Carboxylic acid derivatives suppress the growth of Aspergillus flavus through the inhibition of fungal alpha-amylase

Abstract Aspergillus favus (A. flavus) is a saprophytic fungus and a pathogen affecting several important foods and crops, including maize. A. flavus produces a toxic secondary metabolite called aflatoxin. Alpha-amylase (α-amylase), a hydrolytic enzyme produced by A. Flavus helps in the production of aflatoxin by hydrolysing the starch molecules in to simple sugars such as glucose and maltose. These simple sugars induce the production of aflatoxin. Inhibition of α-amylase has been proven as a potential way to reduce the production of aflatoxin. In the present study, we investigated the effect of selected carboxylic acid derivatives such as cinnamic acid (CA), 2, 4-dichlorophenoxyacetic acid (2,4-D), and 3-(4-hydroxyphenyl)-propionic acid (3,4-HPPA) on the fungal growth and for the α-amylase inhibitory activity. The binding potentials of these compounds with α-amylase have been confirmed by enzyme kinetics and isothermal titration calorimetry. Molecular docking and MD simulation studies were also performed to deduce the atomic level interaction between the protein and selected ligands. The results indicated that CA, 2,4-D and 3,4-HPPA can inhibit the fungal growth which could be partly due to the inhibition on fungal α-amylase activity. Communicated by Ramaswamy H. Sarma


Introduction
A wide variety of filamentous fungi produce mycotoxins that are toxic secondary metabolites.The most important mycotoxins are produced by species of the genera Aspergillus, Fusarium, Alternaria, and Penicillium (Pitt et al., 2000).Worldwide, crops are contaminated by toxic molds and their mycotoxins (Bennett and Klich, 2003).
One of the mycotoxins, aflatoxins are highly toxic and hepatocarcinogenic compound, produced by many fungi, including A.flavus (Hedayati et al., 2007).Contamination with aflatoxin continues to be a serious problem for agricultural crops, food, and feed around the world.Simple sugars such as glucose, maltose, and maltotriose induce the formation of aflatoxin (Woloshuk et al., 1997).These simple sugars are formed by cleaving the a-1,4-linkages in starch by a-amylase, a hydrolytic enzyme that further stimulates the production of aflatoxins and supports the fungal growth (Franco et al., 2000).Inhibition of a-amylase is considered to be an effective way to limit the aflatoxin production and fungal growth.A lectin-like protein from Lablab purpureus has the ability to inhibit a-amylase activity and regulate the growth of A flavus (Fakhoury and Woloshuk, 2001).Chen et al. (1999) have reported that a corn trypsin inhibitor with antifungal activity inhibit a-amylase of A. flavus.Small molecule inhibitors of fungal a-amylase such as berberine, 6-gingerol, IAA & IBA were also reported in our previous studies (Dileep et al., 2012;Tintu, Dileep, Augustine, et al., 2012;Tintu, Dileep, Remya, et al. 2012).
The CA (Figure 1A) is one of the most abundant phenolic acids in fruits and vegetables.Antioxidant (Sova, 2012), antiinflammatory (de C� assia da Silveira et al., 2014), and anti-cancer (Anantharaju et al., 2016), antifungal (Kim et al., 2004), and antidiabetic properties (Alam et al., 2016) of CA has been reported previously.Similarly, (2,4-D) (Figure 1A) is one of the most extensively used phenoxy acid herbicides in agriculture.Its auxin-like activity alters normal protein synthesis and cell division in plant meristems and leaves ([Tayeb et al., 2010).The 3,4-HPPA (Figure 1A) is an antioxidant.The antimicrobial activity of 3,4-HPAA against Salmonella enterica subsp.has been reported previously (Narayana et al., 2007).In the present study carboxylic acid derivatives such as CA, 2,4-D and 3,4-HPPA were investigated for their ability to inhibit the fungal growth in the in vitro conditions.a-amylase inhibitory activity of these compounds was also investigated through molecular modelling and biophysical studies.

