Identification of potential inhibitors against Alzheimer-related proteins in Cordyceps militaris ethanol extract: experimental evidence and computational analyses

Abstract

Laboratory experiments were carried out to identify the chemical composition of Cordyceps militaris and reveal the first evidence of their Alzheimer-related potential. Liquid chromatography–mass spectrometry analysis identified 21 bioactive compounds in the ethanol extract (1–21). High-performance liquid chromatography quantified the content of cordycepin (0.32%). Bioassays revealed the overall anti-Alzheimer potential of the extract against acetylcholinesterase (IC₅₀ = 115.9 ± 11.16 µg mL⁻¹). Multi-platform computations were utilized to predict the biological inhibitory effects of its phytochemical components against Alzheimer-related protein structures: acetylcholinesterase (PDB-4EY7) and β-amyloid protein (PDB-2LMN). In particular, 7 is considered as a most effective inhibitor predicted by its chemical stability in dipole-based environments (ground state − 467.26302 a.u.; dipole moment 11.598 Debye), inhibitory effectiveness ($\overline{\mathrm{DS}}$ − 13.6 kcal mol⁻¹), polarized compatibility (polarizability 25.8 ų; logP − 1.01), and brain penetrability (logBB − 0.244; logPS − 3.047). Besides, 3 is promising as a brain-penetrating agent (logBB − 0.257; logPS − 2.400). The results preliminarily suggest further experimental attempts to verify the pro-cognitive effects of l(−)-carnitine (7).

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-023-03714-9.

Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that causes memory and behavioral problems in elderly people (Kumar and Singh 2015). Over the past decades, AD has become the most common cause of dementia, accounting for over 70% cases (Fiest et al. 2016).

To date, the treatments for AD are mainly based on the inactivation of related protein structures. Many findings have indicated that AD is frequently associated with decreasing levels of the neurotransmitter acetylcholine (ACh) in the brain (Schelterns and Feldman 2003), which is hydrolyzed under the catalysis of acetylcholinesterase (AChE) (Pohanka 2012). This means AChE inhibitors serve as major therapeutic agents for treatment of AD given their activity against AChE, thus maintaining ACh in the brain to sufficient levels (Kabir et al. 2019; Sharma 2019; Akıncıoğlu and Gülçin 2020). In the market, several drugs targeting AChE have been approved, such as donepezil, galantamine, and rivastigmine (Whitehouse 1993; Kelly et al. 1997; Gottwald and Rozanski 1999; Scott and Goa 2000; Colovic et al. 2013; Marucci et al. 2021). However, these drugs are also known for their low bioavailability and gastrointestinal side effects (Colovic et al. 2013; Marucci et al. 2021). From another approach, the extracellular accumulation and deposition of amyloid-β (Aβ) peptides in the brain is considered as a main contributor to the progression of AD (Hardy and Selkoe 2002; Jack et al. 2010; Sun et al. 2015). As a result, simultaneously deterring abnormal Aβ aggregation and inducing Aβ clearance is considered as a promising therapeutic strategy (Nie et al. 2011; Yiannopoulou and Papageorgiou 2020; Jeremic et al. 2021). In mechanism, Aβ is created from amyloid precursor protein (APP) by sequential proteolytic cleavages of β-secretase and γ-secretase enzymes (Chen et al. 2017). This means the reduction of Aβ levels can be achievable by effective inhibition of β-secretase and γ-secretase activities. In the market, a variety of Aβ peptide aggregation inhibitors have been proposed, including organic molecules (Jiang et al. 2013; Xie et al. 2014; Doig and Derreumaux 2015), peptides (Sievers et al. 2011; Viet et al. 2011; Xi and Hansmann 2019), antibodies (Manoutcharian et al. 2004), and nanoparticles (Bellucci et al. 2016); nevertheless, the downside refers to their low binding affinity and brain penetration (Galimberti and Scarpini 2016). The evidences suggest that an efficacious inhibition on the activities of these proteins, to some degree, might cut off the contributive factors of Alzheimer-related progression, thus possibly providing prevention of impaired learning and memory. Therefore, looking for more effective and body-friendly anti-Alzheimer agents (either AChE-based, amyloid-based, or both) is still necessary; and, natural products are promising with the advantages of low side effects and cost.

Cordyceps has been known for thousands of years for its important medicinal benefits given by diverse bioactive molecules with various pharmacological properties (Yue et al. 2013; Olatunji et al. 2018; Das et al. 2021). The fungus family is mostly found in Asian countries and other humid temperate and tropical habitats in North and South America, and Europe (Mains 1958). In 1753, Carl Linnaeus first described a species in the family of the genus Cordyceps with the name Clavaria militaris, which is popularly known as Cordyceps militaris. Despite having been long used by the folk experiences (especially by Asian cultures) for its wide variety of medicinal benefits, the fungus (and its genus in general) only attracts attention from the scientific community in recent decades.

In general, the extracts of C. militaris (by various solvent conditions) have been known to have a high diversity of bioactive molecules, giving the characteristic bioavailability. In fact, numerous biological families of phytochemicals have been determined, such as cordycepin, adenosine, ergosterol, trehalose, mannitol, polysaccharides, nucleosides, vitamins, and amino acids (Olatunji et al. 2018; Zhang et al. 2019; Sagaama et al. 2020; Zhong et al. 2020). They, in turn, build up the diversity of important pharmacological properties, such as hypoglycemic, hypolipidemic, anti-inflammatory, antitumor, antibacterial, antifungal, antiviral, antimalarial, antiprotozoal, neuroprotective, antioxidant, and immuno-protective activities (Olatunji et al. 2018; Zhang et al. 2019; Das et al. 2021; Jędrejko et al. 2021; Phull et al. 2022). On the downside, the bioavailability and biodiversity inevitably present challenges to the determination of biopotential from the standpoint of component-activity specification if relied heavily on laboratory-based wet blind tests. Fortunately, an appropriate utilization of time-efficient and cost-reduced computational platforms coupled with simple qualitative experimental characterizations can provide reliable predictions, thus further guiding more in-depth in-practice implementations on the most promising candidates. Especially, molecular docking simulation has been proven to be a reliable technique to identify the most effective protein-specified ligands regarding a wide variety of biological aspects, such as extracellular regulation (Singh et al. 2022b), intracellular conversion (Singh et al. 2022a), or even viral infection (Purohit et al. 2008).

Currently, C. militaris is extensively used as a crude drug and a supplemental product in Asia. Many recipes and procedures for the growth of fruit bodies and mycelia of C. militaris have been developed to overcome the scarcity and high price of wild C. militaris (Kontogiannatos et al. 2021), such as solid-state fermentation, submerged static fermentation, repeated batch or one-step non-static fermentation in liquid media, and larval/pupal-implanted cultivation. This helps the products more affordable to middle-class consumers. From folk experiences, extensive improvement in cognitive function and consciousness is one of the most transparent benefits reported. From scientific observations, behavioral neuroscience has provided some first evidences for the improvement of learning and memory by mouse-model psychological tests (Cai et al. 2013; Yuan et al. 2018). However, a solid mechanism is neither fully understood nor widely accepted. Although there have been several contemplations raised, the one on the inhibitory effects against Alzheimer-related enzymes has not been regarded.

