Monascin and monascinol, azaphilonoid pigments from Mortierella polycephala AM1: in silico and in vitro targeting of the angiogenic VEGFR2 kinase.

Keywords: azaphilonoid pigments; Mortierella polycephala; molecular docking; VEGFR2 kinase inhibitors

Despite significant advances in biomedical research and the efforts that have been made in the search for novel drugs and treatments during the last few decades, cancer remains a major cause of morbidity, mortality, and public health problem worldwide, with a likelihood of increasing prevalence. According to the World Bank income groups, it has been estimated that the incidence of 12.7 million new cancer cases in 2008, will rise to 21.4 million by 2030 [[1]]. On the other hand, cancer is responsible for approximately 7.6 million deaths worldwide, which is expected to increase to 13.1 million by 2030 [[3]]. Therefore, it is imperative to find novel drugs and treatments to overcome this predicted situation. Fungi represent one of the most prolific sources of bioactive compounds, acting as reservoirs of novel bioactive secondary metabolites, which are endowed by complex structures difficult to be chemically synthesized [[5]]. Owing to their huge diversity and habitation, they present a golden mine for research in the field of drug discovery, possessing unique biocatalytic machinery that provides countless complex natural product scaffolds. Hence, they offer a pool of potentially useful medicinal entities and rich sources of bioactive natural products [[6]], [[7]], [[8]], [[9]], [[10]].

During our search for diverse bioactive compounds from the terrestrial fungus Mortierella polycephala AM1 isolated from Egyptian poultry feather waste, HPLC-MS visualization of the obtained extract through its cultivation on rice solid medium revealed the presence of several closely related compounds shown as orange zones during thin-layer chromatography (TLC) in daylight. Working up of the strain's extract using a series of chromatographic purifications afforded two major compounds belonging to azaphilonoid pigments, assigned as monascin (1) and monascinol (2) (Figure 1). Structures of both compounds were identified on the basis of 1D (1H&13C) and 2D (1H–1H COSY, HMBC) NMR and HRESI-MS spectroscopic data. The antimicrobial properties of compounds 1–2 against a set of microorganisms were studied. Monascin and monascinol (monascuspiloin) have been reported for their potential anticancer activities [[11]].

Graph: Figure 1: Chemical structures of monascin (1) and monascinol (2).

The molecular modeling technique became a useful tool to refine suggested antitumor mechanisms of the compounds. Topoisomerases (Topo) IIα and IIβ were excluded as probable targets whereas VEGFR2 kinase was proposed as an interactive docking environment of monascinol (2) with a moderate affinity and a reasonable number of hydrogen bonds. In the light of our docking investigations, we examined the isolated compounds against the angiogenic VEGFR2 enzyme. The strain was taxonomically identified using morphological and genetic assessments as well.

Poultry industry waste is a very rich medium pushing many fungal species to grow and proliferate. To the best of our knowledge, this is the first record of M. polycephala to be isolated from this waste. According to growth characteristics, the fungal strain AM1 showed hairy colonies of greyish brown coloration with age, displaying reverse brown, reaching 6 cm colony diameter in four days. Microscopic examination revealed that sporangia are globose with 55.5 μm in diameter; collumella is absent; sporangiophores are single or present in small groups of 14.0 μm diameter; and sporangiospores are ovoidal, irregular in shape with 7 × 5.5 μm diameter (Figures 2 and 3). Accordingly, the strain was taxonomically identified as M. polycephala AM1.

Graph: Figure 2: Colonies of M. polycephala AM1.A) Poultry feather waste before incubation; B) colonies of M. polycephala AM1 (red ones) and other fungi after incubation of poultry feather waste; and C) pure colony of M. polycephala AM1 on Czapek-Dox agar plates.

Graph: Figure 3: Sporangia of M. polycephala AM1.A (25×), B and C (100×).

