Molecular Insight into Binding Behavior of Caffeine with Lactoferrin: Spectroscopic, Molecular Docking, and Simulation Study

The majority of bioactive substances in the human diet come from polyphenols. Here, we use spectroscopy, molecular docking, molecular dynamics simulations, and in vitro digestion to look at the relationship be-tween caffeine (CAF) and bovine lactoferrin (BLF). The correlation analysis of the CAF-BLF fluorescence quenching process revealed that the reaction was spon-taneous and that the CAF-BLF fluorescence quenching process may have been static. The predominant intrinsic binding forces were hydrogen bonds and van der Waals forces, which were also supported by molecular docking and molecular dynamics simulations. Through fourier infrared and circular dichroism spectroscopy experiments, it was found that CAF changed the secondary structure of BLF and might bind to the hydrophobic amino acids of BLF. Compared with BLF, CAF-BLF showed inhibitory effects on digestion in simulated in vitro digestion. It will be helpful to better understand the interaction between CAF and BLF and provide the basis for the development of innovative dairy products.


INTRODUCTION
An iron-binding glycoprotein found in milk called bovine lactoferrin (BLF) has a high capacity to bind iron (Jia, et al., 2021).Its molecular weight is around 80 kDa (Krolitzki, et al., 2022).BLF is utilized extensively in the food and medical industries as a high-value dietary protein (Krolitzki, et al., 2022).Studies have shown that BLF has antibacterial, anti-inflammatory, and anti-tumor properties in addition to its ability to regulate the immune system of mammals, lessen gastrointestinal discomfort, maintain iron balance in the body, and promote the absorption of calcium and magnesium and other nutrients (Ochoa, et al., 2020;Jing, et al., 2021;Cao, et al., 2022).BLF controls cell growth and eliminates damaging free radicals (Zhang, et al., 2020).The BLF that enters the small intestine guards against oxygen-free radicals in the body (Wang, et al., 2019).Yet if the gastric fluids degrade the BLF, then BLF cannot fulfill its purpose.In this situation, BLFs lost some of their biofunctional characteristics and became less bioavailable (Strate, et al., 2001;Wang, et al., 2017;Wang, et al., 2021).The gastrointestinal system must therefore receive BLF in a structurally sound and biologically functional state.
Proteins and polyphenols can easily form complexes and, as is well known, have a very strong affinity for one another.The structure and functional characteristics of proteins may change as a result (Jiang, et al., 2019).Covalent and noncovalent interactions are the 2 main forms of interactions between polyphenols and proteins.Nevertheless, quinone polymers and other unfavorable byproducts such as covalent protein-polyphenol interactions could develop (Yildirim-Elikoglu, et al., 2017).High temperatures and alkaline environments can significantly influence the stability of phenolic compounds, and covalent polyphenol-protein complexes can reduce the biological activity and bioavailability of dietary polyphenols (Trombley, et al., 2011;Cao, et al., 2017).Non-covalent protein-polyphenol interactions are more common than covalent bonds, which include hydrogen bonding, hydrophobic interactions, and van der Waals gravity (Yildirim-Elikoglu, et al., 2017).Khan et al. studied the interaction between sunset yellow dye and β-lactoglobulin using various spectral and molecular modeling techniques.The thermodynamic results showed that the interaction between sunset yellow dye and β-lactoglobulin was mainly hydrophobic (Khan, et al., 2022).Al-Shabib et al. clarified the binding mechanism of catechin and β-lactoglobulin and found through spectral tests that the interaction between β-lactoglobulin and catechin occurred spontaneously mainly through hydrophobic interaction (Al-Shabib, et al., 2020).The ΔH and ΔS measured by Al-Shabib et al. by fluorescence indicate that the main interaction between the natural polyphenol compound rutin and β-lactoglobulin is hydrophobic (Al-Shabib, et al., 2018).Precupas et al. used isothermal titration calorimetry, differential scanning calorimetry, circular dichroism, and molecular docking methods to study the interaction between caffeic acid and bovine serum albumin (BSA), and the results showed that the presence of caffeic acid enhanced the stability of BSA (Precupas, A., et al., 2017).
To stimulate the central nervous system, caffeine (CAF) is used.It boosts brain activity and the body's circulation of hormones like cortisol and adrenaline (Belscak-Cvitanovic, et al., 2015;Correa, et al., 2021).As a result, caffeine is regarded as a superb non-volatile alkaloid substance (Rashidinejad, et al., 2022).Unfortunately, caffeine has a bad taste, and consuming too much of it can harm your nervous and cardiovascular systems.According to studies, adding protein can help solve solubility issues and lessen the unpleasant sensation (Santa Rosa, et al., 2021).To comprehend the forces that bind protein-polyphenol complexes together, this study explored the molecular interactions between BLF and CAF.To aid in future research on the functional qualities of these complexes in food, we also detailed the structural characteristics of the complexes generated by BLF and CAF.Finally, an in vitro simulation of the CAF-BLF complex and BLF digestion alone was carried out to investigate the protein-polyphenol complex's protection and stability maintenance method of protein function.

