Molecular Dynamics Simulation for Food Products and Processes

Version	Summary	Created by	Modification	Content Size	Created at	Operation
1		Xin Dong	+ 3260 word(s)	3260	2022-01-10 07:44:40	\|
2	update layout and reference	Rita Xu	Meta information modification	3260	2022-01-20 02:57:54	\|

This entry is adapted from the peer-reviewed paper 10.3390/pr10010119

Molecular dynamics (MD) simulation is a particularly useful technique in food processing. Normally, food processing techniques can be optimized to favor the creation of higher-quality, safer, more functional, and more nutritionally valuable food products. Modeling food processes through the application of MD simulations, namely, the Groningen Machine for Chemical Simulations (GROMACS) software package, is helpful in achieving a better understanding of the structural changes occurring at the molecular level to the biomolecules present in food products during processing.

molecular dynamics simulation GROMACS protein thermal treatment static electric field

1. Introduction

The motivation behind and technology applied in food processing have dramatically changed over humanity’s history, often based on the socio-economic context of the time ^[1]. Currently, modern processing techniques are most commonly applied to improve the safety, shelf life, convenience, and nutritional, functional, and organoleptic qualities of processed food products ^[2]. Food processes no longer only involve treatment by conventional thermal, chemical, or mechanical means but may equally include novel treatment methods utilizing high hydrostatic pressure, radiation, microwaves, ultrasound, pulsed electric fields, or cold plasma ^[3]. Understanding the consequences that the application of these external stressors has on the biomolecules present in food products, namely, the carbohydrates, proteins, and lipids, is critical in determining optimal processing conditions. This is where food process engineering is involved, as its primary goal is to better understand the process in order to better manipulate it ^[4]. By better controlling food processes, a more desirable finished product that better meets the population’s needs can be offered. Although this is easily said, this, in reality, is no simple task, seeing as food processes are often dynamic and involve highly complex mechanisms ^[4]. Simulation modeling is an important tool that helps to bridge this gap, as it actively combines knowledge of physics, statistics, applied mathematics, and computing to predict the result of a given processing treatment with sufficient accuracy for the application ^[4]^[5]. The development of an accurate mathematical model that simulates the real-life mechanism with minimal data noise can eliminate the need for repetitive, isolated experimental trials ^[5]. It can equally allow for continuous improvement and learning, while also serving to improve experimental design efficiency, thus limiting costs and time invested ^[5].

One such computational methodology is the use of atomistic molecular dynamics (MD) simulation. Atomistic MD simulation is, at its core, based on Newton’s equation of motion ^[6]. Each atom, part of the molecular system being modeled, is assigned a position vector defined as a function of time ^[6]. This position vector then fluctuates based on the stresses acting on the atom through time ^[6]. The stresses that can be applied to each atom include any relevant forces, originating from “inter-atomic interactions, [treatment] energy and interactions between the molecular system and [its] environment”, namely, the presence of water or salt molecules ^[6]. Near the end of the 1950s, the concept of molecular dynamics was initially proposed by two physicists, Berni Alder and Tom Wainwright, who were studying statistical mechanics at Lawrence Radiation Laboratory in Livermore, California ^[7]. Molecular dynamics was a substitute for the earlier Monte Carlo simulation ^[7]. In a time where few physicists and chemists trusted computer simulations over traditional theoretical methods, Alder and Wainwright were able to successfully model, using both molecular dynamics simulation and Monte Carlo, the interactions and consequent phase transition in a hard-sphere generic molecular system ^[7]. Later, between 1960 and 1970, other researchers applied MD simulation to model increasingly realistic material systems, including solid copper and liquid argon ^[8]. However, in September 1974, chemist Frank Stillinger and physicist Aneesur Rahman refined this initial model to simulate the molecular interactions of liquid water molecules at four different temperatures ^[9]. McCammon et al. ^[10] further expanded the applications of MD simulation, running a 9.2-picosecond simulation of the dynamics of a small, folded globular protein, a bovine pancreatic trypsin inhibitor held in vacuum. The same study was reworked eleven years after with the bovine trypsin inhibitor now solvated in water, and an improved simulation length of 210 picoseconds was achieved ^[11].

