New insight into the alcohol induced conformational change and aggregation of the alkaline unfolded state of bovine β-lactoglobulin

Abstract

Accumulation of ordered protein aggregates (or amyloids) is responsible for several neurodegenerative diseases. β-Lactoglobulin (β-lg) an important globular milk protein, self-assembles to form amyloid-like fibrils on heating at low pH. But here we report for first time the self-assembly of β-lg from its alkaline unfolded state. The present work describes the folding and self-assembly of β-lg from a reversible unfolded state at pH 10.5 in the presence of methanol, 2-propanol, t-butanol and 2,2,2-trifluoroethanol (TFE). The extent of secondary and tertiary structure formation is in the order methanol < 2-propanol < t-butanol < TFE. Exposure of the hydrophobic core of the protein molecules in an apolar environment of TFE seems to promote intermolecular cluster formation. Methanol and TFE induce aggregation through the α-helical structure whereas isopropanol and t-butanol favour the formation of the β-structure leading to aggregation at higher concentrations. In vitro aggregation generates various nanometer structures such as nanofibrils, nanovesicles and nanotubes depending on the nature and concentration of the alcohols.

---

1 Introduction

During the last few decades the stabilization of proteins has become one of the big concerns of scientists. Due to the high complexity of protein structures, stabilization of protein molecules plays an important role in controlling the protein aggregation. Protein molecules are stabilized through a balance between the intramolecular interactions of functional groups and their interactions with the solvent environment.^1–4 Protein stability depends on electrostatic interactions, steric interactions, hydrogen bonding and hydrophobic interactions which are disturbed by the addition of osmolytes or co-solvents.^5–8 The instability of the protein structure leads to the formation of pre-molten globule states, molten globules (MGs), partially folded intermediates, and aggregates of native proteins. These misfolded, aggregated amyloid protein states are either a cause or a major evidential symptom for a variety of current neurodegenerative disease states which includes Alzheimer's disease, Huntington's disease, Parkinson's disease and diabetic-related disorders.^9,10 Protein aggregates are used to develop novel biomaterials for a wide range of application in biomedicine and biotechnology, also these protein aggregates are utilised in dairy industry to recover the whey proteins.^11a,b

Alcohols also play an important role to destabilize the native structure of proteins although some of them are used as protein precipitant during the purification of human serum albumin from human plasma. Alcohols such as methanol, ethanol, or 2,2,2-trifluoroethanol (TFE) when used as co-solvent, denature the tertiary structures of proteins while enhancing their helicity.^12–15 Several alcohols weaken non-local hydrophobic interactions while promoting local polar interactions (e.g., hydrogen bonds), stabilizing the extended helical rods in which hydrophobic side chains are exposed while polar amide groups are shielded from the solvent.^16–18 In comparison with non-fluorinated alcohols, TFE is often preferred for such studies because of its high potential to stabilize α-helical structures.^19,20 The effectiveness of TFE in inducing secondary structures appears to be related to its three fluorine atoms.²¹ Methanol, ethanol, or 2,2,2-trifluoroethanol (TFE) not only stabilize the native-like secondary structure but also transform other structural elements of the protein leading to the formation of non-native structures. The secondary structures stabilized by TFE are believed to reflect the conformations existing in early stages of protein folding.^20,22,23 Molten globule (MG), a compact collapsed form of a protein with a pronounced secondary structure but lacking rigid tertiary structure, is such an intermediate conformation.^24,25 However, in several instances, the MG has been shown to possess well-defined tertiary contacts, i.e., a pre-molten globule state.^26,27 Study of these structural intermediates of a protein could provide clues to the nature of protein folding.

The whey proteins, β-lactoglobulin (major), α-lactalbumin, immunoglobulin, bovine serum albumin, lactoferrin, enzymes etc., have excellent nutritional value, are eliminated in traditional cheese-making processes during drainage. In the native state or partly denatured state of these proteins have superb foaming and emulsion forming property and in completely denatured state they display gelation and high water retention properties. Denaturation of the whey proteins by heating of milk in different suitable pH and aggregation of it with casein fraction is one of the way to recover the whey protein part in cheese technology.^28a,b This prompted us to find new technical access to optimise the whey proteins recovery without heating.

Bovine β-lg is a well known globular whey protein (18 000 Da, pI 5.2). It comprises 162 amino acid residues and is predominantly a β-sheet protein. Upon heating above 60 °C, the native structure of β-lg begins to change, with a decrease in amount of ordered zones and an increase in the exposure of the tryptophan and free thiol group.^29,30

At pH 7.0 the protein exists as a reversible dimer and extent of dimerization depends on pH, temperature, protein concentration and ionic strength of the medium. Furthermore, β-lg forms soluble aggregates of polymerized protein after heating 80 °C for 1 h at pH ∼ 7.0.³¹ Therefore, the most of the studies on structural and aggregation properties of β-lg have been done in the acidic or neutral pH under thermal conditions. Our previous studies demonstrated co-solvent induced conformational changes and aggregation of bovine β-lg at pH 7.0.³² In the present study, we have demonstrated the effect of non-fluorinated and fluorinated alcohols on the refolding and aggregation of alkaline denatured bovine β-lg at pH 10.5. To the best of our knowledge the study of folding and aggregation property of β-lg at alkaline pH is not yet done properly.

2 Experimental

2.1. Materials

Bovine β-lactoglobulin (β-lg) was isolated and purified from cow milk as described earlier.³³ Methanol and different fluorescent probes, viz., 8-anilinonaphthalene, 1-sulfonic acid ammonium salt (ANS), Congo Red (CR) as well as Thioflavin T (ThT) were obtained from Sigma Chemical Co. (St. Louis, USA) and used as received without further purification. 2-Propanol (IP), t-butanol, sodium phosphate dibasic, glycine and KOH were purchased from Merck (Mumbai, India) (all were AR grade). 2,2,2-Trifluoroethanol (TFE) (spectroscopic grade) was purchased from Spectrochem Pvt. Ltd. (India). The other chemicals used were of highest purity available.

