In Silico Mining of the Streptome Database for Hunting Putative Candidates to Allosterically Inhibit the Dengue Virus (Serotype 2) RdRp

Abstract

Background/Objectives: In the last few decades, the dengue virus, a prevalent flavivirus, has demonstrated various epidemiological, economic, and health impacts around the world. Dengue virus serotype 2 (DENV2) plays a vital role in dengue-associated mortality. The RNA-dependent RNA polymerase (RdRp) of DENV2 is a charming druggable target owing to its crucial function in viral reproduction. In recent years, streptomycetes natural products (NPs) have attracted considerable attention as a potential source of antiviral drugs. Methods: Seeking prospective inhibitors that inhibit the DENV2 RdRp allosteric site, in silico mining of the Streptome database was executed. AutoDock4.2.6 software performance in predicting docking poses of the inspected inhibitors was initially conducted according to existing experimental data. Upon the assessed docking parameters, the Streptome database was virtually screened against DENV2 RdRp allosteric site. The streptomycetes NPs with docking scores less than the positive control (68T; calc. −35.6 kJ.mol⁻¹) were advanced for molecular dynamics simulations (MDS), and their binding affinities were computed by employing the MM/GBSA approach. Results: SDB9818 and SDB4806 unveiled superior inhibitor activities against DENV2 RdRp upon MM/GBSA//300 ns MDS than 68T with ΔG_binding values of −246.4, −242.3, and −150.6 kJ.mol⁻¹, respectively. A great consistency was found in both the energetic and structural analyses of the identified inhibitors within the DENV2 RdRp allosteric site. Furthermore, the physicochemical characteristics of the identified inhibitors demonstrated good oral bioavailability. Eventually, quantum mechanical computations were carried out to evaluate the chemical reactivity of the identified inhibitors. Conclusions: As determined by in silico computations, the identified streptomycetes NPs may act as DENV2 RdRp allosteric inhibitors and mandate further experimental assays.

1. Introduction

The dengue virus (DENV) is categorized within the Flaviviridae family and is part of the Orthoflavivirus genus [1]. The Orthoflavivirus genus also involves other viruses, like Zika and West Nile, that transmit human diseases through ticks and mosquitoes [2,3]. Typically, DENV is transferred to individuals by bites from female mosquitoes, especially A. aegypti [4]. It has been recently reported that DENV is responsible for more than 25,000 fatalities and 100 million sicknesses every year throughout the world, indicating that the DENV catastrophe is having a significant impact on global healthcare [5,6,7]. The earliest known serotype of DENV was DENV1, which was isolated for the first time in 1943 [8]. Dengue contagion is induced by four types of viruses (DENV1–4) that participate in approximately 65% of their genome [8]. In spite of this genetic resemblance, each serotype still exhibits significant genetic discrepancy [9]. Among the serotypes, DENV2 plays a key function in dengue-associated mortality, as it is not only described by its specific antigenic properties but also shares similarities with other types [10]. The genetic makeup of DENV2 consists of an 11-kilobase RNA molecule [11]. DENV2 RNA is surrounded by open reading frames (ORFs) as well as 5′ and 3′ untranslated regions (UTRs) that are important for translation, replication, and packaging [12]. In spite of the serious repercussions and high mortality rates of the DENV2 contagion, particular therapeutic choices other than symptomatic relief remain scarce. This challenge is caused by a variety of factors, including DENV2’s rapid replication, repeated alterations, and the utilization of multiple strategies to pervert the host immune system [13,14,15]. Considering these challenges, continuous research on curative techniques for DENV2 contagion is specifically complicated.

Furthermore, about three structural proteins, including envelope, membrane, and capsid protein, as well as seven nonstructural proteins (NSP1–5), are encoded by the DENV2 genomic constituent [16]. These components make up the external sheath of DENV2, which is essential for interaction with the host cell and the escape of immune ripostes [17,18]. Of note, each of the seven NSPs plays a distinct role. In addition to viral repetition, NSP1 also contributes to immune evasion, while NSP2A participates in viral congregation and RNA replication [19]. The NSP2B works as a cofactor and is necessary for the protease activity and viral replication of NSP3 [19]. NSP4A expedites host cell membrane realignment and viral reproduction, whereas NSP4B pitches in forming viral repetition complexes and regulates the host immune retorts [20]. NSP5 acts as an RNA-dependent RNA polymerase (RdRp) and methyltransferase and is crucial for viral RNA synthesis and capping [21]. RdRp is in charge of both positive- and negative-strand RNA synthesis through the replication and transcription of its genome [22,23,24]. There is 65–70% sequence homology among the four DENV serotypes [25]. Due to the absence of a mammalian equivalent for DENV RdRp, this offers an appealing chance for discovering effective and novel antiviral medicines [26,27]. Recently, DENV2 RdRp has attracted increasing attention, and many RdRp inhibitors have been proposed and investigated [28,29,30,31,32]. RdRp inhibitors typically target two fundamental kinds of binding sites, namely, the allosteric and the catalytic binding sites [33]. Nucleoside inhibitors (NIs), including remdesivir, favipiravir, BCX4430, and ribavirin, demonstrate wide-spectrum antiviral activity via inhibiting the RdRp binding site [34,35,36,37]. NIs are phosphorylated into triphosphate analogs, which causes them to integrate into the growing viral RNA strand, which, in turn, terminates the RNA replication process. Nevertheless, the inhibition mechanism employed by NIs frequently results in off-target side effects [38]. Conversely, non-nucleoside inhibitors (NNIs) bind to the RdRp allosteric site and demonstrate antiviral activity by inhibiting the configurational changes necessary for viral RNA transcription [39]. NNIs have attracted significant interest in antiviral medication evolution because of their low toxicity and low side effects [40]. A promising allosteric inhibitor of DENV2 RdRp has been identified by the Novartis Institute for Tropical Diseases, namely an 8-quinolyl sulfonamide (27/68T), demonstrating significant activity against all DENV serotypes with an average IC₅₀ value ranging from 0.013 to 0.074 μM [41]. However, this compound did not succeed in safety or effectiveness evaluations in clinical trials [42]. Consequently, the discovery of effective drug candidates that inhibit the DENV2 RdRp remains a formidable challenge.