An application of the fungus to a static large-scale fermentation on rice solid medium for 14 days at 28 ± 2 °C afforded a reddish-orange pigment, which after extraction using methanol followed by ethyl acetate and concentration in vacuo delivered the corresponding crude extract. HPLC-MS analysis of the extract displayed multi UV-visible peaks of low-middle polarity, which appeared as yellowish-orange zones during TLC in daylight. However, they did not show color changing on exposing either to aqueous sodium hydroxide or conc sulfuric acid, excluding any of the typical classes of colored compounds produced by the microorganism, i.e. prodigiosins, peri-hydroxyquinones, polyenes (e.g. carotenoids), and actinomycin derivatives. Working up and purification of the strain extract starting with Isolera One (visualized by TLC), preparative thin-layer chromatography (PTLC), open silica gel, and Sephadex LH-20 columns afforded monascin (1) and monascinol (2) as two reddish-orange solids of middle polarity.

As reddish-orange semi-solid, compound 1 was obtained showing no color changing on exposing to sodium hydroxide or conc sulfuric excluding its nature to belong to any of the usual-colored classes of compounds aforementioned, referring to a further different class of colored compounds. According to HR-ESI MS, the molecular formula of compound 1 was deduced as C21H26O5 with a corresponding molecular weight of 358 Da, and 9 double bond equivalents (DBE). Full assignment of compound 1 on the basis of 1H/13C/HSQC NMR (Table 1) and long-range coupling using H,H COSY and HMBC experiments (see Supplementary Material, Figure S9) and comparison with corresponding literature [[13]] assigned it as monascin.

Table 1: 13C (100 MHz) and 1H (400 MHz) NMR data of monascin (1) and monascinol (2) in CDCl3 (δ in ppm).

As further reddish-orange semi-solid, compound 2 with similar chromatographic properties to 1, was obtained with tinny higher polarity according to TLC monitoring. Based on its HR-ESI MS, the molecular formula of 2 was deduced as C21H28O5 with a corresponding molecular weight of 360 Da, and 8 DBE, indicating the strong close relation between both compounds with just two 2H extra in 2 rather than 1, and hence reduction of one double bond being for carbonyl or olefinic bond of 1 was noticed to afford 2.

According to 1H/13C and HSQC NMR (Table 1) spectral data, compound 2 displayed the same NMR pattern of 1, except that the acetonide carbonyl (C-12) shown at δ 202.5 in 1 disappeared and replaced by a hydroxy methine shown at δ 69.3 (δH 4.17), meanwhile the remaining NMR resonances were shown as the same in 1 with slight up-field shifting for the surrounding lactone methine H-11 (δH 2.72; δC 48.9) and methylene group H2-13 (δH 1.55, δC 35.0) of the pentyl chain to C-11. Details of full assignments for structure 2 were done on the basis of long-range 2D NMR (H,H COSY and HMBC) experiments (see Supplementary Material, Figure S17), confirming it as monascinol. Comparison of the NMR data of 2 with those reported recently from Monascus-fermented rice [[14]], confirmed their identical data, however, some positions in the reported one have been incorrectly assigned e.g. 4a- and 8a-positions have been mistakenly exchanged due probably to a missing full 2D NMR assignment in the reported one. In addition, compound 2 was first reported as patent [[15]], reporting no details regarding the structure assignment. So, we preferably proposed to report the full assignment of 2 herein.