Since then, with the increasing availability of computing power, simulation lengths of MD simulations have increased dramatically to between 10 and 300 nanoseconds for increasingly long polypeptide chains ^[11]^[12]. Currently, polished atomistic MD simulations are being applied in a variety of domains for the modeling of diverse biological molecules, namely, DNA, RNA, and proteins ^[6]. Atomistic MD simulations have been routinely applied in the pharmaceutical industry for drug discovery and drug design ^[13]. However, in recent years, they have been increasingly applied in food process engineering to provide greater insight into the molecular interactions and consequent conformational and functional property changes that take place during the processing of food products ^[2]. The effect of food processes on the structure and function of food carbohydrates, lipids, and other small molecules has been investigated using MD simulations ^[14]. However, recent emphasis has been placed on MD simulations of food proteins, due to their unique properties, including their diverse conformational states, ability to denature and unfold, ability to interact among themselves and with other biomolecules, their important influence on the functional properties of the foods they are present in, and their capacity to function as an enzyme ^[14]. Proteins not only play a vital nutritional role in the body following their consumption, but they equally confer certain dynamic functional properties to the food products themselves prior to consumption ^[15]. Protein structure, which dictates the protein’s function, may influence viscosity, texture, emulsification, gelation, and water absorption and retention. It has equally been suggested that changes in proteins’ conformational state, notably their surface area, limits access to allergenic epitopes, which, when in contact with a T cell or immunoglobulin E antibody, cause allergic reactions by the immune system ^[16].

Studies targeting food proteins that have already utilized atomistic MD simulations include Ara h 6 in peanuts ^[17], trypsin inhibitor in soybean ^[18], bovine β-lactoglobulin in cow’s milk ^[16], profilin in hazelnuts ^[12], Act d 2 in kiwifruit ^[19], avidin in egg whites ^[20], and myofibrillar proteins in pale, soft, exudative chicken breast meat ^[21]. MD simulations have equally been employed to simulate the effects of diverse food processes, including the application of thermal and electric fields, both static and oscillating ^[17].

2. Application of Molecular Dynamics Simulation for Food Products and Processes

As previously described, proteins undergoing MD simulation are often subject to a number of different distortional forces. They are not only agitated by external factors, including surrounding solvent particles, ions, and the chosen treatment techniques, but also by intramolecular forces and forces due to the environment at both the edge and beyond the simulation box. External food processes imposed during the thermal, electrical, or chemical treatment of food products can effectively modify the higher levels of protein structure without redesigning the primary amino acid sequence ^[2]. The overall stability of a protein undergoing a stress treatment is often measured by its inherent ability to resist structural deviations. It has been observed that the presence of hydrophobic amino acids, as well as the formation of salt bridges or disulfide bonds, greatly increases the functional stability of proteins undergoing treatment ^[2]. Conversely, a high concentration of water molecules, surrounding the target protein, is often linked to decreased stability, as the water molecules penetrate within the protein’s structure, aiding in the destruction of the higher-order protein structure ^[2]. The application of MD simulation technology to model food processes has been popularized in recent years, which has allowed us to gain new insight into how pH, solvent composition and ionic strength, temperature, and exposure to electric fields can affect the denaturation process of the various proteins in diverse food products ^[2]. Many recent studies have investigated the effects of food processes, namely, treatment using heat, static, or oscillating electric fields or a combination of these, on different food products using GROMACS (Table 1). These molecular modeling studies often allow the researchers to make important associations between the proposed process parameters and potential improvements in the bioavailability, functionality, or allergenic potential of the protein. These conclusions drawn from computer-based MD simulations successfully connect the changes observed in the protein’s molecular structure to the food product’s final macroscopic functions. This new and deeper understanding of the key relationship between structure and function is significant in supporting efforts to improve overall food safety and quality.

Table 1. Molecular dynamics studies using GROMACS applied to diverse food proteins.

