qNMR quantification and in silico analysis of isobrucein B and neosergeolide from Picrolemma sprucei as potential inhibitors of SARS-CoV-2 protease (3CLpro) and RNA-dependent RNA polymerase (RdRp) and pharmacokinetic and toxicological properties

Objective: To quantify the quassinoids of P. sprucei, a medicinal plant that is native to the Amazon region, using qNMR and investigate the inhibitory potential of isobrucein B and neosergeolide on the 3CLpro and RdRp targets of SARS-CoV-2 through in silico approaches. Methods: the quantification was performed in a fraction (F2-F3) enriched with the quassinoids isobrucein B and neosergeolide using the PULCON method. In silico assays were performed using molecular docking to assess interactions and binding affinity between neosergeolide and isobrucein B ligands with SARS-CoV-2 3CLpro and RdRp targets, and online servers were used to estimate pharmacokinetic and toxicity. Results: It was possible to determine the quantity of the two quassinoids isobrucein B and neosergeolide in the F2-F3 fraction (769.6 mg), which were present in significant amounts in the PsMeOH extract (5.46%). The results of the docking analysis, based on the crystallized structures of RdRp and 3CLpro, indicated that isobrucein B and neosergeolide are potential inhibitors of the two proteins evaluated, as well as showing the importance of hydrogen bonding and pi (π) interactions for the active sites foreseen for each target. Conclusion: The results suggest that P. Research, Society and Development, v. 10, n. 16, e69101623220, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i16.23220 2 sprucei quassinoids may interact with 3CLpro and RdRp targets. In vitro and in vivo experiments are needed to confirm the results of molecular docking and investigate the risks of using P. sprucei as a medicinal plant against COVID-19.


Introduction
The SARS-CoV-2 coronavirus pandemic emerged in Wuhan, China, with the first infections being reported in December 2019 . Due to its rapid spread, by June 2021, more than 178 million cases of the disease had been registered worldwide, and deaths already surpass three million people (World Health Organization [WHO], 2021).
Representing an unprecedented challenge, extensive research into developing new antiviral drugs for COVID-19 has identified potential targets that play an important role in blocking viral infection (Ghahremanpour et al., 2020).
The 3Clpro and RdRp targets have great importance in the viral replication of SARS-CoV-2 and stand out as essential targets in computational strategies such as molecular docking (Qamar et al., 2020;Wu et al., 2020;Das et al., 2020;Yu et al., 2020). Molecular docking is currently being used as a powerful tool in addition to being rational and low-cost for screening new antiviral agents against COVID-19 and allows us to understand how these NSPs are important targets of SARS-Cov-2 and how they will interact with ligands at the active site (Qamar et al., 2020;Wu et al., 2020;Das et al., 2020;Yu et al., 2020;.
Regarding ligands, drugs used against other human diseases and natural products present in medicinal plants, such as flavonoids, and other bioactive compounds, such as quassinoids, have been used as candidates in virtual screening approaches.
(Da Silva et al., 2020;Neves et al., 2021). In addition, the quassinoids are potential candidates in virtual screening approaches in the search for new agents against COVID-19 (Hasan et al., 2020;Qamar et al., 2020;. Because they do not have easy access to medicines, a large part of the world population, including indigenous tribes, quilombolas, and riverine communities in the Amazon, resort to the use of medicinal plants, which in most cases are the only source of medicine for the treatment or cure of diseases (Le Cointe, 1947;Da Silva et al., 2020;) Picrolemma sprucei Hook, from the Simaroubaceae family, is a native plant of the Amazon region that is popularly used as a medicinal plant, mainly for the treatment of malaria (Nunomura, 2006;Pohlit et al., 2009b;Ribeiro et al., 1999). It is popularly known in Brazil as "caferana" due to its similarity to Coffea arabica L. It can be found in street markets and through herbalists in the region and is marketed as an antimalarial and for the treatment of fevers, in the Adolpho Lisboa Municipal Market in Manaus (Cavalcante, 1983;Thomas 1990;Ribeiro et al., 1999;Ferreira, 2000;Saraiva et.al., 2002).
The main constituents of P. sprucei are the quassinoids, which are only produced by species of this family.
Quassinoids are biodegraded triterpenes and are highly oxygenated with a wide range of biological activities. Many quassinoids exhibit a wide range of biological activities in vitro and/or in vivo, including antiviral, antitumor, antimalarial, antiinflammatory, antifeedant, insecticidal, amoebicidal, antiulcer, and herbicidal activities (Guo et al., 2005).
Previous research on the roots and stems of P. sprucei resulted in the isolation of the main quassinoids of this species, neosergeolide and isobrucein B (Moretti et al., 1982;Zukerman-schpecto &;Castellano, 1994, Vieira et al., 2000, Saraiva, 2001, Amorim, 2009. Nunomura et al. (2006) demonstrated in vitro anthelmintic activity of isobrucein B and neosergeolide isolated from the chloroform fraction of P. sprucei against Haemonchus contortus. This nematode parasitizes the gastrointestinal system of sheep and other animals. Other in vitro studies report antitumor, antimalarial, anthelmintic, cytotoxic, insecticidal, and anti-leishmanial activities of neosergeolide and isobrucein B (Fukamiya, et. al., 2005;Almeida et al., 2007;Andrade-Neto, et al., 2007;Silva et al., 2009b). Therefore, the present study quantified the quassinoids neosergeolide and isobrucein B of P. sprucei using NMR and the PUlse Length-based CONcentration determination methodology (PULCON). qNMR has shown to be a less laborious and high precision alternative in the simultaneous quantification of organic compounds in a complex mixture (Oliveira et al., 2021). In addition, the study carried out the tracking analysis of these quassinoids using molecular docking, in order to test them as potential inhibitors of SARS-CoV-23CLpro and RdRp and is the first work to report this analysis of neosergeolide and isobrucein B against COVID-19 target enzymes.

