Glycation as a Tool To Probe the Mechanism of Î²-Lactoglobulin

Article pubs.acs.org/JAFC

Glycation as a Tool To Probe the Mechanism of β‑Lactoglobulin Nanofibril Self-Assembly Anant C. Dave,† Simon M. Loveday,*,† Skelte G. Anema,†,§ Geoffrey B. Jameson,†,‡ and Harjinder Singh† †

Riddet Institute and ‡Institute of Fundamental Sciences, Massey University, Private Bag 11222, Palmerston North, New Zealand § Fonterra Research and Development Centre, Dairy Farm Road, Private Bag 11029, Palmerston North, New Zealand S Supporting Information *

ABSTRACT: In this study we investigated the effects of different levels of glucosylation and lactosylation on β-Lg self-assembly into nanofibrils at 80 °C and pH 2. Fibrils in heated samples were detected with the thioflavin T assay and transmission electron microscopy, while SDS−PAGE was used to investigate the composition of the heated solutions and fibrils. Glycation had different effects in the nucleation and growth phases. The effect of glycation on the nucleation phase depended on the degree of glycation but not the sugar type, whereas both the type of sugar and the degree of glycation affected the rate of fibril growth. Glycation by either sugar strongly inhibited self-assembly in the growth phase, and lactosylation produced a much stronger effect than glucosylation. We suggest that the varying glycation susceptibility of different lysine residues can explain these observations. The large, polar sugar residues on the glycated fibrillogenic peptides may inhibit fibril assembly by imposing steric restrictions and disrupting hydrophobic interactions. KEYWORDS: β-lactoglobulin, nanofibril, self-assembly, glucosylation, lactosylation, dry glycation, circular dichroism, mass spectrometry

■

INTRODUCTION Self-assembled nanofibrils from bovine β-lactoglobulin (β-Lg) have recently been the subject of extensive investigations owing to their potential applications in biotechnology, e.g., for enzyme immobilization,1 in nanoelectronics and biosensors,2 or as food ingredients with modified functionality.3,4 Most studies have produced fibrils by heating β-Lg above its denaturation temperature, at a low pH and low ionic strength. Several studies have examined the effects of the solution conditions on self-assembly kinetics and fibril morphology.5−9 Another approach, which has largely been unexplored, is to investigate the self-assembly of chemically modified β-Lg. Chemical modification of β-Lg involves the interactions of reactive functional groups of amino acid side chains, e.g., amine groups of lysine and arginine, and this reaction has been used as a tool to modify the functional properties of β-Lg.10−12 Bovine β-Lg is one of the major globular proteins of whey and mainly exists in polymorphic forms A and B. It has a chain length of 162 amino acid residues and an approximate molecular mass of 18.3 kDa. Structural studies on β-Lg13,14 have shown that it has a well-defined secondary and tertiary structure and mainly exists as a monomer at pH 2 and low ionic strength,15 showing high heat stability.16 Self-assembly into fibrils at pH 2 and 80 °C follows a nucleation-dependent polymerization mechanism with a characteristic three-phase sigmoidal growth curve consisting of the lag, growth, and stationary phases.17 During the lag phase β-Lg monomers first unfold rapidly18 and undergo heat-induced acid hydrolysis.19−21 The hydrolysis of monomers imparts greater conformational freedom to fibrillogenic regions, allowing their self-assembly into assembly-promoting nuclei.18 The exact structure of the nuclei and the interactions involved in their formation are not yet fully understood, perhaps due to the reversibility of selfassembly during the lag phase.22 © 2014 American Chemical Society

However, several studies have shown that the time required for nucleation can be greatly reduced by varying the parameters of self-assembly. Factors that can reduce the nucleation time include higher temperatures,21 microwave radiation,23 divalent cations such as Ca2+,7,8,24 secondary nucleation either by shearing while heating5,6 and/or by seeding with preformed fibrils.3,5 In the growth phase, monomers and larger peptides continue to undergo hydrolysis during heating, resulting in peptides that can self-assemble onto the growing fibrils. However, peptides do not undergo further hydrolysis once incorporated into fibrils.18 It is not known what effect chemical modifications of different amino acid side-chain functional groups of β-Lg, e.g., glycation of lysine residues, may have on its self-assembly. Glycation of β-Lg by the Maillard reaction or nonenzymatic browning involves the interaction of the reactive side-chain amine groups with the free carbonyl groups of reducing sugars. The details of the various steps involved in Maillard reaction of milk proteins have been reviewed and described previously.25 βLg has 19 free amine groups that are available for Maillard reactions: 15 from lysine, 3 from arginine residues and the Nterminal amine group. Studies using lactosylation have shown that lysine residues at positions 47,26−28 91,27−29 and 10026 are particularly susceptible to glycation. Glycation of β-Lg can be achieved by incubating the protein−sugar mixture at pH 7 under controlled conditions in the dry state or aqueous state. Glycation in the dry state, also referred to as solid-state glycation, is much faster than aqueous glycation.27,29−31 Received: Revised: Accepted: Published: 3269

December March 11, March 12, March 12,

3, 2013 2014 2014 2014

dx.doi.org/10.1021/jf405441g | J. Agric. Food Chem. 2014, 62, 3269−3278

Journal of Agricultural and Food Chemistry

Article

Solid-state glycation of β-Lg conserves its native-like structure29,30,32 and its association tendency29 and yields a heterogeneous mixture of glycoforms that differ in the degree and sites of glycation.33 In contrast, aqueous glycation induces significant changes in the native β-Lg structure29,34 and leads to sugar-induced protein cross-linking.27,29,35 The effects of β-Lg glycation include improved heat stability,10,36−38 lowered isoelectric point,39,40 decreased surface hydrophobicity due to polar sugar groups,10,37 and inhibition of aggregation due to the steric effect of sugar residues.37,38 Modification of the amine groups of amyloidogenic proteins α-synuclein41 and insulin42 by dicarbonyl compounds slowed their self-assembly into fibrils in vitro. On the other hand, glycation promoted the self-assembly of bovine serum albumin.43 Glycation is likely to affect β-Lg self-assembly since almost all of the potential glycation sites are located on the regions involved in fibril formation.18,20,23 Liu and Zhong44 recently prepared nanofibrils from whey protein isolate after solid-state incubation with lactose. However, the incubation produced changes in secondary structure, and Maillard reactions progressed well beyond the addition of integer numbers of sugars, due to the extreme incubation conditions (80 °C for 2 h at 70% relative humidity). Lactose-reacted β-Lg fibrils had modified functional properties, but little mechanistic information could be obtained due to the heterogeneity of the starting material. In this study, we focused on how glycation affects the structural and chemical events involved in fibril formation and examined both the extent of glycation and the sugar type (monosaccharide or disaccharide). We glycated β-Lg under mild conditions to prevent advanced Maillard reactions and produce a well-defined starting material without marked changes to the secondary or tertiary structure of the protein. This systematic approach has produced new insights into the mechanism of amyloid-like fibril self-assembly.

