Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Evidence of conformational changes in oil molecules with protein aggregation and conformational changes at oil–‘protein solution’ interface Partha Patra ∗ , Ponisseril Somasundaran Langmuir Center of Colloids and Interfaces, Columbia University, New York, NY 10027, USA

a r t i c l e

i n f o

Article history: Received 4 November 2013 Received in revised form 4 March 2014 Accepted 25 March 2014 Available online 23 May 2014 Keywords: Interface Protein film Conformational change Adsorption Red/blue shift

a b s t r a c t Time-dependent conformational changes of proteins and oil molecules at oil–‘protein solution’ interface were studied using ATR (Attenuated Total Reflection)-FTIR spectroscopic technique for the case of Bacillus subtilis extracellular proteins (BSEPs) and bovine serum albumin (BSA) in hexane–‘protein solution’ system. The IR spectra collected on the protein aggregate – film – formed at the hexane–‘protein solution’ interface demonstrated time-dependent conformational changes of the proteins through changes in the shapes and positions of the H2 O–‘amide I’ cross peaks and the amide II peaks as a function of time (0–90th minute). Hexane–protein intermolecular association in the film was evident as the CH stretching vibration peaks of hexane were present along with the amide peaks in all the spectra collected over a period of 90 min. Conformational changes of the hexane molecules, along with that of the proteins, were observed via variations (broadening and red/blue shifts) in the CH stretching vibration peaks of the CH3 and the CH2 groups of hexane. The red/blue shifts of the CH stretching vibration peaks of hexane were different with BSEPs and BSA, further indicating that the conformational changes of hexane molecules being protein specific. As similar to the protein types considered here, at oil–‘protein solution’ interfaces, conformational changes of the oil molecules appear to be a regular phenomenon. Published by Elsevier B.V.

1. Introduction Protein conformational dynamics at oil/water interfaces are of fundamental importance to many processes in biology (e.g., protein–lipid interactions) [1–4], food (e.g., role of milk proteins as emulsifiers) [5,6], and pharmaceutical industries [7,8]. Proteins undergo conformational changes at the oil/water interfaces and such changes often are associated with protein aggregation; within certain cases (e.g., milk proteins such as casein), skin-like patterns form with a dense inner layer immediately at the interface and a diffuse layer away from the interface [9,10]. Numerous studies probing the oil–‘protein solution’ interfaces have elucidated protein conformational changes by using different techniques such as ellipsometry [11], fluorescence [12] and IR/Raman spectroscopy [13,14]. A number of proteins that exhibit conformational changes and as a consequence aggregate to form a skin-like coat at oil–‘protein solution’ interfaces are such as apomyoglobin, ␤-casein, ␣-casein, lysozyme, bovine serum albumin, ␬-casein, ␤lactoglobulin, and, many bio-peptides – especially with relevance

∗ Corresponding author. Tel.: +1 212 854 2926; fax: +1 212 854 8362. E-mail addresses: [email protected], [email protected] (P. Patra). http://dx.doi.org/10.1016/j.colsurfb.2014.03.045 0927-7765/Published by Elsevier B.V.

to their interactions with cell membranes [15,16]. Protein conformational dynamics and aggregation at oil/water interface [17], diffusivity of proteins across an interfacial protein aggregated layer [18], and rheological/elastic behavior of the skin-like aggregates [19] have been studied extensively. Protein aggregation at oil/water interfaces is a complex process and involves van der Waals, hydrophobic and electrostatic interactions, and hydrogen bonding [20,21]. While protein conformational changes at oil–water interfaces are often seen, there are a few evidences that indicate conformational changes of the oil molecules at oil–‘water (protein solution)’ interfaces [22–27]. The role of the oil/‘protein solution’ interface on the protein conformational dynamics, specifically accounting for the conformational changes of the oil molecules that are part of the interface has been demonstrated through interfacial properties such as interfacial tension and Laplace pressure [20,21,28]. A number of earlier publications provide evidences of the conformational changes of oil/lipid molecules with emphasis on such changes occurring due to temperature perturbations, adsorption onto different types of solid substrates, and nano-confinement of oil molecules, e.g., the CH stretching vibrations of hydrocarbon molecules such as cyclohexane, ethane and benzene exhibit changes (e.g., red/blue shift) when subjected to temperature variations (reasonably higher than the room temperature) [29–31]. Such