2.2. Methods and instrumentation

2.2.1. Protein samples preparation. All the measurements were carried out at room temperature. Since the extinction coefficient of β-lg (0.96 mg⁻¹ ml⁻¹ cm⁻¹ at 280 nm) is known, different concentrations of protein samples were prepared by dissolving β-lg samples in buffer solution of pH 10.5. All the experiments for native β-lg were performed in 10 mM sodium phosphate buffer pH 7.0. For the study at alkaline pH, β-lg was dissolved in 3 mM glycine–KOH buffer of pH 10.5 respectively to give a stock solution of 271 μM in each case. To the stock protein solutions, different volumes of buffer were added first, followed by the addition of methanol, 2-propanol, t-butanol and TFE (assumed to be 100%, v/v) to get a desired concentration of co-solvent in each case. The final solution mixture (1.0 ml) was incubated for 4 h at room temperature before each experiment. The required protein concentrations were adjusted in each experiment.

2.2.2. Analysis of secondary structures by CD spectroscopy. To investigate the conformational changes of β-lg in presence of different non-fluorinated and fluorinated alcohols, circular dichroism measurements were carried out on a Jasco spectropolarimeter (J-815) at 20 °C in the far-UV region (190–260 nm) using the rectangular cell of 1 mm path length. The β-lg solutions equilibrated separately with non-fluorinated and fluorinated alcohols having protein concentration 13.6 μM. All the spectra are average of three scans. The final spectrum was obtained after the subtraction of corresponding solvent spectrum. The far UV-CD curves were fitted into a curve-fitting program CDNN 2.1 to determine the amount of secondary structures present in β-lg in presence of methanol, 2-propanol, t-butanol and TFE acting as co-solvents.

2.2.3. Intrinsic fluorescence study. Fluorescence measurements were performed on a Shimadzu spectrofluorimeter (Shimadzu 5301 PC). The fluorescence spectra were collected at 25 °C using 1 cm path-length cell and protein concentration 13.6 μM in absence and presence of different non-fluorinated and fluorinated alcohols. The excitation and emission slits were set at 5 nm. Intrinsic fluorescence spectra were recorded in the wavelength region 310 to 400 nm after exciting the protein sample at wavelength of 295 nm.

2.2.4. 1-Anilinonaphthalene-8-sulfonate (ANS) fluorescence studies to monitor the hydrophobicity. Exposure of hydrophobic patches in protein during the aggregation process was monitored using polarity sensitive fluorescent probe 1-anilinonaphthalene-8-sulfonate (ANS).³⁴ A stock solution of ANS was added to each aliquot of β-lg solution (both in absence and presence of different non-fluorinated and fluorinated alcohols) so that the final ANS concentration in each aliquot was 30 μM. Typically, ANS concentration was 50 molar excess of protein concentration. The ANS-fluorescence intensities were measured using Shimadzu RF-5301 PC with excitation at 370 nm (ref. 35 and 36) and scanning the emission wavelength from 400 nm to 650 nm. Slit widths were set at 5 nm for both excitation and emission. Each spectrum was blank corrected. Data points were the average of triplicate measurements.

2.2.5. Thioflavin T (ThT) assay. ThT is a dye which shows enhanced fluorescence at 480 nm when bound to amyloid fibrils.³⁷ Thus to investigate and compare the aggregates formed by β-lg in presence of various non-fluorinated and fluorinated alcohols, the following assay was employed. Briefly 250 μl of β-lg samples having concentration 54.3 μM was taken. It was then added to 40 μl ThT solution (stock 3.13 mM ThT in 10 mM sodium phosphate buffer, pH 7.0) containing a mixture of buffer and varying amount of different alcohols, mixed thoroughly and incubated for 30 min.³⁸ The final concentration of protein was 6.8 μM while the concentration of ThT was 30 μM. The assay solution was excited at 450 nm (ref. 39) and the emissions were measured over the range 460 to 600 nm. Slit widths for both excitation and emission were kept at 5 nm. Three replicates were performed and the data were averaged.

2.2.6. Congo red assay. The formation of aggregates in presence of alcohols was probed by measuring the shift in absorbance of Congo red in the region 400–650 nm. For this experiment, 240 μl (27.2 μM) aliquots of the protein solutions were withdrawn and mixed with 260 μl of a solution containing 40 μM Congo red solution. Final volume (2 ml) was adjusted with 3 mM glycine–KOH buffer of pH 10.5 and varying concentrations of various non-fluorinated and fluorinated alcohols.⁴⁰

2.2.7. Dynamic light scattering (DLS) measurements. The diffusion of nanosized particulates in solution induces fluctuations in the intensity of the scattered light. DLS detects these fluctuations using an auto correlator on a microsecond time scale and is used to analyze the distribution of the molecules and supramolecular aggregates as it is very sensitive to particle size.⁴¹ Different sizes of molecules in the solution can be observed in different peaks provided their sizes vary sufficiently. In our experiment, DLS measurements were performed with β-lg solutions in absence and presence of methanol, 2-propanol, t-butanol and TFE employing Zetasizer Nanos (Malvern Instrument, U.K.) equipped with 633 nm laser and using 2 ml rectangular cuvette (path length 10 mm). Measurements were done at 20 °C taking 250 μl of β-lg sample in 1.75 ml 3 mM glycine–KOH buffers of pH 10.5. Then different amounts of non-fluorinated and fluorinated alcohols (6–80%) were mixed thoroughly and allowing to equilibrate the solutions. The time-dependent auto correlation function was acquired with twelve acquisitions for each run. Each data is an average of five such acquisitions.

