Synthesis, antifungal activity, and molecular dynamics study of novel geranyl aromatic sulfonamide compounds as potential complex III inhibitors

Essential oils (EOs), as unique natural products, are promising resources for the discovery of green agrochemicals. The main ingredient geraniol of citronella oil was found to exhibit substantial antifungal activities in this study. Therefore, a series of novel geranyl aromatic sulfonamide compounds were synthesized and found to display considerable antifungal activities. Two geranyl thiofuran-sulfonamide compounds 4c-1 (median effective concentration (EC50) against Rhizoctonia solani: 24.97 mg/L and EC50 against Sclerotinia sclerotiorum: 27.26 mg/L), 4c-2 (EC50 against S. sclerotiorum: 18.53 mg/L) and one geranyl pyridine-sulfonamide compound 4d-2 (EC50 against R. solani: 29.31 mg/L and EC50 against S. sclerotiorum: 29.98 mg/L) were screened as “star molecules” due to their excellent antifungal activities. The preliminary structure-activity relationship (SAR) study revealed that the introduction of various aromatic heterocycles maybe an efficient protocol to improve the fungicidal activities of geranyl aromatic sulfonamide compounds. The molecular mechanisms of the geranyl aromatic sulfonamide compounds were clarified by performing molecular docking and molecular dynamics (MD) simulations. Three “star molecules” of these geranyl aromatic sulfonamide compounds were found to bind to Complex III through several hydrogen bonds and π-interactions with crucial residues TRP17, GLY20 etc. Their binding free energies were calculated to be strong ranging from −50.60 to −39.44 kcal/mol by MM/GBSA method, which suggested the geranyl aromatic sulfonamide compounds were potential Complex III inhibitors. The main component originating from the natural plant EOs ought to be studied in the future to discover novel pathogenic fungicidal candidates. Graphical abstract Graphical abstract


Introduction
In recent years, plant diseases have received increasing attention because of their negative effects not only on the yields but also on the quality of crops worldwide [1]. Numerous fungicides are continuously used to control various plant diseases, causing the development of extraordinary pathogen resistance to known fungicides [2]. Therefore, the discovery of novel fungicides is particularly crucial in modern plant disease control.
Plant essential oils (EOs) are some natural volatile oily liquids and display sufficient biosafety to be applied widely in cosmetics and antimicrobials of humans [3][4][5][6][7][8]. Many special ingredients and combinations of EOs were recognized and reported to possess some pharmacological properties such as insecticidal activities [9,10], antibacterial activities [11,12], and antifungal activities as well [13,14]. Citronella oil, a well-known EO, was found to be mosquito repellent [15] or insecticide against Myzus persicae [16]. In addition, citronella oil also exhibited excellent antibacterial or antifungal activity in the area of medicine [17][18][19]. As one of the main components of citronella oil, geraniol was also reported to possess valuable antifungal property against Trichophyton rubrum or sensitize fungi to commercial antifungal agent such as fluconazole in humans [20,21]. However, to date, studies on citronella oil or its main ingredient geraniol as agricultural fungicides are still limited. As a consequence, the use of natural citronella oil or its main components remains prospective in the development of new green antifungal agents.
The cytochrome bc1 complex (Complex III) is a vital protein in fungal organisms that comprises the middle part of the mitochondrial respiratory chain [22,23]. Complex III could catalyze reversible electron transferring from ubiquinone to cytochrome c, further coupled with proton translocation through the "Q-cycle" across the inner mitochondrial membrane. When inhibitors bind to the ubiquinone reaction site of Complex III and thus modulate the Q-cycle to further exhibit the antifungal activities [24]. As is known to us, the sulfonamide fungicides Amisulbrom, Cyazofamid, and natural-sourced Antimycin A (Scheme 1) were reported to both bind to the Qi site of Complex III to block the mitochondrial electron transfer between cytochrome b and c [23,25]. Notably, for the sulfonamide fungicides, the sulfonamide azolyl pharmacophores may be responsible for their efficient fungicidal activities [26]. Additionally, benzene sulfonamide compounds were reported to display higher binding affinity to Complex III than triazole sulfonamides [27]. Hence, the fragment of aromatic sulfonamides should be noted to construct new Complex III inhibitor candidate.
Consequently, natural citronella oil and its main ingredients were selected for the antifungal activity evaluation against different phytopathogenic fungi in this study. A variety of geranyl aromatic sulfonamide compounds were designed, synthesized, and subsequently evaluated for their fungicidal activities. Three "star molecules" of the geranyl aromatic sulfonamide compounds were screened successfully possessing considerable fungicidal activities. Molecular docking, molecular dynamics simulations, and MM/ GBSA analysis were performed to clarify the binding mechanisms between the geranyl aromatic sulfonamide compounds and their potential target, Complex III. Conclusively, these geranyl aromatic sulfonamide compounds are found to become fungicidal candidates as potential Complex III inhibitors, which ought to be noted in future researches.

