Molecular dynamic (in silico) modeling of structure–function of glutelin type-B 5-like from proso millet storage protein: effects of temperature and pressure

Abstract

Molecular dynamic (MD) simulation provides an insight into the behavior of a protein under applied processing at the molecular level. The behavior of glutelin type-B 5-like protein, a type of glutelin protein from proso millet was studied, in solution under different temperatures (300, 350, and 400 K) and pressure (1 bar, 3 kbar, and 6 kbar) levels using a molecular dynamics simulation approach. The combined treatment effect (400 K, 6 kbar) increased the compaction of the protein compared to the level at (300 K, 1 bar) as shown by the decreased radius of gyration values from 3.26 to 2.92 nm, decreased solvent accessibility surface area from 327.47 to 311.06 nm² and decreased volume from 108.35 to 105.04 nm³. The root means square deviation increased with increasing temperature but decreased with increasing pressure while the root means square fluctuations increased significantly with increased in temperature and pressure. A snapshot of the three-dimensional structure of the protein revealed compression of its occluded cavities at higher pressure levels but no obvious disruption to the secondary structure elements of the protein was observed, except for the loss of a few amino acid residues that comprise the secondary structure element.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13197-022-05594-y.

Introduction

Molecular dynamic (MD) modeling simulation is gaining traction in elucidating changes in structural properties or conformation of foods at the molecular level, especially proteins as they are subjected to external food processing stresses (Fenwick et al. 2019; Vagadia et al. 2016; Vanga et al. 2018). Traditionally, food proteins are subjected to various physical (heating, high-pressure processing, ultrasound, and extrusion), chemical (glycosylation, and phosphorylation), and biological (hydrolysis, crosslinking, and fermentation) modification and/or processing techniques to improve their functionality, nutrient, bioavailability or suppress undesired property, by exploring the bio-physicochemical and structural properties of these proteins. Application of such processing techniques leads to changes in the native conformation of the food molecules such as protein (secondary structures, bond angles, and, bond length), however minute (Phillips 2013).

These changes in the protein structure–function modification are still at the meso- and macro-levels (Foegeding 2015; Withana-Gamage & Wanasundara 2012). Additionally, the various techniques (such as Spectroscopic: Fourier transformation infrared, Raman, Circular dichroism, Florescence, and Ultraviolet radiations, Nuclear magnetic resonance, X-ray crystallography, Cryo-electron microscopy, and Small-angle, x-ray scattering) used to quantify these alterations in food protein structural properties are limited to evaluating the static conformation (Akharume et al. 2019b; Wang et al. 2017) of the protein structures before and after stress application and not dynamic conformation that reveals the real-time behavior of the protein under applied stress.

MD simulations provide an opportunity to study and evaluate food protein dynamics and changes in conformation under real-time applied stress at the atomic and molecular levels (Singh et al. 2013; Withana-Gamage & Wanasundara 2012). This understanding provides insights into target protein and allows for precision suppression or amplification of such during specific applications such as reduction of anti-nutritional or allergenic attributes of foods.

A few authors have studied the effects of temperature and pressure in MD simulation on the structural changes of selected food proteins. Vanga et al. (2019) revealed from their study on MD simulation of Gly m 4 soy allergen protein under temperature and pressure, that there were significant changes recorded in the structure of the protein molecules, particularly with residues D-27, and T-51. For instance, the root mean square deviation (RMSD) of the protein molecule increased from 0.254 ± 0.03 to 0.324 ± 0.039 nm as the temperature moved from 300 to 373 K while the RMSD decreased from 0.324 ± 0.039 to 0.257 ± 0.037 nm at temperature 373 K as the pressure changes from 1 bar to 6kbar (Vanga et al. 2019). In another study, simulated temperature, and pressure of soybean trypsin inhibitor (STI) lead to changes in the radius of gyration, RMSD, and solvent accessibility surface area (SASA) of the protein molecule. For instance, both temperature (300–373 K) and pressure (1 bar–6 kbar) combinations were observed to reduce the radius of gyration of SPI from 1.581 ± 0.01 nm to 1.567 ± 0.006 nm at 300 K, 1.591 ± 0.01 nm to 0.1557 ± 0.008 nm at 345 K, and 1.590 ± 0.012 nm to 1.564 nm ± 0.008 nm at 373 K, as the pressure increased from 1 bar to 6 kbar, leading to a more compact STI molecule (Vanga et al. 2018).

