This study aims to investigate the potential of selected cannabinoid and stilbenoid compounds as antifungal agents by evaluating their binding affinities and interaction profiles with eight fungal proteins. The eight fungal proteins investigated are associated with signaling pathways and biochemical functions that play critical roles in cellular homeostasis and pathogenic processes. Important information is included to uncover the antifungal potential of these targets by describing the molecular roles, biological mechanisms, and cellular locations of these proteins. The selected proteins include GPI, Enolase, M2DH, GMP Synthase, DHODH, Hsp90, Chitin Synthase, and M1P5DH. These proteins were included in the study because they contribute to a wide range of biological functions. In particular, GPI plays a central role in GPI anchor biosynthesis by regulating the binding of proteins to the cell surface and maintains the stability of the cellular structure (Álvarez-Sánchez et al., 2024). Therefore, GPI was selected as an antifungal target due to its potential to interfere with cell wall-associated biosynthetic processes. Enolase is involved in the basic pathways of energy metabolism such as glycolysis and ribonucleotide catabolism, and inhibition of these processes may prevent fungal growth by stopping energy production (Dasari et al., 2019; Díaz-Ramos et al., 2012). M2DH contributes to energy storage and osmotic balance by participating in the metabolism of mannitol and other polyol derivatives, which makes it an important strategic target that can weaken cellular stress resistance (Upadhyay et al., 2015). GMP Synthase plays a critical role in nucleotide biosynthesis pathways, especially ribonucleotide and purine biosynthesis, and inhibition of these processes can directly affect cell division (Oliver et al., 2013). DHODH supports de novo pyrimidine biosynthesis by localizing to the inner mitochondrial membrane, which is important for the production of nucleotides required for DNA and RNA synthesis (Hai et al., 2024; Rawls et al., 2000). Hsp90 regulates the resistance of pathogenic fungi to environmental stresses by acting in protein folding, stabilization and biological regulation processes, and therefore has significant potential as an antifungal target (Hoter et al., 2018; Wei et al., 2024). CaChs2 plays a critical role in cell wall biosynthesis through chitin synthesis and glycosyltransferase activities; since chitin is one of the main components of the fungal cell wall, inhibition of this protein may lead to weakening of the cell wall (Lenardon et al., 2010; Merzendorfer, 2011). M1P5DH is localized in the cytoplasm and contributes to energy storage and osmotic balance processes through mannitol metabolism; therefore, inhibition of mannitol may weaken the resistance of fungi to osmotic stress (Lim et al., 2021). Selecting proteins from different fungal strains increases the versatility and scope of our study. There are phylogenetic differences between these strains, and it is anticipated that proteins from each strain may exhibit unique properties in ligand binding performance. This approach has allowed us to better understand the diversity of ligand trafficking in proteins from different strains and to conduct comprehensive analyses in a wide biological spectrum. In addition, this diversity is crucial for increasing the potential of the selected ligands to work in a wide range of fungi as their biological targets, thereby enhancing the overall applicability of our findings. Furthermore, the multi-target nature of this study highlights the potential for synergistic effects by simultaneously inhibiting multiple fungal proteins involved in distinct yet complementary pathways. For instance, targeting proteins associated with energy metabolism, cell wall synthesis, and stress response simultaneously could enhance antifungal efficacy by disrupting multiple critical functions (Fisher et al., 2024).In our study, the selected protein targets were evaluated as protein models that could provide potential therapeutic effects on clinically important pathogenic fungal species. These protein targets are particularly associated with pathogenic species such as Candida albicans and Aspergillus fumigatus, which are the main causes of hospital-associated infections and can develop resistance to treatment. For example, Hsp90 plays a critical therapeutic role in important fungal pathogens such as Candida albicans, and this pathogen is known for its nosocomial infections and resistance to treatment. Aspergillus fumigatus, where the other target protein is GMP synthase, is another important pathogen that can become resistant to treatment and commonly cause infections in hospitals. In addition, proteins such as CaChs2 and Enolase AfEno1 are proteins that can potentially contribute to treatment by targeting the resistance mechanisms exhibited by pathogens such as Candida spp. and Aspergillus spp. against antifungal treatment. The selected protein targets are of greater clinical importance due to the resistance of fungal infections to treatment. For example, GPI is a critical component of the cell wall of Candida albicans, while enzymes such as M2DH and M1P5DH are part of critical metabolic pathways in Aspergillus fumigatus, and DHODH is an important biosynthetic enzyme in other fungal species such as Aspergillus nidulans. Selection of these proteins is directly linked to resistance to therapy in different fungal species, and inhibition of these proteins may allow the development of new therapeutic strategies.

