The Conformational Changes of Bovine Serum Albumin at the Air/Water Interface: HDX-MS and Interfacial Rheology Analysis

The characterization and dynamics of protein structures upon adsorption at the air/water interface are important for understanding the mechanism of the foamability of proteins. Hydrogen–deuterium exchange, coupled with mass spectrometry (HDX-MS), is an advantageous technique for providing conformational information for proteins. In this work, an air/water interface, HDX-MS, for the adsorbed proteins at the interface was developed. The model protein bovine serum albumin (BSA) was deuterium-labeled at the air/water interface in situ for different predetermined times (10 min and 4 h), and then the resulting mass shifts were analyzed by MS. The results indicated that peptides 54–63, 227–236, and 355–366 of BSA might be involved in the adsorption to the air/water interface. Moreover, the residues L55, H63, R232, A233, L234, K235, A236, R359, and V366 of these peptides might interact with the air/water interface through hydrophobic and electrostatic interactions. Meanwhile, the results showed that conformational changes of peptides 54–63, 227–236, and 355–366 could lead to structural changes in their surrounding peptides, 204–208 and 349–354, which could cause the reduction of the content of helical structures in the rearrangement process of interfacial proteins. Therefore, our air/water interface HDX-MS method could provide new and meaningful insights into the spatial conformational changes of proteins at the air/water interface, which could help us to further understand the mechanism of protein foaming properties.


Introduction
Proteins are amphiphilic biopolymers that can quickly adsorb to the air/water interface, reduce surface tension, and further stabilize foam-based food systems [1]. The solubility, hydrophobicity, flexibility, and charge of protein molecules are the major factors impacting the interfacial adsorption process [2][3][4]. The interfacial adsorption of proteins includes two basic steps: (i) the protein molecules move rapidly to the air/water interface and their attachment; (ii) the polymer chains uncoil and rearrange (surface denaturation) [5]. Previous research has demonstrated that proteins adsorb to the air/water interface by hydrophobic and electrostatic interactions [6]. Therefore, hydrophobic and positively charged amino acid residues are the potential sites interacting with the interface. After the adsorption of proteins at the interface, changes are achieved not only quantitatively but also conformationally [7,8].
The degree of conformational changes of proteins at the air/water interface strongly affects the rheological properties of the interfacial films formed by proteins [9,10]. The

Solution Preparation
The BSA solution was prepared in phosphate buffer saline (PBS, 10 mM, pH 7) and stirred for 2 h at room temperature. The BSA dispersion was maintained at 4 • C for 12 h for full hydration.

Foaming Properties
The optimal protein concentration for the interfacial experiments was determined according to the foaming properties. With a homogenizer (T18, IKA) set to 8000 rpm, BSA solutions under the various concentrations (1,3,5,7, and 9%, m/v, 20 mL) were foamed for 2 min at room temperature. The foams were poured into 50 mL measuring cylinders as soon as possible. According to an earlier method, foamability (FA) and foam stability (FS) calculations were made [30]. In short, FA was calculated by comparing the foam volume at 2 min to the beginning liquid volume of the samples (20 mL), whereas FS was evaluated by comparing the foam volume at 60 min to the foam volume at 2 min.
where V 2 is the volume of foam at 2 min, and V 60 is the volume of foam at 60 min. On a Nikon instrument, the visual foam volume was captured in photos after being kept at room temperature for 2, 5, 10, 15, 30, 45, 60, and 90 min. Three measurements were made for FA and FS.

Interfacial Rheology Analysis
The interfacial rheology of the BSA at the air/water interface was determined by a drop tensiometer (Tracker Teclis/IT Concept, Longessaigne, France) [31]. Following the injection of the BSA solution into a vessel, an axisymmetric rising bubble with a constant volume of 4 µL was produced at the inverted tip of an air-filled syringe. A camera with a charge-coupled device (CCD) was adopted to capture the photographs of the bubble. It took 7200 s to measure each sample. The surface tension of the PBS was measured to make sure there were no additional surfactants present. The interfacial pressure (π) was obtained from Equation (3) π = γ 0 − γ s (3) where π represents the interfacial pressure, γ 0 represents the interfacial tension of the PBS, and γ s represents the interfacial tension of the protein solution. At a constant temperature of 25 • C ± 1 • C, at least three replicates of the measurements were made for each sample.

