Antitrypanosomal potential of the red sea soft coral Nephthea mollis supported by metabolomics profiling and molecular docking studies

The crude extract and its derived ethyl acetate fraction of the soft coral Nephthea mollis displayed remarkable in vitro antitrypanosomal activity against Trypanosoma brucei with IC 50 values of 6.4 and 3.7 µg/ml, respectively. Consequently, the crude extract was subjected to LC–HR–ESI–MS metabolomics profiling to identify the constituents that may underlie their bioactivities. As a result, 33 secondary metabolites were characterized, among which sesquiterpenes and diterpenes prevailed. In silico molecular modeling revealed the high binding affinity for the ornithine decarboxylase active site, with five dereplicated compounds having docking scores higher than the cocrystallized ligand. These results highlight N. mollis as a rich source of compounds that might help develop therapies for Human African trypanosomiasis.


INTRODUCTION
Human African trypanosomiasis (HAT) is one of the neglected tropical diseases that represent a serious threat to humans, especially in low-income countries of Africa, Asia, and South America (Molyneux et al., 2017). The geographical distribution of the sleeping sickness or HAT is restricted by a distribution of tsetse flies, which is found in the rural areas of more than 36 African countries (Brun and Blum, 2012;De Koning, 2020;Gayer et al., 2007). This vector-borne parasitic disease, caused by two subspecies of Trypanosoma brucei parasite: Trypanosoma brucei gambiense (accounts for 97% of cases) and Trypanosoma brucei rhodesiense, is considered fatal if left untreated (Pays, 1999;Steverding, 2010). Drug toxicities, resistance, and route of administration (either intramuscular and/or intravenous) raise the urgent need for new drugs orally effective against HAT (Álvarez-Bardón et al., 2020). Although extracts of natural products present many challenges to drug discovery programs worldwide (because of the chemical diversity found in a single extract), it is still known that bioactive compounds from natural products offer an incomparable and safer source for various disease therapies.
Moreover, the genus Nephthea has been reported to be rich in secondary metabolites such as sterols, sesquiterpenes, and diterpenes. Many of these metabolites recently showed antiprotozoal activities against various parasitic species such as trypanosome, malaria, and leishmania (Allam et al., 2021;Álvarez-Bardón et al., 2020).
Metabolomics is a valuable analytical tool used to comprehensively and highly accurately determine metabolites in a given metabolome (Bakr et al., 2021;Jacob et al., 2019;Said et al., 2021). It allows rapid identification (dereplication) and quantification of secondary metabolites in extracts. The combination of biological activity testing of crude extracts with metabolomics accelerates drug discovery, partly because crude extracts show higher biological activity and partly due to the ability to discriminate between complex mixtures of metabolites Kjer et al., 2010;Lotfy et al., 2021;Tawfike et al., 2013;Yuliana et al., 2011).
Ornithine decarboxylase (ODC), a proven drug target for the treatment of sleeping sickness, is the enzyme responsible for the metabolism of polyamines (such as putrescine, spermidine, and spermine) which are essential for cellular replication and hence parasite survival (Ibezim et al., 2017a). Trypanosome-specific metabolic and cellular pathways represent excellent molecular targets. The ability to synthesize polyamines, putrescine, and spermidine is vital for the proliferation of bloodstream forms in trypanosomes. In this process, ODC has a crucial function. This enzyme is considered the best-validated drug target in T. brucei, which is the target of eflornithine, a drug used clinically for treating HAT.
TTherefore, we aim to evaluate the in vitro antitrypanosomal activity of the crude extract and its derived fractions; petroleum ether and ethyl acetate of Nephthea mollis assisted by Liquid Chromatography-High Resolution-Electrospray Ionization -Mass Spectrometry (LC-HR-ESI-MS) metabolomics profiling to identify the secondary metabolites responsible for the antitrypanosomal potential of the soft coral. Additionally, we performed in silico molecular docking simulations within ODC active site to explore the binding mode of dereplicated compounds.

Chemicals and reagents
Solvents used in this study (purchased from El-Nasr Company for Pharmaceuticals and Chemicals, Egypt), including ethanol, ethyl acetate, and petroleum ether (b.p. 60°C-80°C), were of analytical grade and distilled before use. Solvents of HPLC grade, such as acetonitrile and methanol, were obtained from SDFCL sd fine-Chem Limited, India. Suramin (Germanin ® ) was obtained from Bayer AG, Leverkusen, Germany.