In light of recent comprehensive reviews, such as (Rabaan et al., 2023; Fisher et al., 2022), which emphasize the increasing prevalence of antifungal resistance, the importance of multi-target strategies becomes even more apparent. Antifungal resistance mechanisms, including efflux pump overexpression, biofilm formation, and genetic alterations in target sites, significantly hinder the efficacy of current therapies. Addressing these mechanisms through a multi-target approach not only has the potential to mitigate resistance but also aligns with global initiatives aimed at developing novel antifungal agents with broader activity spectra and reduced resistance potential. While the potential for antagonistic interactions between targets must be carefully evaluated to optimize combination therapies, such strategies represent a promising avenue for enhancing treatment efficacy and overcoming the limitations of current antifungal agents.

Functional network analyses have revealed potential synergistic interactions and shared functions of these proteins in biological processes. For example, Enolase and GMP Synthase play similar roles in energy metabolism by participating in molecular functions such as “ATP binding”; Chitin Synthase and GPI share similar functions in maintaining cellular stability by participating in cell wall biosynthesis. Hsp90 and GMP Synthase are localized in the cytosol and plasma membrane, supporting cellular resistance mechanisms. These findings are important for our understanding of how proteins participate in various biological pathways and molecular functions and the coordination between them. By employing MD and MDS, ligands with the highest binding affinities were identified, and their pharmacokinetic profiles were analyzed using ADMET. The docking studies indicated that GMP synthase had the highest binding affinity with Cannabistilbene I (−9.1 kcal/mol), reflecting hydrophobic solid interactions and multiple hydrogen bonds. Similarly, CaChs2 demonstrated strong binding with Cannabistilbene I, suggesting its potential as an inhibitor. In contrast, ligands such as Cannabinolic acid and 8-hydroxycannabinolic acid exhibited moderate binding affinities with the target proteins, emphasizing the variability in interaction strengths among different compounds (Appendino et al., 2011; Kogan and Mechoulam, 2007).

ADMET analysis provided important information on the pharmacokinetic properties of the ligands. Hexahydrocannabinol and Δ9-cis-THC showed high lipophilicity and significant intestinal permeability; however, these ligands also carry the risk of HERG channel inhibition, which may limit their therapeutic applications. In contrast, ligands such as Cannabistilbene I and Cannabiglendol stand out as more suitable candidates, exhibiting lower risk of HERG inhibition and moderate lipophilicity. Solvent exposure analysis revealed significant differences in the binding environments of protein-ligand complexes. In particular, solvent exposure was observed for DHODH-Cannabicitran and GMP synthase-Δ9-cis-THC complexes, indicating potential changes in ligand stability and protein interactions in these environments (Jiang et al., 2019).