Streptomyces have been recognized as an important source of medicinal drugs [43]. Precisely, the Streptomyces species is known to produce over two-thirds of clinically used antibiotics and other pharmacologically important compounds [44]. Streptomycetes have demonstrated potential antiviral activity by targeting various viral enzymes and proteins, including virantmycin and alanosine [45]. Narasin, an ionophore isolated from Streptomyces aureofaciens, has been found to inhibit the replication of DENV [46]. It has been documented that the use of narasin as a treatment for DENV2-infected cells from 12 to 48 h after infection did not decrease the levels of both positive-strand and negative-strand DENV2 RNA, suggesting that narasin does not impede DENV2 RNA replication [46,47]. Unlike broader natural product libraries (e.g., ZINC Natural Products, NPASS, or SuperNatural II) that include a wide range of compounds from diverse sources, the Streptome database uniquely provides the most comprehensive collection of streptomyces-derived natural products [48]. This enhances the likelihood of identifying biologically active scaffolds with established antimicrobial or therapeutic potential. Seeking effective anti-DENV2 drug candidates, the Streptome database, containing > 6500 streptomycetes natural products (NPs), was mined to hunt novel allosteric inhibitors toward DENV2 RdRp with a higher potency than 68T. Upon the docking computations, the most potent streptomycetes NPs were picked up and advanced for molecular dynamics simulations (MDS) over 300 ns, accompanied by binding energies utilizing the MM/GBSA approach. Post-MD analyses were executed on the most promising streptomycetes NPs bound to the DENV2 RdRp. Additionally, the drug-like features of the identified streptomycetes NPs were predicted. Furthermore, quantum mechanical computations were employed to obtain a deeper understanding of the geometrical and electronic properties of the identified streptomycetes NPs. A schematic diagram of the applied computational approaches to filter the Streptome database is portrayed in Figure 1. The obtained results highlight the potential of the identified streptomycetes NPs as allosteric DENV2 RdRp inhibitors and offer viable curative candidates for future experimental assays.

2. Results and Discussion

2.1. Docking Protocol Assessment

Prior to data generation, the AutoDock4.2.6 software’s performance in predicting the binding pose of the co-crystallized compound 27/68T inside the DENV2 RdRp allosteric site was validated based on the available experimental data. The anticipated binding pose of compound 27/68T was compared to its original binding pose (PDB entry: 5K5M [41]) (Figure 2). According to the literature, the RMSD values between the anticipated binding pose and the resolved original binding pose should be <2.0 Å [49,50,51]. From Figure 2, the predicted pose closely matched the experimental structure with an RMSD value of 0.79 Å. Examining the binding pose of 68T demonstrated three fundamental H-bonds with ARG729 (3.11 Å), TRP795 (3.40 Å), and GLU802 (2.67 Å). As well, the 68T exhibited a pi-pi T-shaped interaction with HIS711 and a pi-cation interaction with ARG729. Re-docking results showed that the AutoDock2.4.6 software correctly predicted the binding pose of ligand-RdRp complexes. As a result, the AutoDock4.2.6 software was used to mine the Streptome database for hunting putative DENV2 RdRp ligands.

2.2. Virtual Screening of the Streptome Database

A virtual screening technique can be used to hunt prospective bioactive inhibitors at the early stages of the drug design process [52]. Herein, the Streptome database encompassing > 6500 NPs was mined using standard docking parameters of eval = 5 million and GA = 50. On the basis of standard docking computations, only 151 streptomycetes NPs unveiled docking scores equal to or lower than that of 68T (calc. −35.6 kJ.mol⁻¹). As a result, these 151 streptomycetes NPs were re-docked toward the DENV2 RdRp utilizing expensive parameters (i.e., eval = 25 million and GA = 250) (Table S1). As reported in Table S1, only 39 streptomycetes NPs revealed lower docking scores in comparison with 68T (calc. −35.6 kJ.mol⁻¹). Figure S1 depicts the 2D illustrations for the anticipated binding poses of these 39 streptomycetes NPs within the DENV2 RdRp allosteric site. As shown in Figure S1, all investigated streptomycetes NPs within the DENV2 RdRp allosteric site displayed a fundamental H-bond with ARG729 and TRP795. Pi-Sigma, pi-pi T-shaped, amide-pi stacked, carbon H-bond, and pi-cation interactions were also monitored between the inspected streptomycetes NPs and the main residues in the DENV2 RdRp allosteric site. 2D Chemical structures, computed standard and expensive docking scores, and the intramolecular H-bond of the top 10 scoring streptomycetes NPs toward DENV2 RdRp are registered in

Table 1. 2D Chemical structures, computed docking scores, and intramolecular H-bond of the top 10 scoring streptomycetes NPs toward the DENV2 RdRp allosteric site ^a.

