Formation of β-Lactoglobulin Nanofibrils by Microwave Heating Gives

School of Chemical Sciences, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand ... Publication Date (Web): August 10, 2012 ... An ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Biomac

Formation of β‑Lactoglobulin Nanofibrils by Microwave Heating Gives a Peptide Composition Different from Conventional Heating Charith A. Hettiarachchi,†,‡ Laurence D. Melton,*,†,‡ Juliet A. Gerrard,†,§,∥ and Simon M. Loveday† †

Riddet Institute, Private Bag 11222, Palmerston North 4442, New Zealand School of Chemical Sciences, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand § Biomolecular Interaction Centre and School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand ∥ Industrial Research Limited, PO Box 31310, Lower Hutt 5040, New Zealand ‡

S Supporting Information *

ABSTRACT: A novel procedure involving microwave heating (MH) at 80 °C can be used to induce self-assembly of βlactoglobulin (β-lg) into amyloid-like nanofibrils at low pH. We examined the self-assembly induced by MH, and evaluated structural and compositional differences between MH fibrils and those formed by conventional heating (CH). MH significantly accelerated the self-assembly of β-lg, resulting in fully developed fibrils in ≤2 h. However, longer MH caused irreversible disintegration of fibrils. An increase in the fibril yield was observed during the storage of the 2 h MH sample, which gave a yield similar to that of 16 h CH sample. Fourier transform infrared (FTIR) and circular dichroism (CD) spectra suggested that the fibrils formed by the two methods do not show significant differences in their secondary structure components. However, they exhibited differences in surface hydrophobicity, and mass spectrometry showed that the MH fibrils contained larger peptides than CH fibrils, including intact β-lg monomers, providing evidence for a different composition between the MH and CH fibrils, in spite of no observed differences in their morphology. We suggest MH initially accelerates the self-assembly of β-lg due to its nonthermal effects on unfolding, nucleation, and subsequent stacking of β-sheets, rather than promoting partial hydrolysis. Thus, MH fibrils are composed of larger peptides, and the observed higher surface hydrophobicity for the MH fibrils was attributed to the parts of the larger peptides extending out of the core structure of the fibrils.



INTRODUCTION β-Lactoglobulin (β-lg) accounts for 60% w/w of the whey protein fraction of bovine milk, and is composed of 162 amino acids, with a molecular weight of ∼18.3 KDa.1 The peptide chain of β-lg has nine antiparallel β-strands and an α-helix (Figure 1).2,3 Eight of the β-strands assemble to form a calyx, a characteristic structure of the lipocalin family,4 while the ninth, which follows the α-helix, is involved in the dimerization of βlg.4,5 β-Lg can self-assemble into nanofibrils under different conditions. Extensive research has been conducted to understand the physicochemical properties of these nanofibrils due to their potential to be used as food ingredients to manipulate the structural properties of food. Investigating more efficient methods, which can induce self-assembly of β-lg is worthwhile, © 2012 American Chemical Society

as the existing methods require prolonged heating or incubation periods. Heating an acidic solution of β-lg above its denaturation temperature for an extended period is the most commonly used method for inducing self-assembly in vitro,6−8 and this is referred to as conventional heating (CH) in the present work. In addition, β-lg has also shown the ability to form fibrils when incubated with urea or potassium thiocyanate,9 or with alcohols10 for a period of 30 days or longer. Moreover, digesting β-lg with certain protease enzymes has also resulted in fibrils.11 Each method had a distinct kinetic pathway, and the Received: June 12, 2012 Revised: July 30, 2012 Published: August 10, 2012 2868

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

peptides resulting from partial hydrolysis are present in the fibrils.26,27 In the current study, we investigated microwave heating (MH) as an alternative method for preparing β-lg nanofibrils. Microwave radiation at GHz frequencies heats aqueous solutions by inducing continuous reorientation of dipolar water molecules. However, nonthermal effects on proteins in solution have also been postulated. Radiation in the high megahertz to low gigahertz range may directly excite torsional vibrations in the protein backbone, and thereby impact the rate of tertiary structure rearrangements such as denaturation and refolding.28 Nonthermal effects of microwave radiation have been shown to accelerate the refolding of cold-denatured β-lg at 12 °C,28 substantially lower the denaturation temperature of collagen (20−40 °C),29 reversibly diminish the activity of a trypsin inhibitor at 25 °C,30 and accelerate enzyme-catalyzed deglycosylation of glycoproteins at 37 °C.31 At 60 °C, nonthermal effects of microwave radiation can accelerate the aggregation of bovine insulin, and drive the aggregation reaction toward amyloid-like fibrils in preference to random aggregates.32 The effects of microwave radiation on proteins at higher temperature have received attention in the proteomics literature in the context of methods to rapidly hydrolyze proteins in preparation for mass-spectrometric analysis.33 MH of acidic solutions of protein to 130−140 °C hydrolyses proteins into peptides of a few thousand daltons within minutes, and with a high specificity for cleavage either side of aspartic acid (D) residues.34,35 In the present work, we studied the kinetics of self-assembly during MH and CH, and the effect of storage on the β-lg fibril samples at different stages of the self-assembly process. We extended our experiments to investigate whether differences exist in the proportions of secondary structure components and the peptide composition between the fibrils formed by the two methods.

Figure 1. Structure of the β-lg monomer (variant A) obtained from the structure deposited by Qin et al.2 (PDB: 3BLG). α-Helix is given in red. Nine β-strands (A−I) are in blue. Tryptophans (W) 19 and 61 (responsible for tryptophan fluorescence) are given in yellow. Lysines (K) 47 and 100 (susceptible for lactosylation) are given in green. The amino acid sequence (obtained from UniProtKB; www.uniprot.org) for the β-lg variant A is given below the structure. Aspartic acid (D) at 64 and valine (V) at 118 are replaced by glycine (G) and alanine (A), respectively in variant B.

resulting fibrils generally exhibited different morphologies and physicochemical properties, compared to CH fibrils. β-Lg nanofibrils formed by the CH method are amyloid-like as judged by their diagnostic cross-β X-ray diffraction pattern,12 and because they readily bind thioflavinT (ThT)7,8 and Congo red dyes,12,13 in a similar manner to amyloid fibrils of diseaserelated amyloid plaques formed in vivo. The fibrils obtained by the CH method generally have a width of ∼7 nm and a length varying from 0.1 to a few micrometers.6,14−16 The fibrils are composed of two or more protofilaments aligned to form a ribbon, which twists along its axis, exhibiting periodicity.17 The period (or pitch) of the fibrils is dependent on the number of protofilaments making up the fibril and the ionic strength of the medium.18 Atomic force microscopy (AFM) and small-angle neutron scattering (SANS) have shown that the fibrils formed by heating an acidic β-lg solution at 90 °C for 5 h comprise three protofilaments.19,20 However, with prolonged heating (24−30 h), the number of protofilaments assembling into a single fibril increased, reaching a maximum of 16.21 Heat-induced β-lg nanofibril formation appears to be nucleation-limited, resulting in a three-phase sigmoidal growth curve,7,22 which is typical for the amyloid fibril formation. It was previously thought that partially unfolded β-lg monomers could align and aggregate into fibrils via β-sheet interactions,6,23 but subsequent evidence showed that the fibrils formed after prolonged heating consisted of 2−8 kDa peptides, suggesting that hydrolysis was required for self-assembly.8 It is now accepted that hydrolysis promotes self-assembly, but other fibril formation pathways involving little or no hydrolysis also exist.24,25 In the case of heat-induced lysozyme fibril formation, it was reported that both intact lysozyme molecules and