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

conformational changes are attributed to hyperconjugative type interactions that involve ␴* antibonding and ␴ bonding orbitals of the CH of alkanes and the substrate moiety. Among bio-peptides, some strains of Bacillus subtilis have been reported to produce mechanically flexible protein coat, which protects the spores from the toxic molecules [32–34]. Earlier studies suggest that BSA readily adsorbs at oil–water interface and undergoes conformational changes [35]. Here, the stability of the B. subtilis extracellular protein (BSEPs) and Bovine Serum Albumin (BSA) aggregates – film – at the hexane–‘protein solution’ interface was studied along with the time-dependent conformational changes of the proteins in the film. Time-dependent conformational changes of the BSEPs/BSA were followed using ATR-IR (Attenuated Total Reflection-Infrared) spectroscopic technique. We have adopted an in situ ATR-IR spectroscopic approach (explained in Section 2) which enabled the concurrent estimation of the conformational changes of proteins and hexane molecules in the film. The H2 O–‘amide I’ cross peaks and the amide II peaks in the spectra collected on the film over a period of time (0–90 min) were analyzed to follow the time-dependent conformational changes of the proteins. From the same spectra, information on the time-dependent conformational changes of the hexane molecules were also obtained by analyzing the changes in the CH stretching (symmetric and asymmetric) vibration peaks of the CH3 and the CH2 groups, and thus, led to the concurrent estimation of the conformational changes of hexane along with that of the proteins molecules in the film. 2. Experimental 2.1. Materials The extracellular proteins were isolated from 72 h grown bacterial culture of B. subtilis 168 strain acquired from American type culture collections (ATCC 6633). The bacteria were grown in Bromfield [33] medium, and isolation of the extracellular proteins from the culture broth was carried out according to the methods described elsewhere [32,34,36]. In brief, the cells were separated from the broth by centrifugation of the broth at 6000 rpm and the BSEPs (B. subtilis extracellular proteins) were precipitated from the cells-free supernatant using (NH4 )2 SO4 (at 80 wt% saturation). Purification of the precipitated BSEPs was carried out by dialyzing (2000 MWCO dialysis membrane) the precipitate against 50 mM Tris–HCl Buffer (10% glycerol, 2 mM EDTA, 50 mM KCl). Lyophilized BSA protein [assay: ≥98% (agarose gel electrophoresis)] powder, HPLC high purity grade n-hexane and ACS reagent grade (≥99.0%) (NH4 )2 SO4 were obtained from Sigma–Aldrich. Reagent used for the estimation of proteins concentrations included Bradford reagents which were obtained from Sigma–Aldrich. 2.2. Droplet size distribution Hexane-in-water (oil-in-water emulsion – O/W) and hexane-in-hexane (oil-in-oil emulsion – O/O) emulsions were prepared using BSEPs and BSA protein solutions (proteins in water, pH 7–7.5) having proteins concentrations in the range of 100–2000 ppm. Hexane was gently added to the protein solution at 1:1 (v/v) ratio in 1.5 mL Eppendorf tubes and at room temperature (25 ◦ C). The tubes were capped, and the two liquids (oil and ‘protein solution’) were allowed to mix rigorously for 1 min in a vortex shaker (rotation speed of 2000 rpm), which resulted in two types of emulsions: (a) ‘protein coated hexane droplets’ in water: transparent solution at the bottom part of the tube, designated as sample-1, and (b) ‘protein coated hexane droplets’ in hexane – a phase with foamy appearance on the top part of the tube, designated as sample-2. Vortexing led to the

133

mechanical dispersion of hexane in the protein solution, resulting in the formation of the hexane–water interfaces, and consequently led to adsorption of the proteins at the interfaces. Extended length gel pipette tips were used to acquire samples (sample-1) from the bottom part of the Eppendorf tubes, whereas, to acquire samples from the top part (sample-2) pipettes with wider opening (tips of the pipettes were sawed-off to ∼5 mm with a blade) were used. Pipettes with wider opening were used to maintain the integrity (foamy appearing phase) of the samples. Droplet size distributions (DSD) of all the sample-1 types were determined using Malvern Zetasizer Nano-ZS, and that of sample-2 by estimating the droplet sizes from the images of the droplets captured through a phase contrast microscope. 2.3. Topology of protein film at hexane–‘protein solution’ interface Topological features of the proteins films which encapsulated the emulsified hexane droplets in sample-2 types were studied by observing the sample-2 droplets under a phase contrast microscope. Approximately 500 ␮L of the sample carefully acquired from an Eppendorf tube was gently placed on a glass slide. The glass slide placed under the phase contrast microscope was encased within a thermostat which enabled the sample temperature being maintained at ∼25◦ C. A transparent glass window on the top part of the thermostat allowed observation of the droplets under the microscope. The area on the sample that was considered for observation was around the center part (after gaining focus as desired) of the sample, spanning an area of approximately 5 mm × 5 mm, where different sections in the 5 mm × 5 mm window were scanned to acquire the droplet images. The droplets in different sections were continuously monitored for a longer (up to 120 min) duration, during which the droplet images were captured using a Hitachi CCD camera attached to the microscope. Over time, the protein film exhibited its characteristic topological features before rupturing and eventually leaving a residue on the glass slide. By choosing the center part of the sample and scanning within a 5 mm × 5 mm window, the droplets at the periphery of the samples were avoided, where the droplets ruptured irregularly and at a faster rate compared to the droplets that were at the center part of the sample. Complete rupture time (CRT) of a film was recorded as the time from the moment the sample-2 (droplets in oil) types were placed on a glass slide to the time the droplet collapsed due to the rupturing of the film. CRT values were determined for droplets prepared with protein (BSEPs and BSA) solutions having a range (100–2000 ppm) of protein concentrations. For the samples prepared with lower protein concentrations (10 days). Until the time the DSD differed from that determined at the beginning (when the droplets were devised) was considered as the CRTs of the droplets in the samples, a measure of emulsion stability. 2.4. Protein adsorption behavior at hexane–‘protein solution’ interface Protein adsorption densities at the hexane–‘protein solution’ interfaces in the sample-2 types were determined as a function of