■

glycation reaction, the samples were dissolved in Milli-Q water (pH 7) and dialyzed (molecular mass cutoff of 6−8 kDa, Spectra/Por, Spectrum Laboratories Inc., Rancho Dominguez, CA) for 48 h at 4 °C to remove the excess unreacted sugar. The dialyzed protein solution was then freeze-dried to obtain the glycated β-Lg. Determination of the Degree of Glycation. Glycated β-Lg was dissolved in water (pH 2) and filtered using a 0.2 μm syringe filter before being subjected to RP-HPLC (C18, Jupiter, Torrance, CA) at a flow rate of 1 mL/min. The compositions of the buffers were 10% (v/ v) acetonitrile (MeCN) in water containing 0.1% trifluoroacetic acid (TFA) for buffer A and 90% MeCN with 0.08% TFA for buffer B. Upon loading, the sample was eluted using the following program: 2 min wash with buffer A followed by a linear gradient of buffer B from 0% to 65% in 13 min and then to 100% B in 1 min. This was followed by isocratic elution of buffer B for 5 min and then a gradient to 0% B in 1 min, after which a wash with buffer A for 2 min followed. A single large peak was eluted when the MeCN concentration was 38%, and this peak was collected and concentrated to approximately 100 μL (Savant SpeedVac SC100, Holbrook, NY) under vacuum (Varian DS 102, Varian Vacuum Technologies, Torino, Italy). The samples were stored at −86 °C until further analysis. The degree of glycation in modified β-Lg was analyzed by liquid chromatography−mass spectrometry. RP-HPLC-purified samples (2 μL) were autoinjected into an Agilent 1200 LC system (Agilent Technologies, Santa Clara, CA) on bypass mode. The isocratic mobile phase was composed of 0.1% (v/v) formic acid/50% MeCN/water at a flow rate of 100 μL/min. The LC eluant was directed to the electrospray ionization (ESI) source at a capillary voltage of 3.4 kV of the mass spectrometer (Agilent 6520 QTOF, Agilent Technologies, Hanover, Germany). Sample analysis was performed in positive ion mode, and the data were stored in profile mode. MS scans (100−1700 m/z) were obtained with fragmentor and skimmer voltages at 175 and 65 V, respectively. Other source parameters were a drying gas temperature of 325 °C, gas flow rate of 6 L/min, and nebulizer gas pressure of 30 psig. The total ion chromatograms (TICs) obtained were examined and spectra deconvoluted using Agilent MassHunter Workstation Qualitative Analysis software (version B.03.01) with a BioConfirm plugin (Agilent Technologies, Santa Clara, CA). The intensity of the peaks was normalized by assigning the most intense peak in the spectrum to 100%. Preparation of Samples for Self-Assembly Studies. The unmodified β-Lg showed only two peaks corresponding to variants A and B (Figure S1, Supporting Information). The degrees of glycation for β-Lg with glucose and lactose after different glycation times are shown in deconvoluted MS spectra in Figures S2 and S3 (Supporting Information), respectively. The degrees of glycation after 8 h of glucosylation and 3 days of lactosylation were similar, and these glycation times were selected for further investigation. From this point onward, samples made using these heating times will be referred to as low-glucose (LowGlu) and low-lactose (LowLac) β-Lg, respectively. Similarly, for the high degree of glycation, the durations selected were 36 h for glucosylation and 7 days for lactosylation. The samples made using these heating time points will be referred to as high-glucose (HighGlu) and high-lactose (HighLac) β-Lg, respectively. Fresh glycated samples were made using the selected glycation time points and analyzed by MS using the above methodology. For each of the glycation mixtures, a β-Lg sample without sugar was used as a control. The freeze-dried glycated proteins prepared as described above were stored at −18 °C until further analysis. Self-Assembly from β-Lg. β-Lg samples intended for selfassembly were made according to a method described previously.24 Protein concentrations in the samples were determined by spectrophotometry (Ultraspec 2000, Pharmacia Biotech, Cambridge, U.K.) using an extinction coefficient of 0.94 cm2/mg.45 β-Lg (with or without glycation; 1%, w/v) samples, at pH 2.00 ± 0.02 were heated in a temperature-controlled water bath (Jeio Tech, Seoul, South Korea) maintained at 80.0 ± 0.2 °C. The control sample used in the study was untreated and unglycated β-Lg. Since we were interested in investigating the effects of glycation on nucleation, heating was done

MATERIALS AND METHODS

Unless otherwise mentioned, all chemicals were of analytical grade and were purchased from Sigma-Aldrich (St. Louis, MO). Unless otherwise stated, Milli-Q water was used for preparing the reagents. Isolation of β-Lactoglobulin. β-Lg was isolated in-house from whey protein isolate 8855 (Fonterra Cooperative Ltd., Auckland, New Zealand) using the method described previously.18 The β-Lg powder had a protein content of 97% and was a mixture of variants A and B, which did not have any glycation (Figure S1, Supporting Information). Preparation of Glycated β-Lg. Preliminary Experiments for Selection of Glycation Times. A stock solution of β-Lg (pH 7) was allowed to hydrate at 4 °C overnight. The solution was then centrifuged at 44000g for 30 min at 20 °C, and the supernatant was filtered using a syringe filter (pore size 0.2 μm, Minisart CE, Sartorius Stedim Biotech GmbH, Goettingen, Germany). Glucose and lactose powders were mixed with the filtered β-Lg solution (molar ratio βLg:sugar =1:100), and the pH values of the solutions were readjusted to 7.00 ± 0.02 using 1 N NaOH. The samples were then frozen at −18 °C overnight and freeze-dried. For glycation, the freeze-dried powders were placed in a glass Petri dish and transferred to separate desiccators with a saturated potassium bromide (KBr) solution. The desiccators were sealed, and a vacuum of 125 mbar was applied to them. Then the desiccators were transferred to an oven maintained at 40.0 ± 0.5 °C. The saturated KBr solution gave a relative humidity of 80% inside the desiccators. For glucosylation, the samples were removed from the desiccators after 8, 16, or 36 h, while for lactosylation the durations were 3, 5, and 7 days. Longer heating times were not investigated since samples heated for longer than the above-mentioned times showed extensive brown discoloration indicating the presence of advanced Maillard reaction products in the samples. After completion of the 3270