EXPERIMENTAL SECTION

Materials. Bovine β-lg of 97% purity was purchased from Sigma (St Louis, MO, USA) and was a mixture of variants A and B. A mass spectrum obtained for the β-lg revealed that the variants A and B were present in a ratio of 3:2. ThT, Congo red, and 8-anilino-1-naphthalene sulfonic acid (ANS) were also purchased from Sigma. All the other chemicals used were of analytical grade. Milli-Q water (18.2 MΩ.cm) was used throughout the experiment. pH adjustments were done to an accuracy of ±0.05 from the required value using a pH meter (Orion320, Boston, MA, USA), which was calibrated daily. Fibril Formation. β-Lg was dissolved in water, under gentle stirring for 40 min at 4 °C. The solution was then centrifuged at 22600g for 30 min at 4 °C (Sorvall RC-285, Newtown, CT, USA) and subsequently filtered using MillexGV 0.22 μm syringe filters (Millipore, Carrigtwohill, Ireland) to remove any undissolved protein. Filtered solution was stored at 4 °C and was used within 7 days. For the fibril formation experiments, β-lg solution was first allowed to reach room temperature, and the pH was adjusted to 2 from its original pH (7.3) using 1 M HCl. The protein concentration of the pH-adjusted solutions used for the fibril formation experiments was 14.7 ± 0.2 mg/mL, as determined by 280 nm absorbance (UV mini 1240, Shimadzu, Japan). The solution was then subjected to MH at 80 °C for 2 or 16 h using a CEM Discover-S class microwave system (Matthews, NC, USA) set at 4 W, as the maximum power level. The standard tube (10 mL) provided with the machine was used to hold the solution. Alternatively, the β-lg solution (pH 2) was subjected to CH in Pyrex screw-capped test tubes using a digitally controlled 2869

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

thermostatic heat block (Grant QBT2, Royston, England) for the same time periods at 80 °C. Heating times were predetermined based on preliminary experiment results (Figure 2A, inset).

(Heraeus Labofuge 400, Hanau, Germany). The retentate was washed twice by vortexing and then centrifuging with 1000 μL fractions of pH 2 HCl to ensure that any nonfibril proteins and peptides remaining in the retentate was removed. The net weight of the filtrate was obtained by subtracting the weight of the filtrate collection tube from the total weight (filtrate collection tube plus filtrate), and the weight was converted to a volume by assuming a density of 1 g/mL. The concentration of the filtrate which contained the nonfibril proteins and peptides was then determined by measuring absorbance at 280 nm. Percent fibril yield was estimated as given below: Fibril yield(%) =

[X] × VX − [Y] × VY × 100 [X] × VX

where [X] = β-Lg concentration of the fibril sample (mg/mL, assumed to be similar to the β-lg concentration before heating); VX = volume of the β-lg fibril sample (mL); [Y] = concentration of the filtrate (mg/ mL); and VY = volume of the filtrate (mL) Transmission Electron Microscopy (TEM). A glow-discharged carbon-coated 400 gauge copper grid was placed on the 5-fold diluted sample for a fixed time, and then it was washed once with water and transferred onto a 2% w/v aqueous uranyl acetate droplet for negative staining. TEM imaging was performed using a Philips CM 12 microscope (Eindhoven, The Netherlands) operating at 120 KV, equipped with a Gatan 792 BioScan camera (Pleasanton, CA, USA). Images were selected from ≥7 representative images obtained from scanning one or a few grids per sample. Congo Red Birefringence. Macroscopic aggregates (suspended particles in the fibril samples that are visible to the eye) were assessed for Congo red birefringence. The fibril samples (16 h CH and 2 h MH) were first diluted with pH 2 HCl to a protein concentration of 1 mg/mL, and the macroscopic aggregates were then separated by centrifuging at 200g. After removal of the supernatant, aggregates remaining in the bottom of the centrifuge tube were washed twice by means of resuspension and centrifugation. The aggregates were placed on a glass slide and dried under a stream of nitrogen. A Congo red staining solution, which was freshly prepared according to Nilsson,38 was applied to the dried protein sample. The slide was allowed to dry and observed between the crossed polarizer and the analyzer of a Leica DMRE microscope (Wetzlar, Germany) for birefringence. Images were obtained using a Leica DC 500 CCD camera, connected to the microscope. These macroscopic aggregates were removed prior to the CD and mass spectrometry experiments given below. Intrinsic Tryptophan Fluorescence and Hydrophobicity Measurements (S0). The fibril samples were diluted to a protein concentration of 1 mg/mL with 50 mM phosphate buffer (pH 7) for the intrinsic tryptophan fluorescence measurements. Samples were excited at 295 nm, and emission was measured across the range of 310−450 nm with both excitation and emission slit widths set at 3 nm.39,40 Hydrophobicity (S0) of the fibrils was assessed according to a modification of the method used by Kato and Nakai.41 A dilution series was prepared for each fibril sample (0.001−0.02 mg/mL) in 50 mM phosphate buffer at pH 7. Then, 1 mL from the each dilution was mixed with 20 μL of 2 mM ANS in the same buffer, and the mixture was kept for 15 min in the dark before measuring the fluorescence. Samples were excited at 390 nm, and emission was recorded for the range of 420−530 nm, with both excitation and emission slit widths set at 4.2 nm. Both tryptophan and ANS fluorescence measurements were made using the same luminescence spectrophotometer. Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) Spectroscopy. FTIR spectra for the fibril samples were acquired with a resolution of 4 cm−1 using a Thermo Nicolet 8700 spectrometer (Madison, WI, USA) equipped with a mercury cadmium telluride (MCT) detector, cooled with liquid nitrogen. After collecting the background spectrum, freeze-dried samples were placed on the diamond ATR accessory fitted to the spectrophotometer, and the spectrum for each sample was obtained by averaging 64 scans recorded between 400 and 6000 cm−1. Second derivative spectra (Savitsky-