134

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

initial protein concentrations. Hexane was gently added to the protein solutions at 1:1 (v/v) ratio in 1.5 mL Eppendorf tubes, resulting in two separate phases, hexane on the top of the protein solutions. Subsequent to addition, the Eppendorf tubes were vortexed for 1 min and then allowed to equilibrate for the next 1 min. Vortexing led to mechanical dispersion of hexane in the protein solution, and in form of numerous droplets. The total hexane–‘protein solution’ interfacial area generated during vortexing – mechanical dispersion – depends on the vortexing speed, the physicochemical properties of the two liquids, and the holdup fraction of the dispersed phase. Here, the vortexing speed was significantly higher, 2000 rpm, to overcome the contributions of the physicochemical properties of protein solutions (prepared with varying protein concentrations) and the holdup factors to the variations in the interfacial areas. Thus, irrespective of protein concentrations, the total interfacial area for all the samples was expected to be approximately the same. In addition, the samples prepared with different protein concentrations exhibited similar DSD, suggesting that the variations in the interfacial areas with different protein concentrations were negligible. Similarities in DSD values, even with variation in protein concentrations, indicated that the differences in the interfacial areas were negligible, therefore, the adsorption densities are reported as amounts of proteins adsorbed per an arbitrarily defined constant interfacial area, i.e., protein concentrations in sample-2. Approximately 10 mL of a sample-2 type acquired from a set of 10–15 Eppendorf tubes (having sample-2 types prepared with same initial protein concentrations) was poured into a 20 mL glass vial. The samples were subjected to heating at 60–70 ◦ C in a temperature-controlled furnace for 15–20 min, which enabled denaturation of the proteins in the films followed by destabilization of the emulsion – sample-2; hexane on top of the protein solution. The protein concentration in the solution was determined by Lowry Protein assay [37] using Bradford reagent, and BSA was used as protein standard. 2.5. Use of ATR-FTIR spectroscopy to follow time-dependent conformational changes of protein/s and hexane at hexane–‘protein solution’ interface ATR-FTIR spectroscopic measurements were performed on a ZnSe trough via internal reflection element, with 10 internal reflections and at an angle of incidence 45◦ , and using a Perkin Elmer. Samples were placed on the ZnSe trough that was encased in a thermostat hood, which enabled the sample temperature being maintained at ∼25◦ C. All samples were studied between frequency ranges of 4000–650 cm−1 , and spectra were collected at 4 cm−1 resolution. A set of spectra collected on the native protein solution (1000 ppm) was used as a reference for the estimation of time-dependent conformational changes of the proteins at the hexane–‘protein solution’ interfaces. 2 mL of the native protein solution was placed on the ZnSe trough and the spectra were collected at a regular interval of 3–5 min during dehydration. It was difficult to identify the characteristic amide peaks from the spectra that were collected at the beginning, however, due to dehydration induced increase in the solution protein concentrations the amide peaks could be detected later. At a much later stage, the proteins concentrations were significantly higher, leading to either protein aggregation or crystallization, which was evident through the slight shifts and/or broadening of the amide peaks. These spectra were not considered for reference purpose. A set of spectra which exhibited similarities in their line shapes including that of the amide peaks was used for reference. IR spectra were collected on sample2 types prepared with 1000, 1500 and 2000 ppm of initial protein concentrations. Approximately 2 mL of sample was placed on the ZnSe trough and the spectra were collected at regular intervals of

3–5 min, and for longer durations (0–120 min). As the droplets in the sample-2 types were hexane droplets coated with protein films (water and protein) the droplet density was relatively higher than that of hexane (solvent phase in sample-2), and therefore, located at the bottom part of the ZnSe trough, which enabled the IR beam to probe the protein films and hexane in the sample, thus, providing an option to follow the conformational changes of the proteins and hexane molecules simultaneously. Since the aim was to obtain semi-quantitative information on the time-dependent protein conformational changes, the changes in the shapes and positions of the amide (I, II and III) peaks were analyzed for in the spectra collected over time. Therefore, the spectra were not normalized according to H2 O bending vibration peak intensities, and the changes in the shapes and positions of the H2 O–‘amide I’ cross peaks were analyzed instead. However, in order to obtain an approximate estimate of the alpha helix and the beta sheet content in the protein film, the de-convolution of the amide I peaks was carried out using Gaussian distribution function and by using Origin Pro 8. Three peaks representing alpha helix (1655 cm−1 ), beta sheets (1620 cm−1 ) and random coil-beta (1680 cm−1 ) was considered for multiple peak fitting. In order to follow the time-dependent conformational changes in hexane, all the spectra collected on the sample-2 types were also analyzed for changes in the CH stretching vibrations. 3. Results and discussions 3.1. BSEPs/BSA film formation at the hexane–‘protein solution’ interface Addition of BSEP solutions to hexane followed by vortexing resulted in two types of emulsions, ‘protein-film coated hexane droplets (PCHD)’ dispersed in water (O/W emulsion) and hexane (O/O emulsion). For all the emulsion samples prepared with varying (100–2000 ppm) protein concentrations, the droplets in PCHD-inwater (O/W) emulsion spanned from 0.1 to 5 ␮m, whereas, the droplets in PCHD-in-hexane (sample-2 – O/O) emulsion spanned a larger-size range of 10–2000 ␮m (Fig. 1). The DSD in the emulsions prepared with BSA were similar to that of the emulsions prepared with BSEPs. The IR spectra collected on sample-2 types O/O emulsion showed the characteristic amide peaks of BSEP/BSA and the CH stretching vibration peaks of hexane, evidencing the existence of a protein film on a PCHD. 3.2. Topological features of the protein film at hexane–‘protein solution’ interface Changes in the topological features of the protein (BSEPs) films, upon exposure of the PCHD droplets in sample-2 (O/O emulsion) to the thermostat ambient, were studied as a function of time (0–120th minute). An image taken at around 30th minute showed the existence of two different layers in the protein film, the top layer appeared as a torn layer (the torn edge of the layer is indicated by arrows in Fig. 2) spread across the periphery of the droplet and an undisturbed shiny (designated as a primary layer here) layer beneath. The torn layer of the film (designated as a secondary layer) progressively contracted (demonstrated through a series of snapshots images taken over time, shown in Fig. 2) over the primary layer. Tearing and contraction of the secondary layer can be attributed to evaporation of water from the film and that of hexane from the droplet. After approximately 60 min the droplet collapsed due to the rupturing of the primary layer, where the droplet collapsed slowly spanning a few seconds. As the PCHDs (sample-2) were exposed to the thermostat ambient, two interfaces exist: the hexane–‘protein film’ interface and the ‘protein film’–air interface. For the case of the ‘protein film’–air interface, dehydration of water from the film at the interface is relatively faster compared to that

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

135

Fig. 1. Addition of BSEPs solution (1000 ppm) to hexane followed by vortexing for 1 min led to formation of two types of emulsions. In the Eppendorf tube, emulsion of one type (PCHD-in-hexane) with foamy appearance is seen on the top of the protein solution, whereas, the protein solution itself represented an emulsion (PCHD-in-‘protein solution’) of another type (water soluble red dye was used to distinguish water from transparent hexane phase). A confocal image taken on the PCHD-in-‘protein solution’ clearly demonstrates the formation of a protein film around the droplet. Phase contrast microscopy images of the two emulsions types and the droplet size distribution in the emulsions are shown. Bar indicates 10 ␮m.

away from the film–air interface. Thus, faster dehydration from the secondary film is likely and is associated with increase in the protein concentration therein, resulting in higher protein concentration at the immediate vicinity of the ‘protein film’–air interface compared to that away from the interface. Such differences in the protein concentrations across the film cause significant differences in the film pressure can be attributed to the tearing of the secondary layer. Overall, the topological evolution of the BSEP film (formed with initial protein concentrations ≥ 1000 ppm), upon exposure of the PCHD-in-hexane emulsion to the ambient (thermostat ambient), included three sequential phenomena: (1) within the first 5–10 min, tearing of a secondary film existing at the periphery of the PCHD, (2) spanning about 30 min duration, contraction of the torn secondary layer over a stable primary layer, and (3) tearing of the primary layer leading to the collapsing of the droplet. In case of BSA, a shiny primary layer was evident on a droplet, whereas, a secondary film was not observed.