In this work, C. militaris was subject to a theory–experiment study. Laboratory-based experiments were carried out for its phytochemical characterization (by liquid chromatography) and anti-Alzheimer evidence (by enzymatic bioassays); the results were used as the input and justification for theoretical attempts. Computer-based analyses were implemented for bio-physicochemical properties (by quantum calculation and QSAR-based models) and pharmacological potential (by docking technique and ADMET models). Altogether, theoretical arguments specify the promising candidates suggested for further consideration and can contribute evidences for a possible strategy to tackle Alzheimer’s disease from the point of view of enzymatic inhibition.

Methods

Experiments

Biological materials

The strain Cordyceps militaris Cordy-NH-1 was supplied by the Department of Chemistry, University of Sciences, Hue University. The fungus was identified by nuclear ribosomal internal transcribed spacer (ITS) gene sequence analysis. The specific ITS region (ITS1 + 5,8S + ITS2) was amplified by the universal fungal primers ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′). ITS gene sequence of the fungus was deposited in GenBank NCBI under ID number OP575950 (https://www.ncbi.nlm.nih.gov/nuccore/OP575950.1). Fresh fruiting body of C. militaris Cordy-NH-1 used in the present study was cultured by the Institute of Applied Research in Science and Technology, University of Sciences, Hue University. The C. militaris sample used in this study is shown in Fig. 1.

Fig. 1.

Cordyceps militaris morphology: a fresh fruiting body; b dried fruiting body; c dried powder

Extraction of cordycepin and adenosine from C. militaris

Freeze-dried fruiting body of C. militaris was ground into powder (approximately 50 meshes) using a blender. The powder (1 g) was then immersed in water:methanol (10 mL; 85:15 v/v), submerged in an ultrasonic bath (75 W), and then centrifugated (4500 rpm); this procedure was carried out in triplicate. The supernatant obtained was then accumulated and subjected to exact volume quantification. Finally, the sample was filtered through a microfilter (0.45 µm) before further analysis.

Determination of cordycepin and adenosine

High-performance liquid chromatography (HPLC) analysis was performed to quantify the contents of cordycepin and adenosine; equipment: Agilent 1260 Infinity II (Agilent Technologies, Santa Clara, CA, USA) equipped with an Agilent 1260 Photo-Diode Array Detector (DAD) WR; auto injector; Agilent Shield RP C18/4.6 150 mm; 4-µm reverse phase column. The standard chemicals (adenosine and cordycepin) were repeatedly injected six times for draw calibration curves; the injection volume was 1, 10, 25, 50, 100, and 200 mg L⁻¹, consecutively. The mobile phase was water–methanol mixture (85:15 v/v). The separation was conducted in isocratic elution with a flow rate of 0.7 mL min⁻¹. The detection wavelength of photo-diode array was set at 260 nm. The column temperature was 30 °C. The injection volume in the HPLC system was 20 µL. Data collection and analysis were performed using Open Lab CDS 2.X (Agilent Technology); the results were calculated as the average of triplicate iterations.

Standard chemicals: cordycepin and adenosine were purchased from Sigma (St. louis, MO, USA). Each standard was prepared as a stock solution (20 mg mL⁻¹) in methanol. All solutions were degassed after being filtered via a 0.45 mm membrane. A working solution (0.2 mg mL⁻¹) was prepared by diluting the stock solution with methanol. Formic acid used was analytical grade. The mobile phase was made using Milli-Q quality water and methanol of HPLC grade.

Preparation of ethanol extract of C. militaris

The dried powder of C. militaris (20 g) was immersed in absolute ethanol (100 mL; 3 days; room temperature). The extract was then collected, filtered, and evaporated (reduced pressure; 50 °C). The procedure was performed in triplicate; the extracts were gathered and concentrated to the yield of 1.12 ± 0.18 g.

Identification of ethanol extract composition

Liquid chromatography–mass spectrometry (LC–MS) analysis was performed to identify the major chemical components present in the extract of C. militaris. LC–MS was performed on Agilent 1100 ion trap MSD mass spectrometer with ESI source in negative mode equipped with a degasser (G1379A), binary pump (G1312A), auto sampler (G1315B) (Waldbronn, Germany). The data was acquired and processed using LC/MSD trap ChemStation software 4.2. Column conditions: ambient temperature (23 ± 1 °C); nitrogen as nebulizer and curtain gas; collision induced dissociation using helium gas. Ion source conditions: mass range 50–600 m/z; temperature 325 °C; nebulizer gas 40 psi; dry gas 8.0 L min⁻¹; ion spray voltage 5000 V; collision energy 33 V; dwell time 200 ms. Chromatographic configuration: column (Waters XSELECT CSH Phenyl-Hexyl, 5 µm, 4.6 × 150 mm); mobile phase (acetonitrile and water 50:50 with 0.1% formic acid, flow rate 4.0 mL min⁻¹, detection DAD 210 nm).

In vitro AChE inhibitory assay of ethanol extract

Inhibition of acetylcholinesterase activity was determined using the modified Ellman’s colorimetric method (Eldeen et al. 2005). A 96-well plate contained the following: (1) ATCI (25 µL, 15 mmol L⁻¹) in water; (2) DTNB (125 µL, 3 mmol L⁻¹) in Buffer A (Tris–HCl 50 mmol L⁻¹, NaCl 0.1 mol L⁻¹, MgCl₂·6H₂O 0.02 mol L⁻¹, pH 8); (3) Buffer B (bovine serum albumin 0.1% 50 mmol L⁻¹, pH 8); (4) plant extract (25 µL; concentrations: 25 µg mL⁻¹, 25 µg mL⁻¹, 50 µg mL⁻¹, 100 µg mL⁻¹, 200 µg mL⁻¹, 400 µg mL⁻¹). Thereafter, AChE (0.2 U mL⁻¹) was added to the wells and the absorbance was measured at 415 nm; galantamine was a positive control. Any increase in absorbance due to the spontaneous hydrolysis of the substrate was corrected by subtracting the absorbance before adding the enzyme from the absorbance after adding the enzyme. The percentage inhibition was calculated using the equation: $\text{Inhibition}(\%)=1-(A_{\text{sample}}/A_{\text{control}})\times 100$, where A_sample is the absorbance of the sample extracts and A_control is the absorbance of the blank (ethanol in 50 mmol L⁻¹ Tris–HCl, pH 8). Extract concentration providing 50% inhibition (IC₅₀) was obtained by plotting the percentage inhibition against extract concentration.

Computations

The experimental findings were used as the input for multi-platform computational implementations, whose output serves as theoretical arguments to allocate the most promising inhibitors. In particular, quantum calculation provided intrinsic properties of the candidates, such as ground-state energy and dipole moment providing representations for bio-chemical stability and compatibility; molecular docking simulated ligand–protein complex configuration with the associated docking score value representing pseudo Gibbs formation energy; QSARIS-based physicochemical properties coupled with Lipinski's references were used for drug-likeness arguments; ADMET-based pharmacokinetics coupled with Pires' interpretations were used for consideration of further medicinal development.