Figure 2. Self-assembly of β-lg monitored by ThT fluorescence and TEM. (A) ThT fluorescence intensity recorded for β-lg fibril samples after the requisite heating time. Error bars represent the standard deviation for triplicate measurements made separately on two individually prepared samples. Inset shows the ThT intensities recorded during the preliminary experiments and X,Y axes names are similar to the bar graph. (B) TEM images obtained for the MH or CH fibril samples at each time point. Arrows show the smaller particles. Samples were diluted 5-fold with pH 2 HCl and stained with uranyl acetate prior to TEM. Scale bars represent 0.5 μm and apply to both images in each column. After MH or CH of the β-lg solution for the requisite time, samples were immediately cooled by submerging in a water-ice bath. The experiment was further extended to investigate the effect of storage on the β-lg fibril samples by storing them for 7 days at 4 °C. ThT Fluorescence. An aliquot of the β-lg fibril sample (10 μL) was mixed with 800 μL of 60 μM ThT in 10 mM phosphate buffer containing 150 mM NaCl (pH 7)22,36 and kept for 40 min in the dark prior to measurement. They were excited at 440 nm and emission was measured at 482 nm using a Perkin-Elmer LS 55 luminescence spectrophotometer (Buckinghamshire, England) with excitation and emission slit widths set at 5 and 8 nm, respectively. The intensity of the fluorescence emission is presented as ‘net ThT fluorescence’ after the subtracting the fluorescence intensity for the buffer with ThT, in the absence of β-lg. Estimation of Fibril Yield. Fibril yield was estimated according to the centrifugal filtration method described by Bolder et al.37 An aliquot of the β-lg fibril sample was diluted 5-fold with pH 2 HCl. Next, 2000 μL of the diluted sample was transferred into a prewashed 100 KDa molecular weight cutoff (MWCO) centrifugal filter (Amicon Ultra, Millipore, Billerica, MA, USA) and centrifuged at 2383g for 20 min 2870

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

MS/MS scan (100−1600 m/z). The collision energy was automatically decided based on the m/z ratio and the charged state of the peptide. The obtained MS/MS data were searched against the amino acid sequences of β-lg variants A and B obtained from UniProt KB (available from: www.uniprot.org) using the ProteinPilot software (V2.0.1, Applied Biosystems). In addition to MS/MS data, total ion chromatograms obtained for the both dissociated fibril samples were assessed to obtain complete peptide mass profiles of the samples. TICs were deconvoluted to the peptide masses using AnalystQS (V1.1, Applied Biosystems) for the total peptide elution period, using a window of 2 min at a time. Identification of Higher Molecular Weight Species. The peptide sequences corresponding to the higher molecular mass species (>12000 Da) was carried out using the FindPept tool of the ExPASy proteomics server (available from: www.expasy.org) with a tolerance of ±0.5 Da. All the cysteines (C) were considered to be carbamidomethylated (+57.05 Da per C). Oxidation of methionines (M) (+15.99 Da per M) and deamidation of asparagines (N) (+0.98 Da per N) and glutamines (Q) (+0.98 Da per Q) were set as optional modifications. Searching was carried out in the following order: First the observed masses were matched against the amino acid sequences of both β-lg variants (Var A and B). If a probable match was not available, 648.6 or 324.3 Da was subtracted from the observed mass before searching against the variants, to account for mono- or dilactosylation of β-lg. Priority was given for the aspartic acid (D)cleavage and trimming of the complete β-lg sequence at the C or N terminal when multiple matches were present.

Golay derivative) were obtained using the Omnic software to determine the component peak positions (see Supporting Information, Figure S1). Fitting of the components was then carried out for the 1450−1800 cm−1 region using the Grams32 software (V4.0; Salem, NH, USA). A Gaussian profile was used for the fitted curves, and all the given fits converged with the original spectra. Fitting was not limited to the amide-I region (1600−1700 cm−1) to overcome the errors that can occur due to overlapping/extending components from the other regions to the amide-I region and to get a proper baseline correction. Circular Dichroism (CD) Spectroscopy. Far-UV CD spectra were recorded at 20 °C using an Applied Photophysics piStar 180 spectropolarimeter (Leatherhead, England). Fibril samples were diluted to 0.08 mg/mL in 10 mM phosphate buffer (pH 7) and centrifuged at 200g for 5 min to sediment any macroscopic aggregates. CD spectra of the supernatants were recorded from 180 to 260 nm using a quartz cuvette with a 1 mm path length. Five scans were averaged, background subtracted, and then smoothed using polynomial regression and Gaussian density function. A CD spectrum was also acquired for a 0.08 mg/mL unheated β-lg solution in the same buffer. Secondary structure estimations were then carried out using the CDSSTR program 42 in the CDPro Package (available from lamar.colostate.edu/∼sreeram/CDPro) with reference protein sets SP43 and SP37. Separation and Disintegration of Fibrils. β-Lg fibril samples were diluted 5-fold, and macroscopic aggregates were removed by centrifuging at 200g for 5 min. Afterward, the fibrils in the solutions were separated from nonfibril proteins and peptides using 100 KDa MWCO centrifugal filters following the method used for the estimation of fibril yield. The retentate (containing the fibrils) was resuspended in water, and in order to disintegrate the fibrils, solid guanidine hydrochloride (GuHCl) was added to an aliquot (500 μL) of the suspension to give 8 M concentration. Next, the mixture was placed on an orbital shaker for 6 h and stored at 4 °C for >4 days before the mass spectrometry.27 TEM images were obtained to observe the fibril disintegration. Carbamidomethylation and Solid Phase Extraction (SPE) of Peptides. An aliquot of 5 μL obtained from the GuHCl-treated fibril sample was mixed with 20 μL of 50 mM NaHCO3 buffer (pH 8), dithiothreitol (DTT) was added to give a final concentration of 10 mM, and the mixture was heated at 50 °C for 30 min to reduce any disulfide bonds. Next, the sample was treated with iodoacetamide (50 mM final concentration) and kept in the dark for 30 min to modify the reactive sulfhydryl groups by carbamidomethylation. DTT was again added to quench the excess iodoacetamide. Then the sample was made up to 500 μL with 0.1% v/v formic acid. Peptides present in this solution were separated by SPE using Waters Oasis HLB cartridges (Milford, MA, USA). Elution was carried out with 60% v/v acetonitrile after washing the cartridge once with 0.1% v/v formic acid. The eluted solution was concentrated using a Savant SpeedVac vacuum concentrator (Holbrook, NY, USA) to a final volume of ∼100 μL, and it was diluted 2-fold with 0.1% v/v formic acid before injecting into the liquid chromatography (LC) system. LC and Tandem Mass Spectrometry (MS/MS). Samples (10 μL) were autoinjected into a LC Packings-Dionex, UltiMate Capillary/ Nano LC System (Amsterdam, Netherlands) carrying an Agilent Zorbax 300SB (Santa Clara, CA, USA) C18 reverse phase column (10 mm × 300 μm; 3.5 μm particle size; 300 Å pore size) with a LC Packings C18 PepMop trap column (5 mm ×300 μm; 5 μm particle size; 300 Å pore size). The mobile phase was comprised of water (A) and acetonitrile (B) both containing 0.1% v/v formic acid and a gradient of 10−55% v/v B over 55 min was employed. The LC eluant (flow rate: ∼6 μL/min) was directed to the electrospray ionization (ESI) source (spray voltage: 5 kV) of the mass spectrometer (QStar XL-Applied Biosystems, Foster City, CA, USA). Peptide analysis was performed in the data-dependent acquisition mode. The three most intense ions present in each MS scan (300− 1600 m/z) were sequentially selected by the first quadrupole and fragmented at the second quadrupole by collision with nitrogen for the