Results and discussion
Screening the fungicidal activities of Citronella oil Citronella oil was chosen as the basic material in this study owing to its previously reported multiple excellent bioactivities compared to other EOs [17][18][19]. The fungicidal activities of the entire original citronella oil and its three main components, geraniol, citronellol and citronellal (Scheme 2), were evaluated against two different plant pathogenic fungi, R. solani, and Gibberella saubinetii, at a concentration of 50 mg/L. Their corresponding fungicidal activities are demonstrated in Table S1 and Fig. 1. The entire original citronella oil displayed satisfactory activities against R. solani (inhibition rate greater than 30%) and G. saubinetii (inhibition rate greater than 20%). These results were slightly better than the previous report on citronella oil with an~15% inhibitory rate against R. solani [28]. Remarkably, one of the main ingredients of citronella oil, geraniol, exhibited the best fungicidal activities aginst R. solani and G.saubinetii (inhibition rate approximately 40% and 20% separately) when compared to the other two ones, namely citronellol and citronellal, which were even better than the corresponding activities of the entire original citronella oil itself. Therefore, geraniol, as the main source of the fungicidal activities of citronella oil, was chosen to be a suitable lead to be further optimized in the following study to explore novel fungicidal agents.

Synthesis and fungicidal activities evaluation of novel geranyl aromatic sulfonamide compounds
Sulfonamide compounds have aroused increasing interest in recent years due to their extensive antifungal activities [26,27]. Therefore, a series of novel geranyl benzenesulfonamide compounds were designed and synthesized firstly based on the chemical structure of lead, geraniol in our study. The synthesized procedures were shown in Scheme 3. Geraniol was selected as the original material 1 to obtain intermediate 2 with phthalimide via the Mitsunobu reaction (DIAD and PPh 3 served as the acid-binding agent and catalyst, respectively). The dry reaction environment was vital for the high yields of intermediate 2. Subsequently, intermediate 2 was reduced to obtain the key intermediate 3, geranylamine, catalyzed by hydrazine hydrate. All target compounds of series 4a were prepared from geranylamine through a condensation reaction utilizing the corresponding benzene sulfonyl chloride analogs with trimethylamine as the acid-binding agent. Then three compounds of series 4a (4a-3, 4a-5, and 4a-8) were methylated to generate the corresponding compounds of series 5a under the circumstances of sodium ethoxide and iodomethane. All structures of the synthesized geranyl benzenesulfonamide compounds were confirmed by recording 1 H-NMR, 13 C-NMR, and ESI-MS spectrums, which were shown in Figs. S1-S32 in supplementary information.
The bioactivities of sixteen firstly synthesized geranyl benzenesulfonamide compounds were evaluated against two plant pathogenic fungi, R. solani and G. saubinetii, at a concentration of 50 mg/L in the present study. The results of the fungicidal activities are presented in Table 1 and Fig.  S49. Most of the geranyl benzenesulfonamide compounds appeared to exhibit considerable antifungal activities against R. solani and G. saubinetii, with inhibitory rates ranging from 40% to 60%, which were generally better than the lead compound, geraniol. Among the synthesized geranyl benzenesulfonamides, four compounds, 4a-1, 4a-3, 4a-4, 4a-5, and 4a-11 showed a~50% inhibitory rate against R. solani. Two compounds, 4a-3 and 4a-5, also showed greater than or approximately 50% inhibitory rates against G. saubinetii. Nevertheless, compounds 5a-1, 5a-2, and 5a-3, in which the sulfonamide group was substituted with a methyl group, exhibited substantially decreased antifungal activities, with only 10-30% inhibitory rate left. Hence, it was boldly inferred that the aromatic sulfonamides with the fragment of geraniol might have considerable fungicidal activities to be of worth for further optimization.
Therefore, some aromatic heterocyclic groups were introduced to further design and synthesize novel geranyl aromatic sulfonamide analogs 4b-4d using the three-step reaction mentioned above (Scheme 3) in the present study. All structures of the synthesized compounds 4b-4d were also confirmed by 1 H-NMR, 13 C-NMR and ESI-MS spectrums, which were shown in Figs. S33-S48 in supplementary information.