Quantum chemical calculation

Molecular quantum properties were obtained from density functional theory (DFT) calculation on Gaussian 09 (Gaussian, Inc.; Wallingford CT, Connecticut, USA): no symmetry constraints; level of theory M052X/6–311++ G(d,p) (Kassel 1988; Schäfer et al. 1992). The global minimum on the potential energy surface (PES) was confirmed by vibrational frequencies. The frozen core approximation for non-valence shell electrons and the resolution-of-identity (RI) approximation were applied. The criterion of convergence for the SCF energy was set to 10^–8 a.u. and the modified integration grid “m4” was used. The configurations were used for the calculation of optimized geometries, potential energy surface (PES), and dipole moments at the level of theory M052X/def2-TZVPP. The molecular electrostatic potential (MEP) in the molecule was calculated and visualized by GaussView 05 software.

Molecular docking simulation

Ligand–protein static inhibitability can be evaluated using MOE 2015.10 (Chemical Computing Group ULC.; Montreal, Quebec, Canada) based on the molecular docking technique. In a typical procedure, the simulation follows four steps and results in ligand–protein complex structures, namely: (1) input preparation; (2) docking simulation; (3) re-docking iteration; (4) theoretical interpretation. First, crystal structures of human acetylcholinesterase (Cheung et al. 2012), and β-amyloid protein 40 (Paravastu et al. 2008) were downloaded from RCSB Protein Data Bank, under the IDs PDB-4EY7 (https://doi.org/10.2210/pdb4EY7/pdb) and PDB-2LMN (https://doi.org/10.2210/pdb2LMN/pdb), respectively; all co-crystal molecules (if any) were removed. Ligand structures were from those determined by LC–MS in this work; geometrical optimization: Conj Grad algorithm; energy change termination: 0.0001 kcal mol⁻¹; charge assignment: Gasteiger–Huckel method. Second, ligand–protein interaction was simulated; active gird range: 4.5 Å from amino acids; force field: MMFF94x; tether-receptor strength: 5000; energy resolution: 0.0001 kcal mol⁻¹ Å⁻¹; number of retaining poses = 10; maximum solutions per iteration = 1000; maximum solutions per fragmentation = 200. Third, the inhibitory components (ligand and protein) were separated, then re-docked; the accuracy of the docking protocol is justified if RMSD values (of docked and re-docked conformations) are all under 2 Å. Finally, the primary parameters for inhibitory effectiveness are used for discussion, including docking score (DS) energy, root-mean-square deviation (RMSD) value, and numbers of hydrophilic binding (hydrogen-like bonds). Besides, ligand–protein interactions and in-pose arrangement were mapped and rendered on 2D and 3D visualization, respectively. In this work, five sites corresponding to the highest docking score values of each ligand–complex duo was selected for discussion on the inhibitory effectiveness.

Figure 2 shows the crystal assemblies of the protein used in this work and the control drug; the data were referenced from RCSB Protein Data Bank. Figure 3 presents the input for ligands in this work; the structural formulae were from the experiments (extraction and LC–MS).

Fig. 2.

Biological assemblies of a 4EY7 (human acetylcholinesterase), b 2LMN (β-amyloid protein 40), and c structural formula of galantamine (D)

Fig. 3.

The structural formulae of investigated compounds (1–21) detected from C. militaris ethanol extract

QSARIS analysis

Drug-likeness properties of phytochemicals were predicted by a combinational model, including (1) parameters: QSARIS-derived physical properties [based on Gasteiger–Marsili method (Gasteiger and Marsili 1980)]; (2) reference: Lipinski's rule of five (Lipinski et al. 1997). The former includes molecular mass (Da), polarizability (ų) and size (Å), and dispersion coefficients (logP and logS); on the other side, the rule set criteria for a well membrane-permeable candidate, i.e., (1) molecular mass < 500 Da; (2) hydrogen bond donors ≤ 5; (3) hydrogen bond acceptors ≤ 10; (4) logP <  + 5 (Mazumdera et al. 2009; Ahsan et al. 2011)^.

ADMET analysis

ADMET properties (absorption, distribution, metabolism, excretion, and toxicity) were obtained from a web-based regressive model developed and maintained by the Molecular Modeling Group, Swiss Institute of Bioinformatics: SwissADME (http://www.swissadme.ch/; May 6th, 2023). The theoretical interpretations of output pharmacokinetic parameters were described by Pires et al. (2015).

Results and discussion

Experimental

C. militaris cordycepin and adenosine contents

Nucleosides cordycepin and adenosine are considered as main active constituents in the genus Cordyceps, whose contents in this study are given in Table 1. Although the latter was not detected by HPLC analysis, the former registered an especially high value (0.32% in dried fruiting body) compared to those often seen in this natural source. Regarding its value, cordycepin has been proven to exhibit a variety of bio-medical activities, for example: neuroprotection (Schmidt et al. 2003); lung and kidney protection (Nakamura et al. 2005); anticancer and antileukemic activities (Yoshikawa et al. 2004; Weil and Chen 2011); antioxidant activity (Li et al. 2006); antibacterial and antifungal activities (Lee et al. 2012); anti-inflammatory and immunomodulatory effects (Yu et al. 2006; Lee and Hong 2011); pro-sexual activity (Lim et al. 2012). Meanwhile, adenosine is known to have the ability to modulate cell proliferation, differentiation, and apoptosis (Yang et al. 2010; Jozsef Szentmiklosi et al. 2015). However, it is also often seen that the contents of these compounds widely vary depending on different factors, such as culture media, culture conditions, and fungal strains.

Table 1.

Contents of adenosine and cordycepin in C. militaris

Bioactive ingredients	Content (%)
Cordycepin	0.32
Adenosine	–

–, not detected

Inhibition against AChE

The inhibitory effects of C. militaris ethanol extract against AChE were demonstrated by in vitro bioassays, whose results are given in Table 2. The corresponding IC₅₀ value is 115.9 ± 11.16 µg mL⁻¹; those by the positive control (galantamine) is 1.458 ± 0.089 µg mL⁻¹. Although the results revealed the potential inhibitory effects of the total extract against an Alzheimer-related protein (AChE), the effectiveness is considered relatively incompetent, compared to those given by a commercial drug (galantamine). This indicates the need for further screening research to allocate the potent components in the source (C. militaris ethanol extract), whose bio-pharmacological potentiality deserves further experimental efforts.

Table 2.

Inhibition of C. militaris ethanol extract against AChE

Extract/compound	IC50 (µg mL−1)
Ethanol extract of C. militaris	115.9 ± 11.16
Galantamine	1.458 ± 0.089

From the literature, the anti-Alzheimer potential of a variety of C. militaris extracts has been preliminarily evidenced by in vitro and in vivo studies. For instance, a polypeptide isolated from C. militaris was reported to improve memory in mice (Yuan et al. 2018), whose mechanism is apparently related to the decrease of acetylcholinesterase (AChE) activity and the lowering of GABA (gamma-aminobutyric acid) contents in the brain of model group mice. In another work, the source was thought to exhibit antioxidant and neuroprotective potential by increasing SOD activity and reducing MDA concentration (Yuan et al. 2018). In recognition tests on laboratory mice, the extract of C. militaris improved pro-cognitive functions of the objects in the water maze modeling (He et al. 2019). On rodents treated with doxorubicin chemotherapy, the extract of C. militaris was able to reduce chemotherapy-induced oxidative stress; in particular, the postmortem suggested that C. militaris extract might increase the concentration of ATP in the brain and inhibit the activity of AChE (Veena et al. 2020). However, other than the overall effects of the total extracts, the findings have been unlikely to specify any solid suggestion on the associated component–activity relationship.