Materials
Yuanye Biotechnology Co., LTD (Shanghai, China) was paid for bovine lactoferrin (BLF, 95%), and crude bovine lactoferrin was purchased from Beston Pure Dairies Pty LTD.(South Australia, Australia).Caffeine (CAF, 99.96%) came from Chengdu must bio-technology Co., LTD.Kangwei Century Biotechnology (Beijing) Co., LTD.and Beijing Solaibao Technology Co., LTD.both sold BCA protein quantification kits and SDS-PAGE gel preparation kits.Coomas Bright Blue G-250 was purchased from Tianjin Comeiou Chemical Reagent Co., LTD.Rainbow 130 Broad-spectrum Protein marker (15-130 kDa) was purchased from Beijing Soleibao Technology Co., LTD.It was decided to buy sodium dodecyl sulfate (SDS) from Tianjin Fuchen Chemical Reagent Factory.Protein was dissolved in phosphate-buffered saline (PBS, 0.1 mol L −1 , pH 7.4).Analytical grade reagents were also employed in the experiment, and ultra-pure water was the water used in the experiment.In further studies, none of the samples underwent more purification.

Reagent Configuration
With the aid of a PBS buffer, the BLF protein solution was made into a 10 µM working solution.Caffeine was produced with a PBS buffer to create a working solution with a 500 µM concentration.This solution was maintained in brown glass tubes and diluted as needed for further experiments.In a certain proportion, water, glacial acetic acid, and ethanol made up the decolorizing solution.A specific ratio of Tris, glycine, SDS, and water made up the electrophoresis buffer.Following preparation, all reagents were kept in a refrigerator at 4°C.

Fluorescence Spectrum
The fluorescence spectrum of a 4 mL sample of the BLF protein at a concentration of 10 µM was scanned using an RF-6000 fluorescence spectrophotometer.The protein was then added dropwise to the CAF solution at a concentration of 500 µM to generate 9 distinct BLF: CAF molar ratios (1:0, 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:8, 1:10).At 280 nm, the excitation wavelength was set.The scanning range of 290-450 nm data was used to determine the emission wavelength.The excitation and emission light's slit width were set to 5 nm.Three distinct temperatures-290 K, 300 K, and 310 K-were used in the trials.

Fourier transform infrared spectroscopy
To create FTIR samples, BLF (10 µM) and CAF (500 µM) solutions were combined in ratios of 1:0 and 1:8.The test sample was smeared onto Spectral grade KBr tablets for analysis after being pulverized and compressed into tablets.The measurements, which had a resolution of 4 cm −1 and covered the range from 4,000 to 400 cm −1 , were made using a BRUKER INVENNIO S Fourier infrared spectrometer.