Plant material processing
The plant material was dried in a forced air circulation oven at 40 ºC until completely dry. After drying, the material was separated into roots, stems, leaves, and fruits. Then, the part of the plant of interest in this work, the roots, was chopped into lengths of about 10 cm using pruning scissors and a hatchet, followed by fine grinding in a knife mill to obtain a fine powder (302.54 g).

Preparation of the plant extract
The P. sprucei root powder (302.54 g) was subjected to cold maceration (2 weeks) with hexane (1.0 L) with the aid of ultrasound to remove non-polar substances. This material was later extracted with methanol (1.0 L) via cold maceration. The methanol extract of P. sprucei root powder (PsMeOH) was obtained by exhaustive cold maceration (3 weeks). The PsMeOH was concentrated in a rotary evaporator (Fisatom ® 801) at a temperature of 40 °C to remove all the solvent. After the complete drying of the PsMeoH extract, it was weighed to calculate the percentage of yield.

Analysis of fractions obtained from fractionation of PsMeOH extract by mass spectrometry (MS)
The spectrometry analyses of the fractions obtained from the fractionation of the PsMeOH extract were carried out at the Multidisciplinary Support Center -Analytical Center -CAM, located at the Federal University of Amazonas. The mass spectra were obtained by analysis in a triple-quadrupole spectrometer, (model TSQ Quantum Access Thermo scientific), equipped with an atmospheric pressure chemical ionization (APCI) chamber. For the analysis, 30 µL of each fraction was pipetted (filtered) and transferred to Eppendorf tubes and diluted in 1000 µL MeOH; this procedure was performed for the six samples. After this procedure, the samples were injected into the mass spectrometer with the APCI source in positive acquisition mode in a mass range (m/z) of 100-1000 Da. The obtained spectra were analyzed using Xcalibur ® 2.0.7 software.