Compound Name/ID	2D Chemical Structure	Docking Score (kJ.mol−1)		Intermolecular H-Bond
		Standard	Expensive
68T		−35.6	−35.6	ARG729 (3.11 Å), TRP795 (3.40 Å), GLU802 (2.67 Å)
SDB9818		−46.9	−46.9	ARG737 (2.04; 2.62; 1.83 Å), SER796 (3.01 Å), CYS709 (1.83 Å), ASP664 (1.74 Å), TRP795 (1.80 Å), SER710 (1.70 Å)
SDB4806		−42.7	−45.6	GLU459 (1.63 Å), ASP664 (1.84 Å), ARG729 (2.16 Å), ARG737 (2.84 Å), THR794 (1.97; 2.44 Å), TRP795 (2.54 Å), SER796 (2.83 Å)
SDB895		–41.8	–45.2	LYS461 (2.90 Å), ASP664 (1.81 Å), ARG737 (3.72 Å), SER796 (2.22; 2.33; 2.08 Å),
SDB12947		–35.6	–45.2	ASP664 (1.88 Å), HIS798 (3.18 Å), SER796 (3.02; 3.18 Å), CYS709 (2.19 Å)
SDB13026		–46.9	–45.2	LYS461 (2.09; 2.12; 2.22 Å), ASP664 (2.45 Å), ARG729 (2.33; 2.35 Å), TYR766 (2.83 Å), ARG737 (1.73 Å),
SDB9891		–42.3	–45.2	LYS461 (1.90 Å), SER796 (1.72; 2.25; 3.15 Å), TYR766 (2.02; 2.15 Å), HIS798 (3.08 Å), CYS709 (1.83 Å)
SDB10285		–41.8	–42.3	LYS461 (2.24 Å), ARG472 (2.95 Å), ARG737 (2.53 Å), GLU802 (1.86; 2.36 Å)
SDB993		–41.8	–42.3	SER796 (2.65 Å), ARG737 (2.01 Å)
SDB1014		–41.4	–40.2	LYS461 (2.60 Å), SER796 (2.84 Å)
SDB827		–41.4	–39.3	ASP664 (2.13; 2.23 Å), ARG729 (3.07 Å), ARG737 (1.99, 2.57 Å), TYR766 (2.09, 2.01 Å), THR794 (2.41 Å), SER796 (2.12, 2.19 Å)

^a Streptomycetes NPs were ordered according to the expensive docking computations.

. As well, the 3D and 2D depictions for the predicted binding poses of two outstanding streptomycetes NPs within the DENV2 RdRp allosteric site are illustrated in Figure 3. Furthermore, the most potent two streptomycetes NPs were re-docked using GA = 1000 against the DENV2 RdRp allosteric site. Comparing results obtained with a GA value of 250 to those using a GA value of 1000, no notable differences were observed in the docking poses and scores of these two streptomycetes NPs within the DENV2 RdRp allosteric site (Figure S2). Of note, these two promising streptomycetes NPs were selected according to the evaluated binding affinity throughout 300 ns MDS, as evidenced in the MDS section.

According to data listed in Table 1, SDB9818 manifested a significantly favorable docking score of −46.9 kJ.mol⁻¹ against the DENV2 RdRp allosteric site. SDB9818 is a natural product isolated from Streptomyces sp. VITBRK2, which shows a high activity against MRSA [53]. Structural insights into the docking pose of SDB9818 against the DENV2 RdRp allosteric site showed that the CO of SDB9818 formed three H-bonds with the NH₂ of ARG737 (2.04, 2.62, and 1.83 Å). Moreover, the two OH groups of (2R,3S)-butane-2,3-diol exhibited three H-bonds with the OH of SER796 (3.01 Å), CO of CYS709 (1.83 Å), and carboxylate of ASP664 (1.74 Å). Ultimately, the NH and CO groups of formamide showed two H-bonds with the CO of TRP795 (1.80 Å) and the OH of SER710 (1.70 Å) (Figure 3).

SDB4806 is an anthracycline antibiotic extracted from Streptomyces peucetius var. aureus, which demonstrates anticarcinoma activity by impeding RNA and DNA synthesis [54]. SDB4806 revealed the second-lowest docking score of −45.6 kJ.mol⁻¹ against the DENV2 RdRp allosteric site. According to Figure 3, SDB4806 displayed eight H-bonds with the essential residues of the DENV2 RdRp allosteric site. More precisely, the OH groups of SDB4806 established two H-bonds with the carboxylate of GLU459 (1.63 Å) and ASP664 (1.84 Å). Moreover, the CO groups of SDB4806 exhibited six H-bonds with the NH₂ of ARG729 (2.16 Å), NH₂ of ARG737 (2.84 Å), OH of THR794 (1.97 and 2.44 Å), NH of TRP795 (2.54 Å), and NH of SER796 (2.83 Å).