Circular Dichroism Experiment
A Characin CD spectrometer was used to gather CD spectral information from the protein solution.A dropwise addition of CAF solution with an amount of 500 µM was made to the pure sample of protein after adding 1 mL of the 10 µM BLF sample of protein solution to the quartz cuvette to produce 3 distinct BLF: molar ratios of CAF (1:0, 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:8,  1:10).Far UV spectrum data was gathered by scanning the wavelength region between 190 nm and 260 nm.To create the circular chromatograms for each system and the contents of each protein secondary structure component, the data was processed using CDNN software.

Molecular Docking Simulation
Molecular docking simulation tests were carried out with the help of Auto Dock 4.2.In this experiment, the crystal structure of BLF (PDB ID: 1BIY) was downloaded from the Protein Data Bank (PDB) website charge (Wang, et al., 2021).After being loaded into pymol, it underwent a series of modifications to become an automatic docking-compatible system file, including the removal of water, the addition of hydrogen atoms, the correction of erroneous amino acids, and the adjustment of charge.PubChem (https: / / pubchem .ncbi.nlm.nih.gov/ ) provided the CAF structure, which was then translated to PDB format using Open Babel (PubChem CID: 2519) (Santa Rosa, et al., 2021).The Lamarckian Genetic Algorithm (GA) was used for semi-flexible molecular docking, and we selected the grid box size to cover the complete protein at 126 × 126 × 114 Å.To find the best conformation after docking and further examine the microenvironment of the CAF-BLF complex binding site, pymol was employed to visually investigate the predictions.

Molecular Dynamics Simulation
Gromacs2022.3 software was used for molecular dynamics simulation (Van Der Spoel, et al., 2005).Gaussian 16W was applied to hydrogenate small compounds and determine RESP potential, and AmberTools22 was utilized to apply the GAFF force field to tiny molecules.Potential information would be added to the molecular dynamics system's topology file.Atmosphere pressure of 1 bar and a static temperature of 298 K were used for the simulation.Amber99sb-ildn was utilized as the force field, water molecules were used as the solvent (Tip3p water model), and an appropriate amount of Na + ions were added to balance out the system's overall charge.With a coupling constant of 0.1 ps and a duration of 100 ps, the simulation system first used the steepest descent method to minimize energy before running through the isothermal isovolumic ensemble (NVT) equilibrium and isothermal isobaric ensemble (NPT) equilibrium for 100000 steps in each case.The simulation of free molecular dynamics was then completed.The operation took 10 ns to complete and had 5000000 steps with a step length of 2 fs.The software's built-in tool was applied to examine the trajectory when the simulation was complete.Each amino acid trajectory's root-mean-square variance (RMSD), root-mean-square fluctuation (RMSF), and protein rotation radius were merged with the free energy (MMPBSA), free energy topography, and other data (Abraham, et al., 2015).

Simulated in Vitro Digestion Experiment
An in vitro digestion system with oral, gastric, and intestinal phases was used to track the digestive patterns of BLF and the CAF-BLF complex.The activity of digestive enzymes and the concentration of bile salts at each stage of digestion were measured experimentally or by recommended standardized measurements (Brodkorb, et al., 2019;Zhang, et al., 2019;Wan, et al., 2023).INFOGEST method was used to simulate the digestive process in vitro (Brodkorb, et al., 2019).
Five mL of 300 µg/mL BLF and CAF-BLF samples were mixed with 4 mL of oral digestive fluid and 1 mL of 750 U/ml α-salivary amylase, respectively, and shaken for 2 min at a low speed (40-60 r/ min) at 37°C and away from light.After the completion of oral digestion, a 1 mL reaction sample was immediately placed on ice to terminate digestion and was used as the oral digestion group (SSF).The rest of the solution continued gastric digestion.The remaining samples after oral digestion were mixed with 6.5 mL gastric digestive fluid and 2.5 mL of 15000 U/mL pepsin, and the pH was adjusted to 3.0.Shake at a low speed (40-60 r/ min) for 120 min at 37°C and avoid light.After the completion of gastric cavity digestion, a 1 mL reaction sample was immediately placed on ice to terminate digestion and was used as the gastric cavity digestion group (SGF).The rest of the solution continued intestinal digestion.Add 13.5 mL intestinal digestive fluid, 3.5 mL 1000 U/ mL trypsin, and 0.15 g bile salt to the remaining sample after gastric digestion, and adjust the pH to 7.0.Shake at a low speed (40-60 r/ min) for 120 min at 37°C and avoid light.After digestion, the reaction samples were placed on ice immediately to terminate digestion and were used as the intestinal cavity digestion group (SIF).
The digestive liquids at different stages were centrifuged at 4°C, 3 000 × g, for 5 min, and the supernatant was taken and refrigerated at −80°C for subsequent analysis.

