The effect of heat treatment on the lactosylation of milk proteins

Protein lactosylation is a significant modification that occurs during the heat treatment of dairy products, causing changes in proteins’ physical-chemical and nutritional properties. Knowledge of the detailed lac-tosylation information on milk proteins under various heat treatments is important for selecting appropriate thermo-processing and identifying markers to monitor heat load in dairy products. In the present study, we used proteomics techniques to investigate lactosylated proteins under different heating temperatures. We observed a total of 123 lactosylated lysines in 65 proteins, with lactosylation even occurring in raw milk. The number of lactosylated lysines and proteins increased moderately at 75°C to 130°C, but dramatically at 140°C. We found that 6 out of 10, 9 out of 16, 6 out of 12, and 5 out of 15 lysine residues in κ-casein, β-lactoglobulin, α-lactalbumin, and α S1 -casein, respectively, were lacto-sylated under the applied heating treatment. Moreover, different lactosylation states of individual lysines and proteins can indicate the intensity of heating processes. Lactosylation of K14 in β-lactoglobulin could distinguish pasteurized and UHT milk, while lactosylation of lactotransferrin can reflect moderate heat treatment of products.


INTRODUCTION
Heat treatment is a crucial process in dairy product manufacturing to ensure microbiological safety and extend shelf life.However, during heat treatments, the physical and chemical properties of milk are altered, which can affect the quality, flavor, and nutritional value of the final product (Krishna et al., 2021).One of the most significant changes that occur during heat treatment is the Maillard reaction (MR), which in-volves a series of nonenzymatic reactions initiated by the redox reaction between reducing sugars and ε-AA.This reaction produces Amadori products, primarily lactulosyl-lysines, in milk (Van Boekel, 1998), leading to a reduction in the availability of lysines in milk proteins and compromising the nutritional value of milk.Meanwhile, the Amadori products cannot be efficiently digested, affecting the absorption of milk proteins (Nyakayiru et al., 2020).
The extent of the different stages of the MR can be estimated using the content of furosine (Rabineau and Dhainaut, 1981), carboxymethyllysine (CML; Badoud et al., 1990, Badoud et al., 1991) and hydroxymethylfurfural (HMF ;Furth, 1988).However, all these methods encountered common disadvantages: (1) they only indicated the overall amount of MR products; (2) the conversion factors used to calculate the MR product content are inconsistent (Van Boekel, 1998).(3) they do not provide information about the modification of individual lysines and proteins.Detailed information about protein modification sites is crucial as it can help identify newly formed immunogenic and allergenic epitopes (Gazi et al., 2022).
In recent years, mass spectrometry has been used to study the lactosylation of milk proteins (Chandra Roy et al., 2020;Meltretter et al., 2020b;Wölk et al., 2020Wölk et al., , 2021)).Specifically, researchers have focused on changes in protein lactosylation under various storage conditions of different milk powder products, such as skim milk powder, whey protein concentrate, and micellular casein powder (Nasser et al., 2018;Gasparini et al., 2020a;Gazi et al., 2022).In addition, studies have also focused on investigating commercial dairy products, including raw milk, pasteurized milk, UHT milk, and milk powder, as they represent different types of heating treatments commonly used in milk processing (Milkovska-Stamenova andHoffmann, 2016a,b, 2017;Milkovska-Stamenova et al., 2017;Gasparini et al., 2020b;Meltretter et al., 2020a;Wölk et al., 2020a,b).However, it is worth noting that the specific heating conditions may differ among products from different companies, leading to a large variation in protein lactosylation even within the same type of product (Milkovska-Stamenova and Hoffmann, 2016b).Therefore, understanding the underlying mechanisms linking heating and protein lactosylation is crucial.However, this information has been lacking due to the lack of availability of heating facilities to control heating treatments.In the present study, we developed a self-designed small scale heating equipment to apply serial heating temperatures to liquid milk and investigate the effect of heating temperature on milk lactosylation using proteomics techniques.The results provide detailed information on milk protein lactosylation, which can help optimize heating processes and provide biomarkers to estimate applied heating conditions in dairy products.

Sample Collection and Heating Treatments
All procedures were conducted in accordance with the "Guiding Principles in the Care and Use of Animals" (China) and were approved by the Animal Ethics Committee of Beijing Technology and Business University.
Tank milk of 4,510 lactating cows was obtained from Sino farm with the permission of farm manager (Beijing, China).Raw milk samples were heated at 75°C for 15 s, 100°C for 10 s, 115°C for 10 s, 130°C for 10 s, and 140°C for 10 s in self-designed heating apparatus (Figure 1).The temperature was monitored by controlling the pressure (Figure 1c) and the heating time was monitored by adjusting length of holding tube (Figure 1e).All heating treatments were conducted in triplicates.