According to sequence and structure comparisons, the DENV2 RdRp allosteric site is not homologous to any known domains in human DNA/RNA polymerases or bacterial polymerases, reducing off-target interactions [55]. Moreover, SDB9818 and SDB4806 often do not exhibit significant activity against eukaryotic polymerases unless they are specifically selected or designed to have this effect.

2.3. Molecular Dynamics Simulations (MDS)

To examine receptor–ligand stability and conformational variations, MDS has been widely used to analyze the dynamic properties of the inhibitor complexed with receptors. The most promising 39 streptomycetes NPs with docking scores less than 68T (calc. −35.6 kJ.mol⁻¹) in the complex with the DENV2 RdRp allosteric site were advanced for MDS over 10 ns. Table S2 lists the corresponding MM/GBSA binding energies over the 10 ns MDS. As found in Table S2, only five streptomycetes NPs unveiled higher binding affinities than 68T (ΔG_binding = −167.8 kJ.mol⁻¹) toward the DENV2 RdRp allosteric site. In order to achieve more reliable outcomes, these five streptomycetes NPs in complex with the DENV2 RdRp allosteric site were introduced to longer MDS throughout 200 ns, accompanied by MM/GBSA binding energy estimations (Figure 4). As depicted in Figure 4, two streptomycetes NPs, namely SDB9818 and SDB4806, demonstrated lower binding energies compared to 68T (ΔG_binding = −152.7 kJ.mol⁻¹). Moreover, a 300 ns MDS was executed for these two streptomycetes NPs bound to the DENV2 RdRp allosteric site, followed by binding affinity computations (Figure 4). From Figure 4, it is obvious that no significant variations were observed in the evaluated binding energies over 200 and 300 ns MDS for the identified two streptomycetes NPs complexed with the DENV2 RdRp allosteric site. In comparison with 68T (ΔG_binding = −150.6 kJ.mol⁻¹), SDB9818 and SDB4806 unveiled superior binding affinities toward the DENV2 RdRp allosteric site throughout 300 ns MDS with ΔG_binding values of −246.4 and −242.3 kJ.mol⁻¹, respectively. Moreover, the standard deviation and standard error of the mean for the binding energy (ΔG_binding) of the two identified streptomycetes NPs and 68T complexed with the DENV2 allosteric site over 300 ns MDS were computed (Table S3). As registered in Table S3, the low standard deviation and standard error of mean values indicated that the binding energy remained relatively stable throughout the simulation time. These outcomes illuminated the efficiency of the identified streptomycetes NPs as promising anti-DENV2 drug candidates.

The main limitation of the current research is the lack of MDS runs replication for the identified streptomycetes NPs as promising anti-DENV2 drug candidates. As a consequence, the future study will consider multiple independent MDS with a variety of initial velocities in order to verify the consistency of the findings.

The decomposition of the computed binding energies into separate components was then executed and is depicted in Figure 5. As depicted in Figure 5, the E_vdW was the predominant participation in the ΔG_binding of SDB9818, SDB4806, and 68T complexed with the DENV2 RdRp allosteric site, with mean values of −1358.5, −977.4, and −228.9 kJ.mol⁻¹, respectively. Moreover, the E_ele demonstrated a favorable contribution to ΔG_binding with the mean values of −160.7, −196.2, and −224.3 kJ.mol⁻¹ for SDB4806, SDB9818, and 68T complexed with the DENV2 RdRp allosteric site, respectively.

To obtain more insights into the fundamental amino acids included in the inhibition of the DENV2 RdRp allosteric site, the per-residue energy decomposition was performed and plotted in Figure 6. It is worth noting that only residues with ΔG_binding < −2.1 kJ.mol⁻¹ were taken into account. From the delineated per-residue energy decomposition in Figure 6, ARG729, THR794, TRP795, SER796, and ARG737 disclosed promising participations in the binding of SDB9818, SDB4806, SDB895, and 68T with the allosteric site of the DENV2 RdRp. The significant participation of ARG729 in the allosteric site of DENV RdRp evinced −9.6, −6.7, −7.1, and −2.5 kJ.mol⁻¹ for SDB9818-, SDB4806-, and 68T-RdRp complexes, respectively.

2.4. Post-MD Analyses

In order to further investigate the steadiness and behavior of SDB9818-, SDB4806-, and 68T-RdRp, post-MD analyses were accomplished over the course of 300 ns MDS. Six properties were evaluated, including root-mean-square fluctuation and deviation (RMSF and RMSD), radius of gyration (Rg), binding energy per trajectory, H-bond analysis, and solvent-accessible surface area (SASA).

2.4.1. Binding Energy per Trajectory

The energetic persistence of SDB9818, SDB4806, and 68T bound to the DENV2 RdRp allosteric site was evaluated by measuring the binding energy versus time correlation (Figure 7a). As illustrated in Figure 7a, the comprehensive constancy for SDB9818, SDB4806, and 68T bound to the DENV2 RdRp allosteric site was observed with average ΔG_binding values of −246.4, −242.3, and −150.6 kJ.mol⁻¹, respectively. The most striking outcome to emerge from this graph is that all inspected complexes conserved steadiness throughout 300 ns MDS.