Article

Figure 1. Deconvoluted MS spectra of glucosylated (A) LowGlu β-Lg and (B) HighGlu β-Lg and lactosylated (C) LowLac β-Lg and (D) HighLac β-Lg. Peak labels show the number of attached glucose or lactose residues. without agitation to minimize secondary nucleation, which would otherwise obscure other phenomena. After being heated for different

holding times (2−24 h), the samples were cooled rapidly using an ice bath. The cooled samples were analyzed by the thioflavin T assay and 3271



Article

SDS−PAGE. The lactose used in the study was found to be stable under conditions of glycation and fibril formation (data not shown). Thioflavin T (ThT) Fluorescence Assay. Fibril formation from βLg was studied by the ThT assay as previously described.24 The fluorescence intensities in the samples were corrected against that of the unheated sample and ThT solution without any protein sample. The corrected intensities were then fitted using the mathematical model given in the following equation: β γ

ft = α −

1+

β αγ

λ = [θ ]M

⎛ Δ[θ ]t ⎞ ⎜ ⎟ ⎝ Δ[θ ]0 ⎠ (1)

⎞ αγ 1 ⎛ ⎛ αγ ⎞ ⎜ln⎜ ⎟ − 4 + 2⎟ β + αγ ⎝ ⎝ β ⎠ β + αγ ⎠

(

ln 2 + t1/2max =

⎛ df ⎞ ⎜ ⎟ = ⎝ dt ⎠max

αγ β

(

β γ

(2)

(3)

)

+ α (β + αγ ) 4

(4)

where tlag represents the duration of the lag phase, t1/2max is the time required for attaining half of the maximum fluorescence value, and (df/dt)max represents the maximum rate of fluorescence increase. The value of the empirical constant α represents the fitted maximum fluorescence in the stationary phase, and is relabeled f max. Equation 1 was fitted to data using SigmaPlot 12.1, and the standard deviations of tlag, t1/2max, and (df/dt)max were calculated using partial derivatives and the covariance matrix, as detailed in the Supporting Information. Separation of Fibrils. Fibrils in the samples were separated using ultracentrifugation and washed using the method described previously.18 Briefly, 9 mL volumes of heated and cooled samples were subjected to ultracentrifugation at (2.4 × 105)g for 1 h. The pellets obtained were dispersed in 9 mL of water and vortexed, followed by ultracentrifugation. These washing and ultracentrifugation steps were repeated twice. The washed pellet was used for analysis. SDS−PAGE. Heated samples were analyzed by tricine SDS−PAGE methodology described elsewhere.18 Heated samples were diluted 1:10 in reducing PAGE sample buffer (4% SDS and 100 mM dithiothreitol (DTT)). For centrifuged samples, the supernatants were diluted 1:10 while the pellets were suspended directly in the PAGE sample buffer. The PAGE gels were scanned and analyzed using Image Lab software (Bio-Rad Laboratories, Hercules, CA). The quantification of the hydrolysis of β-Lg was done as per the method described in the reference. For each of the samples, the hydrolysis rate constants (kh) were calculated from the following equation: I = I0e−kht

293

[θ ]t − [θ ]60 [θ ]0 − [θ ]60

(7)

(8)

θ0 represents the ellipticity in the unheated sample, θt that at time t (min) during heating, and θ60 that at the end of the heating period after 60 min. Transmission Electron Microscopy (TEM). Fibrils in heated solutions were separated using the methodology previously described.24 A 200 μL volume of the heated solution was diluted 10-fold using water (pH 2) and centrifuged at 3000g for 15 min in a centrifuge filter (Amicon Ultra-4, Millipore) with a nominal cutoff of 100 kDa, rinsed previously with water (pH 2). The sample was washed with pH 2 water three times and recentrifuged to get fibrils in the retentate. The fibrils in the retentate were then suspended in 1 mL of the pH 2 water. For TEM, 200 μL of the suspension was mixed with 200 μL of water (pH 2) and 20 μL of BSA (0.5%, w/v) in a polypropylene tube before being stained by a negative staining protocol that is described in the reference.

)

β + αγ

=

⎛ Δ[θ ]t ⎞ [θ ]t − [θ ]0 ⎜ ⎟ = [θ ]60 − [θ ]0 ⎝ Δ[θ ]60 ⎠208

where α, β, and γ are empirical constants and f t denotes fluorescence at a given time t (h). The kinetic parameters for β-Lg self-assembly for each of the samples were calculated using the following equations:24

tlag =

(6)

where θλ is the measured ellipticity (deg) at a given wavelength λ, L represents the path length of the cuvette used (cm), and C is the protein concentration (dmol/L). The data were smoothed in SigmaPlot 12.1, and the relative changes in ellipticities at 293 and 208 nm were calculated from the following equations:18

+α

exp[t(β + αγ )]