2.2.8. Monitoring of secondary structure of β-lg during aggregation by FT-IR spectroscopy. For FT-IR measurements, 50 μl β-lg sample solutions (in absence and presence of methanol, 2-propanol, t-butanol and TFE) having concentrations 1087 μM were taken in a microcon filter device and diluted with 200 μl of D₂O. It was then quickly centrifuged at 4000 × g for 8 min until the volume reached ∼50 μl. After that 200 μl of D₂O was added again and centrifuged for another 8–10 min. This process of D₂O exchange was repeated 3–4 times.³⁹ Finally, the D₂O exchanged β-lg samples were placed between two CaF₂ windows separated by a 50 μm thick Teflon spacer. FT-IR scans were collected in the range of 1550–1750 cm⁻¹ at a resolution of 2 cm⁻¹ in N₂ environment using a Spectrum 100 FT-IR spectrometer (Perkin Elmer). Spectrum of D₂O at pD 7.0 was collected and subtracted from sample spectrum.

2.2.9. Morphological studies with SEM and TEM.
2.2.9.1. Field emission scanning electron microscope (FESEM). The morphology of the aggregates of β-lg generating from U_A state in various alcoholic co-solvents at different concentrations were investigated using FE-SEM (Hitachi S-4800, JAPAN) operating with a voltage of 20 kV. For this study the protein sample (U_A state) was incubated with MeOH, i-PrOH, t-BuOH and TFE and the concentrations of proteins were maintained at 10 μM. One drop of the sample solution was taken on a glass slide. It was dried by slow evaporation in open air and then under vacuum and gold coated for imaging.
2.2.9.2. Transmission electron microscopy (TEM). The morphology and size of the aggregates of β-lg obtained from U_A state after incubation separately with non-fluorinated and fluorinated alcohols were investigated by high resolution transmission electron microscopy (Jeol-HRTEM-2011, Tokyo, Japan) with an accelerating voltage of 80–85 kV in different magnifications. The sample solutions were diluted 50 times in 3 mM glycine–KOH buffer of pH 10.5. A droplet of the diluted sample was put on a carbon coated copper grid of mesh size 300C (Pro Sci Tech). After 20 s the droplet was removed with a filter paper followed by a droplet of 2% uranyl acetate (Sigma, Steinheim, Germany) solution put on the grid and finally removed after 15 s and left for air dry and used for imaging purpose. Before taking the image all the samples were incubated 6 h.

3 Results and discussion

3.1. Stabilization of secondary structures in alkaline unfolded U_A state of β-lg by alcohols

CD-spectropolarimetry is an important biophysical technique to elucidate the secondary structures of proteins and peptides in solution.⁴² Using the spectra-structure correlation method, the change in the secondary structure of β-lg in the alkaline unfolded U_A state with varying concentrations of non-fluorinated and fluorinated alcohols monitored by far-UV CD spectroscopy. Alcohols are well-known secondary structure inducer in proteins.^11–17 Alcohols weaken non-local hydrophobic interactions at the cost of promoting local polar interactions (i.e., hydrogen bond) in proteins. Alcohol-induced rearrangement is accompanied by stabilization of the extended helical rods in which hydrophobic side chains are solvent-accessible and polar amide groups are withdrawn from the solvent.⁴³ Fig. 1 displays the far-UV CD spectra in the region between 190 and 260 nm range of β-lg at pH 10.5 in the presence of various concentrations of MeOH (A), i-PrOH (B), t-BuOH (C) and TFE (D). Result shows that both the non-fluorinated and fluorinated alcohols have induced the secondary structures in U_A state of β-lg (Fig. 1A–D). Significant change in the CD spectra, either shape or intensity, was observed by the addition of alcohols. Alkaline unfolded state (U_A) of β-lg shows a single negative peak near 208 nm, whereas the peak near 215 nm, which is a characteristic feature of β-sheet structure is absent in all the cases in absence of alcohols at this alkaline pH.⁴⁴ In the presence of higher concentrations (70% and 80%) (v/v) methanol, two minima near 208 nm and 222 nm, which are the characteristics features for α-helical structure, are observed (Fig. 1A). The minima at 215 nm became prominent when 60–70% (v/v) i-PrOH and t-BuOH have been added indicating the formation of β-sheet structure with the concomitant loss of α-helical structure (Fig. 1B and C). The percentage of α-helix gradually increases along with the increasing concentrations of i-PrOH and t-BuOH up to 60% (v/v) and 50% (v/v) and then it decreases probably due to the increase in bulkness from MeOH to t-BuOH (Fig. 1E). Interestingly, such helical structure formation has been started only in the presence of 15% (v/v) TFE (Fig. 1D) and the percentage of α-helix increased slowly even at 30% (v/v) TFE concentration (Fig. 1F). Thus the fluorinated alcohol TFE seems to be more effective than the non-fluorinated alcohols in inducing secondary structure in the protein. Several β-sheet and α/β proteins show similar trend.⁴⁵

Fig. 1 Far-UV CD spectra of β-lg (13.6 μM) in native state in 10 mM sodium phosphate buffer pH-7.0 (curve a) and in alkaline unfolded state (curve b) in presence of 20%, 40%, 50%, 60%, 70% and 80% (v/v) MeOH (curve c–h) (A), 10–70% (v/v) i-PrOH (curve c–i) (B), 10–60% (v/v) t-BuOH (curve c–h) (C) and with 1%, 3%, 6%, 8%, 10%, 12%, 14%, 15%, 17%, 18%, 20%, 25% and 30% (v/v) TFE (curve c–o) (D) at pH 10.5 in 3 mM glycine–KOH buffer. Percentage of α-helix formed at 222 nm in the presence of MeOH, i-PrOH and t-BuOH (E) and TFE (F) in alkaline unfolded β-lg. The far-UV CD spectra were recorded between 200 nm and 260 nm and the path length was 1 mm.