Glutelin type-B 5-like protein is a type of glutelin protein from proso millet. Proso glutelin proteins are not widely used as food protein ingredients owing to their poor physicochemical and functional properties (Akharume et al. 2019a), although they have promising nutritional and health benefits (Saleh et al. 2013) as the glutelin has been reported to reduce the concentration of low-density lipoprotein (LDL) cholesterol and increase the concentration of high-density lipoprotein (HDL) cholesterol in several mice studies (Nishizawa & Fudamoto 1995; Park et al. 2008). Some modification to improve proteins usually involve the application of stresses like temperature and pressure. These modifications impact the structural conformation of the protein which are better understood at the molecular level.

Thermal treatment beyond the protein denaturation temperature (varies with ambient conditions − 82.1 ± 3.5 °C for proso millet protein fractions (Akharume et al. 2019a) and 73.3–82.2 °C for rice protein fractions (Ju et al. 2001)) leads to rupturing of the protein’s intra- and inter-molecular bonds, and loss of the protein secondary and tertiary structures (Sun-Waterhouse et al. 2014). Similarly, high pressure (usually beyond 200 MPa) leads to the loss of the protein secondary and tertiary structure but is not able to rupture its covalent bonds (Yang & Powers 2016). With the understanding that the changes in protein structure conformation may confer improved functionality, our objective was to evaluate the effect of selected processing stress within the range that are applicable in for food application, such as temperature (300, 350, and 400 K) and pressure (1 bar, 3 kbar, and 6 kbar) levels on the structure–function formation by considering changes at the secondary structural level, root mean square deviation (RMSD), root mean square fluctuations (RMSF) per residue, the radius of gyration (Rg), surface electrostatic potential, and solvent accessibility surface area (SASA) of the glutelin type-B 5-like protein. The selected temperature levels were to understand the protein behavior at room and atmospheric pressure and then to simulate the temperature within the range of pasteurization, and pressures range reported to cause protein modification (200–700 MPa) (Akharume et al. 2021). This understanding will provide insights into the behavior of the proteins at molecular/micro level from the standpoint of impact of different processing stresses and a range of desired changes in proteins during conversion processes.

Materials and methods

Molecular dynamic simulations

The MD simulations of glutelin type-B 5-like in water was carried out by the Groningen machine for chemical structure (GROMACS) software package (version 2018.1, Stockholm Center for Biomembrane Research, Stockholm, Sweden) (van Der Spoel et al. 2005). The glutelin type-B 5-like protein used in this study was developed in a previous study and can be accessed from the protein model database (PMDB) with the identity number of PM0083241. The protein (Fig. 1a) is a protomer of three non-covalently linked monomers with each monomer chain comprising 16.2% alpha-helix, 2.4% 3/10 helix, 33.5% beta-sheet, and 48% coils/turns/bends/bridges. The protein-containing 7713 atoms from 507 residues were kept in a cubic box of dimension 12.019 × 12.019 × 12.019 nm as a periodic boundary condition and the protein was solvated with 163,155 atoms of water molecules neutralized with 3 sodium ions (Fig. 1b).

Fig. 1.

A snapshot of the three-dimensional structure of glutelin type-B 5-like protein molecule (A) in a vacuum (B) in neutralized water

The OPLS-AA/L all-atom force field (Robertson et al. 2015) and SPC/E water model (Kusalik & Svishchev 1994) were selected to provide potential energy function and water parameters to the system. Following neutralization of the systems to mimic the physiological state of the protein, the systems were energy minimized to converge at maximum force value < 1000 kJ/mol/nm using stepwise descent minimization algorithm for 100,000 steps and the systems was further equilibrated to a constant number of particles, volume, and temperature (NVT) and a constant number of particles, pressure, and temperature (NPT) for 200 ps at 300 K and I bar.