Sample Preparation
The milk samples were centrifuged at 2,000 ×g for 20 min at 4°C to remove the milk fat.Then the protein concentration was determined by using BCA assay (Thermo Scientific Pierce BCA protein assay kit).

Protein Digestion
The proteins were digested as described previously with modifications (Ding et al., 2013).Proteins (1 mg) were reduced by dithiothreitol (final concentration 5 mM in 0.05 M NH 4 HCO 3 ) and incubated at 56°C for 30 min.Subsequently, the sample was alkylated by iodoacetamide (final concentration 10 mM in 0.05 M NH 4 HCO 3 ) and incubated at room temperature in the dark for 30 min.Proteins were digested by trypsin with mass ratio of 1:100 (enzyme/protein) and incubated at 37°C overnight.The reaction was stopped by adding 1% formic acid.The digested proteins were lyophilized.

Glycated Peptides Enrichment
The enrichment of glycated peptides was performed as previously described with some modifications (Arena et al., 2010): m-Aminophenylboronic acid-agarose (Sigma) was packed into 1mL polypropylene column and equilibrated with washing buffer (250 mM ammonium acetate, 50 mM MgCl 2 , pH 8).Subsequently, 500 mg of tryptic digested peptides dissolved in washing buffer were loaded at 4°C and incubated for 1 h.Then 10 mL of washing buffer was used to wash the column.Finally, the lactosylated peptides were eluted using 250 mM acetic acid (pH 2.8).The elute was concentrated by using speed-vacuum to final volume of around 200 μL, desalted by homemade column and frozen at −80°C until analyzed.

LC-MS/MS
Enriched glycated peptides were injected to a homemade C18 pre-column (100 μm inner diameter, 360 μm outer diameter × 2 cm, 5μm 150 Å pore size, Durashell C18 particle) followed by separating over a C18 analytical column (75 μm inner diameter, 360 μm outer diameter × 15 cm, 3 μm 150 Å pore size, Durashell C18).The peptides were eluted with series linear gradient (mobile phase A: 0.2% formic acid in water; mobile phase B: 0.2 formic acid in acetonitrile) with a flow rate of 380 nL/min.The Q-Exactive Plus were set as follows: The source was run at 2.0 kV, with no sheath gas flow and the temperature of ion transfer tube was set as 350°C.The data were obtained in a dependent acquisition mode.The survey scan was from m/z 300 to 1,400 with resolution 70,000 at m/z 400.The 20 most intense peaks (≧2 + ) were acquired with HCD with normalized collision energy of 27%, fragment ions were transferred into the Orbitrap analyzer operating at a resolution of 17,500 at m/z 400.

Data Analysis
All MS/MS spectra were analyzed by MaxQuant 1.6.14.0.The reference proteome of Bos taurus from Swiss-Uniprot was used as the protein database with reverse sequences generated by Max-Quant.Carbamidomethylated cysteine was set as fixed modification; oxidation of methionine, N-terminal acetylation, deamidation of asparagines or glutamine and lactosylated lysines (+324.10)were set as variable modifications for both identification and quantification.A mass deviation of 0.5 Da was set as maximum allowed for MS/ MS peaks, and a maximum of 2 missed cleavages was allowed.Maximum false discovery rates were set to 1% both on peptide and protein levels.Minimum required peptide length was 6 AA for both identification and quantification.A minimum of 2 peptides for each protein were required for reliable identification and quantification.

Furosine Analysis
The content of furosine was detected according to Agricultural Industry standard of People republic of China (NY/T 939-2016) with some modifications.Briefly, 6 mL of 6 M HCl were added to 2 mL of sample and the sample tube was filled with nitrogen and sealed.It was heated at 110°C for 24 h and filtered.The protein content in filtrate was analyzed by Kjeldahl method.Sep-pak C18 cartridge was used to enriched furosine and 3 mL of eluents washed by 3M HCl was analyzed by HPLC using Sunfire C18 column (5 μm 100 Å pore size, 4.6 × 250 mm) with series gradients (mobile phase A methanol, mobile phase B 0.1% TFA in water).

Solvent Accessibility Calculation
The solvent accessibility was predicted by Sable II (Adamczak et al., 2005).

Statistical Analysis
The one-way ANOVA and Tukey's test were applied to compare means among different treatments (Graph-Pad Prism 8.4.3).