RESULTS MH Changes β-Lg Self-Assembly Kinetics. β-Lg samples (∼15 mg/mL; pH 2) either MH or CH at 80 °C for different time periods were assessed for ThT fluorescence to monitor the fibril formation. MH samples showed the highest fluorescence intensity at 2 h with a ThT fluorescence intensity similar to the intensity recorded for the 8−10 h samples (Figure 2A, Inset). Upon further MH, a marked reduction in the fluorescence emission was observed, suggesting that extended MH does not favor fibril formation. An increase in the ThT fluorescence emission with time was observed for the CH β-lg samples, as expected (Figure 2A). Fluorescence intensities of the 16 h CH samples were within the range corresponding to the saturation (or stationary) phase of self-assembly, as determined during the preliminary experiments (Figure 2A, inset). The relative abundance of fibrils observed in TEM images obtained for the β-lg samples at different stages of MH or CH were in agreement with the ThT fluorescence measurements (Figure 2B). Smaller (roughly spherical) particles were present in the 2 h MH and 2 h CH samples in addition to the fibrils (Figure 2B, 2 h MH, 2 h CH). The 16 h MH sample showed that most of the fibrils were shorter than they were in the 2 h MH sample (Figure 2B, 16 h MH). More TEM images obtained for the 16 h MH sample are given in the Supporting Information (Figure S2). The fibril yield was indirectly estimated by measuring the nonfibril protein and peptide content of the fibril samples. A fibril yield of ∼40 and 20% w/w was calculated for the 2 and 16 h MH samples, respectively (Figure 3). This explained the decrease in ThT fluorescence between 2 and 16 h MH samples. The increase in nonfibril protein and peptide content in the 16 h MH sample suggested disintegration of a considerable amount of fibrils during extended MH. TEM images obtained for the 16 h MH sample (Figures 2B, 16 h MH and S2) are consistent with this, showing shortened fibrils. Fibril formation increased with time for CH samples, yielding ∼15 and 55% w/w fibrils in 2 and 16 h, respectively (Figure 3). 2871

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

Figure 3. Estimation of fibril yield. Fibril yield was calculated as mentioned under the Estimation of Fibril Yield, in the Experimental Section. Error bars represent the standard deviation for triplicate measurements made separately on two individually prepared samples.

Table 1 presents previously reported values for β-lg fibril yield. The results obtained in this study are within the range of those Figure 4. Congo red birefringence assay for the macroscopic aggregates formed in the β-lg fibril solutions. Aggregates formed in (A) 2 h MH and (C) 16 h CH fibril solutions stained with Congo red. (B,D) The same areas observed between the crossed polarizer and analyzer. Magnification ×200.

Table 1. β-Lg Fibril Yields Estimated During Previous Studies protein and the concentration

fibril formation conditions

β-Lg (Sigma); 3% w/v

pH 2, 85 °C, 20 h; Shear rate (323 S1−) pH 2, 80 °C; 10 h

b

WPI 1−3, 4 and 5% w/w WPIb 2% w/w β-Lg (Sigma); 1, 2 and 3% w/v β-Lg (isolated from fresh milk); 0.5 and 1% w/v

pH 2, 80 °C; 10 h; at rest or stirred pH 2, 80 °C; 10 h pH 2.5; 80 °C; 4 h; 0.1 M NaCl

fibril yield % w/w 28

6000 Da) were present in MHS fibrils, in contrast to CH fibrils, which were predominantly made of shorter peptides having a mass of 12000 Da) present in 2874

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

Table 2. Analysis of Components Fitted to the Amide-I Region of the FTIR Spectra MHS assignment

a

component

β-Sheets turns turns α-Helices β-Sheets β-Sheets

−1

b

1 2 3 4 5 6

unheated β-lgc

16 h CH −1

center (cm )

area %

center (cm )

area %

center (cm−1)

area %

1693 1680 1667 1655 1639 1621

13 8 14 16 32 17

1694 1680 1668 1657 1639 1617

10 9 8 16 37 20

1693 1682 1666 1651 1632

3 14 23 20 41

a

Assignment of the components was carried out according to Dong et al.,48 Dong and Caughey,49 and Byler and Sushi.50 bComponent numbers correspond to the numbering in Figure 8. c1611 cm−1 component of the unheated β-lg amide I band (Component 7) was not considered for the area calculation, as it arises from side chain vibrations.48



DISCUSSION ThT fluorescence, TEM images, and estimation of fibril yield together suggested that MH markedly accelerates β-lg selfassembly kinetics during the first 2 h (Figures 2 and 3), even though the temperature of the sample was the same as that of the CH sample. There was an unusually short lag phase for MH, and we were able to observe fibrils by TEM even after MH for 1 h (see Supporting Information, Figure S6). The increased fibril yield observed after the storage of the 2 h MH and 2 h CH samples at 4 °C suggests that modified monomers and/or peptides can continue assembling into fibrils during storage. However, exposing the β-lg solution to MH for more than 2 h at 80 °C did not result in an increase in fibril content. Hence, extended MH may alter the structure of these primary units preventing further assembly, in addition to the disintegration of already formed fibrils. Storage of the 2 h MH β-lg sample for ≥4 days gave a fibril yield (52 ± 3% w/w) similar to the 16 h CH sample (Figure 5). Previously reported fibril yields depended on the fibril formation conditions and method of estimation (Table 1), and it was difficult to find a consensus among the reported values, e.g., shearing is known to enhance the β-lg fibril formation,7 yet a lower yield was reported by Akkermans et al.8 for a β-lg solution heated and sheared for 20 h, in comparison to the yield obtained by Veerman et al.15 for a β-lg solution of a similar concentration, heated for 10 h without stirring (Table 1). Intrinsic fluorescence spectra indicated that the tryptophan (W) residues in both 16 h CH and MHS fibrils are exposed to a more hydrophilic environment than they are in the native structure (Figures 1, 6). In the intrinsic fluorescence spectrum, the wavelength of maximum emission (λmax) is an indication of the polarity of the environment surrounding the indole group of tryptophan residues. It shifts to a higher wavelength when tryptophan residues are exposed to a more polar environment,53 and the quantum yield (Fmax) decreases when they interact with quenching agents, either in the solvent or in the protein itself.39 Similar intrinsic fluorescence results were reported during the self-assembly of α-chymotrypsin in the presence of trifluoroethanol (TFE)54 and self-assembly of β-lg in the presence of 5 M urea.9 The observed red shift in the λmax and decrease in Fmax during the present work can be attributed to the unfolding and/or partial hydrolysis of β-lg during the self-assembly process, which leads to greater exposure of one or both tryptophan residues to the solvent. The MHS fibril sample showed a higher Fmax in comparison to the 16 h CH fibrils, suggesting that tryptophan residues in the former may exist in a more hydrophobic environment. However, this might not

Figure 9. Far-UV CD Spectra obtained for fibril samples and β-lg. CD spectra were recorded for the samples diluted in 10 mM phosphate buffer (pH 7) to a final concentration of 0.08 mg/mL. Five scans were averaged for each spectrum, background subtracted, and smoothed by Gaussian density function.

Table 3. Estimation of Secondary Structures Based on CD Data % secondary structuresa sample

α-helices

β-sheets

turns and unordered

MHS 16 h CH unheated β-lg

7 6 20

30 33 28

63 60 52

a

Secondary structure estimation was carried out using the SP43 protein set in the CDPro software. Protein set SP37 gave similar results to SP43.