The bioactivities of the newly synthesized eight geranyl aromatic sulfonamide compounds and their lead geraniol were firstly evaluated against the same two plant pathogenic fungi R. solani and G. saubinetii tested above at a concentration of 50 mg/L. The fungicidal activities are demonstrated in Table S2 and Fig. 2. As shown in Fig. 2, almost all of the newly synthesized geranyl aromatic sulfonamide compounds displayed considerably higher fungicidal activities than geraniol and most of them showed similar fungicidal activities to those synthesized geranyl benzenesulfonamide compounds. Meanwhile, the bioactivities of these compounds were also evaluated against other four different plant pathogenic fungi S. sclerotiorum, Valsa mali, Gibberella fujikuroi, and Magnaporthe oryzae (Table  S2 and Fig. 2). It was surprising that the majority of newly synthesized geranyl aromatic sulfonamide compounds, for example, compounds 4c-1, 4c-2, 4c-3, 4c-4, and 4d-2 (with inhibition rates higher than 90%) displayed good and specific fungicidal activities against S. sclerotiorum. Especially, compound 4c-2 and 4d-2 displayed inhibition rates of 99% and 100% against S. sclerotiorum, respectively, whose fungicidal activities were similar to that of the commercial fungicide Pyraclostrobin (100%). Consequently, the EC 50 values of highly effective synthesized geranyl aromatic sulfonamide compounds were further determined in this study. The results are shown in Table 2. Satisfactorily, the five synthesized geranyl aromatic sulfonamide compounds possessed considerable EC 50 values against S. sclerotiorum, particularly compound 4c-2, which exhibited the lowest EC 50 value of only 18.53 mg/L. Moreover, compound 4c-1 displayed good fungicidal activities against not only S. sclerotiorum but also R. solani with the EC 50 values of 27.26 and 24.97 mg/L, respectively. Compound 4d-2, whose fungicidal activities were slightly inferior to that of 4c-1, possessing EC 50 values of 29.98 and 29.31 mg/L against S. sclerotiorum and R. solani, respectively. Thus, the aforementioned compounds, 4c-1, 4c-2, and 4d-2, were designated as "star molecules" among the synthesized geranyl aromatic sulfonamide compounds, which was shown in Scheme 4.
The preliminary structure-activity relationship (SAR) of the 24 synthesized geranyl aromatic sulfonamide compounds and their bioactivities were then discussed in this study. Some compounds with electron-donating substituted groups, like compounds 4a-1, 4a-2 4a-3, and 4a-8, 4a-11, whose benzene ring was substituted by methyl, methoxy or tert-buty, displayed moderate fungicidal activities regardless  of the position of the substituent groups. However, the introduction of multiple electron-donating substituted groups, like compounds 4a-9 and 4a-12, resulted in decrease or elimination in the bioactivities. On the other hand, some compounds whose hydrogen atoms of benzene rings were substituted by an electron-withdrawing group generally exhibited better fungicidal activities than those benzene ring substituted by an electron-donating group. For example, 4a-4 and 4a-5 with strong electron-withdrawing trifluoromethyl or nitro groups both exhibited the highest inhibitory rates against R. solani. Similarly, the introduction of multiple electron-withdrawing groups, like chlorine substituents (4a-6), resulted in disappearance of fungicidal activities as well. Moreover, if the hydrogen atom of the amine group in geranylamine was replaced, such as in compounds 5a-1, 5a-2, and 5a-3, the fungicidal activities were decreased substantially compared to their corresponding unsubstituted compounds 4a-3, 4a-5, and 4a-8, which may suggest that it was detrimental to increase the steric hindrance of the sulfonamide moiety for the fungicidal activities of compounds. Besides, it was worth noting that the geranyl sulfonamide compounds with aromatic heterocycles thiofuran or pyridyl groups substituted universally exhibited good bioactivities against six tested phytopathogenic fungi except for compound 4d-1, especially for S. sclerotioru, which was almost completely inhibited by compounds 4c and 4d at a concentration of 50 mg/L. Consequently, it was speculated that it may be a beneficial strategy to introduce various substituted aromatic heterocycles for improving fungicidal activities of geranyl sulfonamide compounds.