Chemical constituents

Chemical profile of C. militaris ethanol extract was analyzed by LC–MS, whose identified components are summarized in Table 3. In general, the results reveal a diverse chemical composition including 21 compounds detected. More particularly, the phytochemicals belong to a variety of classes, including nucleosides, fatty acids, amino acids, phenolics, carboxylic acids, polyols, and saccharides, in which carboxylic acids are the substances of most abundance. Also, cordycepin was found, while adenosine was not detected. In fact, the concentration and distribution of bioactive compounds are not uniform in the fruiting bodies. The outer parts of C. militaris fruiting bodies have the highest concentration of nucleosides, polysaccharides, carotenoids, and selenium organic compounds. In addition, the optimal drying temperature for C. militaris is 60 °C. A higher temperature causes a loss of the content of cordycepin and phenolic compounds (Wu et al. 2019). The biological availability and phytochemical diversity give the fungus family extensive applicability, yet put challenges onto the scientific efforts to specify the most promising candidates.

Table 3.

Bioactive compounds 1–21 detected from C. militaris ethanol extract

Notation	Compounds	Formula	Type
1	Fumosoroseain A	C28H52NO6	Fatty acids
2	l-Azatyrosine	C8H10N2O3	Amino acids
3	(Z)-3-(2-hydroxyphenyl)acrylic acid	C9H8O3	Carboxylic acids
4	Ethyl 4-(1-hydroxy-1-methylethyl)-2-propyl-imidazole-5-carboxylate	C12H20N2O3	Heterocycles
5	Cordycepin	C10H13N5O3	Nucleosides
6	4-((4-Methylacridin-9-yl)amino)benzoic acid hydrochloride	C21H17ClN2O2	Benzenoids
7	l(−)-Carnitine	C7H15NO3	Carboxylic acids
8	l-Ornithine, N5-(4,5-dihydro-4-methyl-5-oxo-1H-imidazol-2-yl)	C9H16N4O3	Amino acids
9	l-Histidine, 2-amino	C6H10N4O2	Amino acids
10	2,5,8,11-Tetramethyl-3,6,9,12-tetraoxapentadecane-1,14-diol	C15H32O6	Polyols
11	Adenine,9-[3-deoxy-3-(2,6-diaminohexanamido)-b-d-arabinofuranosyl]-, l-(8CI)	C16H26N8O4	Nucleosides
12	d-Gluco-Deconic acid,6,10-diamino-4,6,7,8,9,10-hexadeoxy	C10H22N2O5	Saccharides
13	2-Pyrrolidinone,1-benzoyl-5-(1-pyrrolidinylcarbonyl)-, (5S)	C16H18N2O3	Pyrrolidines
14	Butanedioic acid,1-[2-[3-(1,3,5-hexatrien-1-yl)phenyl]hydrazide]	C16H18N2O3	Carboxylic acids
15	Carbamic acid, (3-methyl-4H-1,2,4-triazol-4-yl)-, ethyl ester (9CI)	C6H10N4O2	Organic carbonic acids
16	d-Glucaric acid,6,3-lactone	C6H8O7	Saccharides
17	d-Gluconamide,N-(2-mercaptoethyl)	C8H17NO6S	Saccharides
18	N-(tert-Butoxycarbonyl)-l-threonine	C9H17NO5	Carboxylic acids
19	Benzoic acid,3,4,5-trihydroxy-, dodecyl ester	C19H30O5	Phenolic acids
20	allo-Heptopyranoside,methyl 3,7-dideoxy-4-O-methyl-2-O-[(2-methylphenyl)methyl]-(9CI)	C17H26O5	Saccharides
21	Ser-His-Asp	C13H19N5O7	Carboxylic acids

Besides cordycepin (5), many compounds detected in the ethanol extract of C. militaris by LC–MS analysis in this work have been also reported as having promising biological activities by preceding studies. For typical instances, fumosoroseain A (2) possessed anti-aging activity (Wei et al. 2022); l(−)-carnitine (7) exhibited antioxidant effects (Cao et al. 2011; Li et al. 2012) and prevented cardiovascular disease (DiNicolantonio et al. 2013; Song et al. 2017). However, to our knowledge, there has been no attempt to investigate the anti-Alzheimer ability of the C. militaris ethanol components. This means that the potential is still vastly untouched by common research. Therefore, this stage of qualitative analysis is sufficient to obtain the structural input for computer-based screening, whose output in turn might serve as the justification for more in-depth quantitative attempts or trial-and-error tests.

Computational

Quantum-based chemical properties

Quantum calculation resulted in the optimized structural geometry and electronic onfigurations of each bioactive compound (1–21). The information can be utilized for argument of their chemical stability in dipole-based environments and intermolecular behaviour in general.

Figure 4 shows the optimized geometries of the studied structures. In general, the convergence regarding each formula was reached without noticeable constraints or unusual parameters; simply put, all the geometrical angles and bonding lengths meet those of characteristic ranges known to common knowledge. This confirms the structural stability existing in natural sources.

Fig. 4.

Geometrically optimized structures of 1–21 by DFT at the level of theory M052X/6–311++ g(d,p); unit of bond length: Å, unit of bond angle: °

Table 4 summarizes the main electronic-based properties. In particular, the ground state energy represents electronic stability, and the dipole moment is defined by the separation magnitude of positive–negative charge in a molecule. In essence, the former indicates higher chemical inactiveness with lower values, thus considered more suitable for applications based on non-chemical induction (by van der Waals forces and ionic bonds); on the other hand, the latter suggests the compatibility of the host molecule when applied in dipole-solvent environments (such as physio-chemical media). Given these theoretical arguments, the most chemically stable candidates are: 17–21 (under − 600 a.u.), while those with the most predominant values regarding dipole moment are: 7 (11.598 Debye) > 15 (7.865 Debye) > 17 (6.295 Debye) > 21 (6.423 Debye) > 19 (5.108 Debye); nevertheless, it is noteworthy that the ranks are simply for relative comparison, not for executive disregard of the others. To our experience in this field, all the structures can be considered with appropriate chemical inertia; the dipole moment of 7 is exceptionally significant, while that of 20 (0.687 Debye) is exclusively unfavorable for dipole–dipole applicability. Considering cordycepin (5), the figures are considered moderate values, − 888.49 a.u. for ground state energy and 3.779 Debye for dipole moment.

Table 4.