MHS fibrils conducted using proteomics tools of the ExPASy server showed the presence of β-lg monomers or monomers lacking a few residues from either C or N terminal in MHS fibrils (Table 6). Lactosylated forms were also present. Searching for the lactosylated peptides in 16 h CH fibrils was limited to the MS/MS confirmed peptide sequences, as several sequences were possible for a single mass, due to their lower molecular weight. We selected the sequences including lysine (K) 47 and/or K100 from the MS/MS list (Table 4), which are the most susceptible sites for the lactosylation,51,52 and manually searched within the total peptide list for the presence of lactose-bound forms. We found the mass corresponding to the lactose bound 34−62 sequence, and other possibilities may exist within the total list of peptide masses, especially those that include K47 and/or K100 in their sequence. The location of the K47 and K100 in β-lg is given in Figure 1. 2875

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

Table 4. Peptides Present in Fibrils as Confirmed by MS/MS Data amino acid sequencea

presence

position

LIVTQTMKGL (D) LIVTQTMKGLD (I) LIVTQTMKGLDIQKVAGTWYSLAMAASD (I) LIVTQTMKGLDIQKVAGTWYSLAMAASDISLLDAQSAPLRVYVEELKPTPEG (D) LIVTQTMKGLDIQKVAGTWYSLAMAASDISLLDAQSAPLRVYVEELKPTPEGD (L) (D) IQKVAGTWYSLAMAAS (D) (D) IQKVAGTWYSLAMAASDISLLDAQSAPLRVYVEELKPTPEGD (L) (D) AQSAPLRVYVEELKPTPEG (D) (D) AQSAPLRVYVEELKPTPEGD (L) (D) AQSAPLRVYVEELKPTPEGDLEILLQKWE (N)b (D) AQSAPLRVYVEELKPTPEGDLEILLQKWEN (G or D) (D) AQSAPLRVYVEELKPTPEGDLEILLQKWEND (E) − var. A (D) LEILLQKWEND (E) − var. A (D) LEILLQKWENGECAQKKIIAEKTKIPAVFKI (D) − var. B (D) LEILLQKWENDECAQKKIIAEKTKIPAVFKI (D) − var. A (D) LEILLQKWENGECAQKKIIAEKTKIPAVFKID (A) − var. B (D) LEILLQKWENDECAQKKIIAEKTKIPAVFKID (A) − var. A (G or D) ECAQKKIIAEKTKIPAVFKID (A) (D) ALNENKVLVL (D) (D) ALNENKVLVLD (T) (D) ALNENKVLVLDTD (Y) (D) DEALEKFDKALKALPMHIRLSFNPTQLEEQCHI (D) EALEKFDKALKALPMHIRLSFNPTQLEEQCHI (D) KALKALPMHIRLSFNPTQLEEQCHI (N) PTQLEEQCHI

MHS, 16 h CH MHS, 16 h CH MHS MHS MHS 16 h CH MHS MHS, 16 h CH MHS, 16 h CH 16 h CH 16 h CH MHS, 16 h CH 16 h CH 16 h CH MHS, 16 h CH 16 h CH MHS, 16 h CH MHS, 16 h CH MHS, 16 h CH MHS, 16 h CH MHS MHS, 16 h CH MHS, 16 h CH MHS, 16 h CH MHS, 16 h CH

1−10 1−11 1−28 1−52 1−53 12−27 12−53 34−52 34−53 34−62 34−63 34−64 54−64 54−84 54−84 54−85 54−85 65−85 86−95 86−96 86−98 130−162 131−162 138−162 153−162

a Adjacent amino acid residues of the peptides are given within brackets. The following modifications were commonly observed: deamidation of asparagines (N) and glutamine (Q) residues, oxidation of methionine (M), pyroglutamic acid formation at the glutamic (E) or glutamine (Q) residues at the N-terminal. Carbamidomethylation was observed for all cysteines (C). bManual searching among the peptide masses of the CH fibrils showed the presence of lactose bound form (3466.9 + 324.3 Da).

the core, they may still bind with ANS, giving a higher S0 to the MHS fibrils. FTIR and CD experiments were conducted to determine the changes that take place in the secondary structure composition during the two self-assembly processes. The amide-I band of the FTIR spectrum obtained for the unheated β-lg is characteristic of a β-sheet rich protein with a peak value of 1632 cm−1 (Figure 8C). Although the classical shift of the amide-I peak to a lower wavenumber that takes place during amyloid fibril formation57 was not prominent for the β-lg fibrils, we were able to observe the appearance of a new peak at the ∼1620 cm−1 region (Figure 8A,B, Component 6) revealing the formation of intermolecular β-sheets with stronger hydrogen bonding, in parallel with the self-assembly of β-lg. Secondary structure estimations showed that no marked differences exist in the β-sheet or α-helix content between the MHS and 16 h CH fibrils, suggesting that the peptides stack in a similar manner in the both fibril types, irrespective of their relative size. This is consistent with the observed morphological similarity between the MHS and 16 h CH fibrils. Moreover, a new shoulder in the 1720 cm−1 region was observed for the both MHS and 16 h CH fibril samples (Figure 8A,B). Vibrations in this region can be assigned to the protonated carboxyl groups of aspartic (D) and glutamic (E) residues, considering the given pH conditions.58 D-cleavage, which was prominently observed during the fibril formation (Table 4) explains the observed shoulder, due to the increased exposure of their carboxyl group. CD spectra obtained for the unheated β-lg are typical of a βstructure-rich protein,59 with a negative band showing a

Table 5. Accumulated Area Under the Peaks for Different Mass Ranges % from total peak area fibril type

12000 Da

MHS 16 h CH

62 90

30 10

8 0

provide conclusive evidence for the relative locations of tryptophan residues in the two types of fibrils, due to the possibility of difference in the amount of macroscopic aggregates formed in the two samples, which may respond differently compared to free fibrils. ANS is a hydrophobic fluorescent probe that is used to assess the hydrophobicity (S0) of proteins regardless of their molecular form.55 Self-assembly of β-lg by both methods increased the S0 (Figure 7) compared to its native state, suggesting an increase in accessibility to the hydrophobic areas. Similar results were observed by Pallarés et al.54 during the selfassembly of α-chymotrypsin. MHS fibrils showed a higher S0 than 16 h CH fibrils (Figure 7), revealing that they have more sites available for the binding of the probe and/or greater accessibility toward the hydrophobic areas in MHS fibrils than 16 h CH fibrils. Mass spectrometry showed that MHS fibrils are made of larger peptides, including intact monomers. Thus, more loops (that connect β-strands in the same or different βsheets) and/or parts of the peptides extending out of the core structure of the fibrils24,56 are probably available on the MHS fibril surface (Figure 11). Although these protrusions can be relatively less hydrophobic than the peptides assembling into 2876

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

Figure 10. Mass reconstruction of +TOF-MS obtained for the dissociated MHS fibrils. Fibrils were separated using 100 KDa centrifugal filters and washed twice to remove nonfibril materials. Fibrils were dissociated using GuHCl (8M). Samples were treated with DTT and iodoacetamide to modify the sulfhydryl groups. Peptides were then extracted with 60% v/v acetonitrile, concentrated, and injected into the LC-MS system. Probable matches identified for the major peaks (area ≥200 au) are listed in Table 6.