Hydrogen-Deuterium Exchange in the Solution
Purchased deuterium oxide (RHAWN, Shanghai, China) was used directly to carry out the HDX experiments. To initiate the HDX experiments in the solution, 10 µL of the prepared BSA solution (0.1%, m/v) was added to 140 µL of D 2 O and allowed to be labeled continuously at room temperature for predetermined times, specifically 0 s, 10 min, and 4 h, representing medium and long exchange times, respectively [32]. Then, the exchange reaction was quenched by mixing samples with an equal volume of prechilled quench buffer containing 0.8% (m/v) formic acid and 0.8 M of guanidine hydrochloride in water, to lower the pH to 2.5, on ice for 1 min. A pH meter was used to determine the pH of the system, with no isotope corrections applied [20].

System Design and Circulation Experiments
To perform the hydrogen-deuterium exchange experiments at the air/water interface, a self-assembly device was utilized. Several ports were machined into opposite sides of a trough (acrylic), and additional Teflon tubing was connected to a peristaltic pump to efficiently and quickly exchange the aqueous subphase (the buffer under the adsorbed protein layers at the air/water interface) with the D 2 O. Figure 1A displays a schematic drawing of the experimental setup. To measure the minimum amount of subphase buffer needed for full subphase exchange, circulation experiments were conducted. Absorbance measurements of the tartrazine dye-containing buffer at 425 nm were made, both before and after circulation [33]. The efficiency of the subphase exchange was calculated by Equation (4) where A 1 is the absorbance of the subphase buffer after circulation, and A 0 is the absorbance of the subphase buffer before circulation.

Hydrogen-Deuterium Exchange at the Air/Water Interface
Hydrogen-deuterium exchange experiments at the air/water interface were carried out utilizing the system described above. An amount of 330 μL of the BSA solution was

Hydrogen-Deuterium Exchange at the Air/Water Interface
Hydrogen-deuterium exchange experiments at the air/water interface were carried out utilizing the system described above. An amount of 330 µL of the BSA solution was spread onto the surface of 3 mL of aqueous subphase (PBS, 10 mM, pH 7) in the trough. A microsyringe was used to carry out the procedure of depositing the protein layer at the air/water interface. The protein layer was considered to reach dynamic equilibrium after 10 min [34]. Following that, a peristaltic pump was used to circulate 48 mL of subphase buffer and 12 mL of D 2 O through the trough, which took 5 min and did not disturb the adsorbed proteins at the air/water interface. The adsorbed proteins were labeled for predetermined times-0 s, 10 min, and 4 h-identical to the solution hydrogen-deuterium exchange experiments. Immediately after the labeling period, 300 µL of the adsorbed proteins and subphase buffer (final concentration, 5.78 mg/L) were sucked into a sample tube containing 300 µL of precooling quench buffer.