Monascin (1) and monascinol (2) are organic hetero-tricyclic compounds, named as azaphilonoid pigments, γ-lactones, α,β-unsaturated ketonic polyketides [[16]]. They have been reported in the extracts of the fungus Monascus pilosus-fermented rice (red-mold rice) [[11]]. The reddish-orange pigments, monascin analogs were reported as constituent metabolites of Monascus purpureus, M. pilosus, and M. ruber. Biologically, they are potent inhibitors of carcinogenesis measured against chemical- or UV-initiated, phorbol-promoted mouse skin tumors, and have roles as antineoplastic agents, peroxisome proliferator-activated receptor (PPAR) gamma agonists, antilipemic drugs, and fungal metabolites [[16]]. Monascin (1) strongly, inhibited differentiation and lipogenesis and stimulated lipolysis effects in a 3T3-L1 preadipocyte model. It had additionally a significant antiobesity effect and represents the first agent found to suppress Niemann-Pick C1 Like 1 (NPC1L1) protein expression associated with small intestine tissue lipid absorption. Importantly, monascin is not like monacolin K, increasing creatine phosphokinase activity, known as a rhabdomyolysis indicator [[16]], [[17]], [[18]], [[19]]. Monascin exhibits further multiple beneficial effects including anti inflammation, antioxidation, antidiabetes, immunomodulation, attenuation of Alzheimer's disease risk factor, and antitumorigenic effects. Monascin not only possesses pleiotropic bioactivity but is also more potent than monacolin K in lowering lipid levels and has lower toxicity [[16]], and acts as the activator of PPARγ agonist/Nrf-2 that subsequently, ameliorate metabolic syndrome [[16]].

The in silico investigation addressed two important targets in the field of direct chemotherapeutic and combined therapy, regarding topoisomerase II [[20]] and the antiangiogenic VEGFR2 kinase inhibition [[21]], respectively.

The compounds under investigation were designed in the spatial structure and minimized to the most stable conformation. The proteins were downloaded from the protein data bank (PDB) [[22]] and prepared by the known protocol. The native ligands were designed using the same manner of constructing the three dimensional (3D) structures of the investigated compounds. Autodock Vina [[23]] was employed for the docking step. The validation of the molecular docking tools was performed by the standard method of the redocking of the native ligand. The orientation of the re-docked native ligand was compared with the real binding pose.

Topoisomerase II could be inhibited by the catalytic pathway or by protein poisoning. The first mechanism is executed by interference with adenosine triphosphatase (ATPase) interaction in the site of activity in a competitive manner, noncompetitively through stabilizes the ATPase dimer [[24]], or by suppression of the DNA synthesis [[25]]. On the other hand, topo II poisoning takes place through the stabilizing of the topo II-DNA complex. The antiproliferative agents in this case act as a bridge between the DNA helix and the protein resulting in preventing their separation which leads to apoptosis [[26]]. The available crystal structures on the PDB [[22]] allowed us to test the competitive inhibition and topo II poisons activities.

The 3D structure of topoisomerase IIα is in interaction with the intrinsic native ligand and topoisomerase IIβ in a complex with DNA during the unwinding of the double helix were obtained. They were the macromolecules targets of the docking process to test the topo II inhibitory activities of the analyzed structures.

During the simulation approach validation of topoisomerase IIα, a slight deviation from the original mode of interaction of the native ligand was noticed with the re-docked one. The binding energy was relatively high (−10.8 kcal/mol) and the pattern of hydrogen bond interaction was preserved (Figure 4). The same parameters were applied for the tested structures, and the results showed a considerable affinity toward the target regarding the binding energy. The number of formed hydrogen bonds of compounds 1 and 2 was little that implied the difficulty of replacing the native ligand on the interaction pocket of the protein.

Graph: Figure 4: The redocking process of the native ligand inside the topo IIα catalytic site, on the right the co-crystallized (green and orange stick) and the re-docked structures (yellow and orange stick) are shown superimposed with a slight deviation.

The available crystal structure of topo IIβ shows its interaction with DNA. It indicates the role of etoposide which is a known anticancer agent in stabilizing the topo IIβ-DNA complex by forming hydrogen bonds with both the DNA helix and the protein. Etoposide was attempted for the docking process evaluation and the re-docked structure showed a good matching with the co-crystallized compound in the pose and the number of hydrogen bonds with stable binding energy of −13.3 kcal/mol (Figure 5). The designed compounds were tried on the site of this interaction and it was found that compounds 1 and 2 formed numerous hydrogen bonds with the nucleotides of DNA and no bond with the protein, which may mean the incapability of them to stabilize the topo IIβ-complex.