Identification and quantification of quassinoids via the PULCON method
The nuclear magnetic resonance (NMR) experiments were carried out at the Nuclear Magnetic Resonance Laboratory (NMRLab) of the Analytical Center (CA) of the Federal University of Amazonas (UFAM). One-dimensional ( 1 H, 13 C, and DEPT135) and two-dimensional (COSY, HSQC, and HMBC) NMR spectra were obtained in a nuclear magnetic resonance spectrometer (Bruker Avance III HD, 500.13 MHz for 1 H and 125.8 MHz for 13 C, BBFO Plus SmartProbe™). Chemical shifts were expressed in ppm () and coupling constants (J) were recorded in (Hz). The TopSpin 4.1 program was used to manipulate the NMR spectra.
The qNMR of the quassinoid fraction based on the PULCON method was evaluated using the ERETIC2 (Electronic Referencing to Access In Vivo Concentrations) tool in TopSpin 4.1 software (Tyburn & Coutant, 2016). The pulse sequence used was the zg90, with time-domain data points (TD) of 64 k, spectral width (SWH) of 10 kHz, relaxation delay (D1) of 20.7 seconds, acquisition time (AQ) of 3.27 seconds, 16 scan numbers (NS) with DS of 2, decomposition resolution of 0.30 Hz, receiver gain constant at 32 with offset frequency set to 3088, 30 Hz. The calibration pulse (P1 10.313 s) was generated with PLW9 of 7.183.10-5 W. The estimated value of D1 (20.7 s) was seven times the value of T1 for the isobrucein B quassinoid signal (6.03 ppm). Baselines and phases of the samples' 1H NMR spectra were automatically corrected using TopSpin 4.1.
The quantification of the quassinoid fraction was performed in triplicate and consisted of preparing an aliquot (10.1 mg) solubilized in 550 mL of MeOD with TMS as a reference standard (0.0 ppm). The lengths of the 90 o pulse were 10.313 s of the signals of neosergeolide ( 5.75) and isobrucein ( 6.03) and were quantified using the PULCON method. The standard quinine solution (98.0% purity) was prepared (n=3) in concentrations of 13.66 mM in CD3OD, with TMS as an internal reference (0.0 ppm). The lengths of the 90 o pulse were 9.976 s of the doublet in 8.65 ppm of quinine. The longitudinal relaxation constant (T1) of the hydrogens of each of the two quassinoids and the quinine to be quantified were determined using the inversion recovery experiment (t1ir1d). The measured acquisitions of quinine were 17.11 s (> 7xT1).

Molecular docking studies of quassinoids and targets of SARS-CoV-2 2.7.1 Chemical structures
The 2D structures of isobrucein B and neosergeolide were drawn on Chemdraw Ultra v.12.0. The structure's SMILES were obtained using the interface of the website http://swissadme.ch/. The 2D structures were imported to Discovery Studio Visualizer v.20 for the obtention of the 3D structures. The same procedure was done for the SARS-CoV-2 3CLpro and RdRp inhibitors, namely X77 and theaflavin, respectively.

Pharmacokinetic predictions
The pharmacokinetic properties and drug-likeness predictions of the quassinoids were performed using SwissADME (Daina et al, 2017) from the structure's SMILES. Predictions included an evaluation of the following parameters: Lipinski's rule of five (molecular weight -MW; hydrogen bond acceptor -HBA; hydrogen bond donor -HBD; and octanol/water partition coefficient -LogP) and several other parameters/properties such as molecular polar surface area (TPSA), logarithm of solubility (LogS), gastrointestinal (GI) absorption, and blood-brain barrier (BBB) permeation. The same calculations were performed with X77 and theaflavin to compare the pharmacokinetic predictions with the SARS-CoV-2 inhibitors.

Toxicity predictions
The compound's toxicity was predicted using the pKCSM (Pires et al., 2015), AdmetSAR (Cheng et al., 2012), Pred-hERG (Braga et al., 2015), and Pred-Skin (Braga et al., 2017) computational tools based on the structure's SMILES. For this purpose, several toxicological parameters were considered, namely Ames toxicity, carcinogenicity, hERG inhibition, skin sensitization, and hepatotoxicity. The results were expressed as "-" for non-detected or "+" for detected risk. The same calculations were performed with X77 and theaflavin.