Bicinchoninic Acid Assay
10% TCA was uniformly distributed throughout a portion of the digestive solutions from each system, and the combination was then submerged in ice water at 4°C for 30 min.In a frozen centrifuge, the solution was removed and centrifuged for 10 min at 4000 rpm at 4°C.After removing the supernatant, the digestibility of proteins in each system was evaluated with a BCA test.To get the average, the experiment was run 3 times.

SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE)
SDS-PAGE was used to evaluate the digestion materials in 12% (wt/vol) acrylamide gels.Eight distinct sample types were examined: The above samples were loaded into the ep tubes after being thoroughly mixed with the 4:1 volume-tovolume 5 × SDS-PAGE loading buffer.They were then heated in a water bath at 100°C for 3-5 min.It was then chilled for subsequent use in an ice bath.Each lane was loaded with aliquots of the following samples: Marker, SSF(BLF), SSF(CAF-BLF), SGF(BLF), SGF(CAF-BLF), SIF(BLF), and SIF (CAF-BLF).Different settings were established at different voltages for separation and electrophoresis, respectively.The electrophoresis stopped when the indicator contacted the ground.The gel was scraped after electrophoresis, washed with deionized water, and then stained with Coomassie Brilliant Blue G-250 for 2 h on a shaker.The gel was taken out, washed with water that had been twice distilled, and then allowed to soak in a decolorizing solution all night.The decoloring solution should be swapped out periodically.

Experimental Analysis of UV-Vis's Spectra
An inexpensive, straightforward, adaptable, and nondestructive technique for analyzing organic materials and some inorganic stuff is UV-visible absorption spectroscopy (Rocha, et al., 2018).It is used to investigate how proteins change structurally and how their complexes form.In the UV-Vis' spectroscopy, proteins and polyphenols can be absorbed.The structure of the protein skeleton is reflected at 200 nm.It is an aromatic amino acid at 280 nm.Typically, electron transitions between π-molecular orbitals are responsible for their UV-visible absorption spectra (Poklar Ulrih, 2017).Hence, UV-Vis spectroscopy investigations were used to investigate the interaction mechanism between CAF and BLF.
The UV-visible spectra of BLF at various caffeine doses are depicted in Figure 1A.At first, it was largely possible to overlook the CAF's absorption peak strength in the protein UV spectrum's wavelength range.Second, the UV absorption intensity of the BLF amino residue at 280 nm dramatically increased with the rise in CAF content.The CAF-BLF complex system's UV absorption spectrum in the enlarged image showed a little red shift.This redshift of the absorption peak might be explained by CAF interacting with the hydrophobic amino acid residues of BLF molecules to form a hydrophobic cavity (Poklar Ulrih, et al., 2017).
Regular changes in the UV absorption spectra of CAF upon addition further suggested that CAF and BLF could interact to create complexes, while further research would be required to confirm the mechanism of the complex system's interaction.Next, we will use a variable temperature fluorescence experiment to further calculate the thermodynamic values of the binding process.