Extract preparation
The freeze-dried soft coral material (45 g) was extracted with ethanol until exhaustion. The crude ethanol extract (15 g, 33%) was suspended in distilled water and fractionated successively with petroleum ether and ethyl acetate. The organic phase in each step was separately concentrated under a vacuum, yielding the petroleum ether fraction (6 g), the ethyl acetate fraction (1 g), and the aqueous fraction (5 g). The crude extract and its derived fractions, petroleum ether and ethyl acetate, were kept at 4°C ± 1°C for antitrypanosomal assay and other analysis.

In vitro antitrypanosomal assay
The antitrypanosomal activity was tested against T. brucei following the protocol of Huber and Koella (1993). A number of 10 4 trypanosomes per milliliter of T. brucei strain TC 221 was cultivated in a complete Baltz medium. Trypanosomes were tested in 96-well plate chambers against different concentrations of test extracts (20 µL of stock solutions 10-200 µg/ml) diluted with 1 % Dimethyl Sulfoxide (DMSO) to a final volume of 200 µL. For control, 1% DMSO, as well as parasites without any test extracts, were used simultaneously in each plate to show an absence of any activity due to 1% DMSO. The plates were then incubated at 37°C ± 1 °C in an atmosphere of 5% CO 2 for 24 hours. After adding 20 µl of AlamarBlue, the activity was measured after 48 and 72 hours by light absorption using an MR 700 Microplate Reader at a wavelength of 550 nm with a reference wavelength of 650 nm. Linear interpolating three independent measurements quantified IC 50 values of the tested extracts. Suramin was used as a positive control (IC 50 0.23 µg/ml) (Abdelmohsen et al., 2010;Cos et al., 2006).

LC-MS metabolomics profiling
Metabolomics profiling of the crude extract of N. mollis (Elmaidomy et al., 2019) using an ACQUITY ultra-performance liquid chromatography system connected to a SYNAPT G2 HDMS quadrupole time-of-flight hybrid mass spectrometer (Waters, Milford, USA). Chromatographic separation was carried out using a BEH C18 column (2.1 × 100 mm, 1.7 μm particle size; Waters, Milford, USA) accompanied by a guard column (2.1 × 5 mm, 1.7 μm particle size). The mobile phase used during the separation consisted of purified water (A) and acetonitrile (B) with 0.1% formic acid added to each solvent. A gradient elution started at a flow rate of 300 μl/minute with 10% B linearly increasing to 100% B within 30 minutes and remaining isocratic for the next 5 minutes before linearly decreasing back to 10% B for the following 1 minute. The volume injected was 2 μl, and the column temperature was adjusted to 40°C ± 1°C. The obtained raw data were separated into positive and negative ionization modes using MSConvert software. Accordingly, the files were imported to the data mining software MZmine version 2.10 for peak picking, followed by deconvolution, deisotoping, alignment, and formula prediction. In MZmine, a framework for differential analysis of MS data, the RAW data are imported by selecting the ProteoWizard-converted positive or negative files in mzML format was carried out as described by et al. (2010) Abdelmohsen followed by peaks detection in the sample using the chromatogram builder with a minimum time span set to 0.2 minutes, the minimum height and m/z tolerance to 1.0 × 10 4 and 0.001 m/z or 5.0 ppm, respectively. Mass ion peaks were isolated with a centroid detector threshold, greater than the noise level set to 1.0 × 10 4 and an MS level of one. For all remaining steps, select all files under peak lists before executing each step. Chromatogram deconvolution was then performed to detect the individual peaks. The local minimum search algorithm (chromatographic threshold: 95%, search minimum in RT range: 0.4 minutes, minimum relative height: 5%, minimum absolute height: 3.0 × 10 4 , minimum ratio of peak top/edge: 3 and peak duration range: 0.2-5 minute) was applied. The isotopic peaks grouper was used (m/z tolerance: 0.001 m/z or 5.0 ppm, retention time tolerance: 0.1 absolute (minute), maximum charge: 2, and representative isotope: most intense) to identify the isotopes. This step will only deisotope peaks detected in the original search, i.e., those assigned a peak ID.
Filtering is helpful to set certain parameters when only considering a certain RT window, e.g., 5-40 minutes or m/z range window, or discard IDs that are only present in the sample. For chromatographic alignment and gap-filling, the retention time normalizer (m/z tolerance: 0.001 m/z or 5.0 ppm, retention time tolerance: 0.5 absolute (minute), and minimum standard intensity: 5.0 × 10 3 ) was used to reduce interbatch variation. The peak lists were all aligned using the join aligner parameters set to m/z tolerance: 0.001 m/z or 5.0 ppm, weight for m/z: 20, retention time tolerance: 5.0 relative (%), and weight for RT: 20. The values for the weight of m/z and RT should be kept the same; this means that both RT and m/z are given equal importance. Missing peaks were detected using the gap-filling peak finder with an intensity tolerance of 25%, m/z tolerance of 0.001 m/z or 5.0 ppm and retention time tolerance of 0.5 absolute (minute), followed by creating a file called "Neg-gap filled" if negative mode and "Posgap filled" if positive mode. An adduct search was performed for Na-H, K-H, NH 4 , formate, and ACN+H (RT tolerance: 0.2 absolute (minute), m/z tolerance: 0.001 m/z or 5.0 ppm, max relative adduct peak height: 30%).
Additionally, a complex search was performed (ionization method: [M+H] + for ESI positive mode and [M-H] − for ESI negative mode, retention time tolerance: 0.2 absolute (minute), m/z tolerance: 0.001 m/z or 5.0 ppm and with a maximum complex peak height of 50%). The processed data set was then subjected to molecular formula prediction and peak identification to search for unidentified features and select atoms C, H, N, O, and other elements. Adjust parameters with heuristics element count with all three suboptions to get the isotope pattern filter working with all features with isotope peaks (Mahmoud et al., 2021).
Excel macros were written to enable the subtraction of background peaks. The Excel macro was used to dereplicate each m/z ion peak with compounds in the customized database (using RT and m/z threshold of ±3 ppm), which provided details on the putative identities of all metabolites. The macro was then utilized to identify the top 20 features (ranked by peak intensity) and to correspond putative identities in the sample by creating a list for the sample.
The Dictionary of Natural Products (DNP), METLIN, and MarinLit databases were used to identify compounds (Ancheeva et al., 2018).