Molecular docking analysis of novel synthesized geranyl aromatic sulfonamide compounds
The sulfonamide fragment was regarded as an important pharmacophore of fungicides inhibiting complex III [26]. As a consequence, the complex III was selected as the potential target to investigate the molecular mechanisms of the newly synthesized geranyl aromatic sulfonamide compounds in this study [29]. First of all, the re-docking of the crystal structure of Complex III co-crystallized with Antimycin A (PDB ID: 1PPJ) was performed in this work. The RMSD value of Antimycin A was calculated to be 0.76 Å between its co-crystallized pose and the docked pose, indicating docking was a reliable protocol to predict the interaction modes of novel ligands and Complex III. Subsequently, "star molecules" 4c-1, 4c-2, 4d-2, and the commercial sulfonamide fungicide Amisulbrom were docked separately into the binding site (Qi site) of Complex III. Results of Surflex Docking score and different interactions (hydrogen bonds) with specified amino acids of the binding pocket were presented in Table 3. Obviously, Antimycin A with the top-ranked docking score exhibited the most stable interaction with Complex III through four hydrogen bonds.
Three "star molecules" demonstrated moderate binding abilities to Complex III with their sulfonamide moieties forming hydrogen bonds with GLY20, SER21, ARG86 and ASP214. However, there were no apparent hydrogen bonds existing between the commercial fungicide Amisulbrom and Complex III, corresponding to its lowest docking score.

