection of COVID-19 and dengue: A serious public health threat

Evidence that the N-terminal domain of nonstructural protein NS3 from yellow fever virus is a serine protease responsible for site-specific cleavages in the viral polyprotein

In vitro processing of dengue virus type 2 nonstructural proteins NS2A, NS2B, and NS3

Functional profiling of recombinant NS3 proteases from all four serotypes of dengue virus using tetrapeptide and octapeptide substrate libraries

Coronavirus diversity, phylogeny and interspecies jumping

Is the discovery of the novel human betacoronavirus 2c EMC/2012 (HCoV-EMC) the beginning of another SARS-like pandemic?

The emerging novel Middle East respiratory syndrome coronavirus: The "knowns" and "unknowns

The species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2

Genome Composition and Divergence of the Novel Coronavirus (2019-nCoV) Originating in China

An overview of severe acute respiratory syndrome-coronavirus (SARS-CoV) 3CL protease inhibitors: Peptidomimetics and small molecule chemotherapy

Design of wide-spectrum inhibitors targeting coronavirus main proteases

From SARS to MERS: crystallographic studies on coronaviral proteases enable antiviral drug design

Crystal structure of SARS-coV-2 main protease provides a basis for design of improved α-ketoamide inhibitors. Science (80-)

α-Ketoamides as Broad-Spectrum Inhibitors of Coronavirus and Enterovirus Replication: Structure-Based Design, Synthesis, and Activity Assessment

NS2B-NS3 Protease from Zika Virus

The SARS-CoV-2 main protease as drug target

Profiling of substrate specificity of SARS-CoV 3CLpro

Conservation of substrate specificities among coronavirus main proteases

The crystal structures of severe acute respiratory syndrome virus main protease and its complex with an inhibitor

Structure-based design of antiviral drug candidates targeting the SARS-CoV-2 main protease. Science (80-)

Structure-Based Virtual Screening to Discover Potential Lead Molecules for the SARS-CoV-2 Main Protease

Progress in developing inhibitors of sars-cov-2 3c-like protease

C-terminal residue optimization and fragment merging: Discovery of a potent peptide-hybrid inhibitor of dengue protease

Retro peptide-hybrids as selective inhibitors of the Dengue virus NS2B-NS3 protease

Thiazolidinonepeptide hybrids as dengue virus protease inhibitors with antiviral activity in cell culture

Phenylalanine and Phenylglycine Analogues as Arginine Mimetics in Dengue Protease Inhibitors

Discovery of Nanomolar Dengue and West Nile Virus Protease Inhibitors Containing a 4-Benzyloxyphenylglycine Residue

Dual inhibitors of the dengue and West Nile virus NS2B-NS3 proteases: Synthesis, biological evaluation and docking studies of novel peptide-hybrids

Dengue Protease Substrate Recognition: Binding of the Prime Side

Dengue Virus NS2B/NS3

Protease Inhibitors Exploiting the Prime Side

Efficient delivery of cyclic peptides into mammalian cells with short sequence motifs

Early endosomal escape of a cyclic cell-penetrating peptide allows effective cytosolic cargo delivery

Discovery of novel cyclic peptide inhibitors of dengue virus NS2B-NS3 protease with antiviral activity

Binding of low molecular weight inhibitors promotes large conformational changes in the dengue virus ns2b-ns3 protease: Fold analysis by pseudocontact shifts

HTS identifies novel and specific uncompetitive inhibitors of the two-component NS2B-NS3 proteinase of West Nile virus

Identification of covalent active site inhibitors of dengue virus protease

Inhibitory activity of cyclohexenyl chalcone derivatives and flavonoids of fingerroot, Boesenbergia rotunda (L.), towards dengue-2 virus NS3 protease

Allosteric inhibition of the NS2B-NS3 protease from dengue virus

Allosteric pocket of the dengue virus (serotype 2) NS2B/NS3 protease: In silico ligand screening and molecular dynamics studies of inhibition

Flavonoids as noncompetitive inhibitors of Dengue virus NS2B-NS3 protease: Inhibition kinetics and docking studies

Molecular docking and pharmacokinetics study for selected leaf phytochemicals from <i>Carica papaya<\i> Linn. against dengue virus protein

Anti-dengue infectivity evaluation of bioflavonoid from Azadirachta indica by dengue virus serine protease inhibition

Exploration of Carica papaya bioactive compounds as potential inhibitors of dengue NS2B, NS3 and NS5 protease

The identification of active compounds in Ganoderma lucidum var. antler extract inhibiting dengue virus serine protease and its computational studies