3.3. Stability of ‘protein coated hexane droplets’-in-hexane O/O emulsion Stability of the droplets (in O/O emulsion: PCHD-in-hexane) devised with varying (100–2000 ppm) initial protein concentrations were determined by estimating the complete rupture times (CRT) of the protein films. For the droplets prepared with protein concentrations greater than 500 ppm of either BSEPs or BSA, upon exposure of the droplets to the ambient, the droplets prepared with BSEPs were stable for longer time (CRT > 30 min) compared to that with BSA (CRT < 5 min). For the droplets in the emulsion phase, the CRT values of the protein films were higher, especially in comparison to that when the droplets were exposed to the ambient (Fig. 3). For the case of sample 2 O/O emulsions that were prepared with BSEP solution having concentrations above 1000 ppm, the DSD in the emulsions were almost same for more than 10 days as compared to 600 ppm. As the protein densities in the two layers were different it is indicative of two different types of adsorption mechanisms. Several factors are responsible for such differences are such as: type of proteins (flexible or globular), conformation of protein molecules, and time dependent conformational entropy. Various theoretical models have been proposed to explain adsorption of macromolecules, e.g., a flexible polymer chain – using a phenomenological model on surface thermodynamics. Springer calculated the total number of various conformations for a flexiblechain polymer located within a surface that is modeled as a two-dimensional quasi-crystal [38,39]. In his model, any intermolecular interaction of polymer–surface and polymer–solvent were neglected, and the contributions of the configurational entropy to the free energy were correlated as:

2

(x − x¯ ) /(n − 1) and n = 3.









RT z ω0 2 − ln(1 − 1 ) + 1− ln 1 − 1 ω0 2 ω1 z

 (1)

where ˘ is the surface pressure, R is the gas law constant, T is the temperature,  1 = ω1  1 is the surface coverage,  1 is the adsorption density, ω1 is the molar area of the macromolecule unfolded at the surface, ω0 is the molar area of a quasi-crystal cell, and z is the co-ordination number of the lattice. A common understanding has been that conformational changes in proteins contribute to

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

(a)

(b)

Native BSEPs 0th min,

Droplets th

1700

1650

th

0 min,

th

5 min

60/90th min

30 min,

1750

Native BSA Droplets

10th min

137

1600

-1

1750

1550

1700

1650

1600

1550

-1

cm

cm

Fig. 4. ATR-FTIR spectra collected over time (0–90 min) on sample-2 (PCHD-in-hexane droplets) formed by BSEPs and BSA. H2 O–‘amide I’ cross peaks in the spectra at 0th, 10th, and 30th minutes shifted to lower wave numbers compared to those in the spectrum of native BSEPs (a), and the cross peaks were observed to shift to the higher wave numbers in the spectra at 60th and 90th minutes compared to those at 0th, 15th, and 30th minutes; the amide II peaks were broadened in the spectra at 30th minute compared to those at 60th minute. H2 O–‘amide I’ cross peaks of BSA (b) proteins in the spectra at 0th and 5th minutes shifted to higher wave numbers compared to those in the spectra of native BSA protein. The amide II peaks in the spectra at 0th and 5th minutes shifted to lower wave numbers compared to those of native BSA protein.

their interfacial packing density (represented as molecular interfacial area it occupies). In a two-dimensional solution model, if partial molar areas, ωi , is defined in analogy with the partial molar volumes in a solution, then for proteins of n number of states, the components of the surface mixture that cover the interfacial area, A, under saturation conditions can be equated as: ω0 N0S + ω1 N1S + ω2 N2S + · · · + ωn NnS = A

(2)

where the adsorption density of the ith state is i =

NSi A

(3)

Saturation in adsorption density is attained when ω0 0 + ω1 1 + ω2 2 + · · · + ωn n = 1

(4)

Fig. 3 shows that the adsorption density increases with increase in initial protein concentrations and up to a value of 1000 ppm. With increase in the protein concentration, increase in adsorption density follows a relationship described by Ward and Thordai [40]:

 

t = 2

D 

 C0√t −

√ t



√ C(0, t − t  )d( t)

(5)

0

where  indicates adsorption density, D as the diffusion co-efficient and C0 as the initial protein concentration. It can be observed from Fig. 3 that beyond initial protein concentration of ∼1000 ppm the protein adsorption density decreases significantly as compared that below, indicating saturation of the interfacial layer with proteins – the primary layer, and formation of a secondary layer at the ‘primary layer’–‘protein solution’ interface. Proteins participating in the formation of primary layer exhibit irreversible adsorption behavior. The secondary layer is formed at the ‘primary layer’/‘protein solution’ interface, and protein adsorption in the secondary layer is postulated to be reversible [41]. As the protein adsorption density relates to the protein conformational changes at interfaces, understanding of the protein conformational changes at the hexane–‘protein solution’ interface provide clues to such differences in the adsorption densities observed with BSEPs and BSA. Conformation changes of the proteins in the sample-2 protein films were studied using ATR-FTIR spectroscopic technique, in which sample-2 O/O emulsion samples were prepared with different protein concentrations: 1000, 1500, and 2000 ppm.