Fluorescence Quenching
Fluorescence Quenching Type between Caffeine and Lactoferrin.The fluorescence emission of a protein solution containing these fluorophores is typically decreased when ligands are added due to the fluorescence quenching of small molecules.However, the inner-filter effect (IFE), which occurs when the ligand appropriately absorbs light in the excitation and emission wavelength range, has a significant impact on the UV-Vis absorption capabilities of the ligand and the fluorescence attenuation of the protein (Rubio, et al., 1986).The fluorescence intensity was adjusted using the following equation to remove the inner-filter effects: A ex and A em are the quencher and lactoferrin's absorption intensities at excitation and emission wavelengths, and F cor and F obs are the fluorescence intensities observed and corrected, respectively.The behavior can be explained by a variety of quenching methods, but the 2 that are most frequently used are dynamic quenching and static quenching.Dynamic quenching is the reduction of fluorescence intensity without changing the structure of the protein caused by the interaction of quenching small molecules with  (Zhu, et al., 2016).Since greater temperatures might lead to bigger diffusion coefficients, the bimolecular quenching constants will rise in the dynamic quenching process as the temperature rises.In contrast, because complexes become less stable as temperature rises, the static quenching constants are anticipated to decrease (Zhang, et al., 2013).Dynamic quenching and static quenching can be identified based on these traits.To further explore, used the Stern-Volmer equation (2) to process experimental data: Where F 0 and F are lactoferrin's strengths with and without a quenching agent, respectively.The quencher concentration is Q.Stern-Volmer quenching constant is K sv .The quenching rate constant is K q .It is lactoferrin's typical half-life, which is 10 ns when the quencher is not present.By doing a linear regression on the F 0 /F and Q pictures, the Stern-Volmer equation may calculate the value of K sv .
Figure 1B shows the effect of caffeine on lactoferrin at temperatures of 300K.As seen in the image, the interaction between caffeine and lactoferrin resulted in a considerable drop in lactoferrin fluorescence intensity with an increase in caffeine concentration.
The relationship between F/F 0 and caffeine concentration is plotted from the data in Figure 1B (see Figure 2).There was a good linear relationship between them.At 290 K, 300 K, and 310 K temperatures, the quenching process rate constant Kq was computed (see Table 1).It could be seen from the experimental data that the value of K q was much higher than the dynamic quenching constant 2.0 × 10 10 L mol −1 s −1 of the maximum diffusion collision.The technique by which caffeine quenched the fluorescence of lactoferrin might therefore be static quenching rather than dynamic quenching.
Binding Constant and Binding Site Analysis between Caffeine and Lactoferrin.Equation (3) can be used to explain static quenching.The slope and intercept of log(F 0 -F)/F and log[Q] in the doublelogarithm equation can be used to calculate the binding constant K A and the binding site n for static quenching: Where [P] is the concentration of BLF, [K A ] is the binding constant, n is the number of binding sites, and F 0 and F are the strengths of lactoferrin with or without a quenching agent.The quencher concentration is Q.
Figure 3 shows the log((F 0 -F)/F) and log(1/([Q]−[P] (F 0 −F)/F 0 )) relationships during quenching at different temperatures, the K A and n can be determined from the y-axis intercept and slope, which are shown in Table 1.It could observe that the K A decreased with the rising temperature, which further suggested that the quenching between CAF and BLF was static quenching.Additionally, a medicine that is substantially protein- bound will normally have a K A value between 10 5 and 10 7 L mol −1 , while a drug that is moderately or weakly protein-bound will have a K A value between 10 2 and 10 4 L mol −1 (Yamashita, et al., 2013).The binding force between the CAF and BLF was therefore modest to moderate in vivo.
Thermodynamic Analysis and Interaction Force of Caffeine and Lactoferrin.Hydrogen bonds, electrostatic forces, van der Waals forces, and hydrophobic interactions are among the mechanisms through which tiny molecules and proteins interact.One way to assess this interaction is to examine the change in thermodynamic characteristics.The calculation equation was as follows: Where K A is the connection constant in equation ( 3), T is the thermodynamic temperature, R is the gas's stability, ΔH and ΔS are the changes in enthalpy and the amount of entropy shift, respectively, and ΔG is the free energy of Gibbs.
In Figure 4, the log K A vs. 1/T curve is displayed.After linearly fitting the data, Table 1 lists the values of ΔH and ΔS.Using Eq. ( 5), the ΔG value is determined and listed in Table 1.The hydrogen bonds and van der Waals force appeared to be the primary internal bonding forces between caffeine and lactoferrin as CAF and BLF's ΔH and ΔS values were −40.54 kJ mol −1 and −63.3 kJ mol −1 , respectively (Ali, et al., 2017).ΔH < 0 indicated that the reaction was exothermic, and ΔG < 0 indicated that the reaction between caffeine and lactoferrin was spontaneous.
FTIR spectroscopy.To determine the secondary structure of proteins, Fourier transform infrared spectroscopy (FTIR) is performed.FTIR can provide comprehensive information regarding chemical composition (Yang, et al., 2022).The advantages of FTIR over circular dichroism spectroscopy include the ability to analyze protein structure in solids, crystals, and aqueous solutions, as well as the effectiveness of each wave number's absorbance.PeakFit 4.12 software was used to calculate the area of each peak of the amido I band     2. The secondary structure of lactoferrin was changed after the binding of lactoferrin with caffeine, in which the proportion of α-helix and β-turns decreased, and the proportion of β-sheet and random coil increased.It was possible that TA's binding to hydrophobic areas of the BLF molecule, which disrupted some hydrogen bond networks, was what caused the observed reduction in α-helix structure (Xiong, et al., 2016).Three important characteristic absorption bands are included in the FTIR spectrum, 1700cm −1 -1600cm −1 (amide I band) represents the tensile vibration of C = O, 1600cm −1 -1500cm −1 (amide II band) represents the bending vibration of N-H and C-H, and 1300cm −1 -1260cm −1 represents the vibration of C-O and C-O-C (Wang, et al., 2022).As shown in Figure 5(A), the addition of caffeine increases the peak value of lactoferrin, and C = O is stretched.
Circular Dichroism.The interaction of caffeine with BLF may have an impact on the structure of BLF by a variety of spectral approaches.As a result, the BLF's structure needs to be verified.The secondary structure of lactoferrin and caffeine was examined using a circular dichroic spectrometer.The alteration of secondary structure can be seen in the far UV band (200-250 nm) (Khalil, et al., 2023).A common technique for determining the secondary structure of proteins is circular dichroism.
The modifications to the α-helix, β-sheet, β-turn, and random coil architectures are depicted in Figure 5(B).With a rise in CAF concentration, the contents of the β-sheet and random curling structures showed an upward trend (P < 0.05), while the α-helix and β-turn structures showed a decreasing trend.It was consistent with earlier observations to show that the far-UV circular dichroism spectrum had a minimum at 208 nm and was primarily α-helix in structure.The α-helix's p-p* transition can be found in the 208 nm band (Jash, et al., 2015, Liu, et al., 2021).The intensity of the far UV CD spectrum of BLF decreased during titration as CAF concentration increased, but the peak value remained constant, indicating a reduction in helical structure (Figure 5(B)), indicating that caffeine binding caused secondary structural changes in the conformation of BLF.These findings were similar to those of Gong et al., who found that binding purple potato anthocyanins to casein and whey protein resulted in decreased helix and turn contents and increased sheet and random coil contents (Gong, et al., 2021).Because CAF-BLF was more likely to result in a more compact and structured β-sheet than BLF, this suggested that CAF-BLF's protein structure might be more stable (Jing et al., 2021).
Molecular Modeling of CAF and BLF Interactions.The development of CAF-BLF complexes can be understood at the molecular level using thermodynamic information from thermodynamic experiments.Yet, a clever and effective technique for learning more about binding locations is molecular modeling technology.The thermodynamics of interactions between relevant molecules can be precisely predicted using molecular simulations (Nunes, et al., 2020).To evaluate the binding energy and contact force of protein-ligand complexes, identify the ligand's presumed binding site, and investigate the microenvironment of the binding site, molecular docking was used (Shanmugaraj, et al., 2015).On the other hand, by replicating experimental parameters like solvents, pH, and the presence of ions, molecular dynamics could forecast interactions (Rocklin, et al., 2013).
In Figure 6, the mimicked docking model that used the smallest binding energies for CAF and BLF was shown (the combination had the greatest stability when   in this manner).It was obvious that the CAF and BLF active sites were intimately connected and integrated with it.Figure 6 displays the molecular docking simulation's binding mode results.The results of the docking study showed that the residues of lactoferrin, including Arg 603, Leu600, Gly387, Arg323, Thr326, and Arg342, were involved in the interaction with CAF, which bonded to Arg603 by hydrogen bonds, and the remaining 5, Leu600, Gly387, Arg323, Thr326, and Arg342, by hydrophobic forces.