Consistent with the study by Mulia et al., ∆(9)-THC derivatives exhibited pharmacokinetic properties such as high lipophilicity (logP >5), intestinal permeability, and moderate brain penetration (Mulia et al., 2021). In our study, the calculated QPlogPo/w (5.663) and QPPCaco (4571.459) values for Δ9-cis-THC were found to be consistent with these results. The study of Thomas et al. examined the pharmacological properties of Δ(9)-THC analogs in more depth, and showed the effects of lipophilicity on a compound. In particular, it was stated that the addition of long side chains increased lipophilicity by approximately 3 times per CH2 group, and although no direct correlation with pharmacological activity was found, it was emphasized that lipophilicity plays a critical role in pharmacokinetic processes (Thomas et al., 1990). The QPlogPo/w value in our study (5.663) confirms the natural high lipophilicity of ∆(9)-THC, and as stated by Thomas et al., it also draws attention to the role of this lipophilicity in pharmacokinetic processes. However, the QPlogHERG value of Δ9-cis-THC (−4.894) indicates the risk of cardiotoxicity, complementing the findings of both Mulia and Thomas. Chetia et al., addressing the toxic effects of THC on brain dopamine levels and the attenuation of these effects with Cannabidiol, emphasized that THC affects the dopaminergic (DAergic) system in the brain, especially stimulating mesolimbic DA-containing neurons and increasing striatal dopamine levels. However, thanks to the anxiolytic and antipsychotic properties of CBD, it has been stated that THC has the potential to improve these variations on the DAergic system (Chetia and Borah, 2020). Δ9-cis-THC (Δ9-cis-THC) is an isomer that has attracted attention in low-THC Cannabis sativa cultivars, where it is found at levels comparable to Δ9-trans-THC (Jagannathan, 2020). This ligand was detected especially in low THC fiber hemp species but not in high THC medical cannabis varieties. This suggests that the natural abundance of Δ9-cis-THC may be related to genetic or environmental factors. It has been stated that Δ9-cis-THC has high enantiomeric purity (80%–90%) and this feature provides a significant advantage in the investigation of its biological effects (Schafroth et al., 2021). In our study, the selection of Δ9-cis-THC to evaluate its antifungal potential reveals the possible importance of both its natural abundance and chemical properties in interaction with target proteins. In particular, Δ9-cis-THC, which shows strong binding to targets such as GMP synthase, is considered a promising candidate for the development of new antifungal agents due to its potential to be easily obtained at low cost and its stability in biological systems. In this context, information on its natural abundance increases the importance of our study in terms of the sustainability and large-scale production of the compound.

The study by O'Croinin et al. addressing the pharmaceutical potential and predictive pharmacokinetic properties of stilbenes provides valuable data regarding the anti-inflammatory, antioxidant and anticancer effects of stilbenes, especially those derived from Cannabis sativa (O’Croinin et al., 2023). In the study, the pharmacokinetic profiles of Cannabistilbene I and other stilbenes were evaluated with parameters such as the potential to cross the BBB and the ability to pass through human skin and jejunal tissues. In this context, the high BBB permeability of Cannabistilbene I (84%) and its compliance with optimal Lipinski parameters indicate that the molecule is a promising candidate pharmacologically for neurological targets. In our study, ADMET analyses of Cannabistilbene I support its high lipophilicity and reliable bioavailability properties, and also emphasize the effects of the molecule especially on inflammatory mechanisms. The findings of both studies indicate that the physiochemical and pharmacokinetic properties of stilbenes increase the therapeutic potential, strengthening the consistency of the data in the current literature. These results reveal the need for further research to develop pharmaceutical formulations of stilbenes. The effect of Cannabistilbene I on Angiotensin II (Ang II)-induced cardiac hypertrophy was studied and how this effect was modulated by cytochrome P450 (CYP) enzymes and arachidonic acid (AA) metabolites. The findings indicate that Cannabistilbene I attenuates the cardiac hypertrophy-inducing effects of Ang II, reduces the increase in cellular surface area, and regulates the expression of hypertrophic marker genes. In addition, Cannabistilbene I was found to provide cardioprotective enzymatic activity by increasing CYP1A1 gene expression, increasing the levels of the metabolite 19(S)-HETE, and reversing the decline induced by Ang II (Alammari et al., 2024).

In our study, it was determined that stilbene derivatives such as Cannabistilbene I exhibit high lipophilicity and anti-inflammatory properties, and also show potential therapeutic effects in the cardiovascular system. In particular, ADMET analyses emphasized that Cannabistilbene I has BBB permeability, but its low lipophilic profile may reduce cardiotoxicity. The effects on Ang II-induced cardiac hypertrophy support the protective role of Cannabistilbene I against cardiovascular diseases and are consistent with our ADMET-based pharmacokinetic findings. In particular, stilbenes' effects on inflammation and cellular stress response were emphasized, demonstrating a cardioprotective potential at both molecular and cellular levels. In our study, the ADMET profile of Cannabidiolic Acid (CBDA) was evaluated in detail. Our findings regarding the pharmacokinetic and pharmacodynamic properties of CBDA largely overlap with the research conducted by Formato et al. Formato et al. In their study, the molecular weight (MW) of CBDA was determined as 358.48, the logP value measuring hydrophobicity was 6.43, the number of hydrogen bond acceptors (HBA) was 4 and the number of donors (HBD) was 3. In addition, the topological polar surface area (TPSA) of CBDA was calculated as 77.75 and it was determined that it contained 7 rotatable bonds (NRTOB). They emphasized that these parameters were positive in terms of oral bioavailability and showed that CBDA has a good absorption potential (Formato et al., 2020).