3.5. ATR-FTIR spectroscopic evaluation of time-dependent conformational changes in protein/s at hexane–‘protein solution’ interface For PCHD-in-hexane O/O emulsion prepared with BSEPs (initial protein concentration as 1000 ppm), the H2 O–‘amide I’ cross peaks and the strong peaks pertaining to the CH stretching (3000–2850 cm−1 ) and bending vibrations (∼1400 cm−1 ) of the CH3 /CH2 groups of hexane were seen in the spectrum collected on the emulsion sample at 0th minute – as soon as the sample was placed on the ZnSe trough and a spectrum was acquired. A spectrum collected on the same sample at 15th minute showed the amide II/III peaks due to sufficient increase in the protein concentration upon evaporation of water from the protein film. The amide peaks in the spectra were representative of the proteins in the film formed at the hexane–‘protein solution’ interface in the emulsion. The amide III peaks in the spectra collected in the duration of 15–90th minutes were similar (in terms of peak positions) to that of the spectrum collected on native BSEPs. Thus, the possibility of protein denaturation upon their adsorption at the interface was eliminated. As a function of time, in the 0–90th minute period, the shifts in both the cross peaks and the amide II peaks in the spectra were compared with reference to the corresponding amide peaks in the spectrum collected on native protein. The spectrum collected on native protein (BSEPs) exhibited H2 O–‘amide I’ cross peak (Fig. 4a) prominent at 1652 cm−1 , suggesting that ␣-helix structures are the predominant motifs in their native form. In comparison to the cross peak in the spectrum of native protein the cross peaks in the spectra collected on sample-2 at 0th, 15th, and 30th minutes exhibited shift toward lower wave numbers. Such shifts in the cross peaks from 1652 cm−1 to lower wave number 1630 cm−1 indicated transformation of ␣-helix to ␤-sheets and random coils. The spectra collected at 60th and 90th minutes (Fig. 4a) showed that the cross peaks shifted further to higher wave numbers from that seen at 30th minute (1630 cm−1 to 1665 cm−1 ), indicating time-dependent conformational reconstruction of the secondary structure (␤-sheets) of the BSEPs in the film. As in the 30th, 60th, and 90th minute spectra there was no evidence of ␣-helix peaks; it is obvious that during structural reconstruction ␤-sheet transformed to their different forms and turn structures. Such time-dependent reconstruction at the interface also included the broadening of the amide II peak (1530 cm−1 ) at 30th minute, especially in comparison to the relatively sharp peaks at 60th and

138

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

90th minutes (Fig. 4a). Similar changes in the peak shifts and broadening were also observed in the spectra collected on sample-2 prepared with 1500 and 2000 ppm of BSEP solutions. For initial protein concentrations of values 1000 ppm, where the formation of the secondary layer is likely, as ‘beta sheet’/‘alpha helix’ ratio decreased with increase in initial protein concentrations, it suggested that with increase in the initial protein concentrations the proteins undergo minimal conformational changes in the secondary layer. For all the spectra collected on the PCHD-in-hexane O/O emulsion prepared with BSA (initial protein concentration as 1000 ppm), strong peaks pertaining to the CH stretching, rocking, and scissoring vibrations of the CH3 and CH2 groups of hexane were noticed in the spectra along with the cross peaks and the amide II/III peaks (Fig. 4b). The positions of the cross peaks in the spectra collected at 0th and 5th minutes shifted to higher wave numbers (1660 cm−1 ) with a shoulder at around 1630 cm−1 , especially when compared to that (cross peak centered at ∼1645 cm−1 , typically seen for BSA) in the spectrum collected on the native BSA (Fig. 4b), further indicating conformational changes of BSA at the oil–‘protein solution’ interface. Native form of BSA predominantly consists of ␣-helical conformation, and since a significant shift in the cross peak position toward higher wave numbers was observed, it suggested transformation of ␣-helical structures to ␤-sheets (1630 cm−1 ) and ␤-turns. Conformational changes were also noticed via the shift in amide II peaks collected at 0th and 5th minutes to lower wave numbers. Time-dependent conformational reconstruction was difficult to infer from the cross peaks positions in the spectra (at 0th and 5th minutes) as the peak positions (1652 cm−1 ) were close to each other. The amide II peaks at 0th and 5th minute were observed to shift to a higher wave numbers (from 1540 cm−1 to 1550 cm−1 ) in comparison to that of the corresponding peak collected on the native protein. 3.6. Conformational changes of hexane at hexane–‘protein solution’ interface in O/O emulsion A spectrum collected on the sample-2 O/O emulsion devised with BSEPs showed the CH stretching vibration peaks of hexane along with the amide peaks at 90th minute (the sample-2 exhibited a wrinkled skin-like appearance on the ZnSe trough). Notably, at 90th minute, it was expected that complete evaporation of hexane from the sample would lead to the disappearance of the CH stretching vibration peaks of hexane from the spectrum, instead, the presence of the CH stretching vibration peaks of hexane in the spectrum suggested hexane–protein intermolecular association. In addition, there were notable changes in the shapes and positions of both the CH symmetric and the CH asymmetric stretching peaks as compared to that of pure hexane, which suggested conformational changes of hexane upon their molecular-level association with the proteins in the film. While it is obvious that proteins would undergo conformational changes at oil–water interfaces, conformational changes of the hexane molecules along with that of the conformational changes of the proteins were unexpected. Conformational changes of the hexane molecules were studied with importance of such changes occurring at a later stage, as indications of the conformational changes of the hexane molecules were expected to be at a molecular level proximity of the proteins in the film. The CH stretching vibrations peaks of hexane are typically seen as two doublets in the range of 3000–2850 cm−1 [CH symmetric

Fig. 5. The CH symmetric and asymmetric stretching vibrations peaks of hexane collected on sample-2 devised with BSEPs and BSA. Red and blue shifts of the peaks are shown as red and blue pillars, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