The Effect of Temperature on Lactosylated Protein and Peptides
In present study, we identified 123 lactosylated peptides in 65 milk proteins.The number of lactosylated lysines and proteins ranged from 26 to 38 and 15 to 21, respectively, in milk samples heated at temperatures lower than 140°C.In control group, which is raw milk without heating, we observed 26 lactosylated sites in 15 proteins, including K24, K46, K86 in κ-CN, K47 and K48 in β-LG, K83 and K126 in α-LA, K7, K34, K105, K132 in α S1 -CN, K32 in β-CN, and K32 and K113 in α S2 -CN and so on.These lysines could be lactosylated during milking and transporting at room temperature, which indicated the susceptibility of these sites to temperature.Compared with raw milk, heating at 75°C to 130°C increased the number of lactosylated sites and proteins.But the increase was not significant (Figure 2a).When the temperature increased to 140°C, 94 lysines from 43 proteins were lactosylated in milk, which was almost 3 times higher than those of the other heated milk samples (Figure 2a).Of note, 21 lactosylated proteins, including osteopontin (SPP1), IGHM protein (IGHM), glycoprotein 2 (GP2), and xanthine dehydrogenase/oxidase (XDH), and so on, were specifically identified when heating temperature increased to 140°C (Table 1).These were mainly low abundant proteins in milk.In raw milk, we observed 1 to 4 lactosylated sites per protein (Figure 2b).We detected 4 lactosylated lysines in α S1 -CN (Table 1).Proteins with more than 5 lactosylated lysines were detected when the samples were heated at 100°C and above.We observed proteins with 7 to 9 lysines lactosylated in samples heated at 140°C, including 9 lactosylated lysines in β-LG, 8 lactosylated lysines in serum albumin and 6 lactosylated lysines in κ-CN and lactoferrin (Table 1).Furosine content, the conventional indicator of lactosylation level, was analyzed in milk heated at different temperatures (Figure 2c).It is significantly different in milk heated at 140°C for 10 s as compared with the other samples.In addition, we summed the intensity of lactosylated peptides to estimate the extent of lactosylation (Figure 2d).We found similar results between summed lactosylated peptides intensity and furosine content.The level of lactosylation of proteins significantly increased when the heating temperature increased to 140°C.And it is not different among samples heated at lower temperatures (Figure 2c and 2d).

Lactosylation Profiling of Individual Proteins and Lysines
The percentage of lactosylated lysines of individual protein increased as the increase of heating temperature (Table 1).In the present study, the heating treatment resulted in lactosylation of 60% (6 out of 10) lysine residues in κ-CN, followed by 56.25% (9 out of 16) lysine residues in β-LG, 50% (6 out of 12) lysine residues in α-LA, 33.33% (5 out of 15) lysine residues in α S1 -CN, and 33.33% (4 out of 12) lysine residues in β-CN.In addition to these major milk proteins, lactadherin, one of the proteins in milk fat globule membrane, showed lactosylation of 31.25% (5 out of 16) lysine residues, with 4 of them only lactosylated at 140°C.Moreover, lactotransferrin had 12.72% (8 out of 56) of its lysine residues lactosylated when the heating temperature was above 115°C.
The lactosylation condition of individual lysines varied in each protein (Table 1).In κ-CN, K86 was lactosylated in raw milk.K63 started to be lactosylated at 75°C for 15 s.K111 and k112 can only be lactosylated at 140°C.In β-LG, K47 was lactosylated in raw and heated milk samples.K117 was lactosylated in heated milk samples.K91 and K60 can be lactosylated when the heating temperature is higher than 100°C.K14 and K100 were identified to be lactosylated at 140°C.K83 in α-LA was lactosylated in raw and heated milk samples.In α S1 -CN, K7, K34, and K105 were lactosylated in milk of raw and heated milk.K32 of α S2 -CN and K32 of β-CN were lactosylated in raw and heated milk samples.The accessibilities of lysines to lactosylation varied within protein.We tried to evaluate the solvent accessibilities of all lactosylated lysines using Sable II.However, we could not find a clear trend in present study.

Markedly Number of Proteins Influenced at Higher Temperature
Large number of newly identified lactosylated sites and proteins can only be observed in milk samples heated at 140°C.These proteins included xanthine      (Lu et al., 2014).It indicated that the proteins in lower abundance, mainly milk fat globule membrane proteins, were more thermo-stable to lactosylation than major whey proteins.The function of these proteins is important and diverse including host defense, lipid transport and lipid metabolic process.The importance of milk fat globule membrane proteins in human health especially for infant nutrition is emerging (Brink and Lönnerdal, 2020;Kosmerl et al., 2021) (Kosmerl et al., 2021) and MFGM proteins are now used as functional ingredients in food industry (da Silva et al., 2021;Wang et al., 2022).However, the effect of lactosylation on the functionality of these proteins during heat treatment has not been well investigated.It is already known that the lysines could be blocked by lactosylation and the blockage influences the nutritional value of proteins (Van Boekel, 1998;Zenker et al., 2020).The attachments of lactose on lysines hinder the access of proteases on proteins (Kastrup Dalsgaard et al., 2007;Zenker et al., 2020), which could influence the release of functional peptides and induce the potential allergen of long-length peptides.Thus, it is important to find a temperature that could eliminate the microbiological hazard without intensive lactosyaltion of proteins.This temperature could between 130°C to 140°C based on our observation.However, these need to be further investigated.