100θλ LC

■

RESULTS Characterization of Glycated β-Lg. The deconvoluted MS spectra of glycated β-Lg are shown in Figure 1. All samples showed a series of peak pairs with mass increments of either 162 Da (Figure 1A,B, glucosylation) or 324 Da (Figure 1C,D, lactosylation), indicating that both variants B (18277 Da) and A (18364 Da) participated in glycation. Peaks corresponding to unglycated β-Lg appeared only in the scans of LowLac β-Lg, and at low levels in that case. Variants A and B showed a similar propensity for glycation, and the glucosylation of both variants was faster than lactosylation. The presence of multiple peaks in the spectrum suggests heterogeneity of glycoforms of β-Lg in samples with respect to the number of sugar residues attached to the β-Lg. These results are in agreement with previous studies which show similar reactivities of β-Lg A and B,47 a heterogeneous mixture of glycoforms in samples upon glycation with respect to the number of attached sugar residues,27,30,33,36,39 and faster rates of glucosylation than those of lactosylation of β-Lg.10 Unfolding Behavior of Glycated β-Lg. The effect of glycation on the β-Lg native structure and thermal unfolding was investigated using CD spectroscopy with in situ heating at 80 °C, as described previously.18 In the NUV region, all unheated samples showed characteristic troughs at 293 and 286 nm (Figure S4A, Supporting Information) arising due to the tryptophan residue at position 19,48 indicating that the environment around Trp19 was unaffected by glycation. In addition, the scans showed a small trough at 265 nm contributed by the phenylalanine residues in the native β-Lg structure. These troughs in the NUV region have been shown to rapidly decrease in intensity upon heating β-Lg under fibrilforming conditions.18 Heating of glycated samples resulted in rapid loss of the characteristic troughs at 293, 286, and 265 nm. Figure 2A shows an example of scans recorded for LowGlu β-

(5)

Circular Dichroism (CD) Spectroscopy. The β-Lg solutions for CD spectroscopy were prepared in HPLC-grade water at pH 2.00 ± 0.02 according to the method described in a previous report.18 The βLg concentrations for near-UV (NUV) scans (260−320 nm) and farUV (FUV) scans (180−250 nm) were 1 and 0.01 mg/mL. Scans were recorded using a CD spectrometer (Chirascan, Applied Photophysics Ltd., Surrey, U.K.) and a cuvette of path length 10 mm either at 20 °C or continuously for 1 h at 80 °C. The time required for raising the temperature from 20 to 80 °C was excluded, and the time at which the temperature reached 80 °C was designated as 0 min. Scans were averaged sequentially to obtain data at 2 min intervals and corrected by subtracting baseline scans from each of the sample scans. Molar ellipticity was calculated using the following equation46 with data from the β-Lg scans after baseline subtraction: 3272



Article

Lg during heating, and Figure 3A shows the comparison of relative loss of ellipticities at 293 nm upon heating.

Figure 3. Relative change in ellipticity at (A) 293 nm in the NUV region and (B) 208 nm in the FUV region. Solid lines are for visual reference. For details of the y-axis parameters, see the Materials and Methods.

Figure 2. CD spectra of LowGlu β-Lg in the (A) NUV region (10 mg/ mL β-Lg) and (B) FUV region (0.01 mg/mL β-Lg). The numbers in the legends indicate the heating time in minutes.

In the FUV region, glycated samples showed a broad trough with a minimum at 217 nm (Figure S4B, Supporting Information), indicating that in all the samples β-sheets were the dominant secondary structural conformation.48 It has been shown that heating β-Lg at 80 °C and pH 2 resulted in a shift of the trough toward lower wavelengths,18 representing unfolding of the secondary structure. All glycated samples showed a rapid shift in the trough during heating. Figure 2B illustrates the shift of the trough in the FUV region for LowGlu β-Lg during heating. The plot of ([Δθt]/[Δθ60])208 against time comparing the change in the ellipticity at 208 nm upon heating is shown in Figure 3B. Heat-Induced Hydrolysis. To investigate the effect of glycation on the acid hydrolysis of β-Lg, the composition of samples heated for different times was analyzed by tricine SDS−PAGE under reducing conditions. All samples showed hydrolysis of the β-Lg monomer with heating. A comparison of PAGE profiles of samples after heating for 12 h at 80 °C is shown in Figure 4. The peptide bands in glycated samples appeared less distinct than in control β-Lg. This may be attributed to the presence of a heterogeneous mixture of peptides with different degrees of glycation in the samples. In addition, the modification of amine groups on peptides may

Figure 4. Tricine SDS−PAGE profiles of control and glycated samples after heating for 12 h at 80 °C. Key: M0, molecular mass markers (kDa); C, heated control.

have altered the affinity of these peptides to bind to the staining dye.49 Among the glycated samples, the peptide bands in lactosylated samples were the least well-defined. The intensity of the band corresponding to the unhydrolyzed β-Lg in the SDS−PAGE gels at different times was quantified and plotted against the heating times. Hydrolysis of the monomer in all samples followed first-order exponential decay kinetics (Figure S5, Supporting Information), and the rate 3273



Article

constants (kh) were calculated using eq 5; these are listed in Table 1. Glycation reduced kh by ∼15%, and this effect was Table 1. Rate Constants of Acid Hydrolysis of β-Lg Calculated from Eq 5 sample details control LowGlu β-Lga HighGlu-β-Lg LowLac β-Lga HighLac β-Lg

kh (min−1 × 10−3) 2.8 2.4 2.3 2.3 2.5

(0.10) (0.08) (0.07) (0.10) (0.11)

b

adj R2

Sy/xc × 10−3

0.9921 0.9925 0.9752 0.9609 0.9746

0.50 0.41 0.40 0.55 0.60

The degrees of glycation in LowGlu and LowLac β-Lg were similar. Numbers in parentheses indicate standard deviations. cStandard error of regression. a b