(2872 cm−1 for CH3 , and, 2853 cm−1 for CH2 ), and asymmetric (2962 cm−1 for CH3 , and, 2926 cm−1 for CH2 )] [34]. The first three spectra (0th, 5th and 10th minute) collected on the sample-2 (BSEPs/BSA) exhibited strong peaks pertaining to the CH symmetric and asymmetric stretching vibrations. For BSEPs, a spectrum (Fig. 5) collected on sample-2 at 30th minute showed broadening of the CH symmetric stretching peaks of the CH3 and the CH2 groups of hexane, and slight broadening of the CH asymmetric stretching peaks of the CH3 and the CH2 groups. In the spectra collected at 60th and 90th minutes, there were notable shifts of both the CH symmetric and asymmetric stretching vibration peaks of the CH3 and the CH2 groups. Fig. 5 shows that in comparison to the spectra collected at 0th and 30th minute, the CH asymmetric stretching peaks of the CH3 group of hexane exhibited a red-shift at 60th and 90th minutes, whereas, the CH asymmetric stretching peaks of the CH2 group of hexane exhibited slight blue shift. Thus, at 60th and 90th minutes, CH asymmetric stretching peaks exhibited blue and red shifts, respectively, for CH3 and CH2 groups, which suggested conformational changes of hexane. At 60th and 90th minutes, the CH stretching vibration peaks of hexane are more representative of the hexane molecules having intermolecular association with the proteins, thereof, the conformational changes of hexane are obviously due to molecular level interactions with the proteins. In addition to the variation observed in the CH asymmetric stretching vibrations, Fig. 5 shows that the CH symmetric stretching vibrations of the CH3 and the CH2 groups broadened in the spectra collected at 30th, 60th, and 90th minutes, and, in particular, red shifts of the CH stretching vibrational peaks of the CH2 group were observed at 30th, 60th, and 90th minutes, further indicating conformational change of hexane and protein–hexane intermolecular association. For collection of spectra (Fig. 5) on sample-2 prepared with BSA, the CRT values of the droplets were less than 5 min, however, spectra could be collected on the sample existing on the ZnSe trough for a relatively longer period of time (until 15th minute). In the spectrum collected at 15th minute, a significant shift and/or broadening of the peaks of both the CH symmetric and asymmetric stretching vibrations of the CH3 and CH2 groups of hexane were observed. The broadening of the CH asymmetric stretching vibration peaks of both the CH3 and CH2 groups were noted along with blue shift at 15th minute. In comparison to the peaks at 0th minute, the blue shift of

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

139

Fig. 6. Red and blue shift in CH stretching vibrations of hexane observed in samples (sample-2 type) prepared with initial protein concentrations of 1000, 1500 and 2000 ppm. The red and blue shifts are indicated in red and blue arrows, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

the CH asymmetric stretching vibrational peaks of CH3 and CH2 at 15th minute indicated conformational change of hexane. In addition to the variations in the CH asymmetric vibration peaks, for the CH symmetric stretching vibration of the CH3 group, a significant red shifted peak was observed at 15th minute compared to that at 0th/5th/10th minute, whereas, for CH2 group, slight broadening of the peaks was observed and the red shifts were not prominent. Thus broadening and shifting in both the CH asymmetric and symmetric stretching vibrations indicated conformational changes of the hexane molecules with that of BSA at the oil–‘protein solution’ interface. Red/blue shifts in the CH stretching vibrations were followed with varying initial protein concentrations: 1000, 1500 and 2000 ppm. It can be observed from Fig. 6a that irrespective of the initial protein (BSEPs) concentrations, the CH asymmetric stretching vibrations of CH3 and CH2 groups of hexane exhibited red and blue shifts respectively, further indicating that conformation change of hexane is dependent on specific moieties of proteins and is not concentration dependent. Fig. 6b shows that in case of BSA the CH asymmetric stretching vibrations of both the CH3 and CH2 groups of hexane exhibited blue shifts, as opposed to both blue and red shifts observed with BSEPs. Thus, even though similar concentrations of proteins were used there were notable differences (in terms of red/blue shifts) in the conformational changes of hexane with BSEPs and BSA, further indicating that conformational change of hexane is dependent on the electronic environment particular to a protein type. 3.7. Protein–hexane interactions in protein aggregated film As for BSEPs, ‘beta sheet’/‘alpha helix’ ratio increased upon adsorption, which suggested that beta sheets were transformed from alpha helix upon adsorption. It is likely that the beta sheets are possibly merged in the oil phase, and being hydrophobic in nature, contributed to protein packing through inter-protein hydrophobic interactions. Notably, transformation of alpha helix to beta sheet was also observed in case of the emulsions prepared with BSA. The CH stretching vibration peaks in the IR spectra of the alkanes exhibit broadening and/or red/blue shift upon adsorption of alkanes onto substrates, which are attributed to the environmental effects and interactions with the substrate [42–45]. Such broadening and shift in the CH stretching vibrations are mostly attributed to hyper-conjugative interactions that involve ␴* antibonding and ␴ bonding orbitals of CH of alkanes and the substrate moiety. Broadening and red/blue shifts in the CH stretching

vibration peaks of hexane with BSEPs and BSA also suggested the possibility of a hyperconjugation type interaction. Also, though broadening/shifts in the CH symmetric and asymmetric stretching vibration peaks were equally prominent for the CH3 and the CH2 groups, such variations were different with BSEPs and BSA, further indicating that the nature of hyper-conjugative interactions are primarily influenced by types of moieties specific to a protein type. It is postulated that hexane–protein intermolecular localizations lead to the formation of X [(C H) of CH2 and CH3 ]· · ·Yn (X = hexane, Yn as protein and ‘n’ indicates active moiety) type complex, where hyperconjugation play an important role in influencing the structural and the electronic properties of the complexes. 4. Conclusions and notes Formation of a multilayered film of protein [B. subtilis extracellular proteins (BSEPs)] aggregates was observed at the hexane–‘protein solution’ interface. At higher (>600 ppm) film protein concentrations, a film was found to be stable for a longer duration, >10 days. Such a stable film, which encapsulated the hexane droplets, led to the formation of two types of emulsions, ‘protein-coated hexane droplet’-in-water (O/W emulsion) and ‘protein-coated hexane droplet’-in-oil (O/O emulsion). The droplets smaller (0.05–5 ␮m) in sizes were dispersed in water and the larger (>5 ␮m) in hexane. Adsorption of the proteins at the oil–‘protein solution’ interface led to their conformational change as a function of time, transformation of the ␣-helical structures to ␤-sheets and turns. Time-dependent conformational changes of hexane, observed via variations in the CH stretching vibrations, were significant along with that of the protein conformational changes. The CH symmetric and asymmetric stretching vibrations of both the CH3 and the CH2 groups of hexane broadened in the spectra over time, and in cases, exhibited either blue or red shifts. Conformational changes of hexane with BSEPs were different from that with BSA, indicating that the conformational changes of the hexane molecules are dependent on the electronic environment particular to a protein type, and thus, on their conformational states. Furthermore, broadening and shift in the CH stretching vibration peaks of the end terminal CH3 groups as well as centered CH2 groups of hexane indicated that both the CH3 and CH2 groups of hexane participate in hexane–protein intermolecular interactions. Such interactions are also suggestive of molecular level localization of hexane in the protein aggregates/films. As it is unlikely that the interactions between hexane (CH3 and CH2 ) and protein