3.2. Influence of non-fluorinated and fluorinated alcohols on the tertiary structure of β-lg in the U_A state

The effect of non-fluorinated and fluorinated alcohols on the tertiary structure of β-lg in the alkali unfolded (U_A) state was monitored using steady-state tryptophan fluorescence spectroscopy. It is known that the fluorescence properties of proteins are sensitive to the microenvironment of the three amino acids, tryptophan (Trp), tyrosine (Tyr) and phenylalanine (Phe), which have intrinsic fluorescence properties. However, the contributions from Tyr and Phe residues are insignificant when the protein is excited at or above 295 nm. Thus, in the present study, to minimize the complexity arising from the contributions from the Tyr and Phe residues, the fluorophores of β-lg were excited at 295 nm. Generally, the two parameters, emission wavelength maximum (λ_max) and emission intensity maximum (I_max), were used to obtain information on the structure and dynamics of proteins.^46,47 The polarity around the tryptophan in the protein is monitored through the measurement of λ_max and it is sensitive to the tertiary structure. The I_max is high when the Trp residue is exposed to hydrophobic environment while it shows low when Trp is exposed to hydrophilic environment. The λ_max of tryptophan derivative in solution is also very sensitive to the polarity of the solvent.⁴³ In general, the λ_max of Trp residue shifts toward a lower wavelength (blue shift) when the polarity of the solvent diminishes.⁴⁸

There are two tryptophan residues in bovine β-lg at residues 19 and 61, respectively. Report on X-ray structure reveals that Trp19 is located within the central hydrophobic calyx of the protein whereas Trp61 is close to the disulfide bond Cys160–66, and lies close to the protein surface. It was suggested by Mills, 1976 that the neighbouring disulfide bond, Cys160–66, considerably quenches the fluorescence intensity of Trp61.⁴⁹ Trp19 therefore mostly dictates tryptophan fluorescence of bovine β-lg.

In our study, the U_A state of β-lg at pH 10.5 exhibits a prominent emission peak around 350 nm in the wavelength range of 300–400 nm when excited at 295 nm. Hence the relative change in the fluorescence intensities around 350 nm can be used as a probe for the change of microenvironment in the vicinity of the tryptophan residues of protein arising due to altered conformations. Fig. 2A reveals the normalized relative fluorescence intensities at 350 nm of the unfolded β-lg (U_A) against increasing concentrations of various alcohols after exciting it at 295 nm. The λ_max of emission shifted from 332 at pH 7.0 to 350 nm at pH 10.5. The observed increase of λ_max in the pH region 10.5 can be attributed to the loss of tertiary structure resulting from unfolding of the protein. Except TFE, the other non-fluorinated alcohols have induced significant and continuous increase in the fluorescence intensity at concentrations range up to 20%. TFE showed similar change at very lower concentration [nearly at 5% (v/v)]. Then the gradual lowering of fluorescence intensities was observed after reaching the higher concentrations of the alcohols [70% (v/v) for i-PrOH and 60% (v/v) for t-BuOH] except the MeOH (Fig. 2A) indicating the formation of tertiary structures from the alkaline unfolded U_A structure. Here, the highly unfolded protein may first regained its secondary structure and then becomes stacked in less polar environment leading to aggregation at very high concentrations.

Fig. 2 (A) Normalized relative fluorescence intensity of tryptophenyl residues (emitted at 350 nm) of alkaline unfolded β-lg (pH 10.5) with increasing concentrations of MeOH, i-PrOH, t-BuOH and TFE, when the sample was excited at 295 nm. Protein concentration was 13.6 μM. (B) Relative fluorescence intensity at 480 nm of alkaline (pH 10.5) unfolded β-lg–ANS complex with increasing concentration of MeOH, i-PrOH, t-BuOH and TFE after exciting at 370 nm.

Except methanol, other alcohols have induced both secondary and tertiary conformations at higher range of concentrations (>20%). Methanol being more polar can stabilize the secondary structure only. The highly randomized U_A state of β-lg seems to acquire secondary as well as tertiary contacts in the presence of alcohols. The lesser dielectric constant of alcohols as compared to water has favored the intermolecular hydrogen bonding and electrostatic interactions. The higher alcohol concentrations weaken the hydrogen bonding between water and protein molecules and favor the formation of strong hydrogen bonds between the main chain amide and carbonyl group. The highly randomized U_A state seems to acquire secondary as well as tertiary structures in the presence of alcohols in the order of effectiveness MeOH < i-PrOH < t-BuOH < TFE. TFE, even at very low concentration, enhances hydrophobic interactions,⁵⁰ which may be responsible for the stabilization of protein structure and also enhances the aggregation at higher concentration regions. The lesser polarity of TFE may have favored intermolecular hydrogen bonding and electrostatic interactions.

3.3. Influence of non-fluorinated and fluorinated alcohols on ANS-fluorescence

The fluorescence emission spectra arising from molecules of an external probe, ANS were measured at different alcohol concentrations (0–80%) (v/v) at pH 10.5 to study global conformational changes in the presence of non-fluorinated and fluorinated alcohols. The source of ANS interaction with β-lg may be due to both electrostatic and hydrophobic interactions.⁵¹ ANS interacting with β-lg shows an emission maximum around 480 nm. Earlier reports show binding of ANS to the molten globule state of lysozyme at alkaline pH produced a large increase in fluorescence intensity compared to native state due to exposure of sizeable amount of hydrophobic pockets in the unfolded state.⁵² Fig. 2B shows relative fluorescence intensities (RFI) at 480 nm of β-lg–ANS complex at pH 10.5 against the increasing concentrations of MeOH, i-PrOH, t-BuOH and TFE. As alcohols are hydrophobic in nature we have done proper baseline correction with the appropriate control solution. No significant ANS bindings were observed for i-PrOH and t-BuOH up to 20% (v/v) concentrations, and then it increased continuously in the concentration range 20–80% (v/v). But above 40% (v/v), it has shown continuous increase in fluorescence intensity for i-PrOH and t-BuOH. The sharp increase of ANS fluorescence in the presence of higher concentrations of i-PrOH and t-BuOH might be an indication of formation of some intermediate state which is more hydrophobic than alkali unfolded (U_A) state of β-lg. Methanol being more polar is thus less effective in the formation of such intermediate state. Fig. 2B also shows the emission spectra of ANS in the presence of β-lg at different TFE concentrations. TFE-induced conformational changes in the U_A state of β-lg led to a gradual but nevertheless significant increase in ANS binding even at very lower concentration 5% (v/v) than the non-fluorinated alcohols. This shows that more hydrophobic clusters of buried region are exposed in the TFE relative to U_A state.