The MD simulations were run for 1000 ps (pico second) using a leap-frog integrator algorithm during which the temperature of the systems was maintained using a modified Berendsen thermostat, and the pressure was maintained using the Parrinello-Rahman barostat (Berendsen et al. 1984; Parrinello & Rahman 1980). A total of nine simulations were ran for the pressure processing levels and the temperature levels (Table 1s—Supplementary file). The results of simulations such as RMSD, the radius of gyration, and SASA were analyzed using GROMACS inbuilt tools. GROMACS inbuilt DSSP (Kabsch and Sander 1983; Touw et al. 2015) was used for the secondary structure analysis and virtual molecular dynamics (VMD) (Humphrey et al. 1996) was used to visualize protein conformational change.

Results and discussion

Root mean square deviation (RMSD)

When a protein molecule is subjected to external simulated processing stress such as temperature and pressure, there are changes in their original conformation because of the applied stresses. The RMSD provides information on the deviation of the protein during simulations from its original conformation at the start of the simulation (time zero). Mostly the deviation from the protein backbone or main chain’s atom for an atom is usually considered as in the case in our experiment and such deviation is calculated using Eq. 1.

$$RMSD=\sqrt{\frac{1}{N}\sum _{i=1}^N{|r_{\mathit{final}}\left(i\right)-r_{\mathit{initial}}(i)|}^2}$$ 1

where $r_{\mathit{final}}$ is the final coordinates of atom i, and $r_{\mathit{initial}}$ is the initial coordinate of the atom i, and N is the number of atoms. Table 1 summarizes the average RMSD for glutelin type-B 5-like protein after 1 ns simulation from different combinations of temperature and pressure. The average values of RMSD for the glutelin type-B 5-like protein decreased significantly (p < 0.0001) with increasing pressure. When the temperature was constant at 300 K, the RMSD decreased with increasing pressure (1 bar–6 kbar) from 0.51 ± 0.16 to 0.37 ± 0.10 nm while at a temperature of 350 K, the RMSD decreased from 0.64 ± 0.21 to 0.47 ± 0.18 nm. Similarly, at a temperature of 400 K, the RMSD decreased from 0.82 ± 0.29 to 0.55 ± 0.12 nm. In addition, the average RMSD increased significantly (p < 0.0001) with increasing temperature from 0.51 ± 0.16 to 0.82 ± 0.29 nm, from 0.39 ± 0.14 to 0.66 ± 0.18 nm, and from 0.37 ± 0.10 to 0.55 ± 0.12 for pressure levels of 1 bar, 3kbar, and 6 kbar respectively. Vanga et al. (2018) reported a similar trend for soybean trypsin inhibitor protein where the RMSD values decreased from 0.269 ± 0.026 nm to 0.225 ± 0.019 nm when pressure changed from 1 bar to 6 kbar and similarly at other temperature levels of 345 and 373 K. Figure 2 presents the behavior of the glutelin type-B 5-like protein with simulation time. It can be seen from Fig. 2 that the RMSD increased with simulation time. At each pressure levels (Fig. 2a–c), simulation at 400 K showed the highest RMSD and 300 K showed the lowest RMSD. Conversely, at each temperature levels (Fig. 2d–f), the simulation at 1 bar gave the highest RMSD value and 6 kbar the lowest. This is very much expected, as the temperature increases the atoms of the protein molecules gain more energy for mobility, breaking of bonds, unfolding, and eventual denaturation of the protein molecule. However, the effect of pressure on protein is based on the Le Chatelier’s principle where a change in volume is accompanied by a change in pressure, that is as the pressure increases the volume of the protein decreases because the pressure pushes out the occluded cavities in the protein molecules (Galazka, 2000; Messens et al. 1997; Yang & Powers 2016) as can be seen later in Fig. 5. This might have been responsible for the decrease in the RMSD molecules.

Table 1.