Ground state electronic energy and dipole moment values of 1–21 by DFT at level of theory M052X/6–311++ g(d,p)

Compound	Ground state electronic energy (a.u.)	Dipole moment (Debye)
1	− 1604.54004	2.913
2	− 646.17173	1.540
3	− 573.53618	3.476
4	− 804.64971	4.729
5	− 888.48736	3.779
6	− 1070.03951	2.039
7	− 556.25339	11.598
8	− 797.44653	2.376
9	− 604.24322	2.746
10	− 1042.32245	2.272
11	− 1364.53247	3.992
12	− 880.09643	3.594
13	− 955.90723	3.859
14	− 955.82482	2.808
15	− 604.17609	7.865
16	− 760.18473	3.630
17	− 1219.53617	6.295
18	− 784.25886	2.789
19	− 1118.41776	5.108
20	− 1039.71697	0.687
21	− 1307.46795	6.423

Figure 5 exhibits the molecular electronic potential (MEP) maps of the optimized structure, which are equal to molecular electronic distributions. By convention, electron density is higher in reddish space, lower in bluish space, and neutral in greenish space, which are equivalent to nucleophilic, electrophilic, and van der Waals tendencies. In theory, those with neutral potential possess low flexibility for external interactions with complex structures (like proteins), and thus are unfavorable for inhibitory applications; for example, 1, 4, 6, 19 are not highly recommended. Comparing the magnitude of potential difference, the rank is: 7 (± 9.570 × 10^–2 a.u.) > 4 (± 8.715 × 10^–2 a.u.) > 17 (± 8.127 × 10^–2 a.u.); however, it is noteworthy that regarding the argument on the flexibility, these values are not useful themselves, since higher figures might be due to too localized distributions. The most reasonable approach is a combination of heuristic and quantitative evaluations.

Fig. 5.

Molecular electrostatic potential (MEP) formed by mapping of total density over the electrostatic potential of 1–21; reddish region: negative electrostatic potential, bluish region: positive electrostatic potential, greenish region: null electrostatic potential

In summary, C. militaris ethanol extract is a promising source for bio-inhibitors from the standpoint of intrinsic quantum properties. For brief selection, 7, 15, 17, 19, 21 can be picked out as the most potential candidates.

Docking-based inhibitory potential

The docking technique resulted in the affinity of each bioactive compound (1–21) toward each protein structure (4EY7 or 2LMN). This provides the view for argument on the inhibitory potential in the context of complex static interactions. From this standpoint, the total docking score (DS) values (the pseudo Gibbs free energy) and the number of hydrogen-like bonds (strong intermolecular interaction) are selected as the main indicators for inhibitory effectiveness.

Figure 6 highlights the susceptible sites of the targeted proteins to the ligands and Table 5 summarizes the corresponding primary docking parameters; the control drug (D) is galantamine. In general, the protein structures are likely more vulnerable at sites 1 and 2 than the others. In practice, the biological inhibition by natural products often occurs as a multi-site process rather than being selective to any specific site; thus, the average DS value might be a reasonable indicator for inhibitory effectiveness. In essence, sufficient distortion forces pressuring over the secondary–tertiary structures would lead to the denaturation of enzyme shape and stopping of enzymatic activity overall. Given this argument, the most effective inhibitors against 4EY7 (acetylcholinesterase) are predicted by the order: 9-4EY7 (DS_average − 11.7 kcal mol⁻¹) > 7-4EY7 (DS_average − 10.8 kcal mol⁻¹) > 6-4EY7 ≈ 11-4EY7 (DS_average − 10.7 kcal mol⁻¹) > 21-4EY7 (DS_average − 10.6 kcal mol⁻¹) > 17-4EY7 (DS_average − 10.5 kcal mol⁻¹) > 3-4EY7 ≈ 10-4EY7 (DS_average − 10.1 kcal mol⁻¹); the corresponding order for 2LMN (β-amyloid protein) is: 7-2LMN (DS_average − 12.0 kcal mol⁻¹) > 17-2LMN (DS_average − 11.4 kcal mol⁻¹) > 5-2LMN (DS_average − 10.9 kcal mol⁻¹) > 11-, 1-, 4-, 8-, 21-2LMN (DS_average ca. − 10 kcal mol⁻¹). Overall, these figures are markedly predominant compared to those given by the control drug, D-4EY9 (DS_average − 8.6 kcal mol⁻¹) and D-2LMN (DS_average − 8.4 kcal mol⁻¹), and thus predicted to have elevated efficacy in in-practice applications. When considering the most stable ligand–protein inhibitory complexes, equivalent to the in-practice inhibitory products, there are slight differences in the ranks. Regarding 4EY7 (acetylcholinesterase), the highest figures correspond to the ligands: 9 > 21 > 11 > 7 > 6 > 10 ≈ 17 > 5 (DS > − 12 kcal mol⁻¹; hydrophilic bonds > 3); D (DS − 11.2 kcal mol⁻¹; hydrophilic bonds 2). Regarding 2LMN (β-amyloid protein), the most effective inhibitors are: 7 > 17 > 1 > 8 ≈ 12 ≈ 18 > 11 > 5 ≈ 19 (DS > − 12 kcal mol⁻¹; hydrophilic bonds > 3); D (DS − 10.2 kcal mol⁻¹; hydrophilic bonds 1). In summary, the repeating rise of 5, 7, 9, 11, 17 implies their potential as inhibitors against Alzheimer-related proteins and the DS values of significance, compared to those given by D, predicting their inhibitory effectiveness; the justification is based on the standpoint of ligand–protein static interaction.

Fig. 6.

Quaternary structures of 4EY7 (human acetylcholinesterase) and 2LMN (β-amyloid protein 40) with the approachable sites by 1–21; yellow: site 1, gray: site 2, blue: site 3, orange: site 4, cyan; site 5

Table 5.

Screening results on inhibitability of ligands (1–21) and controlled drug (D) toward different sites of 4EY7 and 2LMN