independent of the type of sugar or the degree of glycation. However, the glycation effect was not significant (p = 0.29) in the likelihood ratio test. The SDS−PAGE analysis of β-Lg glycation controls showed that hydrolysis rates were unaffected (p = 0.25, likelihood ratio test) by sample treatments in the absence of sugars (Figure S6 and Table S1, Supporting Information). Self-Assembly from Glycated β-Lg. Self-assembly from glycated samples followed a three-phase sigmoidal growth pattern (Figure 5A), which is typical of β-Lg self-assembly under unstirred conditions. The ThT data were fitted well by eq 1, and results from a likelihood ratio test confirmed that the effect of glycation treatments on the parameters was statistically significant (p < 0.0001). The kinetic parameters of selfassembly calculated from eqs 2−4 are listed in Table 2. A comparison of kinetic parameters of self-assembly is shown in Figure 5B. LowGlu β-Lg and LowLac β-Lg showed only a small increase in tlag (6 and 7 h, respectively, against 5.6 h for the control), whereas in HighGlu β-Lg and HighLac β-Lg the tlag values were nearly twice that of the control (11.3 and 11 h, respectively). In the growth phase, glycation greatly decreased (df/dt)max. For glucosylated samples, this effect was more pronounced at high degrees of glycation. In contrast, lactosylation had a strong effect irrespective of the degree of glycation. The effect of glycation on the stationary phase fluorescence (f max) followed a similar trend. Control β-Lg samples were freeze-dried, incubated, and dialyzed as with glycated samples but without sugars, and these controls self-assembled normally (Figure S7, Supporting Information). Thus, glucosylation and lactosylation had different effects on the different stages of self-assembly, and the effect of lactosylation was stronger than that of glucosylation. Composition of Fibrils. Fibrils in samples heated for 24 h were separated using ultracentrifugation and analyzed using SDS−PAGE under reducing conditions. The residual ThT intensities in the supernatant samples remained low, indicating complete separation of fibrils upon ultracentrifugation (Figure S8, Supporting Information). HighGlu and HighLac β-Lg showed long tlag and lower (df/dt)max, so a heating time of 24 h was selected for studying the composition of the fibrils. Figure 6 shows the peptide bands in centrifugally-separated fibrils from glycated β-Lg. Figure 4 shows the results of analyzing noncentrifuged solutions, which contain both fibril and nonfibril material, whereas only the fibril fraction (the pellet) was analyzed in Figure 6. All pellet samples contained only peptides, indicating that the building blocks of fibrils were

Figure 5. Self-assembly from glycated β-Lg. (A) Thioflavin T fluorescence intensity of control and glycated β-lg (1%, w/v) during heating at pH 2 and 80 °C. Each data point is an average of triplicates from two independent experiments, and error bars are standard errors. Solid lines indicate the fit to eq 1. (B) Normalized parameters of selfassembly calculated from eqs 2−4. The kinetic parameters of individual samples were normalized by those of the control sample.

predominantly peptides under the conditions used here (1% (w/v) protein, pH 2, heating at 80 °C). The unmodified control β-Lg showed seven peptide bands. The same bands appeared in all glycated samples, but they were less distinct in glycated samples and showed slightly reduced mobility, indicating that these peptides may be glycated. The lighter overall intensity in LowLac β-Lg and HighLac β-Lg lanes may result from lower fibril yields (Figure 5). Morphology of Fibrils. The morphology of fibrils from glycated β-Lg was examined by TEM (Figure 7). The morphology of the fibrils was unaffected by glycation. The fibrils appeared long, straight, and unbranched, similar to those in unmodified β-Lg (Figure S9, Supporting Information), irrespective of glucosylation or lactosylation. Furthermore, all samples showed both long and short fibrils.

■

DISCUSSION The CD studies showed that glucosylation and lactosylation had only very minor effects on the native structure of β-Lg. The structural changes in β-Lg upon heating were very fast, and CD results did not show differences between control and/or 3274



Article

Table 2. Kinetic Parameters Describing β-Lg Self-Assembly Calculated Using Eqs 1−4 sample details control LowGlu β-Lga HighGlu-β-Lg LowLac β-Lga HighLac β-Lg a

tlag (h) b

5.6 (0.2) 6.0 (0.2) 11.3 (0.7) 7.0 (0.8) 11.0 (0.4)

(df/dt)max (FU·h−1)

t1/2max (h)

f max (FU)

158.0 (8.4) 56.0 (2.9) 51.0 (10.7) 15.5 (2.0) 32.7 (4.8)

9.1 (0.1) 10.7 (0.2) 14.3 (0.4) 12.0 (0.5) 13.8 (0.3)

1117 (16.0) 523 (6.7) 305 (17.0) 155 (7.0) 178 (6.0)

The degrees of glycation in LowGlu and LowLac β-Lg were similar. bNumbers in parentheses indicate standard deviations.

number of lysine residues modified (Figure 1). The longer tlag values for HighGlu β-Lg and HighLac β-Lg suggest that the amount of glycation had to reach a threshold before nucleation was adversely affected, and the size of the adduct had no effect. In contrast, fibril growth (related to (df/dt)max) was affected by even low levels of glucosylation or lactosylation, and this effect was stronger for lactosylation than for glucosylation. In the stationary phase, increasing glucosylation progressively decreased the fibril yield (i.e., f max), whereas even a small amount of lactosylation dramatically reduced the yield. Lactose is approximately twice the size of glucose, and the stronger impact of lactose on the growth kinetics and yield indicates that the effects of glycation on self-assembly are related to the size of the sugar residue. Steric hindrance from bound sugar residues is thought to inhibit whey protein aggregation at pH 7,37,38 and it is likely that glycation inhibited self-assembly of peptides at pH 2 by a similar mechanism. The spacing between adjacent β-strands in an amyloid fibril is on the order of 4.87 Å.51 The hydrodynamic diameters of glucose and lactose molecules in solution are 7 and 9 Å, respectively,52 so these adducts would certainly disrupt β-strand packing and thereby inhibit fibril assembly. In addition, the hydrophilic −OH side groups of sugar residues may interfere with hydrophobic interactions between fibrilforming peptides, which contain many hydrophobic residues.18,20 Glycation appears to reduce or abolish the ability of peptides to self-assemble, and glycated peptides may instead undergo further hydrolysis and/or may assemble into nonfibril aggregates,53 reducing the potential yield of fibrils. A schematic illustrating our hypothesis is shown in Figure 8. It may be questioned why fibril growth was sensitive to low levels of glycation, whereas nucleation was affected only by high levels of glycation. The answer lies in considering that the susceptibility of lysine residues to glycation varies from site to site. At shorter glycation times, lysines 47, 91, and 100 are preferentially glycated.26,27,29 With an average of three bound sugar residues in LowGlu β-Lg and LowLac β-Lg (Figure1A,C), it is likely that these lysine residues were glycated. More extensive glycation in HighGlu and HighLac would also affect amine groups in the N-terminal region (Lys8 and Lys14 and the N-terminal amine). We suggest the N-terminal region containing these late-glycating lysines may be the main region involved in nucleation, whereas regions containing lysines that are preferentially glycated first are involved mainly in fibril growth. This work has shown that even heavily glucosylated or lactosylated β-Lg retains the secondary and tertiary structures of native unglycated β-Lg. However, the steric constraints introduced by glycation substantially inhibit the self-assembly of glycated β-Lg fragments into fibrils, the effect being greater for the larger lactose moiety, even though the extent of lactosylation was smaller than that of glucosylation. We