Average root means square deviation (RMSD), radius of gyration (Rg), solvent accessibility surface area (SASA), volume, and density of glutelin type-B 5-like protein after 1000 ps simulation under temperature and pressure conditions

Treatments	RMSD (nm)	Rg (nm)	SASA (nm2)	Volume (nm3)	Density (g/l)
300 K, 1 bar	0.51 ± 0.16aj1	3.26 ± 0.07aj	327.47 ± 2.66 aj	108.35 ± 1.00 aj	835.27 ± 7.69 aj
300 K, 3 kbar	0.39 ± 0.14bm	3.07 ± 0.03bk	317.06 ± 2.98 bk	105.84 ± 1.08 bk	855.09 ± 8.71 bk
300 K, 6 kbar	0.37 ± 0.10co	3.06 ± 0.03 cl	313.45 ± 3.76 cl	104.95 ± 1.05 cl	862.41 ± 8.58 cl
350 K, 1 bar	0.64 ± 0.21dk	3.26 ± 0.05 dm	331.65 ± 3.18 dm	108.39 ± 1.09 dm	835.03 ± 8.39 dm
350 K, 3 kbar	0.40 ± 0.10em	3.12 ± 0.05en	317.33 ± 3.47 en	106.22 ± 1.11 en	852.05 ± 8.85 en
350 K, 6 kbar	0.47 ± 0.18fo	3.14 ± 0.05fo	317.98 ± 3.82 fo	105.33 ± 1.00 fo	859.26 ± 8.01 fo
400 K, 1 bar	0.82 ± 0.29gl	3.35 ± 0.08gp	330.84 ± 4.32 gp	108.82 ± 1.19 gp	831.71 ± 9.12 gp
400 K, 3 kbar	0.66 ± 0.18hn	3.09 ± 0.06hq	313.23 ± 10.68 hq	106.05 ± 1.44 hq	853.52 ± 11.58 hq
400 K, 6 kbar	0.55 ± 0.12io	2.92 ± 0.06ir	311.06 ± 5.31 ir	105.04 ± 1.16 ir	861.65 ± 9.44 ir

For a given attribute, values with different letter subscript are significantly different while values with common letter subscript are not significantly different

Fig. 2.

Root mean square deviation of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1000 ps (pico second) simulation time

Fig. 5.

Solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1000 ps simulation time

Root mean square fluctuation (RMSF)

The root mean square fluctuations per residue for all the three chains of glutelin type-B 5-like protein under different simulated temperature and pressure combinations for 1000 ps are presented in Fig. 1s (See Supplementary file). The RMSF provides information on the flexible and the rigid regions of the protein molecules. We observed that intensity of the fluctuation increases with increasing temperature and increasing pressure. For instance, at the N-terminal of chain A (Q77), the fluctuation recorded for 300 K 1 bar, 300 K 3 bar to 400 K 6kbar were 0.09, 0.32, 0.53; 0.75, 1.00, 1.21; 1.56, 1.73, 2.00 nm respectively and on chain B were 0.11, 0.27, 0.54; 0.80, 0.97, 1.22; 1.50, 1.70, 1.96 nm respectively. Similarly, for the alpha helix region on chain A (A233–K240), the average fluctuations recorded for all treatment combinations (300 K 1 bar, 300 K 3 bar … 400 K 6kbar) were 0.08, 0.29, 0.50; 0.80, 0.99, 1.20; and 1.50, 1.70, 1.90 nm respectively. Additionally, we observed that the N- and C- terminals on each of the protein chains as well as the alpha helix regions of the protein showed higher fluctuations compared to the beta sheets. For instance, on chain C we recorded average fluctuation values of 0.05, 0.28, 0.50; 0.77, 1.00, 1.25; 1.51, 1.76, 1.91 nm for alpha helix region (V222 -A226) and 0.06, 0.28, 0.51; 0.81, 0.97, 1.27; 1.51, 1.71, 1.93 nm for alpha helix region (A233–K240) and 0.04, 0.29, 0.50; 0.74, 0.97, 1.18; 1.44, 1.67, 1.90 nm for the β-sheet region (S122–T127) and 0.04, 0.28, 0.50; 0.76, 0.97, 1.20; 1.44, 1.66, 1.89 nm for β-sheet region (S122–T127) and we recorded 0.03, 0.27, 0.55; 0.84, 0.98, 1.27; 1.65, 1.67, 2.04 for the C-terminal (G245) for all treatment combinations (300 K 1 bar, 300 K 3 bar … 400 K 6kbar) respectively.