minimum of 217 nm (Figure 9). For both β-lg fibril samples, an increase in the intensity of negative signal and the shift of the spectral minimum toward ∼203 nm was observed due to the unfolding of β-lg.60,61 The shape of the CD spectra obtained for the fibril samples indicated an increase in random coils,59 which is consistent with the observed increase in turns and unordered structures (Table 3). Although the CD data showed a similar trend to FTIR, the increase in β-sheet content during fibril formation was not large, as was observed with FTIR results (Table 2). FTIR experiments were conducted on total fibril samples, which contained macroscopic aggregates, while the CD experiments were conducted in the absence of these macroscopic aggregates. The relative proportion of fibrils present is therefore less in the samples subjected to CD analysis. Peptides with C or N terminal aspartic (D) were prominent in both MHS and 16 h CH fibrils (Table 4), in agreement with the results of Akkermans et al.8 and Bateman et al.62 for β-lg fibrils and also with those of Frare et al.,63 who observed partial hydrolysis at D residues prominently during the heat-induced self-assembly of hen lysozyme. The observed specificity can be attributed to the susceptibility of the peptide bonds on either side of the D residue to high-temperature acid hydrolysis.64 Microwave-accelerated acid hydrolysis of ovalbumin showed similar specificity.35 Mass spectrometric analysis further suggested that 16 h CH fibrils are mainly composed of peptides fragments with a mass ≤6000 Da (90%), in agreement with Akkermans et al.8 and Bateman et al (Table 5).62 In contrast, MHS fibrils were composed of larger peptides, including intact β-lg monomers (Tables 5 and 6). β-Lg fibrils are considered to be amyloid-like,12 thus these fibrils (or the protofilaments assembling into fibrils) are thought to have βsheets running parallel to the fibril (or protofilament) axis. The β-sheets are made by stacking of β-strands. Larger peptides and unfolded monomers will consist of more than a single β-strand, and these strands can be a part of two or more β-sheets. The βstrands may be connected by loops, and the existence of such loops has been previously reported for amyloid fibrils.65,66

Moreover, a large peptide may provide more than one β-strand for the same β-sheet.66 In addition to the loops, there can be also hairs due to the non-β-strand parts of the peptides. The proposed structures for the MHS and 16 h CH fibrils are illustrated in Figure 11. Fibril formation continued at 4 °C during storage of MH samples, which indicates the presence of assembly-competent species (peptides and unfolded monomers) at the end of 2 h MH, in addition to fibrils formed by that time. Assemblycompetent species may have accumulated during MH because their assembly into fibrils was slower than their creation. It is reasonable to assume that any native β-lg remained at the end of 2 h MH did not denature during storage at 4 °C, therefore fibrils formed during storage would comprise the same building blocks as those formed during 2 h MH. De Pomerai et al.32 reported that nonthermal effects of microwave radiation promoted the self-assembly of bovine insulin into amyloid-like fibrils at 60 °C, but did not suggest a mechanism. On the basis of the classical nucleation-polymerization model suggested for amyloid fibril formation60,67 and recent insights into the self-assembly of proteins,8,21,27,63 we believe there are 4 stages at which microwave radiation could have affected self-assembly associated with fibril formation: (1) protein unfolding, (2) nucleation, (3) hydrolysis, and (4) fibril growth. The temperature was maintained at 80 ± 2 °C during the MH process, which was consistent with the temperature maintained during the CH process, thus it is unlikely that any differences in temperature between the two methods affected the results (see Supporting Information, Figure S8). Nonthermal effects of microwave radiation on protein folding28,29 via putative induced torsional vibrations28 have been reported, so it is possible that β-lg unfolded faster at 80 °C during MH than during CH. Nucleation also requires tertiary structure rearrangements, and would probably be accelerated by any increased backbone mobility that resulted from microwave excitation. The effects of microwave radiation on hydrolysis probably depend on the type and frequency of vibrational modes affected 2877

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

Table 6. Probable Matches for the Prominent Higher Molecular Weight Species (>12000 Da) Identified for the MHS Fibrils assignmenta

position

theoretical mass (Da)b

observed mass (Da)b,c

(A) PLRV....QCHI (Var B) (S) APLR....EQCH (I) (Var A) LIVT....PEVD (D) (Var A) (D) ISLL.... EQCH (I) (Var B)+ lactose (S) DISL....QCHI (Var B) (T) WYSL....QCHI (Var B) (D) IQKV....PTQL (E) (Var A) (D) IQKV....FNPT (Q) (Var B)+2 lactose LIVT....HIRL (S) (Var B) (K) GLDI....EQCH (I) (Var A) (L) IVTQ....EQCH (I) (Var A)d LIVT....EEQC (H) (Var A) (T) QTMK....QCHI (Var B)+lactosed (T) QTMK....EQCH (I) (Var A)+lactose (I) VTQT....EEQC(H) (Var A)+lactosed (L) IVTQ....QCHI (Var A) (L) IVTQ.... EEQC (H) (Var B)+lactose LIVT....QCHI (Var B)d (T) MKGL.... QCHI (Var B)+2 lactose (L) IVTQ.... EEQC (H) (Var A)+lactose LIVT....QCHI (Var A) LIVT....EEQC (H) (Var A)+lactose (L) IVTQ....EQCHI (Var B)+lactosed (I) VTQT....EEQC (H) (Var A)+2 lactose (L) IVTQ....QCHI (Var A)+lactose

38−162

14689.9

14689.8

37−161

14717.9

14717.8

1−129 29−161

14747.1 15191.5

28−162 19−162

15695.9 16430.9

14747.2 15516.0 (15191.7 + 324.3) 15696.2 16430.8

12−156

16659.1

16659.0

12−154

16360.8

1−149 9−161

17013.7 17375.9

17009.3 (16360.7 + 648.6) 17014.2 17376.2

2−161

18435.1

18435.0

1−160

18451.2

18451.5

5−162

18155.9

5−161

18166.7

3−160

18210.9

2−162 2−160

18541.4 18238.9

1−162 7−162

18566.5 17947.6

2−160

18326.0

1−162 1−160

18653.5 18451.2

2−162

18454.3

3−160

18213.8

2−162

18592.3

18480.4 (18156.1 + 324.3) 18490.9 (18166.6 + 324.3) 18535.1 (18210.8 + 324.3) 18541.5 18562.9 (18238.6 + 324.3) 18566.8 18595.7 (17947.1 + 648.6) 18650.3 (18326.0 + 324.3) 18653.5 18775.0 (18450.7 + 324.3) 18778.1 (18453.8 + 324.3) 18861.8 (18213.2 + 648.6) 18917.1 (18592.8 + 324.3)

Figure 11. Schematic representation of the proposed structure for MHS and 16 h CH fibrils. Large peptides are shown in blue. Intermediate peptides are in red, and smaller peptides are in green. Arrows show the axis of fibrils and the dashed rectangles show 2 βsheets lying parallel to the fibril axis.

large peptides, nuclei, or fibrils) diffusing together, and this process would be affected little by high-frequency protein backbone dynamics. However, the subsequent alignment of adjacent fibril building blocks into the correct arrangement to form stacked β-sheets69 may well be accelerated by microwaveinduced vibrations. Our results show that 2 h MH produced approximately the same amount of fibrils as 10 h CH (Figure 2A inset), and MHS fibrils contained larger peptides than 16 h CH fibrils (Table 5). This suggests that accelerated self-assembly was not due to accelerated hydrolysis. We propose a scenario in which nonthermal effects of microwave radiation accelerate protein unfolding, nucleation, and/or fibril growth more than they accelerate hydrolysis, thus partially decoupling hydrolysis from the other processes. On the basis of the evidence presented here, we cannot quantify the individual effects of microwave radiation on protein unfolding, nucleation, and fibril growth, and this remains a question for future investigation.