Figure 6. Tricine SDS−PAGE profiles of centrifugally precipitated fibrils made by heating glycated β-lg at 80 °C for 24 h. Key: M0, molecular mass markers (kDa); C, nonglycated control fibrils made with the same heat treatment.

glycated samples. However, the CD method used in this study took 40−60 s to collect a spectrum, which meant that the temporal resolution was not high enough to detect subtle changes in the kinetics of unfolding. More targeted scanning of only selected wavelengths would be faster and could potentially quantify the rate of unfolding with higher precision. An innate limitation of this method is that structural changes take place during the heating of the sample to 80 °C, thus making it impossible to define an accurate t = 0. Glycation produced a consistent ∼15% decrease in kh, which we are unable to explain. However, the proportion of unhydrolyzed β-Lg monomer at any one time (“I/I0” in Figure S5, Supporting Information) never spanned more than 10 percentage points across all treatments. Thus, glycation had a minor impact on the amount of hydrolyzed material present, and the effects of glycation on fibril formation were primarily at the self-assembly stage. Glycation of the lysine residues on β-Lg would reduce the net charge at pH 2 by +1 per lysine, and a charge reduction alone would be expected to increase the rate of self-assembly by reducing the electrostatic repulsion between peptides. However, glycation always reduced the rate of peptide self-assembly, indicating that the charge reduction effect was counterbalanced by a more dominant assembly-inhibiting factor. Using succinylated ovalbumin, Weijers et al.50 have shown that the net charge on the protein influences the morphology of its aggregates. However, glycation of β-lg did not appear to modify the fibril morphology (Figure 7), even though MS spectra and SDS−PAGE patterns indicated that the fibrils must be composed of glycated peptides. The kinetic parameters shown in Table 1 and Figure 5 reveal some important details about the effects of glycation on selfassembly. The rate of nucleation (related to tlag) was largely unaffected in LowGlu and LowLac β-Lg, which had a similar 3275



Article

Figure 8. Proposed mechanism of inhibition of self-assembly of glycated β-Lg by the steric effect of the sugar adduct.

■

ASSOCIATED CONTENT

■

AUTHOR INFORMATION

S Supporting Information *

Deconvoluted MS spectra of unglycated β-Lg (Figure S1), glucosylated and lactosylated β-Lg at different glycation times at 40 °C (Figures S2 and S3), CD spectra showing the effect of glycation on the β-Lg structure (Figure S4), residual intensities of the β-Lg band quantified from the SDS−PAGE of glycated samples (Figure S5) and β-Lg samples used as glycation controls (Figure S6), ThT intensities of β-Lg used as a glycation control (Figure S7), residual ThT intensities for supernatants for samples in Figure 6 (Figure S8), TEM image of fibrils from untreated control β-Lg after heating for 24 h (Figure S9), Tables S1 and S2 showing the kinetic parameters of acid hydrolysis and self-assembly calculated using eqs 1−5 from the data in Figures S6 and S7, and variance calculations for composite parameters. This material is available free of charge via the Internet at http://pubs.acs.org. Corresponding Author

*E-mail: [email protected]. Phone: +64-6-3569099, ext 81375. Fax: +64-6-3505655. Funding

This work was funded by Fonterra Cooperative Ltd. and the New Zealand Foundation for Research Science and Technology, Contract DRIX0701. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge Mr. Trevor Loo at the Institute of Fundamental Sciences (IFS) at Massey University for the MS analysis of glycated samples, and we thank Mr Doug Hopcroft at the Manawatu Microscopy and Imaging Centre Massey University for his help with the TEM imaging. We also thank Associate Professor Geoff Jones (IFS) for his guidance with statistical analysis and Dr. Gillian Norris (IFS) for useful discussions.

■

ABBREVIATIONS USED SDS−PAGE, sodium dodecyl sulfate−polyacrylamide gel electrophoresis; HPLC, high-performance liquid chromatography; UV, ultraviolet; MS, mass spectrometry; Q-TOF, quadrupole time-of-flight; ESI, electron spray ionization

Figure 7. TEM images of fibrils from glycated β-Lg after heating for 24 h at 80 °C: (A) LowGlu β-Lg; (B) HighGlu β-Lg; (C) LowLac β-Lg; (D) HighLac β-Lg.

■

propose that the inhibitory effect of glycation is influenced by the location of glycated residues, which is related to the varying glycation susceptibility of different lysine residues in the natively folded protein.

REFERENCES

(1) Pilkington, S. M.; Roberts, S. J.; Meade, S. J.; Gerrard, J. A. Amyloid fibrils as a nanoscaffold for enzyme immobilization. Biotechnol. Prog. 2010, 26, 93−100.