Food Protein Studied	Food Process Studied	Simulation Lengths	Reference
Ara h 6 peanut protein	9 simulations: 27 °C, 107 °C, and 151 °C with no electric field, with a static electric field (0.05 V/nm), or with an oscillating electric field (0.05 V/nm, 2.45 GHz).	1 ns	^[17]
Soybean trypsin inhibitor	8 simulations: 27 °C, 70 °C, 100 °C, and 121 °C with no electric field or with an oscillating electric field (0.5 V/nm, 2.45 GHz).	5 ns	^[18]
β-lactoglobulin in cow’s milk	6 simulations: 60 °C, 75 °C, and 90 °C with no electric field or with an oscillating electric field (0.5 V/nm, 2.45 GHz).	2 ns	^[16]
Act d 2 in kiwifruit	8 simulations: 27 °C, 52 °C, 77 °C, and 102 °C with no electric field or with an oscillating electric field (0.05 V/nm, 2.45 GHz).	2 ns	^[19]
Cora a 2 profilin in hazelnuts	5 simulations: 27 °C, 77 °C, 127 °C, 177 °C, and 227 °C.	200 ns	^[12]
Avidin in egg whites	8 simulations: 27 °C, 60 °C, 70 °C, and 80 °C with no electric field or with an oscillating electric field (0.05 V/nm, 2.45 GHz).	2 ns	^[20]
Myofibrillar proteins from pale, soft, exudative chicken breast meat	4 simulations: 0 V/nm, 0.0008 V/nm, 0.0018 V/nm, or 0.0028 V/nm pulsed electric field.	30 ns	^[21]

Vanga et al. ^[17] authored a study in which the Ara h 6 peanut protein was modeled using GROMACS MD simulation. In this study, the Ara h 6 peanut protein, composed of 127 individual amino acids, was subject to the CHARMM27 force field. The protein molecule was enclosed in a cubic box with sides 10.215 nm long, solvated using the water model TIP3P, and neutralized using three sodium atoms ^[17]. This simulation setup was used to run nine separate MD simulations, in which the Ara h 6 protein was subject to thermal treatments at 27 °C, 107 °C, and 151 °C, with either no electric field imposed, a static electric field at an intensity of 0.05 V/nm, or an oscillating electric field at an intensity of 0.05 V/nm and a frequency of 2450 MHz ^[17]. The results of the STRIDE and RMSD analysis showed that important deviations in the secondary structure of the Ara h 6 protein occurred under both the static and oscillating electric fields as the treatment temperature rose to 151 °C ^[17]. Most interestingly, analysis of the radius of gyration of the modeled protein suggested that these modifications in the protein’s higher order structure led to the compaction of the protein ^[17]. SASA calculations confirmed this result, as the accessible surface area for intermolecular interaction was diminished ^[17]. It has been suggested that this observed compaction of the Ara h 6 protein imparts its superior functional properties, including “better thermal stability, protein solubility, emulsifying and foam properties” ^[22]. Zhu et al. ^[20], in their later MD modeling study of the avidin protein in egg whites, equally determined that the undoing of hydrogen bonds in the protein structure due to the imposed external stresses would affect the functional properties of the product. However, the researchers were not able to specify the nature of the functional properties that would be affected but suggested this would be a topic of future research ^[20]. Dong, Tian et al. ^[21], in their molecular modeling study of myofibrillar proteins in pale, soft, and exudative (PSE-like) chicken breast meat, observed that under the stress of a pulsed electric field, hydrogen bonds within the myofibrillar proteins increased and the radius of gyration decreased. These conformational changes, observed through 30-nanosecond MD simulations and combined with experimental knowledge of observed macroscale changes, allowed the researchers to conclude that pulsed electric field treatment could be useful in enhancing the gelation properties of myofibrillar proteins in PSE-like chicken breast meat ^[21].

Vagadia et al. ^[18] performed a similar modeling study on the Kunitz-type trypsin inhibitor protein in soybeans, a uniquely stable protein. It is critical to understand how processing techniques can efficiently denature the trypsin inhibitor in soybean, as trypsin inhibitors are a known antinutritional factor that inhibits the function of human pancreatic enzymes trypsin and chymotrypsin, which are crucial for the absorption of proteins from the diet ^[23]. Furthermore, this same negative effect of ingested nondenatured trypsin inhibitors also applies to livestock animals and therefore is equally relevant in the processing of animal feed. Similar setup parameters were specified for this simulation as in the study performed by Vanga et al. ^[18], as the CHARMM27 force field and TIP3P water model were also used. However, to accommodate the larger protein of 181 amino acids, the cubic simulation cell was built instead with longer 11.005 nm sides and the system was then neutralized using seven sodium atoms ^[18]. The results of this particular study helped to confirm an initial hypothesis regarding the stabilizing effect of some secondary structures associated with the trypsin inhibitor protein. The trypsin inhibitor protein performed as expected, demonstrating its inherent stability as it produced few structural deviations during its MD simulations ^[18]. The presence of β-sheets, lack of α-helices, and presence of two disulfide bonds provide a unique stability to the trypsin inhibitor protein that allow it to better resist denaturation ^[18]. However, the MD simulations in this study could only be performed over a period lasting 5 nanoseconds, and therefore denaturation of the soybean trypsin inhibitor may have occurred later in the simulations if they could have been performed over a longer duration ^[18]. This highlights a rather important weakness of current MD simulations, as with limited computational capacity, they can only be run over short time periods that are not necessarily comparable to the time period over which a unit operation may be performed in a food processing plant.