Digestion, Desalting, and Storage of Samples
The following steps were the same for the quenched samples in the solution and at the air/water interface ( Figure 1B). Pepsin and aspergillopepsin (final concentrations 0.020% and 0.023%, m/v, respectively, dissolved in water) were added, and the samples were digested on ice for 5 min. The digested samples were desalted using Waters Sep-Pak Vac 1cc (50 mg) C18 Cartridges. Finally, the desalted peptides were dried with a CentriVap vacuum concentrator (Labconco, Kansas City, MO, USA) and kept at −80 • C for the following separation and mass analysis.
Using a Q Exactive Orbitrap mass spectrometer (Thermo Scientific), a data-dependent MS/MS analysis was carried out. A distal 2.5-kV spray voltage was used to electrospray the eluted peptides from the LC column straight into the mass spectrometer. A cycle of one full scan of the MS spectrum (m/z 300-1800) was obtained. Then, the top 20 MS/MS events were created on the first to the twentieth most intense ions chosen from the full MS spectrum at a 30% normalized collision energy. With an automated gain control (AGC) target of 3 × 10 6 , the full scan resolution was set at 70,000. With an isolation window of 1.8 m/z and an AGC target of 1 × 10 5 , the MS/MS scan resolution was set at 17,500. For both the MS and MS/MS scans, there was one microscan, and the maximum ion injection times were 50 and 100 ms, respectively. The parameters for the dynamic exclusion were as follows: exclusion time, 30 s; exclude isotopes, on; and charge exclusion, 1 and >8. The Xcalibur data system (Thermo Scientific) managed the LC solvent gradients and MS scan operations. Using the HDExaminer software (Version 3.3, Sierra Analytics), the deuterium incorporation (# Deut, Da) and deuteration percentage (% Deut, %) of each tested peptide at various exchange time points were computed, both of which represented the deuteration level. The deuterium incorporation was the number of deuterons corresponding to the difference between the deuterated and undeuterated centroids, while the deuteration percentage was the ratio of the deuterium incorporation to the theoretical maximum deuteration (Max D). The deuterium incorporation (# Deut, Da) and deuteration percentage (% Deut, %) for each peptide were calculated by Equations (5) and (6), respectively.
where # Deut represents the deuterium incorporation (Da), D centroid represents the deuterated centroids, UD centroid represents the undeuterated centroids, z is the constant coefficient, % Deut is the deuteration percentage (%), and Max D represents the theoretical maximum deuteration (Da) for each peptide. The relative deuterium incorporation (r # Deut, Da) of each tested peptide was calculated by subtracting the deuterium incorporation of the peptide in the solution from the deuterium incorporation of the peptide at the air/water interface, while the relative deuteration percentage (r % Deut, %) was calculated by subtracting the deuteration percentage of the peptide in the solution from the deuteration percentage of the peptide at the air/water interface, as indicated by Equations (7) and (8) r # Deut = # Deut at the air−water inter f ace − # Deut in the solution (7) r % Deut = % Deut at the air−water inter f ace − % Deut in the solution (8) where r # Deut is the relative deuterium incorporation, and r % Deut is the relative deuteration percentage. The resulting numbers provided by HDExaminer were then imported into PyMOL (The PyMOL Molecular Graphics System, Version 2.3, Schrodinger, LLC) to map the regions of increased or decreased exchange of BSA.

Statistical Analysis
The statistical analysis was performed with a variance (ANOVA) using the SPSS 26.0 statistical analysis program. The differences were defined through the means of trials by a significant difference test (p < 0.05). All the measurements were replicated three times unless stated otherwise.

Optimization of Protein Concentration Used for HDX-MS at the Air/Water Interface and Interfacial Rheology Analysis Based on Foaming Properties
The foaming characteristics and foam volume of the BSA under the different concentrations (1, 3, 5, 7, and 9%, m/v) are presented in Figure 2. Based on Figure 2A, the FA was constant at BSA concentrations from 5% (m/v) onwards. The FS was not affected by the concentration, which was likely due to the high protein concentration used. With the increase in the BSA concentration, more BSA molecules were adsorbed to the air/water interface, resulting in the ascent of the FA [35,36]. Nevertheless, with the continuous elevation of the protein concentration, the adsorption equilibrium of the BSA at the air/water interface was reached, leading to a constant FA [37,38]. In Figure 2B, the results demonstrate that the foam volume for all the samples steadily reduced from 2 to 90 min. The foaming properties of the BSA were often correlated to the interfacial adsorption and interfacial structures. Considering the results of the FA and FS conjunctively, we decided to use the concentration of 5% (m/v) in the following interfacial rheology analysis and interfacial HDX-MS experiments. demonstrate that the foam volume for all the samples steadily reduced from 2 to 90 m The foaming properties of the BSA were often correlated to the interfacial adsorption a interfacial structures. Considering the results of the FA and FS conjunctively, we decid to use the concentration of 5% (m/v) in the following interfacial rheology analysis a interfacial HDX-MS experiments.

Determination of the Volume of D2O Utilized for HDX-MS at the Air/Water Interface by Circulation Experiments
After the device was assembled, subphase exchange (circulation) experiments we carried out [33]. As shown in Figure 3, the subphase exchange with buffer volumes of 9, and 12 mL (2, 3, and 4 times the volume of the subphase buffer in the trough) result in 68.6%, 78.2%, and 89.8% exchange ratios of the subphase, respectively. Although a concentration of D2O could be used for HDX experiments, higher concentrations are ten used to obtain a higher deuteration level [39]. The 89.8% subphase exchange ra could meet the requirement of D2O concentration (typically 80-90%, v/v) in the labeli experiments [20]. Moreover, to minimize the structural interference of the protein in t solution on the protein at the interface, 48 mL of fresh buffer was exchanged before t D2O to remove the protein in the solution.