3276



Article

(2) Raynes, J. K.; Gerrard, J. A. Amyloid fibrils as bionanomaterials. In Bionanotechnology: Biological Self-Assembly and Its Applications; Rehm, B., Ed.; Horizon Scientific Press: Norfolk, U.K., 2012. (3) Loveday, S. M.; Su, J.; Rao, M. A.; Anema, S. G.; Singh, H. Whey protein nanofibrils: the environment−morphology−functionality relationship in lyophilization, rehydration, and seeding. J. Agric. Food Chem. 2012, 60, 5229−5236. (4) Oboroceanu, D.; Wang, L.; Magner, E.; Auty, M. A. E. Fibrillization of whey proteins improves foaming capacity and foam stability at low protein concentrations. J. Food Eng. 2014, 121, 102− 111. (5) Bolder, S. G.; Sagis, L. M. C.; Venema, P.; van der Linden, E. Effect of stirring and seeding on whey protein fibril formation. J. Agric. Food Chem. 2007, 55, 5661−5669. (6) Hill, E. K.; Krebs, B.; Goodall, D. G.; Howlett, G. J.; Dunstan, D. E. Shear flow induces amyloid fibril formation. Biomacromolecules 2006, 7, 10−13. (7) Loveday, S. M.; Su, J.; Rao, M. A.; Anema, S. G.; Singh, H. Effect of calcium on the morphology and functionality of whey protein nanofibrils. Biomacromolecules 2011, 12, 3780−3788. (8) Loveday, S. M.; Wang, X. L.; Rao, M. A.; Anema, S. G.; Singh, H. Effect of pH, NaCl, CaCl2 and temperature on self-assembly of βlactoglobulin into nanofibrils: a central composite design study. J. Agric. Food Chem. 2011, 59, 8467−8474. (9) Loveday, S. M.; Wang, X. L.; Rao, M. A.; Anema, S. G.; Singh, H. β-Lactoglobulin nanofibrils: effect of temperature on fibril formation kinetics, fibril morphology and the rheological properties of fibril dispersions. Food Hydrocolloids 2012, 27, 242−249. (10) Medrano, A.; Abirached, C.; Panizzolo, L.; Moyna, P.; Anon, M. C. The effect of glycation on foam and structural properties of βlactoglobulin. Food Chem. 2009, 113, 127−133. (11) Chevalier, F.; Chobert, J. M.; Popineau, Y.; Nicolas, M. G.; Haertle, T. Improvement of functional properties of β-lactoglobulin glycated through the Maillard reaction is related to the nature of the sugar. Int. Dairy J. 2001, 11, 145−152. (12) Nacka, F.; Chobert, J. M.; Burova, T.; Leonil, J.; Haertle, T. Induction of new physicochemical and functional properties by the glycosylation of whey proteins. J. Protein Chem. 1998, 17, 495−503. (13) Qi, X. L.; Holt, C.; McNulty, D.; Clarke, D. T.; Brownlow, S.; Jones, G. R. Effect of temperature on the secondary structure of βlactoglobulin at pH 6.7, as determined by CD and IR spectroscopy: a test of the molten globule hypothesis. Biochem. J. 1997, 324, 341−346. (14) Jameson, G. B.; Adams, J. J.; Creamer, L. K. Flexibility, functionality and hydrophobicity of bovine β-lactoglobulin. Int. Dairy J. 2002, 12, 319−329. (15) Uhrínová, S.; Smith, M. H.; Jameson, G. B.; Uhrín, D.; Sawyer, L.; Barlow, P. N. Accompanying pH-induced dissociation of the βlactoglobulin dimer. Biochemistry 2000, 39, 3565−3574. (16) Kella, N. K.; Kinsella, J. E. Enhanced thermodynamic stability of β-lactoglobulin at low pH. A possible mechanism. Biochem. J. 1988, 255, 113−118. (17) Bromley, E. H. C.; Krebs, M. R. H.; Donald, A. M. Aggregation across the length-scales in β-lactoglobulin. Faraday Discuss. 2005, 128, 13−27. (18) Dave, A. C.; Loveday, S. M.; Anema, S. G.; Loo, T. S.; Norris, G. E.; Jameson, G. B.; Singh, H. β-Lactoglobulin self-assembly: structural changes in early stages and disulfide bonding in fibrils. J. Agric. Food Chem. 2013, 61, 7817−7828. (19) Oboroceanu, D.; Wang, L. Z.; Brodkorb, A.; Magner, E.; Auty, M. A. E. Characterization of β-lactoglobulin fibrillar assembly using atomic force microscopy, polyacrylamide gel electrophoresis, and in situ fourier transform infrared spectroscopy. J. Agric. Food Chem. 2010, 58, 3667−3673. (20) Akkermans, C.; Venema, P.; van der Goot, A. J.; Gruppen, H.; Bakx, E. J.; Boom, R. M.; van der Linden, E. Peptides are building blocks of heat-induced fibrillar protein aggregates of β-lactoglobulin formed at pH 2. Biomacromolecules 2008, 9, 1474−1479.

(21) Kroes-Nijboer, A.; Venema, P.; Bouman, J.; van der Linden, E. Influence of protein hydrolysis on the growth kinetics of β-Ig fibrils. Langmuir 2011, 27, 5753−5761. (22) Arnaudov, L. N.; deVries, R.; Ippel, H.; vanMierlo, C. P. M. Multiple steps during the formation of β-lactoglobulin fibrils. Biomacromolecules 2003, 4, 1614−1622. (23) Hettiarachchi, C. A.; Melton, L. D.; Gerrard, J. A.; Loveday, S. M. Formation of β-lactoglobulin nanofibrils by microwave heating gives a peptide composition different from conventional heating. Biomacromolecules 2012, 13, 2868−2880. (24) Loveday, S. M.; Wang, X. L.; Rao, M. A.; Anema, S. G.; Creamer, L. K.; Singh, H. Tuning the properties of β-lactoglobulin nanofibrils with pH, NaCl and CaCl2. Int. Dairy J. 2010, 20, 571−579. (25) O’Brien, J. Heat-Induced Changes in Lactose: Isomerization, Degradation, Maillard Browning. International Dairy Federation (IDF): Brussels, Belgium, 1995; pp 134−170. (26) Fogliano, V.; Monti, S. M.; Visconti, A.; Randazzo, G.; Facchiano, A. M.; Colonna, G.; Ritieni, A. Identification of a βlactoglobulin lactosylation site. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1998, 1388, 295−304. (27) Morgan, F.; Bouhallab, S.; Molle, D.; Henry, G.; Maubois, J. L.; Leonil, J. Lactolation of β-lactoglobulin monitored by electrospray ionisation mass spectrometry. Int. Dairy J. 1998, 8, 95−98. (28) Fenaille, F.; Morgan, F.; Parisod, V.; Tabet, J. C.; Guy, P. A. Solid-state glycation of β-lactoglobulin by lactose and galactose: localization of the modified amino acids using mass spectrometric techniques. J. Mass Spectrom. 2004, 39, 16−28. (29) Morgan, F.; Molle, D.; Henry, G.; Venien, A.; Leonil, J.; Peltre, G.; Levieux, D.; Maubois, J. L.; Bouhallab, S. Glycation of bovine βlactoglobulin: effect on the protein structure. Int. J. Food Sci. Technol. 1999, 34, 429−435. (30) Morgan, F.; Leonil, J.; Molle, D.; Bouhallab, S. Modification of bovine β-lactoglobulin by glycation in a powdered state or in an aqueous solution: effect on association behavior and protein conformation. J. Agric. Food Chem. 1999, 47, 83−91. (31) French, S. J.; Harper, W. J.; Kleinholz, N. M.; Jones, R. B.; Green-Church, K. B. Maillard reaction induced lactose attachment to bovine β-lactoglobulin: electrospray ionization and matrix-assisted laser desorption/ionization examination. J. Agric. Food Chem. 2002, 50, 820−823. (32) van Teeffelen, A. M. M.; Broersen, K.; De Jongh, H. H. J. Glucosylation of β-lactoglobulin lowers the heat capacity change of unfolding; a unique way to affect protein thermodynamics. Protein Sci. 2005, 14, 2187−2194. (33) Morgan, F.; Leonil, J.; Molle, D.; Bouhallab, S. Nonenzymatic lactosylation of bovine β-lactoglobulin under mild heat treatment leads to structural heterogeneity of the glycoforms. Biochem. Biophys. Res. Commun. 1997, 236, 413−417. (34) Chevalier, F.; Chobert, J. M.; Dalgalarrondo, M.; Choiset, Y.; Haertle, T. Maillard glycation of β-lactoglobulin induces conformation changes. Nahrung-Food 2002, 46, 58−63. (35) Da Silva Pinto, M.; Bouhallab, S.; De Carvalho, A. F.; Henry, G.; Putaux, J. L.; Leonil, J. Glucose slows down the heat-induced aggregation of β-lactoglobulin at neutral pH. J. Agric. Food Chem. 2012, 60, 214−219. (36) Broersen, K.; Voragen, A. G. J.; Hamer, R. J.; de Jongh, H. H. J. Glycoforms of β-lactoglobulin with improved thermostability and preserved structural packing. Biotechnol. Bioeng. 2004, 86, 78−87. (37) Mulsow, B. B.; Jacob, M.; Henle, T. Studies on the impact of glycation on the denaturation of whey proteins. Eur. Food Res. Technol. 2009, 228, 643−649. (38) Liu, G.; Zhong, Q. Thermal aggregation properties of whey protein glycated with various saccharides. Food Hydrocolloids 2013, 32, 87−96. (39) Broersen, K.; Elshof, M.; de Groot, J.; Voragen, A. G. J.; Hamer, R. J.; de Jongh, H. H. J. Aggregation of β-lactoglobulin regulated by glucosylation. J. Agric. Food Chem. 2007, 55, 2431−2437. 3277