Molecular Dynamics Simulation of CAF and BLF Interaction.
To evaluate the stability of CAF molecules binding to BLF and the interaction mechanism, 100000 ps molecular dynamics simulations were performed.The root means square deviation (RMSD) value between the main atom of the protein and the initial structure was analyzed to test the trajectory stability of CAF molecules after binding to the protein.As shown in Figure 7A, after 30000-50000ps simulation, RMSD values of CAF molecule and protein tended to be stable and low, which was close to the stability of protein molecule, indicating that CAF molecule had high binding stability.The outcomes agreed with stoichiometric outcomes from thermodynamic tests (Van Der Spoel, et al., 2005;Abraham, et al., 2015).
RMSF, which measures the average of atomic position changes over time, can describe the flexibility and degree of motion of protein amino acids over the whole simulation process.As shown in Figure 7B, the RMSF of the protein was lower and slightly increased after binding with CAF, indicating that it had certain changes in protein structure but still had high stability.(Van Der Spoel, et al., 2005;Abraham, et al., 2015).
Solvent-accessible surface area (SASA) is the surface area of biomolecules that are exposed to the solvent.During the simulation, the scale of the probe radius is less than 1.4 Å (Van Der Spoel, et al., 2005;Abraham, et al., 2015).As shown in Figure 7C, the changes in SASA indicated that the microenvironment and surface hydrophobicity of the CAF-BLF complex changed and finally reached stability during the molecular simulation.
Cyclotron radius can be used to describe how tightly packed a protein structure is as well as how the protein's peptide chain looseness changes over the simulation process.In the molecular dynamic simulation, the tightness and rigidity of the protein skeleton were evaluated by Rg, which could also reflect the strength of the protein stability at different temperatures (Van Der Spoel, et al., 2005;Abraham, et al., 2015).As can be seen from Figure 7D, the structure of the CAF-BLF complex was relatively stable in the time range of 30000ps-54000ps.
Binding energy values and interactions between ligands and proteins are calculated by MM/PBSA, and the results are shown in Table 3.The binding energy between lactoferrin-caffeine was −66.26 kJ mol −1 , which indicated that lactoferrin was automatically bound to CAF.The electrostatic force between lactoferrin-caffeine was −12.31 kJ mol −1 and the van der Waals force was −101.88 kJ mol −1 .Also, there was one hydrogen bond between lactoferrin and caffeine.It showed that the electrostatic effect of lactoferrin and caffeine was much smaller than van der Waals forces and negligible.To bind CAF to lactoferrin, hydrogen bonds, and van der Waals contributions were crucial.Thus, the results confirm the findings of quenching fluorescence: both molecular dynamics simulations and thermodynamic experiments showed that van der Waals forces and hydrogen bonds were the main binding forces between caffeine and lactoferrin.