Fungal growth inhibition
A. flavus strain was obtained from the culture collection of the Institute of Microbial Technology (MTCC), Chandigarh, India.The organism was grown and sub-cultured repeatedly in Czapek Yeast Extract Agar (CYA).Inhibition of the fungal mycelial growth was determined on potato dextrose agar medium (PDA) using the well diffusion method by measuring CONTACT K.V. Dileep dileepkvijayan@gmail.com, dileepvijayan@jmmc.ac.inSupplemental data for this article can be accessed online at https://doi.org/10.1080/07391102.2023.2214235.
the radius of the inhibition zone in the presence of different compounds such as CA, 2,4-D, and 3,4-HPPA at 100 mM concentration (prepared in 0.1% of DMSO).A 4-mm well is made at the centre of PDA plate and streaked with the 10-day-old fungal culture using a loop and incubated at 25-28 � C for 2-4 days.The radial mycelial growth was measured.

Purification of a-amylase
Fifty milliliters of the starch medium [composition of the medium is 1.4 g of KH 2 PO 4 , 10 g of NH 4 NO 3 , 0.5 g of KCL, 0.1 g of MgSO 4 7H 2 O, 0.01 g of FeSO 4 7H 2 O, and 20 g of soluble starch dissolved in 1000 mL of distilled water and adjust pH to 6.5] in a 250 mL of cotton plugged sterilized conical flask was inoculated with the 1 mL spore suspension of fungi A. flavus.The flask was aerated using a rotary shaker at 200 rpm for 72 h.Since a-amylase is an extracellular protein, the protein was extracted from the medium.The medium was then dialyzed against distilled water to remove the residual sugars.The enzyme was isolated by the salting out process with ammonium sulphate.The precipitant of 30% ammonium sulphate has been discarded initially and the precipitate of 85% saturation of ammonium sulphate was collected and dialyzed for 20 h against phosphate buffer (pH 6.9).Further, the enzyme was concentrated upto 10 mg/mL solution and was used for the enzyme assay.The activity of a-amylase was assayed by a method of Bernfeld [30] with a slight modification.

Enzyme inhibition and kinetic studies
To evaluate the activity of the selected compounds, enzyme kinetic studies have been performed with the purified a-amylase.Almost 1 mL of enzyme solution (0.01 mM) was prepared and added to 1 mL of the substrate (prepared in 0.02 M potassium phosphate buffer pH 6.9) at different concentrations (0.50, 0.75, 1.00,1.25,1.50, and 1.75 mg/mL) and incubated for 3 min at 25 � C. Two mL of reagent was added to each test tube to stop the enzyme activity.Tubes were heated in a boiling water bath for 15 min to monitor the color change and then cooled with running tap water.Samples were made of up to 10 mL of water.Absorbance at 540 nm was read on a UV-Vis Spectrophotometer.The blank was prepared by substituting the amylase solution with distilled water.The whole experiment was repeated in the presence of selected compounds such as CA, 2,4-D and 3,4-HPPA at concentrations 0.2 mM.To ascertain the type of inhibition, Line weaver-Burk plots were drawn for the native and inhibited enzyme.The Michaelis constant (K m ) and maximal velocity (V max ) were determined from the Lineweaver-Burk plot.The inhibition constant, K i was calculated using the Equation [31].
From the K i value obtained, the IC 50 was calculated using Cheng-Prusoff Equation [32]

Isothermal titration calorimetric assay
To identify the binding constant of ligands (CA; 2,4-D; 3,4-HPPA) with a amylase, isothermal titration calorimetric (ITC) analysis was performed using Microcal ITC.The concentration of protein and ligands were 0.01 mM and 0.2 mM respectively.Both protein and ligands were prepared in DMSO (0.1%) water mixture.1.8 mL of the purified a-amylase solution was taken in the sample cell, and a total of 300 mL of ligand solution was injected in steps and stirred using the same syringe at 307 rpm.All solutions were degassed and loaded into the cells without any bubble formation.The volume of the 1 st injection was set 2 mL to avoid inaccuracy and the volume of the followed injections is 10 mL each.Total of 29 injections were made.Between two consecutive injections, a constant time interval of 2 min was maintained to stabilize the base line.

Molecular modelling and docking
The atomic level of interaction of CA, 2,4-D and 3,4-HPPA were determined by induced fit docking (IFD) study.IFD was performed using Schr€ odinger 9.1.(Schr€ odinger, LLC, New York, USA).The crystal structure fungal a-amylase (Source A. niger) in complex with acarbose (PDB ID 7TAA) was downloaded from the PDB and taken as the target for the docking studies.Prior to the protein preparation the water molecules bound in the crystal structure were deleted.Polar hydrogens were added and protein structure was optimized.Energy minimization was also done by applying a cut of 0.30 Å using a force field OPLS 2001.A Grid was set with a box dimension of 10 � 10 � 10 Å from the centre of the crystallographic ligand.In IFD protocol, the ligands were flexibly aligned to the grid of the protein using a Glide program.The top 20 poses were selected based on the energy and the plasticity was applied to the protein residues using Prime module.The 20 new target conformations were taken and docking was again performed using Glide XP (extra -precision) method.Finally, the best pose was selected based on the Glide Score.