Graph: Figure 5: Etoposide in the interface between the DNA helix and topo IIβ, the magnification of the bridging position showed the accuracy of the molecular docking process.

The available crystal structure of VEGFR2 kinase contains a pyrrolo-pyrimidine structure as an inhibitor. The redocking process of this inhibitor showed some deviation from the co-crystallized pose. The number of H-bonds was preserved, and the affinity was relatively high which reflected the reliability of the method. It was noticed that the number of formed hydrogen bonds in the native ligand-protein complex was not so much (only four) and the affinity was −9.5 kcal/mol (Figure 6). The promising results achieved by the docking of our compounds were obtained from compound 2. The number of formed hydrogen bonds of this compound was five and the affinity of the selected mode was moderate (−7.8 kcal/mol) (Figure 7). The inhibitory activity of monascinol (2) is still a possibility while compound 1 did not perform as a good inhibitory candidate during the docking simulation.

Graph: Figure 6: A slight difference in the binding orientation between the theoretical (light pink stick) and the real (white stick) inhibitors inside the binding pocket of VEGFR2 kinase.

Graph: Figure 7: The binding orientation of compound 2 (green stick) inside the VEGFR2 kinase pocket. The amino acids involved in the hydrogen bond interaction (yellow dashes) with the docked structure are shown as orange sticks.

Interestingly, our in vitro results showed a similar trend to the in silico study, against VEGFR2 kinase. Monascinol (2) rather than monascin (1) strongly impacts the enzyme while its inhibitory concentration; IC50 was approximately equipotent to the sorafenib (Figure 8). In contrast, about 3-fold was the determined IC50 of compound 1 when compared to the same standard.

Graph: Figure 8: IC50 of M. polycephala AM1 isolated compounds compared with sorafenib as inhibitors to the VEGFR2 kinase.

The antimicrobial activity of the fungal extract in comparison with its produced metabolites (1–2) was carried out against a set of 10 diverse microorganisms (Staphylococcus aureus ATCC 259239, Escherichia coli ATCC 8739, Pseudomonas aeruginosa ATCC 9027, Bacillus subtilis ATCC 6633, Enterococcus faecalis ATCC 29212, Klebsiella pneumoniae ATCC 13883, Vibrio damsel, Vibrio fluvialis, Candida albicans A13, and Aspergillus flavus MM6.) using agar diffusion assay on paper disk (9 mm) at concentrations of 25, 50, 75, and 100 μg/disk. However, no activity was observed against the whole test organisms.

NMR spectra were recorded on 400 MHz Bruker Avance Neo (av 402) and Bruker AMX 300 (300.135 MHz) spectrometers equipped with a TCI CryoProbe using standard pulse sequences. NMR data were processed using MestReNova 11.0. UHPLC-HRMS was performed on an Agilent Infinity 1290 UHPLC system equipped with a diode array detector. UV–vis spectra were recorded from 190 to 640 nm. TLC analysis was performed on silica gel plates (Sil G/UV254, 0.20 mm, Macherey-Nagel). Biotage Isolera One Flash Chromatography system was used for flash chromatography and performed on silica gel 60 (Merck, 0.04–0.063 mm, 230–400 mesh ASTM). Sephadex LH-20 was bought from Pharmacia. All solvents and chemicals used for HRMS and chromatography were VWR Chemicals LC-MS grade, while for metabolites extraction; the solvents were of HPLC grade (VWR Chemicals).

The fungus M. polycephala AM1 was isolated from Egyptian poultry feather waste. Three gm of the waste was put in a sterile Petri dish, moistened with 3 mL of sterile tap water, and incubated at 28 ± 2 °C for five days. The growing colonies were purified by transfer to Czapek-Dox agar medium (g/L: 30 sucrose, 3 NaNO3, 1 K2HPO4, 0.5 KCl, 0.5 MgSO4, 0.01 FeSO4, and 20 agar-agar) and fortified with 0.1 gm chloramphenicol to prevent bacterial growth. The reported identification was based on current universal keys using an image analysis system [[27]].