Determination of the Volume of D 2 O Utilized for HDX-MS at the Air/Water Interface by Circulation Experiments
After the device was assembled, subphase exchange (circulation) experiments were carried out [33]. As shown in Figure 3, the subphase exchange with buffer volumes of 6, 9, and 12 mL (2, 3, and 4 times the volume of the subphase buffer in the trough) resulted in 68.6%, 78.2%, and 89.8% exchange ratios of the subphase, respectively. Although any concentration of D 2 O could be used for HDX experiments, higher concentrations are often used to obtain a higher deuteration level [39]. The 89.8% subphase exchange ratio could meet the requirement of D 2 O concentration (typically 80-90%, v/v) in the labeling experiments [20]. Moreover, to minimize the structural interference of the protein in the solution on the protein at the interface, 48 mL of fresh buffer was exchanged before the D 2 O to remove the protein in the solution.

Structural Changes of Adsorbed BSA at the Air/Water Interface Analyzed by HDX-MS and Interfacial Rheology
To investigate conformational changes of proteins at the air/water interface, the method of HDX-MS for interfacial proteins was developed. Setting BSA as the model protein, its deuterium incorporation (Da) at the air/water interface and in the solution was compared. Due to its accessibility, BSA has been one of the most studied proteins for decades, contributing to a detailed characterization. With three homologous domains, the three-dimensional structure of BSA is available. In the neutral environment, the percentage of α-helices in BSA is 48% [40]. As a protein, BSA also has ideal functional properties, such as foaming and emulsifying, leading to a wide range of applications in the food field. The available structural information and good features at interfaces for BSA have made it an ideal model of a structured globular protein. Thus, in this study, we adopted BSA as the