3.4. Formation of β-lg self-assembly with non-fluorinated and fluorinated alcohols studied by ThT assay

The formations of β-lg self-assembly from alkaline unfolded U_A state at pH 10.5 in presence of various concentrations of alcohols were studied by ThT assay. ThT is a cationic benzothiazole dye showing enhanced fluorescence upon binding to protein assembly in vitro. It binds to intermolecular β-sheet present in the aggregates.⁵³ Here, the alkaline unfolded β-lg may first regains its secondary structure and then becomes stacked in less polar environment of the alcohols leading to self-assembles when incubated with different non-fluorinated and fluorinated alcohols. Hence in order to study the self-assembles and hydrophobic residue in β-lg during incubation with varying amounts alcohols, a solution of U_A state containing ThT was incubated in absence and presence of non-fluorinated and fluorinated alcohols and the change in fluorescence was monitored (Fig. 3A and B). The ThT fluorescence intensities did not vary widely up to 40% (v/v) addition of MeOH, i-PrOH and t-BuOH (Fig. 3A). The result might be an indicative to a process of the gradual regaining the native conformation of alkali induced unfolded U_A state through the stabilization of secondary and hence the tertiary structure in less polar environment of the alcohols in the pre-self-assembly step. The t-BuOH has induced greater effects due to its larger size and low polarity. Similar result is achieved when U_A is incubated in presence of only 5% (v/v) TFE (Fig. 3B). The more hydrophobic environment of TFE may have favoured intermolecular hydrogen bonding and electrostatic interactions during the stabilization of secondary and tertiary structures.

Fig. 3 Bar diagram of the end-point ThT intensity versus (A) MeOH, i-PrOH and t-BuOH concentrations (B) TFE concentrations in ThT assay to study the aggregation of β-lg in 3 mM glycine–KOH at pH 10.5 buffer. Fluorescence emissions were monitored in the wavelength range 460–600 nm after excitation at 450 nm. Standard deviations are within the range of ±3.0. A control experiment of ThT without β-lg has been performed.

Then the ThT fluorescence intensities increased regularly on gradual addition of MeOH, i-PrOH and t-BuOH from 40% (v/v) to 80% (v/v) whereas only 30% (v/v) TFE shows even larger ThT fluorescence intensities. At moderate concentrations, (40–60%) (v/v) of non-fluorinated and (5–15%) (v/v) of TFE, the protein molecules further unfolds and undergoes a rapid aggregation when incubated with 60–80% (v/v) t-BuOH, i-PrOH, MeOH, or 30% (v/v) TFE. Therefore TFE is more effective in promoting the amyloid-like protein aggregation at much lower concentration and thus causing the increase in Th-T fluorescence emission. TFE has previously been shown to induce aggregates distinct from amyloids.⁵⁴ Given the induction of a non-native helical structure in the presence of 30% (v/v) TFE, it seems possible that TFE at 30% (v/v) promote intermolecular cluster formation.

3.5. Congo red assay to detect the protein assembly

The change of conformation and hence the aggregation of β-lg from the U_A state at the higher concentrations of MeOH, i-PrOH, t-BuOH (60–80%) (v/v) and TFE (20–30%) (v/v) can also be verified with the Congo red assay. Congo Red (CR) is a dye that binds preferentially with β-sheet structure of the aggregates but not to the native. CR is structurally similar to ThT and equally employed to detect protein aggregates. Binding of CR with the aggregates induces a characteristic increase in the absorption maxima from 480 nm to 490 nm.⁵⁵ To investigate the formation of such aggregates from alkaline unfolded U_A state of β-lg, the CR absorption spectra of UA state in the presence of various concentrations MeOH, i-PrOH, t-BuOH and TFE were recorded (Fig. 4). The U_A state showed absorption maxima at 495 nm (curve a, Fig. 4A). The addition of 10% MeOH causing the enhancement of CR absorption intensities and then no significant change was observed up to 30% addition (Fig. 4A) indicating the process of stabilization of the secondary and tertiary structures of U_A in less polar solvent. Again the increase of CR absorption intensities and red shift of absorption maxima were observed at higher concentrations (70–80%) (v/v) of MeOH, showing the formation of β-lg aggregates at this concentrations. The CR dye was trapped into the cluster of the aggregates. Our observation is in agreement with the earlier report of the aggregation of β-lg with MeCN.³² Similar intensity enhancement and red shift of absorption maxima were observed with other non-fluorinated and fluorinated alcohols like i-PrOH, t-BuOH and TFE (Fig. 4B–D). But t-BuOH seems to be more effective in inducing secondary structure and self-assembly in the protein at lower concentrations 60% (v/v) than MeOH, and i-PrOH due to the hydrophobic nature of t-butyl group. TFE, being a well-known α-helix inducer, has induced α-helix and enhances the self-assembly at very low concentration regions compared to other non-fluorinated alcohols.

Fig. 4 Congo red absorption spectra of β-lg in native state in 10 mM sodium phosphate buffer pH-7.0 (curve a) and in alkaline unfolded state (curve b) in 3 mM glycine–KOH buffer at pH 10.5 in presence of (10–80%) (v/v) of MeOH (curve c–j) (A), i-PrOH (curve c–j) (B), t-BuOH (curve c–j) (C) with interval of 10 and in presence of 1%, 6%, 10%, 12%, 15%, 20%, 25% and 30% TFE (curve c–j) (D). The absorption spectra were recorded from 400 to 600 nm. The protein concentration was 27.2 μM.