The Possible Indicator of Heating Process
The identification of reconstitute milk and the differentiation of pasteurized and UHT milk are of great importance in dairy markets.The change of lactosylated proteins and lysines could be used as an indicator for the harshness of heating.In present study, K14 in β-LG can only be lactosylated at 140°C.In constant, K14 was also observed in UHT milk and products undergo severe heating treatments but not in raw, pasteurized and ESL milk (Meltretter et al., 2014).The detection of lactosylation in K14 of β-LG could be used to distinguish pasteurized and UHT milk or the adulteration of UHT milk in pasteurized milk.Another possible indicator could be lactotransferrin, which is robust to heat treatments (Dyer et al., 2016).In present study, the most sensitive lysines in lactotransferrin were modified at 115°C.Five lysines were only modified at 140°C.The lactosylation of lactotransferrin might be used to indicate the harshness of moderate heat treatment.And the lactosylation of K28/K174/K301/K520/K637 could be used to indicate the severe heat treatment of milk.K28 was also proposed as a sensitive marker to track and evaluate the modification of lactotransferrin during processing and food preparation (Dyer et al., 2016).The peptides derived from major proteins have been extensively studied and proposed as indicators of heating treatments.(Pinto et al., 2012;Meltretter et al., 2014;Dyer et al., 2016).However, the 26 newly lactosylated proteins at 140°C could also be valuable indicators for the severe heat treatments.Considering the feasibility of analysis, proteins with higher abundance among these 26 proteins might be better candidates, such as xanthine dehydrogenase/oxidase, lactoperoxidase and perilipin.And, this needs to be further investigated.Lactosylated peptides or proteins have been suggested as an alternative indicator of heat treatment, instead of the widely used furosine content.Furosine is an artificial AA that is derived from the acid hydrolysis of Amadori products, and it is an indirect indicator of MR products (Van Boekel, 1998).However, the conversion rate of Amadori products to furosine varies between 30 and 40% depending on the type of Amadori product and the consistency of analytical conditions (Erbersdobler and Somoza, 2007;Aalaei et al., 2019), which limits its accuracy.Although it is worth noting that the detection of furosine content is easier and requires less sophisticated equipment compared with mass spectrometry, we analyzed Amadori products directly and identified possible indicators for heat treatment that may provide more accurate readouts.

CONCLUSIONS
The lactosylation of milk proteins was investigated under serial heat treatments, to elucidate the effect of heating on protein modifications.The number of lactosylated proteins and lysines increased significantly when the heating temperature increased to 140°C.We found that the major milk proteins including κ-CN, β-LG, and α-LA were lactosylated at lower temperature, while the low abundant proteins were lactosylated at higher temperature, mainly under conditions comparable to normally applied UHT treatment in dairy industry.K14 in β-LG, 5 lactosylated lysines in lactotransferrin, and several low abundant proteins could be the indicators for the harshness of heat treatments in milk processing.By employing proteomic techniques, we conducted a comprehensive analysis of lactosylated proteins and their sites, which provided valuable insights into protein modifications during heat treatment.This information can be useful in developing appropriate thermal-treatment protocols for the processing of dairy products.

Figure 2 .
Figure 2. The relationship between milk protein lactosylation and heating treatments.(a) Number of lactosylated lysines and proteins.(b) Distribution of singly and multiply lactosylated proteins.(c) Furosine content of milk samples.(d) Summed intensity of lactosylated peptides.Error bars represent the standard deviation of triplicate experiments.
Identified lactosylated proteins, listed as descending percentage of lactosylated lysines in each protein Continued Lu et al.: HEAT TREATMENT AND MILK PROTEIN LACTOSYLATION

Table 1 .
Lu et al.:HEAT TREATMENT AND MILK PROTEIN LACTOSYLATION Identified lactosylated proteins, listed as descending percentage of lactosylated lysines in each protein Gene name

Table 1 (
Continued).Identified lactosylated proteins, listed as descending percentage of lactosylated lysines in each protein dehydrogenase/oxidase, perilipin, acetyl-CoA carboxylase, and apolipoprotein C-III et al., which are mainly milk fat globule membrane proteins Lu et al.: HEAT TREATMENT AND MILK PROTEIN LACTOSYLATION