CONCLUSIONS MH accelerates the self-assembly of β-lg, but extending the microwaving period >2 h leads to the disintegration of fibrils. Storage of the 2 h MH β-lg sample increases the fibril yield to the maximum level that can be attained by the conventional method. Fibrils formed by the microwave method consist of a notable amount of large peptides (>6000 Da), which were absent from conventional fibrils. We believe that those parts of peptides not involved in β-sheets within the fibril core will appear as loops and hairs at the surface of fibrils. The appearance of large peptides and even intact β-lg monomers in microwave fibrils is attributed to a partial decoupling of hydrolysis from assembly, which appears to result from nonthermal effects of microwave radiation.

Peptide mass fingerprinting was done using the FindPept tool of the ExPasy proteomics server for the prominent peaks (area ≥100au). See Identification of Higher Molecular Weight Species under the Experimental Section for more details. bAverage masses are given. c No match was found for the masses 18506.8, 18920.6, and 18946.2 Da (not listed), probably due to unknown modifications. dSame sequence was identified for 2 or 3 other observed molecular masses (not given) due to the differences in the optional modifications. a

by microwaves, and how those vibrations impact specially on the first part of the C or N terminal hydrolysis of aspartic acid (D), which involves the formation of an intramolecular anhydride or a cyclic imide intermediate between the βcarboxyl group (which acts as a proton donor) and amide groups at either terminal.68 Whether microwave excitation would accelerate or decelerate hydrolysis is therefore hard to predict. Addition of a building block to a growing fibril or fibril nucleus starts with relatively large molecular species (proteins,



ASSOCIATED CONTENT

S Supporting Information *

Second derivative FTIR spectra for β-lg fibrils and β-lg. Additional TEM images for the 16 h MH sample. Effect of storage on the 2 h MH β-lg sample. TEM images for the effect of GuHCl on MHS and 16 h CH fibrils. TICs for the MHS and 16 h CH fibril samples. TEM image of 1 h MH β-lg sample. S0 2878

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

of MHS and 16 h CH fibrils at pH 3. Temperature and power profiles for the 2 h MH sample. This material is available free of charge via the Internet at http://pubs.acs.org.



(22) Loveday, S. M.; Wang, X. L.; Rao, M. A.; Anema, S. G.; Creamer, L. K.; Singh, H. Int. Dairy J. 2010, 20, 571−579. (23) Lefèvre, T.; Subirade, M. Biopolymers 2000, 54, 578−586. (24) Loveday, S. M.; Wang, X. L.; Rao, M. A.; Anema, S. G.; Singh, H. Food Hydrocolloids 2012, 27, 242−249. (25) Adamcik, J.; Mezzenga, R. Macromolecules 2012, 45, 1137− 1150. (26) Krebs, M. R. H.; Wilkins, D. K.; Chung, E. W.; Pitkeathly, M. C.; Chamberlain, A. K.; Zurdo, J.; Robinson, C. V.; Dobson, C. M. J. Mol. Biol. 2000, 300, 541−549. (27) Mishra, R.; Sörgjerd, K.; Nyström, S.; Nordigården, A.; Yu, Y. C.; Hammarström, P. J. Mol. Biol. 2007, 366, 1029−1044. (28) Bohr, H.; Bohr, J. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 2000, 61 (4 B), 4310−4314. (29) Li, D. F.; Mu, C. D.; Zhang, Q.; Zhou, Y. L.; Lin, W. J. Food Biochem. 2010, 34, 1319−1331. (30) Ponne, C. T.; Meijer, M. M. T.; Bartels, P. V. J. Agric. Food Chem. 1994, 42 (11), 2583−2588. (31) Sandoval, W. N.; Arellano, F.; Arnott, D.; Raab, H.; Vandlen, R.; Lill, J. R. Int. J. Mass Spectrom. 2007, 259, 117−123. (32) De Pomerai, D. I.; Smith, B.; Dawe, A.; North, K.; Smith, T.; Archer, D. B.; Duce, I. R.; Jones, D.; Candido, E. P. M. FEBS Lett. 2003, 543, 93−97. (33) Lill, J. R.; Ingle, E. S.; Liu, P. S.; Pham, V.; Sandoval, W. N. Mass Spectrom. Rev. 2007, 26, 657−671. (34) Hauser, N. J.; Basile, F. J. Proteome Res. 2008, 7, 1012−1026. (35) Swatkoski, S.; Gutierrez, P.; Wynne, C.; Petrov, A.; Dinman, J. D.; Edwards, N.; Fenselau, C. J. Proteome Res. 2008, 7, 579−586. (36) Kroes-Nijboer, A.; Venema, P.; Bouman, J.; Van Der Linden, E. Food Biophys. 2009, 4, 59−63. (37) Bolder, S. G.; Vasbinder, A. J.; Sagis, L. M. C.; van der Linden, E. Int. Dairy J. 2007, 17, 846−853. (38) Nilsson, M. R. Methods 2004, 34, 151−160. (39) Mills, O. E. Biochim. Biophys. Acta 1976, 434, 324−332. (40) Grothe, H. L.; Little, M. R.; Cho, A. S.; Huang, A. J.; Yuan, C. Mol. Vision 2009, 15, 2617−2626. (41) Kato, A.; Nakai, S. Biochim. Biophys. Acta 1980, 624, 13−20. (42) Sreerama, N.; Woody, R. W. Anal. Biochem. 2000, 287, 252− 260. (43) Bolder, S. G.; Sagis, L. M. C.; Venema, P.; Van Linden, E. D Langmuir 2007, 23, 4144−4147. (44) Schokker, E. P.; Singh, H.; Pinder, D. N.; Creamer, L. K. Int. Dairy. J. 2000, 10, 233−240. (45) Puchtler, H.; Sweat, F.; Levine, M. J. Histochem. Cytochem. 1962, 10, 355−364. (46) Howie, A. J.; Brewer, D. B. Micron 2009, 40, 285−301. (47) Alizadeh−Pasdar, N.; Li−Chan, E. C. Y. J. Agric. Food Chem. 2000, 48, 328−334. (48) Dong, A.; Matsuura, J.; Allison, S. D.; Chrisman, E.; Manning, M. C.; Carpenter, J. F. Biochemistry 1996, 35 (5), 1450−1457. (49) Dong, A.; Caughey, W. S. Methods Enzymol. 1994, 232, 139− 175. (50) Byler, D. M.; Susi, H. Biopolymers 1986, 25, 469−487. (51) Leonil, J.; Molle, D.; Fauquant, J.; Maubois, J. L.; Pearce, R. J.; Bouhallab, S. J. Dairy Sci. 1997, 80 (10), 2270−2281. (52) Fogliano, V.; Monti, S. M.; Visconti, A.; Randazzo, G.; Facchiano, A. M.; Colonna, G.; Ritieni, A. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1998, 1388, 295−304. (53) Van Durren, B. L. J. Org. Chem. 1961, 26, 2954−2960. (54) Pallarès, I.; Vendrell, J.; Avilés, F. X.; Ventura, S. J. Mol. Biol. 2004, 342, 321−331. (55) Iametti, S.; De Gregori, B.; Vecchio, G.; Bonomi, F. Eur. J. Biochem. 1996, 237, 106−112. (56) Török, M.; Milton, S.; Kayed, R.; Wu, P.; McIntire, T.; Glabe, C. G.; Langen, R. J. Biol. Chem. 2002, 277, 40810−40815. (57) Zandomeneghi, G.; Krebs, M. R. H.; McCammon, M. G.; Fändrich, M. Protein Sci. 2004, 13, 3314−3321. (58) Andreas, B. Prog. Biophys. Mol. Biol. 2000, 74, 141−173.