3.6. Analysis of secondary structure of β-lg in presence of non-fluorinated and fluorinated alcohols

The FT-IR spectra of β-lg samples (U_A state and the self-assembled states in higher concentrations of MeOH, i-PrOH, t-BuOH and TFE) in D₂O-buffer are shown in Fig. 5A. The resulting spectra for U_A state (curve-a) and the self-assembled states (curve b–e) of β-lg show the amide-I band at around 1650 cm⁻¹. Native β-lg molecule in absence of any alcohols shows the amide-I band at around 1632 cm⁻¹ which is the characteristic features for the protein like β-lg having predominant β-sheet structure.⁵⁶ Most of the information on the protein's secondary structure is contained in the amide-I band in the region 1700–1600 cm⁻¹. Thus the shifting of amide-I band of native β-lg occurs from 1632 cm⁻¹ to 1650 cm⁻¹ in U_A state and also in the self-assembled states in presence of high concentration of alcohols, showing a clear transition of β-sheet to α-helix structure. The observation is consistent with the far UV-CD result. This intermediate α-helix structure is unstable and latter gets transformed into the intermolecular β-sheet structure during its aggregation pathway when β-lg is subjected to 70% (v/v) of i-PrOH, 60% (v/v) of t-BuOH. TFE has induced α-helix and enhances the self-assembly at very low concentration regions compared to other non-fluorinated alcohols. Furthermore the amide-II band at 1480 cm⁻¹, a characteristic for the protein having predominant β-sheet structures^57,58 in the aggregates was observed in all the cases except the U_A state.

Fig. 5 (A) Fourier transform infra red (FTIR) spectra of alkaline unfolded β-lg (curve a) and native β-lg (inset), in the presence of 80% (v/v) MeOH (curve b), 70% (v/v) i-PrOH (curve c), 60% (v/v) t-BuOH (curve d) and 30% (v/v) TFE (curve e) respectively at pH 10.5 in 3 mM glycine–KOH. Protein concentrations were 1087 μM. Each spectrum is an average of 32 scans in D₂O solvent at 25 °C. (B) Number-particle size distribution profile of native β-lg (curve a), alkaline unfolded β-lg (curve b) in the presence of 80% (v/v) MeOH (curve c), 70% (v/v) i-PrOH (curve d), 60% (v/v) t-BuOH (curve e) and 6% (v/v), 15% (v/v) and 30% (v/v) (curve f–h) TFE respectively in same buffer. Each of these spectra is an average of 48 scans.

3.7. Dynamic light scattering (DLS) study

To investigate the formation of β-lg aggregate in the presence of non-fluorinated and fluorinated alcohols DLS study was employed (Fig. 5B). DLS is frequently employed to measure the hydrodynamic radius of spherical particles in solution and the intensity of scattered light is dominated by large particles present in solution. Although the β-lg aggregates are heterogeneous, either spherical or rod-like fibrous shapes, DLS data analysis can provide a quantitative estimation of aggregates distribution in MeOH, i-PrOH, t-BuOH and TFE. The data were plotted as scattered light intensity versus size in radius. In absence of alcohols, the hydrodynamic diameter of U_A state of β-lg is in agreement with that of native protein as previously reported. The size distribution profile of β-lg in the presence of higher concentrations of MeOH 80% (v/v) showed the formation of bigger molecules and aggregates of different size ranging from 10 nm to 50 nm. Similarly the particles of hydrodynamic radius greater than 50 nm were dominated when U_A was incubated either with 70% (v/v) i-PrOH or 60% (v/v) t-BuOH. Interestingly, the oligomerization of β-lg was initiated only with 6% TFE and further addition of 15% (v/v) and 30% (v/v) TFE made wider distribution of aggregate molecules with larger hydrodynamic radius greater than 100 nm and 200 nm respectively. The width of the DLS peaks both in the presence of TFE become wider indicating the heterogeneity in the oligomeric state leading to the self assembly of β-lg.

3.8. Morphological studies with SEM and TEM

The morphological characteristics of the aggregates of β-lg arising from alkaline unfolded U_A state in various alcoholic co-solvents at different concentrations are shown in the Fig. 6a–f. The SEM image (Fig. 6a) of U_A does not show any distinct self assembled structure whereas some elongated worm-like aggregates of 50–60 nm in apparent diameter are formed upon addition of 80% (v/v) MeOH (Fig. 6b). Again Fig. 6c shows the formation of large number of very small spherical shaped β-lg nanoparticles (5–10 nm) when U_A is incubated with 70% (v/v) i-PrOH. These short species bind together to form higher order aggregates. In 60% (v/v) t-BuOH, a different long sheet-like morphology of the aggregates was observed [Fig. 6d]. In contrast, flake-like aggregates were formed in 15% (v/v) TFE and at 30% (v/v), nanotubes-like aggregates (average diameter 5–10 nm) are found to form to generate multilayer three dimensional frameworks [Fig. 6f]. Thus the morphology of the aggregates depends on the nature and concentration of the alcohols. The multilayer-mediated self-assembly mechanism supports the formation of tubes.

Fig. 6 FE-SEM images showing the formation of distinct self assembled structure of β-lg at pH 10.5 and at: β-lg in absence of any alcohol (a) and native β-lg (inset), worm-like aggregates of 50–60 nm diameter in 80% (v/v) MeOH (b), smaller spherical nanoparticles in 70% (v/v) i-PrOH (c), long sheet-like aggregates in 60% (v/v) t-BuOH (d), flake-like and nanotube-like morphology with 15% (v/v) (e) and 30% (v/v) TFE (f).

The heterogeneity in the morphology of protein aggregates in various alcohols is revealed by TEM. Results showed (Fig. 7b) that morphology of the β-lg aggregates is completely amyloid fibrillar in nature at 80% (v/v) concentration of MeOH. The amyloid fibrillar network connects to generate the nanovesicular like structures (average diameter 40–60 nm) when the alkaline unfolded state is incubated with 70% (v/v) i-PrOH as confirmed by TEM imaging (Fig. 7c). TEM image supports the formation of large number of spherical shaped nanoparticles along with some vesicular structures when U_A is treated with 60% (v/v) t-BuOH (Fig. 7d). The less polar and more hydrophobic fluorinated alcohol TFE stabilizes the secondary structures and enhances the self-assembly formation at much lower concentration. A drastically different image (Fig. 7e) is visible in presence of 30% (v/v) TFE where the self-assembly leads to the formation of nanotube like structure of different length and of the average thickness of 5 nm. These morphological differences are consistent with the ThT binding capacity.