Elsewhere, Saxena et al. ^[16] applied thermal and oscillating electric fields to the common allergen β-lactoglobulin protein present in cow’s milk. In this study, the AMBER99SB-ILDN force field was utilized, and, considering the net negative charge (−14e) of the β-lactoglobulin protein, it was neutralized with fourteen positively charged sodium ions ^[16]. This particular modeling study identified the important effect the treatments had on epitopes that lead to IgE-mediated allergic reactions by the immune system. Epitopes on β-lactoglobulin are specific regions on the protein that readily interact with immunoglobulin E antibodies present in the body, thus causing allergic reactions by the immune system ^[16]. At lower thermal treatment temperatures, Saxena et al. ^[16] observed that the allergenic epitopes began gradually losing their rigidity and opened up new sites for the binding of immunoglobulin E antibodies, which may further increase the risk of an allergic reaction. However, the opposite effect was observed when β-lactoglobulin was treated at higher temperatures, as the allergenic epitopes became increasingly rigid and moved into the hydrophobic core of the protein where they could no longer be accessed by the immunoglobulin E antibodies ^[16]. Saxena et al. ^[16] linked this augmented rigidity and consequent restricted access to allergenic epitopes to a potential reduction in protein digestibility due to restricted access for digestive enzymes, and to a reduction in the risk of allergic reaction due to restricted access for immunoglobulin E antibodies.

The effect of combined thermal and oscillating electric field treatment on allergenicity was also observed by Wang et al. ^[19], who investigated the Act d 2 allergenic protein in kiwifruit using GROMACS MD simulations. The model was generated using the CHARMM27 force field and TIP3P water model in a cubic simulation box of length 7.799 nm ^[19]. The results of this study effectively demonstrated the remarkable thermal stability of the Act d 2 protein. There were no significant changes in the secondary structures of the Act d 2 protein, even under the highest thermal treatment of 102 °C ^[19]. Furthermore, these minute structural changes, observed during the MD simulations of only the thermal treatment, were proven to have no significant effect on modifying the Act d 2 protein’s allergenicity, as determined by experimental ELISA tests ^[19]. However, when the thermal treatment was applied in conjunction with an oscillating electric field, turns and α-helices in Act d 2’s structure were undone ^[19]. The oscillating electric field contributes significantly to the unfolding of the secondary structures of the Act d 2 protein, as the protein actively deforms in an attempt to align itself in the direction of the electric field ^[17]. The ability of antibodies to bind to the immunoglobulin E epitopes of Act d 2 to cause an allergic reaction was diminished by a remarkable 75.2% under the combined treatment at 102 °C and using the oscillating electric field ^[19]. Similar to the Act d 2 protein in kiwifruit and trypsin inhibitor in soybeans, Barazorda-Ccahuana et al. ^[12] equally reported the impressive thermal stability of the Cor a 2 profilin in hazelnuts due to the presence of β-sheets and salt bridges, in an MD modeling study. In this study, they applied the OPLS-AA force field and TIP4P water model in GROMACS ^[12]. They found that Cor a 2 did not show complete denaturation of the β-sheets, even when treated at temperatures as high as 227 °C, and only demonstrated partial, not complete, loss of allergenicity ^[12].