2.4.2. RMSD Analysis

In order to supply valuable insights into ligand–receptor structural steadiness, RMSD for the backbone atoms relative to the initial position was measured over 300 ns MDS (Figure 7b). Based on Figure 7b, the mean RMSD values were 0.32, 0.23, and 0.27 nm for SDB9818-, SDB4806-, and 68T-RdRp complexes, respectively. These results demonstrated that these inspected streptomycetes NPs were bound tightly and were overall stable within the RdRp allosteric site. Furthermore, multiple snapshots taken at consistent intervals throughout the MDS were extracted to further assess the stability of the identified streptomycetes NPs within the DENV2 RdRp allosteric site. The 3D binding patterns of the identified streptomycetes NPs inside the allosteric site of DENV2 RdRp are depicted in Figure S3. As illustrated in Figure S3, the identified streptomycetes NPs sustained stable interactions at the allosteric site of streptomycetes NPs, resulting in the formation of a stable RdRp-inhibitor complex.

2.4.3. Rg Analysis

The Rg is one of the parameters utilized to gauge the equilibrium configurations of the whole system, and this assists in determining receptor compactness during MDS. Protein compactness is dependent on their residue sequence constitution, and it changes during receptor–ligand interactions [56]. As a result, the Rg for SDB4806, SDB9818, and 68T complexed with the DENV2 RdRp allosteric site was estimated throughout 300 ns MDS in order to investigate how residue flexibility affects enzyme compactness (Figure 8a). The measured average Rg values were 2.57, 2.66, 2.65, and 2.66 nm for apo-, SDB4806-, SDB9818-, and 68T-RdRp, respectively. These results implied that the RdRp structure was well-compacted and more steady after the complexation with SDB4806 and SDB9818.

2.4.4. RMSF Analysis

RMSF is a parameter utilized for determining the elasticity of every residue of the target during the MDS. RMSF explains variations in the structural configurations of the target from the starting position until the termination of the simulations according to the residue fluctuations [57]. Thus, the RMSF was evaluated over 300 ns MDS to grasp how structural deviations affect target elasticity (Figure 8b). It is known that greater RMSF demonstrates higher elasticity of the receptor during simulations. As illustrated in Figure 8b, it was noticed that the 68T-RdRp had a greater appearance of oscillating residues compared to the SDB4806 and SDB9818 complexed with the DENV2 RdRp allosteric site. The average RMSF values were 0.117, 0.118, 0.124, and 0.122 nm for apo-, SDB9818-, SDB4806-, and 68T-RdRp, respectively. Based on RMSF results, most of the residues of the DENV2 RdRp allosteric site exhibited higher stability after complexation with SDB4806, SDB9818, and 68T.

2.4.5. SASA Analysis

Water plays a substantial role in the determination of the stability, structure, role, and dynamics of targets. SASA is an indicator that is utilized for measuring the attainability of protein amino acids to encompass water molecules, and any alterations in the attainability could impact protein functions, dynamics, and structure [58]. Herein, SASA was employed to gauge the binding impact of the investigated inhibitor on the configurational behavior of the DENV2 RdRp allosteric site. Figure 8c illustrates the SASA of the DENV2 RdRp allosteric site upon binding with SDB4806, SDB9818, and 68T throughout 300 ns MDS. As can be seen from the SASA plots, SDB4806, SDB9818, and 68T demonstrated similar frame patterns after binding to the RdRp allosteric site. Nevertheless, the SASA of RdRp increased upon binding with the 68T (Figure 8c). In contrast, as SDB4806 and SDB9818 were bound to the RdRp allosteric site, the SASA value slightly decreased. The average SASA values were 275.9, 277.0, 277.8, and 282.9 nm² for apo-, SDB4806-, SDB9818-, and 68T-RdRp, respectively. SASA outcomes indicated that the streptomycetes NPs and 68T had little impact on the solvent accessibility of the DENV2 RdRp enzyme.

2.4.6. H-Bond Analysis

The steadiness and binding affinity of inhibitors with the receptor were investigated by gauging the number of conventional H-bonds established during the MDS. Figure 9 portrays the number of H-bonds between the investigated streptomycetes NPs and the key residues of the DENV2 RdRp allosteric site throughout 300 ns MDS. As shown in Figure 9, the average number of H-bonds was 7, 6, and 2 for SDB9818-, SDB4806-, and 68T-RdRp complexes, respectively. Notably, the average number of H-bonds for the identified streptomycetes NPs was more remarkable compared to the co-crystallized 68T complexed with the DENV2 RdRp, indicating the higher stability of these streptomycetes NPs than 68T within the DENV2 RdRp allosteric site over 300 ns MDS. These results firmly endorsed the anticipated H-bond interactions between the investigated streptomycetes NPs and the allosteric site amino acids of RdRp via molecular docking computations.

2.5. Physicochemical Characteristics

The drug-likeness of a molecule is estimated in accordance with Lipinski’s rule of five (Ro5), which detects whether the compound is orally active or not.