The antifungal activities and ADMET profiles of selected compounds such as Cannabistilbene I and Δ9-cis-THC were compared with those of the widely used clinical antifungal agents fluconazole (Kauthale et al., 2017) and amphotericin B (Khosravi et al., 2022) in this study. In the literature, it has been reported that the minimum inhibitory concentration (MIC) values of fluconazole against fungal pathogens such as Candida albicans, Aspergillus niger, and Aspergillus flavus are between 3.12 and 25 μg/mL, while amphotericin B, despite its high activity, has serious disadvantages such as nephrotoxicity and high cost. In our study, Cannabistilbene I exhibited high binding affinity with GMP Synthase (−9.1 kcal/mol) and CaChs2 (−9.1 kcal/mol), while Δ9-cis-THC showed a lower binding affinity with GMP Synthase (−7.2 kcal/mol). ADMET analyses show that Cannabistilbene I has safety properties similar to fluconazole with a low risk of cardiotoxicity and a high bioavailability profile. On the other hand, Δ9-cis-THC carries pharmacokinetic advantages such as high lipophilicity (QPlogPo/w = 5.663) and intestinal permeability (QPPCaco = 4571.459), while its higher hERG channel inhibition potential (−4.894) compared to fluconazole and amphotericin B poses a possible risk in terms of cardiotoxicity. While the efficacy of fluconazole is supported by MIC values, the antifungal activities of the compounds in our study were based on their binding affinities, which requires in vitro and in vivo validation to better understand their clinical potential. Cannabistilbene I and Δ9-cis-THC can be considered as innovative and potential candidates in antifungal therapy; However, comprehensive pharmacokinetic and toxicological studies as well as experimental confirmations are needed before it can be considered as an alternative to fluconazole and amphotericin B (Elias et al., 2022; Ghobadi et al., 2023; Martínez et al., 2024; Parveen and Balamurugan, 2023).

In our study, similarly strong oral bioavailability signals were found for CBDA and it was shown that the logP value was calculated as 6.2, TPSA as 79.1, HBA as 4, and HBD as 3. In addition, in our protein target analyses, CBDA’s G-protein coupled receptor (GPCR) binding score was evaluated as −0.41, ion channel modulator score as −0.06, kinase inhibitor score as −0.72, and nuclear receptor binding score as −0.28. These findings are very close to the values of −0.39, −0.05, −0.74, and −0.30 in the study by Formato et al. In addition, minimal deviations were observed in protease inhibitor and enzyme inhibitor activities, and the protease inhibitor score was found as −0.65 in our study and −0.63 in Formato et al. The pharmacokinetic profiles and antiseizure effects of phytocannabinoid acids (including CBDA) were examined in detail in the study by Lyndsey L. Anderson et al. (Anderson et al., 2019). In this study, it was reported that CBDA was rapidly absorbed and its brain-to-plasma ratio was low (≤0.04). However, when a different Tween 80-based carrier was used, this ratio increased to 1.9, emphasizing that this increased the brain penetration of CBDA. In addition, it was shown that CBDA delayed the occurrence of generalized tonic-clonic seizures by increasing the seizure threshold in mice with Dravet syndrome. Our study, on the other hand, examined the pharmacokinetic properties and biological activities of CBDA at the molecular level, including the evaluation of important pharmacokinetic parameters of CBDA such as QPlogPo/w, TPSA and NRTOB. The report of a low brain-to-plasma ratio by Anderson et al. and their emphasis on anti-seizure effects are parallel to our results confirming the pharmacokinetic potential of CBDA. In particular, the finding that the bioavailability and brain penetration of CBDA may vary depending on the carrier supports the relevance of our pharmacokinetic assessments for clinical applications. Current studies on cannabicitran have indicated that this compound is often found at levels of up to 10% in commercial “purified” CBD extracts. The literature suggests that cannabicitran is racemic in nature and may be formed during herbal extraction processes (Wood et al., 2023). In contrast, the evaluations made on the pharmacokinetic profile of cannabicitran in our study provided new information, especially on its structural and pharmacological properties. ADMET results obtained from the literature provide critical information regarding the bioavailability of cannabicitran and its interaction with target receptors. For example, it is known that this compound has affinity for receptors such as 5-HT1A. While our study provides preliminary information for a better understanding of the metabolic and biological effects of cannabicitran, it can be expanded to provide full compliance with the literature with more comprehensive ADMET analyses and abundance data. In this context, in-depth studies are required at both the biosynthesis and biological effects levels to evaluate the pharmacological potential of cannabicitran.