Simulated in Vitro Digestion
It was crucial to understanding variations in the gastrointestinal stability of BLF because BLF was easily digested in the stomach and broken down into tiny particles by the human body, ensuring that nutrients were taken in the intestine and transported through the blood.The retention rate of CAF-BLF in the stomach and intestine was higher than that of BLF, although the retention rate of CAF-BLF in the oral was lower.This is clear from Figure 8(A).This demonstrated that BLF's physiological activity in the intestine benefited from CAF's ability to reduce the amount of digestion that took place in the stomach.The molecular weight of digestive fluid proteins, may be seen following the completion of various phases of digestion, as shown in Figure 8(B).The BLF protein and the CAF-BLF complex were broken down into minute peptide or amino acid fragments, the majority of which had a molecular weight of 50 KDa, during digestion in the mouth and stomach.However, when the BLF protein and CAF-BLF complex entered the intestinal digestion stage, they were broken down into smaller peptide and amino acid fragments, and there were hardly any noticeable bands on the electrophoretic map.
It was suggested that CAF might bind to pepsin and inhibit pepsin because hydrophobic interactions between polyphenols and digestive enzymes might decrease the digestion of proteins (Jing, et al., 2021).When combined with the secondary structure test results, it was clear that CAF stabilized and compactified the structure of BLF, which might make pepsin's activity on the active site of BLF difficult (Jing, et al., 2021).The CAF-BLF complex was less digestible as a result.These findings conclusively showed that the CAF-BLF complex, which was created by combining CAF and BLF, could prevent the body from absorbing BLF.The findings served as a theoretical foundation for the preparation and manufacture of beneficial dairy products.
In conclusion, it has been shown through thermodynamic experiments, Fourier transforms infrared spectroscopy, circular dichroism spectra experiments, molecular docking, molecular dynamics simulation, and in vitro digestion experiments that CAF can spontaneously bind to BLF through hydrogen bonding and van der Waals forces.It will alleviate the problem where BLF is predigested by the stomach and unable to perform its function in the intestine.More research will be required to determine how CAF affects BLF's functional qualities other than digestibility.Additionally, because infant formula milk powder and middle-aged and elderly dairy products received many reports on BLF, it would be important to consider whether BLF can lessen the stimulation of the central nervous system and other negative effects caused by CAF to better protect benefit groups.Also, it will increase the CAF-BLF complex's consumer market and the economic value of functional dairy products.