Docking simulations -Ligand preparation
The 3D structures obtained for the compounds were subjected to geometry optimization by the Dreiding-like forcefield (Hahn, 1995) method in Discovery Studio Visualizer v.20, and the results were saved in Protein Data Bank (PDB) format.
These structures were prepared using Discovery Studio Visualizer v.20 tools. Water molecules and bound ligands were removed, polar hydrogens and Kollman charges were added, and the non-polar hydrogens were merged, and the results were saved in PDB format. Pymol v.2.3.4. was then used to visualize the docking results.

Docking calculations
Docking simulations of compounds with SARS-CoV-2 3CLpro and RdRp structures were performed using the DockThor server. For 3CLpro, the grid box was centered at x = -20.810, y = 19.141 and z = -29.186, with x = 24 Å, y = 24 Å and z = 24 Å size. For RdRp, the grid box was centered at the x = 117.382, y = 111.853 and z = 121.073, with x = 24 Å, y = 24 Å and z = 24 Å size. The most energetically favorable conformations were selected for the analysis. The advanced options were maintained as default. The docking protocol was validated by redocking the ligand X77 complexed with the 3CLpro structure. Validation was considered successful if the 10 top-ranked redocked orientations showed heavy atom RMSD values ≤ 2.0 when compared with crystallographic orientation. Due to the lack of models with ligands for RdRp in the RCSB PDB, only the 3CLpro was tested for redocking.

Statistical analysis
The quantitative NMR data value were compared using the T-test for single comparison and ANOVA (one-way) with a Tukey test (95 % confidence) for multiple comparison using the Minitab ® 18.1 program (State College, PA, USA), for which the results were expressed as mean ± standard deviation.