Fig. 7 Selected TEM images of β-lg aggregates at pH 10.5 and at 30 °C: β-lg in absence of any alcohol (a), amyloid fibrillar network in 80% (v/v) MeOH (b), nanovesicular structures in 70% (v/v) i-PrOH (c), spherical nanoparticles-shaped structures in 60% (v/v) t-BuOH (d), formation of nanotube of the average thickness of 5 nm in 30% (v/v) TFE (e). Images were taken after 6 h incubation and the protein concentrations were 10 μM.

4 Conclusion

The structural changes along the path of folding of alkaline unfolded beta lactoglobulin (β-lg) in presence of three non-fluorinated alcohols MeOH, i-PrOH, t-BuOH and a fluorinated alcohol, 2,2,2-trifluoroethanol (TFE) have been assessed. Result shows that all the alcohols have stabilized the secondary and tertiary structures of alkaline unfolded β-lg. The highly randomized U_A state seems to regain the secondary as well as tertiary structures in the presence of alcohols in the order of effectiveness MeOH < i-PrOH < t-BuOH < TFE. The t-BuOH being less polar and more bulky can induce secondary and tertiary structures at lower concentration. Similarly TFE at much lower concentration favored the intermolecular hydrogen bonding by lowering the dielectric constants. All of them have increased α-helicity and stabilized the secondary structure at lower concentrations. The β-sheet structure formation occurs with the concomitant loss of α-helical structure at 70% (v/v) and 60% (v/v) of i-PrOH and t-BuOH. The higher concentration of both non-fluorinated and fluorinated alcohols accelerated the formation of non-native secondary structure leading to the formation of protein self-assembly. Interestingly TFE has induced aggregation of β-lg through the formation of only α-helical structure. Similar evidences were observed with other proteins like bovine serum albumin, bovine serum fetuin etc.^59,60 TEM images showed the heterogeneity in the morphology of protein aggregates formed in various alcohols. Morphologically heterogeneous species of protein aggregates ranging from nanofibrillar to nanotube-like structures of different length and thickness were generated depending on the nature and concentration of the alcohols. TFE being less polar and more hydrophobic stabilizes the secondary structures and enhances the self-assembly formation at much lower concentration. Thus we performed a systematic study on the conformational transitions of alkaline unfolded state of β-lg at pH 10.5 in the presence of TFE and three other non-fluorinated alcohols MeOH, i-PrOH and t-BuOH. The unfolded state undergoes a three-step structural transition. Firstly TFE and other alcohols stabilize the secondary and tertiary structures by increasing the α-helicity. This non native structure further unfolds at moderate concentration and undergoes aggregation at higher concentrations forming either non-native α-helix or β-sheet structure. The novel approach of our work shows how a protein like β-lg undergoes aggregation under alkaline condition at room temperature without the application of thermal energy. Our findings might also open up the insight of the mechanism of alcohol induced protein folding and aggregation which finds application in dairy industry to prevent the drainage of nutritionally important whey proteins during cheese manufacturing technology.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

Financial support of University Grant Commission (UGC)-CAS-II and DST-PURSE-II Program of Jadavpur University, Kolkata are greatly acknowledged. Sanhita Maity, SRF, is a recipient of UGC-NET Research Fellowship in Chemistry. The authors wish to acknowledge Prof. Dipankar Chakraborty, Department of Applied Chemistry, Kolkata University for providing the TEM instrumental facility.