CONCLUSION
The effects of caffeine on the structure and digestive properties of lactoferrin were studied by multispectral techniques, computational molecular docking, and molecular dynamics simulations.The reaction of caffeine with lactoferrin was exothermic, a spontaneous binding process of static quenching.The formation of CAF-BLF and the conformational change of BLF were further confirmed by fluorescence, UV, FTIR, and circular dichroism spectra.The interaction between CAF and several hydrophobic amino acid residues in the pocket of the active site of BLF was predicted using molecular docking and molecular dynamics simulations.Van der Waals and hydrogen bonds accounted for many of the forces that interacted between CAF and BLF.In vitro digestion results showed that CAF could inhibit BLF digestion and make it function better.In consequence, these studies will add to the potential of lactoferrin and caffeine as nutritional molecules in functional foods.
Li et al.: Molecular Insight into… (A) Oral digestion of BLF; (B) Oral digestion of CAF-BLF; (C) Oral and gastric digestion of BLF; (D) Oral and gastric digestion of CAF-BLF; (E) Oral, gastric, and intestinal digestion of BLF; and (F) Oral, gastric, and intestinal digestion of CAF-BLF.
Li et al.: Molecular Insight into… fluorescent excited state proteins.The development of a non-luminescent ground state complex between fluorescent molecules and ground state quenching molecules results in static quenching of the fluorescence quenching process, which weakens fluorescence and alters the molecular structure of proteins Figure 1.UV absorption spectrum (A) and.fluorescence quenching spectra (B) of CAF-BLF complex at 300K Figure 2. Stern-Volmer diagram of caffeine quenching of lactoferrin at different temperatures

Figure 4 .
Figure 4. Van't Hoff diagram of caffeine binding to lactoferrin Figure 5. (A) FT-IR spectra of BLF and complex at 298K.(B) Circular dichroic spectra of CAF-BLF at 3 concentrations

Figure 6 .
Figure 6.Molecular docking of CAF-BLF Figure 7. (A)Root means square deviation (RMSD) of caffeine molecule interaction with lactoferrin.(B)root mean square fluctuation (RMSF) of caffeine molecule interaction with lactoferrin.(C)SASA of caffeine molecule interaction with lactoferrin.(D) cyclotron radius of caffeine molecule interaction with lactoferrin