AUTHOR INFORMATION

Corresponding Author

*Phone: (+64) 9 3737599, ext. 86658. E-mail: l.melton@ auckland.ac.nz. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We wish to thank the Riddet Institute for funding and a Ph.D. Scholarship (C.A.H.), and the University of Auckland for a contribution to financial support. We acknowledge Dr. Adrian Turner and Martin Middleditch (SBS, University of Auckland) for providing assistance with electron microscopy and mass spectrometry, respectively. Michel Nieuwoudt and Dr. Celine Valery are thanked for their helpful discussions on FTIR analysis.



REFERENCES

(1) Edwards, P. B.; Creamer, L. K.; Jameson, G. B. In Milk Proteins: From Expression to Food; Thompson, A., Boland, M., Singh, H., Eds.; Academic Press: San Diego, CA, 2008; p 164. (2) Qin, B. Y.; Bewley, M. C.; Creamer, L. K.; Baker, H. M.; Baker, E. N.; Jameson, G. B. Biochemistry 1998, 37, 14014−14023. (3) Vijayalakshmi, L.; Krishna, R.; Sankaranarayanan, R.; Vijayan, M. Proteins: Struct., Funct., Genet. 2008, 71, 241−249. (4) Flower, D. R. Biochem. J. 1996, 318, 1−14. (5) Mercadante, D.; Melton, L. D.; Norris, G. E.; Loo, T. S.; Williams, M. A. K.; Dobson, R. C. J.; Jameson, G. B. Biophys. J. 2012, 103, 303−312. (6) Arnaudov, L. N.; de Vries, R.; Ippel, H.; van Mierlo, C. P. M. Biomacromolecules 2003, 4, 1614−1622. (7) Hill, E. K.; Krebs, B.; Goodall, D. G.; Howlett, G. J.; Dunstan, D. E. Biomacromolecules 2006, 7, 10−13. (8) Akkermans, C.; Venema, P.; van der Goot, A. J.; Gruppen, H.; Bakx, E. J.; Boom, R. M.; van der Linden, E. Biomacromolecules 2008, 9, 1474−1479. (9) Rasmussen, P.; Barbiroli, A.; Bonomi, F.; Faoro, F.; Ferranti, P.; Iriti, M.; Picariello, G.; Iametti, S. Biopolymers 2007, 86, 57−72. (10) Gosal, W. S.; Clark, A. H.; Ross−Murphy, S. B. Biomacromolecules 2004, 5, 2408−2419. (11) Akkermans, C.; Venema, P.; Van Der Goot, A. J.; Boom, R. M.; Van Der Linden, E. Food Biophys. 2008, 3, 390−394. (12) Bromley, E. H. C.; Krebs, M. R. H.; Donald, A. M. Faraday Discuss. 2005, 128, 13−27. (13) Ghadami, S. A.; Khodarahmi, R.; Ghobadi, S.; Ghasemi, M.; Pirmoradi, S. Biophys. Chem. 2011, 159, 311−320. (14) Gosal, W. S.; Clark, A. H.; Pudney, P. D. A.; Ross−Murphy, S. B. Langmuir 2002, 18, 7174−7181. (15) Veerman, C.; Ruis, H.; Sagis, L. M. C.; van der Linden, E. Biomacromolecules 2002, 3, 869−873. (16) Rogers, S. S.; Venema, P.; Sagis, L. M. C.; Van Der Linden, E.; Donald, A. M. Macromolecules 2005, 38, 2948−2958. (17) Adamcik, J.; Jung, J. M.; Flakowski, J.; De Los Rios, P.; Dietler, G.; Mezzenga, R. Nat. Nanotechnol. 2010, 5, 423−428. (18) Adamcik, J.; Mezzenga, R. Soft Matter 2011, 7, 5437−5443. (19) Bolisetty, S.; Adamcik, J.; Mezzenga, R. Soft Matter 2011, 7, 493−499. (20) Jones, O. G.; Handschin, S.; Adamcik, J.; Harnau, L.; Bolisetty, S.; Mezzenga, R. Biomacromolecules 2011, 12, 3056−3065. (21) Lara, C.; Adamcik, J.; Jordens, S.; Mezzenga, R. Biomacromolecules 2011, 12, 1868−1875. 2879

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880

Biomacromolecules

Article

(59) Juban, M. M.; Javadpour, M. M.; Barkley, M. D. Methods Mol. Biol. 1997, 78, 73−78. (60) Kad, N. M.; Thomson, N. H.; Smith, D. P.; Smith, D. A.; Radford, S. E. J. Mol. Biol. 2001, 313, 559−571. (61) Aouzelleg, A.; Bull, L. A.; Price, N. C.; Kelly, S. M. J. Sci. Food Agric. 2004, 84, 398−404. (62) Bateman, L.; Ye, A.; Singh, H. J. Agric. Food Chem. 2010, 58, 9800−9808. (63) Frare, E.; Polverino de Laureto, P.; Zurdo, J.; Dobson, C. M.; Fontana, A. J. Mol. Biol. 2004, 340, 1153−1165. (64) Inglis, A. S. Methods Enzymol. 1983, 91, 324−332. (65) Luca, S.; Yau, W. M.; Leapman, R.; Tycko, R. Biochemistry 2007, 46, 13505−13522. (66) Ritter, C.; Maddelein, M. L.; Siemer, A. B.; Luhrs, T.; Ernst, M.; Meier, B. H.; Saupe, S. J.; Riek, R. Nature 2005, 435, 844−848. (67) Harper, J. D.; Lansbury, P. T. Annu. Rev. Biochem. 1997, 66, 385−407. (68) Li, A.; Sowder, Ii, R. C.; Henderson, L. E.; Moore, S. P.; Garfinkel, D. J.; Fisher, R. J. Anal. Chem. 2001, 73, 5395−5402. (69) Giurleo, J. T.; He, X.; Talaga, D. S. J. Mol. Biol. 2008, 381, 1332−1348.

2880

dx.doi.org/10.1021/bm300896r | Biomacromolecules 2012, 13, 2868−2880