MD simulations and MM/GBSA calculation of novel geranyl aromatic sulfonamide compounds and complex III co-structures
For the purpose of obtaining dynamics information of the docked ligand interacting with Complex III and validating the crucial amino acids residues in the binding processes of each ligand, five independent 100 ns MD simulations of compounds 4c-1, 4c-2, 4d-2, Amisulbrom and Antimycin A with Complex III co-bound structures were performed in this study. The root-mean-square deviation (RMSD) values of ligand-Complex III were calculated with initial structures selected as reference poses to illustrate the conformational change and evaluate the dynamics stability of each system [30,31]. As shown in Fig. 3A, it was evident that the average RMSD value for Antimycin A-Complex III was lowest compared to the remaining four systems, which was corresponded with the conclusion that Antimycin A had the most stabilized binding affinity with Complex III obtained from the described-above docking results. The RMSD values of 4c-1-Complex III, 4d-2-Complex III, and Amisulbrom-Complex III were finally converged at~3.25, 3 and~3.5 Å after 20, 30 and 40 ns, respectively. In addition, it was worth noting that the RMSD values of 4c-2-Complex III fluctuate most severely and did not reach a good convergence throughout the 100 ns simulation, implying the instable interaction between ligand and receptor, as its slight lower docking score mentioned above. Apart from RMSD, the Radius of Gyration (RG) values of each complex were also calculated, as shown in Fig. 3B, the RG of 4c-2-Complex III system possessed significant fluctuations above and below the average, confirming the suppose that 4c-2 had the most instable binding affinity with Complex III, which were also in good agreement with the RMSD results of the corresponding system.
In order to probe the stable interaction modes between each ligand and Complex III, the averaged conformations of those five systems were extracted from the whole trajectories through clustering analysis, and the binding patterns were presented in Fig. 4. As shown in Fig. 4A, the benzamide moiety of Antimycin A fitted in the polar pocket of Complex III forming a hydrogen bond with ASP214, and a π-π T-shaped interaction with PHE206. The macrolide moiety extended into the hydrophobic area of Complex III with a hydrogen bond formed with HIS187. The predicted interaction mode between Antimycin A and Complex III was roughly consistent with its co-crystallized structure, suggesting interaction models obtained from simulations were reliable in some distance. As shown in Fig. 4B, the triazole ring of Amisulbrom was buried inside the binding site and interacted with ILE5 through a hydrogen bond. In addition, the indole ring of Amisulbrom which was solventoriented formed a π-Sulfur interaction with MET176, which was completely different from the interaction pattern of Antimycin A and Complex III.
As shown in Fig. 4A-C, the newly synthesized geranyl aromatic sulfonamide compounds exhibited similar interaction mechanisms with Complex III, whereas were distinct from the mechanism of sulfonamide commercial fungicides, Amisulbrom. Evidently, compound 4c-1 and 4d-2 bound tightly to Complex III in more stretched conformations, whose aromatic rings extended into a hydrophobic pocket composed of ALA9, SER21, LEU27, LEU80, MET180 etc. inside the binding cavity, and the aliphatic chains towards outside. The sulfonamide moieties of 4c-1 and 4d-2 were stably anchored in the binding cavity by hydrogen bonds. For instance, the sulfonamide moiety of 4d-2 can form one hydrogen bond with GLY20 at a distance of 1.9 Å and there were three hydrogen bonds could be observed between 4c-1 and GLY20 and ARG86, which implied 4c-1 may possess much stronger binding affinity with Complex III than 4d-2. Apart from hydrophobic interactions and hydrogen bonds, the π interactions existing between 4c-1, 4d-2 and TRP17, PHE206, TRY210 also contributed a lot to their binding. Unexpectedly, compound 4c-2 interacted with Complex III in a seemed unstable conformation, with its sulfonamide moiety formed three hydrogen bonds with GLY20, LEU23 and GLY24 near the poplar cavity, however, its thiophene ring and hydrophobic chain were wiggly within the binding site. To sum up, our synthesized geranyl aromatic sulfonamide compounds could bind to the Qi active site in Complex III through forming hydrogen bonds and π-interactions with crucial amino acid residues mentioned above, such as TRP17, GLY20, ARG86 and PHE206 etc.
The binding free energies of Complex III with compounds 4c-1, 4c-2, 4d-2, Amisulbrom and Antimycin A were finally calculated adapting MM/GBSA protocol, and the results were presented in Fig. 5 and Table S3. Apparently, Antimycin A showed the strongest binding ability (the ΔG total value of −51.07 kcal/mol) with Complex III among five compounds, which confirmed the hypothesis proposed above. The total binding free energies of 4c-1 and 4d-2 with complex III (the ΔG total value of −50.60 and −47.18 kcal/mol) were slightly lower than the energy of Amisulbrom (the ΔG total value of −44.22 kcal/mol), whereas the ΔG total of 4c-2 with Complex III was highest among four sulfonamide compounds, which may be resulted from the its unstable binding confrontation. The results proposed above indicated that there were stronger binding affinities existing between the 4c-1, 4d-2 and complex III, when compared with Amisulbrom and 4c-2. Besides, the Surflex Docking score of those five compounds were consistent with their binding free energy qualitatively, explaining both protocols were reasonable for prediction. Consequently, from the perspective of molecular simulation, it was concluded that the novel aromatic geranyl sulfonamide compounds, particularly 4c-1 and 4d-2 can not only interacted with their potential target, complex III in brand new mechanisms, but also exhibited higher binding affinities than sulfonamide commercial fungicide, Amisulbrom, which made them become potential inhibitors of complex III.