Other food protein modeling studies have also been performed, researching the interactions between food proteins and smaller molecules in their environment. β-lactoglobulin is an important bovine whey protein, present in cow’s milk and a common by-product of cheese production, that is widely available and inexpensive ^[24]. Huang et al. ^[25] performed MD simulations of 150 ns using the GROMACS software package, GROMOS54a7 force field, and SPC water model to gain insight into how β-lactoglobulin would deform under high-pressure treatment (600 MPa) when bound at one of two sites to a small molecule, namely, epigallocatechin gallate (EGCG), a component of tea polyphenols with important benefits to human health. Huang et al. ^[25] noted in their concluding remarks the importance of the results of their study in “further improving the quality of tea milk beverage and the application of high-pressure technology in milk beverage”. Sahihi et al. ^[26] previously explored, using MD simulations, a similar interaction between β-lactoglobulin and natural polyphenolic compounds, including quercetin, quercitrin, and rutin, believed to capture harmful reactive oxygen and nitrogen in the body. They utilized the GROMACS software package, the GROMOS96 43a1 force field, and the SPC water model, over a simulation length of 10 ns ^[26]. Abdollahi et al. ^[24] also investigated the interaction between β-lactoglobulin and a major beneficial phenolic acid present in plants, ferulic acid, using MD simulation, at a neutral pH (7.3) and an acidic pH (2.4). Their MD model used the GROMACS software package with the CHARMM27 force field and TIP3P water model, over a simulation length of at least 50 ns ^[24]. Abdollahi et al. ^[24] expressed the relevance of their MD results to the food processing industry, as they could help potentially “enhance associative interactions […] of bioactives”, and in the development of novel and more nutritious consumer products.

As has been seen, the potential applications of molecular dynamics simulation studies in the food processing industry are vast. Performing MD simulations using the GROMACS software package provides deeper insight into the more minute molecular changes that occur within the food products being treated that cannot be observed macroscopically. MD simulations, as applied to food products and processes, are a unique, inexpensive, and user-friendly tool that researchers can adopt to help guide them in the development of their initial hypotheses and their methodologies used to perform their practical experiments. Moreover, MD simulations can be a particularly useful tool in improving the management of available resources for experimentation and in lowering the unintentional wasting of time. The list of modeling studies performed for the diverse food products discussed in the section above is not exhaustive but provides a general idea of how molecular dynamics is being used to determine more ideal treatment conditions, needed to achieve the desired final product state or quality. Molecular dynamics play a significant role in determining the necessary processing conditions required to improve the functional properties and attenuate antinutritional components in food products, while equally further reducing the allergenicity and improving the bioavailability of food proteins.