Results and Discussion
After the complete drying of the PsMeOH extract, its weight was 10.4g with a yield of 3.4% for the mass of plant material (302.54 g). The spectrometric analysis (Figure 1)    The PULCON method correlates absolute intensities of two different spectra and using the principle of reciprocity, which indicates that pulse lengths are inversely proportional to the intensity of the NMR signal, and the analyte concentration is determined (Burton et al, 2005). .03), and it was possible to observe the ratio of 55.9:44.1, respectively. A relative concentration was used as a function of a known compound to determine the mass of each quassinoid in the total mass. Table 1 shows the mass of each quassinoid considering the total mass of the F2-F3 fraction.  The quantitative analysis of quassinoids in root-based teas carried out by Nunomura et.al. (2012) provided results different from those presented in this work since ours showed a higher concentration of isobrucein B than neosergeolide.
However, it should be considered that since isobrucein B is more polar than neosergeolide, it is possibly better extracted by using water as a solvent. The quantification showed the quassinoids as the most representative compounds present in the PsMeOH extract of the P. sprucei roots since the percentage of the two quassinoids adds up to 5.46% of the total mass (10.4 g) of the PsMeOH extract.
Considering that the quassinoids present in the species are potentially active for several biological activities, a theoretical study of these compounds against the proteins of SARS-CoV-2 was carried out. However, being highly bioactive is not a good enough criterion to qualify the compound as a good candidate (Guan et al., 2019). A better pharmacokinetic profile is extremely important for a novel compound examined in the process of drug/drug-like compound discovery. Hence, it is imperative to evaluate the ADMET profile of new compounds earlier to avoid a possible waste of time/resources (Shou, 2020).
The ADMET analysis results revealed the physicochemical properties of the quassinoids and the SARS-CoV-2 3CLpro and RdRp inhibitors, as represented in Table 2. According to Lipinski's rule of five (Lipinski et al., 2001), most of the compounds followed the rules by causing no more than one violation. That is to say, the MW, HBA, HBD, and LogP are within the acceptable range. However, the quassinoids were predicted as having low GI absorption, and thus not good in an orally bioavailable sense, and this may be due to its polar character (LogS and TPSA calculated). Only X77 was predicted as having high GI absorption. On the other hand, none of the compounds were predicted to have BBB permeation. In general, the quassinoids and the SARS-CoV-2 inhibitors possess a good pharmacokinetic profile. The toxicophorical analysis of the compounds was performed with tools that are capable of identifying the main chemical substructures related to toxicity data, thus guiding the design of inhibitors with low toxic potential (Whu & Wang, 2018). This strategy intends to minimize common toxicity failures in the early stages of drug discovery (Rim, 2020). The result of the toxicity analysis of the quassinoids and the SARS-CoV-2 3CLpro and RdRp inhibitors is represented in Table 3.
Prediction data revealed no risks regarding carcinogenicity. In contrast, a predicted risk of skin sensitization was attributed to the compounds, which is a potential adverse effect for dermally applied products (Kleinstreuer et al., 2018). Only Neosergeolide and X77 were predicted as non-inhibitors of the hERG channel, for which inhibition may lead to ventricular arrhythmia (Priest et al., 2008). On the other hand, a detected risk of Ames toxicity and hepatoxicity was predicted for X77.
The Ames test is a short-term bacterial reverse test for mutagenicity prediction (Mortelmans & Zeiger, 2000). In general, we can observe that the quassinoids possess lower toxicological potential than the SARS-CoV-2 inhibitors. Molecular docking analysis were performed to verify the conformation of the quassinoids within SARS-CoV-2 targets and the binding affinities. First, the redocking procedure with the ligand X77 complexed in SARS-CoV-2 3CLpro binding site revealed that all the 10-top ranked orientations presented heavy atoms RMSD < 2 when compared with the crystallographic conformation ( Figure 5). The docking protocol was able to reproduce the conformation and the binding mode of X77, thus validating our docking simulations. Docking simulations between the compounds and SARS-CoV-2 3CLpro revealed that they bind at the receptorbinding pocket mainly by hydrogen bond (HB) and Van der Waals interactions (Figure 7 and 8). The compounds presented interactions with the receptor residues such as those observed in the inhibitor's binding mode ( Figure 6 and Table 4) and are Source: Authors (2021).