Article

(40) Chevalier, F.; Chobert, J. M.; Dalgalarrondo, M.; Haertle, T. Characterization of the Maillard reaction products of β-lactoglobulin glucosylated in mild conditions. J. Food Biochem. 2001, 25, 33−55. (41) Lee, D.; Park, C. W.; Paik, S. R.; Choi, K. Y. The modification of α-synuclein by dicarbonyl compounds inhibits its fibril-forming process. Biochim. Biophys. Acta, Proteins Proteomics 2009, 1794, 421− 430. (42) Oliveira, L. M.; Lages, A.; Gomes, R. A.; Neves, H.; Família, C.; Coelho, A. V.; Quintas, A. Insulin glycation by methylglyoxal results in native-like aggregation and inhibition of fibril formation. BMC Biochem. 2011, 12, 41. (43) Bouma, B.; Kroon-Batenburg, L. M. J.; Wu, Y. P.; Brunjes, B.; Posthuma, G.; Kranenburg, O.; de Groot, P. G.; Voest, E. E.; Gebbink, M. Glycation induces formation of amyloid cross-β structure in albumin. J. Biol. Chem. 2003, 278, 41810−41819. (44) Liu, G.; Zhong, Q. Dispersible and thermal stable nanofibrils derived from glycated whey protein. Biomacromolecules 2013, 14, 2146−2153. (45) Swaisgood, H. E., Chemistry of milk protein. In Developments in Dairy Chemistry, Fox, P. F., Ed.; Elsevier Applied Science: London, U.K., 1982; Vol. 1, pp 1−59. (46) Woody, R. W. Circular-dichroism. Methods Enzymol. 1995, 246, 34−71. (47) Meltretter, J.; Wüst, J.; Pischetsrieder, M. Comprehensive analysis of nonenzymatic post-translational β-lactoglobulin modifications in processed milk by ultrahigh-performance liquid chromatography-tandem mass spectrometry. J. Agric. Food Chem. 2013, 61, 6971−6981. (48) Manderson, G. A.; Creamer, L. K.; Hardman, M. J. Effect of heat treatment on the circular dichroism spectra of bovine β-lactoglobulin A, B, and C. J. Agric. Food Chem. 1999, 47, 4557−4567. (49) Tal, M.; Silberstein, A.; Nusser, E. Why does Coomassie Brilliant Blue R interact differently with different proteins? A partial answer. J. Biol. Chem. 1985, 260, 9976−9980. (50) Weijers, M.; Broersen, K.; Barneveld, P. A.; Stuart, M. A. C.; Hamer, R. J.; De Jongh, H. J.; Visschers, R. W. Net charge affects morphology and visual properties of ovalbumin aggregates. Biomacromolecules 2008, 9, 3165−3172. (51) Nelson, R.; Sawaya, M. R.; Balbirnie, M.; Madsen, A. O.; Riekel, C.; Grothe, R.; Eisenberg, D. Structure of the cross-beta spine of amyloid-like fibrils. Nature 2005, 435, 773−778. (52) Ribeiro, A. C. F.; Ortona, O.; Simões, S. M. N.; Santos, C. I. A. V.; Prazeres, P. M. R. A.; Valente, A. J. M.; Lobo, V. M. M.; Burrows, H. D. Binary mutual diffusion coefficients of aqueous solutions of sucrose, lactose, glucose, and fructose in the temperature range from (298.15 to 328.15) K. J. Chem. Eng. Data 2006, 51, 1836−1840. (53) Dave, A. C.; Loveday, S. M.; Anema, S. G.; Jameson, G. B.; Singh, H. Modulating β-lactoglobulin nanofibril self-assembly at pH 2 using glycerol or sorbitol. Biomacromolecules 2014, 15, 95−103.

3278


Glycation as a Tool To Probe the Mechanism of Î²-Lactoglobulin

Recommend Documents