Structural Changes of Adsorbed BSA at the Air/Water Interface Analyz Interfacial Rheology
To investigate conformational changes of proteins at the air/w method of HDX-MS for interfacial proteins was developed. Setting protein, its deuterium incorporation (Da) at the air/water interface a was compared. Due to its accessibility, BSA has been one of the most s decades, contributing to a detailed characterization. With three homol three-dimensional structure of BSA is available. In the neutral env centage of α-helices in BSA is 48% [40]. As a protein, BSA also has ide erties, such as foaming and emulsifying, leading to a wide range of food field. The available structural information and good features at have made it an ideal model of a structured globular protein. Thus adopted BSA as the model protein to analyze protein structural altera tion to the air/water interface by HDX-MS.
Our results from the HDX-MS analysis show that, in total, 36 d the BSA were detected, leading to a 67.1% sequence coverage (Figure 4 results of the 10 min labeling time showed that the degree of deuterium most peptides of the BSA was quite low, both in the solution and at face. This was because the rates of deuterium incorporation depended the protein structures [25]. The flexible proteins fundamentally te deuterium incorporation rates, contributing to a larger mass shift in time. However, for the typical globular protein, BSA, which contains and is relatively stable, there were extensive regions with a low deute rate. When the labeling time increased to 4 h ( Figure 4B), the majori Our results from the HDX-MS analysis show that, in total, 36 digested peptides of the BSA were detected, leading to a 67.1% sequence coverage (Figure 4). In Figure 4A, the results of the 10 min labeling time showed that the degree of deuterium incorporation for most peptides of the BSA was quite low, both in the solution and at the air/water interface. This was because the rates of deuterium incorporation depended on the stability of the protein structures [25]. The flexible proteins fundamentally tended to show fast deuterium incorporation rates, contributing to a larger mass shift in a shorter labeling time. However, for the typical globular protein, BSA, which contains 17 disulfide bonds and is relatively stable, there were extensive regions with a low deuterium incorporation rate. When the labeling time increased to 4 h ( Figure 4B), the majority of BSA peptides showed increased deuterium incorporation. The deuterium incorporation of peptides 54-63, 97-103, 165-174, 227-236, 229-251, 355-366, 361-369, 370-377, 561-567, and 567-574 in the solution labeling at 4 h showed a significant increase compared to the labeling at 10 min. The results indicated that these peptides were located on the surface of the BSA and more exposed to the solvent (D 2 O), which facilitated the adsorption of these peptides to the air/water interface, leading to a certain orientation of BSA at the air/water interface [6]. The peptides 54-63, 165-174, and 355-366 all contained random coil structures, proving the reliability of the HDX-MS results [25]. However, the deuterium incorporation of peptides 448-456 and 519-529 in the solution labeling at 4 h showed a large decrease compared to the labeling at 10 min, as a result of steric solvent shielding (caused by the random protein-protein interaction) and conformational changes [41,42]. As shown in Figure 4, there were significant differences for some peptides in deuterium incorporation between the protein at the air/water interface and in the solution for the interface effect, such as peptides 37-43,  Meanwhile, the interfacial rheology of BSA was analyzed. After adsorbing to the interface, proteins can greatly lower the interfacial tension to stabilize foam systems. At pH 7.0, the BSA (5%, m/v) reduced the interfacial tension from 71.85 mNm −1 to 45.13 mNm −1 after two hours of adsorption. When the adsorption time reached 4000 s, the interfacial tension of the BSA was relatively stable, indicating that this stage was close to the dynamic adsorption equilibrium. The results in Figure 5A show the curve of interfacial pressure (π) versus the square root of time (t 1/2 ) for the BSA protein at 5% (m/v). Within 100 s, the diffusion-controlled fast adsorption process took place. To determine the diffusion rate (k diff ), the slope of the linear plot of the curve between 0 and 100 s was calculated [0.58 (0.98) mNm −1 s −0.5 ]. The high molecular weight of BSA (66 k Da) explains why the value was smaller than that of other proteins [40]. In Figure 5B, the typical plot of ln[(π 7200 − π t )/(π 7200 − π 0 )] with the time of the BSA at 5% (m/v) is displayed. The penetration rate (k P ) of the protein is represented by the slope of the first linear area (2000-6000 s), while its rearrangement rate (k R ) is shown by the slope of the second linear region (6000-7200 s). The second linear area demonstrates that the BSA completed the structural uncoiling and rearrangement within 2 h. The values of k P and k R were 5.84 × 10 −4 (0.98) s −1 and 18 × 10 −4 (0.96) s −1 , respectively, smaller than other flexible proteins, indicating that relatively less conformational rearrangement of the BSA occurred at the air/water interface [43,44].  Meanwhile, the interfacial rheology of BSA was analyzed. After adsorbing to t interface, proteins can greatly lower the interfacial tension to stabilize foam systems. pH 7.0, the BSA (5%, m/v) reduced the interfacial tension from 71.85 mNm −1 to 45 mNm −1 after two hours of adsorption. When the adsorption time reached 4000 s, the terfacial tension of the BSA was relatively stable, indicating that this stage was close  As analysis of the interfacial rheology results of the BSA testifies, the HDX exper ment results of the BSA labeling at 10 min revealed structural changes during the ad sorption. Meanwhile, the HDX experiment results of the BSA labeling at 4h reflect th structural state after the uncoiling and rearrangement process.