In silico molecular docking
In silico molecular docking simulations were performed for dereplicated compounds of N. mollis within the ODC (PDB ID: 1NJJ) active site. RCSB PDB deposited crystal structure resolution was 2.45 Å and it was used as a PDB entry for ODC since it was cocrystallized with a small molecule natural substrate capable of stimulating the activity of ODC in a biochemical and cell-based assay. Structures of dereplicated compounds were prepared in ChemDraw ® Ultra (v. 8, 2013) and were transferred as smiles to Builder software embedded in Molecular Operating Environment (MOE 2014.0901; Chemical Computing Group, Montreal, QC, Canada) software. Protein-ligand complex was retrieved from the protein data bank, and the cocrystallized water molecules and nonessential small molecules were deleted. After carefully examining the topologies, the retained protein-ligand complex was subjected to energy minimization using GROMACS 4.5.5 and the GROMOS force field 53a6 (Berman et al., 2000;Scholz et al., 2015). The cocrystallized ligands were removed from the protein and redocked using the London dG scoring function implemented in MOE. Their structures were protonated at a cut-off value of 15 Å. Validation of the docking process of cocrystallized ligand (ornithine ORX) and MD simulations of all test compounds from N. mollis were done by the Triangle Matcher method, which was preferred because it has been reported to be more systematic in its ligand atoms alignment than other methods like Alpha Triangle. The poses generated from the placement stage were refined using the Forcefield refinement scheme according to its default settings. The scheme uses the molecular mechanics' setup and the reaction field functional form to perform energy minimization and treat the solvent effect (Ibezim 2017a(Ibezim , 2017b.
The free energy of binding of a given pose is estimated by the London dG scoring function from the sum of each of these terms: average gain/loss of rotational/translational entropy, energy due to loss of flexibility of the ligand, H-bond energies, energy due to metal-ligation and desolvation energy of atoms. London dG rescores 1 for only 10 retained docking poses of each compound, taking the lowest free energy of binding.
The resultant docking poses for each compound were examined and arranged according to their free binding energy value (S; Kcal/mol) and the root-mean-square deviation (RMSD; Å), using the lowest RMSD values relative to the crystal structure binding poses (Jackson et al., 2003;Shady et al., 2020).