Notes and references

1. M. Buck, Q. Rev. Biophys., 1998, 31, 297.
2. N. Kumar and N. Kishore, Biophys. Chem., 2013, 171, 9.
3. P. Venkatesu, M. J. Lee and H. Lin, J. Phys. Chem. B, 2009, 113, 5327.
4. M. W. Pantoliano, M. Whitlow, J. F. Wood, S. W. Dodd, K. D. Hardman and M. L. Rollence, Biochemistry, 1989, 28, 7205.
5. R. H. Khan and M. S. Shabnum, Biochemistry, 2001, 66, 1042.
6. P. Venkatesu, M. J. Lee and H. M. Lin, J. Phys. Chem. B, 2009, 113, 5327.
7. N. Kumar and N. Kishore, J. Chem. Phys., 2013, 139, 115104.
8. A. Kumar and P. Venkatesu, Chem. Rev., 2012, 112, 4283.
9. L. Wei, P. Jiang, W. X. Xu, H. Li, H. Zhang, L. Y. Yan, M. B. Chan-Park, X. W. Liu, K. Tang, Y. G. Mu and K. Pervushin, J. Biol. Chem., 2011, 286, 6291.
10. S. B. Padrick and A. D. Miranker, Biochemistry, 2002, 41, 4694.
11. (a) G. W. Jameson and B. J. Sutherland, US Pat., US5356639, 1994; (b) R. C. Lawrence and J. Lelievre, Proceedings of the XXXIIIth International Dairy Congress, 1990, vol. 3, p. 1880.
12. P. Luo and R. L. Baldwin, Biochemistry, 1997, 36, 8413.
13. B. Ahmad, S. K. Haq, A. Varshney, A. A. Moosavi-Movahedi and R. H. Khan, Biochemistry, 2010, 75, 486.
14. A. Miranker, S. E. Radford, M. Karplus and C. M. Dobson, Nature, 1991, 349, 633.
15. M. Hameed, B. Ahmad, K. M. Fazili, K. Andrabi and R. H. Khan, J. Biochem., 2007, 141, 573.
16. Y. Liu and D. W. Bolen, Biochemistry, 1995, 34, 12884.
17. K. Shiraki, K. Nishikawa and Y. Goto, J. Mol. Biol., 1995, 245, 180.
18. P. D. Thomas and K. A. Dill, Protein Sci., 1993, 2, 2050.
19. M. Buck, H. Schwalbe and C. M. Dobson, Biochemistry, 1995, 34, 13219.
20. D. Hamada, Y. Kuroda, T. Tanaka and Y. Goto, J. Mol. Biol., 1995, 254, 737.
21. N. Hirota, K. Mizuno and Y. Goto, Protein Sci., 1997, 6, 416.
22. K. Shiraki, K. Nishikawa and Y. Goto, J. Mol. Biol., 1995, 245, 180.
23. S. Samaddar, A. K. Mandal, S. K. Mondal, K. Sahu, K. Bhattacharyya and S. Roy, J. Phys. Chem. B, 2006, 110, 21210.
24. K. Kuwajima, Proteins, 1989, 6, 87.
25. P. Gupta, R. H. Khan and M. Saleemuddin, Arch. Biochem. Biophys., 2003, 413, 199.
26. M. S. Kay and R. L. Baldwin, Nat. Struct. Biol., 1996, 3, 439.
27. Y. Goto, L. J. Calciano and A. L. Fink, Proc. Natl. Acad. Sci. U. S. A., 1990, 87, 573.
28. (a) H. Singh and A. Waungana, Int. Dairy J., 2001, 11, 543; (b) P. Zoon, Int. Dairy Fed., 1994, 9402, 402.
29. O. E. Mills, Biochim. Biophys. Acta, 1976, 434, 324.
30. U. C. Halder, J. Chakraborty, N. Das and S. Bose, J. Photochem. Photobiol., B, 2012, 109, 50.
31. M. R. H. Kerbs, G. L. Derlin and A. M. Donald, Biophys. J., 2009, 96, 5013.
32. S. Pal, S. Maity, S. Sardar, J. Chakraborty and U. C. Halder, Int. J. Biol. Macromol., 2016, 84, 121.
33. J. Chakraborty, N. Das, K. P. Das and U. C. Halder, Int. Dairy J., 2009, 19, 43.
34. G. V. Semisotnov, N. A. Radionava, O. I. Razgulyaev, V. N. Uversky, A. F. Gripas and R. I. Gilmanshin, Biopolymers, 1991, 31, 119.
35. J. R. Lakowicz, Principles of fluorescence spectroscopy, Springer, Singapore, 3rd edn, 2006, vol. I, p. 230.
36. J. M. Khan, A. Qadeer, E. Ahmad, R. Ashraf, B. Bhushan, S. K. Chaturvedi, G. Rabbani and R. H. Khan, PLoS One, 2013, 8, 62428.
37. M. R. Nilsson, Methods, 2004, 34, 151.
38. J. M. Khan, S. K. Chaturvedi, S. K. Rahman, M. Ishtikhar, A. Qadeer, E. Ahmad and R. H. Khan, Soft Matter, 2014, 10, 2591.
39. V. Bannerjee and K. P. Das, Colloids Surf., B, 2012, 92, 142.
40. S. A. Hudson, H. Ecroyd, T. W. Kee and J. A. Carver, FEBS J., 2009, 276, 5960.
41. W. W. Yu, E. Chang, J. C. Falkner, J. Zhang, A. M. Sl-Somali, C. M. Sayes, J. Johns, R. Drezek and V. L. Colvin, J. Am. Chem. Soc., 2007, 129, 2871.
42. G. D. Fasman, Circular Dichroism and the Conformational Analysis of Biomolecules, Plenum Press, New York, 1996.
43. V. Bhakuni, Arch. Biochem. Biophys., 1998, 357, 274.
44. N. Taulier and T. V. Chalikian, J. Mol. Biol., 2001, 314, 873.
45. N. Sreerama and R. W. Woody, Protein Sci., 2003, 12, 384.
46. S. Ghisla, V. Massey, J. M. Lhoste and S. G. Mayhew, Biochemistry, 1974, 13, 589.
47. E. Di Stasio, P. Bizzarri, F. Misiti, E. Pavoni and A. Brancaccio, Biophys. Chem., 2004, 107, 197.
48. T. Kijima and T. Izumi, Biotechnol. Lett., 2004, 26, 655.
49. O. E. Mills, Biochim. Biophys. Acta, 1976, 434, 324.
50. F. Khan, R. H. Khan and S. Muzammil, Biochim. Biophys. Acta, 2000, 1481, 229.
51. M. Collini, L. D'Alfanso, H. Molinari, L. Ragona, M. Catalano and G. Baldini, Protein Sci., 2003, 12, 1596.
52. M. Hameed, B. Ahmad, K. M. Fazili, K. Andrabi and R. H. Khan, J. Biochem., 2007, 141, 573.
53. A. Naeem, K. A. Khan and R. H. Khan, Arch. Biochem. Biophys., 2004, 432, 79.
54. S. Dave, S. Mahajan, V. Chandra and P. Gupta, Int. J. Biol. Macromol., 2011, 49, 536.
55. S. A. Hudson, H. Ecroyd, T. W. Kee and J. A. Carver, FEBS J., 2009, 276, 5960.
56. S. Sardar, S. Pal, S. Maity, J. Chakraborty and U. C. Halder, Int. J. Biol. Macromol., 2014, 69, 137.
57. A. Dong, J. Matsuura, S. D. Allison, E. Chrisman, M. C. Manning and J. F. Carpenter, Biochemistry, 1996, 35, 1450.
58. C. Bahttacharjee, S. Saha, A. Biswas, M. Kundu, L. Ghosh and K. P. Das, Protein J., 2005, 24, 27.
59. P. Sen, B. Ahmad, G. Rabbani and R. H. Khan, Int. J. Biol. Macromol., 2010, 46, 250.
60. B. Ahmad, Z. Islam, A. Varshney and R. H. Khan, Protein Pept. Lett., 2010, 17, 660.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra12057a

---