The analyses made on CBDVA in our study evaluated the pharmacokinetic properties and potential anticonvulsant effects of the compound. In our study, CBDVA significantly increased the thermal seizure threshold at certain doses (e.g., 100 mg kg⁻¹), indicating the antiseizure effect of the compound. These findings support a study by Anderson et al. (2021). In the study by Anderson et al., it was reported that CBDVA, as well as other phytocannabinoid acids such as CBGA and CBGVA, exhibited protective effects against thermal seizures in the Dravet syndrome mouse model. In particular, this study showed that CBDVA had a neuroprotective effect against thermal seizures after acute intraperitoneal administration, but Δ9-THCV (at a dose of 3 mg kg⁻¹) had a proconvulsant effect in contrast. Our study suggests that CBDVA may be a promising agent in the treatment of epileptic diseases such as Dravet syndrome, and supports the fact that this compound should be considered in future drug development processes for epilepsy, together with the results of Anderson et al. Among the 16 ligands evaluated in our study, considering that certain ligands may be higher in natural abundance than others, this situation stands out as an important finding when considered together with RMSD and RMSF analyses. In particular, ligands exhibiting high RMSF stability and low RMSD values supported strong interactions with the target protein. These ligands, which are naturally more abundant, offer the potential to improve both their availability in biological systems and their pharmacokinetic profiles. Considering this relationship, the findings supported by protein-ligand contact and ADMET predictions suggest that naturally more abundant compounds may be more pharmacologically suitable candidates. For example, although the natural abundances of ligands such as cannabicitran have been reported in the literature, in this study they were found to support this assumption with their strong stability and interaction profiles. These findings demonstrate that our study is not only relevant to pharmacological profile assessments, but also to the effect of ligand abundance on therapeutic potential. RMSD analyses are crucial for evaluating the structural stability of protein-ligand complexes over time. In the GMP synthase-Δ9-cis-Tetrahydrocannabinol (Δ9-cis-THC) complex, the protein RMSD values remain stable between 2–4 Å, indicating the structural stability of the protein-ligand interaction. This suggests that the ligand binds firmly and stably to the protein. On the other hand, in the CaChs2-Cannabistilbene I complex, the protein RMSD values fluctuate between 6–10 Å, indicating some flexibility in the protein structure. This flexibility allows more movement in the ligand binding site, leading to higher ligand RMSD values (4–7 Å). In the DHODH-Cannabicitran complex, the protein RMSD values remain steady between 1–2 Å, demonstrating high structural stability; however, the ligand RMSD values fluctuate widely between 10–20 Å, indicating significant movements within the ligand binding site, which might be related to solvent exposure.

RMSF analyses are critical for evaluating protein-ligand interactions' dynamic flexibility and stability. During the interaction between CaChs2 and Cannabistilbene I, high protein stability was observed, with slight increases in ligand RMSF values at specific atoms. In contrast, the interaction with Cannabinolic acid showed very low ligand RMSF values, indicating that Cannabinolic acid stabilizes CaChs2, resulting in less dynamic flexibility in the protein-ligand interaction. In the interactions of DHODH protein with Cannabinolic acid and Cannabiglendol, moderate protein mobility was observed, but the ligand RMSF values were lower compared to the protein. These results highlight the impact of ligands on protein stability and reveal dynamic interactions in specific regions.