Conclusions
Geraniol, as one of the main components of citronella oil, exhibited good fungicidal activities against two plant pathogenic fungi R. solani and G. saubinetii. 24 novel geranyl aromatic sulfonamide derivatives were synthesized based on geraniol as lead and their fungicidal activities were evaluated in this study. Some of the synthesized geranyl aromatic sulfonamide compounds exhibited good fungicidal activities with inhibitory rates greater than 50% against R. solani, G. saubinetii, and S. sclerotiorum at a concentration of 50 mg/L. The preliminary SAR analysis suggested that the introduction of various aromatic heterocycle into  The total binding free energy (ΔG total ) of Complex III with the novel geranyl aromatic sulfonamide compounds, commercial fungicide Amisulbrom and Antimycin A calculated using the MM/ GBSA method geranylamine may have favorable effects on their fungicidal activities. The geranyl thiofuran-sulfonamide compounds 4c-1 and 4c-2 showed excellent antifungal activities against R. solani with an EC 50 value of 24.97 mg/L and against S. sclerotiorum with an EC 50 value of 18.53 mg/L, respectively. Another compound 4d-2, which contained a chlorinesubstituted piperidine, also displayed good fungicidal activities against R. solani with an EC 50 value of 29.98 mg/L and against S. sclerotiorum with an EC 50 value of 29.31 mg/L. The potential molecular mechanism of these geranyl aromatic sulfonamide compounds were proposed to possibly stably bound to Complex III via hydrogen bonds and multiple π-interactions with some vital residues TRP17, GLY20, ARG86 and PHE 206 etc. The binding affinities of compounds 4c-1 and 4d-2 to Complex III were calculated to be much lower than that of sulfonamide commercial fungicide, Amisulbrom, but will be furtherly confirmed through biochemistry and molecular biology studies. Finally, three geranyl aromatic sulfonamide compounds, 4c-1, 4c-2, and 4d-2, were screened as "star molecules" and potential inhibitors of Complex III needed to be noted in a follow-up study because of their excellent antifungal activities and their good predicted binding affinities to the target Complex III.

Instruments and chemicals
Citronella oil was purchased from AFU (Beijing, China) as a mixture without purification. Geraniol, citronellol and citronellal were purchased from MACKLIN (China). Pyraclostrobin was purchased from J&K (China). For all reactions, the solvents and chemical reagents, obtained from China National Medicines Co., Ltd, Beijing, China, were of analytical or synthetic grade and were used without purification. Column chromatography was performed to purify all of the intermediates and final products, using silica gel in a solvent system of petroleum ether/ethyl acetate = 10/1. NMR spectra were obtained using a Bruker Avance DPX300 spectrometer with tetramethylsilane as the internal standard. Mass spectra were obtained with an Agilent 1100 LC-MSD-Trap mass spectrometer equipped with a standard electrospray ionization (ESI) source.

General synthesis procedure for target compounds 4a and 4b-4d
The plant germ-resisting compounds 4a-4d were prepared via a crosslinking reaction with different combinations of intermediate 3 and acyl chloride compounds or sulfonyl chloride compounds. Taking the synthetic method of compound 4a-1 as an example, intermediate 3 (700 mg, 4.57 mmol) and triethylamine (TEA, 920 mg, 9.14 mmol) were dissolved in dichloromethane (DCM, 30 mL), and 2-methylbenzenesulfonyl chloride (1045 mg, 5.48 mmol) was slowly added dropwise (1 mL/min) to the mixture in an ice bath. The mixture was incubated at 25°C for 15 h. Compound 4a-1 was purified using column chromatography (petroleum ether/ethyl acetate = 10/1) to give a yield of 51.4%. All of the other target compounds were prepared using the similar procedures.

General Synthesis Procedure for Target Compounds 5a
Taking the synthetic method of compound 5a-1 as an example. 4a-1 (300 mg, 0.98 mmol) was dissolved in EtOH (10 mL) followed by adding sodium ethoxide (EtONa, 20% v/v, 180 mg, 1.50 mmol) and then the mixture was stirred for 0.5 h at room temperature. Iodomethane (1 mL) was dripped into the mixture and then it was stirred for 3 h at 25°C. After the reaction was completed, the mixture was concentrated under vacuum and was extracted using ethyl acetate and saturated NaCl solution. Finally, the organic phase was dried by Na 2 SO 4 , and then was concentrated to give pure 5a-1 in a yield of 73.4%. All of the other target compounds were prepared using the similar procedures.