Secondary structure analysis

Figure 3 presents the average number of amino acid residues that make the secondary structure elements of the glutelin type-B 5-like protein molecules after a 1000 ps simulation under different simulation conditions. We expect that as the proteins unfold the number of residues for helices and beta sheets should decrease and adds to the number are coils, turns, bends, or bridges. We observed that the number of residues in coils, bends, and turns (6 kbar only) increases with temperature at constant pressure while in the alpha-helix and β-sheets the residue decreases with the temperature only at a constant pressure of 1 bar. However, as the pressure ramped up from 1 bar to 6 kbar the number of residues in the alpha-helix and β-sheets increased with temperature which may be a result of the refolding of the secondary structure elements in the protein or new aggregation which are a common characteristic of high-pressure treatment.

Fig. 3.

The numbers of residues present in the secondary structure of glutelin type-B 5-like protein molecules under different simulation conditions (temperature and pressure) after 1000 ps simulation time

The snap shots of the protein molecule after simulation (Fig. 2s in the Supplementary file) showed a compression in the cavities with the chains of the protein with pressure. However, no obvious disruption of the secondary structure elements of helices and beta sheets were observed in all the treatment combinations, perhaps the short time simulation may not have been enough to cause an irreversible unfolding or permanent denaturation of the protein secondary structure. Additionally, high-pressure processing is only known to rupture the non-covalent interactions (intramolecular hydrophobic and electrostatic interactions) and forms new non-covalent/semi-covalent associations without breaking of hydrogen bonds (Galazka et al. 2000; Messens et al. 1997; Yang & Powers 2016).

Radius of gyration (Rg)

The degree of spread or compaction of the glutelin type-B 5-like protein after stimulation was measured by the radius of gyration using the Eq. (2). The Rg provides information on the deviation of the atoms of the protein molecule with respect to its center of mass.

$$Rg=\sqrt{\frac{1}{N}\sum _{i=1}^N{|r\left(i\right)-r_{\mathit{centre}}|}^2}$$ 2

where $r\left(i\right)$ is the coordinate of atom i, $r_{\mathit{centre}}$ is the coordinate of the center of mass of the protein, and N is the number of atoms. We observed (Table 1) that at a constant temperature the Rg values of the protein decreased significantly (p < 0.0001) with an increase in pressure from 3.26 ± 0.07 to 3.06 ± 0.03 nm, 3.26 ± 0.05 to 3.14 ± 0.05 nm, and from 3.35 ± 0.08 to 2.92 ± 0.06 nm for 300, 350, and 400 K temperature levels, respectively. At constant pressure, the Rg values increases with temperature from 3.26 ± 0.07 to 3.35 ± 0.08 at 1 bar and from 3.07 ± 0.03 to 3.09 ± 0.06 at 3 kbar but decreased from 3.06 ± 0.03 to 2.92 ± 0.06 at 6 kbar. In summary, at low pressure (1 bar) higher temperatures lead to unfolding and spreading of the protein (Fig. 4a), however at a higher pressure of 3 and 6 kbar (Fig. 4b and c), the protein begins to experience some compaction even with high-temperature levels.

Fig. 4.

Radius of gyration of glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1000 ps simulation time

Solvent accessibility surface area (SASA)

The summary of the solvent accessibility surface area of glutelin type-B 5-like protein under different simulation conditions are presented in Table 1. The SASA provides information on the hydrophobicity of the protein. Hydrophobicity of protein is inversely related to its SASA (Gromiha et al. 2019). The average SASA decreased significantly (P < 0.0001) with pressure at a constant temperature as presented in Table 1. For example, at 300 K we recorded SAS values of 327.47 ± 2.66 nm² for 1 bar, 317.06 ± 2.98 nm² for 3 kbar, and 313.45 ± 3.76 for 6 kbar. Similarly, at a constant pressure of 1 bar, the SASA values increased with temperature showing that the pressure effects were not strong on SASA at this level. However, as the pressure was increased to 3 and 6 kbar the SAS values decreased, except at 350 K, 6 bar which points to the fact that the pressure compaction effect was less pronounced at this level. The trends in SASA with simulation time is presented in Fig. 5 and it can be observed that not much fluctuations were observed in the SASA value at low pressure (Fig. 5a) at higher pressure (Fig. 5b and c), the SASA showed initial upward trend up to 700 ps and then begins to decline greatly which may suggest initial unfolding and revelations of the buried hydrophobic residue followed by new hydrophobic aggregation the hides some of the protein residues form solvent accessibility.