140

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141

are of the nature of the formation of chemical bonds or hydrogen bonding, the conformational changes in the hexane and protein molecules thereof led to the consideration that the interactions are of hyperconjugation type; formation of hexane–protein complex [CH of hexane] Yn (‘Yn ’ represents protein motif and ‘n’ indicates motifs in proteins participating in interaction with hexane). Thus, at oil–‘protein solution’ interface, changes in the protein’s conformation are likely to be reliant upon the conformational changes of the oil molecules, where the nature of hexane–protein interaction, in the context of hyperconjugative type interactions need to be investigated further. Conformational changes and intermolecular packing of proteins at oil–water interfaces govern emulsion stability in formulations of personal care, pharmaceutical and food products. Oil-in-water and oil-in-oil emulsions were stable for several days with BSEPs, especially in comparison to that of several minutes for the emulsions prepared with BSA. As the adsorption density of BSEPs at oil–‘protein solution’ interface were significantly higher than that of BSA, the emulsions prepared with BSEPs were stable for longer duration. The protein adsorption density for BSEPs was higher than that of BSA, which indicated that the oil–‘protein solution’ interfacial area occupied by BSEP molecules is significantly lower than that of BSA and led to compact packing. For both the cases of BSEPs and BSA, the proteins exhibited conformational changes and reconstruction as a function of time. Such conformational changes are key drivers toward protein adsorption and packing at the interfaces. As conformational changes in the oil/hexane molecules were observed along with that of proteins, it suggested that protein-hexane interactions possibly play an important role in regulating the protein conformational changes, packing at the oil–water interfaces, and emulsion stability. Acknowledgement The authors are grateful for the grant (CBET-1052697) provided by National Science Foundation (USA) under RAPID program to conduct the research work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb. 2014.03.045. References [1] S. Adams, A.M. Higgins, R.A.L. Jones, Surface-mediated folding and misfolding of proteins at lipid/water interfaces, Langmuir 18 (12) (2002) 4854–4861. [2] K. Nakatsukasa, et al., Dissecting the ER-associated degradation of a misfolded polytopic membrane protein, Cell 132 (1) (2008) 101–112. [3] Y. Sonntag, et al., Mutual adaptation of a membrane protein and its lipid bilayer during conformational changes, Nat. Commun. 2 (2011) 304. [4] M.R. Watry, G.L. Richmond, Orientation and conformation of amino acids in monolayers adsorbed at an oil/water interface as determined by vibrational sum-frequency spectroscopy, J. Phys. Chem. B 106 (48) (2002) 12517–12523. [5] E. Dickinson, L. Denis (Eds.), Recent Trends in Food Colloids Research. Food Macromolecules and Colloids, Royal Society of Chemistry, Cambridge, 1995, p. 380. [6] A.N. Mauri, M.C. Anon, Mechanical and physical properties of soy protein films with pH-modified microstructures, Food Sci. Technol. Int. 14 (2) (2008) 119–125. [7] E.Y. Chi, et al., Physical stability of proteins in aqueous solution: mechanism and driving forces in nonnative protein aggregation, Pharm. Res. 20 (9) (2003) 1325–1336. [8] W.B. Teng, J. Cappello, X.Y. Wu, Physical crosslinking modulates sustained drug release from recombinant silk-elastinlike protein polymer for ophthalmic applications, J. Control. Release 156 (2) (2011) 186–194. [9] E. Dickinson, et al., A neutron reflectivity study of the adsorption of beta-casein at fluid interfaces, Langmuir 9 (1) (1993) 242–248. [10] G. Fragneto, et al., Neutron reflection study of bovine beta-casein adsorbed on OTS self-assembled monolayers, Science 267 (5198) (1995) 657–660.