The results of HDX-MS of the BSA in the solution showed that peptide 54-63 e changed a few deuterium at 10 min labeling, but its deuterium incorporation showed a increase of >4.0 Da at 4 h labeling ( Figure 4). However, the same peptide, 54-63, at th air/water interface showed low deuterium incorporation labeling at both 10 min and 4 indicating that the air/water interface offered protection from deuterium exchange. Sim ilar results were obtained from the analysis of peptides 227-236 and 355-366. Compare to the BSA in the solution labeling at 4 h, peptide 227-236 at the air/water interfa showed a reduction >3.5 Da, and peptide 355-366 demonstrated a decrease >6 Da in th deuterium incorporation ( Figure 4B). Figure 6A,B displays the 3D structures of BS mapped by the relative deuteration percentage labeling at 10 min and 4 h, respectivel From the spatial positions of peptides 54-63, 227-236, and 355-366 in Figure 6C, it w observed that they were on a plane, where the protein interacted with the air/water i terface. Thus, we deduced that peptides 54-63, 227-236, and 355-366 were involved the adsorption of the BSA to the interface, resulting in the certain orientation of the BS at the air/water interface after 4 h labeling [45]. Our results are in agreement with th current theories: when proteins (especially structured globular proteins) adsorb to th air/water interface, most of the protein molecules remain in the aqueous phase, and on a small part mosaic on the air/water interface [46,47]. As analysis of the interfacial rheology results of the BSA testifies, the HDX experiment results of the BSA labeling at 10 min revealed structural changes during the adsorption. Meanwhile, the HDX experiment results of the BSA labeling at 4h reflect the structural state after the uncoiling and rearrangement process.
The results of HDX-MS of the BSA in the solution showed that peptide 54-63 exchanged a few deuterium at 10 min labeling, but its deuterium incorporation showed an increase of >4.0 Da at 4 h labeling ( Figure 4). However, the same peptide, 54-63, at the air/water interface showed low deuterium incorporation labeling at both 10 min and 4 h, indicating that the air/water interface offered protection from deuterium exchange. Similar results were obtained from the analysis of peptides 227-236 and 355-366. Compared to the BSA in the solution labeling at 4 h, peptide 227-236 at the air/water interface showed a reduction >3.5 Da, and peptide 355-366 demonstrated a decrease >6 Da in the deuterium incorporation ( Figure 4B). Figure 6A,B displays the 3D structures of BSA mapped by the relative deuteration percentage labeling at 10 min and 4 h, respectively. From the spatial positions of peptides 54-63, 227-236, and 355-366 in Figure 6C, it was observed that they were on a plane, where the protein interacted with the air/water interface. Thus, we deduced that peptides 54-63, 227-236, and 355-366 were involved in the adsorption of the BSA to the interface, resulting in the certain orientation of the BSA at the air/water interface after 4 h labeling [45]. Our results are in agreement with the current theories: when proteins (especially structured globular proteins) adsorb to the air/water interface, most of the protein molecules remain in the aqueous phase, and only a small part mosaic on the air/water interface [46,47].
Previous research has shown the role of hydrophobic interaction. In that theory, the hydrophobic part of proteins extends to the air phase in the interfacial adsorption [6]. Hydrophobic amino acids (non-polar amino acids, containing Ala, Leu, Ile, Val, Pro, Trp, Phe, and Met) were favorable for protein adsorption to the interface. Moreover, Li et al. [45] examined cryo-electron microscopy information from frozen proteins using various detergents (anionic, cationic, nonionic, and zwitterionic detergents) and found that the electrostatic interaction between proteins and air/water interface carrying negative charges also had an influence on the protein adsorption to the interface. The peptides possessing positively charged amino acid residues (Lys, Arg, and His) were more likely to adsorb to the interface, leading to a preferred orientation of the protein at the interface. Therefore, peptides composed of hydrophobic and positively charged amino acid residues are likely to be involved in the protein adsorption to the air/water interface. These interactions are supposed to affect the orientation of adsorbed proteins on air/water interfaces. Previous research has shown the role of hydrophobic interaction. In that theory, the hydrophobic part of proteins extends to the air phase in the interfacial adsorption [6]. Hydrophobic amino acids (non-polar amino acids, containing Ala, Leu, Ile, Val, Pro, Trp, Phe, and Met) were favorable for protein adsorption to the interface. Moreover, Li et al. [45] examined cryo-electron microscopy information from frozen proteins using various detergents (anionic, cationic, nonionic, and zwitterionic detergents) and found that the electrostatic interaction between proteins and air/water interface carrying negative charges also had an influence on the protein adsorption to the interface. The peptides possessing positively charged amino acid residues (Lys, Arg, and His) were more likely to adsorb to the interface, leading to a preferred orientation of the protein at the interface. Therefore, peptides composed of hydrophobic and positively charged amino acid residues are likely to be involved in the protein adsorption to the air/water interface. These interactions are supposed to affect the orientation of adsorbed proteins on air/water interfaces.