Volume and density

The volume of protein under pressure stresses is expected to decrease and the density is expected to increase according to Le Chatelier’s principle. The average values of the volume and density of the glutelin type-B 5-like protein are presented in Table 1 and the trends with simulation time are presented in Fig. 6. The trends in the volume of the protein for all treatment combinations were relatively stable over the simulation time. However, the average volume at constant temperature levels decreased significantly (p < 0.0001) with pressure while at constant pressure, the volume increased significantly (P < 0.0050) with temperatures. Conversely, the average density at constant temperature levels increased significantly (p < 0.0001) with pressure while at constant pressure, the volume decreased with temperature.

Fig. 6.

Volume changes observed for glutelin type-B 5-like protein under different simulation conditions (temperature and pressure) for 1000 ps simulation time

Conclusion

In summary, we evaluated the effect of temperature and pressure combinations (300 K, 350 K, and 400 K for pressure levels 1 bar, 3 kbar, and 6 kbar) on the behavior of glutelin type-B 5-like protein in molecular dynamic simulation environment successfully. We observed that at a low pressure of 1 bar and high pressure of 3 kbar the temperature effects on the glutelin type-B 5-like protein is well pronounced and the stress at this levels led to increasing RMSD, RMSF, SASA, Rg and volume, but decreased density which reveals that the protein may not be experiencing much aggregation or compaction at the initial stage of simulation (600 ps) but as the simulation proceeds to 1000 ps, there may be compaction at 3 kbar levels (at all levels of temperature) for at the rest of the simulation time as a result of the pressure effects. However, at a higher pressure of 6 kbar, the temperature effects became less impactful on the values of Rg, SASA, and volume of the glutelin type-B 5-like protein so that the value of Rg decreased from 3.06 ± 0.03 to 2.92 ± 0.06 nm, and SASA values decreased from 313.45 ± 3.76 to 311.06 ± 5.31 nm² and volume decreased 104.95 ± 1.05 to 105.04 ± 1.16 nm³ which shows that the protein treated at this level of pressure and temperature may aggregate and compact. Secondary structure analysis reveals loss in the numbers of residues of the beta-sheet and alpha-helix with a corresponding increase in the number of the residue of coils, turns and bends with increasing temperature while the alpha-helix and beta-sheet only reduced with the temperature at a constant pressure of 1 bar and increased at higher pressure levels confirming aggregation or refolding at high-pressure levels. We opined that increasing the MD simulation length could be necessary for the disruptions of the protein secondary structure which can expose whether there is a permanent denaturation or aggregation at higher pressure. This understanding provides useful information that can be combined with Fourier Transform Infrared (FTIR) spectroscopy to understand the conformation changes due to the applied stresses.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MD simulation

Molecular dynamic simulation

GTB

Glutelin type-B5-like

RMSD

Root mean square deviation

RMSF

Root mean square fluctuations

STI

Soybean trypsin inhibitor

SASA

Solvent accessibility surface area

LDL

Low-density lipoprotein

NVT

Number of particles, volume, and temperature

NPT

Number of particles, pressure, and temperature

GROMACS

Groningen machine for chemical structure

PMDB

Protein model database

Rg

Radius of gyration

Author contributions

Dr. FA was responsible for the design of experiment, data collection and analysis as well as writing the manuscript. Dr AA was responsible for ideation, supervision of the study and correcting the manuscript.

Funding

This work was supported by the Kentucky Agricultural Experiment Station (KAES), and the National Institute of Food and Agriculture (NIFA), U.S. Department of Agriculture, Hatch- Multistate project #: 1024529.

Declarations

Conflict of interest

The authors declare no conflict of interest in the preparation, submission, and publication of this manuscript. All authors approved the submission and publication of this manuscript to the Journal of Food Science and Technology. Authors also acknowledge that this manuscript has not be previously published or being considered by other journal for publication.

Availability of Data and Materials

The dataset used for the protein model can be found in protein model database (PMDB) with the identity number of PM0083241. Additional data can be provided by the authors if requested.

Ethical approval

Not Applicable.

Consent to Participate

Not Applicable.

Consent for Publication

Not Applicable.

Code Availability

Not Applicable.