Molecular dynamic simulation
Further to understand the ligand stability within the active site, we have performed molecular dynamics (MD) studies.
The best scored pose from the docking was selected for the study.All MD simulations were performed by GPU accelerated Desmond software.Prior to MD simulations, the system was set up using the 'System Builder' in Desmond.The protein-ligand complex was immersed into an orthorhombic box with dimensions of 10 � 10 � 10 Å containing TIP3P explicit solvent model.The systems were further neutralized by adding Na þ or Cl -ions.In addition to that, 0.15 M of NaCl was also added to the system to provide a physiological environment.Using the OPLS-2005 force field, the MD simulation was performed for 500 ns with a recording interval of 500 ps.
The NPT ensemble was employed with a temperature fixed at 300 K and pressure at 1.01 bar.The thermostat and barostat methods used in the studies were 'Nose-Hoover chain' and 'Martyna-Tobias-Klein' respectively.RESPA was used for the integration and the integration time step was set at 2fs.Default settings were used for all other parameters.

Selected carboxylic acid derivatives inhibits the growth of A. flavus
It was found that at 100 mM concentration of the carboxylic acids such as CA, 2,4-D and 3,4-HPPA produced inhibition around the well.We first marked the zone of inhibition of three consecutive experiments of each carboxylic acid derivatives (Figure 1 B-E and S1) and calculated the area under the zone of inhibition.Further, we compared each other to understand the effect of these selected carboxylic acids derivatives.It was found that the area is directly proportional to the inhibitory potency of the compound.According to the results, CA has the highest and 3,4-HPPA has lowest activities on the growth of A. flavus.The inhibitory activity of 0.1% DMSO was also determined and found no activity on the fungal growth.

In vitro and biophysical studies confirmed the a-amylase inhibitory activity of three carboxylic acids derivatives
To check whether the ligands can inhibit fungal a-amylase or not, we purified and concentrated the enzyme up to 10 mg/mL and subjected to inhibition assay.The activity of purified enzyme was confirmed by performing normal enzyme kinetic reactions.It has been found that all the selected ligands have competitive mode of binding towards the a-amylase (Figure S2).While comparing the IC50 values, of three compounds, it has been found that the 2,4-D has slightly improved IC50 value (0.127 ± 0.29 mM) when compared to the CA (0.164 ± 0.19 mM) and 3,4-HPAA (0.151 ± 0.12 mM) (Figure 1F and S2).The other parameters such as K' m , K m , V max and IC 50 were calculated from Lineweaver burk plot is shown in Table 1.
We also performed ITC for all the ligands against a-amylase to find out the binding affinity and binding free energy.The ITC experiment suggests that all ligands binding to a-amylase with a stoichiometric value approximately equal to 1 (Table 2).The competitive mode of inhibition and stoichiometric value explicitly explains the active site binding of three carboxylic acid derivatives.The thermodynamics parameters of the ligands while binding to a-amylase are shown in Table 2.