Antitrypanosomal activity
Using the protocol of Huber and Koella (1993), the crude extract, and its derived fractions, petroleum ether, and ethyl acetate of N. mollis were tested for their in vitro antitrypanosomal effects against T. brucei. The results revealed that the crude extract and ethyl acetate fraction of the soft coral N. mollis displayed remarkable in vitro antitrypanosomal effect against T. brucei with IC 50 values of 6.4 and 3.7 µg/ml, respectively, while the petroleum ether fraction shows a weak antitrypanosomal activity with IC 50 value of (>50 µg/ml). This encouraged us to undergo further metabolomics analysis of the active crude extract to reveal its constituents.
The proposed mechanism of action of sesquiterpenes in the parasite cells may be due to the formation of thiol adducts with components found in the intracellular medium (namely trypanothione, glutathione, and thiol groups in proteins). The parasite's cells become more vulnerable to oxidative stress with a reduction in trypanothione and, hence, it was suggested that compounds induced programmed cell death in the tested parasite (Zeouk et al., 2020). In another way, the proposed mechanism of action of diterpenes in the parasite cells may be due to their considerable permeabilization in the membrane of the parasite (Ueno et al., 2018).

Metabolomics analysis
The crude extract of N. mollis was subjected to metabolomics profiling for the first time. The chemical profiling of dereplicated compounds was consistent with literature data on secondary metabolites of N. mollis (Figs. 1 and 2). The dereplicated compounds were identified using databases (DNP, METLIN, and MarinLit databases) (Macintyre et al., 2014). Thirty-three compounds were preliminarily identified based on their mass and summarized in Table 1 and Figure 3. Most of the identified  compounds (27 compounds) were identified as sesquiterpenes type of compounds, in addition to a few diterpenes and terpenoidrelated carboxylic acids compounds.

Molecular docking
For further exploration and to get a better idea about the possible target affected by dereplicated compounds of N. mollis to afford their antitrypanosomal activity, we performed in silico molecular docking simulations within ODC. The X-ray structure of ODC showed ligands; D-ornithine, ORX, a substrate analog, was placed within the ODC active site, and Geneticin (G418) as a weak non-competitive inhibitor occupying an allosteric site of ODC. Molecular docking simulations performed within the active site of ODC showed a number of binding interactions (H-bonding and H-pi interactions) between ORX and a number of amino acid residues (GLU 274,ASP 332,GLY 276,ARG 277,TYR 389,GLY 237,and HIS 197) with binding free energy (expressed as docking score, S) of −4.5058 kcal/mol, as listed in Table 2.
Additionally, as found with ORX, most of the test compounds showed either H-bonding and/or H-pi binding interactions with key amino acid residues: GLU 274, ARG 277, and HIS 197 (Table 3 and Fig. 4). Finally, a visual inspection of the best docking poses of 6,7-epoxy-4-hydroxyguaian-10-ene, chabrolidione A, and nephtheoxydiol showed a good overlay with ORX within ODC active site, as shown in Figure 5. Overall, the docking results showed how good the overlay of the dereplicated compounds of N. mollis with ORX within ODC active site was, which could be used for explaining their antitrypanosomal activity. Table 3. Binding free energy (S; kcal/mol) and binding interactions for co-crystallized ligand (ORX) and Nephthea mollis within ornithine decarboxylase (ODC) active site (PDB ID: 1NJJ; 2.45 Å).

CONCLUSION AND FUTURE PERSPECTIVES
This study highlighted the promising antitrypanosomal activity of the soft coral N. mollis extracts as a possible natural therapy for HAT. The crude extract and its derived ethyl acetate fraction displayed remarkable in vitro antitrypanosomal activity against T. brucei. Thirty-three compounds of sesquiterpenes and diterpenes class of compounds were mined with the help of LC-MSbased metabolomics. Docking studies within ODC crystal structure provided insights about how the dereplicated compounds bind and interact with various amino acid residues lining ODC active site, which was further evidenced by in vitro assays against T. brucei. Thus, further work on the extract of N. mollis will be included in our plans to purify the compounds responsible for their antitrypanosomal activity and perform in vivo experiments and safety studies for the purified compounds. Moreover, it should be further investigated for a supportive role in the pharmaceutical field toward developing new antitrypanosomal drugs.

FUNDING
There is no funding to report.

ETHICAL APPROVALS
This study does not involve experiments on animals or human subjects.

DATA AVAILABILITY
All data generated and analyzed are included in this research article.

PUBLISHER'S NOTE
This journal remains neutral concerning jurisdictional claims in published institutional affiliation.

AUTHORS' CONTRIBUTIONS
K.M.A. did conceptualization, data curation, and writingoriginal draft. Y.A.M. did formal analysis and methodology. U.R.A. carried out writing review and editing. A.I.K. did investigation, A.E.A. validation, M.A.F. project administration, and E.S.E. supervision. All authors discussed the results and contributed to the final manuscript.