References

Huebbe, P.; Rimbach, G. Historical reflection of food processing and the role of legumes as part of a healthy balanced diet. Foods 2020, 9, 1056.
Singh, A.; Vanga, S.K.; Orsat, V.; Raghavan, V. Application of molecular dynamic simulation to study food proteins: A review. Crit. Rev. Food Sci. Nutr. 2018, 58, 2779–2789.
Dong, X.; Wang, J.; Raghavan, V. Critical reviews and recent advances of novel non-thermal processing techniques on the modification of food allergens. Crit. Rev. Food Sci. Nutr. 2021, 61, 196–210.
Trystram, G. Modelling of food and food processes. J. Food Eng. 2012, 110, 269–277.
Ekins, S. Computer Applications in Pharmaceutical Research and Development; Wiley-Interscience: Hoboken, NJ, USA, 2006.
Eom, K. Computer simulation of protein materials at multiple length scales: From single proteins to protein assemblies. Multiscale Sci. Eng. 2019, 1, 1–25.
Battimelli, G.; Ciccotti, G. Berni Alder and the pioneering times of molecular simulation. Eur. Phys. J. H Hist. Perspect. Contemp. Phys. 2018, 43, 303–335.
Oluwajobi, A. Molecular Dynamics Simulation of Nanoscale Machining. In Molecular Dynamics—Studies of Synthetic and Biological; Wang, L., Ed.; InTech: London, UK, 2012.
Stillinger, F.H.; Rahman, A. Improved simulation of liquid water by molecular dynamics. J. Chem. Phys. 1974, 60, 1545–1557.
McCammon, J.; Gelin, B.; Karplus, M. Dynamics of folded proteins. Nature 1977, 267, 585–590.
Karplus, M.; Kuriyan, J. Molecular dynamics and protein function. Proc. Natl. Acad. Sci. USA 2005, 102, 6679–6685.
Barazorda-Ccahuana, H.L.; Theiss-De-Rosso, V.; Valencia, D.E.; Gómez, B. Heat-stable hazelnut profilin: Molecular dynamics simulations and immunoinformatics analysis. Polymers 2020, 12, 1742.
De, V.M.; Masetti, M.; Bottegoni, G.; Cavalli, A. Role of molecular dynamics and related methods in drug discovery. J. Med. Chem. 2016, 59, 4035–4061.
Chen, G.; Huang, K.; Miao, M.; Feng, B.; Campanella, O.H. Molecular dynamics simulation for mechanism elucidation of food processing and safety: State of the art: Md simulation for mechanism elucidation…. Compr. Rev. Food Sci. Food Saf. 2019, 18, 243–263.
Titchenal, A.; Calabrese, A.; Gibby, C.; Revilla, M.K.F.; Meinke, W. Human Nutrition; University of Hawaii at Mānoa Food Science and Human Nutrition Program: Honolulu, HI, USA, 2018; Chapter 6; pp. 231–233. Available online: http://pressbooks.oer.hawaii.edu/humannutrition/ (accessed on 5 January 2022).
Saxena, R.; Vanga, S.K.; Raghavan, V. Effect of thermal and microwave processing on secondary structure of bovine β-lactoglobulin: A molecular modeling study. J. Food Biochem. 2019, 43, e12898.
Vanga, S.K.; Singh, A.; Raghavan, V. Effect of thermal and electric field treatment on the conformation of ara h 6 peanut protein allergen. Innov. Food Sci. Emerg. Technol. 2015, 30, 79–88.
Vagadia, B.H.; Vanga, S.K.; Singh, A.; Raghavan, V. Effects of thermal and electric fields on soybean trypsin inhibitor protein: A molecular modelling study. Innov. Food Sci. Emerg. Technol. 2016, 35, 9–20.
Wang, J.; Vanga, S.K.; Raghavan, V. Structural responses of kiwifruit allergen act d 2 to thermal and electric field stresses based on molecular dynamics simulations and experiments. Food Funct. 2020, 11, 1373–1384.
Zhu, Y.; Wang, J.; Vanga, S.K.; Raghavan, V. Visualizing structural changes of egg avidin to thermal and electric field stresses by molecular dynamics simulation. LWT 2021, 151, 112139.
Dong, M.; Tian, H.; Xu, Y.; Han, M.; Xu, X. Effects of pulsed electric fields on the conformation and gelation properties of myofibrillar proteins isolated from pale, soft, exudative (PSE)-like chicken breast meat: A molecular dynamics study. Food Chem. 2021, 342, 128306.
Liu, Y.; Zhao, G.; Zhao, M.; Ren, J.; Yang, B. Improvement of functional properties of peanut protein isolate by conjugation with dextran through Maillard reaction. Food Chem. 2012, 131, 901–906.
Avilés-Gaxiola, S.; Chuck-Hernández, C.; Serna Saldívar, S.O. Inactivation methods of trypsin inhibitor in legumes: A review. J. Food Sci. 2018, 83, 17–29.
Abdollahi, K.; Ince, C.; Condict, L.; Hung, A.; Kasapis, S. Combined spectroscopic and molecular docking study on the pH dependence of molecular interactions between β-lactoglobulin and ferulic acid. Food Hydrocoll. 2020, 101, 105461.
Huang, Y.; Zhang, X.; Suo, H.; Bello Ramírez, M. Interaction between β-lactoglobulin and EGCG under high-pressure by molecular dynamics simulation. PLoS ONE 2021, 16, e0255866.
Sahihi, M.; Heidari-Koholi, Z.; Bordbar, A.-K. The interaction of polyphenol flavonoids with β-lactoglobulin: Molecular docking and molecular dynamics simulation studies. J. Macromol. Sci. Part B 2012, 51, 2311–2323.

© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.

Upload a video for this entry

Information

Subjects: Food Science & Technology

Contributor MDPI registered users' name will be linked to their SciProfiles pages. To register with us, please refer to https://encyclopedia.pub/register : Xin Dong

View Times: 1.3K

Update Date: 20 Jan 2022

Table of Contents

Video Upload Options

Confirm

1. Introduction

2. Application of Molecular Dynamics Simulation for Food Products and Processes

References