For Enolase, interactions with 8-hydroxycannabinolic acid showed increased mobility at certain atoms, whereas interactions with Cannabinolic acid indicated that the ligand stabilized the protein. The Glycosylphosphatidylinositol protein did not show significant mobility differences in interactions with 3-Butyl-Δ9-THC and Δ9-Tetrahydrocannabinorcol. During the interaction of GMP synthase with Cannabistilbene I, an increase in protein RMSF values was observed, while no significant difference was noted with Δ9-cis-THC. The Hsp90 protein displayed high mobility when interacting with 5-Hydroxy-7-methoxyindan-1-spiro cyclohexane, whereas it exhibited low RMSF values with Hexahydrocannabinol. The M1P5DH protein showed low RMSF values and stability in interactions with both 8-hydroxycannabinolic acid and Cannabinolic acid. However, in the interaction of Mannitol-2-dehydrogenase protein with 8-Hydroxy-Δ9-Tetrahydrocannabinol, a significant increase in RMSF was observed, particularly after atom 25, whereas no significant difference was noted with Δ9-THC. These results elucidate the impact of ligands on protein stability and reveal dynamic interactions in specific regions.

The differences between the binding affinities obtained in our study are due to the diversity in the chemical and structural properties of the protein targets and ligands. In particular, the hydrogen bonding potentials, hydrophobic surface areas and electrostatic interactions of the ligands are among the determining factors of these differences (Chen et al., 2016; Patil et al., 2010). Differences in the active sites of protein targets may also contribute to the changes observed in ligand binding affinities. For example, while extensive hydrogen bonding networks in proteins such as Hsp90 lead to high binding affinities, hydrophobic regions in targets such as GMP Synthase may be effective in ligand selection (Hoxie and Street, 2020). In similar studies in the literature, studies comparing the binding properties of different ligands to the same protein targets show that the changes in binding affinities are affected by the chemical modifications of the ligands or the conformational flexibility of the protein targets (Ferenczy and Kellermayer, 2022; Madushanka et al., 2023; Patil et al., 2010). In future studies, detailed examinations are planned with methods such as molecular dynamics simulations and energy component analysis of ligand-protein complexes to better understand the differences in binding affinities. This approach may contribute to the design of structural modifications that will increase the binding affinities of ligands.

The choice to employ MD and MDS in this study is well-founded due to their complementary nature in evaluating protein-ligand interactions. Molecular docking serves as an initial screening tool, allowing for the identification of the most promising ligand candidates based on binding affinities (Agu et al., 2023; Salmaso and Moro, 2018). However, docking alone cannot fully capture these interactions' dynamic and temporal aspects, which are crucial for understanding the stability and behavior of protein-ligand complexes in a physiological environment. Therefore, MDS was utilized to provide a more comprehensive assessment of the binding stability and dynamic interactions over time (Shukla and Tripathi, 2021). Furthermore, the decision to utilize a 100 ns simulation duration for MDS was based on a balance between computational feasibility and the ability to capture significant conformational changes and stabilization patterns within the protein-ligand complexes (Koshy et al., 2010). While longer simulations could potentially provide additional insights, previous studies have demonstrated that a 100 ns timeframe is sufficient to observe key interaction features, stabilize RMSD values, and gain meaningful insights into protein flexibility and ligand binding modes. Therefore, this duration was considered appropriate for effectively investigating the dynamic behaviors of the complexes and their potential antifungal properties (Al-Karmalawy et al., 2021). The combination of MD followed by MDS ensured a robust evaluation of the selected ligands’ potential as antifungal agents, supporting their future experimental exploration as promising drug candidates.

This study provides a comprehensive analysis of interactions between various fungal proteins and ligands, revealing the antifungal potential of selected compounds. Cannabistilbene I exhibited strong binding affinities (−9.1 kcal/mol) and stable dynamic profiles with GMP Synthase and CaChs2, while ADMET analyses highlighted its high bioavailability and low cardiotoxicity risk. Similarly, Δ9-cis-THC stood out as another promising candidate due to its natural abundance and protein-ligand stability. These findings suggest that both molecules are promising candidates for antifungal drug development and underscore the importance of further experimental validation.