Table 2. The drug-like characteristics of the identified streptomycetes NPs and 68T as prospective anti-DENV2 drug candidates.

Compound Name/ID	LogP	MW (g/mol)	HBD	HBA
68T	4.23	487.55	2	7
SDB9818	0.40	423.46	6	8
SDB4806	2.24	511.52	4	10

lists the drug-likeness features of the identified streptomycetes NPs and 68T. According to Ro5, a molecule failing to meet more than two of its criteria is likely to have poor absorption in preclinical testing [59]. According to data listed in Table 2, SDB9818 violates only the HBD limit (6 vs. 5), which may marginally impact permeability but not necessarily disqualify it as a lead, especially given its favorable solubility (LogP = 0.40). SDB4806 slightly exceeds the MW threshold (511.52 g/mol) but remains within acceptable bounds for lead optimization. The LogP and H-bond of SDB4806 counts were favorable. 68T complied fully with Ro5, supporting its drug-like profile. More precisely, the MW was in the range of 423.46 to 511.52 g/mol for the investigated streptomycetes NPs and 68T. Moreover, the HBD was 6, 4, and 2 for SDB9818, SDB4806, and 68T, respectively. SDB4806, SDB9818, and 68T were found to have a number of HBA with values ranging from 7 to 10. The logP was less than 5 for the investigated streptomycetes NPs and 68T. Notably, these results underscore the promise of the identified streptomycetes NPs as potential allosteric inhibitors of DENV2 RdRp and present them as viable candidates for future experimental validation.

Figure 10a displays the bioavailability radar plots of the identified streptomycetes NPs and 68T and their drug-likeness characteristics. The pink area inside the hexagonal shape represents the ideal range for the investigated compounds. The recommended parameters for the drug-likeness of a small molecule are as follows: (i) insaturation (INSITU): a fraction of carbons exhibiting sp³ hybridization of at least 0.25; (ii) insolubility (INSOLE): a log S not exceeding 6; (iii) hydrophobicity (LIPO): ranging from −0.7 to +5.0; (iv) rotatable bonds (FLEXI): a maximum of 9 rotatable bonds; (v) molecular weight (SIZE): between 150 and 500 g/mol; and (vi) polar surface area (POLAR): between 20 and 130 g/mol, along with polar surface area (POLAR): within the range of 20 and 130 Ų. According to data presented in Figure 10a, the investigated streptomycetes NPs and 68T lie within the passable range, indicating their favorable drug-like characteristics.

In addition, the pharmacokinetic features of the identified streptomycetes NPs and 68T were investigated using a BOILED-Egg model (Figure 10b). The BOILED-Egg model demonstrates its utility in simultaneously forecasting two significant pharmacokinetic properties, specifically, passive gastrointestinal absorption (HIA) and penetration through the blood–brain barrier (BBB). As shown in Figure 10b, SDB9818, SDB4806, and 68T were detected in the grey region, indicating that these compounds demonstrated low passive HIA and BBB. Consequently, the investigated compounds would gain better bioavailability profiles during a drug development process.

2.6. QM Computations

ESP analysis is a reliable approach to demonstrate the negative and positive potentials on the surface of a chemical molecule. For the optimized last snapshot of streptomycetes NPs and 68T retrieved from MDS, the MEP maps were generated and are graphed in Figure 11. As delineated in Figure 11, red regions were noticed above the N and O atoms of the investigated streptomycetes NPs and 68T, indicating their nucleophilic nature. Additionally, blue regions were marked above the H atoms of the inspected streptomycetes NPs and 68T, implying their electrophilic nature. According to the MEP maps, streptomycetes NPs and 68T displayed the capability of forming H-bonds with fundamental residues within the allosteric site of DENV2 RdRp.

Based on Figure 12, HOMO levels were mainly found around the electron-rich zones of the investigated compounds (e.g., O and N atoms). In addition, LUMO levels were concentrated around the electron-deficient regions in the investigated streptomycetes NPs and 68T (e.g., H atoms). As numerical evidence from

Table 3. The computed E_HOMO, E_LUMO, E_FL, and E_gap in eV for the most promising streptomycetes NPs and 68T.

Compound Name/ID	EHOMO	ELUMO	EFL	Egap
68T	−6.88	−0.83	−3.85	6.05
SDB9818	−7.99	−1.34	−4.67	6.65
SDB4806	−8.06	−2.33	−5.19	5.74

, the E_HOMO/E_LUMO values were −7.99/−1.34, −8.06/−2.33, and −6.88/−0.83 eV for SDB9818, SDB4806, and 68T, respectively. Moreover, SDB9818, SDB4806, and 68T revealed E_FL values in the range of −3.85 to −5.19 eV. SDB9818, SDB4806, and 68T displayed low E_gap with values of 6.65, 5.74, and 6.05 eV, respectively. Of note, the low E_gap values indicated the considerable chemical reactivity of the investigated streptomycetes NPs and 68T.

According to the indisputable role of electronic parameters, global descriptors were computed for SDB9818, SDB4806, and 68T (

Table 4. Computed global descriptors for SDB4806, SDB9818, and 68T.