P	L	Site 1		Site 2		Site 3		Site 4		Site 5		Average
		E	N	E	N	E	N	E	N	E	N	E
4EY7	1	− 9.4	1	− 11.0a	2a	− 9.8	1	− 8.0	0	− 7.6	0	− 9.2
	2	− 11.5a	3a	− 10.8	2	− 9.5	1	− 9.1	1	− 7.9	0	− 9.8
	3	− 11.8a	3a	− 10.3	2	− 10.0	2	− 9.4	1	− 9.0	1	− 10.1
	4	− 11.6a	3a	− 10.1	2	− 9.6	1	− 9.3	1	− 8.0	0	− 9.7
	5	− 12.0a	3a	− 9.2	1	− 9.0	1	− 9.4	1	− 7.8	0	− 9.5
	6	− 11.4	2	− 13.0a	4a	− 10.9	2	− 9.2	1	− 8.9	1	− 10.7
	7	− 11.0	2	− 13.2a	4aa	− 10.8	2	− 9.9	2	− 9.1	1	− 10.8
	8	− 10.1a	2a	− 8.9	1	− 9.0	1	− 7.4	0	− 7.1	0	− 8.5
	9	− 11.9	3	− 13.9a	6a	− 11.4	3	− 10.8	2	− 10.4	2	− 11.7
	10	− 10.9	2	− 12.4a	4a	− 9.9	2	− 8.7	1	− 8.5	1	− 10.1
	11	− 10.8	2	− 13.5a	5a	− 9.7	2	− 9.9	2	− 9.4	2	− 10.7
	12	− 11.0a	3a	− 8.5	1	− 8.7	1	− 7.3	0	− 7.0	0	− 8.5
	13	− 9.1	1	− 10.9a	2a	− 9.0	1	− 7.1	0	− 6.9	0	− 8.6
	14	− 8.8	1	− 11.9a	3a	− 8.3	1	− 8.6	1	7.6	0	− 6.0
	15	− 8.5	1	− 9.3a	2a	− 7.8	0	− 7.5	0	− 7.0	0	− 8.0
	16	− 10.9a	3a	− 9.6	2	− 8.0	1	− 7.1	0	− 6.5	0	− 8.4
	17	− 11.0	3	− 12.4a	5a	− 10.2	2	− 9.9	2	− 9.2	1	− 10.5
	18	− 10.4	2	− 11.1a	3a	− 9.7	1	− 8.1	0	− 7.4	1	− 9.3
	19	− 10.9a	2a	− 9.0	1	− 8.1	0	− 7.2	0	− 6.8	0	− 8.4
	20	− 9.0	1	− 10.0a	2a	− 8.0	0	− 7.8	0	− 7.1	0	− 8.4
	21	− 10.0	2	− 13.7a	5a	− 11.3	3	− 9.2	1	− 8.8	1	− 10.6
	D	− 11.2a	2a	− 9.6	1	− 8.0	0	− 7.1	0	− 6.9	0	− 8.6
2LMN	1	− 12.9a	4a	− 9.8	2	− 9.2	2	− 9.6	2	− 8.3	1	− 10.0
	2	− 11.6a	3a	− 10.3	2	− 9.4	1	− 9.0	1	− 8.7	1	− 9.8
	3	− 10.0a	2a	− 9.2	1	− 8.0	0	− 7.7	0	7.0	0	− 5.6
	4	− 11.4a	3a	− 10.6	2	− 9.8	1	− 9.7	1	− 8.3	0	− 10.0
	5	− 12.0a	3a	− 10.9	1	− 11.3	2	− 10.9	2	− 9.2	1	− 10.9
	6	− 10.1	2	− 11.0a	3a	− 9.0	1	− 8.9	1	− 8.0	0	− 9.4
	7	− 14.0a	7a	− 12.1	4	− 11.9	3	− 11.7	3	− 10.1	2	− 12.0
	8	− 12.7a	4a	− 10.7	2	− 10.3	2	− 9.1	1	− 7.9	0	− 10.1
	9	− 11.9a	3a	− 10.0	2	− 9.6	1	− 7.5	0	− 7.7	0	− 9.3
	10	− 9.4	1	− 10.9a	2a	− 8.1	0	− 7.4	0	− 6.5	0	− 8.5
	11	− 10.1	2	− 12.4a	4a	− 10.3	2	− 9.2	1	− 8.9	1	− 10.2
	12	− 10.0	2	− 12.8a	4a	− 9.8	1	10.3	2	− 9.4	1	− 6.3
	13	− 9.9a	2a	− 9.0	1	− 8.4	0	− 8.1	0	− 7.7	0	− 8.6
	14	− 8.6	1	− 11.7a	3a	− 10.6	2	− 8.7	1	− 8.0	0	− 9.5
	15	− 10.4a	2a	− 8.5	1	− 8.0	0	− 7.7	0	− 7.2	0	− 8.4
	16	− 9.2	1	− 10.1a	2a	− 8.6	1	− 8.0	0	− 7.1	0	− 8.6
	17	− 13.8a	6a	− 11.6	4	− 11.0	3	− 9.7	2	− 10.9	3	− 11.4
	18	− 10.0	2	− 12.6a	4a	− 9.9	2	− 8.2	1	− 8.4	1	− 9.8
	19	− 12.0a	3a	− 11.2	2	− 9.0	1	− 8.6	1	− 8.0	0	− 9.8
	20	− 10.3a	2a	− 9.0	1	− 8.1	0	− 7.3	0	− 7.0	0	− 8.3
	21	− 10.0	2	− 12.3a	4a	− 9.9	2	− 9.6	2	− 8.7	1	− 10.1
	D	− 10.2a	1a	− 9.6	1	− 8.0	0	− 7.5	0	− 6.8	0	− 8.4

P protein structure (PDB-ID), C ligand, E DS value (kcal mol⁻¹), N number of hydrophilic interactions

^aMost effective ligand–protein complexes (in-detail data provided in Supplementary Material)

For further reference, the in-detail data for each of the most effective inhibitory systems regarding each ligand–protein inhibition is provided in Supporting Information. The corresponding visualizations of in-site arrangements and interaction maps are presented in Fig. 7 (ligand–4EY7) and Fig. 8 (ligand–2LMN). Regarding 3D in-pose morphology, the sites are rather tight and close, compared to the compound structures. This means that further functionalization with significant increase of molecular size is not recommended. From 2D maps, the ligands are likely to have good conformational fitness with the in-site features given by the continuousness of dashed contours.

Fig. 7.

Visual presentation and in-pose interaction map of ligand–4EY7 complexes; dashed arrow: hydrogen-like bonding, blurry purple: van der Waals interaction, dashed contour: conformational fitness

Fig. 8.

Visual presentation and in-pose interaction map of ligand–2LMN complexes; dashed arrow: hydrogen-like bonding, blurry purple: van der Waals interaction, dashed contour: conformational fitness

QSARIS-based physicochemical properties

Table 6 summarizes the physicochemical properties of the compounds (retrieved from the QSARIS) and the number of hydrogen bonds (counted from docking-based results). The first theoretical argument is based on Lipinski's rule of five for drug-like candidates. Except for 1 (mass 498.4 amu) and 19 (logP 4.84), C. militaris ethanol extract components register good satisfaction with the criteria, which are proposed to ensure appropriate behavior of the candidates in biocompatible environments. Extensively, among the promising inhibitors predicted by docking-based simulation, 19 (polarizability 58.8 ų) > 11 (polarizability 57.5 ų) > 5 ≈ 17 (polarizability 34.7 and 34.2 ų) are characterized by their sensitivity to external electric fields such as those created by other polarized agents (such as amino acid-based protein structures); 9 (logP − 4.13) > 21 (logP − 4.03) > 17 (logP − 3.12) are characterized by their aqueous dispersibility. The unit conversion of polarizability is given by the Claussius–Mossotti relation: ${10}^6/4\pi {ϵ}_0[{\text{A}}^2{\text{s}}^4{\text{kg}}^{-1}]\equiv 1[{\text{cm}}^3]$ (Feynman 2010); logP represents the octane–water partition. Therefore, these indicators might serve as the representative of pre-inhibition conditions. For further argument, 7 ($\overline{\mathrm{DS}}$ − 13.6 kcal mol⁻¹) > 17 ($\overline{\mathrm{DS}}$ − 13.1 kcal mol⁻¹) > 11 ≈ 21 ($\overline{\mathrm{DS}}$ − 13.0 kcal mol⁻¹) can be thought of as versatile bio-inhibitors against different protein structures (4EY7 and 2LMN in the scope of this work). In general, those with promising inhibitability are also highly compatible with the physicochemical environments.

Table 6.