In silico studies demonstrated the binding of selected carboxylic acid derivatives
Molecular modeling and docking were carried out, in order to get the atomic level details of the interaction between the selected carboxylic acid derivatives and a-amylase.The a-amylase of A. oryzae (PDB ID: 7TAA) was taken as template for modelling since it shares high structural similarities.Tintu et al. reported that the active site consists of a number of charged and polar amino acids such as Y82, D117, H122, Y155, R204, D206, T207, K209, H210, E230, D233, H296, D297, and R344 (Tintu, Dileep, Augustine, et al., 2012).Docking studies have given a number of poses from which the most favourable pose was selected based on the free energy of binding and number of hydrogen bonds.The binding energy obtained for 2, 4-D is À 31.54 kcal/mol which is higher than CA (À 21.54 kcal/mol) and (3, 4-HPP À 11 kcal/mol).In all the ligands except 3,4-HPPA, the carboxylic groups are interacting with K209, H210 & L232 through hydrogen bonds.On the other hand, for 3, 4-HPPA, there is only one hydrogen bond formed between the carboxylic group and L232.Furthermore, all ligands were observed to form p-p-stacking interactions with H210 and also exhibits hydrophobic interactions with residues L166, Y155, H210, K209, V213, and L232 (Figure 1 H-J).
MD simulation study was performed to assess the residence of time of the selected ligands on the protein for a period of time (i.e.500 ns).The root means square deviation (RMSD) with respect to the initial structure was calculated for all the ligands during the simulations (Figure 1K-M).
It has been observed that 2,4-D and CA has retained majority of the interactions (that produced in the docking studies) in the simulations.However, the p-p-stacking interaction with H210 was observed solely in the case of 2,4-D, whereas the interaction with L232 was found to be retained for only 67% of the simulation time in the case of 3,4-HPP.Nevertheless, 3,4-HPP formed a stable binding at the active site of a-amylase by creating a water-mediated interaction and a hydrogen bond with Y155 and K209 through its hydroxyl group and carboxyl moiety, respectively.These interactions were sustained for more than 50% of the simulation time, thus contributing to the stability of the protein-ligand complex.This stability was also confirmed by the RMSD analysis, as the RMSD values for all the ligands with respect to the protein were below 0.5 Å, which is likely due to the stable interactions formed between the ligands and the protein throughout the simulation period.The mean RMSD values for the ligands with respect to their initial position were as follows: 2,4-D (0.15 ± 0.03 Å), CA (0.14 ± 0.04 Å), and 3,4-HPP (0.40 ± 0.15 Å).Therefore, the simulations suggest that all the ligands exhibit low entropy during the simulations.We additionally calculated the binding free energies (MM-GBSA) by utilizing the intermediate snap shots of 500 ns simulation data.The order of the ligands according to the binding energy is as follows, 2,4-D (À 31.42 ± 0.90 kcal/mol), CA (À 23.62 ± 1.70 kcal/mol) and 3, 4-HPP (À 14.35 ± 0.17 kcal/mol).It has been found that the binding energies obtained from the simulations studies and docking studies were perfectly correlated.

Conclusion
The current study investigated the inhibition profiles and binding potentials of three carboxylic acid derivatives such as CA, 2,4-D and 3,4-HPP.Action of these compounds on fungal mycelial growth suggested that CA has better inhibition profiles when compared to other ligands.At the same time, the enzyme inhibition results shows that 2,4-D has slightly better inhibitory activity than other molecules.The enzyme kinetic studies also indicated that all of these compounds competitively bind to the active site and block the enzyme activity.The binding of these compounds with fungal a-amylase is further confirmed by ITC experiment and obtained a stoichiometric value approximately equal to 1 confirming a one site binding.Molecular docking and MD simulation analysis suggested that all the selected compounds exhibited similar pattern of binding.These studies also reveal the better binding affinity, stability, and structural conformation at the binding site of three compounds against the a-amylase receptor.Based on the results of biophysical and molecular modelling studies we conclude that the inhibition of A. flavus is partly due to the inhibition of a-amylase.Moreover, it also revealed that, the compound can serve as a potential lead compound for the development of an effective fungal a-amylase inhibitors, which can be used against A. flavus.Hence, a-amylase could be used as a target to develop novel anti-fungal therapies.

Figure 1 .
Figure 1.Structures of selected carboxylic acid derivatives and its action against A. flavus and a-amylase.(A) 2D diagrams of CA, 2,4-D and 3,4-HPPA.(B-D) Zone of inhibition of CA, 2,4-D and 3,4-HPPA against A. flavus.(E) Area of the zone of inhibition.(F) IC50 values of CA, 2,4-D and 3,4-HPPA in mM (G) The RMSD of the ligands deduced from the 500 ns simulations.Mode of binding of (H) CA (cyan), (I) 2,4-D (yellow) and (J) 3,4-HPPA (pink) with fungal a-amylase (green).Both protein and ligands are shown in ball and stick models.The hydrogen bonds were shown in black dashed lines.(K-M) The RMSD (for a period of 500 ns) of a amylase in complex with CA, 2,4-D and 3,4-HPPA respectively.

Table 2 .
Thermodynamic parameters calculated for the binding of CA, 2,4-D an 3,4-HPPA from ITC analysis.