To determine the residues involved in the interactions between the protein residues and the air/water interface, the relative deuteration percentage of each residue was calculated by subtracting the deuteration percentage of the BSA in the solution from that at the air/water interface. We assumed that the obvious decrease in the residue's relative deuteration percentage (<−90%) was caused by the interaction between the protein residues and the air/water interface. In our experiment, peptides 54-63, 227-236, and 355-366 showed significant reductions in their relative deuteration percentage. The results in Figure 7 display the difference heatmap denoted by the relative deuteration percentage after labeling at 10 min and 4 h. In peptide 54-63, one positively charged amino acid residue, H63, was located at the end of the peptide. The relative deuteration percentage of L55 was <−90%, which was considered to be involved in the process of adsorption to the air/water interface. For peptide 227-236, the hydrophobic amino acid residues (F229, A233, L234, and A236), and the positively charged amino acid residues (K228, R232, and K235) accounted for 70%. Moreover, the relative deuteration percentages of A233, L234, K235, and A236 were <-90% and the relative deuteration percentage of R232 was <−70%. Peptide 355-366 had three hydrophobic amino acid residues (P362, A365, and V366) and three positively charged amino acid residues (R359, R360, and H361), playing a key role in the adsorption process. Furthermore, the relative deuteration percentage of R359 was To determine the residues involved in the interactions between the protein residues and the air/water interface, the relative deuteration percentage of each residue was calculated by subtracting the deuteration percentage of the BSA in the solution from that at the air/water interface. We assumed that the obvious decrease in the residue's relative deuteration percentage (<−90%) was caused by the interaction between the protein residues and the air/water interface. In our experiment, peptides 54-63, 227-236, and 355-366 showed significant reductions in their relative deuteration percentage. The results in Figure 7 display the difference heatmap denoted by the relative deuteration percentage after labeling at 10 min and 4 h. In peptide 54-63, one positively charged amino acid residue, H63, was located at the end of the peptide. The relative deuteration percentage of L55 was <−90%, which was considered to be involved in the process of adsorption to the air/water interface. For peptide 227-236, the hydrophobic amino acid residues (F229, A233, L234, and A236), and the positively charged amino acid residues (K228, R232, and K235) accounted for 70%. Moreover, the relative deuteration percentages of A233, L234, K235, and A236 were <−90% and the relative deuteration percentage of R232 was <−70%. Peptide 355-366 had three hydrophobic amino acid residues (P362, A365, and V366) and three positively charged amino acid residues (R359, R360, and H361), playing a key role in the adsorption process. Furthermore, the relative deuteration percentage of R359 was <−70%, and the relative deuteration percentage of V366 was <−50%. Thus, residues L55, H63, R232, A233, L234, K235, A236, R359, and V366 from peptides 54-63, 227-236, and 355-366 were considered as the main sites interacting with the air/water interface. The hydrophobic amino acid residues and positively charged amino acid residues accounted for 78% among the amino acid residues, whose relative deuteration percentages were <−70% from peptides 54-63, 227-236, and 355-366, which was consistent with the theory that the protein interacts with the negatively charged air/water interface via hydrophobic and electrostatic interactions.
When the fragments of BSA adsorbed to the air/water interface, the conformation and solvent accessibility of the surrounding peptides were also affected. In Figures 4B and  6C, peptide 204-208, located in the vicinity of peptide 227-236, displayed an increment of >3 Da in the deuterium incorporation on the interface at 4 h labeling. When peptide 227-236 adsorbed to the interface, the conformation of nearby peptides changed because of the interfacial interaction and the interfacial tension. In addition, next to peptide 355-366, peptide 349-354 showed an increased deuterium incorporation, indicating the improvement of its solvent accessibility. 023, 12, x FOR PEER REVIEW 12 for 78% among the amino acid residues, whose relative deuteration percentages <−70% from peptides 54-63, 227-236, and 355-366, which was consistent with the th that the protein interacts with the negatively charged air/water interface via hydrop and electrostatic interactions. When the fragments of BSA adsorbed to the air/water interface, the conform and solvent accessibility of the surrounding peptides were also affected. In Figur and 6C, peptide 204-208, located in the vicinity of peptide 227-236, displayed an ment of >3 Da in the deuterium incorporation on the interface at 4 h labeling. When tide 227-236 adsorbed to the interface, the conformation of nearby peptides change cause of the interfacial interaction and the interfacial tension. In addition, next to pe 355-366, peptide 349-354 showed an increased deuterium incorporation, indicatin improvement of its solvent accessibility.