Under aseptic conditions, the spore suspension of M. polycephala AM1 (106 spores/mL) has been used to inoculate sterilized bottles (volume 100 mL) containing commercial rice 10 gm; 15 mL distilled water. The bottles were incubated for 14 days at 37 °C. After cultivation, to each bottle 30 mL methanol were added for socking, followed by shaking and filtration under vacuum. After filtration, the water/methanol fraction was evaporated to remove methanol using a rotary evaporator (Heidolph), and afforded water residue was re-extracted by ethyl acetate. The obtained ethyl acetate extract was finally concentrated in vacuo till dryness and then applied to work up stages.

The crude extract (4.0 g) was applied to fractionation using Isolera One with aid of Biotag SNAP Cartridge (50 g) and eluted with using DCM-MeOH gradient starting from 100% DCM, with a slow increase of MeOH from 1 to 15%, and finally washed with pure MeOH (200 mL). According to TLC monitoring, four reddish-orange fractions were afforded: FI (141 mg), FII (42 mg), FIII (0.8 g), FIV (1.4 g). Fractions I and II were combined due to their similarities during TLC and followed by sub-fractionation on Sephadex LH-20 (DCM/50% MeOH), and the obtained components were purified by PTLC (DCM/20% MeOH) and Sephadex LH-20 (DCM/50% MeOH) affording monascin (1, 3.1 mg) and monascinol (2, 18.6 mg) as reddish-orange solids. The remaining fractions were excluded as they contain multicolored components of very low quantities, which will be re-investigated in a fellow work after re-cultivation of the strain in higher large-scale fermentation to get much more quantities of such desired components. Details of spectral data of monascin (1) and monascinol (2) are found in "Supplementary Material".

Antimicrobial assays using the agar diffusion test were performed utilizing the disc agar method [[1]]. Different concentrations of the fungal extract and obtained two compounds (1–2) (dissolved in CH2Cl2/20% MeOH) namely: 25, 50, 75, and 100 μg/mL were soaked on paper discs (9 mm ∅, no. 2668, Schleicher & Schüll, Germany) to obtain 25, 50,75, and 100 μg/disk, respectively, and dried for 1 h under sterilized conditions. Then, they were placed on inoculated agar plats and incubated for 24 h at 37 °C for bacterial and 48 h (28 °C) for the fungal isolates. Inhibition zones were measured in mm and recorded. The antimicrobial activity of the fungal extract and pure compounds were examined against the test microbes: S. aureus ATCC 259239, E. coli ATCC 8739, P. aeruginosa ATCC 9027, B. subtilis ATCC 6633, Enterococcus faecalis ATCC 29212, K. pneumonia ATCC 13883, V. damsel, V. fluvialis, C. albicans A13, and A. flavus MM6.

All the crystal structures were fetched from the PDB site [[22]] and the grid box parameters were adjusted according to the information in Table 2. The protein preparation (removing water, deletion of unwanted atoms... etc.) was done by VEGA ZZ 3.1.1.4 [[31]]. The compounds (1–2) and the native ligands were drawn in 2D and nonminimized 3D forms using ChemBioDraw Ultra 14.0 and ChemBio3D Ultra 14.0, respectively. The best conformers were approached by the MOPAC algorithm of energy minimization. The docking simulation was performed by Autodock Vina [[23]] after writing the file formats in PDBQT using AutoDock Tools 1.5.7 [[32]]. Visualization and exporting images were accomplished by Pymol software [[33]].

Table 2: The crystal structure codes of the macromolecules and the grid box parameters of the docking.

The instruction manual of VEGFR2 (KDR) Kinase Assay Kit Cat. # 40325 was followed. A series of 10-fold dilutions were prepared from monascin, monascinol as well as sorafenib as a standard (Cat. No: 284461-73-0; Santa Cruz). The IC50; the concentration that inactivates 50% of the VEGFR2 enzyme, was determined using a curve fitting software; Prism, version 6.