Fungicidal activity assay
First of all, the purity of each compound was detected through HPLC before the fungicidal activity assay was performed, and the result implied that the purity of all tested compounds reached 95% or higher. Then, the fungicidal activities of tested compounds against G. saubinetii, R. solani, G. fujikuroi, V. mali, S. sclerotiorum, and M. oryzae were screened in vitro at a concentration of 50 mg/L through performing the previously reported method [32]. The fungi were provided by the College of Plant Protection, China Agricultural University (Beijing, China). The commercial fungicide Pyraclostrobin was selected as the positive control because of its broad-spectrum antifungal property, and the pure DMSO solvent was applied as the blank control. The antifungal rates of tested compounds were calculated using the following formula.
where the D blank represents the average diameter of fungal growth on untreated PDA, D compound represents the average diameter of fungal growth on treated PDA and the D Disk represents the diameter of mycelia disks inoculated in the center of the Petri dishes.
In the precision antifungal test, the median effective concentration (EC 50 ) of compounds with good inhibition activities over 60% were evaluated by three parallel repeats according to previously reported procedure [32]. The values of each EC 50 were determined with a standard procedure using SPSS 15.0 software.

Molecular docking
There are several crystal structures of Complex III stored in Protein Data Bank (PDB, https://www1.rcsb.org/). Among them, the crystal structure (PDB ID: 1PPJ), the only structure of Complex III co-crystallized with QiIs commercial fungicide Antimycin A and possessed almost the highest resolution at 2.1 Å was obtained from PDB as the docking protein [22]. The protein structure was prepared, such as to add hydrogens and repair sidechains in Structure Preparation Tool integrated in commercial SYBYL 7.3 software package. For the reason that both Antimycin A and sulfonamides are identified as QiIs fungicides, the biding site, so-called "Protomol" in SYBYL was generated based on the interaction between Antimycin A and Complex III for the following docking process [22]. The 3D structures of five ligands, including the commercial sulfonamide fungicide Amisulbrom, newly synthesized aromatic sulfonamide molecules 4d-1, 4d-2, 4e-2, and the native ligand of 1PPJ: Antimycin A were constructed in Sketch mode and optimized using Tripos force field and the Gasteiger-Huckel charge. Then, these optimized ligands were docked into the generated "Protomol" of the receptor utilizing Surflex-Dock mode with all parameters set as default in the SYBYL 7.3 software. Surflex-Dock score was obtained using Total Score to assess the binding affinities of each ligands to the target protein, and the higher values of score corresponding to stronger binding affinities, through which the complexes between four sulfonamide compounds and complex III protein were finally generated [33].

MD simulation and binding free energy calculation
MD simulations starting from five complexes generated through docking strategy were performed with AMBER 20 package. The AMBER ff99SB force field [34] and the GAFF force field [35] were exerted to receptor and ligands respectively. All of the five complexes were solvated in a 12 Å cubic box [36] with TIP3P explicit water models and the total number of net charges were balanced to neutral by adding Na + cations. The steepest descent method and the conjugate gradient method were applied to minimize the energy of the complexes. Then the temperature of the complexes was heated slowly from 0 to 300 K, followed by a 100 ns equilibration of the whole system in the NPT ensemble. After the simulations were completed, CPPTRAJ module was imported to calculate RMSD value and to perform cluster analysis to evaluate the stabilities of each system and to obtain the stable interaction confrontations between ligands and receptors. The 3D and 2D interaction mode graphics were drawn by PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC. [37], and Discovery Studio 2017 [38], respectively.

MM/GBSA
MM-GBSA is an effective method to calculate binding energy, which uses energy properties of free ligand, free receptor and receptor-ligand complex for binding affinity calculation. For each complex in our study, the generated trajectory of the last 2 ns was used to calculate the binding free energy between receptor and ligand by using the MMPBSA.py script in AMBER 20 software. The binding free energy was calculated utilizing the following formulas. In formula (1), the value of ΔS bind was often neglected for expensive computing costs, as a consequence, the formula (2) was finally adapted.
ΔG bind % ΔH bind ð2Þ For each component in formula (3), the value of E was calculated using the following formulas: where the E gas is the total gas-phase free energy, the E solv is the solvation free energy, the E ele is the electrostatic energy, the E vdw is the van der Waals energy, the E gb is the polar solvation energy, and the E surf is the nonpolar solvation energy.