In silico evaluation of inhibitory potential of triterpenoids from azadirachta indica against therapeutic target of dengue virus, NS2B-NS3 protease

Genome sequence of the model medicinal mushroom Ganoderma lucidum

Anti-HIV-1 and anti-HIV-1-protease substances from Ganoderma lucidum

Discovery of Ganoderma lucidum triterpenoids as potential inhibitors against Dengue virus NS2B-NS3 protease

High-throughput screening for the identification of small-molecule inhibitors of the flaviviral protease

Bioavailability of curcumin: Problems and promises

Stability of curcumin in buffer solutions and characterization of its degradation products

Inhibition of dengue virus by curcuminoids

Synthesis and in silico studies of a benzenesulfonyl curcumin analogue as a new anti dengue virus type 2 (DEN2) NS2B/NS3

Anti-dengue virus constituents from Formosan zoanthid Palythoa mutuki

In silico studies on some dengue viral proteins with selected Allium cepa oil constituents from Romanian source

Novel dengue virus NS2B/NS3 protease inhibitors

A small molecule inhibitor of dengue virus type 2 protease inhibits the replication of all four dengue virus serotypes in cell culture

Structure-guided discovery of a novel non-peptide inhibitor of dengue virus NS2B-NS3 protease

Synthesis and molecular modelling studies of novel sulphonamide derivatives as dengue virus 2 protease inhibitors

Inhibition of dengue virus replication by novel inhibitors of RNA-dependent RNA polymerase and protease activities

Thioguanine-based DENV-2 NS2B/NS3 protease inhibitors: Virtual screening, synthesis, biological evaluation and molecular modelling

Structure-guided screening of chemical database to identify NS3-NS2B inhibitors for effective therapeutic application in dengue infection

The synthetic molecules YK51 and YK73 attenuate replication of dengue virus serotype 2

Synthesis and biological evaluation of certain α,β-unsaturated ketones and their corresponding fused pyridines as antiviral and cytotoxic agents

Novel 3,5-bis(bromohydroxybenzylidene)piperidin-4-ones as coactivator-associated arginine methyltransferase 1 inhibitors: Enzyme selectivity and cellular activity

Synthesis of N-substituted 3,5-bis(arylidene)-4-piperidones with high antitumor and antioxidant activity

Design and synthesis of new piperidone grafted acetylcholinesterase inhibitors

Design of new competitive dengue Ns2b/Ns3 protease inhibitors-a computational approach

5-Bis(arylidene)-4-piperidones as potential dengue protease inhibitors

Development of antiviral inhibitor against dengue 2 targeting Ns3 protein: In vitro and in silico significant studies

A New Class of Dengue and West Nile Virus Protease Inhibitors with

Protease and Viral Replication Assays

Discovery and SAR studies of methionine-proline anilides as dengue virus NS2B-NS3 protease inhibitors

Identification of fused bicyclic derivatives of pyrrolidine and imidazolidinone as dengue virus-2 NS2B-NS3 protease inhibitors

Structure-guided fragment-based in silico drug design of dengue protease inhibitors

Antifertility activity of systemically administered proteinase (acrosin) inhibitors

Characterization of reversible and pseudo-irreversible acetylcholinesterase inhibitors by means of an immobilized enzyme reactor

Synthesis and structure-activity relationships of small-molecular di-basic esters, amides and carbamates as flaviviral protease inhibitors

Antiviral activities of 15 dengue NS2B-NS3 protease inhibitors using a human cell-based viral quantification assay

Reaching beyond HIV/HCV: Nelfinavir as a potential starting point for broad-spectrum protease inhibitors against dengue and chikungunya virus

Determination of Four Related Substances in Policresulen Solution by HPLC

Policresulen, a novel NS2B/NS3 protease inhibitor, effectively inhibits the replication of DENV2 virus in BHK-21 cells

Existing drugs as broad-spectrum and potent inhibitors for Zika virus by targeting NS2B-NS3 interaction

Flavivirus NS2B/NS3 Protease: Structure, Function, and Inhibition

Erythrosin B is a potent and broad-spectrum orthosteric inhibitor of the flavivirus NS2B-NS3 protease

Identification of a Novel Inhibitor of Dengue Virus Protease through Use of a Virtual Screening Drug Discovery Web Portal

The authors would like to thank the funding agency, DST-SERB, Govt. of India, for providing financial support through their project (EMR /2016/005711) dated 7 th August 2017 and Birla Institute of Technology, Mesra, Ranchi, India for providing the necessary infrastructural facilities.

The authors do not have any conflict of interest