[11] U.M. Elofsson, M.A. Paulsson, T. Arnebrant, Adsorption of beta-lactoglobulin A and B in relation to self-association: effect of concentration and pH, Langmuir 13 (6) (1997) 1695–1700. [12] C. Castelain, C. Genot, Conformational-changes of bovine serum-albumin upon its adsorption in dodecane-in-water emulsions as revealed by front-face steady-state fluorescence, Biochim. Biophys. Acta: Gen. Subj. 1199 (1) (1994) 59–64. [13] M. Cho, Infrared spectroscopy: mapping protein–protein contacts, Nat. Chem. 4 (5) (2012) 339–341. [14] H.J. Kim, E.A. Decker, D.J. McClements, Impact of protein surface denaturation on droplet flocculation in hexadecane oil-in-water emulsions stabilized by beta-lactoglobulin, J. Agric. Food Chem. 50 (24) (2002) 7131–7137. [15] G. Castillo, et al., Influence of protein source, type, and concentration, and product form on the protein quality of commercial enteral formulas, J. Food Sci. 67 (1) (2002) 328–334. [16] D.M. Small, L. Wang, M.A. Mitsche, The adsorption of biological peptides and proteins at the oil/water interface. A potentially important but largely unexplored field, J. Lipid Res. 50 (Suppl.) (2009) S329–S334. [17] E. Dickinson, Milk protein interfacial layers and the relationship to emulsion stability and rheology, Colloids Surf. B: Biointerfaces 20 (3) (2001) 197–210. [18] T. Cosgrove, J.S. Phipps, R.M. Richardson, Neutron reflection from a liquid/liquid interface, Colloids Surf. 62 (3) (1992) 199–206. [19] D.E. Graham, M.C. Phillips, Proteins at liquid interfaces. 1. Kinetics of adsorption and surface denaturation, J. Colloid Interface Sci. 70 (3) (1979) 403–414. [20] F. Caruso, H. Mohwald, Protein multilayer formation on colloids through a stepwise self-assembly technique, J. Am. Chem. Soc. 121 (25) (1999) 6039–6046. [21] M.R.R. de Planque, J.A. Killian, Protein–lipid interactions studied with designed transmembrane peptides: role of hydrophobic matching and interfacial anchoring (Review), Mol. Membr. Biol. 20 (4) (2003) 271–284. [22] M.O. Jensen, O.G. Mouritsen, Lipids do influence protein function – the hydrophobic matching hypothesis revisited, Biochim. Biophys. Acta: Biomembranes 1666 (1–2) (2004) 205–226. [23] L.V. Schafer, et al., Lipid packing drives the segregation of transmembrane helices into disordered lipid domains in model membranes, Proc. Natl. Acad. Sci. U. S. A. 108 (4) (2011) 1343–1348. [24] B.T. Wong, et al., Conformational changes to deamidated wheat gliadins and beta-casein upon adsorption to oil–water emulsion interfaces, Food Hydrocolloids 27 (1) (2012) 91–101. [25] J. Zhai, et al., Changes in beta-lactoglobulin conformation at the oil/water interface of emulsions studied by synchrotron radiation circular dichroism spectroscopy, Biomacromolecules 11 (8) (2010) 2136–2142. [26] J.L. Zhai, et al., Conformational changes of alpha-lactalbumin adsorbed at oil–water interfaces: interplay between protein structure and emulsion stability, Langmuir 28 (5) (2012) 2357–2367. [27] J.L. Zhai, et al., Structural rearrangement of beta-lactoglobulin at different oil–water interfaces and its effect on emulsion stability, Langmuir 27 (15) (2011) 9227–9236. [28] R.E. Anderson, V.S. Pande, C.J. Radke, Dynamic lattice Monte Carlo simulation of a model protein at an oil/water interface, J. Chem. Phys. 112 (20) (2000) 9167–9185. [29] T. Koitaya, et al., Low-temperature observation of the softened C H stretching vibrations of cyclohexane on Rh(1 1 1), Phys. Rev. B 80 (19) (2009). [30] A. Matsuda, K.G. Nakamura, K. Kondo, Time-resolved Raman spectroscopy of benzene and cyclohexane under laser-driven shock compression, Phys. Rev. B 65 (17) (2002). [31] H. Matsuura, et al., Experimental evidence for intramolecular blue-shifting C H· · ·O hydrogen bonding by matrix-isolation infrared spectroscopy, J. Am. Chem. Soc. 125 (46) (2003) 13910–13911. [32] H. Antelmann, et al., A proteomic view on genome-based signal peptide predictions, Genome Res. 11 (9) (2001) 1484–1502. [33] S.M. Bromfield, Reduction of ferric compounds by soil bacteria, J. Gen. Microbiol. 11 (1) (1954) 1–6. [34] H. Tjalsma, et al., Proteomics of protein secretion by Bacillus subtilis: separating the “secrets” of the secretome, Microbiol. Mol. Biol. Rev. 68 (2) (2004) 207. [35] L. Day, et al., Conformational changes of globular proteins adsorbed at oilin-water emulsion interfaces examined by Synchrotron Radiation Circular Dichroism, Food Hydrocolloids 34 (1) (2014) 78–87. [36] I. Hirose, et al., Proteome analysis of Bacillus subtilis extracellular proteins: a two-dimensional protein electrophoretic study, Microbiology 146 (2000) 65–75. [37] O.H. Lowry, et al., Protein measurement with the folin phenol reagent, J. Biol. Chem. 193 (1) (1951) 265–275. [38] V.B. Fainerman, E.H. Lucassen-Reynders, R. Miller, Description of the adsorption behaviour of proteins at water/fluid interfaces in the framework of a two-dimensional solution model, Adv. Colloid Interface Sci. 106 (2003) 237–259. [39] S.J. Singer, Note on an equation of state for linear macromolecules in monolayers, J. Chem. Phys. 16 (9) (1948) 872–876. [40] A.F.H. Ward, L. Tordai, Time-dependence of boundary tensions of solutions. 1. The role of diffusion in time-effects, J. Chem. Phys. 14 (7) (1946) 453–461. [41] D.E. Graham, M.C. Phillips, Proteins at liquid interfaces. 2. Adsorptionisotherms, J. Colloid Interface Sci. 70 (3) (1979) 415–426.

P. Patra, P. Somasundaran / Colloids and Surfaces B: Biointerfaces 120 (2014) 132–141 [42] I. Daizadeh, E.S. Medvedev, A.A. Stuchebrukhov, Effect of protein dynamics on biological electron transfer, Proc. Natl. Acad. Sci. U. S. A. 94 (8) (1997) 3703–3708. [43] L.P. DeFlores, A. Tokmakoff, Water penetration into protein secondary structure revealed by hydrogen–deuterium exchange two-dimensional infrared spectroscopy, J. Am. Chem. Soc. 128 (51) (2006) 16520–16521.

141

[44] S.S. Skourtis, D.H. Waldeck, D.N. Beratan, Fluctuations in biological and bioinspired electron-transfer reactions, Annu. Rev. Phys. Chem. 61 (2010) 461–485. [45] M. Yamamoto, et al., Softened CH stretching vibration of a long-chain n-alkane, n-C44H90, physisorbed on a Ag(1 1 1) surface: an infrared reflection absorption spectroscopic study, J. Phys. Chem. B 104 (31) (2000) 7370–7376.