In previous research, the in situ CD spectroscopy technique was utilized to stud conformation (mainly secondary structure) of BSA in the adsorbed state at the air/w interface [16]. The estimation from the surface CD spectrum revealed 35% β-turn random coil, and no α-helix. Compared with the secondary structure of BSA in the tion, composed of 60% α-helix and 40% random coil, the BSA lost its helix structu the interface and turned into a disordered protein. This was consistent with our resu far-ultraviolet CD spectroscopy of BSA in the solution in Table S1. In our research detected differences in the deuterium incorporation between the samples from the tion and the air/water interface mostly belong to the peptides located in the α structure. Thus, we speculated that the helical structures in peptides 54-63, 227-236 354, and 355-366 transformed into random coils in the process of uncoiling and rangement.
Although an air/water interface HDX-MS method was developed to study the formational changes of interfacial proteins in situ in this study, there was still one s coming of the method compared to a standard workflow of HDX-MS experiments The deuteration rates could be affected by many factors, such as temperature an [25]. In order to achieve better separation of peptides, the temperature of the H column was set at 55 °C, which may change the deuteration levels of peptides (bac change). However, the control samples were set up and the deuteration levels in bot control and experimental samples were affected in the same way. The results in study, which were calculated by subtracting the deuteration levels of the experim samples from those of the control samples, have eliminated the effect of back exch To improve the sensitivity of the method and increase the intensity of the results fu the temperature of the HPLC column will be set at 0 °C in the following experiment In previous research, the in situ CD spectroscopy technique was utilized to study the conformation (mainly secondary structure) of BSA in the adsorbed state at the air/water interface [16]. The estimation from the surface CD spectrum revealed 35% β-turn, 65% random coil, and no α-helix. Compared with the secondary structure of BSA in the solution, composed of 60% α-helix and 40% random coil, the BSA lost its helix structure on the interface and turned into a disordered protein. This was consistent with our results of far-ultraviolet CD spectroscopy of BSA in the solution in Table S1. In our research, the detected differences in the deuterium incorporation between the samples from the solution and the air/water interface mostly belong to the peptides located in the α-helix structure. Thus, we speculated that the helical structures in peptides 54-63, 227-236, 349-354, and 355-366 transformed into random coils in the process of uncoiling and rearrangement.
Although an air/water interface HDX-MS method was developed to study the conformational changes of interfacial proteins in situ in this study, there was still one shortcoming of the method compared to a standard workflow of HDX-MS experiments [20]. The deuteration rates could be affected by many factors, such as temperature and pH [25]. In order to achieve better separation of peptides, the temperature of the HPLC column was set at 55 • C, which may change the deuteration levels of peptides (back exchange). However, the control samples were set up and the deuteration levels in both the control and experimental samples were affected in the same way. The results in this study, which were calculated by subtracting the deuteration levels of the experimental samples from those of the control samples, have eliminated the effect of back exchange. To improve the sensitivity of the method and increase the intensity of the results further, the temperature of the HPLC column will be set at 0 • C in the following experiments.

Conclusions
In our project, we successfully established a process of HDX-MS analysis for food systems to monitor the conformational changes of a model protein, BSA, at the air/water interface in situ. Peptides with great differences in deuterium incorporation between the samples in the solution and at the air/water interface were analyzed. The results suggested that the peptides 54-63, 227-236, and 355-366 might be involved in the adsorption of BSA to the air/water interface, and their residues L55, H63, R232, A233, L234, K235, A236, R359, and V366 might interact with the air/water interface through hydrophobic and electrostatic interactions. Furthermore, the conformational changes of peptides 54-63, 227-236, and 355-366 resulted in conformational changes of their surrounding peptides 204-208 and 349-354, causing the reduction of helix structures in the uncoiling and rearrangement process of the interfacial adsorbed BSA. This work could provide a new, powerful method for studying the conformational changes of proteins at the air/water interface and help to understand the mechanism of interfacial protein adsorption and rearrangement at the peptide and amino acid residue level.