Physicochemical properties of studied compounds 1–21 and D

Compound	$\overline{\mathrm{DS}}$ (kcal mol−1)	Mass (amu)	Polarizability (Å3)	Size (Å)	Dispersion coefficients		Hydrogen bond (4EY7/2LMN)
					logP	logS	H donor	H acceptor
1	− 12.0	498.4	89.9	598.3	4.12	− 7.15	0/0	1/2
2	− 11.6	182.1	25.4	214.9	− 3.10	1.06	2/1	1/1
3	− 10.9	164.3	23.6	249.0	1.92	− 1.50	1/1	0/1
4	− 11.5	240.4	39.7	358.6	1.89	− 2.30	0/1	0/1
5	− 12.0	251.2	34.7	226.2	− 1.30	− 1.37	0/0	1/1
6	− 12.0	364.5	52.8	503.5	3.99	− 4.69	1/0	1/0
7	− 13.6	161.0	25.8	185.1	− 1.01	0.71	1/0	2/3
8	− 11.4	228.4	33.3	287.6	− 3.27	− 0.65	1/2	1/2
9	− 12.9	170.2	23.2	185.5	− 4.13	− 0.19	0/1	2/1
10	− 11.7	308.1	52.5	495.7	0.14	− 1.39	1/1	2/1
11	− 13.0	394.2	57.5	489.3	− 2.88	− 1.33	0/2	2/1
12	− 11.9	250.3	38.5	367.5	− 1.87	0.83	0/1	3/3
13	− 10.4	286.2	44.8	345.1	1.41	− 2.66	0/1	0/0
14	− 11.8	286.1	44.4	350.3	2.81	− 4.41	0/1	2/2
15	− 9.9	170.3	23.2	210.2	− 0.26	− 0.56	0/0	0/1
16	− 10.5	192.2	21.5	162.9	− 2.39	0.46	2/0	½
17	− 13.1	255.0	34.2	289.7	− 3.12	0.26	1/5	3/1
18	− 11.9	219.4	32.3	305.8	0.43	− 0.94	1/1	1/3
19	− 11.5	338.0	58.8	504.9	4.84	− 5.23	0/0	2/2
20	− 10.2	310.1	51.3	421.0	2.22	− 2.76	1/1	0/1
21	− 13.0	357.2	46.7	480.3	− 4.03	0.12	1/1	2/3
D	− 10.7	287.3	47.4	370.4	2.49	− 2.29	0/1	0/0

$\overline{\mathrm{DS}}$: sverage DS value calculated from all ligand–protein structures (ligand: 1–21 and D; protein: 4EY7, 2LMN)

ADMET-based pharmacokinetics and pharmacology

Table 7 (1–11) and Table 8 (12–21 and D) show the ADMET properties (absorption, distribution, metabolism, excretion, and toxicity) of the bioactive compounds. Overall, the C. militaris ethanol extract is predicted to have certain negative effects on the body: (1) mutagenic potential (5, 9, 13, 20); (2) potential for fatal ventricular arrhythmia as hERG inhibitors (1, 6, 19); (3) potential for hepatotoxicity (1, 4, 9, 13, 15, 20, 21); (4) skin sensitization (6, 19); (5) certain toxicity to bacterium T. Pyriformis (pIGC50 > − 0.5 log μg L⁻¹) and fish flathead minnows (LC50 > − 0.3). Fortunately, the compounds with biological inhibitability and compatibility, 5, 7, 11, and 17, are safe for use in humans; hence, they are better further isolated for use in pure form instead of in the total extract mixture. Furthermore, none of them is predicted to either inhibit the activity of the cytochromes P450 family or be oxidized by the liver (as substrates). However, only 7 (logBB − 0.244; logPS − 3.047) is predicted to readily cross the blood–brain barrier (impermeable threshold: logBB < − 1) and reasonably penetrate the central nervous system (impermeable threshold: logPS < − 3); to a certain degree, the values are comparable to those of D (logBB − 0.081; logPS − 2.511). These parameters are considered especially important since the targeted proteins are brain-expressed enzymes. The compound is likely distributed in plasma, rather than being accumulated in tissue (logVDss − 0.664 <  < − 0.15), and easily absorbed via the intestinal route (over 90%). Besides, there is no potential interaction with P-glycoproteins, meaning that there are no effects on the cellular activity of the toxin/xenobiotic extrusion. Another candidate that can be considered is 3, which is also expected to have similar pharmacokinetics to those of 7, yet predicted with inferior inhibitory effectiveness ($\overline{\mathrm{DS}}$ − 10.9 kcal mol⁻¹) and physicochemical compatibility (polarizability 23.6 Å; logP 1.92).

Table 7.

ADMET-based pharmacokinetics and pharmacology of the studied compounds 1–11

Property	1	2	3	4	5	6	7	8	9	10	11	Unit
Absorption
Water solubility	− 5.218	− 2.198	− 1.56	− 2.105	− 2.325	− 3.346	0.464	− 2.208	− 1.821	− 1.636	− 3.084	(1)
Caco2 permeability	0.749	− 0.287	1.158	0.964	0.119	0.719	1.024	− 0.323	− 0.507	1.402	0.492	(2)
Intestinal absorption	74.557	44.678	91.115	90.109	70.846	90.87	93.184	37.843	40.811	83.922	30.044	(3)
Skin permeability	− 2.732	− 2.735	− 2.323	− 2.735	− 2.735	− 2.746	− 2.735	− 2.735	− 2.735	− 2.845	− 2.792	(4)
P-Glycoprotein substrate	Yes	No	No	Yes	Yes	Yes	No	No	No	No	Yes	(5)
P-Glycoprotein I inhibitor	Yes	No	No	No	No	No	No	No	No	No	No	(5)
P-Glycoprotein II inhibitor	Yes	No	No	No	No	No	No	No	No	No	No	(5)
Distribution
VDss	− 0.291	− 0.641	− 0.406	− 0.472	0.102	− 0.061	− 0.664	− 0.474	− 0.559	− 0.3	− 0.554	(6)
Fraction unbound	0.234	0.872	0.429	0.618	0.699	0.111	0.841	0.73	0.645	0.703	0.933	(6)
BBB permeability	− 1.076	− 0.761	− 0.257	− 1.178	− 1.138	− 1.374	− 0.244	− 1.042	− 1.267	− 0.823	− 1.005	(7)
CNS permeability	− 3.419	− 3.124	− 2.4	− 3.384	− 3.387	− 2.951	− 3.047	− 4.023	− 3.992	− 3.207	− 4.408	(8)
Metabolism
CYP2D6 substrate	No	No	No	No	No	No	Yes	Yes	Yes	No	No	(5)
CYP3A4 substrate	Yes	No	No	No	No	No	No	No	No	No	No	(5)
CYP1A2 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2C19 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2C9 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2D6 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP3A4 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
Excretion
Total clearance	1.355	0.62	0.746	0.915	0.887	1.197	0.329	0.532	0.682	1.879	0.636	(9)
Renal OCT2 substrate	No	No	No	No	No	No	No	No	No	No	No	(5)
Toxicity
AMES toxicity	No	No	No	No	No	No	No	No	Yes	No	No	(5)
Max. tolerated dose	0.114	1.336	0.446	− 0.057	0.959	0.465	1.433	1.946	1.103	0.706	0.96	(10)
hERG I inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
hERG II inhibitor	Yes	No	No	No	No	Yes	No	No	No	No	No	(5)
Oral rat acute toxicity	2.08	1.695	2.402	2.688	1.685	2.196	1.37	2.046	2.244	1.945	2.645	(11)
Oral rat chronic toxicity	2.191	1.889	1.847	2.237	2.518	2.862	2.378	1.636	1.853	1.095	3.25	(12)
Hepatotoxicity	Yes	No	No	Yes	Yes	No	No	No	Yes	No	Yes	(5)
Skin sensitisation	No	No	No	No	No	Yes	No	No	No	No	No	(5)
T. Pyriformis toxicity	0.288	0.27	0.183	0.285	0.285	0.287	0.266	0.285	0.285	− 0.246	0.285	(13)
Minnow toxicity	− 0.231	3.217	1.963	2.44	3.236	− 0.809	3.102	3.44	3.843	3.252	4.709	(14)