Compound Name/ID	IP (eV)	EA (eV)	η (eV)	S (eV−1)
68T	6.88	0.83	3.03	0.33
SDB9818	7.99	1.34	3.32	0.30
SDB4806	8.06	2.33	2.87	0.35

). As reported in Table 4, SDB9818, SDB4806, and 68T demonstrated IP values ranging from 6.88 to 8.71 eV. In addition, the EA values were found to be 1.34, 2.33, and 0.83 eV for SDB9818, SDB4806, and 68T, respectively. The η and S of the investigated compounds can serve as an indicator for their stability and chemical reactivity. The η values for SDB9818, SDB4806, and 68T ranged from 2.87 to 3.32 eV. SDB9818, SDB4806, and 68T unveiled outstanding S values of 0.30, 0.35, and 0.33 eV⁻¹, respectively. Among the three compounds, SDB4806 exhibited the highest softness (0.35 eV⁻¹), suggesting a greater capacity for electronic adaptability. Generally, the obtained η and S values of SDB9818, SDB4806, and 68T indicated promising polarizability and electron-donating/accepting flexibility, which facilitated effective interactions with DENV RdRp allosteric site through non-covalent interactions (e.g., H-bonds, electrostatic interactions, and π–π stacking). Moreover, the computed global descriptors provided an electronic rationale for the biological activity of the investigated compounds and reinforced their potential as effective inhibitors for the DENV2 RdRp allosteric site.

3. Computational Methodology

3.1. RdRp Preparation

The X-ray crystallographic structure of DENV2 RdRp (PDB entry: 5K5M, resolution: 2.1 Å) was chosen and employed as a template for all in silico computations [41]. All heteroatoms, including ions, inhibitors, and water molecules, were extracted for preparation purposes. The ionization status of titratable residues at a pH of 7.4 was determined utilizing the PropKa3 software [60]. All missing H-atoms were consequently inserted.

3.2. Streptome Database Preparation

The Streptome database, involving > 6500 NPs derived from approximately 3300 streptomycetes strains, was retrieved in SDF format [61]. Duplicates were removed in accordance with the International Chemical Identifier (InChIKey) [62]. Omega2 software (version 4.1.1.0) was utilized for generating 3D structures for each streptomycete NP [63,64]. All generated 3D structures were then adopted for an optimization process utilizing the MMFF94S force field implemented in SZYBKI software (version 2.4.0.0) [65,66]. The FixPka tool within the QUACPAC program (version 2.1.3.0) was used to determine the dominant ionization state at a pH of 7.4 [67]. The atomic charges of the streptomycetes NPs were computed with the assistance of the Gasteiger-Marsili method [68]. At www.compchem.net/ccdb, accessed on 13 February 2024, all prepared streptomycetes NPs can be accessed.

3.3. Docking Computation

AutoDock4.2.6 software was applied to accomplish molecular docking computations [69]. For docking calculations, the DENV2 RdRp enzyme was saved in the pdbqt format with the help of MGLTools 1.5.7 [70]. This research included two stages of docking calculations, namely standard and expensive docking computations. For standard and expensive docking computations, the genetic algorithm (GA) runs were set to 50 and 250 iterations, respectively. Moreover, 5 and 25 million energy evaluations (eval) were adjusted to standard and expensive docking computations, respectively. Other docking parameters were left at their defaults. The grid was designed to include the entire allosteric site of the DENV2 RdRP, measuring 40 × 40 × 40 ų, with a spacing value of 0.375 Å. Grid coordinates were located at x = −15.551, y = −44.03, and z = −19.247. The docking conformation with the highest binding affinity was selected from the largest cluster as a representative docking mode.

3.4. MD Simulations (MDS)

All MDS for the most promising streptomycetes NPs complexed with DENV2 RdRp allosteric site were accomplished utilizing AMBER20 software [71]. Parameters for conducting MDS are more fully described elsewhere [72,73,74]. Briefly, DENV2 RdRp was characterized by the AMBER force field 14SB [75]. The general AMBER Force Field (GAFF2) was utilized for the parameterization of the inspected streptomycetes NPs [76]. Using Gaussian09 software, the geometrical optimization for the inspected streptomycetes NPs was executed at the HF/6-31G* level [77]. The restrained electrostatic potential (RESP) was then used for computing the atomic charges for the optimized streptomycetes NPs [78]. In a truncated octahedral periodic box, each RdRp-NP complex was solvated with the TIP3P water model [79]. In order to preserve the electroneutrality of the inspected RdRp-NP complexes, sodium or chloride counterions were added. The isosmotic salt environment was also maintained by the insertion of 0.15 M NaCl. The inspected complexes were advanced for 5000 iterations of energy minimization. Following that, the investigated complexes were gently annealed up to 310 K throughout 50 ps. The heated complexes were submitted to the equilibration stage for 10 ns. Finally, the production runs were carried out on the equilibrated complexes throughout 10, 200, and 300 ns. MD trajectories were recorded every 10 ps for post-MD analyses, which were performed using the CPPTRAJ tool [80]. MDS was accelerated using the PMEMD version of CUDA on GPU cores. All molecular interactions were displayed utilizing the Discovery Studio module of Biovia software (version 2019) [81].