thus capable of interacting with the catalytic dyad (His41 and Cys145) of the receptor (Qamar et al., 2020). The catalytically active 3CLpro is a dimer. Cleavage by 3CLpro occurs in the glutamine residue of the substrate via the protease Cys-145 and His-41 dyad in which the cysteine thiol functions as the nucleophile in the proteolytic process (Anand et al., 2003). Predicted binding affinities showed that, although X77 presented the stronger affinity, the quassinoids showed similar values compared to X77 (Table 4).    Docking simulations with SARS-CoV-2 RdRp revealed that the quassinoids may bind at the nucleoside triphosphates (NTP) binding pocket. This active site includes motif A (residues 611-626), with the classic divalent cation-binding residue 618, and motif C (residues 753-767), with the catalytic residues 759-761 . The compounds isobrucein B, neosergeolide and theaflavin performed interactions with these key residues mainly by HB, Van der Waals, and Pi interactions ( Figure 9-11). The binding across the RNA tunnel potentially blocks the access of NTPs to the active site on one side and of RNA on the other. Without NTPs, RNA replication cannot occur, thus blocking viral replication (Pokhrel et al 2020). The binding affinities of the compounds with the receptor were similar when compared to the values obtained for theaflavin (Table   5).    The results of the molecular docking simulations in this work propose the quassinoids' capacity to interact with key amino acids residues in SARS-CoV-2 targets, which shows the potential of isobrucein B and neosergeolide as possible inhibitors of the SARS-CoV-2 target proteins 3CLpro and RdRp.Few molecular docking studies using quassinoids and SARS-Cov-2 have been reported, and no studies using quassinoids from P. sprucei and SARS-CoV-2 have been reported yet. Some studies have shown that compounds of natural origin have potential SARS-CoV-2 inhibitors.
Rahmatullah et al. (2020) carried out a molecular docking study with the 3CLpro protein from SARS-CoV-2 with 12 Javanicins, which are quassinoids found in the Picrasma javanica species from the Simaroubaceae family. The quassinoid Javanicin G showed the greatest affinity for SARS-CoV-2 3CLpro with a binding energy of -8.2 kcal/mol. This result found by these authors corroborates the result found for neosergeolide and 3CLpro protein, in which neosergeolide also presented an affinity of 8.248 kcal/mol.  reported the interaction of 7 quassinoid javanicolides found in Brucea javanica with 3CLpro.
Javanicolide B showed a higher affinity for 3CLpro with a binding energy of -7.9 kcal/mol. Pele et al. (2020) evaluated some antimalarial compounds of natural origin for SARS-CoV-2 3CLpro. In the study, the following antimalarials quinine, primaquine, atovaquone, artesunate, and proguanil presented binding affinity values -6.508 kcal/mol, -6.361 kcal/mol, -5.493 kcal/mol, -4.862 kcal/mol, and -4.842 kcal /mol, respectively. These values are lower than the quassinoids isobrucein B and neosergeolide, for which the binding affinity value was -8.024kcal/mol and -8.248 kcal/mol for this target. Guedes et al. (2021) evaluated the repositioning of drugs using the SARS-CoV-2 3CLpro and RdRp targets. These drugs are already used in clinical practice and were analyzed against the SARS-CoV-2 targets mentioned above. Some of these drugs had a binding affinity value for 3Clpro and RdRp that was lower than the quassinoids studied in this work, bearing in mind that the more negative (and further from zero) the value of ΔG' o , the more affinity the ligand has for the target. Some antimalarial drugs, such as chloroquine, hydroxychloroquine, and mefloquine, had values of -8.06 kcal/mol, -7.83 kcal/mol, and -7.46 kcal/mol, respectively for the 3CLpro target. In contrast, the quassinoid neosergeolide showed a higher affinity with a value of -8.248 kcal/mol for the same target. Isobrucein B had greater affinity than the values shown by hydroxychloroquine and mefloquine, and a value close to the value of chloroquine with a binding affinity value of -8.024 kcal/mol.
As for the RdRp target of SARS-CoV-2 in the same study by Guedes et al. (2021), the binding affinity values for the drugs chloroquine, hydroxychloroquine, and mefloquine were -7.49 kcal/mol, -7.42 kcal/mol, and -7.03 kcal/mol, respectively.
Isobrucein B had a binding affinity value of -7.500 kcal/mol, slightly higher than those of chloroquine and hydroxychloroquine, and a higher binding affinity for the RdRp target than mefloquine. For this target, the quassinoid neosergeolide had an affinity value a little lower than that of chloroquine and hydroxychloroquine and greater than the value of mefloquine, with a binding affinity of -7.388 kcal/mol.

Conclusion
Using the qNMR technique, which until now had not been used for the quassinoids of the species P. sprucei, it was possible to quantify them in the F2-F3 fraction. The in silico results suggest that the quassinoids isobrucein B and neosergeolide are interesting starting points for obtaining novel SARS-CoV-2 inhibitors with therapeutical potential in the treatment of COVID-19. This is the first study of the binding of P. sprucei quassinoids (by molecular docking) to key proteins of SARS-CoV-2 that play an important role in inhibiting virus replication. Thus, these two quassinoids may play an essential role as inhibitors of SARS-CoV-2 3CLpro and RdRp.
Analyses in vitro and in vivo are needed to corroborate the reported docking results and the proper application and further optimization of these substances. The knowledge obtained in this study may be essential for exploring and developing new candidates for natural therapeutic agents against COVID-19. Furthermore, it is also fundamental that we investigate the risks of P. sprucei as a phytomedicine for use against COVID-19.