⁽¹⁾log mol L⁻¹; ⁽²⁾log Papp (10^–6 cm s⁻¹); ⁽³⁾%; ⁽⁴⁾log Kp; ⁽⁵⁾yes/no; ⁽⁶⁾log L kg⁻¹; ⁽⁷⁾log BB; ⁽⁸⁾log PS; ⁽⁹⁾log mL min⁻¹ kg⁻¹; ⁽¹⁰⁾log mg kg⁻¹ day⁻¹; ⁽¹¹⁾mol kg⁻¹; ⁽¹²⁾log mg kg⁻¹_bw day⁻¹; ⁽¹³⁾log μg L⁻¹; ⁽¹⁴⁾log mM

Table 8.

ADMET-based pharmacokinetics and pharmacology of the compounds 12–21 and D

Property	12	13	14	15	16	17	18	19	20	21	D	Unit
Absorption
Water solubility	− 2.552	− 2.678	− 3.401	− 1.486	− 1.85	− 0.074	− 1.595	− 3.346	− 2.679	− 2.856	− 2.641	(1)
Caco2 permeability	− 0.833	1.19	0.759	0.714	− 0.312	− 0.729	− 0.296	0.719	1.296	− 0.748	1.594	(2)
Intestinal absorption	2.281	95.22	89.387	93.346	9.98	7.469	28.565	90.87	96.452	0	94.994	(3)
Skin permeability	− 2.735	− 3.775	− 2.726	− 2.941	− 2.735	− 3.456	− 2.735	− 2.746	− 3.383	− 2.735	− 3.75	(4)
P-Glycoprotein substrate	No	No	Yes	No	No	No	No	Yes	No	Yes	No	(5)
P-Glycoprotein I inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
P-Glycoprotein II inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
Distribution
VDss	− 1.029	0.264	− 0.807	− 0.538	− 1.024	− 0.573	− 1.566	− 0.061	0.104	− 0.836	0.89	(6)
Fraction unbound	0.72	0.396	0.167	0.595	0.713	0.741	0.538	0.111	0.342	0.628	0.36	(6)
BBB permeability	− 1.659	− 0.059	− 0.728	− 0.481	− 0.861	− 1.453	− 1.223	− 1.374	− 0.396	− 1.562	− 0.081	(7)
CNS permeability	− 5.209	− 2.942	− 2.725	− 3.104	− 3.669	− 5.638	− 4.061	− 2.951	− 2.919	− 4.088	− 2.511	(8)
Metabolism
CYP2D6 substrate	Yes	No	No	No	No	No	No	No	No	No	No	(5)
CYP3A4 substrate	No	No	No	No	No	No	No	No	No	No	Yes	(5)
CYP1A2 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2C19 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2C9 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP2D6 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
CYP3A4 inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
Excretion
Total clearance	0.783	0.324	0.345	0.769	0.727	0.621	1.209	1.197	1.437	0.724	0.991	(9)
Renal OCT2 substrate	No	No	No	No	No	No	No	No	No	No	Yes	(5)
Toxicity
AMES toxicity	No	Yes	No	No	No	No	No	No	Yes	No	No	(5)
Max. tolerated dose	1.176	− 0.51	− 0.614	1.357	1.741	1.976	1.813	0.465	0.826	0.493	− 0.423	(10)
hERG I inhibitor	No	No	No	No	No	No	No	No	No	No	No	(5)
hERG II inhibitor	No	No	No	No	No	No	No	Yes	No	No	Yes	(5)
Oral rat acute toxicity	1.701	2.001	2.253	2.546	1.258	1.8	1.764	2.196	2.496	2.479	2.728	(11)
Oral rat chronic toxicity	4.069	1.13	1.895	0.634	3.319	4.591	2.745	2.862	1.978	2.959	0.966	(12)
Hepatotoxicity	No	Yes	No	Yes	No	No	No	No	Yes	Yes	Yes	(5)
Skin sensitization	No	No	No	No	No	No	No	Yes	No	No	No	(5)
T. Pyriformis toxicity	0.285	0.664	0.576	− 0.052	0.285	0.285	0.285	0.287	0.593	0.285	0.788	(13)
Minnow toxicity	2.902	1.962	0.972	2.254	3.664	3.631	3.075	− 0.809	0.933	6.31	1.675	(14)

⁽¹⁾log mol L⁻¹; ⁽²⁾log Papp (10^–6 cm s⁻¹); ⁽³⁾%; ⁽⁴⁾log Kp; ⁽⁵⁾yes/no; ⁽⁶⁾log L kg⁻¹; ⁽⁷⁾log BB; ⁽⁸⁾log PS; ⁽⁹⁾log mL min⁻¹ kg⁻¹; ⁽¹⁰⁾log mg kg⁻¹ day⁻¹; ⁽¹¹⁾mol kg⁻¹; ⁽¹²⁾log mg kg⁻¹_bw day⁻¹; ⁽¹³⁾log μg L⁻¹; ⁽¹⁴⁾log mM

Conclusions

This work preliminarily screens for promising bio-active agents in C. militaris ethanol extract for anti-Alzheimer progress, in which experimental evidences were used as the justification and input for computational analyses. The contents of cordycepin in the fungus are 0.32%. Liquid chromatography identified 21 bioactive compounds in the extract (1–21). Bioassays revealed the anti-Alzheimer potential of the extract against AChE with IC₅₀ value 115.9 ± 11.16 µg mL⁻¹. Quantum-based electronic properties suggest 7, 15, 17, 19, 21 as promising bio-compatible candidates given by their ground state energy (under − 450 a.u.) and dipole moment (over 5 Debye). Docking-based inhibitory properties suggest 5, 7, 9, 11, 17 as potential effective inhibitors given by their ligand-protein docking score ($\overline{\mathrm{DS}}$ < − 12 kcal mol⁻¹). QSARIS-based physicochemical properties confirm their biological compatibility (polarizability > 30 ų; logP < 0). ADMET-based pharmacological properties select 3 and 7 for their suitable pharmacokinetics on the brain (logBB < − 1; logPS − 3). Altogether, the results encourage further experimental isolation of l(−)-carnitine (7) as a prototype drug for experimental validation of the inhibitory mechanisms or behavioral model testing of pro-cognitive effectiveness.

Supplementary Information

Below is the link to the electronic supplementary material.

Data availability

The main data generated in this work is included in the manuscript and the supplementary information. The raw data can be shared upon request from the corresponding author.

Declarations

Conflict of interest

The authors declare no conflict of interest.