3.5. Binding Energy Computations

To evaluate the binding energy between DENV2 RdRp and streptomycetes NPs, the molecular mechanics/generalized Born surface area (MM/GBSA) approach was used [82]. The binding energy was estimated by the following equation:

$${\Delta G}_{binding}=G_{RdRp-NP }-G_{RdRp} -G_{NP}$$

(1)

where the energy term (G) was numerically assessed as follows:

$$G=E_{vdW}+G_{SA}+G_{GB }+E_{ele}-TS$$

(2)

E_vdW points out van der Waals energy. G_SA and G_GB imply non-polar and polar contributions of the desolvation energy, respectively. Using the modified GB model (iGB = 2) developed by Onufriev et al., the G_GB was computed [83]. E_ele stands for electrostatic energy. T indicates absolute temperature, and S is the entropy participation. Entropic participation was not considered due to its high computation costs [84,85]. Notably, the exclusion of the entropic component had no significant effect on the MM/GBSA binding energy estimations [86].

3.6. Physicochemical Features

The SwissADME web server was employed to estimate the physicochemical characteristics of the identified streptomycetes NPs based on Lipinski’s rule of five (Ro5) conditions [87]. Lipinski’s rule states that (i) the number of H-bond acceptors (HBA) of the investigated compound should be <10, (ii) the molecular weight (MW) should not be >500 daltons, (iii) the number of H-bond donors (HBD) should be <5, and (iv) the partition coefficient (logP) should not be >5. In addition, the BOILED-Egg method was utilized for predicting gastrointestinal absorption and brain permeation of the identified streptomycetes NPs [88]. In the shape of a boiled egg, the plot is composed of white, grey, and yolk regions. More exactly, compounds in white zones are those that are more likely to be absorbed by the gastrointestinal tract, while compounds in yolk zones are those that are more likely to permeate the brain. Meanwhile, compounds in grey zones are those that are less likely to be non-absorbed by the gastrointestinal tract or non-brain permeating. For dot coloring, the blue dots indicate that the compound could be effluxed from the Central Nervous System (CNS) by P-Glycoprotein (PGP), whereas the red dots mean that compounds could not be effluxed from CNS by PGP [88].

3.7. Quantum Mechanical (QM) Computations

In the realm of QM computations, the final frame of the identified streptomycetes NPs elicited from MDS was geometrically optimized utilizing the M062X/6-311+G** level of theory with the assistance of Gaussian09 software [77]. For the optimized streptomycetes NPs, electrostatic potential (ESP) analysis was performed, and the corresponding molecular electrostatic potential (MEP) maps were generated at an electron density envelope of 0.002 au [89]. The frontier molecular orbitals (FMOs) theory was implemented for a deeper understanding of the electronic features of the optimized streptomycetes NPs. As a result, the energetic values and electronic patterns of the highest occupied/lowest unoccupied molecular orbital (i.e., HOMO/LUMO) were generated. Furthermore, the Fermi level (E_FL) energy and energy gap (E_gap) were computed on the basis of E_LUMO and E_HOMO, as shown in Equations (3) and (4), respectively.

$$E_{\mathrm{F}\mathrm{L}}=E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}}+{\displaystyle \frac{E_{\mathrm{L}\mathrm{U}\mathrm{M}\mathrm{O}}-E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}} }{2}}$$

(3)

$$E_{\mathrm{g}\mathrm{a}\mathrm{p}}=E_{\mathrm{L}\mathrm{U}\mathrm{M}\mathrm{O}}-E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}}$$

(4)

As well, further electronic parameters such as global hardness (η), global softness (S), ionization potential (IP), and electron affinity (EA) were computed as follows:

$$\eta ={\displaystyle \frac{E_{\mathrm{L}\mathrm{U}\mathrm{M}\mathrm{O}}-E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}} }{2}}$$

(5)

$$IP=-E_{\mathrm{H}\mathrm{O}\mathrm{M}\mathrm{O}}$$

(6)

$$S={\displaystyle \frac{1}{\eta }}$$

(7)

$$EA=-E_{\mathrm{L}\mathrm{U}\mathrm{M}\mathrm{O}}$$

(8)

4. Conclusions

Recently, the dengue virus, a commonly encountered flavivirus, has shown a range of epidemiological, economic, and health effects globally. DENV2 is a significant factor in mortality associated with dengue. DENV2 RdRp is an attractive druggable target owing to its essential role in the replication of the virus. In the present work, the Streptome database, containing > 6500 compounds, was mined to hunt the most prospective inhibitors against the allosteric site of RdRp, utilizing advanced in silico approaches. Based on docking computations and MDS coupled with the MM/GBSA binding energy evaluations, SDB4806 and SDB9818 revealed promising binding affinities with ΔG_binding < −188.3 kJ.mol⁻¹ against RdRp. The constancy of the identified streptomycetes NPs bound to the DENV2 RdRp allosteric site was confirmed upon the post-MD analyses throughout 300 ns. Based on the drug-likeness features, the identified streptomycetes NPs unveiled good oral bioavailability. Quantum mechanical computations were also performed for the identified streptomycetes NPs, and the outcomes demonstrated their stability and chemical reactivity. These findings provide a solid basis for future experimental validations and in-vitro studies aimed at combating DENV2.